The Annals of Tacitus [1] 0521085845, 9780521085847

Citation preview

CAMBRIDGE

CLASSICAL TEXTS COMMENTARIES

AND

EDITORS

C. 0. BRINK

D. W. LUCAS

F. H. SANDBACH

15

THE

ANNALS TACITUS VOLUME

I

OF

ROSALEAE VXORI CARISSIMAE

THE ANNALS

OF

TACITUS BOOKS 1-6

EDITED

WITH

A COMMENTARY BY

F. R. D. GOODYEAR Hildred CarlileProfessorof LAtin in the Universityof LonMn

VOLUME

I: ANNALS

CAMBRIDGE AT THE UNIVERSITY 1972

I. 1-54

PRESS

Published by the Syndics of the Cambridge University Press Bentley House, 200 Euston Road, London NWI 2DB American Branch: 32 East 57th Street, New York, N.v.10022 © Cambridge University Press 1972 Library of Congress Catalogue Card Number: 72-75300 ISBN:

0 521 08584 5

Printed in Great Britain at the University Printing House, Cambridge (Brooke Crutchley, University Printer)

CONTENTS

..

Preface

page vn

Map: The German frontier

X

INTRODUCTION 1

n 111 IV

TEXT

Codex Laurentianus Mediceus 68. Editions of the Annals On editing Tacitus Aspects of Tacitean historiography AND CRITICAL

1

APPARATUS

3 5 19

25

49 ~

COMMENTARY APPENDIXES Adnotationis criticae additamenta 2 On certain terminology used to describe the language of Tacitus 3 Some remarks on alliteration in Tacitus 4 Tacitus' variation and paraphrase of standard terms 5 Certain aspects of ellipse in Tacitus 1

LIST

OF WORKS

CITED

INDEXES 1

2

3

Lexical Passages discussed General

V

331 334 336 342 346 351

PREFACE This edition with commentary of Tacitus, Annals 1. 1-52, the first of a series of four volumes which I hope to complete within about ten years, is intended to be reasonably full and to give attention equally to text, style, and subject-matter. So vast, however, is the number of modern writings on Tacitus, and so extensive the ramification of problems arising from his works, that a truly comprehensive edition and commentary would be almost impossible to compile and quite intolerable to read: selection of some kind is necessary and desirable. And so I have omitted much which might have been included, often no doubt from mere ignorance, but often too from choice. Sometimes important general questions have been discussed only briefly in this volume, because they will be treated more fully later. The appropriate scale of treatment for historical matters has not been easy to determine. In general I have written at length only of those historical questions which are closely associated with the interpretation of Tacitus' text. Being no historian, I still fear reproach for trespassing even thus far upon alien territory. E. Koestermann's commentary on the Annals has appeared while this book was in preparation. There is much in Koestermann's work which I admire. If he had dealt as successfully with text and language as with historiography, I should perhaps have looked for another field of research. And yet there is in English no full-scale commentary on the Annals since Furneaux's, which is now some eighty years out of date and, for all its very real merits, rather limited in scope. After much irresolution I decided to make my critical apparatus extremely brief and selective. This seemed justifiable in view of the very generous treatment of textual problems in the commentary. For those who want to know

..

vu

PREFACE

more about the errors of the Medicean manuscript, fuller information is provided in an appendix. It was no less difficult to decide upon the form of bibliography to be adopted. To have listed all important modern writings on Tacitus would have been a very large task, and probably a superfluous one too, since S. Borzsak's valuable survey has so recently appeared. I have therefore recorded only works cited in this volume. Even so the list is quite a long one, and perhaps it could have been shortened by the omission of such items as pertain very indirectly to Tacitus. But, though at first I tried to exclude these items, no clear dividing-line was to be found, and in the end it seemed best to include everything, apart from general works of reference, editions, and commentaries. Editorial work on the Annals is described at some length in the second section of the introduction. And, of course, supplementary lists of works cited will be provided in future volumes of this edition. In all references to published opinions I use only surnames (with or without initials), never adding styles or titles. Whenever such nomenclature is used, it indicates that I am reporting opinions privately communicated to me. It remains as my most pleasant task to express, as best I can, my gratitude to all who have helped to bring this book into the light of day. Queens' College, Cambridge, gave me ideal surroundings and ample opportunity to begin my work, and Bedford College, London, for its continuation. The Classical Faculty Board of the University of Cambridge provided funds for the purchase of certain necessary material. Several friends have advised over details: acknowledgments of their assistance will be found in various footnotes. But to three friends my debt is immense, to Charles Brink, for his encouragement from the start, for his wise advice and infinite patience, and, above all, for his example, to John Crook, for reading early drafts of part of the commentary and by so doing saving me from some serious errors, for much constructive criticism, and for several original and conv111

PREFACE

vincing ideas, and to Ronald Martin, of all scholars the most deeply versed in the intricacies of Tacitus' style, for reading almost the whole commentary in a late draft, for numerous perceptive and illuminating comments on it, and for explaining to me so much of what I shall in the following pages try to explain to others. Of course those who have so generously sustained and counselled me have no responsibility at all for the views I have expressed. Finally, I am most grateful to Joan Booth and James Diggle for assistance in the correction of the proofs, to my wife for help in preparing the map, and to the staff of the Cambridge University Press for their invariable kindness, patience, and efficiency.

Bedford College,London January 1972

F.R.D.G.

-------. 50

o 0

50

100

Miles Km

THE GERMAN

FRONTIER

INTRODUCTION

I

GTA

I CODEX

LAVRENTIANVS

MEDICEVS

68.1

The character and provenance of this prince of codices, our sole authority for Annals 1-6, have been discussed so often and so thoroughly by earlier scholars that the briefest outline will be more than sufficient here. 1 I have nothing new to contribute. The first Medicean dates from about the middle of the ninth century. It is written in a Carolingian minuscule, except that, towards the beginning, some Merovingian forms occasionally appear. There are corrections by the original hand and others which seem to be by an early, but different, hand. On this matter it is not always easy to judge, but in general M I is extremely easy to read. It has many recurrent errors, but few, if any, of special palaeographical interest. 2 In its margin we find numerous annotations: they are usually attributable to the first editor of Annals 1-6, Beroaldus the younger. It is commonly and plausibly surmised that M I emanates from Fulda: experts affirm that its script is characteristic of 1

1

The most important aid to our study of M I is, of course, the magnificent facsimile reproduction, published at Leyden in 1902 as vol. 7. 1 of S. de Vries' Codices Graeci et Latini photographice depicti and accompanied by a valuable introduction by H. Rostagno. Amongst other contributions I must mention those of K. Heraeus, Studia critica in Mediceos Taciti codices, Cassellis 1846, I. G. Orelli (aided by I. G. Baiter) in the preface to his edition (Zurich 1846), viiff and in its apparatus, F. Ritter in the preface to his edition (Leipzig 1864), vff, F. Philippi, Philologus 45 ( I 886), 3 76--80, G. Andresen, De codicibus Medi«is annalium Taciti, Berlin 1892, M. Lenchantin de Gubernatis in the preface to his edition of Ann. 1-6 (Rome 1940), vff, and E. Koestermann in his preface, pp. vff of the 1960 edition. These recurrent errors are copiously illustrated by Heraeus, Rostagno, Lenchantin, and Koestermann.

3

1·2

INTRODUCTION

this centre 1 and certainly the monk Rudolph, writing there in the mid ninth century, seems to show knowledge of the early books of the Annals.2 Whatever the truth may be about its provenance and that of its exemplar, 3 M I was in the end discovered in the monastery at Corvey and thence, in about 1508, brought to Rome and ultimately acquired, after passing through several hands, by Pope Leo X. 4 He commissioned its publication by Beroaldus. And so in the year 1515, after centuries of oblivion, Annals 1-6 again emerged into the light of day. 6 1

1 1

4

6

M I was at some time joined with another MS of virtually the same date and character. This MS is now Laurentianus 47. 36 or M of Pliny's letters, 'ea scripturae forma quam Fuldensem esse periti agnoscunt exaratus' (R. A. B. ~fynors, preface to his edition of the letters (Oxford 1963), xvii). When the 11SS were separated is as uncertain as when they were joined. See Lenchantin, op. cit. x. Some scholars have thought, and they could be right, that knowledge of Tacitus came to Fulda from Tours. See Lenchantin, Ann. 1-6, vii-ix. The express testimony of Beroaldus, Soderinius, and Pope Leo, there collected, leaves little doubt that M 1 was not extracted from Germany before the beginning of the sixteenth century. But a few scholars have discounted this testimony and supposed that the l\1S came to Italy much earlier: see e.g. G. Voigt, Die Wiederbelebungdes classisclzenAlterthums, Berlin 1893, 1. 251-3 and L. Pralle, Die Wiederentdeckungdes T acitus, Fulda 1952. Such suppositions are very properly rejected by Lenchantin, by Koestermann, op.cit. v n. 1, and by Mynors, op.cit. xvii n. 1. If I say nothing about the fortunes of Ann. 1-6 between the ninth and sixteenth centuries, it is because there is nothing usefully to be said. The most tantalizing problem about this text's transmission concerns an earlier period: when was book 5 mutilated and was its mutilation purely a matter of chance?

4

II EDITIONS

OF THE ANNALS

Tacitus was singularly fortunate in his early editors and critics, 1 for he attracted the interest of several of the finest scholars of the sixteenth century. 2 The first great age of Tacitean studies begins with Beroaldus' edition in 1515 and ends with Pichena's edition in 1607. In quality, as well as in quantity, the contributions of this period have not been surpassed. That those early in the field reap the easier harvest is true enough, but not always the whole truth. The editioprinceps of Annals 1-6 by Philippus Beroaldus the younger is no perfunctory and commonplace piece of work. To be sure the 1

1

For a convenient survey of editions of Tacitus see C. W. Mendell, Tacitus: the man and his work, New Haven 1957, 349-78. Mendell assemble3 much useful information not readily to be found elsewhere, but some of the opinions he expresses on the value of particular editions are misguided. For much detailed evidence about the work of sixteenth-century scholars on Tacitus see also J. Ruysschaert, Juste Lipsett les Annales de Tacite, Louvain 1949. For a judicious assessment of many aspects of it see C. 0. Brink, 'Justus Lipsius and the text of Tacitus', JRS 41 (1951), 32-51. I should add that J.E. Sandys, A history of classicalscholarship,Cambridge I go~, has disappointingly little to say about the contributions of some of the most important Tacitean scholars of this period, notably Pichena. As to the fifteenth century, many true corrections and many false ones are to be found in the MSS of Ann. 11-16 and Hist. 1-5 descended from M 2. Often enough they passed into the early printed texts, and there was probably some traffic the other way too, from printed texts into MSS. But MSS of this time are rarely, if ever, systematic editions: see E.J. Kenney, in Classicalinfluenceson EuropeancultureA.D. 500-1500 (ed. R. R. Bolgar), Cambridge 1971, 11g-28. The typical humanist MS, as Kenney says ( 1 22), was 'a random hotchpotch of tradition and often wilful and occasionally violent alteration'. And most early printed texts inherited these characteristics from the MSS on which they were based. In recent years Mendell and others have paid far more attention to fifteenth-century MSS of Tacitus than any of them deserve.

