Hostname: page-component-78c5997874-fbnjt Total loading time: 0 Render date: 2024-11-05T05:28:29.343Z Has data issue: false hasContentIssue false
  • English
  • Français

Narrator Interventions in Thucydides*

Published online by Cambridge University Press:  23 February 2012

David Gribble
David Gribble
Affiliation:
London
Get access Rights & Permissions [Opens in a new window]

Extract

The main narrative of Thucydides is characterised by a third person ‘objective’ style where signs of the narrator are concealed. But this predominant narrative mode is punctuated by passages (2. 65, 6. 15, etc.) where the narrator interrupts the main account, referring to himself in the first person and/or to time outside that of the main narrative. These rare intrusions of the voice of the narrator-historian—‘narrator interventions’—are the most quoted and discussed in the whole History. Reaction to them has been of two sorts. They have either been seen as later additions and used as the centrepiece of analyst interpretations of the History, or they have been treated as expressions of the ‘judgement’ of the historian, providing the key to the History's meaning. The result of these approaches is unsatisfactory. The interventions are either bracketed as foreign to the original plan of the historian, or given special status as the exclusive source of his meaning. The effect is to cut them loose from the reading of the rest of the work, as intrusions of another stage of composition or of another voice which no longer narrates, but gives judgement. Worse still, such interpretation compares the decontextualised ‘judgements’ it has isolated from the narrative and declares them inconsistent with each other. Such ‘extrinsic’ approaches to the interventions risk reducing Thucydides’ text to a patchwork of differing and competing voices and opinions.

Type
Research Article
Information
The Journal of Hellenic Studies , Volume 118 , November 1998 , pp. 41 - 67
Copyright
Copyright © The Society for the Promotion of Hellenic Studies 1998

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 As far as I know, there is no general treatment of narrator comment in Thucydides. But note Loraux, N., ‘Thucydides a écrit la guerre du Péloponnèse’, MHTIS: Revue d'Anthropologie du Monde Grec Ancien 1 (1986) 139–61Google Scholar on the persona of the narrator; and Pearson, L., ‘Thucydides as reporter and critic’, TAPA 78 (1947) 37–60Google Scholar, on expressions of ‘personal opinion’ in Thucydides.

2 On the pervasiveness of this model in Western literature, the origins of which can be traced partly to Thucydides himself, see Booth, W.C., The Rhetoric of Fiction2 (Chicago 1983), esp. chaps. 1 and 10CrossRefGoogle Scholar; and Genette, G., ‘Boundaries of narrative’, New Literary History 8 (1976) 1–13 at 11-12CrossRefGoogle Scholar.

3 (n.2) using ‘discourse’ in the restricted sense of utterance by an agent. On the use of this narrative mode in historiography, see Barthes, R., ‘The discourse of history’ in Barthes, , The Rustle of Language (Oxford 1986) 127–40Google Scholar; White, H., ‘The value of narrativity’, in Mitchell, W.J.T (ed.), On Narrative (Chicago 1980) 1–23Google Scholar; and id., ‘The question of narrative in contemporary historical theory’, History and Theory 23 (1984) 1–33CrossRefGoogle Scholar.

4 Barthes (n.3) 132 refers to a ‘referential illusion’.

5 On ‘focalisation’, the narratological term for describing ‘who sees’, cf. Genette, G., Narrative Discourse (Oxford 1980) 185 ff.Google Scholar; Bal, M., Narratology: Introduction to the Theory of Narrative (Toronto, Buffalo & London 1985) 100–15Google Scholar. For other works applying the terminology of narratology (particularly focalisation) to Thucydides, see Rood, T., Interpreting Thucydides: A Narratological Approach (Oxford D. Phil, thesis 1995)Google Scholar; Hornblower, S., ‘Narratology and narrative techniques in Thucydides’, in Hornblower, (ed.), Greek Historiography (Oxford 1994) 131–166Google Scholar.

6 Booth (n.2) 3-20; Genette (n.2) 9-12. Though writers of such narratives may think they are writing in a way that is absolutely objective, there is in fact no pure access to reality through language. To use the terms of Genette (n.5, 189), there is no ‘zero-focalised’ statement; cf. Bal (n.5) 128; Hornblower (n.5) 133, 148 ff.. On the pervasiveness of the historian's voice in classical historiography see Fowler, R.L., ‘Herodotos and his contemporaries’, JHS 116 (1996), 62-87 at 70 ffCrossRefGoogle Scholar.

7 The term is intended to refer only to a narrative mode involving the effacement of the narrator, without any implications as to the level of involvement of narrator or reader in the fate of the characters. On the various narrative strategies available to an author, see Bal (n.5) 123-6.

