1. OLD memoria 7, 8; see above, p. 1.

2. Seneca (Ep. 100.9) says that Livy wrote dialogues that were ‘partway between history and philosophy’; scholars have tended to accept this testimony (so Ogilvie (²1970), 2, Walsh (1961), 4, 205) though no trace of these works has survived. He also is said to have written a ‘letter’ to his son giving advice on reading matter for the would-be orator (Quint. 10.1.39). Livy must have composed the Ab urbe condita at the rate (on average) of three books per year, starting some time in the mid-30s B.C. (he was born ca. 64), and continuing on till his death ca. A.D. 17. For the evidence and discussion of these dates see Luce (1965), 231 with n. 61.

3. Note also Moles (1993), 146 on pro uirili parte: ‘Livy’s historical project is itself exemplary of ‘virile’ individual public service’; on Livy’s borrowings from and challenges to Sallust in the preface see Moles (1993), 155–62 with further references; on Sallust’s history as politics see above, pp. 13–14.

4. On festinantibus see Moles (1993), 146–7: ‘Livy’s readers are readers “in a hurry”, hence the size of the res [Livy’s task] poses difficulties for readers as well as historian . . . the festinantes [have] an unhealthy preoccupation with the present or recent past . . . Livy will advocate . . . critical and slow reading commensurate with the size of his subject-matter – the past from the very beginning all the way down to the present.’ More on Livy’s persona above, pp. 70–4.

5. The motif is illustrated at RICH _s 133–4; see also Moles (1993), 151–2 on the remedia.

6. Praef. 12, the Romans are ‘in love with death both individual and collective’, de Selincourt’s translation of desiderium . . . pereundi perdendique omnia.

7. On time in historiography see Momigliano, A., Essays in Ancient and Modern Historiography (1977), 179–204 Google Scholar and Woodman-Martin (1996), 407; on time in Livy see Chaplin (1993) and Miles (1995), 75–109 (orig. published 1986).

8. On the Trojan legend and Rome see now Cornell (1995), 63–8.

9. The pace is not uniform; for a detailed chart see Stadter (1972), 304–7.

10. The expression is that of Badian (1966), 11–13.

11. This statement requires some qualification. The last 22 books (121-42), seem not to have been organized in the same way as 1–120 (Stadter (1972), 300), while the general scheme of publication in pentads is questioned by some (e.g. Levene, D. S., review of Kraus (1994), CR 46 (1996), 50 Google Scholar). Book 1 may have been separately published (see Luce (1965), 210 n. 2); and many accept the persuasive hypothesis that the first pentad was reissued in a second edition between 27 and 25 B.C. (below, n. 106). We have very broad outlines of the contents of the missing books (with the exception of Books 136–7) in late-antique summaries, or periochae; on them see Begbie, C. M., ‘The epitome of Livy’, CQ 17 (1967), 332–8 CrossRefGoogle Scholar. The history was transmitted in a manuscript tradition of groups often books (‘decades’).

12. RICH, 139 (following Syme).

13. For discussion of the ‘appendix’ and further references see Kraus (1994), 7–8. Diodorus Siculus (above, p. 8 n. 7) curtailed his original plan, ending with 60 rather than 46 B.C.

14. See especially Henderson, J., ‘Livy and the invention of history’, in Cameron, A., ed., History as Text (London, 1989), 66–85 Google Scholar.

15. On Homer as history see RICH, 1 with nn. 1–5 and above, p. 48 n. 108; whether or not the ancients drew a definitive line (as Fornara (1983), 4–10 argues) between legendary and historical time, they do speak of a natural boundary at least partly imposed by literary preservation: cf. Horace’s ‘there lived many great men before Agamemnon’ (but without Homer to write about them, they have effectively never existed: Odes 4.9.25-6). The equation of the city (urbs) with the world (orbis) is an Augustan concept with roots in late-republican thought: see now Muecke, F., ‘Horace’s Rome’, Classicum 21 (1995), 29–35 Google Scholar.

16. Sellar, W. C. and Yeatman, R. J., 1066 and All That (London, 1930), v Google Scholar. On the two temporal perspectives in Livy see further Jaeger (1997).