5

INTRODUCTION

majority of Beroaldus' corrections are such as anyone with a little Latin could have made, given a ninth-century MS never systematically corrected before. But a respectable minority attest perception, judgement, and understanding of Tacitus' usage. Beroaldus has undeservedly been eclipsed by his eminent successors: though he can hardly bulk very large in the history of scholarship, he merits fuller recognition than he has yet been accorded. 1 The edition pirated from Beroaldus and printed at Milan in 1516 need not concern us,2 nor the Froben edition of 1519,3 nor the Juntine of 1527. But the second Froben edition, which appeared at Basel in 1533, is a different matter. This is the first of the great editions ofTacitus. Valuable as had been the services of Puteolanus 4 and Beroaldus, Tacitus now for the first time came into the hands of a scholar of the highest ability and foremost amongst the Latinists of his generation, Beatus Rhenanus. 6 Rhenanus' success may in part be measured by the ready absorption of his conjectures and his observations into the common stock of material upon which his successors worked. But the conjectures at least may conveniently be separated out and assessed. They are naturally far fewer than Beroaldus ', but they are not few, and they show the activity of an exceptionally penetrating and disciplined intellect. Indeed I go 1

1

8

4

6

The multitude of Beroaldus' corrections poses a problem for the scrupulous mode1n editor. Like many others before me, I have tried to find a compromise. Such corrections as seem to me of some merit are recorded in the apparatus. All the others, except only the slightest corrections of spelling, are collected in Appendix I as 'adnotationis criticae additamenta '. The notes of A. Alciatus appended to this edition are the only material it contains of any interest. This edition is of some importance, in that it contains a text of the Germaniarevised and corrected by Rhenanus. In his Milan edition of 1476 and his Venice edition of 1497. Rhenanus deserved well of other Latin historians, and he holds a special place of honour as sospitatorVelleii, badly though his edition of Velleiw was executed. But his Tacitus is his best work.

6

EDITIONS

OF THE

ANNALS

further: if quality of emendations were the only criterion of scholarly excellence, Rhenanus might vie with Pichena and Nipperdey for the second place amongst Tacitean scholars. His conjectures are not often, one must admit, brilliant and exciting, but they possess the more valuable quality of often being right. In his patient attention to grammar and idiom, and his concern with logic and precision, Rhenanus might well be compared with Madvig. Lipsius built upon the foundations which Rhenanus had laid. 1 If his occasional mentions of Rhenanus are mainly adverse criticisms, that is largely the fashion of the time: it was not then conventional to record specific obligations or to express agreement. 2 The Aldine edition (Venice 1534) has, it seems, no independent value. But the Gryphian edition (Lyons 1542) contains, along with earlier material, some notes by A. Ferrettus which are not entirely negligible. 3 And in 1544 Rhenanus produced a second edition (published, like the earlier one, by Froben at Basel), with corrections and modifications.' Nothing of importance can be recorded for the next twenty years and more, 5 but a spate of activity Lipsius based his text on Rhenanus' revised edition of 1544, as Ruysschaert has shown. 1 Rhenanus' conjectures get a somewhat lukewarm appraisal from Sandys, History of classicalscholars/tip2. 263 and Mendell, Tacitus 362, probably because they are usually unspectacular. I do not understand what Mendell means when he says that Rhenanus' corrections 'are on the whole of a conservative character'. In AM. 1-6 at least, while most of his changes are neat and simple, they are not conservative, for as yet there was no firmly agreed text to conserve, and changes do not con.serve a text still undetermined. Changes back from a vulgate to the paradosis may properly be called conservative, but these are not in question. 3 Lipsius used this edition as the starting-point for his commentary. ' That it was no mere reprint has been established by Ruysschaert, Juste Lipse 20-1. 1 This gap is perhaps not wholly the result of chance. A new generation of scholars, such as V ertranius and Li psius, were to offer a new approach to Tacitus, in particular by the use of more diverse historical evidence. Tacitus became less of a literary text, more of an object for scholarly 1

7

INTRODUCTION

begins with the acute and original notes on the Histories and Annals by M. Vertranius Maurus (Lyons 1569).1 And now a commanding personality appears. First and indisputably first of all Tacitean scholars stands Justus Li psi us, whose earliest edition of Tacitus was published by Plantin at Antwerp in 1574. Lipsius cannot rank with Bentley and Heinsius and Housman, but he can bear comparison with those who come next, such as his contemporary Scaliger. If he lacks Scaliger's dazzling ingenuity and breadth of interest, he is by no means narrow in scope nor devoid of imagination. He excels Scaliger in sanity of judgement, though he almost equals him in egotism. Lipsius was pre-eminently qualified to edit and interpret Tacitus, by his versatility in conjecture, by his sense for style, and by his perfect knowledge of Roman history, as far as it could at that time be known. 2 His main contribution is twofold: in a vast improvement of Tacitus' text and in the creation of the first complete commentary on Tacitus. Tiny this commentary may be by the scale of later work, but it was the necessary nucleus upon which a mass of later accretions gathered, some inferior and superfluous. Two hundred and fifty years were to elapse before students of Tacitus could discard from direct consideration that body of learned material, based upon Lipsius, which was the substance of so many variorum editions. If we are fairly to appraise Lipsius' success in conjectural emendation, we must remember that he worked upon a textus receptusoften far removed from M I and M 2. After its initial employment by Beroaldus, M I had in the sixteenth century largely been neglected. And M 2 had never been

1 3

research. Such a change of attitude naturally takes some time to become established and reflected in published works. Sec Ruysschaert, Juste Lipse 38. A. Momigliano, J RS 39 ( 1949), 1go, justly observes that 'his combination of feeling for style with historical knowledge is still a challenge to any editor of an historical text'.

8

EDITIONS

OF THE ANNALS

used directly as the foundation of a printed text. Much of Lipsius' most admirable work was in a sense unnecessary, for in numerous passages he laboured to restore by conjecture what was later found to be transmitted by the Medicean MSS. These conjectures now go unrecorded in critical editions, but, thanks to Ruysschaert, we can at least take account of them in assessing Lipsius' achievement. Even if we were unjustly to set them aside, the number and quality of his other conjectures would still fully justify his reputation as a critic. Unhappily this reputation is not unblemished. It seems likely that Lipsius drew rather too freely on the work of Muretus (see Brink, 'Justus Lipsius' 51), though in this instance conscious plagiarism cannot be proved. Perhaps it can be proved for Lipsius' use of certain unpublished conjectures of Claude Chiffiet: at least Ruysschaert, Juste Lipse 144ff, has produced documentary evidence which strongly suggests that Lipsius knew Chiffiet had prior rights to ideas which he published as his own. If this evidence is valid, we cannot exonerate him completely on the grounds that the age in which he lived accepted less exacting standards of scholarly proprietorship than our own. But it is proper to emphasize that the likely cases of plagiarism are, amongst the multitude of Lipsius' conjectures, very few indeed. 1 One enduring result of Lipsius' work deserves special mention. He established as canonical Vertranius' division 2 1

2

Lipsius' great celebrity led to a good many conjectures being ascribed to him which had in fact been advanced earlier by others. As far as earlier published work is concerned, we may without hesitation restore the correct attribution. Unpublished work poses more of a problem. Perhaps, as Brink says, 'Justus Lipsius' 50, it would be 'an act of justice' to give back his unpublished conjectures to Chiffiet. But later scholars, as well as Lipsius, are here concerned, and we cannot assume that they had access to Chiffiet's notes. To give two names is usually the fairest solution. See Ruysschaert, Juste Lipse 147 n. 4. Incidentally Ferrettus, not Lipsius, first divided books 5 and 6 of the Annals.

9

INTRODUCTION

of Tacitus' historical writings into Annals and Histories.1 No one now has much confidence that this titulature is correct, but it would be fruitless to change a convention so long accepted, when we have nothing more certain with which to replace it. The last quarter of the sixteenth century saw an abundance of excellent work on Tacitus. Particular mention is due to the contributions of M.A. Muretus, in his Variae Lectiones (Antwerp 1580) and elsewhere, of F. Modius, in his Nouantiquae Lectiones (Frankfurt 1584), and of F. Ursinus and I. Mercerus in their Notae, published respectively at Antwerp in 1595 and Paris in 1599. Meanwhile Li psi us' edition was going through several revisions. 2 The last appeared in 1607, the year after his death. This year was indeed an annus mirabilis in Tacitean studies, being additionally distinguished by the appearance at Frankfurt of the immensely important edition by Curtius Pichena and the useful variorum edition by lanus Gruterus, and at Hanover of the acute and original notes of V. Acidalius. Pichena's services to Tacitus are second only to Li psi us'. If Lipsius was a child of his age in his cavalier attitudes to the MSS, Pichena had in no small part grasped one of the great lessons which Bentley was to teach a hundred years later, that MSS must be weighed, not selected at random. He recognized the outstanding importance 3 of the Mediceans, 1

I have discussed these matters in the second note in my commentary. See Ruysschaert, Juste Lipse x-xi and Brink, 'Justus Li psi us' 32 n. 2. • I choose my words very advisedly here. Pichena did not recognize the unique importance of M 2. If he had, he would have discarded from consideration Puteolanus' edition of 1497. Pichena had already, before his edition, published information about the readings of the Medicean MSS, in two collections of notes (Hanover 1600 and 1604). Lipsius knew of this crucial information from these notes: see Brink, 'Justus Lipsius' 33 n. 10. His reaction is typical of his time and personality. He was delighted to find many of his conjectures confirmed, but unable or unwilling to perceive that the whole basis of his text of the Annals and Histories needed to be changed.

1

10

EDITIONS

OF THE ANNALS

and made them one of the main foundations of his text. Indeed he was the first editor of a printed text to use M 2 at all. And he rescued M I from increasing neglect. By his use of these MSS Pichena was able to effect very substantial improvement over all texts earlier current. But he did much more by his own skill, for he was an adroit and careful emendator, endowed with a rare sensitivity to the niceties of Tacitus, language and style. It might have been better for Tacitus if some of Pichena,s successors had followed his example, rather than expending so much labour on the fifteenth-century MSS. Pichena was a prophet who won only limited acknowledgment from his contemporaries, and scant praise from posterity. Gruterus, edition contributes a little of independent value, but, like the Paris variorum edition of 1608, is mainly useful as gathering together the fruits of learned work by scholars of the preceding century. Gruterus is usually credited with establishing the chapter divisions in Tacitus which have become conventional. 1 The next two hundred years brought only slow and intermittent progress in the study of Tacitus: as a whole this was an arid period and in retrospect the more depressing because substantial advances were being made elsewhere. Tacitus had not ceased to be popular. It was rather that, without new information and new scholarly methods, original research on the constitution and interpretation of his text could not be sustained at the level it had attained by the end of the sixteenth century. We must wait until the mid nineteenth century before Tacitean studies will again advance so quickly and on so broad a front as in their first and best period. The edition by M. Bernegger (Strassburg 1638) claims passing mention because it contains notes by I. Freinsheim, 1

On this convention see below, p. 18. According to Mendell, Tacitus 367, the credit for the chapter divisions should go to Pichena, not Gruterus. I I

INTRODUCTION

some of which are good. I. F. Gronovius did not himself produce an edition of Tacitus. His notes, together with notes of earlier scholars, were produced by I. Gronovius in his Amsterdam edition of 1672, and again in the Utrecht edition of 1721. They are sometimes illuminating, but, as a whole, Gronovius' contribution to the criticism of Tacitus is somewhat disappointing. 1 The edition by T. Ryck (Leyden 1687) is, as far as original ideas are concerned, unimportant. But Ryck's extensive use of the MS now called Leidensis BPL 16B deserves to be recorded. Even those who utterly discount the extravagant claims recently made for this MS may freely admit that it is an exceptionally abundant source of true, as of false, corrections. The only edition of any importance which appeared in the eighteenth century 2 was that by I. A. Ernesti (Leipzig 1752), and it fell short of what might have been expected from it. We do indeed, as Mendell says, Tacitus 369, have here 'the first attempt at a critical edition'. Ernesti enjoins proper reporting of the MSS and refuses to attribute any authority to a vulgate text. In practice, however, he systematically employs only Gudianus 118 and fails to perceive, as, after the work of Pichena, he should have perceived, that M 2 alone forms the basis for our text of Ann. 11-16 and Hist. 1-5. In general, though he deserves credit for his discontent with outdated, pre-Bentleian attitudes and though some of his notes show sense and discrimination, Ernesti's tangible services to Tacitus are neither substantial nor very distinguished. I suspect that he was rather unsympathetic to Tacitus' colourful and emotive style: certainly many of his observations seem almost wooden and somehow to miss the point. 1

1

I should add a word about I. Gronovius. He offers a few excellent proposals of his own. Further, the edition of 1721 is most attractive in presentation and convenient to use. The younger Gronovius set the standard of the variorum editions of his time. Those by G. Broticr (Paris 1771) and G. C. Croll (Bipontinc 1779) are negligible.

