1. These present special problems of genre, being neither history in its technical sense (a commentarius was a set of notes, generally intended to provide a basis for a more elaborate narrative) nor in fact – at least in so far as the books by Caesar himself are concerned – proper commentarii: no one dared elaborate them, and both Hirtius, author of De Bello Gallico 8, and Cicero (Brut. 262), declared them perfect just as they were. For an introduction to the problems see Eden, P. T., ‘Caesar’s Style: intelligence versus inheritance’, Giotta 40 (1962), 74–117 Google Scholar; Gotoff, H. C., ‘Towards a practical criticism of Caesar’s prose style’, ICS 9 (1984), 1–18 Google Scholar; Conte (1994), 225–33 and Hammond, C., Caesar: Seven Commentaries on the Gallic War (Oxford, 1996), xi–xlvi.Google Scholar

2. The dates are traditional but probably roughly accurate (Syme (1964), 13–14); we do not know exactly when each monograph was published, though they predate the Historiae.

3. Sallust may have planned to extend it to 60 (Syme (1964), 191–2), or even as far as 40 (RICH, 117); Syme (1964), 179–80 and Rawson (1991), 546–69 have shown that there was a considerable ‘archaeology’ at the beginning that covered – perhaps in quite a sensational manner – the later career of the dictator Sulla. On the archaeologies in the monographs see below, n. 71.

4. Syme (1964), 256. Cf. Earl (1965), 237: ‘Sallust’s aim was ... an exercise in morbid pathology’; Newbold, R. F. determines that Sallust probably had a high level of diffuse anxiety (‘Patterns of anxiety in Sallust, Suetonius, and Procopius,’ AHB 4 (1991), 48 Google Scholar). The prefaces have occasioned critics some difficulties at least since the time of Quintilian (first century A.D.; his criticism is at 3.8.9): on their form see below, n. 24 and Appendix.

5. BC 3.1; BJ 4.1-4; see further above, p. 14 and below, p. 52 on history as a replacement for political action.

6. Macleod, C., Collected Essays (Oxford, 1983), 68.Google Scholar

7. On his political moderation see Syme (1964), 116–17, 253–4, Conte (1994), 237–8.

8. The pervasive idea that history is both practically and morally useful goes back at least as far as Thucydides. One of his precursors makes a programmatic statement that well anticipates Sallust: nam neque alacriores ad rem publicam defendundam neque segniores ad rem perperam faciundam annales libri commouere quosquam possunt (Sempronius Asellio 2P, ‘For annals can inspire no one either to be more eager to defend the state or to be slower to act wrongly’): the implication is that Asellio’s own analytical history, which would explain not only what happened but also why, could do so. See also Oakley (1997), 73–4, and for a negative view of Sallust’s ‘harping’ on the theme of morality and emphasis on the ‘personalities of history’ see Laistner, M. L. W., The Greater Roman Historians (Berkeley, 1947), 52–5.Google Scholar

9. nam saepe ego audiui Q. Maxumum, P. Scipionem, (alios) praeterea ciuitatis nostrae praeclaros uiros solitos ita dicere, quom maiorum imagines intuerentur, uehementissume sibi animum ad uirtutem adcendi. scilicet non ceram illam neque figuram tantam uim in sese habere, sed memoria rerum gestarum earn flammam egregiis uiris in pectore crescere neque prius sedari quam uirtus eorum famam atque gloriom adaequauerit (BJ 4.5-6). See Scanlon (1987), 17, 49–50, 80 n.29 and Paul (1984) ad loc.

10. Earl (1961), 28.

11. Earl (1961), 31, general discussion 18–40; there is an overview of Roman uirtus (in German) by Eisenhut, W., Virtus Romana (Munich, 1973)Google Scholar. Marius’ speech at BJ 85 is the clearest statement of this kind of self-made uirtus – though Marius himself fails to live up to the ideal.

12. On the prefaces and the reader’s attention see Scanlon (as above, n. 9); on compromised uirtus see Syme (1964), 268–9 with Earl (1965), 237–8 and above, pp. 11, 27; on the syncrisis in the BC. see above, p. 19.

13. It has been extensively argued that Sallust’s pessimistic, critical mode, which Tacitus followed, was exceptional: RICH, e.g. 167–8.

14. On Coelius see bibliographical Appendix; for general remarks on monographs (‘a new trend in Roman historiography’: Forsythe (1994), 41) see Syme (1964), 57, Puccioni, G., II Problema della Monografia Storica Latina (1981), and Scanlon, (1987), 66-7Google Scholar. Conte (1994), 235–6 makes a plausible connection between Sallust’s choice of the monograph form and the ‘demand for short works in a refined style that had grown as the result of the neoteric experiment.’

15. Sketched at Leeman (1963), 87 (from whom come the descriptive tags quoted here) and Oakley (1997), 145. The third type, a ‘somewhat heavy and elaborate chancery style’ with frequent use of participles and indirect speech, will surface occasionally in Livy; it is especially noticeable in Caesar, and may derive ultimately from Polybius.

16. Goodyear, CHCL 2.269.

17. Gay (1974), 9, 189; for the ancient version of the same idea see especially Seneca, , Ep. 114 (translated in Russell, D. A. and Winterbottom, M., Ancient literary Criticism (Oxford, 1972), 363–7 Google Scholar). On the author’s persona see the fundamental discussion of Booth (²1983).

18. This is an old topic that has come under much recent discussion, e.g. Plass (1988) and Henderson, J., ‘Tacitus/The world in pieces,’ in Boyle, A. J., ed., The Imperial Muse: Flavian Epicist to Chudian (Victoria, 1990), 167–210 Google Scholar.

