D. Z. Phillips died quite unexpectedly on 25 July last, in Swansea, while searching in the college library for a quotation that he could not track down in his books at home.

Born at Sketty in West Glamorgan on 24 November 1934 Dewi Zephaniah Phillips studied philosophy at Swansea from 1952 to 1958, then at Oxford from 1958 to 1961. His dissertation, begun under the guidance of the Anglican philosopher Michael Foster, turned into his first book, The Concept of Prayer (1965), indebted in the event chiefly to discussion with Rush Rhees. It is, as he says, heavily coloured by his reading of Søren Kierkegaard and Simone Weil. By 1965, after brief stints at Dundee and Bangor, he was back at Swansea, where he taught philosophy until 1996 when he retired. He continued to hold the Danforth Chair in Philosophy of Religion at Claremont Graduate University, California, spending some months there every year and hosting an annual conference on Wittgenstein and religion.

D.Z. Phillips had no formal training in Christian theology. From 1959 until 1961, he was Minister of Fabian Bay Congregational Church, Swansea. He was allowed to become a minister on a probationary basis: he had been preaching to large congregations since his teens, in Welsh, of course, his mother tongue. Through Rhees, one of Ludwig Wittgenstein's closest friends, Phillips inherited a certain way of reading the later Wittgenstein's work. He became the best known of the ‘Swansea Wittgensteinians’: philosophy understood as a kind of intellectual therapy in dark times rather than constructive theorizing; against prevalent aspirations and practice in the discipline, philosophical work as reminding ourselves of things we may overlook but cannot deny, rather than adding to the sum of knowledge by quasi-scientific discoveries.

The name of D.Z. Phillips is associated, in standard philosophy of religion courses, with ‘Wittgensteinian fideism’. This label was invented by Kai Nielsen, a distinguished Canadian philosopher, who wanted to make the case for atheism. He resisted the thesis that religious concepts can be understood only if we have a participant's familiarity with the ‘language game’ (Wittgenstein's term) in which they occur – a thesis, then, supposedly excluding in advance the atheist's would-be objections to the whole enterprise.

Funnily enough, in the original article (in the journal Philosophy 1967), Phillips's name is never mentioned: ‘Wittgensteinian fideism’, in its inventor's mind, was based on ‘certain remarks’ that Nielsen culled from the writings of Peter Winch, G.E. Hughes, Norman Malcolm, Peter T. Geach, Stanley Cavell, J.M. Cameron and Robert C. Coburn – a heterogeneous list, and quite bizarre for those acquainted with the names.

The claim, anyway, is supposedly that argument about the rationality of religion cannot get going because outsiders have no idea what they are talking about: religion is thus ‘invulnerable’ to criticism

Obviously, there is something in the claim. Critiques of religion, and especially refutations of arguments for the existence of God, often focus on accounts of religion, and involve a concept of God, which no religious person recognizes. (Consider the current best seller by Richard Dawkins, The God Delusion, and the lengthy, highly critical review in London Review of Books, 19 October 2006 by Terry Eagleton, citing Herbert McCabe at a key point.) Yet, for Catholic apologists at any rate, allowing that faith is a gift, there remains a good deal that believers, agnostics and atheists should in principle be able to discuss. As for D.Z. Phillips's approach, Rowan Williams made the decisive comment years ago: it ‘does not strike me as cosy, “invulnerabilist” neo-orthodoxy … it does seem to me that he is more concerned than most philosophers of religion to locate the question of faith and unfaith where it belongs – not indeed in an irrationalist enclave, but in the context of questions like, “Is there a real self?”, “Is there a human meaning?”, “Does it make sense to ask if things make sense?” And the securities and vulnerabilities alike of faith are those bound up with the unpredictable and often horrible contingencies of countless particular human stories’ (see Theology May 1980: 205–7).

Unlike most philosophers of religion, Phillips could change his mind and even charge his younger self with saying things that he later regarded as ‘cursory and misleading’. Religion without Explanation (Blackwell 1976) is a good book, one of the few books by a philosopher that discusses issues concerning religion in a way interesting to students in the social sciences as well as in the humanities – which does not mean that they would always agree with his criticisms, of Frazer, Freud, Feuerbach, Durkheim and others. Wanting to revise the book he ended by writing a much better one, Religion and the Hermeneutics of Contemplation (Cambridge University Press 2001). Here, besides deeper critiques of the usual suspects, Phillips took up Paul Ricoeur's distinction between the hermeneutics of suspicion and that of recollection. In matters of religion this is the distinction between those who explain away religion as illusion and fantasy (Marx, Nietzsche, Freud, and, as we may now say, Dawkins and Dan Dennett) and those who shore up religious claims with appeals to experiential evidence, phenomenology, rational justification, and suchlike. Phillips, following the later Wittgenstein as he contends, saw a third way: the hermeneutics of contemplation, as he calls it. This would be a consideration of religious claims and practices which strives to be neither reductive nor apologetical but simply to do justice to the concepts in their complexity, ‘raggedness’ and irreducible difference. ‘Contemplating the world without meddling in it’: the inspiration of this hermeneutics of contemplation ‘comes from wonder at the world in all its variety, and the constant struggle to give a just account of it’. While there will be no more books from him, this conception of philosophical work remains a challenge.