When the Nazi soldier knocks at the door and asks if there are Jews hiding in the house, and there are, is it wrong to reply, ‘No’? It would be hard to find a philosopher or theologian, let alone a member of the general public, who would argue unequivocally that it is. Paul J. Griffiths, however, is bold enough to defend St Augustine's position that every lie is a sin and therefore all lies should be rejected. In order to explain this, he provides an imaginative reading of Augustine's theology, emphasising the creation of humanity in the image of God. He goes on to examine specific arguments about lying from nine other major thinkers, Christian and non-Christian, ranging from Plato to Jerome to Newman, and to criticise each of these from the perspective of Augustine's theory. Few readers, I suspect, will be persuaded to accept Griffiths's (or Augustine's) conclusion, but most will find his arguments thought-provoking and at times unsettling. After all, it does not seem silly to say that if one were able, one would like to persuade even the Nazi to leave without telling him a direct lie. But why would that thought even make sense if there were absolutely nothing wrong with lying to those even who have no right to be told the truth?

Griffiths is interested not in dialectical arguments about the merits of lying but rather in the wider theological and anthropological assumptions that might make sense of a complete ban on lies. Consequently, his account of Augustine is based on a refreshingly wide reading of texts, with De Trinitate playing an unexpectedly important role. In his interpretation of De Mendacio (hereafter DM), he identifies the key problem with lying as duplicity, that is, the saying of something that does not seem to oneself to be true; this drives a wedge, as it were, between the heart and the mouth. Because it begins with the heart, lying is seen as an interior matter; whether or not, and for whatever motive, it may communicate false beliefs to others is, on Griffiths's reading of Augustine, secondary. Lying is also a matter of decision: whereas a truthful expression flows naturally and freely from heart to mouth, a lie takes effort, as one consciously chooses to misrepresent one's thoughts.

The context of our speech is a world created by God, hierarchically ordered in terms of the goodness of different things, each of them utterly dependent on God. Insofar as anything is separated from God, it tends to non-being; therefore, Augustine can pun, what is ‘private’ is also ‘deprived’. What is true, he often remarks, is given by God and a common possession; if an apparent truth is ‘my own’, it must be a lie. Similarly, Augustine can say, ‘all sin is a lie’, in the sense that it involves the will in turning from its true and shared good to a something private that is mistakenly seen to be good. Ultimately for Augustine, Griffiths argues, the only honest form of speech is confession, the adoration of God's absolute goodness and the recognition of one's own absolute worthlessness apart from God. He ties this in with Augustine's theology of trinity and incarnation: just as the incarnate Son clothes the eternal Word with flesh to proved a true representation of the Father, so our voiced words should clothe our inner words with flesh to provide a true representation of our minds.

On the basis of this ambitious reading of Augustine, Griffiths turns to criticise the arguments of his other selected authors. The second part of the book is less satisfying than the first, partly because Griffiths deliberately chooses to analyse short individual texts rather than the wider thought of each author as a whole, and partly because his ‘Augustinian’ readings are narrow and sometimes uncharitable. That is, his interest lies in where these texts fail to match up to Augustine's doctrine rather than in anything of independent value that they might offer, and he tends to read them in as polarised a way as he can. The result is a history that puts a curious encomium by John Chrysostom of a specific lie told ‘for a good purpose’, along with Nietzsche's radical reinterpretation of all speech, on the same side of the fence as Aquinas, ‘against Augustine’. Augustine himself, one feels, would be very surprised to discover that no insights of value had appeared in the 1600 subsequent years of thought on the topic.

Part II of the book confirms the suspicion that Griffiths's reading of Augustine is more Augustinian than the master himself. DM is a hesitant, nuanced, complex and tangled text, and in general Augustine's thought on specific moral topics cannot easily be streamlined or pigeon-holed. The neatness and coherence of Griffiths's interpretation is his own. For example, he criticises Aquinas as un-Augustinian in (a) arguing that the intent to deceive, though part of the ‘fullness’ (perfectio) of a lie, is not essential to its definition, and (b) holding that the degree of sinfulness of a lie can depend on such factors as its topic or motive. On (a), Aquinas clearly follows Augustine in holding that a lie in itself, irrespective of intention, is something bad (‘malum ex genere’, Summa Theologiae 2a2ae 11.3, reply), because ‘words are by nature signs of thought’. For this reason he shares with him the odd conclusion (see Enarratio in Psalmos V.6, hereafter EP) that expressing untruths as a joke even where there is no intent to deceive anyone is a sin, but a minor one (non magna culpa, as Augustine puts it). On (b) Augustine himself clearly holds that the subject-matter of a lie (e.g. religion) can make a difference to its degree of sinfulness, as can one's motives in telling it (see e.g. DM XXI.42). What seems to upset Griffiths is that St Thomas argues that not all lies are mortally sinful; most of us would be grateful for that conclusion, and Augustine himself, as EP V.6 show, holds a similar position.