17. For elements of the translation I draw on the version at RICH, 130.

18. Moles (1993), 153.

19. So for example when at 6.1.2 Livy complains that the events he has just narrated in Books 1–5 were hard to see, we should think back to this sentence; his promise to go on and make things clariora, ‘more famous’ or ‘brighter,’ is a promise to make subsequent history easier for us to see (on the dual application of clarus to literary and physical visibility see OLD 2, 6, 8).

20. On exemplum in Livy see Chaplin (1993), 11–18.

21. On intueri see Kraus (1994), 171, Moles (1993), 166 nn. 61–3. On the importance for Livy of character see below, n. 44.

22. The technical terms for this literary clarity are enargeia (‘graphic presentation’) or euidentia: there is a good discussion at Walsh (1961), 181–90; see also RICH, 233 s.v. ‘vivid description’ (Index) and Vasaly (1993), 20 n. 6. On the audience feeling emotions see Winterbottom, M., ‘On impulse’, in Innes-Hine-Pelling (1995), 315 Google Scholar.

23. On documentum, which ‘always has a didactic component’, see Chaplin (1993), 17; on exemplarity see further Goldhill, S., ‘The failure of exemplarity’ in Jong-Sullivan, de (1994)Google Scholar. For more on the Roman preoccupation with images of the past see Bettini, M., ‘“The future at your back”: spatial representations of time in Latin’ in Anthropology and Roman Culture (Baltimore, 1991), 113–93 Google Scholar.

24. The narrator’s frequent interventions help this process (above, p. 73). The role of the reader and Livy’s text has received much attention recently; see e.g. Levene (1993), 29–30, Moles (1993), passim, Moles (1995), 8–74.

25. On the combination of pleasure (dulce) and usefulness (utile) in the Preface see Moles (1993), 141–2.

26. Chaplin (1993), 125–54; Miles (1995), 249 s.v. ‘T Livius and: historical change and continuity’ (Index); on repetition in Livy see further Kraus, C. S., ‘Repetition and empire in the Ab urbe condita’,Google Scholar forthcoming in a collection edd. C. Foss and P. E. Knox (Stuttgart).

27. On ancient education see Bonner, S. F., Education in Ancient Rome (Berkeley, 1977)Google Scholar; the primary ancient text is Quintilian, Institutio Oratoria.

28. Kraus (1994a), 268 with the references in n. 6; see also Vasaly (1993), 1–5 and the Introduction to Jaeger (1997).

29. 6.1.2, 3 res cum uetustate nimia obscuras uelut quae magno ex interuallo loci uix cernuntur . . . clariora deinceps certioraque ab secunda origine .... urbis gesta . . . exponentur; Rhet. Her. 3.32 tum nec nimis illustres [cf. Praef. 10] nec uehementer obscuros locos habere oportet, ne aut obcaecentur tenebris imagines aut splendore praefulgent. interualla locorum mediocria placet esse . . . nam ut aspectus item cogitatio minus ualet siue nimis procul remoueris siue uehementer prope admoueris id quod oportet uideri. On space, topography, and memory see further Jaeger, M. K., ‘ Custodia fidelis memoriae: Livy’s story of M. Manlius Capitolinus’, Latomus 52 (1993), 350–63 Google Scholar.

30. The monument to which it is most commonly compared is the Forum of Augustus, though that was not built until the end of the century: Luce, T. J., ‘Livy, Augustus, and the Forum Augustum’, in Raaflaub-Toher (1990), 123-38Google Scholar.

31. Vasaly (1993), 26–39; cf. also Sail. BJ 4.5, quoted above, p. 10 (portraits have the same inspirational value).

32. So e.g. Serres, M., Livy: the Book of Foundations (Berkeley, 1991), 140 Google Scholar; for a detailed study of this aspect of Livy see Jaeger (1997).

33. So the poet Martial (first century A.D.) claims that his library cannot hold all of ‘huge [ingens] Livy’ (14.190).

34. Cf. the chronological difficulties that arise when Livy describes different theatres of activity (Greece, Italy, etc.) as the empire expanded: on these, and on Livy’s method of coping with them, see Luce (1977), 33–138.

35. The second pentad has been analysed by Lipovsky, J., A Historiographical Study of Livy Books VI-X (Salem, N. H. 1981)Google Scholar.