12

EDITIONS

OF THE ANNALS

In the early nineteenth century we enter a period of greater activity and move gradually into an age of transition. In 1801 there appeared at Leipzig a revision of Ernesti by F. A. Wolf and I. I. Oberlin :1 it makes some contribution both in criticism and interpretation. G. A. Ruperti's first edition of the Annals soon followed (Gottingen 1804): of Ruperti more in a moment. I. Bekker's edition (Berlin 1825) scarcely deserves mention, for he does no more than gather together earlier work. There is, in contrast, no lack of originality in the edition by G. H. Walther (Halle 1831-3). Walther was a stubborn and often obtuse defender of the paradosis. Though all too often prepared to defend the indefensible and explain the inexplicable, he sometimes rescued Tacitus' text from gratuitous change or timehonoured misunderstanding. Ruperti's second edition (Hanover 1834) is chiefly remarkable for the vast mass of information it presents, bibliographical, critical, and illustrative. It is still useful, though frequently exasperating to the reader, being a strange mixture of old and new, a variorum edition half metamorphosized into an up-to-date edition with commentary. The edition by L. Doederlein (Halle 1841-7) claims passing notice, for he advances a few highly intelligent ideas. I. G. Orelli's edition (Zurich 1846) is on the whole a poor and scrappy piece of work, much inferior to his Horace. But Orelli offers a number of original suggestions, and, more importantly, an account of M I and M 2, based on collations by I. G. Baiter, which is fuller than any available before. The contributions of F. Ritter in two editions (Cambridge 1848 and Leipzig 1864) are not contemptible. He might have achieved much if he had been endowed with stability of judgement. As it is, his work is marred equally by irrational conservatism and by irrational conjecture. By the 1840s one can see four major developments 1n 1

The main part of the work was done by Oberlin.

INTRODUCTION

Tacitean studies: recognition of the authority of M I and M 2, formation of a critical apparatus fit so to be described, discarding of much of the material contained in the variorum editions, and use of new evidence, drawn from more intense investigation of the language, of coins and inscriptions, and of other historical sources. These developments are to continue and accelerate in the decades which follow. The modern period in the editing of Tacitus begins with K. Halm's edition (Leipzig 1850) and K. Nipperdey's edition with commentary of the Annals (Leipzig 1852). Halm's original contributions, though considerable, would not place him in the front rank of Tacitean scholars. But the influence of his modest edition has been immense. It, more than any other, has formed the basis of a modern vulgate, adopted without reconsideration by many editors of recent times. Nipperdey's work has been no less influential and in itself is more important. Nipperdey possessed the most alert and penetrating intellect of all who have worked on Tacitus for any length of time, and he merits the highest praise both as a critic and a commentator. 1 His numerous conjectures, though not always compelling, are invariably of some value or interest, if only diagnostically. And many are right. We owe to him a greater improvement of the text than to any editor since Pichena. Again, he created a new style of commentary on Tacitus, drawing as a matter of course upon the latest work in all relevant fields of Latin studies. 2 His commentary is perhaps unduly selective in choice of topics and problems for discussion, but it is distinguished throughout by independence, insight, and good sense. Later work has, not surprisingly, superseded some of Nipperdey's conclusions 1

1

Naturally I take account here not only of Nipperdey's 1852 edition of the Annals and revisions of it, but also of his edition of Tacitus as a whole (Berlin 187 1-6). Of course we find some up-to-date information in the old variorum editions, and much of it in a few of them. But in these editions outdated and exploded ideas were continually reproduced. Nipperdey broke this bad tradition in commentaries on Tacitus.

EDITIONS

OF THE ANNALS

and occasionally shown his judgement to be errant. For all that, his successors, Furneaux, Koestermann, and I, depend heavily on him: he cleared and marked the path which we have followed. In the second half of the nineteenth century Tacitus' subject-matter, sources, and historiographical technique attracted more intense and sustained scrutiny than ever before. And, no less importantly, E. Wolfflin in the mid 1860s initiated a more exact and systematic investigation of Tacitus' language and style, some of the fruits of which were to be gathered later in the great Lexicon T aciteumof A. Gerber, A. Greef, and C.John (Leipzig 1877-go). 1 The results of all this work brought substantial benefit to commentaries on Tacitus, and some benefit to his text. Three lesser contributions of this period claim brief mention, the edition of the whole of Tacitus by F. Haase (Leipzig 1855), of the Annals by A. Draeger (Leipzig 1868), and of Annals 1-3 by R. Novak (Prague 1890). Haase and Novak offer a little at least which is new and interesting,• and Draeger sometimes shows the independence to be expected from one so deeply versed in Latin usage as a whole and Tacitean usage in particular. The only contribution which, for its intrinsic merit and enduring influence, may in modern times be compared with the work of Nipperdey, is _the edition and commentary by H. Furneaux (Oxford vol. 1 1884, 1 1896, vol. 2 1891, revised by H. F. Pelham and C. D. Fisher 1907).3 Furneaux was one of the most unassuming, level-headed, and judicious of 1

1 1

There is much in Gerber-Greefwith which one may disagree. Naturally so, for it is a scholarly work and full of controversial opinions. I esteem it the more every time I look at the wretched computerized products which now masquerade as lexica and concordances. It was Haase who in Philologus 3 (1848), 152-3 finally settled the problem about the division of books 5 and 6 of the Annals. Mendell, T acitus 373, does gross injustice to Furneaux by setting his edition on a level with the trivial work of Holbrook, now deservedly ignored.

15

INTRODUCTION

commentators. He could quickly grasp the essence of a problem, make a point clearly and briefly, and balance conflicting views without prejudice or confusion. He mastered his subject as a whole. The substantial merits of his commentary are eloquently attested by its survival as a standard work for over eighty years. But the passage of time has inevitably antiquated much of the information Furneaux provides, particularly on historical matters. And there were certain weaknesses in his work from the start. The most serious, I think, is that in discussing the earlier attestation of words and phrases used by Tacitus he often omits essential evidence. One result of these omissions is that the extent ofTacitus' debt to Livy is obscured, 1 while his debt to Vergil and other poets is somewhat exaggerated. Again, Furneaux shows little independence over textual problems, and some of them he treats far too sketchily. And again, he concerns himself hardly at all with Tacitus' presentation of individual scenes and episodes. Though Furneaux's work has these weaknesses and overall falls short of real distinction, it richly deserves the high regard in which it has so long been held. G. Andresen's revision of Nipperdey's edition and commentary (Berlin 1892)2 effected some improvements in detail. Merely by keeping this valuable work up-to-date Andresen performed a useful service. No edition of major importance has appeared in the present century. That by C. D. Fisher (Oxford 1906) has fewer obvious demerits than several later editions, but Fisher's positive contribution is slight in the extreme. 3 Andresen has a little which is new to offer in his revision of 1

1

3

I fear that Furneaux had undue confidence in Lewis and Short' s Latin Dictionary, in which Livy is inadequately represented. See W. B. Anderson, 'Livy and the lexica ', CQ 25 ( 1931), 38-48. Nipperdey had seen his work through several editions. Andresen too produced several editions of his revision, the last of vol. I in I 915 and of vol. 2 in I go8. On Fisher see further p. 20 n. 2.

16

EDITIONS

OF THE ANNALS

Halm (Leipzig 1913) and he was able to make some use of the results of his meticulous and searching study of the Medicean MSS. On the whole both these editions are competently executed and no disgrace to the series in which they were published. The same cannot be said of the edition by H. Goelzer and others (Paris 1922-), which is disfigured by numerous inaccuracies, particularly in the apparatus. In his various revisions of Halm-Andresen (Leipzig 1 1934-) E. Koestermann has not contributed substantially to the emendation ofTacitus' text. 2 But his editions are quite reasonably now regarded as standard, partly because there is nothing better to put in their place, partly because they present, as most other editions do not, a sub-division of chapters highly convenient for reference. 3 One must admit, however, that his editorial work since 1960 is sadly vitiated by his addiction to the Leyden MS :4 nothing so radical and misguided as Koestermann's attempt to set this MS on a level with M 2 has been seen in Tacitean studies since the days of Ross and Hochart. It is in interpretation, rather than in criticism, that Koestermann's main strength lies. In several important papers 5 and in his commentary on the 1

I should perhaps record that I have used Koestermann's 1960-2 edition for my citations from other parts ofTacitus' writings, correcting it not infrequently when it seemed wrong, particularly in punctuation. 1 But see further p. 19 n. 1. 1 The credit for making this sub-division should largely go to Goelzer. But Koestennann more than anyone else has established it as conventional. Fumeaux, before either of them, devised a similar, but different, system of his own: it is still occasionally used by modern scholan, but it has never been adopted generally. Fisher has no subdivision of chapters at all. • See H. Heubner, Gnomon34 (1962), 159-63, R.H. Martin, CQ n.s. 14 (1964), IOg-19, F. R. D. Goodyear, CQ n.s. 15 (1965), 299-322 and 20 (1970), 365-70. ' Historia4 (1955), 72-I06, 6 (1956), 429-79, 7 (1957), 33-75, IO (1961), 330-55. In my estimation the best ofKoestennann's work is to be found here, rather than in his commentary.

INTRODUCTION

Annals (Heidelberg 1963-8) he has greatly advanced our understanding of Tacitus as a historian and stylist. But he has little to say which is not entirely derivative about detailed matters of text and language. 1 The edition of Annals 1-6 by M. Lenchantin de Gubernatis (Rome 1940) is hard to judge fairly. I shall resist the temptation to damn it for that extreme conservatism which R. Syme has inordinately praised. 2 Lenchantin's edition is, for all its faults, the work of a learned man and a scholar. Though his grasp of Latinity is weak and uncertain, he could occasionally produce new ideas of no little cogency. The edition of the Annals by H. Fuchs (Frauenfeld 1946--g) is of only slight importance, but it is not uninteresting. Fuchs possesses one quality which many of his predecessors have lacked, independence. He refuses to accept the easy doctrine that all the textual problems in Tacitus which can be settled were settled long ago. Unfortunately his critical powers are not commensurate with his admirably sane approach to his work, and, as a result, he has imported into the text various dubious or unnecessary conjectures. But he is not always finding unreal problems: in his paragraphing he has rightly broken away from the practice of most recent editors, who have acquiesced too willingly in a convention founded neither on reason nor authority. 3 1

1 1

His dependence on Furneaux for observations on Tacitean usage is hardly less than scandalous, but it has passed largely unnoticed by reviewers. He has at least paid the plagiarist's penalty, for some of the information he copies is erroneous or at least incomplete. One must, I fear, admit that there is a good deal of justice in Syme's bitter remark (JRS 38 (194-8), 123) about 'the sloth and the psittacosis of the commentators'. See the next section of this introduction. It is necessary to keep in mind the distinction between paragraphing on the one hand and division and sub-division of chapters on the other hand. Chapter-divisions and sub-divisions, in Tacitus as in quite a lot of other Latin texts, may primarily now serve to provide a convenient system of reference. They fall within the sphere of convention. But in marking paragraphs, which may or may not be identical with

18

EDITIONS

OF THE

ANNALS

The school-edition of Annals I by N. P. Miller (London 1959) deserves brief notice. Miller offers many just and pertinent observations upon Tacitus' language and style, and some of them are original. If she had consistently thought for herself, particularly on textual matters, her book would have been outstanding amongst its kind. Much work remains to be done, though at present the Leyden MS, like King Charles' head, is diverting attention from the main business. But a 'law of diminishing returns' applies inexorably to the study of such texts as the Annals and Histories:we cannot reasonably expect future contributions as substantial and exciting as those of the past.

