1. For the dating of Pliny’s letters see Sherwin-White, A. N., The Letters of Pliny (Oxford, 1966), 36-8Google Scholar.

2. The preface is of course an integral part of his book or ‘work’ (opus), on the various nuances of which Tacitus is evidently playing.

3. The details of Tacitus’ early career are not quite certain, but it is known that he was praetor in A.D. 88, suffect consul in 97 and proconsul of Asia in 112/13. It has recently been suggested by Alfoldy, G. (MDAI(R) 102 (1995), 251-68)Google Scholar that a fragmentary funerary inscription from Rome (CIL 6.1574) is that of Tacitus: if so, he had had the distinction of being quaestor Augusti.

4. i.e. though the contemporary period may be ‘more productive’, his chosen period nevertheless has its own variety of ‘richness’. The transmitted opimum has often been questioned (e.g. by Syme in RP, 6.205-8), and Wellesley in his edition re-writes the opening of the sentence entirely (Tempus adgredior dirum casibus); but the metaphorical context (continued at 1.3.1 stenle) suggests that the word is correct.

5. See RICH, 165–7.

6. The standard work on the theme of civil war in Latin literature is Jal, P., La Guerre civile à Rome (Paris, 1963)Google Scholar.

7. Of which Tacitus’ work is very full: see Plass (1988), who is principally concerned with the political and other effects of wit. Yet many of Tacitus’ examples seem to have no particular purpose: see e.g. H. 2.22.3 Caecinae haud alienus (Caecina’s cognomen was Alienus) or A. 1.13.6 Tiberius casu an manibus eius impeditus prociderat (where casu is from the same root as prociderat, manibus interacts with the pes root of impeditus, and knees [genud] are mentioned earlier in the sentence).

8. For a survey of opinion see Syme (1958), 145 n. 5.

9. Syme (1958), 145.

10. For a recent discussion see Cole, T., YCS 29 (1992), 231–45 Google Scholar. (The volume is devoted to ‘Beginnings in classical literature’.)

11. Syme (1939), 507.

12. See Leeman, A. D., ‘Structure and meaning in the prologues of Tacitus’, YCS 23 (1973), 169–208 Google Scholar (at 188–9). Remarkably few scholars seem to have recognized Tacitus’ clear procedure.

13. E.g. Syme (1958), 367–71, 427, RP 3.939-40, 1027–8. For a recent discussion of the whole matter see Griffin, M., ‘Tacitus, Tiberius and the principate’, in Leaders and Masses in the Roman World (ed. Malkin, I. & Rubinsohn, Z. W., Leiden, 1995), esp. 33–43 Google Scholar.

14. We also lack most of A. 5, all of A. 7–10, and some of A. 11.

15. Comm. Zach. 3.14 ‘in thirty volumes he wrote down the lives of the Caesars after Augustus up to the death of Domitian’.

16. See pp. 183–4 of Wellesley’s edition of the Histories, where he also argues the likelihood that Historiae was the original title of the Histories. See also Goodyear (1972), 85–7 and, for a fascinating study of the titulature of ancient books, Horsfall (1981).

17. For the textual transmission of Tacitus’ works see Reynolds, L. D. (ed.), Texts and Transmission (Oxford, 1983), 406-11Google Scholar.

18. Goodyear (1970), 17–18.

19. See Syme (1958), 686–7.

20. Martin, R. H., ‘Tacitus and the death of Augustus’, CQ 5 (1955), 123-8CrossRefGoogle Scholar.

21. On the other hand, if Tacitus died before completing the Annals, it would follow from St Jerome that the Histories comprised fourteen books; yet the clear break after H 3 (see p. 96) perhaps suggests that a hexadic structure operated there too, in which case a total of fourteen books would be impossible.

22. It should be borne in mind that ‘the promise of future work(s)’ was a familiar literary motif (Woodman (1975), 287–8).