36. Books 6–15 formed one division of the history (Stadter (1972), 293–4); but though the last Samnite war continued into book 11, Livy did in some ways also articulate 6–10 as a self-contained unit: for argument see Oakley (1997), 112–14.

37. So Sail. BJ 114.2 (quoted above, p. 26) with Paul (1984), 257: ‘fear of the Gauls was traditional at Rome.’

38. On that event, still the subject of much debate among historians, see Oakley (1997) on 6.34.

39. On Camillus see Ogilvie (²1970), 631, 741–3, Burck, E., ‘Die Gestalt des Camillus,’ in Wege zu Livius, ed. Burck, E. (Darmstadt, 1967), 310–28 Google Scholar, and Miles (1995), 110–36 (orig. published 1988). Camillus dies just inside Book 7 (1.8-10), which therefore also looks back, just as the close of 6 looks ahead (the full results of the elections held at 6.42.9 are not announced until 7.1.2). For the many responsions between Books 5 and 6 see Kraus (1994), 348 s.v. ‘Livy: Ab urbe condita 6: ~ Book 5’ (Index).

40. And perhaps playing with earlier texts: Claudius Quadrigarius began with the sack of Rome, where Livy ends a unit. See also below, pp. 88–97.

41. 21-30 may have been unique in their decade pattern (Oakley (1997), 111); see also Luce (1977), 27–8. The fundamental study of these books is Burck, E., Einführung in die dritte Dekade des Livius (Heidelberg, ² 1962)Google Scholar; see also his article, ‘The third decade’ in Dorey, T. A., ed., Livy (London, 1971), 21–46 Google Scholar and Levene (1993), 38–77.

42. ‘Most historians have prefaced their work by stressing the importance of the period they propose to deal with; and I may well, at this point, follow their example and declare that I am now about to tell the story of the most memorable war in history’ (21.1.1); the first unambiguous example is Thucydides 1.1.1, which Livy echoes here (see Walsh, R. G., ed., Livy: Book 21 (repr. Bristol, 1991), 121)Google Scholar. Some of the later books may similarly have been organized by war, though the wars do not correspond to textual decades; so e.g. the periochae for 109–116 are ‘subtitled’ qui est ciuilis belli primus (‘being the first book of the civil war’) etc., perhaps reflecting a pattern perceived by Livian readers: see Stadter (1972), 297–8.

43. Peripeteia is a term borrowed from criticism of drama and frequently used to indicate the turning points, both large and small, in Livy’s narrative; see Walsh (1961), 202, 210, 213 and (more generally) Chausserie-Laprée (1969), 541–648 on ‘Les techniques de rupture.’

44. For Livy’s concentration on people see e.g. Ogilvie, R. M. in CHCL 2.459–60 Google Scholar: ‘he saw history in terms of human personalities and representative individuals ... for Livy, human nature . . . determines the course of human events.’ There is extended, though now somewhat dated, discussion of the moral aspects of Livy’s characterization at Walsh (1961), 82–109. Preoccupation with human actors is found throughout Roman historiography (for Sallust see above, pp. 32–9), as well as in the idea, explicit in Cicero and Livy, that the eternal body of the state consists of the unending series of the mortal bodies and deeds of great men (Kraus (1994), 16–17 with n. 69).

45. These are the only examples of paired exhortations in Livy (Luce (1977), 27 n. 58); on the genre of pre-battle hortationes see the articles listed at Woodman-Martin (1996), 346, adding Ehrhardt, C. T. H. R., ‘Speeches before battle?’, Historia 44 (1995), 120–1 Google Scholar.

46. Livy may also discard suspense to warn us of impending disaster or pathos, or – as with the cavalry skirmish discussed below – to reassure us, a technique also notable in Vergil: examples at Kraus (1994), 166 on 12.11.

47. In fact it occurs only three times in the decade, at beginning, middle (25.2.6), and end. (29.22.10 is an interpolation: Oakley, S. P., CQ 42 (1992), 547–51 CrossRefGoogle Scholar.)

48. Also called ‘Einzelerzählungen’ or ‘individual narratives,’ these are one of Livy’s characteristic compositional units and the focus of much literary analysis since they were first identified (by Witte, K., ‘Über die Form der Darstellung in Livius’ Geschichtswerk’, RhM 65 (1910), 270–305, 359–419Google Scholar); see further Oakley (1997), 125–8. For a sample analysis of an episode see above, pp. 62–70.

