Jump to ContentJump to Main Navigation
Ancient Historiography and its ContextsStudies in Honour of A. J. Woodman$

Christina S. Kraus, John Marincola, and Christopher Pelling

Print publication date: 2010

Print ISBN-13: 9780199558681

Published to Oxford Scholarship Online: May 2010

DOI: 10.1093/acprof:oso/9780199558681.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 24 February 2017

Divide and Conquer: Caesar, De Bello Gallico 7

Divide and Conquer: Caesar, De Bello Gallico 7

(p.40) 2 Divide and Conquer: Caesar, De Bello Gallico 7
Ancient Historiography and its Contexts

Christina Shuttleworth Kraus (Contributor Webpage)

Oxford University Press

Abstract and Keywords

This chapter discusses the work narrative structure does in creating meaning through a case study of BG 7. It analyses how the text invites us to isolate and then pit paired elements against each other: Caesar and Vercingetorix, who structure BG 7 and the history as a whole; Avaricum and Alesia, with their internally-divisive physical structures; and various doubled plot elements at Gergovia. Close reading of the Gergovia narrative shows how the interweaving of plots and the insistent doubling of conventional battle and ethnographical elements create significant echoes that signal to a reader both the constructedness of the narrative and its affinity with ‘proper’ literary historiography. Caesar leads us to consider not only the whats and whys, but most importantly the hows, of narrative. How we divide this (or any) text affects our interpretation; and the more conscious we are of the spaces, and the topoi that structure the narrative, the better we can understand it as historiographical re-presentation.

Keywords:   Vercingetorix, narrative, topoi, historiography, doubling, Gergovia, ethnography, battles, Bellum Gallicum 7

Caesar … having defeated the Ancient Britons by unfair means, such as battering‐rams, tortoises, hippocausts, centipedes, axes, and bundles, set the memorable Latin sentence, ‘Veni, Vidi, Vici’, which the Romans, who were all very well educated, construed correctly. The Britons, however, who of course still used the old pronunciation, understanding him to have called them ‘Weeny, Weedy, and Weaky’, lost heart and gave up the struggle, thinking that he had already divided them All into Three Parts.

Sellar and Yeatman, 1066 and All That

I. Divide

In his memoirs of the great declaimers, Seneca the Elder devotes intensive critique to the diuisio, the art of making correct distinctions among the essential points of a controuersia that marks rhetorical expertise and, consequently, controls the kind of persuasive discourse proper to an elite male.1 The way things are, or can be, divided is a primary concern of those Roman men who in the first century BC were engaged in codifying conquest by articulating their empire through centuriation, mapping, apportionment, (p.41) and confiscation of land, and by identifying distinct parts of a territory for purposes of military conscription.2 Finally, diuisio, considered both in terms of textual and of real space, is a particular concern of Caesar in his Bellum Gallicum, whose Gallia is first and foremost marked by great divides: Gallia est omnis diuisa in partes tres (BG 1.1.1).3 Those famous opening words align the BG with the tradition of geographical commentarii;4 as Caesar continues, by revealing geographical knowledge and designating hierarchical relationships within this foreign territory he pre‐emptively insinuates that it is already marked out ready for conquest.5

Though clear textual segmentation was a historiographical habit, connecting the res written about and the written res, as well as the passage of time with the stages of written time,6 Caesar's interest in division is marked even among historians. This may be a function of his chosen genre, the commentarius, which tends towards the schematic: a kind of writing that emphasizes the anecdotal or the list form, so far as we can tell, it probably avoided large‐scale connected narrative.7 It may also reflect a tendency of Caesar's own personality, which shows a penchant for dividing things—be they peoples, actions, grammatical elements, or sentences—into clear and distinct groups.8 Whatever their ultimate origin, as reflections of rhetorical expertise, cultural attitudes, and personal style, Caesarian sections are worth a closer look.

The macro‐divisions of the BG are, of course, its bookrolls, which correspond to campaign years.9 Within these, Caesar shapes and structures events (p.42) into episodes by various means—deploying unities of space, time, action, or combinations thereof—and links these together with ‘clichés de liaison’ (dum haec geruntur, his rebus gestis, etc.).10 Correspondences among otherwise separate scenes are brought out by linking repetition, either of language (more common than one might expect in this ‘simple’ text11), of design, or of character.12 In particular, topoi are strong connective devices, both intra and intertextually. A topos may serve as shorthand for a particular famous scene—the most widely diffused is Troy, but one also finds Sardis (itself a reflection of Achilles' heel, e.g. at Sall. BJ 93, Livy 5.47) and the battle in Syracuse harbour (e.g. Polyb. 1.44.4–5, Sall. BJ 60.3–4)—or for a type of scene (e.g., cannibalism in a besieged town); both kinds can forge links between parts of a narrative, and across narratives.13 Finally, Caesar builds objects into his text—particularly defensive and offensive opera—that serve as focal points and structural markers, aiding in the construction of a properly organized (divided) text. In this paper, I explore elements of the three great siege narratives of Bellum Gallicum 7, paying special attention to Caesar's use of structure, especially (but not exclusively) the meta‐narrative use of physical objects.14 These echoes, of material both repeating and contrasting, put disparate parts of the text in communication with one another, using the observing reader as a conduit of meaning and interpretation.

I read BG 7 as falling into three major segments articulated by its three major siege‐battles. Each of these receives emphasis from the amount of text devoted to it, and in each Caesar turns topography and space to compositional advantage, marking the episodes off from the surrounding narrative. The three episodes are joined by an introductory section in which the Gallic rebellion takes shape and Caesar and Vercingetorix take each other's measure (1–13); and by a slightly longer section featuring Labienus in the environs of (p.43) Paris (54–67). The overall modulation is that of a tricolon with a smaller central unit and extended third:15

Part 1(12.5 Teubner pages)

1–13 introductory moves: 1–5 C. in Italy, Gallic plotting; 6–13 C. and V., skirmishes

13.3 C. arrives at Avaricum; V. follows

14–31 AVARICUM (Romans victorious)

32.1 C. refits army at Avaricum, departs; 34.2 arrives at Gergovia

Part 2 (9 Teubner pages)

32–53 GERGOVIA (Romans withdraw)

53.4 C. withdraws from Gergovia

Part 3(16.5 Teubner pages)

54–67 interlude: C.'s regrouping, 57–62 Labienus without

C., 63–7 skirmishes

68 V., followed by C., arrives at Alesia

68–90 ALESIA (Romans victorious)

Gergovia follows immediately from Avaricum and leads on to Alesia. Much of the fighting at Gergovia should not have happened at all: since 43.4, when he learns that his chief allies have joined the Gallic rebellion, Caesar has been looking for a way to leave the area without losing face. The battle that eventually takes place does so largely without either his leadership or the obedience of his normally disciplined soldiers. The attempt to storm the besieged oppidum marks the moment at which Caesar moves from being uictor to (all but) uictus, a shift effected partly by strong, ironic echoes of elements from the (victorious) Avaricum narrative. The Alesia narrative, in turn, will pick up on elements from the two earlier ones, in the process of turning Caesar from uictus to uictor once more.

