Although there are a number of problems with this argument, the biggest and most serious is the simple fact that it misrepresents religion so badly. It may be that some very liberal believers' religions are only about values and morals and make no empirical "what" claims about the world around us, but that doesn't describe most religions in the world or religion through most of human history. Apologists are essentially making an argument based on what they'd like religion to be, not on what religion really is.
As pleasant as this orientation may be, I do not know how tenable it is. Incommensurability does not fit with religion as we know it. Can the world’s religions withdraw from all factual claims (other than the existence of God, perhaps)? Many years ago, in an Introduction to the Philosophy of Religion class, a Christian student rejected the incommensurability thesis, remarking that ‘‘if they find the body, the game is up.’’ Can Christians maintain their faith without defending the historical reality of Christ walking on Earth?
In short, I do not know how well religion can stay on the ‘‘ought’’ side of the is–ought dichotomy (assuming that it is a dichotomy).
Source: "Faith and Reason, the Perpetual War: Ruminations of a Fool," by Stewart Shapiro in Philosophers Without Gods
When a religion makes empirical claims, it renders those claims subject to proof or disproof. The only way for the "faith" of a religion might be fairly exempted from the critical scrutiny of science is for it to refrain from making claims that are directly subject to empirical testing or logical analysis. Just how likely is that, however? Not very, at least as far as I can tell.
Religionists claim not only to have a unique access to the grounds of truth (i.e., the ultimate reality of their god), but also to particular truths established by those grounds. This position is incompatible with science and it is incompatible with the principles necessary for an open society. It would be a nice change if believers didn't attempt to latch on to the latest finding of biology of physics in an attempt to "prove" the truth of what they typically insist requires "faith." A little consistency here isn't too much to ask for.
This above misrepresentation of religion is joined with an equally serious and unjustifiable misrepresentation of science. Since when has science — the empirical, rational investigation of our universe — been confined to nothing but "what" questions? Science can contribute a great deal to answering questions dealing with morality and value, if for no other reason than rational answers to such questions must incorporate facts and reasoned arguments which science provides.
Rationality, and certainly philosophy, are not—and should not be—content to stay on the ‘‘is’’ side. The incommensurability thesis is that religion tells us how to live. It instructs us on how to treat each other, how to treat the planet, and so on. And, as noted above, there are indeed deep insights along those lines in the holy texts. However, it seems to me that moral matters are susceptible to rational appraisal. That is, rationality has something to say about the best way to live and does not abandon that arena to the world’s religions.
This is not to insist on moral realism, the thesis that moral discourse is objective. Moral non-cognitivists from Hume to Simon Blackburn insist that moral assertions, such as the wrongness of killing innocent children, do not express matters of fact, or have truth conditions. But it does not follow from such views that moral discourse lies outside the norms of rational appraisal altogether. Hume admits as much, when he discusses the senses in which we hold that a given feeling or action is rational or irrational.
That moral claims are subject to rational evaluation, scrutiny, and appraisal means that science can have something to say about their subject matter. Insofar as a moral claim makes any factual claim about the world or assumes as true any factual claim about the world, science can have some relevant input to offer. The science may come from chemistry and biology or it may come from psychology and sociology; either way, though, it would absurd to insist that scientific research or methods are irrelevant. We don't need to adopt a form of "moral realism" to insist that both our moral reasoning as well as our moral conclusions be firmly grounded in reality.
Science is the result of critical methods. Religion is a matter of fulfilling wishes.
“The only way for the “faith” of a religion might be fairly exempted from the critical scrutiny of science is for it to refrain from making claims that are directly subject to empirical testing or logical analysis.”
The problem about a claim that is not subject to empirical testing or logical analysis is that this may forever remove this claim from the realm of science. This has historically been a mistake since the realm of science grows continuously. For example, people used to think morals and conscious were outside the reach of science, but neuroscience and Darwinian psychology have brought them both into the realm of testable logical analysis.