49. For the first two see below; for the annalistic material, see above, pp. 3–4, 7, Levene (1993) on prodigies, and Kraus (1994), 9–13. The standard model is that of Ginsburg (1981), whose view of Livy is, however, oversimplified; for a challenge to it see next n. but one.

50. 2.1.1 Liben iam hinc populi Romani res pace belloque gestas, annuos magistratus . . . peragam, echoed at Tac. A. 1.1.1 libertatem et consulatum L. Brutus instituit (below, p. 94). See also Wiseman, T. P., Remus (Cambridge, 1995), 103.Google Scholar

51. ‘Structuring Roman history: the consular year and the Roman historical tradition’, Histos (23 October 1996).

52. For his play with traditional structures see e.g. Levene (1993), 38–42 (and often), Kraus (1994), 91 (on 6.1.8: reports of yearly elections).

53. The term is Oakley’s (1997), 128. His discussion (111-51) of Livy’s style, in which he treats nearly all of the elements discussed below, will become standard.

54. Livy’s style was once thought to have become progressively more ‘classical’ after the highly poetic ‘colouring’ of the first books; this view has been modified, and most scholars now see him as adapting his language to the situation. On the variety and evolution of his style see Oakley (1997), 146–7.

55. 36. 10–11. Other close readings of Livy include Moles (1993); Jumeau, R., ‘Remarques sur la structure de l’exposé livien’, RP 65 (1939), 21–43 Google Scholar; Jakel, W., ‘Satzbau und Stilmittel bei Livius’, Gymnasium 66 (1959), 302–17 Google Scholar; Luce, T. J., ‘Design and structure in Livy: 5.32-55’, TAPA 102 (1971), 265–302 Google Scholar; Solodow, J. B., ‘Livy and the story of Horatius, 1.24-26’, TAPA 109 (1979), 251–68 Google Scholar.

56. Text and translation as in Walsh (1990), 26–9.

57. The next explicit narrative signpost, marking the start of the next season, comes at 12.1, ‘at the beginning of spring’ (principio ueris). On Livy’s marking stages see Walsh (1961), 197–9; on chapter beginnings see Walsh (1990) on 12.1.

58. Examples at Kraus (1994), 101 on 6.2.14.

59. While the stratagem is conventional, Larisa’s involvement suggests the Caesarian text. For Livy’s use of and debt to Caesar see Oakley (1997), 129: ‘it is a fact that the sentence-structure of no two other Latin historians is so alike.’ For more complex examples of Livian intertextuality see above, n. 3 and Kraus (1994), 199–200 (on 6.18.5-15); for intertextuality in Tacitus see below, pp. 97–102.

60. Livy often moves the narrative ‘lens’ to follow a character; hence journeys, embassies, etc. become a means of making a transition from one part of the story to another. On the plot as a road see van Baak, J. J., The Place of Space in Narration (Amsterdam, 1983), 98, 118Google Scholar; on the transitional technique see Walsh (1961), 180–1.

61. For a general study of the word see Champeaux, J., ‘ Forte chez Tite-Live’, REL 45 (1967), 363–87 Google Scholar.

62. For how to use campfires to mislead the enemy see Frontinus, Strat. 1.5.3, 22, 24, 2.5.17; for not being fooled see Kraus (1994), 183 on 6.15.7.

63. A further rounding off is obtained by the frame with another praesidiums departure mentioned at 10.3: that was a Larisaean garrison that Antiochus released from Pherae (36.9.14-15).

64. The exception is the ut-clause at 10.4. On sieges in Livy see Walsh (1961), 191–7 and for an example just preceding our extract cf. 36.9.10-12. The capture of cities became the textbook example of a scene that cried out for the techniques of euidentia (above, n. 22; cf. Quint. 8.3.67-71): see Paul (1982) and Purcell, N., ‘On the sacking of Carthage and Corinth’, in Innes-Hine-Pelling (1995), 133–48 Google Scholar.

65. For focalization (roughly equivalent to the concept of ‘point of view’), a term coined by the French narratologist Gerard Genette, see Fowler, D., ‘Deviant focalisation in Virgil’s Aeneid ’ PCPS 36 (1990), 42–63 Google Scholar with references to earlier discussions.

