The Cambridge University Press first suggested to David Lewis in 1987 that it might publish a book of his. He gladly agreed to the publication of a volume of his selected papers, and started to think, and to consult friends, about what ought to be included in it, but he did not come to a final decision on the contents. Shortly before his death, in 1994, he asked me as his literary executor to take the final decision (which I have done after further consultation) and to edit the book.

Chapters 5, 21, 25 and 26 are published here for the first time. They are the papers from amongst his unpublished works which Lewis wanted to be included in this volume. He did some work on them in the last year of his life, but he left it to me to do the final editing, and to write the footnotes (though he gave some indication of what he wanted, and in some of the notes which provide more than bare references I have been able to use his words). Copies of some other unpublished papers are being deposited in selected libraries: see p. 411.

The remaining chapters have been published before, and I am grateful to the editors and publishers concerned for permission to republish them here. I have standardised the form and have done some updating of the references; I have supplied a first footnote (cued by an asterisk rather than by a number) for each chapter, including the details of its first publication; my more substantial interventions elsewhere are enclosed in square brackets.

These three chapters are a version of lectures delivered at Cambridge on 22, 23 and 24 November 1993. The occasion was organised at Clare Hall by Dr Janet Huskinson with unfailing thoughtfulness. It was rendered gracious by the hospitality and by the participation throughout of the President, Sir Anthony Low. The panel of discussants chaired by Keith Hopkins – Peter Garnsey, Robin Lane Fox, Christopher Kelly and Rosamond McKitterick – have not only left me with food for thought for many years to come: they provided us all with a model, for our times, of commentary and disagreement that were as lively as they were courteous. The presence in the audience of so many friends and colleagues – Henry Chadwick, Ian Wood, Robert Markus, William Frend, Andrew Palmer, to mention only a few – guaranteed that the discussion ranged vigorously throughout the entire late Roman and early medieval period. Altogether, I present these chapters with a touch of sadness: they are, simply, the lees of the wine – what survives in print of an unusually vivid and humane occasion.

A shorter version of the first chapter had been delivered, in the previous year, as a Raleigh lecture of the British Academy (Peter Brown, ‘The Problem of Christianisation’, Proceedings of the British Academy: 1992 Lectures and Memoirs 82 (1993), pp. 89–106). The themes of that chapter, and of the two subsequent chapters, emerged in large part as a result of my work for sections of volumes XII and XIV of the Cambridge Ancient History.

Faced by a topic as labyrinthine as the problem of Christianisation, it is a relief to begin with a person for whom the problem apparently caused little trouble. Some time in fourthcentury Britain, Annianus, son of Matutina, had a purse of six silver pieces stolen from him. He placed a leaden curse in the sacred spring of Sulis Minerva at Bath, in order to bring the miscreant to the attention of the goddess. On this tablet, the traditional list of antithetical categories, that would constitute an exhaustive description of all possible suspects – ‘whether man or woman, boy or girl, slave or free’ – begins with a new antithesis: seu gentilis seu christianus quaecumque, ‘whether a gentile or a Christian, whomsoever’. As Roger Tomlin, the alert editor of the tablets, has observed: ‘it is tempting to think that a novel gentilis/christianus pair was added as a tribute to the universal power of Sulis’. Christianisation, at the shrine of Sulis Minerva at Bath, means knowledge of yet another world-wide category of persons whose deeds were open to the eye of an effective goddess of the post-Constantinian age.

Annianus, and many other fourth-century persons, lived in a universe rustling with the presence of many divine beings. In that universe, Christians, even the power of Christ and of his servants, the martyrs, had come to stay. But they appear in a perspective to which our modern eyes take some time to adjust – they are set in an ancient, pre-Christian spiritual landscape.

Some time in the 520s, the great old man Barsanuphius, an Egyptian recluse, wrote from his cell in the vicinity of Gaza, in order to comfort a sick and dispirited monk:

Throughout the Christian world of the fifth and sixth centuries, average Christian believers (like the sick monk, Andrew) were encouraged to draw comfort from the expectation that, somewhere, in their own times, even maybe in their own region, and so directly accessible to their own distress, a chosen few of their fellows (who might be women quite as much as men) had achieved, usually through prolonged ascetic labour, an exceptional degree of closeness to God. God loved them as His favoured children. He would answer their prayers on behalf of the majority of believers, whose own sins kept them at a distance from Him.

I have been prompted to re-think the problem of intolerance in the later empire by the experience of my nephew when, at the age of three, he paid his first visit to the Cincinnati Zoo. He already knew what all the animals looked like. He had seen pictures of elephants, giraffes, lions and walruses. But the only living animal that he had actually seen was his own pet cat, Teddy. Naturally, he assumed, therefore, that all animals were the same size as that cat – and, consequently, that they were smaller than himself: a flattering assumption. The fact that the animals in the zoo were all of very different sizes – some, indeed, immeasurably larger than himself – came as a great surprise to him. On seeing two Bengal tigers eating their lunch, he exclaimed: ‘Big, BIG cats!’, and remained silent, adtonitus, for a full twenty minutes, until a visit to the bird-house restored his overshadowed ego.

Religious intolerance is a phenomenon that bulks large in any history of late antiquity, to such an extent that it colours our perception of the entire quality of the age. The historian who wishes to form a just estimate of its nature, extent and effects has to face the same puzzlement as did this young fellow. We are dealing with a phenomenon that is documented in vivid, and seemingly unambiguous, fragments of evidence. The problem is, how to assess the relative size of the phenomena to which the fragments allude.