23. The contrast between Trajan and Domitian forms a major theme of Plin. Pan. (2-3, 53); likewise Velleius had greeted both Augustus’ principate by contrasting it with the previous age (89.2-6) and Tiberius’ principate by contrasting it with Augustus’ (126.2-5). It is virtually certain that long before Trajan’s time such contrasts had become standard as a method of praising the current ruler (cf. Plin. Ep. 3.13.2, 6.27.3). Tacitus capitalizes on this convention by reversing it at the start of Tiberius’ reign (A. 1.10.7 comparatione deterrima, ‘a very base comparison’; cf. Suet. Tib. 21.2, Dio 56.45.3). Pliny’s Panegyric and the issues it raises, many of them relevant to Tacitus also, are superbly discussed by Bartsch, S., Actors in the Audience (Cambridge, Ma./London, 1994), 148-87CrossRefGoogle Scholar.

24. This would be analogous to what is called ‘figured speech’, for which see the evidence assembled by Dyer, R. R., ‘Rhetoric and Intention in Cicero’s Pro Marcello’, JRS 80 (1990), 26–30 Google Scholar.

25. Syme (1958), 470–4, 826 (index ‘ANNALES, contemporary relevance’).

26. The famous saying of B. Croce: see Collingwood, R. G., The Idea of History (Oxford, 1946), 202 Google Scholar, Marwick, A., The Nature of History (London/Basingstoke, ³ 1989), 291-2CrossRefGoogle Scholar. The precise dates of the composition of the Annals are not known: for the suggestion that A. 4 belongs to A.D. 115 see Martin-Woodman (1997), 102–3 and n. 1 (contra Bowersock, G. W., ‘The Greek–Nabataean bilingual inscription at Ruwwâfa, Saudi Arabia’, in Le Monde grec: Hommages à Claire Préaux (ed. Bingen, J. et al., Brussels, 1975), 518-20Google Scholar).

27. Attacking the dead was a proverbially pointless exercise (see e.g. Plin. NH pref. 31). Braund, S. M. in her new commentary explains the apparent contradiction in terms of persona and as a variation on a traditional satiric motif (Juvenal: Satires 1 (Cambridge 1996), 116-21)Google Scholar.

28. Courtney, E., A Commentary on the Satires of Juvenal (London, 1980), 81 Google Scholar; see also Rudd, N., Themes in Roman Satire (London, 1986), 70–81 Google Scholar. Tacitus’ work constantly suggests comparison with Juvenal, though their relationship awaits further study. See, however, Syme, RP, 3.1135-57.

29. This whole subject is interestingly illustrated by Phaedrus, a writer of fables and contemporary with Tiberius, when he defends himself and his genre in the prologue to his Book 3. He first says that the genre was devised by slaves, who, ‘because they did not dare to say what they wished’ (cf. Tac. H. 1.1.4: above, p. 89), gave vent to their critical feelings through the allegory of fable (33-7). He then apologizes if a reader, because of a bad conscience, identifies himself with the faults portrayed in the fables: it is not the author’s intention to brand individuals (45-50).

30. On the digression see Martin-Woodman (1997), 169–76.

31. There is an excellent discussion of this episode by Cancik-Lindemaier, H. and Cancik, H., Der altsprachliche Unterricht 29 (1986), 4.16–35 Google Scholar.

32. For a useful survey of ‘closure’ (i.e. how works of literature, or sections thereof, end) see Fowler (1989).

33. See above, p. 62.

34. The standard and invaluable discussion of these matters in Ginsburg (1981), who treats only the Tiberian books. Her tabulation of the various types of narrative year (p. 54) perhaps requires some modification, but the point made in the text above remains true. The narrative structure of each of Tacitus’ works is analysed in great detail by Wille, G., Der Aufbau der Werke des Tacitus (Amsterdam, 1983)Google Scholar.

35. See Woodman-Martin (1996), 77–9 and 87–8. (Book 1 of the Annals is clearly a special case: see p. 91.)

36. It is A. 15 which begins prosaically, perhaps as a prelude to the subsequent reversal (see p. 110).

37. Note pp. 1574–5 of Martin, R. H., ‘Structure and interpretation in the “Annals” of Tacitus’, ANRW 2.33.2.1500-81 (1990)Google Scholar, a most valuable introduction to the content and organization of the surviving books; on the later books note too Classen, C. J., ‘Tacitus—historian between republic and principate’, Mnem. 41 (1988), 104-14CrossRefGoogle Scholar.