This narrative economy has distinct advantages. It allows Caesar to organize complex historical information around a limited number of significant moments; to showcase his army in slow motion, as it were, in the epic drama of the stationary siege narrative;16 and, by thus slowing the narrative down for speeches and sieges, to hold the spotlight on himself and Vercingetorix for (p.44) long periods.17 But even as sketchy a structural model as this raises potential difficulties for historical interpretation. If we accept Caesar's invitation to concentrate on the triad of set pieces, this has at least one significant result for our understanding of what happened on the ground: by tucking its narrative into a less emphatically presented section, Caesar minimizes the importance of the cavalry battle at 7.66–7, whose loss forces Vercingetorix to make the disastrous decision to retreat to Alesia.18 Structure imports meaning, in this case by sidelining a causally important event.

Structure imports meaning, as well, by inviting comparison and contrast. More so even than in the earlier books, in BG 7 characters and action alike are repeatedly doubled, mirrored, or split, a multiplicity enacted by recurring narrative elements.19 The most obvious example is the presence of not one but two charismatic commanders, Caesar and Vercingetorix. Pairing himself with (or against) a barbarian leader is a technique Caesar introduces early on, in his depiction of the German leader Ariovistus in BG 1.20 But it is in Book 7 where the device is exploited most fully. Vercingetorix enters the book's Gallic narrative in person (at 4.1) before the character Caesar does (at 6.1), and functions throughout as a fully articulated second focal point. The rivalry between the evenly matched commanders tropes the fight for control of Gallia Comata, which is also—and not incidentally—a fight for control of the narrative.

II. Barring Misfortune

The episodes of Avaricum and Alesia contain physical structures specifically designed to block movement: respectively, a Gallic wall and Roman siege works. From the start, Avaricum is seen as a reflection of its defences. Having burned the neighbouring towns to deny the Romans refuge or food, the Gallic freedom fighters spare Avaricum at the tearful request of its citizens. From Caesar's point (p.45) of view, it is the largest, the best fortified, and most fertile of local settlements; from the Bituriges', the most beautiful in all Gaul:

13.3 Caesar ad oppidum Auaricum, quod erat maximum munitissimumque in finibus Biturigum atque agri fertilissima regione, profectus est

Caesar set off for Avaricum, a very large and well‐fortified town in the territory of the Bituriges, and in a very fertile area of the territory

15.4 procumbunt omnibus Gallis ad pedes Bituriges, ne pulcherrimam prope totius Galliae urbem, quae et praesidio et ornamento sit ciuitati, suis manibus succendere cogantur. facile se loci natura defensuros dicunt

the Bituriges fell at the feet of all the Gauls, lest they be compelled to burn down with their own hands the most beautiful city in nearly the whole of Gaul, which both protected and adorned their nation. They claimed that owing to the nature of the site they would easily defend it

The Bituriges put equal weight on Avaricum's value as a defensive fortress (praesidium) and a decoration (ornamentum); they add that it is ‘easily’ (facile) defensible—but in this text such claims typically mark enemy overconfidence.21 These two complementary passages with their balancing superlatives pit Caesar's practicality and restraint (he considers military and logistical matters, and avoids making global claims) against Gallic fondness for adornment and tendency towards overdone rhetoric.22

The Bituriges are right that Avaricum is naturally defended (17.1), but wrong about the ease of keeping Caesar at bay. Once the Romans invest the oppidum, narrative attention focuses on its wall, to which Caesar devotes an extended, ekphrastic description (22.3–24.2):

totum autem murum ex omni parte turribus contabulauerant atque has coriis intexerant. tum crebris diurnis nocturnisque eruptionibus aut aggeri ignem inferebant aut milites occupatos in opere adoriebantur, et nostrarum turrium altitudinem, quantum has cotidianus agger expresserat, commissis suarum turrium malis adaequabant et apertos cuniculos praeusta et praeacuta materia et pice feruefacta et maximi ponderis saxis morabantur moenibusque adpropinquare prohibebant. [23] Muri autem omnes Gallici hac fere forma sunt. trabes derectae perpetuae in longitudinem paribus interuallis, distantes inter se binos pedes, in solo conlocantur. hae reuinciuntur introrsus et (p.46) multo aggere uestiuntur, ea autem, quae diximus, interualla grandibus in fronte saxis effarciuntur. his conlocatis et coagmentatis alius insuper ordo additur, ut idem illud interuallum seruetur, neque inter se contingant trabes, sed paribus intermissis spatiis singulae singulis saxis interiectis arte contineantur. sic deinceps omne opus contexitur, dum iusta muri altitudo expleatur. hoc cum in speciem uarietatemque opus deforme non est alternis trabibus ac saxis, quae rectis lineis suos ordines seruant, tum ad utilitatem et defensionem urbium summam habet opportunitatem, quod et ab incendio lapis et ab ariete materia defendit, quae perpetuis trabibus pedes quadragenos plerumque introrsus reuincta neque perrumpi neque distrahi potest. [24] His tot rebus impedita oppugnatione milites, cum toto tempore frigore23 et adsiduis imbribus tardarentur, tamen continenti labore omnia haec superauerunt et diebus XXV aggerem latum pedes CCCXXX, altum pedes LXXX exstruxerunt. cum is murum hostium paene contingeret