Again, Griffiths argues that John Cassian's assimilation of promise-breaking to lying is odd and un-Augustinian. In fact, Augustine often makes the same move (once again, EP V.6 is helpful: breaking a promise to return a sword is the example discussed here). Griffiths also criticises Kant on the grounds that he may be limiting in his ban on lies to certain types of statement (Deklarationen and Erklärungen: the evidence for the interpretative claim is not provided). He misses the point that for Augustine too the difference is significant between types of statements that arouse different expectations in the hearer. His distinction in DM (often unnoticed, it is true, by commentators and translators) between type 6, lies told (for a good purpose) as testimony, and type 7, lies told (for a good purpose) not as testimony, hinges on this. Indeed, at DM XVII.36, Augustine explores the meaning of testimonium in a way that makes clear the importance for him of the fact that the purpose of speech is to communicate: when one speaks to God, he says, one embraces the truth only in one's heart; but when one speaks to a human being, one must also express the truth with one's physical mouth, because human beings cannot inspect the heart.

Griffiths's tendency to smooth out the wrinkles in Augustine's thought reveal the fundamental importance to his own position of the claim that communication is not essential to the purpose of speech. This claim seems to me both implausible in its own right and unconvincing as an interpretation of Augustine, who comes to seem far more individualistic and isolated a thinker and a person than he actually was. Indeed, for him our opacity to one another was one of the most disturbing features of our fallen condition. Nor does Griffiths's claim follow from his very interesting exploration of trinity and incarnation: it is a natural thought, and Augustine often suggests something similar, that the purpose of the incarnation was, precisely, that God should be able to communicate with us. Just as the incarnate Christ is the exterior, communicative expression of God's inner being, so our vocalised words are the exterior, communicative expression of our minds.

Such an interpretation would not reduce the gravity of duplicity, but rather explain that it is a grave matter precisely because we are made to communicate with God and one another, in truthfulness and love. If we took that with full seriousness, how would it change our lives? Griffiths concludes with a brief prophetic chapter imagining how politics might be transformed: for both capitalism and democracy, as least as we know them, seem to depend for their existence upon systematic dishonesty and deception. The truly ‘Augustinian’ community would ‘relinquish ownership of speech’ and concentrate on silence and on praise. It would not be tempted by the ‘consequentialist’ attractions of (to use Griffiths's own example) a navigator's misleading the pilot of a plane that is about to drop a bomb on a million innocent people.

Idealism of this sort about political life does not sound much like Augustine, who accepted flawed political structures as a grim necessity, and thought that public officials are regularly required to act in a way that would be wicked for a private person. On the other hand, one does not have to be a consequentialist to allow the navigator his lie; all one needs is a decent theory to explain the circumstances in which lies can be justified. Augustine argued that the goods of the body are of no significance compared to those of the soul, and therefore one ought not to lie even to save another from serious bodily harm. But those who argue that lying to a potential murderer is justified think that such a lie would not in fact harm the soul of the one who tells it.

Griffiths's story is highly selective in its ‘Augustinianism’, and will not convince many readers that all lies are gravely wrong. At the same time, there is something hauntingly attractive in this ideal of absolute truthfulness. It does seem worthwhile, and even admirable, to try to avoid lying even to those who seem to have forfeited the right to be told the truth. We highly value those people whom we know we can ‘take at their word’. It is perhaps significant that English is not unusual in having no morally neutral single word for ‘tell an untruth’: ‘lie’ functions more like ‘murder’, which implies blame, than ‘kill’, which does not. Our own cultural ideal of good living (this is not universal) does not encompass lying; most Christians would feel somewhat uneasy at pondering the idea of Christ himself, for example, telling a direct lie, whatever the circumstances. It seems appropriate for those whose task is to bear witness to the truth, and above all for the one who himself was the truth, to avoid misrepresenting even trivial truths. Again, Griffiths's prophetic question is disturbing: what would the world look like - how much better would it be? - if we simply refused to collaborate in the structures of systematic lying?

When the subject of lies is discussed, most people move all too quickly to cases like the Nazi at the door, without pausing to reflect upon the vast range of less extreme circumstances in which lies are taken for granted, or given prettier names. This book challenges us to examine our collective conscience more thoroughly with reference to the ordinary business of living, and not merely to exceptional crises.