8 On this effect see Genette (n.2) 9, quoting Beneviste (‘the events seem to tell themselves’); and Barthes (n.3) 131-2, 138-40. Such a narrative mode offers not objectivity, but a rhetoric which encourages readers to believe they are being given direct access to the facts. White 1984 (n.3) 4 sees this strategy as a way of concealing the inevitable interpretative element of historiography: the historian resorts to it out of embarrassment that his history includes not just ‘the facts themselves’ but also interpretation. On ‘objective’ narrative in ancient historiography, see Wheeldon, M.J., ‘“True Stories”: the reception of historiography in antiquity’, in Cameron, A. (ed.), History as Text (London 1989) 33–63 at 45 ffGoogle Scholar. On ‘objective’ narrative in Thucydides, see Loraux (n.1) 139 ff.

9 Cf. Plu. Mor. 346f-347c; D.H. Thuc. 15; Woodman, A.J., Rhetoric in Classical Historiography (London & Sydney 1988) 25–8Google Scholar. Cf. also Zanker, G., ‘Enargeia in the ancient criticism of poetry’, RhM 124 (1981) 297–311.Google Scholar

10 Cf. Booth (n.2) 155, 205-9. On the narrator's voice in non-historiographical classical authors, see Block, E., ‘The narrator speaks: apostrophe in Homer and Vergil’, TAPA 92 (1982) 7–21Google Scholar; Byre, C.S., ‘The narrator's addresses to the narratee in Apollonius Rhodius’ Argonautica’, TAPA 121 (1991) 215–27Google Scholar. Cf. also Goldhill, S., The Poet's Voice (Cambridge 1991)Google Scholar, and on Homer, Jong, I.J.F. de, Narrators and Focalizers: The Presentation of the Story in the Iliad (Amsterdam 1987)Google Scholar.

11 Cf. Booth (n.2) 423-4.

12 The role of interventions in fictional works with third-person narrative is apparently somewhat different. Here, adversion to the person of the narrator may be used to remind the reader of the constructedness of the text and the fictionality of the narrative. Such interventions can be compared to the ‘metatheatre’, which reminds the audience of a play of the circumstances of performance and ‘break’ the dramatic illusion. In historiography, on the other hand, interventions are normally designed to reinforce belief in the credibility of the author and reliability of the narrative.

13 Cf. Woodman (n.9); Wheeldon (n.8); and Moles, J.C., ‘Truth and untruth in Herodotus and Thucydides’ and Wiseman, T.P., ‘Lying historians: seven types of mendacity’, both in Gill, C. and Wiseman, T.P. (eds), Lies and Fiction in the Ancient World (Exeter 1993)Google Scholar.

14 On rhetoric in the prologues of ancient historiography, see Woodman (n.9); Wheeldon (n.8); Moles (n.13). On the classical historian's effort to establish authority, see Marincola, J., Authority and Tradition in Ancient Historiography (Cambridge 1997)CrossRefGoogle Scholar.

15 Connor, W.R., ‘Narrative discourse in Thucydides’, in The Greek Historians, Literature and History: Papers Presented to A.E. Raubitschek (Stanford 1985) 1-18 at 6–7.Google Scholar

16 Loraux (n.1) 142, 153-7. Loraux argues that the disappearance of the narrator is already prepared for in the prologue: note the opening third-person at 1. 1.1, with its claim that Thucydides ‘wrote’ the Peloponnesian War, as though the war were an object ‘out there’.

17 On the sphragis, see Woodbury, L., ‘The seal of Theognis’ in Studies in honour of Gilbert Norwood ((Toronto 1952)Google Scholar; Goldhill (n.10) 109-12.

18 On the rhetoric of the prologue in Herodotus and Thucydides, see Moles (n.13) and the works cited there.

19 Not that pleasure is excluded: Woodman (n.9), ch. 1. Contemplation of great achievement and suffering brings both pleasure and understanding: on the lessons of Thucydides’ History, cf. Rutherford, R.B., ‘Learning from history: categories and case studies’, in Osborne, R. and Hornblower, S. (eds), Ritual, Finance, Politics. Athenian Democratic Accounts Presented to David Lewis (Oxford 1994), 53–68.Google Scholar

20 Moles (n.13), who points out that both Herodotus and Thucydides claim to compete with Homer in truthfulness. Moles details other similarities between the two prologues: both begin, for example, with a third person, which suggests the objectivity of the text. In order to depict himself as more truthful than Herodotus, Thucydides thus imitates Herodotus’ own rhetoric.