They had covered every section of their wall with towers and overlaid these with hides. Now with frequent sorties by day and night, they tried to set fire to the earthwork or attack our soldiers while they were working on the siege. They kept equalling the height of our towers (which increased daily as the earthwork was raised higher and higher) by extending the scaffolding on their own towers. They sabotaged the progress of the mines which we had driven by the use of timbers tempered and sharpened at the end, boiling pitch, and very heavy rocks. In this way they prevented us from coming close to the walls of the town. [23] All Gallic walls follow approximately this same pattern: at right angles to the wall, along its entire length, beams are placed in the ground at equal intervals, with a distance between them of two feet. These are secured firmly from the inside, and then clothed with large amounts of rubble; the gaps we mentioned between them are stuffed with large rocks at the front. Once the beams are laid and joined, another level is added on top in such a way that the gap remains the same; nor does the second set of beams touch the first, but, because the spaces between the beams are the same, each one is firmly kept separate by single stones placed between them. The whole construction is built in this manner until the appropriate height of the wall is complete. The finished edifice is not unattractive in appearance or variety, with the alternating beams and rocks in straight lines preserving their own formation; moreover, it offers an excellent chance for the practical defence of cities, since the stone gives protection from fire and the timber from the ram—for continuous beams, usually forty feet long and secured on the inside, can neither be broken through nor torn apart. [24] All these factors hindered the siege operation, but our soldiers, despite continually being obstructed also by the cold and constant rain, by unremitting effort still overcame every obstacle and within 25 days had built an earthwork 330 feet wide and 80 feet high. When this was almost touching the wall…[the Gauls set fire to the agger].

(p.47) Even as the wall takes shape before our eyes, this passage neutralizes it. First, Caesar notes that the wall is typical (‘all Gallic walls follow approximately this same pattern’): that is, this is not (necessarily) a description of the particular wall blocking Caesar's soldiers, but is a kind of ‘standard Gallic wall’, resembling a description one might find, for instance, in a treatise on building.24 In this very rhetoricized scene, the typicality becomes another element inviting us to see the wall as belonging to a plausible, not an actual, world.25 Second, the passage displays both the technical virtuosity required to construct the wall and that needed to write its description,26 reflecting the Bituriges' delight in Avaricum as an adornment—this wall is, after all, a complex and attractive creation (23.5 hoc … in speciem uarietatemque opus deforme non est)—but also Caesar's ability to understand, and hence, inevitably, to defeat, the Gallic munimenta.27 Third, in its apparent clarity but actual difficulty of reproduction, the description renders the wall useless because practically unbuildable.28 Far from presenting an obstacle, it is the act of portraying those walls that allows the Romans to progress from the difficulties of 22.5 (moenibusque adpropinquare prohibebant) to the success of 24.1–2 (omnia haec superauerunt … murum hostium paene contingeret). Instead of constituting a barrier, then, the textual wall opens up a breach.

Framing this description of the muri Gallici are two notices of admirable Gallic action, staged on and further enhancing the thematic significance of Avaricum's wall. The first gives a general overview of their skills (22.1–3):

Against the extraordinary bravery of our soldiers the Gauls used every kind of ingenuity, being an extremely resourceful people, and particularly talented at copying and putting into practice everything they receive from anyone. [Singulari militum (p.48) nostrorum uirtuti consilia cuiusque modi Gallorum occurrebant, ut est summae genus sollertiae atque ad omnia imitanda et efficienda, quae ab quoque traduntur, aptissimum.] For they turned our grappling hooks aside with nooses; once these were made secure, they dragged them inside the walls with ropes. They also started tunnelling beneath our earthwork to undermine it: this was all the more skilfully done because they have [quod apud eos] many iron mines, and so are practised experts in every kind [omne genus] of tunnel.

The second zooms in on a particular act performed by particular Gauls (25):

The night was now at an end, yet fighting continued everywhere…[The Gauls] believed that at that very moment the salvation of Gaul hung in that moment [uestigio] of time; and there then occurred, before our very eyes, something which, being worth remembering, we believed should not be passed over [accidit inspectantibus nobis, quod dignum memoria uisum praetereundum non existimauimus]. A Gaul stood before the town gate, opposite our tower, and was throwing lumps of tallow and pitch—passed from hand to hand—into the fire; he was wounded in the right side by a dart from an artillery machine, breathed his last, and fell. One of the Gauls nearby stepped across the man as he lay there and carried on with the same task; when he was killed in a similar manner by a dart a third man took over, and likewise a fourth succeeded him. Nor was that position abandoned by their defenders before … the fighting was at an end.

Each of these passages is marked as worthy of note, the first by the emphatically fronted singulari and the subsequent ethnographical note on Gallic imitative excellence (continued by the information that ‘they have’ and are experienced with ‘every type’ of tunnelling); the second by one of Caesar's rare first‐person references, in which he puts one foot outside the narrative to function simultaneously as dux and as ethnographer/observer.29 These flags point us towards these cases of remarkable action. On closer attention, however, they turn out—like the wall—to be no more particular to this moment than they are typical of this kind of story. The Gallic consilia that match the ‘unique’ Roman uirtus are ‘characteristic’ (ut est…): these enemies can copy anything (omnia … quae ab quoque traduntur). The story that Caesar relates of the brave Gauls guarding the gate (ante portam) partakes of a topos stretching as far back as Iliad 12.30 In both cases, the idea of (p.49) imitative repetition pulls the narrative towards the typical. In both cases, too, Caesar resists that pull to conventionality by insisting on the ‘unique’ (singulari … occurrebant) and remarkable nature of these actions. That resistance ensures our close attention to this rhetoricized historical ‘record’ staged on the remarkable structure of the muri Gallici. Thus a narrativized structure, Avaricum's defensive works, becomes both the backdrop and the emotional centre of that siege.

The counterpoised narrative of Alesia boasts no fewer than three points of emotional heightening, but each is associated, like the Avaricum wall, with a strongly marked place:31 (1) the centre of the oppidum, in which Critognatus makes his modest proposal to cannibalize the weak (77.1 ‘But those who were besieged in Alesia’); (2) the space between the hill fort and the Roman siege works, in which women and the aged are exposed to die (78.2–5; the space is used again for a disastrous sortie at 81–2); and (3) the Roman opera themselves, which demonstrate Caesar's ingenuity, planning, and capacity for endurance (built at 72–4, narratively deployed at 81–2). These three distinct loci in which vital actions happen are united by their concentric focus on the citadel itself. Here again, physical structures and spaces contribute to narrative exaedificatio. Furthermore, each (physical) locus is an instantiation of a (literary) topos: cannibalism in a besiged town;32 the contested space before a besieged city, a motif that begins with Troy; and opera which, like the murus Gallicus, suggest technical descriptions.33 At both Avaricum and Alesia, then, division and recombination of often typical narrative elements structure and focus Caesar's history.

