One of the problems of an unbalanced dependence on rationality is dissonance. Eventually, specific self justifying systems become more and more removed from actual events. In the start of the chapter "the rational courtesan", John Ralston Saul is highly critical concerning the difference between Robert McNamara's intended outcomes for the arms sales, world bank loans and the eventual outcomes that were diametrically opposite. While one certainly can say that these cases are much greyer than Saul presents, it seems apparent that idealistic projections mirror rational projections. The validty of this mirroring is probably based on the extent to which each fails to deal with human complexities, irrationality and chaotic meddling.
Most people naturally assume that religion is the ultimate self justifying system. Hence many people claim that dissonance is the end result of any sustained belief. In this case it is ironic to see rationalism suffer from the same criticism. Succeptibility to dissonance seems to depend on the degree to which seemingly insignificant facts are omitted. In terms of a scientific critique of religion, one would have to say that religion ommits key facts that, while they appear to be insignificant, are likely become anything but. Similarly a critique of the absolute nature of science would only be valid if science omitted some seemingly insignificant but ultimately chaotic facts. So the real question between science and religion shouldn't really be about who has the right facts at present (certainly things can change and grow), but who can incorportate the necessary facts. As Saul mentions, rationalism has left out chaotic facts about people.
Right now one really would have to conclude science is at the same stage. While it certainly can make some pretty good predictions involving people, it certainly isn't infallible or all reaching. Of course neither is religion. Realistically most people would conclude that science has no need to go down that road. If it doesn't any possible claims involving human interactions must suffer the fate of rationalism; they can be logically consistent, but ultimately only self justifying. Thus science may be able to discover an unchanging world, but it is not able to explain our relation to it - well at least not at any level that can effectively deal with human interaction. It certainly works well in dealing with absolutes, just as it works well when our chaotic nature is removed from the equation. It may leave one wondering if our chaotic nature is in fact real? If it can't be dealt with I don't know how science could say it is anything other that irrational lunacy. While this may certainly be true, an alternative is to treat our chaotic influence as if it was real. (Note I certainly don't mean that the chaotic aspects of human influence can't be noticed, only that they don't correspond to any sort of physical reality (communally accepted and domain independent))
Of course how does religion handle things? I don't see any reason why religion can't assume any of the facts that science does. Certainly it has dogmatic assumptions about the supernatural. But it doesn't seem to have to exclude any claims of science.