III ON EDITING

TACITUS

Conservatism has been prevalent in the editing of Tacitus throughout this present century,1 though recently, thanks not least to the activities of C. W. Mendell and E. Koestermann with the codex Leidensis, we have had two new contenders for supremacy, Chaos and Confusion. Of the worth of the Leidensis I have written at length elsewhere. 9 Here I shall be concerned with the problems of editing Tacitus more generally3 and, in particular, to say something against

1

1 1

chapters, we attempt, however imperfectly, to represent divisions which properly belong to the writings we are editing. And so paragraphing falls within the sphere of criticism. Of course all this complexity derives from a long accretion of scholarly work. Those scholars who made the original chapter-divisions were trying, often rightly, to do exactly what we can now do with our paragraphs, while retaining their numeration. H. Fuchs' divergence from current fashion has been mentioned above. More importantly, E. Kocstermann sometimes exercises independent thought on textual matters, and to good effect. CQ n.s. 15 (1965), 299-322 and 20 (1970), 365-70. For convenience I shall take examples from Ann. I. 1-54.

INTRODUCTION

the pronouncements of that eloquent apostle of reaction, R. Syme, delivered for the benefit of editors ofTacitus everywhere in JRS 38 (1948), 122-31. The arbitrary and violent use of conjecture by several eminent scholars of the mid nineteenth century inevitably fostered a conservative revival,1 in which Tacitus' editors were not slow to take part. 2 But more particular reasons may be advanced for the poverty of work on Tacitus' text over the last two or three generations. Since Nipperdey no critic even of the second rank has edited Tacitus and very few critics have concerned themselves with his text. 3 It is not therefore surprising that something like a textus receptushas established itself. By this I mean specifically that for the generality of editors of Tacitus in the last hundred years many open questions about punctuation, paragraphing, and the correction of scribal error appear to be closed questions: they simply follow their immediate predecessors. Again, the linguistic studies of the so-called 'Swedish school', E. Lofstedt, 4 N. Eriksson, 6 and G. Sorbom, 8 have 1

See Housman, Manilius 1, pp. xli-xliv. C. D. Fisher may well represent them. In the preface to his Annals (Oxford 1906) he smugly asserts that, as far as he dared, he has followed the Medicean MSS through thick and thin. Predictably he wins high praise from R. Syme, J RS 38 ( 1948), 128--g. More surprisingly C. 0. Brink, JRS 41 (1951), 32 n. 1 calls Fisher's editions 'meritorious'. Where their merit lies I do not know, but it certainly lies neither in original contributions nor in the fidelity to M I and M 2 which Fisher claims he has shown. He has, for example, foisted on Tacitw numerous spellings not in M 1, even though that MS' text is often no less acceptable than what he prints: so, quite unaccountably, he often prints -is accusative plurals where M I offers -es (see in the first few pages of the Annals 1. 3. 3, 1. 9. 3 (bis), 1. 9. 5, and 1. 10. 2). 1 Brink's magisterial paper, already mentioned, stands in a class of its own amongst modern writings on the text ofTacitus, unless one should set beside it some of the exceptionally acute discussions of particular textual problems by H. Heubner. ' In particular Phi/o/ogischn-Kommentarzur PeregrinatioAetheriae,Uppsala 1911 and Syntactica,vol. 1\ Lund 1942, vol. 2 Lund 1933. 1 Studim zu den Annalen des Tacitus, Lund 1934. • Variatio semumis Tacitei, Uppsala 1935. 1

20

ON EDITING

TACITUS

lent an appearance of authoritative support to the irrational proceedings of such ultra-conservatives as M. Lenchantin de Gubernatis. 1 Syme (JRS 38 (1948), 123) would have us believe that the Swedes have shown that 'Tacitus is capable of anything if he can avoid the normal, the monotonous, the conventional'. Fortunately no editor of Tacitus has in practice carried this theory to its logical conclusion by reinstating all the corruptions which his MSS offer. Those who wish to contemplate the worst excesses of Korruptelenlcult must for the present look elsewhere. But, when the textus receptusof Tacitus has in recent years been changed at all, the change has usually been calculated to bring it nearer to the MSS. Not all these changes have been for the worse,2 but more often than not return to the MSS has been one and the same thing as departure from what ratioet resipsa suggest that Tacitus wrote. We owe to Lofstedt and his pupils a fuller and more precise understanding of Tacitus' language. But, for all that, their work is open to many criticisms, general and particular. I shall raise some of them in what follows. The most important way in which the 'Swedish school' has influenced editorial practice is by encouraging editors to preserve anomalies and inconsistencies, in orthography, in grammar, and in syntax. Here again we may see the effects of a reaction. In the mid nineteenth century several editors of Latin prose texts, most notably Madvig, showed a persistent inclination to remove anomalies and standardize the usage of the authors they edited. And they went rather too far. 3 More anomalies are to be found in Latin literature, 1

1

1

In particular in his edition of Ann. 1-6 (Rome 1940), described by Syme J RS 38 ( 1948), 128 as 'that model of scholarly conservatism'. For example 1. 4. 4 exulem, rightly restored in place of the conjecture exul, commonly accepted, and 1. 34. I sequeet in place of Sequanos,for long the vulgate: see my notes on these passages. One may admit, I think, that Madvig's attempt to tighten Livy's syntax, by adjusting anomalous passages to accord with Livy's (or 21

INTRODUCTION

even of the best periods, than we can emend away convincingly. But it does not follow that all anomalies are to be admitted. The main basis for judgement on these matters must be provided by examination of the standards and pretensions of each individual author. Ready acceptance of anomalies, provided they may somehow be explained, is a form of linguistic interpretation which Lofstedt applied with conspicuous success to Latin writings of a late period and low stylistic quality. And this same approach may be useful in dealing with such a writer as Propertius, whose language is sometimes as wayward and unpredictable as his thought. But it is most doubtful whether it may properly be applied to the works of such meticulous and self-conscious stylists as, for instance, Cicero in his published speeches and his treatises, or Tacitus. Close examination of their works will show that they imposed on themselves rigorous standards in vocabulary and phraseology and that they consciously kept to those standards. I will risk seeming to labour the obvious and explain why I describe Tacitus as a 'meticulous and self-conscious stylist'. He is such (i) because of his obsessive concern to find words which are, for his purposes, exactly right in quality and tone. He excludes or discards those which are not. 1 And (ii) because he maintains a consistent stylistic level. There is change and development in his language and style, but there are no abrupt movements up or down, either towards a spurious elevation or towards colloquialism. His constant striving for variety in word and phrase is kept in check by a no less constant striving to create a style of the utmost dignity. And (iii) because his innovations are not arbitrary. It would be hard to find in Tacitus an instance of total divergence from usage earlier attested. For the most part

1

even Cicero's) normal usage, was somewhat misguided. But Madvig and his contemporaries were not so misguided as some of our anomalists, to save themselves thought, would like to believe. For some further observations on these matters see Appendix 2. 22

ON EDITING

TACITUS

Tacitus' innovations are no more than extensions, some indeed bolder than others, of linguistic developments already to be traced in the literature surviving from the first century A.D. :1 no doubt, if the historical writings of this period had survived, Tacitus would appear much less of an innovator altogether. Further, his extensions of earlier usage seem rarely to be isolated: we may usually find parallels or analogies elsewhere in his writings. In a word, Tacitus was neither inattentive nor haphazard in his use of the language he had inherited. A stylist such as I have just sketched is not, I submit, 'capable of anything'. He will not, from mere whim or carelessness, depart from the basic rules of grammar and syntax established by the practice of his predecessors over two hundred years and more. Sometimes indeed, for a definite stylistic purpose, he may strain the powers of his language: the 'impressionistic' syntax which Tacitus employs at 1. 41 is an excellent case in point. But the vast majority of the MS readings which the Swedes defended 2 and Lenchantin printed with acclaim from Syme are harsh departures from established usage which serve no explicable purpose whatsoever. Take, for example, the gross solecism transmitted at 1. 9. 4 non aliud discordantispatriae remediumfuisse quamah uno 3 What had Tacitus to gain by omitting ut, except regeretur. the space of two letters? He would have gained nothing in expressiveness, or in emphasis, or even (since equally bad syntax is found on the walls of Pompeii) in novelty. 'A conservative approach to M', says SymeJRS 38 (1948), 123, 'is recommended not merely by disillusion and by dis1

1

1

I have argued or at least illustrated this proposition in several notes in my commentary. The case of glisco, used increasingly freely as an alternative for cresco,is as pertinent as most others, and somewhat better documented: sec on 1. 1. 2 gliscmte. In so doing they were sometimes anticipated by Walther and by W. A. Baehrens, Beitrage zur Lat.Syntax, Philologus,suppl. 12, 1912. On this passage see Brink J RS 41 ( 1951), 45 n.

INTRODUCTION

taste for all the emendations, "magna moles et improspera ", but by sound doctrine.' And later (127) he suggests that we should 'follow the doctrine of the Swedes and endorse the style and syntax of Tacitus to the utter limit of rapidity, acerbity, and abruptness'. Wise men before descending such a precipice have a close look at the credentials of their guides. That is what C. 0. Brink has done inJRS 41 (1951), 32-51 and what I have done in various parts of my commentary. See in particular Brink's discussions of 1. 10. 5 grauius (43-4), 1. 12. 3 sedet (44-5), 1. 28. 3 et si alii, my notes on 1. 4. 4 aliquid, 1. 41. 1 triste, 1. 41. 1 et externaefidei, and Appendix 5, in which Sorbom's views on ellipse are submitted to scrutiny. From examination of the pronouncements of the 'Swedish school' on these passages and many others it becomes abundantly clear that their doctrine is not 'sound', not based on full and judicious inquiry. Indeed I must go further: much of their work is shoddy, illconsidercd, and misleading. 1 Lofstedt and his pupils were not well qualified to deal with the text of Tacitus, for they misjudged the character of his writings, not perceiving that all his works reveal him as immensely painstaking in choice of word and phrase, 2 as largely self-consistent even in his innovations, and as governed by an exacting sense of stylistic propriety. Or, if they perceived any of this, they failed to grasp the consequences for work on Tacitus' text. Our knowledge ofTacitus' high stylistic quality must influence our decisions on all the passages where the MSS present a divergence from his own normal 1

2

Still, as I have said, we have all learned a good deal from it. And I should add in fairness that, though Lofstedt is sometimes inaccurate (see my note on 1. 1. 1 neque... neque) or deficient in judgement (conspicuously so about 1. 41. 1), his work is in a different class from Sorbom's and Eriksson's. Some sections of his work bear more tokens of limae /abor than others, not least in diligent avoidance of repetitions, but there is scarcely a single chapter to be found which could be described as carelessly written.