III. Conquer?

In the central panel of BG 7 (32–53) Caesar narrates his unsuccessful attack on Gergovia. In contrast to Avaricum and Alesia, both the physical and emotional focuses of Gergovia are markedly diffused. Though its defensive struc (p.50) tures play an important part in the story, there is no ekphrastic description corresponding to the murus Gallicus or the Roman opera. Gergovia's murus provides a platform for conventional siege action (see below), but the main story focuses on Caesar's trying to get away from the area, not on its capture.34

Rather than concentrating on the fate of the oppidum (as with Avaricum) or on the concentric circles of besiegers and besieged (as with Alesia), in this episode the diffused narrative focus distracts from the Roman defeat. Two plots—Gergovia [1] and the Aeduan rebellion [2]—are articulated by repetitions of the phrase (ad/ab) Gergouia(m) (34.2, 36.1, 37.1, 38.1, 40.7, 41.1, 42.1, 43.5, 45.4). Our attention (qua readers) shifts from storyline to storyline—as indicated by the bold numbers in the list below—following each one's various permutations:


ch. 32–34.1

Decision to move from Avaricum [1]; war plans interrupted by need to arbitrate among Aeduan royal claimants [2]



March to Gergovia, on river Elaver; misgivings about nature of attack; seizure of hill, building of second castra



Aeduan rebellion; C. removes some troops from oppidum + departs for Aedui; Litaviccus flees to Gergovia



C. returns to Gergovia, rescues harassed troops



Aedui continue rebellion; C. decides to leave Gergovia



Roman ruse to distract enemy; hidden insignia, 45; mismanaged attack; misread insignia, 50



C.'s rebuke of troops; withdraws across river Elaver



Litaviccus & Convictolitavis, progress of Aeduan rebellion; C. departs to join Labienus

Interwoven narratives are a feature of historiography, often efficiently presenting simultaneous action in parallel theatres of war.35 At Gergovia, however, the idea of ‘two’ takes over as both a thematic and a structuring principle: there are not only two storylines, but two commanders, two Aeduan uergobrets, two young Aeduan noblemen, two Roman camps, two manoeuvres with insignia, two centurions, two scenes with Gallic matronae, two Gallic shouts, two parts of town—etc.

(p.51) One storyline ([2]) follows the split of the Aeduan ciuitas between loyal and disloyal factions.36 It begins with two men competing for the position of chief magistrate; their trouble causes Caesar's narrative to take a new tack, and the general himself literally to veer off his main road, after only a sentence introducing storyline [1]. Calculated repetition of vocabulary (indicated typographically below) puts the two movements in close parallel (32.2, 33.1, 2):

Winter was almost over; and though at this very time of year [ipso anni tempore] he was called to wage war [uocaretur] and had decided to set out against the enemy [ad hostem proficisci constituisset]…the leaders of the Aedui came to him as envoys to beg him to help their people in a time of dire necessity [maxime necessario tempore]. [33] Despite [Caesar etsi] believing that it would be to his disadvantage to leave the path [discedere] of the war and the enemy … he decided to set out in person [ipse] for Aeduan territory [in Haeduos proficisci statuit] and summoned [euocauit] their whole senate … to him.37

The doubling of uergobrets is later mirrored by the introduction of two noblemen, Eporedorix and Viridomarus, who—though each supporting one of the rival claimants—work together against the splitting of the Aedui as a whole from Caesar's cause (37–41).38 Even after Caesar physically returns to Gergovia the second time (41.1) the Aeduan problem continues to provoke narrative digression via messages (42.1, 43.1, 43.4), eventually becoming the driving force behind Vercingetorix's final push (63.1). Storyline [2], initially interwoven with the Gergovia narrative, therefore generates continuous narrative pressure, mirroring/mirrored by the military pressure on Caesar to move from one place to another as the Gallic rebellion grows. Downplaying the cavalry battle that causes Vercingetorix to retreat to Alesia (66–7, above) contributes to the sense of pressure: though that victory materially contributes to the ultimate Gallic defeat, Caesar chooses instead to emphasize growing Gallic strength.

(p.52) The battle of Gergovia (storyline [1]) is itself constructed around doublets. These give the episode a coherence that is otherwise threatened by the Aeduan digressions; in addition, they provide communicative and structuring links with the book's earlier episodes. The repeated elements show clearly:


March to Gergovia, on river Elaver; ruse at crossing. Misgivings about nature of attack; Gallic skirmishes; C. seizes hill, builds second camp

37–40 [2]

Aeduan rebellion


C. returns to Gergovia, rescues harassed troops there

41.6–43 [2]

Aeduan rebellion


C. decides to leave Gergovia


Roman ruse to distract enemy; mismanaged attack on oppidum


hidden insignia


women pleading (~ 7.28.4)


Fabius the centurion climbs onto wall (~ 7.27.2–3)


women encouraging (~ 7.26.3–4)


death of Fabius


misread insignia


Petronius the centurion dies in gate to save men (~ 7.28.3 portarum … portis)


C. rebukes troops (~ 7.29); skirmishes to save face (53.2–3); withdraws across river Elaver

Once Caesar can engage with his interrupted plan (32.2 ad hostem proficisci constituisset), he sends Labienus north and leads his legions ad oppidum Gergouiam,39 a hill fort separated from him by a river. In the build‐up to the main action the narrator stages two stratagems, each of which combines topographical knowledge and the manipulation of appearances. The first involves a standoff between Caesar and Vercingetorix for control of a river crossing (35.2–5). In the end Caesar successfully fools his opponent into misjudging the number and location of his men, a feint enabling him to cross over.40 Now, in Caesarian narrative, topography can seem to blend into a kind of indistinct background, an impression aided in no small degree by (p.53) Caesar's deliberately restricted vocabulary. Yet sometimes a flumen is not just a flumen.41 At 5.4–5, for example, the Loire graphically represents the split in loyalty—and the resultant possibility of treachery—between neighbouring tribes: ‘When they had reached the Loire (the river [flumen] separating the Bituriges from the Aedui) … they returned and reported to our legates that they had turned back fearing treachery on the part of the Bituriges.’ Though these tribes have featured earlier in his narrative, Caesar has reserved this particular topographical information until this point—a common ancient narrative technique which here enhances the symbolic use of the geographical feature.42 Readers of Caesar also know from the disaster at Cenabum (11.7–8), where a narrow bridge helps cause a massacre, that control of river crossings is essential. When Caesar and Vercingetorix play their bridge game in 7.35, then, the flatness of the narrative belies the importance of the ability successfully to use topography.43