19. On Sallust and Thucydides see above, pp. 12, 16–18. Cic, Brut. 66 suggests an analogy between Cato (Sallust’s Roman model) and Thucydides: ‘this understandable but highly audacious connection largely determined the later course of Roman historiography’ (Leeman (1963), 72).

20. There are many discussions of Sallust’s style. Conte (1994), 241–2 is brief but good; McGushin (1977), 13–21 has much useful detail; the classic treatment is Syme (1964), 240–73.

21. For antithesis see McGushin (1977), 14, 17, Syme (1964), 265; for political jargon and the speeches’ self-subversion see Syme (1964), 162–3, 198–201, 255–6.

22. The title is not certain. Reynolds (OCT) and Horsfall (1981), 107 prefer a version of De Coniuratione Catilinae; for a defence of Bellum, the term used also by Quintilian (3.8.9), see RICH, 147 n. 1.

23. The extent of the preface is controversial; for the view taken here see e.g. McGushin (1977), 291–2. The sketch of Catiline (5.1-8) is stated explicitly to precede the narrative (4.5 initium narrandt); and the digressive formula at 5.9 (esp. supra repetere) suggests that 6–13 is not part of the main narrative either. It might be argued that the sketch of Catiline continues no further than 5.5 and that the main narrative begins with Hunc (5.6: ‘This man . . .’); but it seems odd to leave the main narrative almost immediately and to launch into a long retrospective digression at 6–13.

24. It has been argued by D. C. Earl that Sallust, influenced by Aristotelian works which were ‘rediscovered’ at Rome during his lifetime, was aiming at a consciously ‘philosophical’ preface (‘Prologue-form in ancient historiography’, ANRW 1.2.842-56 (1972)).

25. As the preface progresses, it becomes clear that silentio = both ‘without speaking’ and ‘without being spoken about’ ( Woodman, A. J., ‘A note on Sallust, Catilina 1.1’, CQ 23 (1973), 310)CrossRefGoogle Scholar.

26. The point, of course, is that auctor often = ‘author (of a literary work)’.

27. Diodorus is often thought to be echoing the Hellenistic historian Duris. Although there is no guarantee ofthat ( Sacks, K. S., Diodorus Siculus and the First Century (Princeton, 1990), 94–5)Google Scholar, as early as the fifth century B.C. it was debated whether, or to what extent, language can describe ‘reality’ (see e.g. Kerferd, G. B., The Sophistic Movement (Cambridge, 1981), 80–2)Google Scholar.

28. See e.g. Isoc. Pan. 13, Diod. 20.2.2, Liv. 6.20.8, Plin. Ep. 8.4.3; Diggle, J., ‘Facta dictis aequare’, PACA 17 (1983), 59–60 Google Scholar. Feeney (1994), however, translates the phrase as meaning ‘deeds have to be made equal to words’.

29. See e.g. Isoc. Plataicus 4.

30. Sallust’s manner of expressing his twofold difficulty evidently gave rise to criticism, against which he is defended in the second century A.D. by Gellius (4.15). For Tacitus’ allusion to this passage of Sallust see p. 93.

31. The insertion of the sketch at this point is not arbitrary but, by revealing that he has retired from politics, is designed to prove that he is now free from partisanship and can thus write fairly, i.e. truthfully. See further RICH, 73–4.

32. carptim alludes to Sallust’s chosen form of the monograph.

33. For the expulsion see Dio 40.63.4; Syme (1964), 33–4.

34. Dio 43.9.2; Syme (1964), 39.

35. Syme (1964), 270 also talks in terms of a ‘conversion’.

36. The usual explanation of the passage is that Sallust had originally attempted to write history before being diverted into politics, but this seems incredibly lame and is not supported by what he himself says at 3.3. For the notion that the historian performs what he describes see esp. Liv. 10.31.10-15, Hor. Odes 2.1.18 (with Nisbet-Hubbard’s п.).

37. See Rutherford, R. B., The Meditations of Marcus Aurelius: a Study (Oxford, 1989), 103–7 Google Scholar.

38. See esp. Renehan, R., ‘A traditional pattern of imitation in Sallust and his sources’, CP 71 (1976), 100–1 Google Scholar. It has recently been observed that in the ancient world describing oneself in terms of a famous predecessor ‘is to be more, not less, oneself’ (Gleason, M. W., Making Men: Sophists and Self-Presentation in Ancient Rome (Princeton, 1995), 154 Google Scholar). It is of course well known that Sallust draws upon many earlier authors in his prefaces: see Perrochat, P., Les Modèles grecs de Salluste (Paris, 1949)Google Scholar; for Plato in particular see also MacQueen, B. D., Plato’s Republic in the Monographs of Sallust (Chicago, 1981 Google Scholar).

39. igitur (‘Therefore’) makes it clear that the reference to truth is to be seen in terms of Sallust’s departure from public life (above, n. 31).

40. Syme (1964), 268–9 has said that ‘Sallust is in his own person a document of concentrated energy and controlled violence’, that he ‘betrays an insight verging on sympathy’ for those whose energy is ‘diverted into criminal paths’, and that he has a ‘keen interest in the psychology of ambition and violence’. Note also D’Huys, V., ‘How to describe violence in historical narrative’, Ane. Soc. 18 (1987), 211.Google Scholar

41. The sketch follows the standard pattern for describing someone’s life: see below, p. 103 and n. 85.

42. For Sallust as an imitator of Thucydides see Sen. Contr. 9.1.13, Veil. 36.2, Quint. 10.1.101; Perrochat (above, n. 38), Scanlon, T. F., The Influence of Thucydides on Sallust (Heidelberg, 1980)Google Scholar, and e.g. Keitel, E., ‘The influence of Thucydides 7.61-71 on Sallust Cat. 20–21’, CJ 82 (1987), 293–300 Google Scholar. The scope of the retrospective digression starting at BJ 5.4 is controversial (see Paul (1984), 23–4); for H. see above, p. 41 n. 3.