38. ‘Do historians . . . have to possess a special metaphorical capacity, a plastic or tactile imagination that can detect shapes and configurations where others less gifted see only jumble and confusion? In what ways is the sort of imagery any great historian chooses integrally related to his personality and his general outlook on the world?’ (Clive (1989), 200–1). On metaphor see further pp. 96 and 108.

39. Syme (1958), 531 and n. 1; in general see Pelling, TTT, 79–81.

40. See Woodman-Martin (1996), 11, 79, 96–7, 489–90.

41. Tacitus has placed Arminius’ obituary anachronistically in order to achieve this link (n. 72 below). For the beginning and ending of A. 4 see Martin-Woodman (1997) 18 and 262. The ending of A. 5 and beginning of A. 6 have not survived; for the ending of the latter see pp. 103–5. On freedom and related concepts in Tacitus see e.g. Morford, M., ‘How Tacitus defined liberty’, ANRW 2.33.5.3420-50 (1991)Google Scholar, and Vielberg, M., Pflichten, Werte, Ideale: eine Untersuchung zu den Wertvorstellungen des Tacitus (Stuttgart, 1987)Google Scholar.

42. See Heubner (1972), 151; also Fraenkel, E., ‘Rome and Greek culture’, Kleine Beiträge (Rome, 1964), 2.594 Google Scholar. There is a recent study of obituary notices by Pomeroy (1991), who discusses those of cities in Appendix 1 (pp. 255–7). For deaths in Tacitus see also Hutchinson, G. O., Latin Literature from Seneca to Juvenal (Oxford, 1993), 257-68Google Scholar; this book contains various sections on Tacitus (see esp. pp. 50–62).

43. Damrosch, L., Fictions of Reality in the Age of Hume and Johnson (Madison/London, 1989), 109 Google Scholar, on the body analogy in Gibbon and others; note also Béranger, J., Recherches sur l’aspect idéologique du principat (Basel, 1953), 218ff.Google Scholar, Nicolet, C., Space, Geography and Politics in the Early Roman Empire (Ann Arbor, 1991), 192 CrossRefGoogle Scholar and n. 9, and, on the symbolism of decapitation, Ash, R., ‘Severed Heads’, in Plutarch and his Intellectual World (ed. Mossman, J., London/Swansea, 1997), 196–200 Google Scholar.

44. See Kraus (1994a) and (1994), 83–8 respectively. On the significance of the City of Rome and its buildings etc. there is much of interest in Vasaly (1993), e.g. 26–39 (with refs. to Camillus’ speech in Livy).

45. There is a study of Vitellius’ death-scene in Levene, D. S., ‘Pity, fear and the historical audience: Tacitus on the fall of Vitellius’, The Passions in Roman Thought and Literature (ed. Braund, & Gill, , Cambridge, 1997)Google Scholar.

46. Pollock, F. & Maitland, F. W., The History of English Law Before the Time of Edward I (Cambridge, ² 1898), 1.1 Google Scholar.

47. ‘Intertextuality’ is a modish word, now about thirty years old, to denote the fact that there is a relationship between texts (for a bibliography on the subject see de Jong-Sullivan (1994), 284). The word has the advantage over more traditional terms, such as ‘imitation’ or ‘allusion’, in that it implies nothing about whether the relationship in any given case is intentional. Nevertheless in the following section, in which literary texts (such as that of Livy) and documentary texts (such as inscriptions) are deliberately discussed side by side, it is assumed that each case is in fact intentional on Tacitus’ part.

48. The third passage (diuersa omnium . . . ciuilium armorum fades) perhaps recalls Lucan, the epic poet of the civil war, in both theme (Luc. 1.1 Bella . . . plus quam ciuilid) and language (9.789 factes . . . diuersa).

49. See Pelling, TTT, 69, referring to Syme (1958), 375.

50. Keitel, E., ‘Principate and civil war in the Annals of Tacitus’, AJP 105 (1984), 306-25Google Scholar (at 325). Note esp. Sen. Clem. 1.1.1 for a similar theme.