During the investment of Gergovia it will become abundantly clear that—as always in Caesar—topographical knowledge and control are vital to success. Caesar is ultimately defeated more by his legions' lack of them than by Gallic superiority in tactics or strength. His successful manipulation of the Elaver crossing will contrast with the Roman failure to conquer the uneven topography of the oppidum itself. Conversely, the scene at the Elaver foreshadows a more serious failure of the Gallic general to make his own specialized knowledge work for him: Vercingetorix is taken in by appearances here (35.3 uti numerus legionum constare uideretur), as he will soon be again in the more consequential matter of the high ground outside Gergovia (to which I now turn).44

This second stratagem involves a more complicated manipulation of the enemy's visual field (45.1–6):

Caesar sent a number of cavalry squadrons to the place; at midnight he ordered them to range in every direction rather noisily [tumultuosius]. At first light he commanded a large number of pack‐horses and mules to be led out of camp … and the muleteers, wearing helmets so as to look and act [specie ac simulatione] like cavalry, to ride (p.54) around the hill. To these he added a few cavalry to range about, for the sake of the show [ostentationis causa45].… All these activities were seen [uidebantur] from far off in the town, for there was a view from Gergovia down into the camp; but at such long range it was impossible to find out anything for sure [neque tanto spatio certi quid esset explorari poterat46]. Caesar sent one of his legions to the same high ground … and concealed it in the woods. Gallic suspicion grew, and their whole force transferred to the place to fortify it.

The narrator stresses the distance from the Roman camps to the town: paradoxically, a straight line of vision (ut erat a Gergouia despectus in castra) does not allow for adequate interpretation of what is seen. As long as the Romans are in the ascendant, Caesar and his men control that interpretation—the muleteers successfully impersonate equites, and the mules, horses. The stratagem succeeds in drawing the Gauls away from the town, giving Caesar the chance to make a face‐saving attack and withdraw quickly before the growing rebellion traps him there (45.7–9):

Seeing that the enemy camp was unoccupied, Caesar had his men cover up their insignia and conceal their military standards; then he took the soldiers in small groups—so as not to attract attention from the town—from the larger camp to the smaller one. He explained to his legates (to each of whom he had given command of a legion) what he wanted done [quid fieri uelit]. In particular he warned them to keep their soldiers under control, lest in an eagerness to fight or in hope of booty they advance too far. He also set out [proponit] the disadvantage of unequal ground [iniquitas loci]. This could be avoided by speed alone: this was a moment for seizing an opportunity, not fighting a full‐scale battle [hoc una celeritate posse uitare: occasionis esse rem, non proeli].

What he orders is topsy‐turvy—moving quickly with unexcited soldiers, a smash‐and‐grab raid on a well‐fortified town—but wholly in line with his deceptive stratagems.47 At first all is well, as the legions with amazing speed (tanta … celeritate) overtake the Gallic castra, even surprising a sleeping king (46.5).48 This is Caesar's chosen end point: consecutus id quod animo propo (p.55) suerat, Caesar receptui cani iussit legionique decimae … continuo signa constituit (47.1 ‘Having achieved what he intended, Caesar ordered the retreat to be sounded; the Tenth legion … halted at once’).49 Overwhelmed by their success, however, the other legions go on autopilot—and once again, the narrative takes off in a new direction.

That direction brings us to a cluster of scenes familiar to readers of epic and history. With the clustering comes also a set of specific references back to the siege of Avaricum—inviting us, once again, to measure a description of experience against one mediated by parallels and topoi. Two scenes of Roman centurions are interwoven with emotive scenes of Gallic women. Women and children first:

47.5–6 The married women began to throw clothing and money from the wall; baring their breasts and leaning forward with outstretched hands they pleaded with the Romans to spare them, not to act as they had at Avaricum, when they did not even refrain from taking the lives of the women and children. Some let themselves down hand by hand from the wall and gave themselves to the soldiers.

48.3 When a large host [of Gauls] had gathered, the married women, who had just now been holding their hands out to the Romans, started to call upon their men and, in Gallic fashion, to unbind and display their hair, and bring their children forward into view.

Their quick transition from pleading for mercy to encouragement tracks the evolving military fortunes of the Gauls, from defensive (47.4) to offensive (48) mode. The town's topography is intimately involved: walls switch from a place of exchange (de muro … iactabant; de muris per manus demissae) that the Romans can scale (47.7 murum ascendit; in murum extulit), to a defensible obstacle (48.2 sub muro consistebat; 50.3 de muro [sc. Romani] praecipitabantur) that is at the same time a venue for spectacular exhortation (48.3 ostentare).

The women's actions are further marked by details that by referring outside the story identify the passages as part of a larger narrative. In their appeal for pity, the matronae hark back to the earlier episode at Avaricum (47.6).50 When describing their subsequent encouragement of their men, the narrator explains that the women are acting more Gallico (48.3). As with the (p.56) descriptions of brave Gallic action discussed in Section II, this material draws attention to the narrative's artificiality: these women—breasts, hair, babies, and all—are typical elements of a siege narrative designed to ratchet up the emotional quotient of the moment.51

Following each of these scenes is a vignette with a centurion, whose actions are similarly conventional:

47.7 It became apparent [constabat] that Lucius Fabius, a centurion from Legion VIII, had declared among his men that he was stirred by the prospect of Avaricum‐like rewards and was not going to enable someone else to scale the wall first. He grabbed three of his fellow soldiers, got them to hoist him up, and climbed the wall; grasping hold of each of them, he in turn hauled them up on to the wall.52

50.4–6 Marcus Petronius, a centurion from the same legion, had tried to break down the gates, but was overwhelmed by the enemy host and abandoned hope of saving himself. Gravely wounded, he called to his soldiers who had followed him: ‘Since’, he said, ‘I cannot save myself along with you, I shall at least take care to secure your survival—for it was my desire for glory which made me lead you into danger. When I give you your chance, watch out for yourselves.’ At the same time he plunged into the midst of the enemy, killing two and forcing the rest a little way back from the gates. To his men, who tried to come to his assistance, he said, ‘In vain you are trying to preserve my life; already my blood and strength fail me. Get away while you can, and return to the legion.’ Fighting in this way he fell not long afterwards and proved the saviour of his men.