43. This statement anticipates that of the emperor Verus writing to his friend and historian, Fronto: ‘my achievements are of course as great as they are, whatever that is; but they will seem as great as you want them to seem’ (Fronto 2.196); see also Plin. Ep. 7.33.10 (below, p. 102), and Woodman-Martin (1996) on Tac. A. 3.65.1.

44. This point seems not to be contradicted by Sallust’s allegation that writers have exaggerated the achievements of Athens, an allegation which he uses as a foil to contrast the lack of writers to describe early Rome (8.5). It is this lack which he himself remedies in the present digression, and both here and in the main narrative it is taken for granted that he has not exaggerated (cf. 3.2, 4.3).

45. See Earl, (1961) and The Moral and Political Tradition of Rome (London, 1967)Google Scholar.

46. Cato 2P (- Cic. Planc. 66). Sallust’s familiarity with Cato’s work may be inferred from the frequency with which he was accused of pillaging its (by now archaic) vocabulary: see Suet. Aug. 86.3, Gramm. 15.2 with Raster (1995), Quint. 8.3.29. There is a full study of Sallust’s archaizing vocabulary in Lebek, W. D., Verba Prisca (Göttingen, 1970)CrossRefGoogle Scholar. On Sallust’s style see further above, pp. 11–13.

47. Letter to Pompey 3, Thuc. 15, 41; see RICH, 40–3.

48. See e.g. Earl (1961), 86, Syme (1964), 66–7 and n. 34.

49. For the various scandalous stories surrounding Catiline see Berry, D. H., Cicero Pro P. Sulla Oratio (Cambridge, 1996), 277–8 Google Scholar.

50. For such superlative expressions elsewhere see Woodman-Martin (1996) on Tac. A. 3.11.2 or Woodman (1983) on Veil. 71.1; they are esp. common in Thucydides and Livy. This is one of only two passages in BC. where Sallust refers to his own experience of events (the other is 48.9, but cf. also 53.6); the digression thereby introduced Occupies the precise centre of the monograph’ (Syme (1964), 68).

51. At 31.1 permoueri is evidently used for the simple moueri (which occurs in the same image at BJ 41.10) but facies (OLD 2a, 7) makes the metaphor clear (see e.g. Tac. A. 4.67.2 antequam Vesuuius mons ardes cens faciem loci uerteret). At 36.5 both uis (OLD 9b) and inuado (OLD 4a) are technical of disease. At 38.1 exagitare (OLD 1da) and incendere (OLD Id) allude to the common image of a torch (OLD fax 8a), which in this case has evidently been smouldering before being shaken into life. Sallust was well known for his metaphors (Suet. Gramm. 10.6 with Raster (1995) here and on 10.2); see also Skard, E., ‘Die Bildersprache des Sallust’, Serta Eitremiana (Oslo, 1942), 141–64 Google Scholar.

52. Granius Licinianus 36A-B (p. 59 Camozzi).

53. See also RICH, 124–7.

54. It is no doubt this oddity which has led some commentators to say that bene dicere is not to be taken with rei publicae.

55. At 3.2 it is implied that some historians will (have) criticize(d): this point, which is interestingly modelled on Pericles’ speech of praise at Thuc. 2.35.2 and seems designed as a foil for the more extensive remarks on praise which follow, presumably constitutes an implied defence of what Sallust himself is about to do in BC.

56. It is not unreasonable to assume that Sallust had read Cicero’s letter: see RICH, 149 n. 30.

57. See Syme (1964), 105–11 for Sallust’s treatment of Cicero and scholarly reaction thereto.

58. For similar inhibitions see Liv. 45.25.3, Tac. A. 15.63.3; in general Brock (1995), esp. 212–13.

59. See Innes, D. C., ‘Quo usque tandem patiemini?’, CQ 27 (1977), 468 CrossRefGoogle Scholar.

60. Embedded within this introduction is a brief summary of an earlier debate: see Heyworth, S. J. and Woodman, A. J., ‘ Sallust, Bellum Catilinae 50.3-5’, LCM 11 (1986), 11–12 Google Scholar (reading the transmitted dixerat at 50.4).

61. It is these two speeches which account for the extraordinarily high percentage of direct speech (almost 27%) in BC. (the other speeches are at 20.2-17, 33, 35 [a letter] and 58). This percentage is significantly higher than that for any other ancient historical work apart from that of Dionysius of Halicarnassus: this statistic seems appropriate in a monograph whose author argues for the importance of dicere as opposed to facere (pp. 13–17), though it is ironical that in the four major speeches the speaker is made to cast doubt on the effectiveness of words (20.15, 51.10, 52.35, 58.1). If the above percentage is added to the amount of space devoted to the preface (1-13), it emerges that almost half the monograph is given over to things other than a narrative of the actual events of the Catilinarian conspiracy.

62. For a survey of views see McGushin (1977), 309–11; add Batstone, W W., ‘The antithesis of virtue: Sallust’s synkrisis and the crisis of the late Republic’, CA 7 (1988), 1–29 Google Scholar.