21 But there is similarity as well as difference in the claims of epic and history to tell the truth. Epic poets also claimed to sing the truth, but they relied on a concept of truth which had more to do with fittingness, inspiration, Or beauty than with accuracy: cf. Macleod, C.W., Collected Essays (Oxford 1983) 4–5Google Scholar, noting how Homeric storytellers are praised for telling things κατὰ μοῖραν (truthfully, fittingly) or ὲπισταμένως (skilfully). Thucydides in the speeches aimed to record not only what was actually said, but also τὰ δέοντα (which includes a notion of what is fitting).

22 Nevertheless (1. 21.1), the further back one goes into the past, the more difficult events are to reconstruct with any certainty: through the passage of time they have ‘entered the mythological realm’. Thucydidean historiography is thus primarily a historiography of the present: Hartog, F., ‘L'oeil de Thucydide et l'histoire “véritable”’ Poétique 13 (1982) 22-30 at 23–4.Google Scholar

23 Cf. Dewald, C., ‘Narrative surface and authorial voice in Herodotus’ Histories’, Arethusa 20 (1987) 147–70 at 164 ff.Google Scholar. On Herodotus’ voice, with its distinctive constant appeal to sources, see now Fowler (n.6).

24 Normally, the narrator does not alert the reader to his own role in organising the events into a narrative (‘as I said earlier’), except 5. 1.1; 6. 94.1, on which see appendix 1, below. Instead, the narrative purports to offer direct access to events which exist ‘out there’, unmediated by the structuring efforts of a narrator (Loraux [n. 1] 142), who appears only in clearly-signalled interventions, and then as commentator rather than organiser. On ‘shifters of organisation’ in historiography, see Barthes (n.3) 128-9.

25 Dewald (n.23) 147-54, noting that the narratorial voice actually points to the multiplicity of logoi and questions their relationship to the truth in various ways: an early, and programmatic, example is the ambiguously distancing attitude of the narrator to the stories explaining the origin of conflict between East and West: Hdt. 1. 5. The result (and perhaps also the cause) of this distancing is a less straightforward attitude to ‘what actually happened’ and the ability of a history to capture this in language. In Thucydides, as I will argue below, the narratorial voice always confirms the main narrative.

26 Hence he hardly ever suggests the possibility of other stories, or casts doubt on his own narrative: Hornblower (n.5) 151; Connor (n.15) 5-6. The ‘arduous’ task of sifting through competing stories (1. 22.3) does not leave a mark on the narrative itself, a trait of Thucydides’ method often overlooked when considering his ‘omissions’.

27 Hornblower (n.5) 149; Loraux (n.1) 157 f..

28 Cf. 1.21.1; 2.23.5 (Thucydides’ account of the origins of the war will free others from the need to investigate the origins.)

29 On this, see further below.

30 Cf. Connor, W.R., Thucydides (Princeton 1984) 14–18Google Scholar; Connor (n.15) 7-8.

31 Loraux (n.1) 155-6. Loraux claims (157) that the ideal reader of Thucydides is characterised by a complete submission to the historian. This goes too far. For a more reasonable account of the impact Thucydides imagined his History would have on its readers see Rutherford (n.19).

32 Cf. de Jong (n.10) 45-53 arguing that it is both poet and Muse who are regarded as singing.

33 Dewald (n.23) 153; Hartog (n.22) 22-3: the narrator's role is ‘dire ce que se dit’.

34 Autopsy and source are not generally cited to guarantee the veracity of the narrative proper. Such ‘shifters of listening’ (Barthes [n. 3] 128) would undermine the sense of objectivity and direct access to the facts.

35 Hartog (n.22) 22-4. This attempt to establish the authority of the text competitively against epic anticipates Plato's attempt to replace epic with philosophy in the Republic: cf. Goldhill (ri.10) 167 ff. Yet whereas Plato claimed a new and better kind of insight for philosophy, Thucydides presents his brand of carefully researched analytical historiography as more able than epic or Herodotean performance historiography to fulfill the educative role (in terms of information and contemplation of suffering) traditionally expected of literature.

36 Though note that 1. 22.4 implies not that the History is not designed ‘for listening’ at all, but that it might be less pleasurable for this purpose. Cf. Hornblower, S., Thucydides (London 1987) 29Google Scholar.

