But experience of the limitations of descriptive or would-be fact-stating language pushes us to see that there are other ways of "representing" the truth of the sometimes "extreme" situations in which we may find and recognise ourselves. Rowan Williams is not only a former Archbishop of Canterbury, he is also an accomplished multi-linguist, a distinguished academic theologian and a poet, with a strong sense of the difference between poems that "ring true" and those that may not; and all this comes through very strongly in this book.

What exactly does all this add up to? Well, for one thing, that the word "exactly" is here as inappropriate as was my earlier use of the word "just" — and if this seems to present some sort of paradox of reflexivity, it is in the context a paradoxically appropriate one. No doubt that all engagement in language has to be seen as a form of what Chapter Four presents as "material practice", that is to say as consisting in its enactments of a series of macro spatio-temporal events, and as such subject to the kind of explanation structured in terms of (preceding) cause and (subsequent) effect that must in principle be presumed to govern all such occurrences. At the same time, to take part in meaning-orientated activity, to think and/or to act linguistically, is to engage in a rule-governed activity, only properly understandable as a goal- or future-directed effort to meet standards of evaluation by which its success or failure in the transmission or deciphering of meanings may be judged. Attempts to show how one might hold together these two schemata of explanation within one and the same framework of humanly rational understanding have lain, and still lie, at the heart of many of the great philosophical enterprises and controversies; and it may indeed be that the best that one can hope for is a rational understanding of why that cannot in the end be done.

Where this would leave us poses, of course, an immense question — to which the only certain answer may be that it leaves us with both the inescapability and the ultimate unanswerability of the question itself. As Rowan Williams puts it towards the end of his concluding chapter, "There are versions of human self-description which in effect make it impossible to understand at all what is going on in the language we actually use — and thus make it, if not impossible, at least unintelligible that we ourselves should speak. And this," he quite rightly comments, "is not an easily sustainable position." And for him "this means that the most comprehensive . . . account we can give of what is recognisably human is deeply implicated in concerns about ‘the sacred' — about what is not yet said, what is not sayable . . . Such an account does not deliver a ‘proof of God's existence' [but it does] enable us to see that what is affirmed in the language of specific religious ritual . . . goes with the grain of what matters most . . . in anything claiming to be an adequate picture of our human speaking."

One of the many striking things about this whole mode of argument is how ecumenical it is in both implication and spirit. Ronald Dworkin's last book was entitled Religion Without God, and he was and is very far from being alone among those unable to attach any clear sense to talk about God, but in finding themselves nevertheless impelled to ask Y a-t-il du sacré dans la nature? (the title of a book, edited by Bérengère Hurand and Catherine Larrère, containing the papers from a conference held in Paris in 2012). It would seem surely to follow from such a line of argument that there can be little clear reason to attach any cognitively (or "factually") distinctive importance to the language and doctrines of any one given religious-cum-theological tradition rather than to that or those of any other.