63. See e.g. Wilkins, A. T., Villain or Hero: Sallust’s Portrayal of Catiline (New York, 1994), 29–70, 131–7Google Scholar, adducing other parallels in addition to those mentioned here.

64. Cicero in fact attacks Catiline for the very qualities which he shares with Caesar (Cat. 1.26 labores tui . . . uigilare, 3.16 uigilaret, Laboraret). It is noteworthy that, while Sallust condemns Catiline for aiming at monarchy (5.6 regnum), he nowhere associates monarchy with Caesar, the most famous and most recent individual to be charged with this aim.

65. Syme (1964), 58. See also Earl (1961), 85–6.

66. Earl (1961), 95; cf. Syme (1964), 68 ‘a fierce and fraudulent oration’.

67. For such activities see Woodman (1983) on Veil. 79.1, 85.4, (1977) on 114.1-3; Kraus (1994), 351 (Index). On good generals see also Eckstein, A. M., Moral Vision in the ‘Histories’ of Polybius (Berkeley, 1995), 161–93.Google Scholar

68. For the second-person subjunctive as a means of achieving vividness see [Long.] Subi. 26; Gilmartin, K., ‘A rhetorical figure in Latin historical style’, TAPA 105 (1975), 99–121 Google Scholar, drawing attention to the unusualness of the present passage (p. 116).

69. For the significance of frontal wounds see BJ 85.29; I. M. Le M. DuQuesnay in AA, 67 and n. 58.

70. See e.g. McGushin (1977), 287, who is nevertheless sceptical of the notion, or Scanlon (1987), 34–5.

71. It is possible that in thus selecting several beginnings he is imitating Polybius, whose ‘main’ history begins in 220 B.C. (1.3.2, 4.2.1) but who identifies other starting points at 387/6 (the beginning of Roman naval power: 1.6, 1.12) and 264–60 (the first time the Romans crossed by sea from Italy: 1.5). Both writers are ultimately influenced by the Thucydidean pattern of archaeology + brief history of the immediate past (the pentecontaetia) + main narrative; see also (above), nn. 3, 42, andFeeney (1994).

72. For the history of the idea see Paul (1984) on 41.2 and Harris, W. V., War and Imperialism in Republican Romey 327–70 B.C. (Oxford, 1985), 127–8, 266–7Google Scholar; for the Histtoriae see below, n. 103.

73. On the monograph form see also n. 14.

74. Though, paradoxically, Catiline is not present in Rome, nor in the narrative, for fully one half of the BC (he departs at 32.1, writes a letter at 34.2-35, but is not seen again until the battle of Pistoria, 56–61). For the ‘neat dramatic arc’ of the BC see Scanlon (1987), 63; for Jugurtha’s person see Levene (1992), 54.

75. For the considerable difficulties with Sallust’s chronologies see Syme (1964), 142–7; Paul (1984) on 37.3 and 43.1.

76. Domestic affairs are depicted primarily in what Scanlon calls ‘reaction narratives,’ scenes at Rome in which the populace and the nobles react to what has happened in Africa (Scanlon (1988), 143–4). These are much more frequent before Metellus takes command (20 OCT pages out of 37 as opposed to eight OCT pages out of 69), a pattern noticed by Earl (1961), 70.

77. For extended discussion see Levene (1992). The introduction of Caesar and Cato in the BC has a similar, though less obvious, effect.

78. The great German historian Mommsen, for example, called it a ‘lionhunt’; see also Paul (1984), 19–20 and 264–8.

79. The fundamental discussion is that of Scanlon (1988); see also Wiedemann, T., ‘Sallust’s Jugurtha: Concord, discord, and the digressions’, G&R 40 (1993), 48–57 Google Scholar.

80. I will argue this fully in ‘Jugurthine disorder,’ to appear in a collection of articles on historiography (Leiden, 1998); on Jugurtha the hero see also Classen, J.-M., ‘Sallust’s Jugurtha-Rebel or freedom fighter? On crossing crocodile-infested waters’, CW 86 (1993), 273-97Google Scholar.

81. There is a large bibliography on the structure of the BJ, much of it in German; for references see Syme (1964), 141 n. 4 and Scanlon (1988), 144–6.

82. On bribery in this period see Paul (1984), 261–3; on money and exchange as symptomatic of slipping categories see von Reden, S., ‘Deceptive readings: poetry and its value reconsidered’, CQ 45 (1995), 47–8 CrossRefGoogle Scholar and Barthes (1975), 40.

83. On the opposition of the noui to the nobiles in Sallust see Earl (1961), 32–5; for the larger issues involved see the dicussions of Wiseman (1971) and Brunt, P. A., ‘ Nobilitas and nouitas ’ JRS 72 (1982), 1–22 Google Scholar.

84. On Marius’ election and Sallust’s misleading report of his early career – the historian suggests that he had an easy time of it, whereas in fact his rise had been slow and difficult – see Earl (1965), 235 and Paul (1984) on 63.5; on Marius see now Evans, R. J., Gaius Marius: a Political Biography (Pretoria, 1994)Google Scholar.

85. On Marius’ reforms see Keppie, L., The Making of the Roman Army: from Republic to Empire (1984), 57–19 Google Scholar and Rich, J. W., ‘The supposed Roman manpower shortage of the later second century B.C.’, Historia 32 (1983), 323–30 Google Scholar; on his affinities with Jugurtha see Earl (1961), 75 and n. 1, Scanlon (1987), 49 and below, n. 93.