37 On the importance of Thucydides’ awareness of the non-oral status of his work, see Gentili, B. and Cerri, G., History and Biography in Ancient Thought (Amsterdam 1988) 11–16.Google Scholar

38 Genette, G., ‘Récit fictionnel, récit factuel’, in his Fiction et Diction (Paris 1991) 65 ff.Google Scholar, argues that in factual narratives the author takes responsibility for the statements of the text, so that there is no point in trying to distinguish a separate narrator: a distinction between narrator and author is, in fact, an aspect of the distinction we make between factual and fictional narratives. But though the adoption of a narratorial mask (‘now I will tell you a story’) may seem to be distinctive of fictional and not factual narratives, Thucydides too says, in effect, ‘now I will tell you a story’: his decision to employ a narrative style, and one which effaces the author (i.e. a strategy precisely parallel to that of epic and many novels) generates, and makes it appropriate for us to speak in terms of, a narrating voice (in this case one which is effaced or concealed). The narrator arises wherever there is an attempt to persuade, or structure reality, through a narrative. For further discussion of the applicability of ‘rhetorical’ analysis to factual narratives, see Booth (n.2) 407-8, 424-5.

39 On the distinction in a fictional text, see Scholes, R. and Kellog, R., The Nature of Narrative (Oxford 1966) 266Google Scholar (‘a projection of the author's empirical virtues’); Bal (n.8) 119-20; De Jong (n.10) 44-5. Booth (n.2) 67-77 uses the somewhat different concept of ‘implied author’.

40 When Thucydides speaks in his text, he speaks as careful observer and acute analyser, as the historian in his work. One might compare the role of the first person in Pindar, which so often stands for the voice of the professional poet rather than the person Pindar: on this, see most recently Goldhill (n.10) 142 ff., with full bibliography.

41 As Genette (n.2) 9-11 demonstrates.

42 A prologue is not an intervention, because the ‘objective’ narrative has not yet been established, the narrator not yet effaced, so there is no sense of an intervention. The beginning of the narrative proper is clearly marked by the change of style at 1. 24.1 (‘Epidamnus is a city …’, which is the signal of narrative: ‘now I will tell you a story’). 5. 26.1 is a clear resumption of the prologic voice of 1. 1.1 (‘These things too were written down by the same Thucydides, an Athenian …’), and the return to narrative mode is clearly marked at 5.27.1 (‘For after the fifty-years peace had been made …’).

43 Booth (n.2) 208.

44 So one cannot restrict the discussion of interventions to ‘first-persons’, as Loraux (n.1) 156 does.

45 On counterfactuals in Thucydides, see Dover, K.J., ‘Thucydides’ historical judgement: Athens and Sicily’, in Dover, , The Greeks and their Legacy (Oxford 1988) 74–82Google Scholar; Flory, S., ‘Thucydides’ hypotheses about the Peloponnesian war’, TAPA 118 (1988) 43–56.Google Scholar

46 This is the only instance where the first person refers to the narrator as an actual agent in the narrative (contrast the use of third persons in 4.104).

47 The most convincing account of this disputed passage is that of Gomme in Gomme, A.W., Andrewes, A., and Dover, K.J., A Historical Commentary on Thucydides (Oxford 1945-1981) 2, 277Google Scholar: the passage is genuine and in its proper place, and άρχομένου τοῦ πολέμου in 17.1 refers to the first years of the war in general, though the conditions described would best fit 430 specifically. Cf. Rood (n.5) ch. 4, 207-10.

48 3. 84, regarded as spurious from a very early date, is rejected by most editors. See Gomme (n.47) ad loc.

49 This passage is most untypical in that it provides information and analysis (especially on Pleistoanax) which could in no way be gleaned from the main narrative: it is perhaps influenced by the suspension of narrative purity and continuity associated with the end of the war, and the imminence of the second preface.

50 Note use of another device here: a rhetorical question, an obvious signal of the reader (or, more technically, the narratee). This passage and 7. 44.1 are the only two instances of the device in Thucydides: see Dover in Gomme (n.47) 5, 400; Hornblower (n.5) 149.

51 The distinction between history and fiction is not the object studied, but rather the method of regarding the object (so historical fiction is not history). As I suggested above, Thucydides’ distinction between his historiography and epic was not so much the objects studied (though lapse of time does tend to push events into the realm of the mythical), but the methodology employed to study them.

52 Hornblower (n.5) 151. Such reference is normally confined to material outside the main narrative: e.g. Hellanicus in the Pentecontaetia (1. 97.2), or popular Athenian tradition on the Peisistratids (6. 54.1). On narrator address to the reader (second persons) in other classical authors, see Gilmartin, K., ‘A rhetorical figure in Latin historical style: the imaginary second person singular’, TAPA 105 (1975) 99–121Google Scholar; Block (n.10) 13 f.; Byre (n.10).

53 The most disputed of all was Alcibiades, and interventions relating to Alcibiades may be influenced by a careful attempt to steer a way between excessively polarised viewpoints. The later reputation of Nicias (whose name was struck off the list ot those killed in Sicily because he surrendered) may also have influenced the narrator intervention about him at 7. 86.5. Other interventions on individuals whose reputation was disputed are 5. 14-16 and 8. 68. On interventions relating to individuals in book 8, see appendix 1.