66. Happy endings, such as we have here, produce a satisfying finale (as for instance in comedy); for emotions as a closural motif cf. also the last words of Sail. BC 61.9 ita uarie per omnem exercitum laetitia maeror, luctus atque gaudia agitabantur (‘And so variously throughout the entire army happiness and sadness, grief and joy were stirred up’) and see above, pp. 26–7.

67. This is an oft-told tale: references to its other occurrences are at Briscoe (1981), 235 on 11.1-4.

68. The idea that a king’s subjects resemble him is conventional; here it explains the wholesale corruption and forgetfulness that overtake Antiochus’ army (cf. also 5.28.4, 36.11.5 and the related motifs illustrated by Woodman (1977), 245 on Vell. 126.5).

69. For tense variation in historiography see Chausserie-Laprée (1969) especially 383–93 (on the combination perfect + imperfect, the reverse of the order we have here).

70. On the good general’s eating habits cf. Veil. 41.2 qui [Julius Caesar] . . . cibo in uitam non in uoluptatem uteretur and Woodman (1983), 53–4 ad loc. Such a general also refrains from sleeping more than is necessary, so Antiochus’ slumber doubly damns him. For Livy’s interest in moral qualities see above, n. 44 and Moore, T. J., Artistry and Ideology: Livy’s Vocabulary of Virtue (Frankfurt, 1989)Google Scholar.

71. The computerized Concordance to Livy by D. W. Packard (Cambridge, Ma., 1968) revolutionized the study of Livian diction; searches for this chapter have been done using it, the Packard Humanities Institute CD-ROM of Latin literature, and the TLL (below, n. 83).

72. So for instance [Longinus], On the Sublime 8.1 (figures, diction, word arrangement): see Innes, D., ‘Longinus, sublimity, and the low emotions’, in Innes-Hine-Pelling (1995), 322 and n. 1Google Scholar.

73. For an introduction to the question of linguistic decorum (the kind of language that is ‘fitting’ in a given situation) see Russell, D. A., Criticism in Antiquity (Berkeley, 1981), 129–47 Google Scholar, with 148–58 on ‘Classification of Literature.’ More on genre in Cairns, F., Generic Composition in Greek and Latin Poetry (Edinburgh, 1972)Google Scholar and Conte, G. B., The Rhetoric of Imitation: Genre and Poetic Memory in Virgil and other Latin Poets (Ithaca, 1986)Google Scholar; for a recent theoretical discussion see Beebee, T. O., The Ideology of Genre (Pennsylvania, 1994 Google Scholar).

74. There are some specialized studies of such language, in Livy and elsewhere: see Bruckmann, H., Die römischen Niederlagen im Geschichtswerk des T. Livius (Diss. Münster, 1938)Google Scholar on narratives of defeat; Wheeler, E. L., Stratagem and the Vocabulary of Military Trickery (Leiden, 1988)Google Scholar; and more generally Walsh (1961), 191–204, Oakley (1997), 83–4.

75. The style, which uses short sentences, asyndeton (lack of connective particles), and passive verbs (often without the auxiliary esse) is not illustrated in our passage. It was identified by Fraenkel, E. (‘Eine Form römischer Kriegsbulletins’, Eranos 54 (1956), 189–94)Google Scholar; see Oakley (1997), 139 n. 146.

76. On the various formulae see Chausserie-Laprée (1969), 17–124 and above, n. 57 on the carefully marked stages.

77. E.g. at 37.17.7, 44.13.4; cf. also Cic. Phil. 13.11, Sil. 2.303 (both using the related verbs).

78. The expression has a metaphorical content, as both words can be used of water (the Larisaeans are tossing on waves); for the metaphor cf. Juv. 12.32, the Elder Sen. C. 1.1.10, Sen. Medea 939, Tac. A. 2.23.2, Apul. Met. 5.21. Before Livy, who uses it more than any other author, the combination is attested only at Lucr. 4.1077.