51. See Woodman, A. J., ‘Self-imitation and the substance of history’, Creative Imitation and Latin Literature (ed. West, David and Woodman, Tony, Cambridge, 1979), 143-55CrossRefGoogle Scholar. Also below, n. 81.

52. ‘This extra-perception [afforded to a historian by his privileged position] allowed him to range backwards and forwards through time, alive to a mysterious unfolding destiny, which is given greater tragic significance by the literary device of anticipatory prolepsis’ (Tulloch (1988), 111 on Lord Acton).

53. For an excellent discussion of the meaningfulness of such intertextual character portrayals see Griffin, J., ‘The creation of characters in the Aeneid ,’ Latin Poets and Roman life (London, 1985), 183-97Google Scholar.

54. See J. Ginsburg, ‘In maiores certamina: past and present in the Annals’, TTT, 88–96; also Woodman-Martin (1996), 283–306.

55. See e.g. RP, 4.207ff.

56. See e.g. Momigliano, A., The Classical Foundations of Modern Historiography (Berkeley/Los Angeles/Oxford, 1990), 110ffGoogle Scholar.

57 Some scholars have detected another reference at A. 2.88.1 (see e.g. Goodyear (1981), 446) but the text is uncertain and the possibility seems remote.

58. There is a helpful edition and translation by P. A. Brunt and J. M. Moore (Oxford, 1967 and often reissued). Good remarks in Galinsky (1996), 10–20, 42–54.

59. This seems to have been pointed out first by Haverfield, F., ‘Four notes on Tacitus’, JRS 2 (1912), 197-9Google Scholar.

60. For the tone of this word see Ogilvie (²1970) on Liv. 2.56.10.

61. This inscription (CIL 13.1668 = ILS 212) is conveniently available in Smallwood, E. M., Documents Illustrating the Principates of Gaius, Claudius and Nero (Oxford, 1967, repr. 1984), no. 369 Google Scholar and is translated in Inscriptions of the Roman Empire A.D. 14–117 (LACTOR 8, 1971), no. 34 or Levick, B., The Government of the Roman Empire: A Sourcebook (London, 1985), no. 159 Google Scholar. For comparisons of the inscription with Tacitus see e.g. Wellesley, K., G&R 1 (1954), 13–33 Google Scholar, Miller, N. P., RhM 99 (1956), 304-15Google Scholar, Griffin, M. T., CQ 32 (1982), 404-18CrossRefGoogle Scholar, 40 (1990), 482–501, von Albrecht (1989), 136–59 (who has translations of each version). Note also Brock (1995), 209–11.

62. Last, D. M. and Ogilvie, R. M., Latomus 17 (1958), 476-87Google Scholar. For Claudius’ connection with Livy see Suet. CL•ud. 41.1 ‘in adolescence he set about writing history under the encouragement of Livy’.

63. A ‘window reference’ occurs when C, who is referring to B (who in turn is referring to A), himself refers to A in order to demonstrate his awareness of B’s procedure and perhaps also to correct it: see Thomas (1986), 188–9. Traces of Claudius’ speech reappear in earlier parts of the Annals (esp. 4.65): Syme (1958), 708–10 argued that Tacitus, even when writing about Tiberius, already knew of Claudius’ speech; G. B. Townend argued that this was unlikely and suggested that Tacitus had used as intermediary the historian Aufidius Bassus, who, being a contemporary of Claudius, would have had reason to refer complimentarily to the emperor’s speech in his own history of the Tiberian years (‘Claudius and the digressions in Tacitus’, RhM 105 (1962), 364–5).

64. EJ, no. 94a, Sherk (1988), pp. 67–71.

65. The inscription was published by Gonzalez, J., ZPE 55 (1984), 55–100 Google Scholar, and is translated in Sherk (1988), no. 36 (pp. 63–7). The two commentators are Koestermann (1963) and Goodyear (1981).