Centurions are Caesar's favourite actors. They focus action, wield significant objects, are granted rare direct speech, and generally serve as the stylized representatives of his legions, who through their leaders speak in (largely) ultra‐brave, ultra‐Roman, ultra‐loyal voices. Like the Gallic women,53 Fabius and Petronius are studies in contrast explicitly flagged as comparandi: they are introduced in precisely the same way; Petronius is eiusdem legionis (50.4); and his story is activated only after Fabius' body is expelled from the town (50.3 ‘at the same time the centurion Lucius Fabius and those who had scaled (p.57) the wall with him were surrounded, killed, and thrown headlong from the wall’).54 It is Petronius who, in sacrificing himself to save his men, is the spiritual descendant of Cato's military tribune55 and the immediate forerunner of Lucan's—and Caesar's—Scaeva (BC 3.53.4–5, Luc. 6.138–262); it is also he who, as a double for the failed Fabius, redeems his mistake. His is the last voice we hear in this battle, which ends swiftly thereafter.

By positioning Petronius where he does, and by recalling such narrative precedents as Cato's tribune, Caesar does not so much mitigate the Roman defeat as recast it in the mode of sacrificial glory, turning our attention away from the disaster towards the memorialization of it through historiography: this is a type of story with which we are very familiar. It is perhaps, then, churlish to note that Petronius' presence is utterly unprepared for. This is not, in a way, surprising, given Caesar's laconic style, and his tendency to provide information only when it is needed. But since one named centurion (Fabius) does go over the wall at 47.7, it is somewhat unexpected that it is not he, but another centurion (pointedly ‘of the same legion’) who directly following Fabius' death at 50.3 receives this aggrandizing notice. The effect—again via doubling—is once more to highlight ostentatiously the rhetoricity of the scene—perhaps even more, given that the scene partakes of the same topos as the Gauls in the gate earlier: one centurion seamlessly replaces another.56

The third prominently doubled scene in the Gergovia narrative reaches outside that storyline, capping the episode's intratextual relationship with Avaricum (nn. 13, 50 above). Caesar's rebuke to his soldiers after their retreat mirrors by its position and its rhetoric Vercingetorix's consolatory rebuke to the Gauls after Avaricum:

52.1–2, 53.1 The following day Caesar summoned an assembly and upbraided the soldiers for their imprudence and over‐eagerness. He explained the significance of disadvantageous ground, which he himself had sensed at Avaricum…However admirable their courage … they deserved reproach for their lack of discipline…[53] He finished his speech by encouraging [confirmatis] the soldiers not to be disturbed because of these events, and not to attribute what had resulted from unfair ground to bravery [uirtuti] on the enemy's part.

29.1–5 The following day Vercingetorix summoned an assembly, spoke words of comfort [consolatus], and urged them not to be too down‐hearted or anxious over (p.58) this setback. After all, he said, the Romans had not won because of their bravery [uirtute] or in open battle, but rather by means of trickery and siegecraft in which they themselves were inexperienced. Anyone who expected all outcomes in time of war to be favourable was mistaken; he himself had never wanted a defence of Avaricum, and they were witness to the fact.…He would, however, swiftly repair the situation with some more significant advances.

Both generals speak after a defeat, using indirect speech introduced with similar formulae (Postero die Caesar contione aduocata ~ Postero die concilio conuocato); both combine criticism with encouragement, praising their fighters' uirtus despite their loss. Both leaders exculpate themselves by arguments drawn from history: Vercingetorix had not wanted to defend Avaricum; Caesar had learned at Avaricum of the dangers of unequal ground. Yet the differences, too, are palpable. Caesar's speech is critical of the legions' licentia and adrogantia; Vercingetorix's, while criticizing the Bituriges' imprudentia, praises the power and might of the Gallic union. Caesar is more critical than Vercingetorix: perhaps an indication of his greater realism, and hence of his greater abilities as a general. Nor must he resort to encomium to maintain the unity of his troops.

The correspondence of these speeches underscores the differences between the generals, to the Roman's advantage. Yet it also, obviously, suggests their similarity, especially in defeat. It is conventional to say that Roman literature magnifies the person and power of significant enemies so as to magnify the Roman victories over them: so Sallust on Jugurtha, Horace on Cleopatra, Livy on Hannibal, Tacitus on Arminius, and everyone on Catiline. That modality is certainly operative here: the more Vercingetorix resembles Caesar, the more difficult and hence glorious Caesar's victory will seem. Furthermore, a defeat in the middle of the book sets up and acts as foil to the greater victory at the end. Caesar was in no way obliged to devote as much time or artistry to what he could easily have presented as a minor skirmish at an unknown oppidum, or indeed left out altogether: this pattern of a triad with a central On the rhetoric of division and conquest in the's presentation of events to be simply a reflection of ‘the way things really were’ (above, n. 15). Yet one could also take Vercingetorix's resemblance to a Roman commander as indicative of a larger anxiety about Roman imperial ideology and its effects on ‘barbarians’. So at the opening of BG 7 the Gauls may use ultra‐Republican, Ciceronian jargon not just for the sake of parody (both of Gallic aspirations and of Cicero, one suspects), but also to suggest that Gallic freedom fighters have some degree of right on their sides.57 And again, we are sent back to BC 1, where Ariovistus' (p.59) long speech (1.44) scores important points against Rome's treatment of its ‘friends’.58

Sympathy with the enemy does not, of course, require one to respect their political integrity. Nor do the similarities traced by repeated structures and diction between Caesar and his Gallo‐German opponents necessarily reduce his military auctoritas. But the preceding exploration has shown, I hope, that Caesar's sophisticated deployment of topoi, narrated structures, and repeated elements of all sorts makes Book 7 not only the cap of the Bellum Gallicum, but also an eloquent representation of the continually challenging work of writing the Roman empire.


(*) Improbably, Tony commented on this paper when I first wrote it; that it has been sitting in a drawer since is at least partly because he pointed out (in the nicest possible way) that it was an incoherent mess. (‘Chris, I'm baffled.’) His reading proved, as usual, invaluable: any merit the piece might now have is due to his sharp eyes. He is owed many thanks, for this and for all his years of unstintingly generous help, wisdom, and care. I am grateful as well to Hans Aili, John Jacobs, Christopher Krebs, John Marincola, Debbie Nousek, Fredrick Oldsjö, and Chris Pelling, for comments and encouragement.