79. For Vergil’s use see Clausen, W., ed., Virgil Eclogues (Oxford, 1994), 101 Google Scholar (on 3.38). The verb has isolated examples in Horace, Grattius, and Ovid, but is favoured by Livy (36x) and after him Curtius (27x), and by the technical writers Pliny the Elder (12x: natural science), Columella (lOx: agriculture), Seneca the Younger (17x: philosophy and natural science), Celsus (llx: medicine), and Justinian (20x: law). This distribution (only in poetry before and contemporaneously with Livy; infrequently so afterward) suggests that his use of a choice term was then picked up and ‘domesticated’ by some imperial prose writers.

80. See Kraus (1994), 226 on 6.22.9.

81. Accius 435 Ribbeck; it does not appear to occur elsewhere outside Livy except at Frontinus Strat. 1.5.19, a story taken from Livy. It may be that the historian has borrowed the tragic expression on the model of a more usual phrase such as fauces portus (‘the mouth of the harbour’, OLD fauces 3a), again producing language that, while reminiscent of ‘ordinary’ military style, is in fact quite extraordinary.

82. See Briscoe (1981), 4.

83. Before this passage and 42.12.4 it is found at Cat. 64.302, Verg. Aen. 7.555; it then occurs sporadically in imperial literature: see the citations at Thesaurus Linguae Latinae (Leipzig, 1900- ) III.4.743.64-71.

84. This sentence is imitated by Tacitus at H. 4.36.2 effusi in luxum et epulas et nocturnos coetus.

85. See Hofmann, J. B. and Szantyr, A., Lateinische Grammatik. Teil 2. Lateinische Syntax und Stilistik (Munich, 1965), 219 Google Scholar; it is found with names of towns, islands, and regions (as here).

86. Scholars have always seen different levels in Livy’s style, but his deliberate use of inconcinnity (lack of parallelism) in the Tacitean sense has been less commented on. See however Catterall, J. L., ‘Variety and inconcinnity of language in the first decade of Livy’, TAPA 49 (1938), 292–318 Google Scholar and Kraus (1994), 352 s.v. ‘uariatio’ (Index).

87. That the two elements are parallel is guaranteed by the pair of comparative expressions (latius quam pro ~ plures quam quot); that pairing, in turn, further emphasizes the inconcinnity of the verb forms. For parallels to our sentence commentators cite 32.14.2 (adortus . . . et ... cum erexisset, where editors often emend to adortus est); cf. the slightly different combinations at Tac. H. 2.34.2 (simulantes ... acne ... . terereť) ana A. 3.3.1 (rati . . . an ne ... intellegerentur), where see Woodman-Martin (1996), 90 ad loc. The combination of participle + cum-clause without the copulative particle is of course entirely regular.

88. One technical name for this device is traducilo (also used for the repetition of a word in connecting sentences; see OLD 5 and von Albrecht (1989), 95, 97): for further examples see Kraus (1994), 353 s.v. ‘word play’ (Index).

89. The term comes from Wilkinson, L. P., Golden Latin Artistry (Cambridge, 1963), 167 Google Scholar; his treatment of prose sentences (167-88) is perhaps the best and clearest available.

90. Had L. wanted to omit the first instance of the name, or combine the two clauses to remove the awkward use of the king in two different cases (ad regem . . . consultanti), he could easily have done so. This kind of parataxis is typical of early Latin prose; see von Albrecht (1989), 25 and 32.

91. The repetition of uel further elevates the tone; repetition of an initial word in sequential units either of verse or prose is called anaphora.

92. ‘The anaphora of nunc, . . . originally poetical, was introduced into prose by Livy’ (Martin-Woodman (1997), 213 on Tac. A. 4.51.1).

93. On this kind of arrangement, including discussion of shapes of tricola, see Kraus (1994), 21–4.

94. For ‘iconic’ language and Livy’s use of postponement see Kraus (1994), 347 s.v. ‘form ~ content’ and 353 s.v. ‘word order’ (Index), and Oakley (1997), 141–2.

95. For technical studies of speech in Livy see Ulimann, R., La Technique des discours dans Salluste, Tite Live, et Tacite (Oslo, 1927)Google Scholar and Étude sur le style des discours de Tite Live (Oslo, 1929); Lambert, A., Die indirekte Rede als Künstlerisches Stilmittel des Uvius (diss. Zürich, 1946)Google Scholar; and Dangel, J., La Phrase oratoire chez Tite-Live (Paris, 1982)Google Scholar; for more general discussions see Walsh (1961), 219–44 and Oakley (1997), 117–20, 139–41.