ON EDITING

TACITUS

usage or from literary Latin as a whole. We shall not, of course, assume that every abnormality must be evened out, that there can be absolutely nothing idiosyncratic in Tacitus; but we shall examine every textual problem in relation to Tacitus' usage in general and, above all, try to weigh the probabilities in each particular case, asking, for instance, 'is corruption or anomaly more likely here?'. The Swedes erred by preferring anomaly everywhere they possibly could, and that was far too often. A more flexible and openminded approach to the editing of Tacitus is now more than overdue.

IV ASPECTS OF TACITEAN HISTORIOGRAPHY My purpose in this essay is to sketch certain characteristics ofT. 's historical writings which I consider important. 1 It may, I hope, be of some use to the reader to have set out here, albeit rather briefly and dogmatically, several themes and lines of interpretation which will be developed further in the commentary. To no small extent the subject-matter of the Histories and Annals was determined by the traditions of the annalistic genre to which these works belong. By choosing to write annalistic history T. accepted an obligation to record the major events of each year within the spheres of political, military, and constitutional history. 2 But, while he was not 1

1

But some important aspects of T.'s historiographical technique, such as his use of his sources, will not be discussed here, since I have no general observations to make upon them which I have not already made in Tacitus, Greece& Rome, Surveys4, Oxford 1970. It was also traditional to record the deaths of leading men, prodigies, and any notable disasters. And T. does indeed record such matters,

INTRODUCTION

at liberty entirely to exclude matters of recognized importance, he enjoyed considerable freedom in detailed selection of material, and in arrangement and treatment of it. The motives which guide T. in the exercise of this freedom are various and complex. One of the most pervasive is his desire, indeed determination, to interest and move and enthral his audience. To achieve this end he is forever seeking novelty of expression and forever trying to avoid monotony in his narrative, speeches, and characterizations. And this same motive disposes him to select for full elaboration such material as is most susceptible of dramatic and moving treatment. Another motive, hardly less influential, is the need T. feels to understand and interpret what happened during the period of history he has chosen for his subject. 1 In feeling this need T. is, of course, in no way exceptional: indeed no historian can fail to acknowledge it. What is rather exceptional is the insidious subtlety with which in T. preconceived interpretations affect the handling of his material in detail: once he has satisfied himself of what 'really happened' and why, he will then, by choice of word or phrase, by inclusion or exclusion of details, by innuendo and indirect comment, and in many other ways, superimpose his interpretation on the facts, without directly perverting them. Hence the notorious discrepancy in so much of T.'s writings

1

though somewhat selectively. In his choice of recipients of obituaries we may occasionally see, or think we see, a reflex.ion of his own special interests: at least, while no few administrators and politicians die unnoticed, T. does not so readily overlook eminent orators or historians. Again, we may detect, above all with prodigies, an unwillingness to abandon the tradition completely combined with a changed attitude towards it. For example, it is not from a sense of his duty as an annalistic historian that T. reports various prodigies at Hist. I. 3. 2 and Ann. 14. 12. 2, but rather to afford himself opportunity for gloomy and sardonic comments. In these passages, as elsewhere, T. is using the tradition for his own purposes, not loyally following it. One may argue that T. 's concern to probe deeply into the causes of the events he narrates made him dissatisfied with his starting-point of A.D. 14: see on 1. 1. 3 pauca. .. tradere.But no starting-point will ever be perfect in historical writing.

ASPECTS

OF TACITEAN

HISTORIOGRAPHY

between 'facts reported' and 'impression conveyed' .1 A third motive, also of recurrent influence, is T.'s conviction, shared with many of his predecessors, that history has a moral and exemplary function :2 3. 65. 1 praecipuummunus annalium rear ne uirtutes sileantur utque prauis dictisfactisque ex posteritate et infamia metus sit. The immediate purpose of this remark is to justify the omission of such senatorial sententiaeas were notable neither as exemplifying virtue nor for depravity. But it has wider relevance. T.'s conception of the moral function of history prompts him to make many evaluative judgements, directly or indirectly, encourages him to present incidents and situations in a stylized manner, and, perhaps most importantly, leads him to see various historical figures as types, embodying good or evil, rather than as individuals. I shall proceed to discuss the operation of these three motives, and some subsidiary ones, and begin with a particularly important example. T.'s account (1. 11-14) of the debate in the Senate over the succession to Augustus is scandalously incompetent by any standards of historical writing which would be acceptable today. T. nowhere provides information about those matters of constitutional law concerning the position of the princeps which may well seem to have been crucial, at this time above all, when the principate was first passed on to a successor. 3 Even the fact that there was a consular relatio about Tiberius' position is introduced belatedly and incidentally. And yet these chapters, like chapters 1-10, are 1

1

1

This whole matter is of major importance for the simple reason that we often have no independent evidence against which to test the reliability of T. 's narrative. Typically T., unlike Sallust and Livy, does not in his prooemia indicate clearly that he holds this view of history, though he implies it in Hist. 1. 2-3. Some historians doubt whether constitutional formalities were of much importance in the transmission of the principate, but few, if any, would deny that T. is culpable for supplying so little information about them.

INTRODUCTION

carefully designed, indeed masterly in execution. 1 Why then is T. so imprecise in reporting the facts and so wilful in his treatment of them? There is only one satisfactory answer. As I shall argue in the commentary, following Syme, T. is convinced that these proceedings in the Senate were a mere charade, that Tiberius took over all effective power when Augustus died 2 and had no intention of relinquishing any of it. And so he will not concern himself with constitutional niceties which he believes to have been of no real importance, and instead presents the debate as the farce he thinks it was, using the opportunity to reveal Tiberius as a humbug and the Senate as both subservient and ridiculous. No better instance could be found of the way in which a preconceived interpretation may colour and indeed largely determine T.'s account of historical events, nor, conversely, of the way in which T.'s interpretation may be elicited from his manner of treatment and selection of material. 3 T. does not directly affirm that the debate was a farce, but we soon discover that this is his view. More. generally the view which T has formed of the personality and the motivation of an individual may override the conclusions which in any particular case the evidence for that individual's actions seems to warrant. The principal case is, of course, Tiberius. Once T. has satisfied himself that Tiberius was a congenital hypocrite, his conception of Tiberius' hypocrisy both leads him to regard with deep 1

2

3

Syme does not convince me that r. 13. 2-3 is artistically objectionable: see ad loc. See I. 7. 5. T.'s view is clearly suggested by his famous phrase primum facinus noui principatus (I. 6. 1). An interesting instance of omission of material has recently become known. Why does T., at r. 14-15 and 1. 81, take no account of the provisions of the lex Valeria Cornelia, now wearisomely familiar to us from the tabula Hebana? Perhaps from failure to research adequately into a matter which he admits to be very complex. More probably T. regarded the canticum comitiorum as negligible and so deliberately neglected the elaborate arrangements which the Jex Valeria Cornelia provided for it. See further on 1. 15. I tum primum . . .jiebant.

ASPECTS

OF TACITEAN

HISTORIOGRAPHY

suspicion any apparently laudable actions by Tiberius and, at the same time, suggests the means, frequent reiteration of the idea of a difference between appearance and reality,1 by which he may reconcile these actions with a generally discreditable motivation. 2 Again, what applies to T.'s prejudiced treatment of certain individuals applies also to his treatment of certain historical developments, notably that of the Lexmaiestatisduring the early principate. 3 His indignation and shame at the later abuse of this law cause him first to misjudge and then, by many devices of style, to misrepresent its use under Tiberius. Here, if anywhere, it is essential, if we go to T. as a historical source, to separate 'fact' and 'impression'. As I have said, T. is at liberty to decide upon the scale of treatment to accord to particular events, scenes, and episodes. He fully exploits this freedom, as I shall attempt to show shortly. But first a proposition of some moment may be advanced, one of the arcanaof Tacitean history: the scale and elaboration of the treatment of historical events in T. bear no necessary relation to their historical importance.' If this proposition is valid, then historians at least should never forget it when reading T., for the obvious reason that he is our main authority for much of the history of the first century A.D. and for some of it our only authority. It is all 1

1

1

4

The manner of thought so neatly formulated in the Greek A6y'l) µtv ... fpy'l) Si is deep rooted and recurrent in T.'s historical writings, and somewhat variously expressed. See R.H. Martin, Eranos 49 (1951), 1 75-6. In my note on 1. 8. 5 remisit. .. moderationeI shall illustrate at length how, by phraseology, by innuendo, by suggestion of alternative motives, and otherwise, T. succeeds in depreciating many, but not all, of the numerous actions by which Tiberius may seem to have displayed moderatio,modestia,and similar virtues. Amongst the many modern writings on this controversial topic I know of no more balanced and helpful contribution than E. Koestermann's paper in HistQria4 (1955), 72-106. On this point, as on several others, I am much indebted to the perceptive observations of D. Timpe in Der Triumph desGmnanicu.s,Bonn 1968.

29

INTRODUCTION

too easy to attach undue weight to matters which in T. receive prominence and emphasis largely for literary reasons. The most notable example of vast elaboration in T. is provided by his account of the mutinies of A.D. 14 in Ann. 1. 16-52. No sufficient explanation of this scale of treatment can be based on the assumption that T. considered the mutinies of special historical importance, since he is so careful to disentangle them from political events at Rome and deliberately to set aside their possible relevance to Tiberius' hesitations over the succession. 1 T. makes so much of the mutinies mainly because they afford him an unrivalled opportunity to deploy all his resources of style in a type of narrative very congenial to him. Here, on a fuller scale than elsewhere, he can combine pictorial and dramatic treatment, presenting a series of vivid and exciting scenes, varied in character and tempo, but held together by recurrent imagery and skilful planning of the stories as a whole. The German campaign of A.D. 15 provides another example of greater elaboration than seems to be required by the intrinsic importance of what is narrated. I think particularly of T.'s account of Caecina's return to the Rhine (Ann. 1. 63-8). When Germanicus' offensive is completed, Caecina has to bring back part of the army. He runs into the enemy in difficult and dangerous ground. There might have been a disaster for the Romans, but discipline and leadership in the end brings them through successfully. The 1

In one sense, however, T. may consider that the mutinies have some importance. As he sees the history of the first century A.O., there are certain crucial times when the power, violence, and hysteria of the troops are decisive. At least this is a dominant theme in his account of the year of the four emperors: sometimes indeed he seems to diverge from his sources in order to give prominence to the influence of the soldiery. Since this theme so fascinated him when he wrote the Histories, it is hardly surprising that he gladly takes it up again in the Annals, even though he must have realized that the mutinies of A.O. 14 had no lasting historical impact.