54 Thus the narratorial voice in Thucydides does not present itself as taking part in a ‘contest of public voices’: contrast Aristophanes’ first persons, skilfully analysed by Goldhill (n.10) 167 ff. The narratorial voice presents itself as rising above the controversies surrounding the interpretation of the events of the war, to address an audience which is not civic but private, timeless, international (cf. 1. 22.4). This does not mean that Thucydides’ work is not sometimes a contribution to an Athenian civic debate (as at 2. 65), merely that a distinction can be made between the narratorial voice in the work and the voice of the (civic) dramatic poet or orator.

55 On the narrator's providing of detail likely to be unfamiliar to readers, see Ridley, R.T., ‘Exegesis and audience in Thucydides’, Hermes 109 (1981) 25–46.Google Scholar

56 Barthes (n.3) 130.

57 The use of intervention to generate pathos in this way is familiar from epic, as in Virgil's apostrophe to the dead Nisus and Euryalus at Aen. 9. 446-9.

58 Macleod (n.21) 11-12 citing Od. 1. 353-5; Timocles CAF 2, 453; Plb. 1. 1.2. The contemplation of suffering in literature is a training of the emotions, consoling us and steeling us through an appreciation of the fact that to be human is to suffer greatly. On insights of this sort in Thucydides, see Macleod (n.21) 103-22, 140-158; Stahl, H.-P., Thukydides: die Stellung des Menschen im geschichtlichen Prozess (Munich 1966) 135 ff., 157Google Scholar.

59 As Thucydides sets out in the prologue (1. 22): the accurate perception of the past leads to understandings of general truths about human behaviour.

60 Cf. also the comment on the fate of Mycalessus at the hands of Thracian mercenaries: 7. 29.5.

61 On this tactic to reinforce an impression of accuracy, see Hornblower (n.5) 150-2; Loraux (n.1) 151. The technique is in fact less surprising than it seems: figures do not produce much emotional effect, and may distract from pathetic climaxes. A similar example is 7. 87.5-6. Here, although the number of those who were captured at the end of the Sicilian campaign is put at ‘about 5000’ (though this too was difficult to determine with accuracy), we are then told that they were ‘wiped out so to speak completely’, and that ‘out of many, few returned home’. Such examples highlight the rhetorical dimension of the narrator's quest for accuracy.

62 Stahl (n.58) 130-7. Some interventions in the Iliad may be designed to provide similar narrative-articulating climactic moments: Iliad 2. 484-93; 12. 176; 17. 260-1: cf. Edwards, M.W., The Iliad: A Commentary. Vol. V: Books 17–20 (Cambridge 1991) 2–7CrossRefGoogle Scholar.

63 For what follows, see Connor (n.30) 50-62.

64 1. 140-44, 2. 13 (in indirect speech), and 2. 60-4.

65 For a subtle analysis of the effect of temporal manipulation in 2. 65 and other narrator interventions, see Rood (n.5), ch. 4, 204 ff.

66 Cf. Connor (n.30) 68-70.

67 The individualism of the Athenians makes the city dynamic and strong, but it also leads to acts of collective folly or selfishness, like the decision to depose Pericles, the refusal of Spartan peace offers after Pylos, the recall of Alcibiades and (perhaps) the launching of the Sicilian expedition; or to the sort of obsession with victory (ϕιλονικία) described at 7.28.3.

68 Cf. Romilly, de, Thucydides and Athenian Imperialism (Oxford 1963) 223Google Scholar: see 2. 65, 3. 17, 4. 108, 6. 15, 7. 28.3,7.787.4-6 + 8.1, 8.24.4-6.

69 7. 28.3, 7. 55.2, cf. 3. 16.2, 3. 113.2-6, 8. 24.5.

70 On paradeigmata and patterns, cf. Hunter, V., Thucydides the Artful Reporter (Toronto 1973) 179–80 and passimGoogle Scholar. However, by paradigmatic action, I mean not the grand or metaphysical patterns which Hunter sees Thucydides’ History as designed to reveal, but simply recurring causal factors.

71 E.g. Naxos was the first allied city to lose its freedom contrary to established agreements (1. 98). The plague was the first beginning of άνομία in Athens (2. 53.1). The political infighting in 415 was the first time affairs in the city were brought into confusion (2.65.11). The Corcyrean stasis was ‘among the first’ of the staseis of the Peloponnesian War (3. 82.1). Brasidas was the first Spartiate to to be sent to help allies overseas, and the best (4. 81.2-3).

