The chapters of this book have their origin in the 2013 Gifford Lectures, a series endowed by Adam Lord Gifford in the ancient Scottish universities in the late nineteenth century ‘to promote and diffuse the study of Natural Theology in the widest sense of the term – in other words, the knowledge of God’. As one might expect, a number of recent Gifford lecturers have been deeply critical of the project of eighteenth- and nineteenth-century natural theology. Rowan Williams, however, seeks to recast the debate by reflecting on a natural human phenomenon: ‘Does the way we talk as human beings tell us anything about God?’ (ix). The poles of revealed and natural knowledge of God ‘both equally sidestep the question of “going on speaking of God” if it is true that speaking of God is characteristically or primitively something quite other than identifying another agent in the universe’ (5). So Williams proposes a different kind of natural theology which does not move from formal features of the world to postulations about God, but examines the theological and philosophical significance of language, given that we do go on speaking about God. A core part of the argument is that speech about God is not, as the logical positivists thought, a disposable aberration lying beyond the meaningful limits of routine language. Our linguistic structures are constantly challenged in the face of the deep actuality of the world because there is always more that needs to be said, whether we are speaking about creatures or God. Williams suggests that a defensible natural theology – one that might include Aquinas's so-called five ways as an exemplary instance – is not so much an argument as a meditative method, ‘a practice of thinking to the edge of what can be said’ (16). Aquinas tests the language of Aristotelian metaphysics and is led to something that cannot belong within the realm of dependent causation because it is the very basis of causation itself. So what frames this speech? It is the language of dependence or contingency that immediately evokes that which is depended on (10).

This is not, however, a book about linguistic theory. By gathering together an eclectic band of interlocutors, Williams wishes to make comments and observations on the habits of language in order to demonstrate how language is both responsive to, and constitutive of, the world. Throughout this subtle and complex work, two extremes are avoided. The first is the postmodern tendency to see language as pure flux and play, having no moorings in the real because there is no extra-linguistic reality. Williams writes, ‘The idea that we are free to give what account of “reality” we choose . . . fails to deal with the actual constraints that speech works with’ (41). The other extreme, perhaps witnessed in scientism or the fantasy that language can be reduced to the merely literal, is the suggestion that what we say is simply dictated by the world. In fact, we labour in our use of language

Williams discusses very suggestively some extreme situations in which the labour of our language becomes particularly arduous; such situations bring to light important aspects of our linguistic habits that might otherwise pass by undetected. In particular, chapter 4 uses the work of the therapist Phoebe Caldwell and her innovative treatment of people with autistic spectrum disorder. The problem with ASD is not an inability to draw ‘normal’ conclusions about the world given a shared set of evidence concerning the world. Rather, those with autism ‘lack the means of absorbing and responding to the environment as part of a shared project’ (113). Caldwell's therapeutic method involves intensive listening to the way the person with ASD talks to him- or herself and entering into that self-enclosed circle of communication by echoing their ‘speech’. Frequently, that communication must be by touch and gesture, but by entering into the self-enclosed sphere of ASD the world becomes a shared project of significance. This reveals that language is relational and is not about the transfer of information from one container to another, but is concerned with ‘establishing a world in common’ (99).

Williams's discussion of ASD is intended to demonstrate not only the shared nature of language and the importance of trust, but also the material nature of language. Speech is an activity of bodies: ‘We select, we check, we coordinate with other perceivers; we are involved in interweaving various material operations in brain and physiology and various relational strategies so as to identify objects that are perceived in common’ (98). Yet language is not a strange aberration within the material universe that belongs only to the realm of conscious human subjects. It is part of a wider pattern of ‘negotiation’ in which systems constantly interact to form ‘a world in common’ or an intelligible universe. So mind and language are aspects of material reality all the way down; language is ‘the natural integrating factor in the evolving material universe. Rather than looking to material processes . . . as the key to understanding what language is, it would be nearer the truth to say that we look to language to show us what matter is’ (102). Williams proposes a broad teleology in which language is an aspect of the material and mental world from top to bottom, reaching a new pitch with human speech that even attempts reference to the infinite. As such, his argument is reminiscent of other recent philosophical interventions, notably Thomas Nagel's Mind in Cosmos, although Williams's resolute refusal of any naturalistic teleology is far more subtle, innovative, and persuasive. This is indeed a complex work of metaphysics and not simply a set of observations and comments on human linguistic practice.