30

ASPECTS

OF TACITEAN

HISTORIOGRAPHY

historical importance of an ambush from which a part of the Roman army thus extricated itself is slight indeed. But T. finds the story highly susceptible of dramatic and romantic treatment, and accordingly develops it very fully. This episode is in fact amongst the most highly wrought stylistically in all his writings: in particular we find here an exceptionally rich accumulation of poeticisms, reminiscent of various poetically coloured scenes in Livy .1 And indeed one may detect in T.'s elaboration of such episodes as this a desire to recreate, as far as his material allows, something of the heroic atmosphere of Livy's narrative in his early books. 2 And so I think that at 1. 63-8, as in the account of the mutinies, there is much which may well be of T.'s own creation or at least very freely adapted from his sources. These observations may now be summarized in another general proposition: in T. material which admits of pictorial, dramatic, or heroic treatment is very likely to receive such treatment, often on an extensive scale. But this generalization is in one respect too sweeping, for the suitability of material for elaboration is not alone 1

Note in eh. 65 alone the following expressions, all in varying degree poetical: truci sonore,oberrarent,peruigiles, quies ( 'dream'), resultantes saltw, paludibus emersum, umentia, propcrus, and lapsantes. The best evidence that such accumulations of poeticisms are not, in T. at least, merely the result of chance is their recurrence in somewhat similar contexts. With 1. 65 we may well compare 14. 30, another scene of remote frontier warfare. It is not, of course, in vocabulary alone that the poetic colour consists: consider, for instance, the apparition of Van.is whom Caecina cernereet audire uisus est uelut uocantem,for this recalls several visions recounted by the poets. This emulation of Livy, whether deliberate or not, is only one aspect of that historian's profound influence on T. Certainly this influence extends far beyond vocabulary and syntax. We may trace it in T.'s treatment as a whole of many scenes and episodes: sec on 1. !23. !2 ac ni .•. aberant.As to T.'s remarks at Ann. 4. 3!2where he enjoins us not to compare his writings with those of the historians qui ueterespopuli Romani res composuere,a pinch of salt is probably required, though doubtl~ much of his material did not give him all the opportunities he would have liked.

=

1

31

INTRODUCTION

enough to ensure that elaboration will be provided. Another factor comes into play, T.'s interest in, indeed involvement with, particular subject-matter. One observes that some topics and some personalities fascinate and excite him, and that others do not. For example, T. is more interested in the wars against northern barbarians (in Britain, Germany, even Thrace) 1 than in those against the Parthians or anywhere on the eastern frontier or in Africa. 2 So, on the whole, the northern wars are more elaborately presented than any others. 3 Again, no other historical figures interest T. as much as Tiberius and Germanicus: at least he accords them a fullness of treatment unparalleled in the surviving parts of his writings.' As far as Tiberius is concerned, this is understandable, for there was no such dominating personality amongst the later Julio-Claudians nor in the troubled year A.D. 69. But what of Germanicus? T. uses all the devices of his art to enable Germanicus to attract and hold the reader's attention. Further, in his final assessment of Germanicus at Ann. 2. 73. 1-3, T. presents us with a view of the prince's calibre which is in all probability far removed from the truth: Germanicus, it is suggested, might be compared with Alexander the Great and come off well from the comparison, and, if he had possessed sole power, he might have settled I have in mind the highly colourful account of a Thracian campaign at Ann. 4. 46-51, a passage remarkable 'inter alia' for exceptionally close clwterings of asyndeta: see on 1. 33. 2 sermoneuultu. 1 The reasons for this interest, already in a different way reflected in the Gmnania, are probably complex. Something may be attributed to T.'s background and experience, or so we are now often asked to believe. But, whatever his background may have been, he could have found much to interest him, as a historian and a rhetorician, in peoples who still possessed the libertas which Rome had lost. 1 I am not, of course, thinking merely of the space these wars occupy in T.'s narrative, though it is considerable. The perennial troubles in and around Armenia also occupy a lot of space, but only occasionally do they obtain even a little of the colourful treatment regularly available for northern campaigns. ' Perhaps Domitian received comparable attention. 1

32

ASPECTS

OF TACITEAN

HISTORIOGRAPHY

the German problem once and for all. 1 And again, when T. records matters which are or may seem discreditable to Germanicus, such as his histrionics and evasion of responsibility during the mutiny 2 or a piece of sheer incompetence during the campaign of A.D. 16, he refrains from any direct and adverse comment. That he records such matters at all and that sometimes adverse comment finds an indirect way into his narrative may be taken to reflect his basic honesty in dealing with a complex tradition. Or, to look at the question another way, we may simply acknowledge that T.'s picture of Germanicus is not altogether self-consistent. 3 However that may be, it remains undeniable that T. often looks at Germanicus through rose-tinted spectacles, and that by so doing he somewhat misrepresents the history of the early years of Tiberius' principate. The explanation is, in my opinion, partly stylistic, partly psychological. From a stylistic point of view it is an immense advantage for T. to be able to set Germanicus in contrast with Tiberius, for this contrast is a very effective form of indirect characterization and contributes much to T.'s denigration of Tiberius as a man and as princeps.But T. is also emotionally involved. He is involved because he finds in Germanicus' actions and personality an embodiment of Roman uirtus, of admirable qualities commonly supposed to have existed in much earlier times, but not often to be found in the period which forms the subject of his history.' Hence idealization, nostalgia, and sentiment. 1

This highly favourable assessment is strictly no more than a report of what some people said at the time. So, in notable contrast with the direct and damning assessment of Tiberiw at Ann. 6. 5 I, T. may seem here to escape all responsibility for the views expressed. But he does not. After such a full and (in general at least) such a' committed' presentation of Germanicus we fairly expect a final summary and, when we get from T. a report of views which amount to such a summary, we no less fairly take him to endone them. 1 Sec on 1. 35. 4 tum . .. attinuissent and 1. 44. 3 nee Caesar. .. inuidia erat. • So I shall tentatively argue in my note on 1. 31-52. • In some respects T.'s Germanicus resembles his Agricola. Add to Agricola's sterling virtues youth, charm, high rank, poetic talent and 2

33

GTA

INTRODUCTION

These attitudes are particularly apparent in T.'s account of Germanicus' death and obsequies, but they also pervade much of Ann. 1-2, leading him to present the German campaigns in incongruously heroic colours 1 and impeding, if not indeed precluding, sober and objective judgement. This leads me back to T.'s conception of the moral purpose of history. T. is concerned not only to discover exemplaof good or bad conduct, but also to single out for special treatment persons who either embody certain virtues or vices or at least represent in their lives a mode of conduct of some general relevance. As to exempla,T. makes his attitude very clear when at Ann. 14. 64. 3, writing of the decrees of the Senate which followed Octavia's murder, he says that, distasteful though it is to do so, he will continue to record si quodsenatusconsultumadulationenouumaut patientiapostremum fuit. 2 And we may compare his remarks at Hist. 1. 3. 1 non tamen adeo uirtutum sterile saeculumut non et bona exemplaprodiderit. Closely allied to explicit mention of exemplaas such is the way in which T. passes judgement on men and their actions. So to do was a historian's recognized duty, and T. discharges his obligation both very fully and very variously. He naturally makes considerable use of direct comment and assessment, in character-sketches and obituaries, as also in those notable passages where he surveys the social or political conditions of a whole period. 3 No less importantly, he employs several oblique methods of evaluation, such as the

1

1

1

there you have a very attractive subject for a moralizing historian and rhetorician. It is almost beside the point that, on closer examination, Germanicus' character and achievements seem as contemptible as his verses. By T.'s own account we may see that these campaigns were to no small extent punitive and retaliatory, exercises of Schrecklichkeitrather than part of a grand, imperialistic enterprise. See further my note on 1. 52. I bellica. . . gloria. Cf., Ann. 3. 65. 1, already mentioned above. E.g. Hist. 1. 4-10, 2. 37-8, and Ann. 3. 55. On the last of these passages see my remarks in BICS 17 (1970), 101-6.

34

ASPECTS

OF TACITEAN

HISTORIOGRAPHY

reporting of rumours and of the alleged opinions of contemporaries. Thus he often produces judgements for which he takes no personal responsibility, an insidious and reprehensible procedure. As to the singling out of individuals who have, as it were, a lesson to teach, M. Lcpidus is an excellent case. No doubt he was a sufficiently important figure in Tiberius' times to have claimed mention more than once. But T. brings him forward very emphatically for discussion and commendation 1 precisely because he is exemplary, representing a mode of conduct necessary and laudable under the principate, namely pursuit of a uia media between intransigence and subservience. 2 For T. the conduct of the senatorial class, in adjustment to changed circumstances, forms a topic of recurrent interest: we may find in this interest a plausible explanation of the prominence given to various persons or groups. 3 It is now time to gather together some of these observations into another general proposition: T.'s selection of material is not invariably guided by its immediate historical importance, but may be influenced greatly by its exemplary value. In T. 's writings we encounter many 'stock' elements: they affect both situations and characterization. Consider, for instance, the manner in which he depicts the Germans and the Britons: the 'noble savage' is opposed to the degenerate Roman, and the opposition is the more effective because the virtues of the savage are similar to those associated traditionally with Rome's remote past. Hence a certain amount of moralizing, implicit or explicit. But in thus presenting 1

1

1

Ann. 4. 20. 2 hunc ego Lepidum temporibusillis grauem et sapientemuirum fuisse comperior. This presentation of Lcpidus is in some ways prefigured by that of Agricola, on which in general see U. Zuccarelli, Psicologiae semantica di Tacito, Brescia 1967, 97ff. And indeed one suspects that T. is always alert to discover elsewhere those qualities which he had known and admired in his father-in-law. I think particularly of the so-called 'Stoic opposition'. See my survey of work on Tacitus, 6-7.

35

INTRODUCTION

Rome's barbarian adversaries T. is just as much a rhetorician, developing a pointed contrast for its own sake, as a moralist, concerned with what the contrast has to teach. 1 Amongst the standard situations we find in T. the most common and also most misleading is that of tyrant and victim, in treason trials particularly: T. evokes hatred for the tyrant and sympathy for the victim without distinction of the facts of each case. A similarly stereotyped situation is that of the step-mother (or step-grandmother) opposed to step-son (or step-grandson), as with Livia and the two princes, Gaius and Lucius, 2 or Agrippina the Younger and Britannicus. Here the influence of declamatory material and treatment may well be surmised, as in T.'s account of Nero's tortured conscience at Ann. 14. 1o. I. But, while rhetorical influences of several kinds, such as that of the declamations and of Hellenistic 'tragic history', 3 are by no means negligible, we should remember that T.'s conception of the edifying purpose of history is highly conducive to standardization and schematization, or, in other words, black and white contrasts. Virtue is to be exalted and an eternal brand of infamy attached to vice. This approach does not readily encourage balanced and impartial assessment. Fortunately, however, T. does not always follow it: in his Otho, for instance, we have a fair and most carefully judged appraisal.' Only a simpleton would suppose that T. himself endorses the eloquent condemnation of Roman imperialism which he puts in the mouth of Calgacus at Agr. 30-2. This speech shows T. at his best as a rhetorician. But occasionally he does seem to show unusual sympathy for the barbarians: at least his tribute toArminius at Ann. 2. 88. 2-3 sounds genuine. 1 Or with her step-grand-daughter, Agrippina the Elder: cf. Ann. I. 33. 3 nouercalibus Liuiae in Agrippinam stimulis. 1 We might have expected more influence on T. of tragedy, Greek and Roman, as well as of ' tragic history', than we in fact find. More such influence might be apparent if we had more evidence, but I doubt it. Certainly T. has the good sense and discrimination to exclude such crudities as Oct. 368ff: see CQn.s. 15 (1965), 319. ' Still a cynical critic may perhaps argue that a 'mixed' character like Otho (cf. Hist. 2. 50. 1) is a subject for rhetorical treatment no less inviting than any other. 1

ASPECTS

OF TACITEAN

HISTORIOGRAPHY

But, though many such exceptions may be found, we must continually be aware that T.'s characters may be types, both more and less than the historical reality, more in that they conform with a pattern and subserve an edifying purpose, less in that they lack the complexities and contradictions of real people. Tiberius, the greatest ofT.'s characters, was too commanding and complicated a personality to be fitted into any familiar type. 1 I now turn to examine this memorable characterization, 2 but first some general observations. Many scholars have held, not unjustifiably, that the Greeks and Romans had no clear idea of change and development of character, that the characters found in ancient literature are generally static. 3 Thus, though their moresmay be revealed to us gradually, their basic characteristics (indoles,ingenium) are fixed from the start. The stylized character-sketches not infrequent in the historians seem to reflect this view: an individual can hardly be sketched definitively at his first appearance if he is later to change. Again, certain alleged examples of development in character are highly questionable.4 On the other hand, there are examples of changed character in tragedy, history, and biography, to mention only three genres. What we rarely find is any attempt to present the process of change: characters simply become different. This inconsistency or lack of transition may be seen epitomized in the celebrated remarks with which 1

Not that there are no stock elements in T.'s Tiberius. Note particularly the idea of the torments of a tyrant's conscience at Ann. 6. 6. 1 Of course the subject is too vast for full discussion here. I shall revert to it many times in the commentary. 1 One might go further and argue that the possibility of an essential change of character was first recognized by the Christians. Of this I am not sure, but I do maintain that, for any full description of change in character, we must wait until the eighteenth-nineteenth-century novel. ' For instance, the contention that Aeneas changes into a wiser and better man in the course of the Aeneid, though primafaci~ attractive, fails on closer scrutiny.