(1) For an introduction see Habinek (2005), 60–71, 118 (further reading); on divisions in rhetoric see Heath (1997) and for Seneca see Fairweather (1981), 152–65 and Huelsenbeck (2009).

(2) On maps and the division of conquered territory see the wide‐ranging discussion of Nicolet (1991) with Wiseman (1992), 22–42, Mattern (1999), 24–80, and Riggsby (2006), 32–45 on surveying and the possession of Gaul. Moatti (1997), 57–95 explores Rome's categorization of the world in its ‘archives of conquest’ (67); see also 217–54 on the development of typologies of thought in the late Republic.

(3) I cite Caesar throughout from Hering's 1987 Teubner edition of the Bellum Gallicum; all translations are from Hammond (1996) with modifications. References to the BG without book number are to Book 7.

(4) Rüpke (1992), 212–14; on space in Caesar see n. 31 below.

(5) On the rhetoric of division and conquest in the BG see Torigian (1998), Krebs (2006), and cf. Braund (1996), 158.

(6) On res gestae see (e.g.) Kraus (1994a), 18; on marking chronological stages in narrative (primo, deinde, etc.) see McDonald (1957), 169; Kraus (1994a), 143–4; for discussion of significant divisions in earlier historiographers see Immerwahr (1966) on Herodotus, Walbank (1975) on Polybius.

(7) Riggsby (2006), 133–56 with extensive bibliography; the matter continues to be debated.

(8) One can start from the (in)famous ‘ueni uidi uici’, Caesar's triumphal titulus after the Pontic campaign. On Caesar's use of syntactical division in the interest of clarity see Eden (1962), 104–6; on his sentences see Spilman (1932); on the De analogia—which seems to have been at least partly concerned with dividing words into categories—see Fantham (2009).

(9) Not, however, to calendar years (Adcock (1956), 35; Wiseman (1998), 5). Six commentarii end with a movement either away from Gallia Comata (1.54.3, 2.35.4, 6.44.3) or into hiberna (3.29.3, 4.38.4, 7.90.7); the exception is 5, which nevertheless suggests withdrawal from military activity: ‘all the forces of the Eburones and Nervii which had mustered dispersed, and soon after Caesar commanded [habuit] a quieter Gaul’ (5.58.7).

(10) Chausserie‐Laprée (1969), 61–108 for their function in historiographical narrative.

(11) See Batstone and Damon (2006), esp. ch. 3 on Caesar's characterization of others in the BC via repeated words, and ch. 5 on his style.

(12) In this Caesar simply behaves like any ancient author of extended narrative. For linkages by repetition of language see Moskalew (1982) (on Virgil), Martin (1955) (Tacitus); for repeating elements of design see Vasaly (2002) (Livy); of character, Griffin (1985), 183–97.

(13) For the former, cf. BG 7.12.5–6 (oppidani turning suddenly from surrender to exhortation, threat to centurions, etc.) which looks forward to Gergovia, itself explicitly linked to Avaricum (see below). On the ‘portability’ of topoi see Hinds (1998), 34–7; Keitel, ch. 19 in this volume; on Troy see Paul (1982), 147–8 and Rossi (2004a), ch. 1.

(14) Two important treatments of physical objects in Latin historiography are Konstan (1986) and Jaeger (1997), 94–131; on built objects in Caesar see Dodington (1980), Scarola (1997), and Kraus (2007).

(15) On the thematic modulation in BG see Nousek (2004); on tricolon shapes see Kraus (1994a), 23–4. There are arguably other ways to divide this narrative: textual segmentation is an interpretative act and as such, contestable. On the issues involved in dividing a text see Fowler (1995), Kraus (2002).

(16) Kraus (2009), 172. ‘Epic’, like ‘tragic’ and ‘dramatic’, is an overused term in historiographical analysis; by it I mean, that a siege looks back to epic precursors (above, n. 13).

(17) Interesting discussions of the varied use of time in (Greek) historiographical narrative are found in de Jong and Nünlist (2007), ch. 710 and 13; an influential treatment of narrative speed and duration is Genette (1980), 86–112.

(18) I owe this point to Chris Pelling; on the battle see Rice Holmes (1911), 790–801 and Stevens (1952), 17–18.

(19) On recurrent events in Caesar's narratives see Damon (1994), and cf. n. 12 above; on structure as argument see Batstone and Damon (2006), ch. 3.

(20) Christ (1974). Ariovistus and Vercingetorix are strongly linked in Caesar's large‐scale narrative, as are—consequently—Books 1 and 7; so too (less often remarked) are Orgetorix (the first opponent: BG 1.2–4) and Vercingetorix; see Torigian (1998), 56–8.

(21) In BG 7 alone, the claim occurs also at 1.7, 14.3, and 64.2. The very first resistance, led by Orgetorix, has similar confidence (1.2.2 perfacile esse … totius Galliae imperio potiri); see Torigian (1998), 55. On ‘easily’ in historiographical narrative see Rood (1998), 34 n. 30; on the locus a facultate (feasibility, including ease) see Heath (1997), 111.

(22) Among other things, their use of urbs instead of oppidum suggests grandeur; for urbs in BG see Bedon (1994). The Gauls were famous for talking and fighting (the two quintessential strengths of a hero: Hom. Il. 9.443 μύθων τε ῥητη̑ρ' ἔμεναι πρηκτη̑ρά τε ἔργων): cf. Cato Orig., HRR F 34: rem militarem et argute loqui with Williams (2001), 79–80.

(23) Hering prints O. Wagner's <so>luto frigore, but that figurative expression does not recur before Amm. Marc. 17.8.1; faute de mieux, I accept the OCT's reading here.

(24) e.g. Vitruvius 1.3.2, 2.8 (Scarola (1997), 195, 199–200). On the real thing, see Dehn (1960); on Caesar's description as an ekphrasis, see Dodington (1980), Scarola (1997).

(25) For plausibility as a criterion in historiographical narrative see the essential arguments of Woodman (1988), 231 (General Index).

(26) Caesar emphasizes the process of creation, not the object itself (Dodington (1980), 76, with an interesting comparison to Homer, Od. 5.243 ff., Odysseus' raft). Damien Nelis points out to me the density of words in 7.23 that are also at home in discussions of rhetorical construction (so trabes, ordo, singulae singulis, opus, species, uarietas, materia): Caesar is clearly playing with the idea of literary as well as defensive aedificatio. In both cases, Roman mastery is unaffected by foreign ‘expertise’.

