Brad DeLong cites some critics of Douthat on social policy, including on abortion, and (via Matt Y.) there is more here, from his college writings. This post, on eugenics, is one of Ross's most controversial (btw overall I like it). I am myself more libertarian than conservative but at the same time I am on Douthat's side in questioning the common presuppositions behind modern opinion. There is a presumption that liberal, tolerant people should have certain views on abortion, stem cell research, and other matters and I am happy to see Douthat breaking the mold. On these issues, the derivation of current liberal policy views from underlying liberal principles is in fact extremely tenuous, even if one views those conclusions as ultimately correct. I view the current alignment of stances on social policy as more of a sociological regularity ("look at how rotten are the people on the other side") then an intellectual necessity.
Take abortion. Let's say that the mainstream modern liberal understanding of when life begins is correct with p = 0.92. That's a pretty high p on a matter where so many intelligent people disagree so vehemently. Does such a "p" provide enough reason to follow through with modern liberal policy conclusions? That's far from obvious. In this debate you'll find lots of fury and very little willingness to apply stochastic reasoning to ethics. There are far too many smart people who offer lip service to the toughness of these questions and then simply go ahead and take sides.