There is so much more to learn from The Edge of Words, but one constant theme that invites comment is Williams's concept of ‘representation’. Language represents the world, but how and with what significance? One straightforward line of thought would suggest that literal speech is the most faithful representation of an object and that metaphorical speech is a mere elaboration. There are many reasons why Williams would regard this as a wholly inadequate account of representation, not least because of the underlying erroneous assumption of a linguistic ‘norm’ – the literal – of which other modes of speech are variants. What Williams has in mind is rather subtler and he makes a crucial distinction between ‘description’ and ‘representation’ (22). By ‘description’ he means the generation of ‘a traceable structural parallel between what we say and what we perceive’. By contrast, ‘representation’ is ‘a way of speaking that may variously be said to seek to embody, translate, make present or re-form what is perceived’. In other words, the representation of language is a crucial aspect of the formation of the world and of our formation by the world. It is a creative participation in the supremely creative divine speech.

The concept of representation is so variegated in the history of philosophy, politics, and theology that Williams adds an appendix to elucidate further his own understanding. Much is owed to a certain interpretation of Hegel's Vorstellung (particularly that of Nicholas Adams), but a crucial aspect of Williams's view seems to be that ‘representation is not something which comes between a reality and its apprehension by a subject – a sort of obstacle to or substitute for the real thing. Rather than seeking to stand in for what's “really” there, a representation is the ‘thereness’ of the object in relation to the subject’ (191). This is an implicit rejection of the notion of knowledge as representation that came to prominence in fourteenth-century scholasticism and that lies at the root of modern scepticism. For Williams, instead of knowing a mere simulacrum of an object rather than the object itself (as in modern representational theories), language makes the object present to us. Indeed, Williams opts for the more traditional Aristotelian-Thomist view of an object's dynamic substantial form acting on the knowing subject (110). This overcomes the dualism between subject and object that plagues modern epistemology:

In drawing together subject and object, linguistic representation is a performance that involves both in creating something new. Subject and object ‘are not two items standing alongside each other needing to be connected by publicly agreed tokens of reference, but two phases in a complex life’ (195).

The immediate theological implications of Williams's comments on language and representation are discussed in the final chapter, ‘Saying the unsayable: where silence happens’. Silence is not simply the absence of speech, for by being silent we are paradoxically saying something. Williams likens this silence to the audience's silence at the end of the performance of a play or piece of music before the applause commences; the length of the pause indicates the significance of the performance and perhaps hints at the intensity and length of the ensuing applause. We might think that the silence that follows the ‘extreme speech’ towards which natural theology drives us is to be filled by revelation; we have nothing more to say so God takes over, leaving behind our speech. Williams suggests that ‘Revelation does not fill the gap, but shows why the gap is there, not resolving difficulty but offering a perspective in which difficulty is what makes sense and what we must become accustomed to’ (180).

What we are left with in The Edge of Words is not the usual product of natural theology, namely a persuasive proof of the existence of an already defined deity. Instead, Williams has used observations about language – a language that is, as it were, naturally supernatural – to undertake a much more important and fundamental task that is so often circumvented by philosophers of religion, namely delineating what on earth it is to speak about God and what that attempt tells us about the God to whom we allegedly refer. As such, he is treading carefully in the footsteps of Thomas Aquinas in not providing a theory of language, but commenting on the use of language. Paradoxically, God is not representable, yet we attempt to speak of God in a way that is meaningful and not arbitrary: ‘we engage in something that is at least analogous to representation’ (148). But in order to do so in a way that avoids idolatry, our language about God will frequently appear eccentric or extreme, involving ‘carefully calculated shocks’. As Williams points out, this is why, for Dionysius and Aquinas, the crudest metaphors for God are often most appropriate because (one hopes) they could not be mistaken for ‘accurate’ or ‘literal’ depictions. For now, the open-ended, performative, and communal aspects of language drive us to the edge and imply an infinite excess of meaning lying beyond those words. The source of that meaning is the Word, the eternal speech of God. In so far as The Edge of Words is a prolegomenon, we await the further elucidation of the theological implications of this profound meditation on the performative metaphysics of language.