86. On Servius Tullius as a nouus homo see Wiseman (1971), 109; for Servius and Marius as exemplifying changes of fortune from humble to high, cf. Val. Max. 3.4 (‘Concerning those who have turned out to be famous though coming from a low station’: Servius is in §3) and 6.9.14 (‘that Marius, who was such a low inhabitant of Arpinum’). Marius was a conventional example of the power of fortuna: see Carney, T. F., A Biography of C. Marius (Assen, 1961), 5.Google Scholar

87. Sallust prefaces the character sketch by saying that Sisenna, in whose Historiae the dictator figured, had described his subject ‘with an insufficiently free voice,’ that is, not critically enough (95.2); Sallust himself suggests that he too would be deterred from writing about Sulla’s later career (‘for as to what he did later, I am unsure whether it would cause me more shame or more revulsion to recount it,’ 95.4). Sulla, of course, wrote his own history in the form of an autobiography, which Sallust used as a source; on Sulla in Sallust’s Historiae see above, n. 3.

88. Per idem tempus aduorsum Gallos ab ducibus nostris Q. Caepione et Cn. Mallio male pugnatum. quo metu Italia omnis contremuerat. illique et inde usque ad nostram memoriam Romani sic habuere, alia omnia uirtuti suae prona esse, cum Gallis pro salute, non pro gloria cenare, sed postquam Bellum in Numidia confectum et Iugurtham Romam uinctum adduci nuntiatum est, Marius consul absensfactus est et ex decreta prouincia Gallia, isque Kalendis Ianuariis magna gloria consul triumphauit. et ea tempestate spes atque opes ciuitatis in ilio sitae (114).

89. On the ending see Levene (1992), 54–5; on double closures see Fowler (1989), 98–100; more on closure below, n. 91.

90. Though Levene rightly emphasizes that we are told neither of Jugurtha’s presence in the triumph nor of his death.

91. On spes (‘hope, expectation’) in the BJ see Scanlon (1987), 37–61. On the narrative effect of the ending cf. Barthes (1975), 76: ‘Expectation thus becomes the basic condition for truth: truth, these narratives tell us, is what is at the end of expectation. This design . . . implies a return to order, for expectation is a disorder: disorder is supplementary, it is what is forever added on without solving anything, without finishing anything; order is complementary, it completes, fills up, saturates, and dismisses everything that risks adding on: truth is what completes, what closes.’ Jugurtha is one manifestation of disorder, which ultimately infects even the form of the narrative itself.

92. Syme (1964), 138, 268–9. While Sallust himself does not credit Jugurtha with uirtus, both Micipsa and Scipio are sure he has it (6.2, 7.2, 9.2-3, 10.2, 8), and Sallust does not contradict them. While Marius’ uirtus is nowhere vouched for by the authorial voice and is attested by less reliable sources – e.g. Marius himself (85.32) and his soldiers (92.2) – in mentioning characteristics that often accompany it (energy, integrity, etc.) the sketch at 63.2-3 implies his uirtus.

93. On Marius and Jugurtha at Numantia see Paul (1984), 5, 31–2. On Marius’ likeness to Jugurtha compare also 63.6 (Marius will rush headlong, praeceps) ~ 8.2 (Jugurtha warned against rushing praeceps); and 64.2 (Metellus warns Marius not to ‘seek from the Roman people that which was denied him by law’ – i.e. the consulate) ~ 8.2 (Scipio warns Jugurtha that ‘it is dangerous to buy from the few that which belongs to the many’ – i.e. power).

94. Because it is restricted in the narrative to the digressions – the parts of the text that are off the main road, as it were – Carthage does not cause the chaos that Jugurtha does, a composite creature fully participant in the narrative proper. On the effect produced by composite entities, beings in which normally distinct qualities are freely mingled or interchanged, cf. Barthes (1975), 215: ‘it is fatal, the text [Balzac’s Sarrasine] says, to remove the dividing line, the paradigmatic slash mark which permits meaning to function (the wall of the Antithesis), life to reproduce (the opposition of the sexes), property to be protected (rule of contract). In short the story represents . . . a generalized collapse of economies: the economy of language, usually protected by the separation of opposites, the economy of genders (the neuter must not lay claim to the human), the economy of the body (its parts cannot be interchanged, the sexes cannot be equivalent), the economy of money. . . . This catastrophic collapse always takes the same form: that of an unrestrained metonymy. By abolishing the paradigmatic barriers, this metonymy abolishes the power of legal substitution on which meaning is based: it is then no longer possible regularly to contrast opposites ... it is no longer possible to safeguard an order of just equivalence; in a word, it is no longer possible to represent, to make things representative, individuated, separate, assigned.’

95. Postea Phoenices .... Hipponem Hadrumeturn Leptim aliasque urbis in ora maritima condidere . . . nam de Cartilagine silere melius puto quant parum dicere, quoniam alio properare tempus monet (19.1–2). Hannibal and the Carthaginians are mentioned earlier, at the very start of the war narrative (5.4 dux Carthaginiensium Hannibal . . . uictis Carthaginiensibus): the whole narrative is thus under the sign of the destroyed Carthage. In a paper delivered at the Classical Association AGM (Nottingham, April 1996) Ellen O’Gorman investigated some further implications of the presence of Carthage in Sallust’s narrative, including the use of the city as a boundary.

96. Igiturad Catabathmon, qui locus Aegyptum ab Africa diuidit, secundo mari prima Cyrene est ... ac deinceps duae Syrtes interque eas Leptis, deinde Philaenon arae, quem locum Aegyptum uorsus finem imperi habuere Carthaginienses, post aliae Punicae urbes (19.3).