96. This sentence is singled out for criticism by Walsh (1990), 88 (‘a much less elegant period-sentence than Livy’s norm’); for other examples of ‘strained’ sentences in Livy see Oakley (1997), 132, quoting the lists made by the great nineteenth-century Livian scholars Madvig, J. N. (Kleine philologische Schuften (Leipzig, 1875), 359–60 Google Scholar) and Riemann, O. (Etudes sur la langue et la grammaire de Tite-Live (Paris, ² 1885), 309–10 Google Scholar).

97. The classic analysis of such periods is by Spilman, M., ‘Cumulative sentence building in Latin historical narrative’, UCPCPh 11 (1932), 153–247 Google Scholar.

98. For the phrase à relance and other narrative sentence-types see Chausserie-Laprée (1969), 129–336.

99. Though this last tricolon consists only of nouns, they form an action-sequence within themselves: banquets with wine are followed by unspecified uoluptates (a polite periphrasis for sex) which are in turn followed by exhaustion and sleep. They can, therefore, be regarded as semicontinuati ve.

100. Galinsky (1996), 280–7; the quotations are from 283 and 286. Cf. also Ogilvie, CHCL 2.466: ‘no other Roman historian was so inventive,’ a remark which may not have been meant to be entirely complimentary.

101. Galinsky (1996), 284. The writer who in many ways is closest to Livy, both in time and in narrative technique (in the Metamorphoses), is Ovid, though Livy has none of the poet’s pervasive iconoclastic cynicism; but the historian is nowhere near as humourless as he has been labelled (e.g. by Walsh, P. G., ‘Livy’s preface and the distortion of history’, AJP 76 (1955), 371–4 Google Scholar, Ogilvie (²1970), 4).

102. So for example Conte (1994), 370 (on the whole a surprisingly conservative and overstated treatment of Livy). The word frequently used to describe this overlap is ‘anachronistic’, but that misses the point.

103. Luce (1977), 230–97. He does not, however, seem to have thought that early Rome might have been radically different either in institutions or in ideals: see further the references at Kraus (1994), 28 n. 119.

104. On the levels of audience and spectacle in the Ab urbe condita see Feldherr, A. M., Spectacle and Society in Livy’s ‘History’ (Diss. Berkeley, 1991)Google Scholar; on internal deployment of exempla see Chaplin (1993).

105. He is also mentioned at 28.12.12, and may have been quoted in Book 59 (perioch. 59): Luce (1965), 209–10.

106. He probably inserted the digression in a second edition of the pentad: so Luce (1965), following a suggestion of Jean Bayet. For a recent discussion of 4.20, followed here, see Miles (1995), 40–7; for a new interpretation see Rich, J. W., Chiron 26 (1996), 85–127 Google Scholar.

107. Miles (1995), 46.

108. Cf. the entertaining anecdote related by Suetonius about Pomponius Porcellus, the ‘extraordinarily obnoxious overseer of the Latin language’, who on one occasion objected to the emperor Tiberius’ unusual diction on the grounds that while Tiberius could give citizenship to people he had not the power to do so to words (see Raster (1995), 26–7, 226 on Gramm. 22.2).

109. There are very few cases of Augustus’ censoring or punishing writers. One, Cassius Severus, was banished and his writings burned for ‘unrestrained hostilities’ (i.e. libel, apparently not directed against Augustus or his family but against leading senators); Ovid was exiled (presumably not primarily for his writing but for the mysterious error, possibly involvement with the princeps’ adulterous granddaughter) but continued to publish; Timagenes, a Greek historian, was banished from Augustus’ house but continued both to live in Rome and to attack the emperor. The older image of Augustus as repressive and tyrannous in his later years cannot be right: see Raaflaub, K. A. and Samons, L. J. III, ‘Opposition to Augustus’, in Raaflaub-Toher (1993), 417–54 Google Scholar.