72 On superlatives, see the general list of interventions, above.

73 Cf. Fowler (n.6) 73-4. Fowler compares the traditional schema of the πρῶτος εύρετής. Cf. also the habit of thought implied by the Greek phrase ἅλλα τε καί, on which see Erbse, H., ‘Der erste Satz im Werke Herodots’, in Festschrift Bruno Snell (Munich 1956) 209–22 at 215-7Google Scholar. On Thucydides’ interest in ‘Gelenkstellen und Krisepunkten’, cf. Stahl (n.58) 92, 129 ff.

74 Cf. de Romilly (n.68) 209; Westlake, H.D., Essays on the Greek Historians and Greek History (Manchester 1969) 168–73Google Scholar; Erbse, H., Thukydides-Interpretationen (Berlin & New York 1989) 84–5.CrossRefGoogle Scholar

75 The expedition was born out of dissension (the Sicilian debate), and undermined by the political sabotage and recriminations of the Herms and Mysteries affair, the main upshot of which was the exile of the most active of the expedition's generals, on its own a ‘subsequent decision’ capable of ‘taking the edge off the expedition’. Moreover, Alcibiades’ advice became crucial in spurring on the Spartans to send Gylippus, and left the ultimately ineffective Nicias in charge. Cf. Dover in Gomme (n.47) 5. 425-7; Liebeschuetz, W., ‘Thucydides and the Sicilian expedition’, Historia 17 (1968) 289–306.Google Scholar

76 It is usually thought that Thucydides must either be referring to 416-15 or 404. Thus Brunt, P.A., ‘Thucydides and Alcibiades’, in Brunt, , Studies in Greek History and Thought (Oxford 1993) 17–46Google Scholar argues (18-19) that he cannot be referring to 406; Dover in Gomme (n.47) 5, 242-5 that the verbs καθεῖλεν and ἔσϕηλαν must refer to 406. But Thucydides saw the defeat in Sicily as so massive that it prefigured the final defeat in 404 (2. 65.11-12; 7. 87-8.1). The two defeats (both preceded by internal dissension and a deposition of Alcibiades) are inextricably linked: both and each may be referred to by καθεῖλεν and ἔσϕηλαν.

77 Other clear illustrations of this type of overlap, apart from those discussed below, are 4. 65.4 (where, as Hunter [n. 70] 78-83 shows, we have already had clear implicit indications of the effect of good fortune in encouraging the Athenians to stretch out for more); 6. 69.1 and 7. 55.2 (the daring and skill of the Syracusans—and also their similiarities to the Athenians—is evident from the narrative); 8. 96.1-3 (the contrast of Athenian and Spartan national characteristics is a constant theme from the first debate in Sparta in book 1 onwards).

78 This aim is certainly present: e.g. the statement that Pericles looked after the city moderately and safely, and under him Athens was at her greatest (2. 65.5), which is not strictly speaking necessary to the main narrative.

79 Objective analysis would have led to an appreciation of Alcibiades’ value for the city. ‘Private’ analysis leads the Athenians to banish him, with disastrous destabilising results for the city.

80 6. 16.6. Though Alcibiades’ presentation of his Mantinean policy is not actually untrue to the facts, the policy was nevertheless a failure, leading to the reassertion of Spartan supremacy in the Peloponnese (5. 75.3).

81 Some scholars (e.g. Brunt [n. 76]) have been worried by the alleged inconsistency between 6. 15 and other passages which imply a more ambivalent view about Alcibiades’ military performance, especially 8. 86.4.

82 De Romilly (n.68) 22-37; Andrewes, A., ‘Thucydides and the causes of the war’, CQ 9 (1959) 223–39.CrossRefGoogle Scholar

83 Ach. 496-556, Peace 603-18. Richardson, J., ‘Thucydides 1. 23.6 and the debate about the Peloponnesian war’, in Craik, E.M. (ed.), Owls to Athens: Essays… Presented to Sir Kenneth Dover (Oxford 1990) 155–62Google Scholar, sees Thucydides 1. 23.6 as a reaction to such popular accounts of the causes of the war.

84 But the precedent of Homer and Herodotus is at least as important a reason for starting the narrative of a war with an account of its causes (as the journal's referee points out to me).

85 E.g. 1. 9, where the Trojan War is explained in terms of imperial expansion rather than the judgement of Paris.

86 Though some hold that 2. 65.11 (‘not so much a mistaken analysis of those against whom the expedition was directed’) is inconsistent with 4. 1.1, which emphasises the Athenians’ ignorance of the scale of the island: see Gomme (n.47) 5. 247; de Romilly (n.68) 109.