66. See Eck, W., Caballos, A. and Fernández, F., Das Senatus Consultum de Cn. Pisone Patre (Munich, 1996)Google Scholar.

67. See Woodman-Martin (1996), 114–18.

68. See Woodman-Martin (1996), 168–72.

69. On rumour in Tacitus see Shatzman, I., Latomus 33 (1974), 549-78Google Scholar.

70. See EJ, pp. 41 and 49.

71. See Woodman-Martin (1996), 68–9, 217–18.

72. For other cases of displacement see Ginsburg (1981), 21–2, 55–72, although in general she believes that Tacitus followed the chronological order of events (55). A good example is Arminius’ obituary at A. 2.88.2-3, which on Tacitus’ own evidence is seemingly placed two years before the man actually died (see Goodyear (1981), 447–8).

73. The words are those of Snotter, D. C. A., Tacitus: Annals IV (Warminster, 1989), 5 Google Scholar.

74. For this view see Woodman-Martin (1996), 70 and n. 1, who discuss the whole problem of dating on pp. 67–77.

75. Purcell (1993), 141; also above, p. 4.

76. For this inference see Woodman-Martin (1996), 70–1 and 114–16.

77. We should also take into account the fact that the speeches of some emperors, such as Tiberius (cf. A. 1.81.1, 2.63.3), were published. That Tacitus reserves a special vocabulary for his Tiberian speeches is argued by Syme (1958), 700–3 and Miller, N. P., ‘Tiberius speaks’, AJP 89 (1968), 1–19 Google Scholar; but whether such vocabulary is authentically Tiberian cannot be proved. On the general issue see Adams, J. N., ‘The vocabulary of the speeches in Tacitus’ historical works’, BICS 20 (1973), 124-44Google Scholar, whose position is more sceptical than those of Syme and Miller.

78. On this see Woodman, A. J., ‘Contemporary history in the classical world’, Contemporary History: Practice and Method (ed. Seldon, A., Oxford, 1988), 157-8Google Scholar.

79. ‘public record’ is a translation of publica acta, a phrase which recurs at Ep. 5.13.8 (and in Tacitus at A. 12.24.2) and is taken to refer to the acta diurna or ‘daily gazette’: this contained a wide variety of information ( Baldwin, B., ‘The acta diurna ’, Chiron 9 (1979), 189–203 Google Scholar) and is referred to by Tacitus himself at A. 3.3.2, 13.31.1, 16.22.3. It is arguably striking that Pliny makes no assumption that Tacitus will have read of the episode in the acta senatus.

80. See J. Ginsburg, ‘In maiores certamina: past and present in the Annals’, TTT, 96–103, and ‘Speech and allusion in Tacitus, Annals 3.49-51 and 14.48-49’, AJP 107 (1986), 525–41 respectively.

81. For a narrative example of intertextual inuentio see Martin-Woodman (1997), 206–7, where it is argued that A. 4.46-51 (war in Thrace) is indebted jointly to Sail. H. 2.87=69 and Caes. BG 7.69-90. Note also e.g. Keitel, E., ‘The function of the Livian reminiscences at Tacitus, Histories 4.58.6 and 62’, CJ 87 (1992), 327-37Google Scholar.

82. For Tacitean innuendo see Ryberg, I. S., TAPA 73 (1942), 383–404 Google Scholar, Sullivan, D., CJ 71 (1975/76), 312-36Google Scholar, Whitehead, D., Latomus 38 (1979), 474-95Google Scholar, Develin, R., Antichthon 17 (1983), 64–95 CrossRefGoogle Scholar.

83. In general on this topic see Walker (1952).

84. Syme (1958), 429. It has been claimed that ‘the tyrants depicted by Tacitus were all himself’ (Miguel de Unamuno, quoted by Gossman, L., Between History and Literature (Cambridge, Ma./London, 1990), 247 Google Scholar).

85. For these and other standard elements of biographical writing see Leo, F., Die griechischrömische Biographie (Leipzig, 1901), 180-2Google Scholar.

86. The ‘enslavement’ (seruiendum) is to be taken as Tiberius’, not as that of the Roman people collectively.

87. See e.g. Goodyear (1972), 37–40, Martin (1994), 105–6. Note also T. J. Luce, ‘Tacitus’ conception of historical change’, in PP, 152–7.

88. Gill, C., ‘The question of character-development: Plutarch and Tacitus’, CQ 33 (1983), 469-87CrossRefGoogle Scholar, esp. 481–7.