97. Ita cum potentia auaritia sine modo modestiaque inuadere, polluere et uastare omnia, nihil pensi neque sancii habere, quoad semet ipsa praecipitauit . . . moueri ciuitas et dissensio ciuilis quasi permixtio terrae oriri coepit (41.9-10). See also above, n. 51.

98. The digression itself marks the centre of one large-scale chiastically arranged structure extending from the battle of the Muthul up to a skirmish near Cirta (48-99: Scanlon (1988), 146–61; on the digression, 161–7). Just as Lepcis is introduced into the narrative by its request for help in maintaining its own political stability, so the story it triggers, forming the centre of a textual unit, appropriately tells of the imposition of form on chaos. On further structural implications and on the literary history of such anecdotes see Paul (1984) on 79.1.

99. The connection between the two is underscored by the fact that Jugurtha is the great-nephew of Masinissa, a Roman ally against Carthage during the war against Hannibal; Numidia’s loyalty had remained fixed since that conflict but all surety was removed by Masinissa’s death, probably in 148 B.C. (5.5). Even on the personal level of the Numidian royal house, then, connection with Carthage produces a reliable universe in which friends are friends and enemies, enemies.

100. Syme (1964), 179 (tempered optimism: Goodyear, CHCL 2.269); Brunt, P. A., ‘On historical fragments and epitomes’, CQ 30 (1980), 494 CrossRefGoogle Scholar with his remarks about the importance of distinguishing between direct quotation (rare) and paraphrase (common) of an ancient author, and on the way the accidents of preservation can materially skew our impressions of a work.

101. McGushin (1992), 5–10 gives an account of the transmission of the Historiae; for Sallust and Plutarch see Syme (1964), 178.

102. McGushin (1992), 14 detects a further, Thucydidean division of the campaigning season into winter and summer. Reconstruction of the shape of the whole is problematic. The most recent attempt (of McGushin) follows a strictly chronological scheme; in response see the reviews by Rich, J. W. in CR 43 (1993), 280–2 Google Scholar and 46 (1996), 250–1. On the ‘archaeology’ in Book 1 see above, n. 3.

103. We cannot say for certain, as Syme did, that no such digression on contemporary politics existed (Earl (1965), 236); ifit did, the point of view expressed may have been broader than in the monographs, as shown e.g. by his less optimistic view of the period before 146 which takes account of the Struggle of the Orders: see McGushin (1992), 74–84, esp. 78–9.

104. For discussion of these conventions see bibliographical Appendix and den Hengst, D., The Prefaces in the Historia Augusta (Amsterdam 1981)Google Scholar. See also above, p. 10 on the prologues to BC. and BJ. A discussion of the prefatory fragments by Scanlon, T. F. (‘Reflexivity and irony in the Proem to Sallust’s Historiae’) is forthcoming in Studies in Latin Literature and Roman History VII, ed. Deroux, C. (Brussels, 1997)Google Scholar.

105. Historiae res gestae, and annales may all refer to both title and contents; on titles see Horsfall (1981), 105–6. L. Calpurnius Piso Frugi (consul 133 B.C.) may have been the first to write annales in Latin (Forsythe (1994), 42).

106. On Sallust’s end point see above, n. 3; for similar gestures in opening sentences cf. Livy 6.1.3 ac deinceps, Tac. A. 1.1.3 et cetera. The monographs’ announced subjects are more neatly delineated (BC 4.3 de Catilinae coniuratione; BJ 5.1-2, bellum. . . quod populus Romanus cum Iugurtha rege Numidarum gessit), though the BJ deliberately inscribes itself into a larger context (above, p. 21).

107. The consuls, whose names are couched in the asyndetic form traditional in Roman dates, may be read not as individuals but as simply indicating Sallust’s starting-point: so the fourth-century scholar Donatus, who discussed this line in his commentary on Verg. Aen. 1.1, compares res and populus R., the subject and actor of the Historiae, with Vergil’s arma uirumque. Sallust is quoting Cato here, from the opening of his Origines (1P): si ques homines sunt quos delectat populi Romani gesta discribere.

108. On the kinship between epic and history see e.g. Wiseman (1979), 145–51, Fornara (1983) s.v. ‘Homer’ (Index) and Feeney, D. C., The Gods in Epic (Oxford, 1992), 42–5, 260–64Google Scholar; on Herodotus and Odysseus see Moles, J. L., ‘Herodotus warns the Athenians’, PLLS 9 (1996), 264–6 Google Scholar, and on others see Walbank (1972), 51–2 and Marincola, J., ‘Odysseus and the Historians’, Histos (23 October 1996)Google Scholar.

109. Sallust himself used a hexameter rhythm at BJ 5.1 (see Paul (1984) ad loc); other examples of rhythmical openings are Livy Praef. 1 Facturusne operae pretium sim (cf. Quint. 9.4.74), Augustus, RG 1.1 Rerum gestarum diui Augusti quibus orbem, Tacitus, A. 1.1.1 Vrbem Romam a principio reges habuere. Scholars have been divided on whether these rhythms are deliberate; see Moles (1993), 157 with n. 77.

110. On Livy’s persona see below, pp. 70–4. The basic treatment of writers’ struggles with their precursors is Bloom, H., The Anxiety of Influence (New York, 1973)Google Scholar; for classical authors see Thomas (1986).

111. Herkommer (1968) has 14 pages of examples, both Greek and Latin (137-51); on the claim see also above, p. 14. Tac. A. 1.1.3 quorum causas procul habeo is a quotation from Sallust: see Woodman, A. J., ‘The preface to Tacitus’ Annals: more Sallust?’, CQ 42 (1992), 567–8 CrossRefGoogle Scholar.