37

INTRODUCTION

Suetonius divides his life of Caligula: hactenusquasi de principe, reliquaut de monstronarrandasunt.1 In a word, we have in ancient literature inconsistent characters, as well as static ones. A few ancient writers, amongst them T., have an occasional inkling of change and development of character. But these intermittent insights have very little practical effect: they may indeed glaringly contradict the picture which is presented as a whole, but they do not alter it. It is ironical that T. at least had at his disposal ideal means for a more subtle and perceptive delineation of character, since, while he often uses the Sallustian type of character-sketch, 2 he is also adept in the indirect form of characterization, whereby the individual personality is built up by stages, piece upon piece, often indeed by suggestion rather than any definite statements of opinion. 3 In the event this complex technique is merely a substitute for the direct approach, not a real alternative. Let us now consider how it is used with Tiberius. Tiberius played too large a part in the story T. narrates to admit of a formal characterization (and, in a sense, dismissal) at his first appearance. But T. is careful to establish the main traits of Tiberius' character as soon as possible: for this purpose he employs at Ann. 1. 4. 3-4 a report of what contemporaries thought about the likely successors of Augustus. And, once he has established these traits of character, he can keep them continually before us by frequent 1

8 1

Again, when we look at his life of Augustus, we see that he, like others, is unable to account adequately for the difference between the callous and blood-thirsty triumvir and the benevolent pater patriae of later years. E.g. Mucianus at Hist. 1. 10, Sejanus at Ann. 4. 1, and Poppaea at Ann. 13. 45. \Ve owe to D. M. Pippidi, in his Tacite et Tibbe, a fuller understanding of T.'s procedure, in particular of the way in which his own comments, his reports of the impressions of contemporaries, and the speeches which he creates or refurbishes combine to fix and then fill out the characters he wants to present.

ASPECTS

OF TACITEAN

HISTORIOGRAPHY

repetition of certain ideas, such as dissimulation and hypocrisy. He commits himself not to present a gradual deterioration of Tiberius' character by his initial insistence on Tiberius' innate vices. He needs to insist upon them if his presentation of Tiberius is to hold together and be credible as a whole. And so he depreciates and indeed blackens Tiberius from the start, not perhaps realizing that there may then be an inconsistency in his assertion that Tiberius' principate changed for the worse in A.D. 23 ( Tiberio mutati in deteriusprincipatus initium ille annus attulit, Ann. 4. 6. 1) and in his acceptance of the view, so monumentally recorded by him in the obituary at Ann. 6. 51, that Tiberius appeared to be different at different stages of his life (morum quoque tempora illi diuersa). At first sight T.'s acknowledgment of these differences might suggest that in fact he recognizes change in Tiberius' character. But he does not, as his concluding remarks at Ann. 6. 51. 3 show: in scelerasimul ac dedecoraprorupit,postquam remotopudore et metu suo tantum ingenioutebatur. The morum temporaare phases of Tiberius' conduct conditioned by external constraints and as seen by superficial observers. T. believes that Tiberius' personality is innately bad and, however effective its camouflage, essentially unchanged from the beginning of his life to the end. Thus the only development which T. accepts and which interests him is the progressive disclosure of what Tiberius was really like, as one after another all constraints are removed. 1 Such at least is the picture of Tiberius which T. presents in the whole of Ann. 1-6: as far as Tiberius himself is concerned, there is no really appreciable change after the beginning of Ann. 4. The keynote had been struck long 1

The idea of external constraint is very interesting. F. Klingner (Studien zur griechischenund romischenLiteratur, Zurich 1964, 658) has plausibly suggested that T. transferred it to Tiberius from a quite different context, the alleged decline of Roman morals after Carthage ceased to be a threat to Rome's security: Sall. Hirt. I. 12 M poJlquamremoto metu Punico,etc.

39

INTRODUCTION

before at 1. 4. 3 multaqueindiciasaeuitiae,quamquampremantur, erumpere.So T.'s characterization of Tiberius is in all essentials static. 1 This prevailing impression is only once contradicted, a little before the concluding obituary, in the last speech of Arruntius (6. 48. 2): cum Tiberiuspost tantam rerum experientiamui dominationisconuulsuset mutatussit. This notion ('power corrupts ... ') goes flatly against T.'s picture of Tiberius as a whole. Whether he saw the contradiction I doubt. Certainly, if he had accepted the opinion he gives to Arruntius, he would have been obliged to recast much of what he had written. The words of Arruntius, together with the observation upon Vespasian at Hist. 1. 50. 4 solusque omnium ante se principum in melius mutatus est, show that T. is indeed not entirely unacquainted with the idea of change in character. But the isolation of these passages, as against numerous references to innate qualities, indicates that the idea is but of slight importance in his thought. T. has often been criticized for concerning himself too much with the personalities of the emperors, too little with the affairs of the empire as a whole. 2 By so doing, it is suggested, he presents an unduly sombre and depressing picture of the period he records. There is a good deal of justice in this criticism, but it is in part at least unfair, and not merely for the obvious reason that the kind of history T.'s critics want was largely unknown in antiquity. 3 By the nature of the principate a very limited number of persons were of dominant influence on events for long periods: consequently not only their actions, but also their personalities, must claim Of course several flaws or inconsistencies in this characterization show through on the surface, and many more are latent. Consider, for instance, Ann. 1. 75. 2 erogandaeper Jumestapecuniaecupiens,quam uirtutnn diu retinuit cum ceterasexueret. 2 The criticism sometimes takes a slightly different form. T. is alleged to be concerned exclusively and anachronistically with Rome, and unable to see beyond its walls. This view is demonstrably untenable. a We should not, however, forget that in fact T. does sometimes show interest in provincial matters and in social history. 1

ASPECTS

OF TACITEAN

HISTORIOGRAPHY

the fullest attention of historians. And so, inevitably and rightly, characterization becomes much more important for T. than it had ever been for the writers of republican history. But T. does not concentrate on any individual, even Tiberius, to such an extent as to compromise the nature of his work, to change it from history into biography. The annalistic framework helps him to avoid this danger, along with his skill in varying his material. There is very little sign in his major writings of a specifically biographical technique. For instance, he seems generally uninterested in minor details of appearance and behaviour 1 or in the sort of brief characterizing anecdotes of which Suetonius is full. He is, however, much given to remarking upon the wider significance or exemplary interest of the behaviour he describes. In so doing he follows the historical tradition. One of the most notable aspects ofT.'s narrative technique is his interposition of general reflections, psychological, political, or moralizing, within the narrative itself.2 Indeed he quite often thus interposes comments in narrative which is vivid and fast moving or in tableaux, as if what interests him most is not the individual scene 'per se ', but its relevance generally to human conduct or historical causation. Occasionally, by comparing T.'s version of a particular episode with that of a writer who drew on the same source or sources, we may establish the probability that any generalizing comments he there offers are his own. 3 In making such comments T. is in no sense an innovator, but he does, I think, introduce them more freely than most earlier historians and he is certainly more ready to coalesce them with his narrative.' 1

He has a special reason for describing Tiberius' appearance (Ann. 4· 57. 2). 1 For some examples see on I. 33. 1 odiis. .. iniquae. 1 I have discussed a well known case (Hist. 3. 84. 4 and Suet. Vit. 16-17) in my survey of work on Tacitus, 27-8. ' Much the same applies, in the Annals at least, to his use of sententiae. In the Histories he tends to separate sentmtiae from the context, to the

41

INTRODUCTION

By eliciting from the events he narrates general lessons about human motivation and psychology T. elevates his history onto a philosophical plane, thus in one way fulfilling his aim to be instructive. Not unrelatedly, he several times questions or rejects superficial and easy explanations, preferring to look for more general and more latent causes. 1 In so far as T.'s conception of the moral purpose of history thus combines with a desire to probe deeply into causes and motives, its effect is beneficial, and we have here some compensation for the disadvantages which the moralizing approach in other ways entails. The foregoing observations may help to explain the strength of T.'s conviction that certain material is inappropriate for history: Nerone iterum L. Pisone consulibus pauca memoratudigna euenere,nisi cui libeat laudandisfundamentis et trabibus, quis molem amphitheatriapud campum Martis Caesar exstruxerat,uoluminaimplere, cum ex dignitatepopuli Romani repertum sit res inlustresannalibus,talia diurnis urbis actis mandare (Ann. 13. 31. 1). No doubt T. had an acute awareness of the dignity of history 2 and of its traditions, but on the whole neither traditionalism nor literary snobbishness has all that much to do with his choice of material in detail. The decisive consideration is the interest which any particular material may hold for him. 3 And, as has already been suggested, he will be specially interested if the material available offers opportunity for elaborate treatment or generalizing comment. From this point of view Nero's amphitheatre was a rather unpromising subject.

1 1

1

extent of making them clauses or sentences in their own right. In the Annals, though the thought is often enough sententious and the expression pointed, it is only occasionally that a sententia is clearly separated out. If my impression is right, this is an interesting change in technique. The matter deserves some further investigation. See, for instance, Hist. 2. 37-8 and Ann. 3. 55. See Hist. 2. 50. 2. For example, Nero's poems can hardly be included under res inlustres.But they interest T. and he has something to say about them (Ann. 14. 16. 1).

ASPECTS

OF TACITEAN

HISTORIOGRAPHY

T. does not always exclude matters of minor importance, unsusceptible of elaboration, for he is continually at pains to sustain his readers' interest by variety. Happily for his purposes the traditional annalistic framework gave him considerable room to manoeuvre, much more, for instance, than would have been afforded by distribution of material into categories, such as political, military, and constitutional history. Within the framework of the year he arranges his material fairly freely, interlacing dynastic affairs, frontier wars, provincial matters, constitutional changes, major and minor senatorial business, and so on. He does not, to be sure, ignore chronological sequence, but he seems to pay more attention to it at some places than at others. 1 The crowded and momentous events of A.D. 69-70 demanded, if they were to be intelligible, careful regard for chronology, particularly because important developments were occurring at the same time in different parts of the empire. T. makes real efforts here to produce a clear account of a complex situation, and on the whole he succeeds. 3 In the Annals, where events do not crowd upon each other so thick and fast, the need for chronological precision is less obviously compelling, and often, it must be admitted, we look for it in vain. Of course, if T. had attempted to adhere rigidly to chronological sequence, week by week and month by month, this would have occasioned a fragmentation of his narrative disastrous stylistically and also, from a historical viewpoint, have made it difficult for us to see the wood as well as the 1

1

Occasionally, I suspect, T.'s arrangement of material is determined by the merest chance: one topic happens in some way to suggest another. At Ann. 14. 41, for instance, we have a reference to the juridical powers of the praifectu.surbis, then in the next chapter an account of the murder of the then praefectu.surbis by one of his own slaves. Similarly at Ann. 14. 17 T. mentions in parenthesis that Livineius Regulus had been removed from the senate (quern motum senatu rettuli), then proceeds in the next chapter to tell how the same fate befell another man (motus senatu et Pedius Blaesu.s). The complicated campaigns in North Italy, narrated in Hist. 2, form a partial exception, being singularly hard to understand in detail.