89. Hands, A. R., ‘ Postremo Suo tantum ingenio utebatur ’, CQ 24 (1974), 312-17, esp. 316–17CrossRefGoogle Scholar.

90. For the interpretation which follows see Woodman, A. J., ‘Tacitus’ obituary of Tiberius’, CQ 39 (1989), 197–205 CrossRefGoogle Scholar.

91. For Germanicus and Drusus jointly, described as the ‘New Dioscuri’ on an inscription from Ephesus (SEG 4.515), see Strabo 6.4.2 (for Germanicus alone note Suet. Tib. 25.3); for Livia see Dio 57.12.2-6; for Sejanus see Veil. 127.3, 128.4, Dio 57.19.7, 58.4.3.

92. Sejanus is not mentioned in the introductory sketch, for the simple reason that, when Germanicus and Drusus were alive, his rise to prominence could not have been foreseen by contemporaries.

93. That ingenio uti can have this meaning is shown by H. 1.90.2 ut in consiliis militiae Suetonio Paulino et Mario Celso, ita in rebus urbanis Galeri Trachali ingenio Othonem uti credebatur. If that is the correct meaning at A. 6.51.3, it follows that Tacitus is there making no reference at all to ‘fixed character’ and that his references elsewhere to Tiberius’ (dis)simulation are to be explained as characteristics of the typical tyrant figure.

94. For the interpretation which follows see Woodman, A. J., ‘A death in the first act’, PLLS 8 (1995), 257-73Google Scholar.

95. On the key significance of this moment see Woodman-Martin (1996), 428–9.

96. For the parallel careers of Germanicus and Drusus see Levick, B., ‘Drusus Caesar and the adoptions of A.D. 4’, Latomus 25 (1966), 239-44Google Scholar; Woodman-Martin (1996), 1–5. See also above, n. 91.

97. See Peiling, ‘Tacitus and Germanicus’, TTT, 59–85, esp. 78–81.

98. The moment is vividly captured in the newly discovered senatorial decree concerning Cn. Piso (above, n. 66), lines 126–30: ‘ [the senate] earnestly asked and sought that, with reference to the concern which he had once shared between his two sons, he should transfer it to the one he still had, and the senate hoped that he who survived would be a greater concern to the immortal gods to the extent that they understood better that all future hope for his father’s post on behalf of the state was placed in one man’.

99. See Martin-Woodman (1997), 90.

100. On the various devices which emphasize both the break at the start of A. 4 and the fact that A.D. 23 represents a turning-point see Martin-Woodman (1997), 14.

101. Heinrichs, A. D., Sejan und das Schicksal Roms in den Annalen des Tacitus (Diss. Marburg/Lahn, 1976)Google Scholar.

102. On Livia see Purcell, N., ‘Livia and the womanhood of Rome’, PCPS 32 (1986), 78–105 Google Scholar.

103. Cf. Liv. 21.43.4 Alpes urgent.

104. Cf. Lucr. 2.263-5 patefactis tempore puncto / carceribus non posse tarnen prorumpere equorum / uim cupidam (‘with the traps opened at a sudden moment, the eager crowd of horses still cannot charge ahead’). For Tacitus’ metaphors see e.g. Walker (1952), 62–6, Martin-Woodman (1997), 22.

105. See e.g. Goodyear (1970), 29, (1972), 27, Martin (1994), 126.

106. Modern texts place commas round the clause quod . . . reor (see quotation).

107. See Luce, T. J., ‘Tacitus on “History’s Highest Function”’, ANRW 2.33.4.2904-27, esp. 2907-14Google Scholar (1991).

108. For this motif see e.g. Herkommer (1968), 164–71.

109. For this interpretation see Woodman, A. J., Mus. Helv. 52 (1995), 111-26Google Scholar.

110. Pliny’s letter to Caninius Rufus, who plans to write an epic poem on Trajan’s Dacian war, is relevant to historiography (Ep. 8.4.2 ‘you will tell of new rivers discharging over the land, and new bridges thrown over the rivers’). Tacitus duly writes (A. 15.9.1): ‘Meanwhile Corbulo occupied the Euphrates’ bank, which he had never neglected, with more frequent garrisons j and, to prevent the enemy squadrons from impeding the throwing over of a bridge (for already they were flying across the surrounding plains in a great display), he deployed across the stream ships which were outstanding in size, connected by beams and augmented with turrets; and with catapults and ballistae he drove back the barbarians’ (complete with a full epic hexameter in the parenthesis: subiectis campis magna specie uolitabant).