112. On character see Walker (1952), 204–43, and Peiling, C., ed., Characterization and Individuality in Greek Literature (Oxford, 1990)Google Scholar, esp. C. Gill on ‘The character-personality distinction’ (1-31); on character in ancient historiography and biography see Wallace-Hadrill, A., Suetonius (London, 1983), 8–22 Google Scholar (on the distinction between the two genres), 142–74. Some histories put less emphasis on the actions of individuals: Cato famously did not name many of his actors, presumably to underscore their importance as parts of the whole. On the longue durée, a term coined by F. Braudel to describe the passage of geographical time, a stretch over which long-term historical tendencies and developments can be observed, see Burke, P., ‘Fernand Braudel’ in Cannon, J., ed., The Historian at Work (London, 1980), 192 Google Scholar.

113. The distinction was originally formulated by I. Bruns in 1898; independent and more accessible is Rimmon-Kenan, S., Narrative Fiction: Contemporary Poetics (London, 1983), 59–70 CrossRefGoogle Scholar (with examples); on characterization through appearance, etc. see Barton, T., ‘The inventio of Nero: Suetonius,’ in Eisner, J. and Masters, J., edd., Reflections of Nero: Culture, History & Representation (London, 1994), 57 Google Scholar. For an overview of character (s) in the Historiae see Syme (1964), 195–6, 207–13.

114. Comprehensive discussion by Pomeroy (1991).

115. in ulteriorem Hispaniam . . . regressus magna gloria concurrentium undique . . . uisebatur. . . . eum . . . ultra Romanum ac mortalium etiam morem curabant, exornatis aedibus per aulaea et insignia, scaenisque ad ostentationem histrionum fabricatis; simul croco sparsa humus et alia in modum templi celeberrimi, praeterea tum sedenti transenna demissum Victoriae simulacrum cum machinato strepitu tonitruum coronam capiti imponebat, tum uenienti ture quasi deo supplicabatur. toga picta plerumque amiculo erat accumbenti, epulae uero quaesitissimae, neque per omnem modo prouinciam, sed trans maria ex Mauretania uolucrum et ferarum incognita antea plura genera, quis rebus aliquantam partem gloriae dempserat, maxime apud ueteres et sanctos uiros superba illa, grauia, indigna Romano imperio aestimantis.

116. Gowers (1993), 18: ‘Writing against luxurious food and the superfluous desires of the body can now be explained as the most immediate and universally intelligible image of Rome’s expansion.’

117. On the fascination the Romans felt for forbidden displays see Barton, C. A., The Sorrows of the Ancient Romans (Princeton, 1993)Google Scholar.

118. The toga picta was the official garment of triumphing generals and worn only during the triumph or by special decree (Pompey was permitted by a decree of 63 to wear it in the Circus, a licence he exercised only once according to Veil. 40.4); on the connection between food and triumphs see Gowers (1993), 37, 39.

119. For general discussions see Plass (1988) and Sinclair, P., Tacitus the Sententious Historian (Pennsylvania, 1995)Google Scholar.

120. The Historiae also contained apparently straightforward judgments (though their epigrammatic twist might have been lost in transmission or excerption), both favourable (e.g. 2.37=36, possibly on Appius Claudius: ‘a serious man and in no pursuit (ars) inferior’) and not (e.g. 4.73=77, probably of Lucullus and echoing BC. 5.1 on Catiline: ‘out of control and overweening’).

121. The phrase ‘loaded alternative’ comes from Whitehead, D., Latomus 38 (1979), 474–95 Google Scholar; see now also Sinclair, P., ‘Rhetorical generalizations in Annales 1–6’, ANRW 2.33.4.2795-2831 (1991) and below, p. 116 n. 82Google Scholar. Rumours can be exploited for similar purposes (see Martin – Woodman (1994) on Tac. A. 4.10-11): the example from book 4 discussed below combines both techniques (the lack of any named source implies a rumour).

122. That the device was an essential element of historiographical characterization is seen from Cic. De orat. 2.63, who in discussing the need to explain motivation suggests a choice among casus, sapientia, and temeritas (accident, wisdom, or rashness).

123. Plass (1988), 62–4 on the unexpected; for the abrupt epigrammatic style cf. Sen. Ep. 114.

124. On the magis-quam and nisi epigrams see Kraus (1994) on Livy 6.10.9, 6.20.14; for snide use of alia cf. also H. 1.116=110.

125. McGushin (1992), 158–9; also the items in n. 113 above.

126. On these activities see Paul (1984) ad loc.

127. ‘Like Thucydides, Sallust uses speeches to provide a variation in narrative method; in each of his works speeches and letters . . . highlight the background and the atmosphere of important events and decisions. ... By using the words of men who play an important role in the events he is dealing with, [he] is able to present a deeper and more varied analysis of the political and social problems which form the essence of his theme’ (McGushin (1992), 14). On Dio Cassius see Gabba, E., ‘The historians and Augustus’, in Millar – Segal (1990), 70–1 Google Scholar.

128. Syme (1964), 200 holds that Sallust’s Cotta, the consul of 75, speaks in a way that ‘conforms admirably’ to the picture of him at Cic. Brut. 202: ‘a plain unadorned style . . . neat and intelligent.’

129. See in general Balsdon (1979), 161–92, esp. 182–5. Accommodating anti-Roman sentiment in Roman historiography can be seen as yet another gesture of imperialism, a form of what has been called ‘imperialist nostalgia’ (Miles (1995), 175).