43

INTRODUCTION

trees. Further, in many parts of his writings, it is doubtful whether T. had access to information about the exact time of the events he records. 1 But, when this much has been said in mitigation, we have still to confess that T. can often be deplorably imprecise, sometimes because, following the annalistic tradition, he simply reports that such and such things happened eo anno or eodem anno, sometimes because, even when he tries to relate events within a year one to another, he is inclined to do so in a vague and unhelpful way, using such phrases as sub idem tempus, per idem tempus, and illis diebus.Again, following the same tradition, T. often diverges from chronological sequence by grouping certain material at the end of each consular year :2 in particular minor senatorial business, prodigies, and obituaries tend to be found in this position. I can now advance yet another general proposition: the sequence of events as narrated by T. under each year does not necessarily correspond with their true historical sequence, but may be influenced by stylistic and other considerations. The annalistic framework is then no self-assumed straitjacket. Albeit at the expense of precision in regard to the time and interrelation of events, T. can usually find within it all the freedom he needs. But, when it becomes at all restrictive, or, in other words, conflicts with the requirements of style and rhetoric, T. readily breaks away from it, as at Ann. 3. 1-2 and 12. 31-40. At the end of the latter passage he offers an explanation: haec, quamquama duobuspro praetoribuspluresper annosgesta, coniunxi,ne diuisa haudperindead memo1

1

Or about their exact location. Perhaps T. was 'the most unmilitary of historians'. But that reproach cannot be founded on his failure to locate such places as mons Graupius or the scene of Boudicca's defeat. It is unreasonable to demand of him precise details of distance and direction in accounts of campaigns conducted in areas remote and at the time largely unexplored. For all this, however, his topography, like his chronology, is sometimes sadly unclear when clarity was attainable. For some examples of, and exceptions to, this regular grouping, see on 1. 53. 1 eodem anno.

44

ASPECTS

OF TACITEAN

HISTORIOGRAPHY

riam sui ualerent. By bringing together here the events of several years T. does indeed make the story more intelligible as a whole, but at the same time he again loses or abandons any chance of chronological accuracy. 1 Indeed, in reading T. even casually and certainly in trying to use him as a historical source, we repeatedly come up against imprecision, not only of the kinds already mentioned, but also in the absence of details or explanations of details or in failure to tie up loose ends. 2 No slight part of the task of a commentator on T. is to diagnose these deficiencies and estimate their gravity. Variety in T. is not, of course, merely a matter of the arrangement of material: it is also attained by diversity of treatment. This diversity is evident in narrative, in speeches, and in characterization. 3 And yet overall T. maintains 1

1 1

It is interesting that T. does not similarly treat the campaigns against Tacfarinas, for they invited such treatment. But they are divided up, according to annalistic structure, at Ann. 2. 52, 3. 20-1, 3. 73-4, and 4. 23-6. Probably, thw divided, they seemed to T. to provide apt means for effecting the variety badly needed in Ann. 1-6, where interest centres so much on Tiberius and the senate. There was less need for conscious use of variety in the later books, for it was already there in the more variow subject-matter. But there may also be a change of technique in the later books: certainly T. there diverges more freely from the annalistic framework. That the reigns of Claudius and Nero are less adaptable to it than Tiberius' is not the whole explanation. Nor were T.'s powers failing. It seems likely that he was increasingly sure of himself and so prepared, in structure as well as in vocabulary and syntax, to experiment further with the medium he had inherited. Some instances are given in my note on 1. 16-30. As good an example as any may be found in T.'s diverse presentation of the murders of three empresses, Mcssalina (Ann. 11. 37-8), Agrippina (Ann. 14. 5-8), and Octavia (Ann. 14. 62-4). T.'s account of Messalina's death is vivid and detailed, but icily cold: he is here an uninvolved narrator. The enormity of Messalina's behaviour seems to have quenched any sympathy he might have had. So there is here no pathos, no tension, no drama. It is otherwise with the death of Agrippina. T.'s treatment of it is highly dramatic, but on the whole stark and bare. What special colour there is goes into the background, into setting the scene. The central action unfolds speedily and inevitably. T. does not make the mistake of accumulating pathos

45

INTRO D UjCjTION

a self-consistent style. A cynic might perhaps assert that the subtle distinctions in expression which recent critics have found between various parts of T.'s narrative and between various speeches in it are figments of their pedantic fancy rather than results of T.'s art. I should assert in turn that, of all the writers of the Silver Age, T. possessed the surest and most subtle command of rhetoric, and that he, if anyone, had the skill to effect the variety which his subject-matter demanded without losing the homogeneity which his own exacting taste enjoined. In T.'s writings style and historical interpretation are inseparable one from the other. The style is part of the interpretation or a result of T.'s attempts to reconcile interpretation with the evidence. The considerable progress in Tacitean studies of the last twenty-five years or so (I think specially of the work of Klingner, Pippidi, and Syme) depends very largely on the recognition of this inseparability. around Agrippina, nor does he develop the horror of matricide with the lurid rhetoric the subject so easily invited. He sees that the plot is too tremendous to need much support or ornament. Octavia's death involves T. completely: he is carried away by emotion and seeks to carry his readers with him: Accordingly his narrative is loaded with pathos, though it is not specially dramatic. T. here attests his involvement by violently indignant comment. In each of these three cases T. seems to me successful in his treatment and to show a perfect sense for what the particular subject-matter required.

TEXT

P. CORNELi I T ACITI LIBER I AB EXCESSV DIVI AVGVSTI Vrbem Romam a principio reges habuere: libertatem et cpnsulatum L. Brutus instituit. dictaturae ad tempus sumebantur; neque decemuiralis potestas ultra biennium neque tribunorum militum consulare ius diu ualuit. non Cinnae, non Sullae longa dominatio; et Pompei Crassique potentia cito in Caesarem, Lepidi atque Antonii arma in Augustum cessere, qui cuncta discordiis ciuilibus fessa nomine principis sub imperium accepit. Sed ueteris populi Romani prospera uel aduersa claris scriptoribus memorata sunt, temporibusque Augusti dicendis non defuere decora ingenia, donec gliscente adulatione deterrerentur: Tiberii Gaique et Claudii ac Neronis res florentibus ipsis ob metum falsae, postquam occiderant recentibus odiis compositae sunt. inde consilium mihi pauca de Augusto et extrema tradere, mox Tiberii principatum et cetera, sine ira et studio, quorum causas procul habeo. 2. Postquam Bruto et Cassio caesis nulla iam publica arma, Pompeius apud Siciliam oppressus, exutoque Lepido, interfecto Antonio ne Iulianis quidem partibus nisi Caesar dux reliquus, posito triumuiri nomine consulem se ferens et ad tuendam plebem tribunicio iure co~tentum, ubi militem donis, populum annona, cunctos dulcedine otii pellexit, insurgere paulatim, munia senatus magistratuum legum in se trahere, nullo aduersante, cum ferocissimi per acies aut proscriptione cecidissent, ceteri nobilium, quanta 1.

5

10

15

20

25

1

2

3

1

in summo marginescriptumest alia atque cotkx ipse manu, sed certe uetusta,TACm autem manu recentiore LIBER 1, quod hie tkest, e subscriptione libri suppletur AB EXCESSV DIVI A VG. VRBEM RO MAM A PRINCIPIO REGES HABV£RE. LIBERTATEM ET CONSVLATVM. L. brutus e.q.s. M

P. CORNELn

49

! __

CORNELII

2

1

2

3

4

5

6

TACITI

quis seru1t10 promptior, opibus et honoribus extollerentur ac nouis ex rebus aucti tuta et praesentia quam uetera et periculosa mallent. neque prouinciae ilium rerum statum abnuebant, suspecto senatus populique imperio oh certamina potentium et auaritiam magistratuum, inualido legum auxilio quae ui ambitu postremo pecunia turbabantur. 3. Ceterum Augustus subsidia dominationi Claudium Marcellum, sororis filium, admodum adulescentem pontificatu et curuli aedilitate, M. Agrippam, ignobilem loco, bonum militia et uictoriae socium, geminatis consulatibus extulit, mox defuncto Marcello generum sumpsit; Tiberi um Neronem et Claudium Drusum priuignos imperatoriis nominibus auxit, integra etiam tum domo sua. nam genitos Agrippa Gaium ac Lucium in familiam Caesarum induxerat, necdum posita puerili praetexta principes iuuentutis appellari, destinari consules specie recusantis flagrantissime cupiuerat. ut Agrippa uita concessit, L. Caesarem euntem ad Hispanienses exercitus, Gaium remeantem Armenia et uulnere inualidum mors fato propera uel nouercae Liuiae dolus abstulit, Drusoque pridem extincto Nero solus e priuignis erat, illuc cuncta uergere: filius, collega imperii, consors tribuniciae potestatis adsumitur omnisque per exercitus ostentatur, non obscuris ut antea matris artibus, sed palam hortatu. nam senem Augustum deuinxerat adeo uti nepotem unicum, Agrippam Postumum, in insulam Planasiam proiecerit, rudcm sane bonarum artium et robore corporis stolide ferocem, nullius tamen flagitii conpertum. at hercule Germanicum, Druso ortum, octo apud Rhenum legionibus inposuit adscirique per adoptionem a Tiberio iussit, quamquam esset in domo Tiberii filius iuuenis, sed quo pluribus munimentis insisteret. Bellum ea tempestate nullum nisi aduersus Germanos supererat, abolendae magis infamiae ob amissum cum ac Beroa/dus(Mmz.): at M: ut Horkel 13 tum Wolf: dum M 16 destinari Acidalius: destinare M 26 proiecerit Ritter: proieceret M: 31 munimentis Muretus et Lipsius: monumentis M proiceret Beroaldus 2

50

5

10

15

20

25

3o

AB EXCESSV

DIVI

A VGVSTI

1. 2-5

Quintilio Varo exercitum quam cupidine proferendi imperil aut txoVTas 61tµeAAE Kal 61fiyev,61Tc.>S µri cp86:ocwresTI veoxµwoeuow EA1T{61 TOV Kal e8EAOVO'IOV

611µ~x1Aias µvpfa6as (i.e.40,000,000), (vi) Suet. Tib. 57. 2 refers to the legataquaeplebei reliquisset(sc. Augustus)and Dio 5 7. 14· 1 says Tc';> 6{}µ~ TO:K6tVTa ViTOTOVAvyou fpy~ h18aO'EVCffil ( neeideoirameiusleniuit)I a>.:>.a 1TOAAO'. KCXI6e1va,rpomx8wv µeTO'. TCXVTa hrcrnecrq,ayri. Kai yap Kai 'TT)V ywai1