111. There is an excellent analysis of this narrative by Gilmartin, K., ‘Corbulo’s campaigns in the East. An analysis of Tacitus’ account’, Histoia 22 (1973), 604-26Google Scholar. For the allusions to Book 9 of Livy (the Caudine Forks, also alluded to in the Histories: above, n. 81), see Woodman in AA, 184.

112. See A. J. Woodman, ‘Nero’s alien capital: Tacitus as paradoxographer’, in AA, 173–88. Paradoxography is a narrative sub-genre which tells of wonders, miracles etc.: see Gabba (1981).

113. See A. J. Woodman, ‘Amateur dramatics at the court of Nero’, in TTT, 104–28; for Nero as scaenicus imperator see Plin. Pan. 46.4. For comedy in the Claudian books (11 and 12) see Dickison, S. K., ‘Claudius: Saturnalicius Princeps’, Latomus 36 (1977), 634-47Google Scholar.

114. The standard work is Sörbom, G., Variano Sermonis Tacitei (Uppsala, 1935)Google Scholar; note also Martin, R. H., ‘Variatio and the development of Tacitus’ style’, Eranos 51 (1953), 89–96 Google Scholar.

115. This phenomenon is studied by Kohl, A., Der Satznachtrag bei Tacitus (Diss. Würzburg, 1959)Google Scholar, with special reference to H. 1, A. 1 and 13. Note also Chausserie-Laprée (1969).

116. For Tacitus’ syntax see Furneaux, H., The Annals of Tacitus (Oxford, ² 1896), 1.42–63 Google Scholar; for his language, Adams, J. N., ‘The language of the later books of Tacitus’ Annals ’, CQ 22 (1972), 350-73CrossRefGoogle Scholar. In general Syme (1958), 340–52 and 711–45, Goodyear, F. R. D., ‘Development of language and style in the Annals of Tacitus’, JRS 58 (1968), 22–31 Google Scholar.

117. Syme (1958), 342–4; for ‘misuse’, for which the technical terms are abusio or catachresis, see e.g. Woodman-Martin (1996), 132.

118. Reversal of names is not peculiar to Tacitus (see Goodyear (1972), 148) but is esp. characteristic of him.

119. Memoirs of the Prince de Talleyrand (ed. De Broglie, Duc, trans. De Beaufort, R. L., New York/London, 1891), 332 Google Scholar.

120. Walker (1952), 241 and n. 1, Martin (1994), 225.

121. See e.g. Gay (1974), 21–34 (esp. 25–6). Gibbon’s relationship to Tacitus has been qualified in various ways by Cartledge, P., ‘The “Tacitism” of Edward Gibbon’, Mediterranean Historical Review 4 (1989), 251-70CrossRefGoogle Scholar; see also the same author’s ‘Vindicating Gibbon’s good faith’, Hermathena 158 (1995), 133–47, which incorporates a discussion of some basic historiographical issues.

122. Clive (1987), 106, cf. 49. Tacitus himself said that ‘disparagement and malice are received favourably’ (H 1.1.2).

123. For a discussion of the relationship between the Dialogus and Tacitus’ historiography see Brink, C. O., ‘History in the “Dialogus de Oratoribus” and Tacitus the historian’, Hermes 121 (1993), 335-49Google Scholar.

124. Clive (1987), 252, quoting Macaulay’s sister, Margaret.

125. Tulloch (1988), 5–6.

126. Respectively Tulloch (1988), 95, Clive (1989), 83, Tulloch (1988), 81.

127. Tacitus refers to Augustus’ Res Gestae 8.5 ‘I restored many examples of our ancestors which were already fading from our period, and I myself handed on examples of many things to be imitated by posterity’.

128. i.e. whether or not posterity does imitate them, something which Tacitus naturally cannot foresee.

129. Syme (1958), 339, 565, 624 n. 3.