87 On this and other ‘not so much’ clauses, see Westlake, H., Essays on the Greek Historians and Greek History (Manchester 1969) 161–7Google Scholar; Hornblower (n.5) 157. Distinguish ‘negation by antithesis’ clauses (‘not x, but y’), which emphatically rule out an option: Rusten, J.S., Thucydides. The Peloponnesian war, Book II (Cambridge 1989) 24–5Google Scholar.

88 Connor (n.30) 15-19, 233-40, pleads for a reading of Thucydides which takes into account the constant development of reader reaction.

89 My thoughts on this passage owe much to some unpublished comments of Chris Pelling.

90 (n.45); cf. Hornblower (n.5) 134.

91 Cf. Fowler, D., ‘Deviant focalisation in Virgil's Aeneid’, PCPS 36 (1990) 42–63.Google Scholar

92 Note the shift from νομίσας; (before the bracketed text), implying a subjective evaluation of affairs, to άνασκοπων (after the bracketed text) implying an objective recognition of an actual state of affairs.

93 Note the echo of the bracketed text (‘when Nicias first arrived he aroused fear’) in the narrative of Demosthenes’ own actions (‘recognising that he too [καὶ αύτός] was most terrible to the enemy on the first day’). This underlines the conclusion that the thoughts are both the narrator's and Demosthenes’.

94 In narratological terminology, a primary focalisation occurs when it is the narrator who sees; a secondary focalisation when it is one of the characters of the story who sees.

95 Barthes (n.3) 137.

96 On the attribution of motivations (many of which cannot have been, or are most unlikely to have been, known to Thucydides) as a way of establishing an interpretative structure cf. Hunter (n.70); Schneider, C., Information und Absicht bei Thukydides (Göttingen 1974)CrossRefGoogle Scholar; Westlake, H.D., ‘Personal motives, aims and feelings in Thucydides’, in Westlake, , Studies in Thucydides and Greek History (Bristol 1989) ch. 14Google Scholar.

97 The echo of ὲϕάινετο in διεϕάνη underlines the contrast between the immediate perception on the part of the allies and the wider perspective of Athenian power revealed in the intervention (the change from imperfect to aorist and the addition of δι - achieving the transition from ‘appearance’ to ‘reality’).

98 Cf. Connor (n.30) 133 and Hornblower, S., A Commentary on Thucydides (Oxford 1991) 2. 341Google Scholar, comparing the structure of 4. 81. The intervention at 7. 55 presents a similar ring structure (Syracusans— Athenians— narrator-Athenians—Syracusans). 7. 42.3 is framed by the focalisation of Demosthenes, 2. 65 of Pericles.

99 Compare the intervention regarding the Chians at 8. 24.4-6, though here the Chians are practically absolved from the fault of miscalculating Athenian resilience, which was so surprising as to be almost too much for prediction by normal reasoning.

100 Cf. Westlake (n.72) 138-44.

101 2. 65 (where the narrator's prediction of future resilience after plague and hardship is already anticipated by Pericles), 4. 108, 7. 28.3, 7. 87.4-6 + 8. 1,8. 24.4-6.

102 8. 1.3. In this case, the whole of book 8 constitutes the demonstration of Athenian resilience.

103 As in 7. 55 (the reaction to the first Syracusan naval victory in Sicily), where the narrator-focalised comments on the similarities between Syracuse and Athens as democracies are likely to b e read to some extent as a continuation of the thoughts of the Athenians. In other key passages, though there is technically secondary focalisation, the feeling of the hand of the narrator is strong: e.g. 7. 18 and 8. 1-2, on which see Connor (n.15) 14. On shifting focalisations in general, see Hornblower (n.5) 164-5.

104 7. 42.3, cf. 2. 65.7, 4. 108.4. Cf. Hornblower (n.5) 134.

105 Cf. Connor (n.15) 8 ff., noting the rapid changes of secondary focalisation in 3. 92.3.

106 On character introductions in Thucydides (a quasi-intervention in themselves, particularly in extended form like 5. 43), see Griffith, G.T., ‘Some habits of Thucydides when introducing persons’, PCPS 8 (1961) 21–33Google Scholar.

107 οι πολλοί: i.e. the Athenian assembly.

108 On the role of 6. 15 in its context, cf. Connor (n.30) 164-6.

109 Cf. Diodotus at Thuc. 3. 43.4-5.

110 On τὸ παράλογον, see above. πρόνοια: 2. 62.5, 2. 65.6, 4. 108.4, 6. 13.1 (cf. 2. 65.13, 2. 89.9, 2. 60.5, 1. 24.4, 1. 138.3).