110. On his detachment see e.g. Conte (1994), 370, Galinsky (1996), 286 (‘distance and independence of mind’), and see above, p. 10 on Sallust. Many of the ideas that Livy shared with Augustus had strong republican roots: see Kraus (1994), 8–9 and most recently Galinsky (1996), 284 (‘the values that Augustus sought to restore were, after all, traditional ones. Livy, in his own way, was just as zealous in this regard as the princeps’) and 285 (‘Like so many other Italians and Romans, Livy saw his own program take shape under Augustus . . . [who] actualized in the public aspirations that had been dormant’). For a judicious review of the problem of Livy and Augustus see Badian, E. in Livius: Aspekte seines Werkes, ed. Schuller, W. (Konstanz, 1993), 9–38 Google Scholar.

111. He is interested above all in the exercise of power and to that end concentrates not only on the res gestae of the Roman military and political aristocracy but on the plebeian crowds as well; in this respect his history may have been quite radical; see Ridley, R. T., ‘Patavinitas among the patricians? Livy and the conflict of the orders’, in Eder, W., ed., Staat und Staatlichkeit in der frühen römischen Republik (Stuttgart, 1990), 132–3 Google Scholar.

112. He ‘remains the most nebulous figure of all the greater historians of the ancient world’ (Walsh (1961), 1). He came from Padua in northern Italy, had at least one son, and worked in Rome, where he seems to have known Augustus and the future emperor Claudius (below, p. 99 n. 62); for what little more we know or can infer see further Kraus (1994), 1–4; for his dates see above, n. 2.

113. For the concept of the ‘implied author’ see Booth (²1983), 530 s.v. (Index).

114. The almost entirely unknown Gracchan historian Cn. Gellius may have written as much: Badian (1966), 11–12.

115. Kraus (1994), 1–2.

116. This is the historiographical topos relata refero, ‘I tell the tale that I heard told’ (see Kraus (1994), 206 on 6.18.16); Livy’s frequent references to it are directly relevant to the notion that his history demands that we think not about what ‘really’ happened but about how and why the Roman traditions formed, and what they mean for the reader.

117. On these authorial interventions, with examples of many more, see Steele, R. B., ‘The historical attitude of Livy’, AJP 25 (1904), 15–44 Google Scholar; on his mentions of contemporary Rome in the first decade see von Haehling, R., Zeitbezüge des T. Livius in der ersten Dekade seines Geschichtswerkes (1989)Google Scholar.

118. Miles (1995), 47; on characters acting like historians see Kraus (1994), 149–51, 191, 213–14 (on Manlius Capitolinus) and Jaeger (1997), esp. on Septimus Marcius.

119. He was probably not the first such professional historian of Rome (see Wiseman (1987), 248–52 on Valerius Antias and others in the first century B.C.), but the Roman suspicion of intellectual activity for its own sake (above, p. 7) had not yet diminished, and the prejudice in favour of the ‘senatorial historian’ (i.e. the man who makes history and then retires to write it) was strong in the ancient world as in the modern; see further Earl (1965), 237–8, Kraus (1994), 5 n. 17.

120. Kraus (1994), 158 on 6.12.2-6; Miles (1995), 47–54.

121. ‘The moralism of Plutarch’s Lives’, in Innes-Hine-Pelling (1995), 219–20.

122. Wallace-Hadrill, A., ‘Time for Augustus’, in Whitby, M., Hardie, P., and Whitby, M., edd., Homo Viator (Bristol, 1987), 221–30 Google Scholar; see now also Evans, J. D., The Art of Persuasion. Political Propaganda from Aeneas to Brutus (Ann Arbor, 1992 Google Scholar) and Galinsky (1996), 38–41.

123. Peiling (last n. but one, 207, 219) makes an analogy with the civic function of Greek tragedy, on which see the essays in Winkler, J. J. and Zeitlin, F. I., edd., Nothing to do with Dionysus? Athenian Drama in its Social Context (Princeton, 1990)Google Scholar.

124. Salmon, E. T., ‘The evolution of Augustus’ Principate’, Historia 5 (1956), 456–78 Google Scholar. For thoughtful discussions of the passage from republic to empire see Eder, W., ‘Augustus and the power of tradition’ in Raaflaub-Toher (1990), 71–122 Google Scholar and Crook, J. A., ‘Augustus: power, authority, achievement’ in Bowman, A. K., Champlin, E., and Lintott, A., edd., The Cambridge Ancient History Vol. X (² 1996), 113–46 Google Scholar.

125. Miles (1995), 68.