The historian's task is to narrate, but he must also win credibility for that narrative: his task is therefore also to persuade his audience that he is the proper person to tell the story and, moreover, that his account is one that should be believed. In his capacity as persuader, the historian will often try to shape the audience's perception of his character and to use this as an additional claim to authority; indeed, among the Roman historians, where explicit professions of research are rarer than with the Greeks, the shaping of the narrator's character takes on a correspondingly larger role. But most of the historians, Greek and Roman, try to shape their audience's perception of their character. Nor is this surprising when we consider the teachings of rhetoric.

THE SCOPE AND SUBJECT OF THE BOOK

This book is a study of the explicit attempts by which the ancient Greek and Roman historians claim the authority to narrate the deeds encompassed in their works. The term ‘authority’ has many meanings over a range of disciplines, but in this book it is used to refer to literary authority, the rhetorical means by which the ancient historian claims the competence to narrate and explain the past, and simultaneously constructs a persona that the audience will find persuasive and believable. The work is thus a study of certain forms and conventions of persuasion employed by the historians. No attempt is made to evaluate the truth or falsity of historians' claims; rather, I try to set out the various claims which are part of the construction of the author's historiographical persona; to see how and why these claims are made; to explain how the tradition of such claims developed; and to show how the tradition moulded the way in which writers claimed historiographical authority.

The writers treated range from Herodotus in the fifth century BC to Ammianus Marcellinus in the fourth century AD. Included in this study are both the surviving (either whole or in part) historians and those whose works have come down to us only in fragments.

In an aside to his audience after narrating the revolt of the Theruingi and the slaughter of the Roman army under Lupicinus in AD 376, the ‘lonely’ historian Ammianus Marcellinus asks the indulgence of his readers on a particularly difficult matter:

The rather poignant parenthesis is consistent with the view that Ammianus presents elsewhere in his history of a public at Rome concerned only with the trivial biographies of emperors and caring more for the details of the private lives of the imperial household than with the grand sweep of res gestae. The last antique historian is indeed a great one, and he may even have been as isolated as is sometimes suggested. But in a larger sense, nearly every ancient historian seeks to portray himself as a lonely seeker of truth, as the only one who has somehow understood the historian's proper task, while his predecessors (as he will frequently remind us) failed in the effort, either because they were ignorant and ill-intentioned from the start, or, though men with the proper attitude and application, they did not yet embody all the qualities necessary for a good historian, or live in a time that was suited to their abilities, or assay a task as grand as the one now placed before the reader.

Up to this point our concern has been how the ancient historian justifies himself before his audience and attempts to portray himself as the proper person for the writing of history, that is, with his role as narrator rerum. The present chapter examines how he approaches his task when a participant in the very deeds he records, and how he reconciles the dual role of actor and auctor rerum. For in fact many historians of the ancient world had the opportunity to be both participant and rememberer. The historian's formal method of presenting himself has received comparatively little attention, yet it is of interest not only because it tells us something of the way that men who wrote history in the ancient world approached the writing of their own deeds, but also what their concerns were in doing so. It is usually assumed that in order to give authority to his account, an historian who narrated his own deeds used the third person and maintained a show of formal impartiality. But a study of the surviving (and partially surviving) historians reveals a variety of approaches and methods, changing with time, the specific type of history written, and the individual intention of the historian himself.

Imitation of predecessors may occur in many ways; I discuss briefly here one of those ways, the manner of an historian in the formal introduction of himself to the audience.

Hesiod, of course, is the first poet to name himself in his work. In a similar way Theognis, although beginning with one (or several) prefaces to the gods, inserted his name into the text as what he calls a ‘seal’ (σφραγίς). The nature of this seal has been much discussed, some seeing it as a mark of poetic pride, others as an attempt to guarantee the authentic character of the text itself. Both in fact may be present. Of the other lyric, elegiac, and iambic poets we can single out Alcman, Phocylides, Demodocus, Solon, and Susarion as those who name themselves in their poems, yet we must also recognise that such a procedure need imply neither Dichterstolz nor the presence of the author as a character in the text itself.

Both pride and presence in the text can first be unmistakably seen in the philosophers, who as a group exert the strongest influence on the early historians. Here names are used with evident pride and as an emphasis on the originality of the work.

An examination of how historians choose from among variant versions sheds some light on the methodology assumed in writing non-contemporary history. As we noted at the outset, the historian differs from the Homeric narrator by his lack of omniscience. Unlike the poet, the historian does not know all, nor does he pretend on each occasion to be able to explain motivation or cause. We may term this entire phenomenon ‘narrative uncertainty’, and it can be as brief as a few words (‘either willingly or unwillingly’), or as lengthy as a full digression and refutation. In the latter cases, one can sometimes find methodological pronouncements. The material will repay a full study, and I present here only an overview of those occasions when the historian explicitly cites two (or more) versions and ascribes them either to unnamed authorities (‘some say’, ‘others believe’, ‘there are some who write’) or to named sources whom he is (presumably) following in part or in toto, and then chooses one over the other. This type gives insight into an historian's methodology. In those cases where an historian does more than merely record the variants, we may note a few recurring ways of choosing or preferring one version over another. For each of these I have given a only few examples, but they are nevertheless typical.