111 Cf. de Romilly (n.68) 223, n. 1 calling Schwartz's automatic branding of interventions as later additions ‘a failure to understand the way Thucydides wrote’. De Romilly also argues persuasively that these passages are fully consistent with the surrounding narrative. She claims, moreover, that certain passages, including the key interventions, reveal a perspective which Thucydides can only have achieved after the end of the war. This is clearly true, though I am not convinced that this perspective is not demonstrated by the surrounding narrative, so that the interventions can be shown to be later (as de Romilly argues: 221-4, 228-9). But in terms of the reading of the work this disagreement is not significant (as de Romilly herself maintains): what is important is that (i) the passages isolated by de Romilly are consistent with the rest of the work, and (ii) the work as we have it now reflects the perspective of the time after the end of the war.

112 Whereas significances in literature are held from the total impression of the whole text, works of history are approached with a view to isolating the judgements of their authors.

113 Connor (n.30) 236.

114 Thus the objective style can actually be seen as ‘a means by which the reader is drawn into the work’: Connor (n. 15) 232.

115 Cf. Luschnat in RE Suppl. xii, 1257: ‘Oder soll man sagen, dass das Vertrauen des Historikers in die Rationalität doch erschüttert ist (wofür sein Rückzug auf das Beschreiben als das Medium einer nicht-rationalen und nach-rationalen Weltbewältigung sprechen könnte)… und dass Stellen wie 11.65 nur Rückzugsgefechte sind, die ohne Überzeugung geführt werden’. But the picture of the narrator interventions as discordant outbreaks of positivism implied here is wrong, as I have tried to show.

116 Cf. Marinatos, N., ‘Thucydides and oracles’, JHS 101 (1981) 138–40CrossRefGoogle Scholar, with full bibliography, and the additional remarks of Dover, K.J., ‘Thucydides on oracles’, in Dover, , The Greeks and their Legacy (Oxford 1988) 65–73Google Scholar. Marinatos may be right to emphasise the fact that Thucydides does not (except perhaps in 5. 26.3) actually call into question the validity of oracles. By adopting an attitude which is agnostic (tending toward the cynical), Thucydides contrives to get the sense of depth and pattern provided by having oracles in the narrative without abandoning his concentration on the human, rather than the divine, as the motor of historical causation. (Even in 5. 26.3, which appears to throw doubt on the worth of oracles in general, the oracle that the war was to last 27 years is nevertheless deployed to give oracular sanction to Thucydides’ own interpretation of the length of the war.)

117 The perspective of the historian puts him in a unique position to determine whether or not oracles were fulfilled. But Thucydides uses this prerogative only tentatively or ambivalently.

118 Rood (n.5) 339 ff., noting the emphasis on familiar themes in some of the book 8 interventions (e.g. Athenian internal dissension in 8.89; the continuing strength of Athens in 8. 97) would prefer to see them as somewhat more integral to the overall narrative strategy of the the work as a whole than my presentation here implies. His arguments about book 8 are to be expanded in the forthcoming book version of his thesis, to be published by OUP.

119 Andrewes in Gomme (n.47) 5. 399-400.

120 Loraux (n.1) 156.

121 ‘ Cf. Connor (n.30) 214-8, and the critique of Andrewes in Erbse, H., Thukydides-Interpretationen (Berlin & New York 1989) 1–82CrossRefGoogle Scholar.

122 Rood (n.5) 344.

123 See Meiggs, R. and Lewis, D., A Collection of Greek Historical Inscriptions2 (Oxford 1988) 85Google Scholar; Gomme (n.47) 5. 309-11.

124 In the case of Alcibiades, this reaction-guiding is part of an ongoing development of a key figure, begun in 5. 43 and 6. 15, which would have been continued in the lost or incomplete section of the History.

125 Though note 1. 128-38 (on the curses associated with Pausanias and Themistocles) and 2. 97 (on the Odrysian empire), where digressive material is introduced in ‘Herodotean’ manner, without the formal apparatus of intervention. Still, the Pausanias/Themistocles digression moves easily into the intervention about Themistocles, at a point where we can start to see a relevance of the digressive material to the main narrative.

126 The Pentecontaetia fills in essential background of Athenian imperialistic growth and the threat this posed to Sparta (1. 88, 1. 118); the Sicilian history sets out the strength and antiquity of the group of cities the Athenians had set themselves against (6. 1.1, 6. 6.1); and the tyrant digression explains Athenian paranoia towards potential tyrants (6. 53.3, 6. 60).

127 This is related to a wider phenomenon in Greek narratives of withholding mention of an event until the moment it becomes relevant: see Fraenkel, E., Aeschylus. Agamemnon (Oxford 1950) 3, 805.Google Scholar