130. Not an uncommon argument, especially from men in power claiming to have unpopular actions forced on them; in a Sallustian-influenced text cf., for example, Livy 6.15.5-6, 38.7, 40.6. Cogere, which also appears frequently in the context of literary patronage, has been described as ‘the mot juste for official encouragement’ (Brink, C. O., Horace on Poetry: Prolegomena to the Literary Epistles (Cambridge, 1963), 191 n. 3)Google Scholar; see Griffin, J., ‘Augustus and the poets’, in Millar – Segal (1990), 189–218 Google Scholar and now White, P., Promised Verse (Cambridge, Ma./London, 1993), 64–91 CrossRefGoogle Scholar with the reservations of Cameron, A., Callimachus and his Critics (Princeton, 1995), 454–93 Google Scholar.

131. For a contemporary equation of Caesar’s and Hannibal’s actions cf. Cic. Att. 7.11.1 ‘Cingulum’ inquit (‘nos tenemus) Anconem amisimus; Labienus discessit a Caesare.’ utrum de imperatore populi Romani an de Hannibale loquimur? It is a comparison that later writers, especially Lucan, make freely.

132. adeste Quintes, et bene iuuantibus diuis M. Aemilium consulem . . . sequimini ad recipiundam libertatem (1.55=48.27); sin libertas et uera magis placent, decernite digna nomine et augete ingenium uirisfortibus (1.77=67.20-1).

133. Sallust’s speeches also echo orations in other authors: a good example is Philippus, who borrows from Cicero’s Philippics (Syme (1964), 221–2). For more on intertextuality see below, p. 65 (and n. 59) and pp. 97–103 (and n. 47).

134. Syme (1964), 193.

135. For an over-schematic but useful introduction to ethnography in historiography see Fornara (1983), 12–16; on Herodotus see Hartog, F., The Mirror of Herodotus (Berkeley, 1988)Google Scholar, Lateiner, D., The Historical Method of Herodotus (Toronto, 1989), 145–62 CrossRefGoogle Scholar; on Homer and ethnography see Hall, E., Inventing the Barbarian (Oxford, 1989), 52–3 Google Scholar. One should emphasize that the ethnographical impulse is inseparable from the historical one: see esp. Immerwahr, H., Form and Thought in Herodotus (Cleveland 1966), 314–26 Google Scholar and Lowenthal, D., The Past is a Foreign Country (Cambridge, 1985)Google Scholar. There is a new Italian monograph on ethnography in Sallust: Oniga, R., Sallustio e l’Etnografia (Pisa, 1995)Google Scholar.

136. Cato: Astin, A. E., Cato the Censor (Oxford, 1978), 212–13, 216–18, 227–30Google Scholar; Sallust: Scanlon (1988), 138–43. On the importance to ethnographers of foreign institutions and ‘moral oddity or aberration’ (staples of Sallust on Rome) see Thomas (1982), 4.

137. That so many fragments from these sections have been preserved may testify only to the preoccupations of Sallust’s ancient readers: but it is equally possible that their abundance points to the historian’s own interests. (By comparison, there are very few traces of Livy’s ethnography of Carthage from Book 16, or of Gaul and Germany from Books 103–4.) Sallust may also have written ethnographies of Mesopotamia (4.61=62, 72=71, 78=75) and of Africa: Syme (1964), 193–5. On Rome and the barbarian see in general Balsdon (1979) and Dauge, Y., Le Barbare (Brussels, 1981)Google Scholar.

138. On this standard way of describing the shape of countries see Thomas (1982), 3 with n. 12.

139. The first writer known to us to have produced such a systematic description – intended to accompany his map of the inhabited world – was the Ionian Hecataeus at the turn of the fifth century B.C. (Pearson (1939), 17–19, 28–31, 34–96); on travellers’ tales and history see Gabba (1981), 60–2 and Fornara (1983), 13.

140. Hecataeus was again a pioneer, writing a genealogy of the gods and heroes that tied them to historical times and probably rationalized the myths (Pearson (1939), 96–106); see Fornara (1983), 4–12 on genealogy and history.

141. On nomadism/migration in ethnographies see Thomas (1982), 4, 23–4 and Shaw, B. D., ‘“Eaters of flesh, drinkers of milk”: the ancient Mediterranean ideology of the pastoral nomad’, in Rulers, Nomads, and Christians in Roman North Africa (Aldershot 1995), VI.5–31 Google Scholar (orig. published 1982–3).

142. Thomas (1982), 4–5; on ethnography as an introduction to conquered foreign lands see Wiseman (1979), 160 and n. 39; for the conventional features see Thomas (1982) s.w. ‘rivers, springs, etc.’ and ‘climate’ (Index).

143. si res eae quas gessimus orbis terrae regionibus definiuntur, cupere debemus, quo hominum nostrorum tela peruenerint, eodem gloriam famamque penetrare, quod cum ipsis populis de quorum rebus scribitur haec ampla sunt, turn eis certe qui de uita gloriae causa dimicant hoc maximum et periculorum incitamentum est et laborum (Arch, 23). On empire and narrative see Quint, D., Epic and Empire (Princeton, 1993)Google Scholar.

144. The desire for the strange or exotic can also serve imperialist urges, as at Tac. Agr. 10.4 incognitas . . . insulas . . . inuenit domuitque. dispecta est et Thule. . . . On the themes of escapism, utopias, and the miraculous see Gabba (1981), 53, 58–60 and Thomas (1982), 21–2.