(27) For the relationship between knowledge and power in Roman ethnography see the items cited in n. 5 above. Scarola (1997), 193 notes that after the wall, the topographical space at Gergovia becomes ‘tactical’—that is, Caesar can use it.

(28) See Rice Holmes (1911), 746–8, esp. 747 ‘Now Caesar does not guarantee the absolute and invariable accuracy of his description. He only professes to describe the general principles of construction’; Dodington (1980), 3 ‘The main problem with the assumption that [the engineering] passages function as historical descriptions is that they accomplish this task so poorly.’ For one possible (re)construction see Winbolt (1910), 54, 55.

(29) In this case he adopts the stance of the observing historian who deploys a typical historiographical justification for his selection of material. On dignum memoria and related expressions see Marincola (1997), 34–43; on singularis and the observer here see Kraus (2009), 161, 162.

(30) Guardians of the gates: Hom. Il. 12.127–94 (Polypoites and Leontes), cf. Virg. Aen. 9.677–8 (Pandarus and Bitias), Sil. It. Pun. 13.191–205 (Numitor, Laurens, Taburnus, cf. 199 tertius aptabat flammis ac sulphure taedas); live defenders replacing dead: Jos. BJ 3.23, Paus. 10.19.10 (Gauls); see Miniconi (1951), 175.

(31) For an introduction to the importance of place and space in literature see Mitchell (1980); for space in Latin historiography see Jaeger (1997), esp. ch. 1; in Caesar, see Rambaud (1974), Riggsby (2006), 24–45 and (on Alesia) Scarola (1984–5).

(32) See Cipriani (1986), 1–41 (a study of Alesia, concentrating on Critognatus and parallels to his speech), and on the motif see Stramaglia (2003), 17–23.

(33) Special emphasis is put on the pit‐traps, whose description—highlighted by the presence of three examples of (rare) sermo castrensis (7.73.4, 8, 9: Mosci Sassi (1983), 72, 77, 84)—balances the Avaricum wall. Like the wall, these have inspired recreations; see (e.g.) Bond and Walpole (1901), 52 and on the opera as a whole see Dodington (1980), 55–6.

(34) At 36.1, C. despairs of expugnatio and wants to defer obsessio; 43.5, plans to leave Gergovia however he can; 47.1 calls for retreat from walls; 53.1, still desires departure; 53.3–4 eventually decamps.

(35) See Walbank (1975), developed by Clarke (1999).

(36) On the Aeduan relationship with Rome see Woolf (2000), 290 (Index); I follow most scholars in referring to this episode as a ‘rebellion’, though see Dyson (1971) for a nuanced treatment. Chris Pelling points out that Aeduan adherence to Rome is a serious issue in BG 1 as well, forming yet another link between the books, and (in 7) helping close the narrative ring.

(37) The Aeduan situation causes Caesar to digress from his main plot both at 33.1 and at 40.1. For discedere of narrative transition, see OLD, 6c; for the phrase Caesar etsi, which often ‘initiates the reader into Caesar's complex thinking’, see Batstone (1990; quotation from 352).

(38) Like many such episodes in Caesar, the story of the Aedui both confirms and questions the stereotype of the disunified, quarrelsome Gauls. So here the avaricious, duplicitous Litaviccus is a fair representative of the stereotype, while Eporedorix and Viridomarus are not. On the stereotype see Jervis (2001); Williams (2001), 85, 90–3.

(39) This disposition of troops at 34.2 ~ 32.2 siue [hostem] ex paludibus siluisque elicere siue obsidione premere posset (‘to see whether he could lure the enemy out of the marshes and woods, or could oppress him with a siege’): Labienus gets the paludes (57.1–58.2), Caesar the obsessio (36.1). The subsequent expansion of the ‘table of contents’ at 34.2 is chiastic (Kraus (1994a), 22).

(40) Other such manipulations of the visual field are collected at Front. Strat. 1.5.

(41) On flumina, see Guillaumin (1987); for Caesar's preferred words see Richter (1977), 180–90. Any ancient military historiography can produce a similar effect, not least because topographical features tend to be described in conventional terms (Horsfall 1982, 1987).

(42) On rivers in Latin literature see Jones (2005), with extensive further bibliography, and cf. Munson (2001), 9–12 and nn.; for the symbolic use of them in military narrative see esp. Masters (1992), 150–78, and on objects in Latin literature see works cited in n. 14 above.

(43) When Caesar crosses the Elaver again (without incident) at 53.4, that neatly signals the end of the episode.

(44) For the fighting at Gergovia and the topographical problems involved see Rice Holmes (1911), 754–67 and Noché (1974); for Gergovia's afterlife see Dietler (1998).

(45) The Roman show here is countered by Gallic ostentatio at 53.3; I hope to explore elsewhere this interplay of showmanship and species in Caesar.

(46) This is the language of autopsy: cf. Liv. 6.1.1 res cum uetustate nimia obscuras uelut quae magno ex interuallo loci uix cernuntur with my note (Kraus 1994a). The Romans are also misled by appearances (similitudine), at 50.2.

(47) Caesar goes out of his way at BC 3.92.4–5 to stress the importance of charging with excited soldiers; his cohortatio at 45.7–9, albeit delivered to commanders, not to the rank and file, is remarkable for the ways in which it negates the norms of exhortation. On these see Yellin (2008), esp. ch. 1; Melchior (forthcoming) discusses Caesar's manipulation of the cohortatio form.

(48) Exact parallels at Liv. 24.40.12–13, Plut. Demetr. 9; for the narrow escape by a single leader cf. BG 1.53.2 (Ariovistus), Sall. BJ 101.9 (Jugurtha) solus inter tela hostium uitabundus erumpit (both similarly ending the battle narrative), Tac. Ann. 3.39.2 regisque opportuna eruptione. For the topos—which does not always involve a king—see Hardie's nn. on Virg. Aen. 9.189 and 314–66, and Martin–Woodman (1989) on Tac. Ann. 4.48.1.

(49) Historians are not entirely sure what, in fact, he wanted; but 47.1 proposuerat ~ 45.9 proponit clinches the impression that his design has indeed been completed.

(50) That battle serves as an explicit reference point at 47.6, 47.7, and 52.2; see also parallels given in the outline, p. 52, above.

(51) On the women in this episode, with many parallels, see Cipriani (1986), 43–73; the link with BG 1.51.3 ‘there they placed their women, who, with outstretched hands and weeping, begged the men setting out into battle not to hand them over into Roman slavery’, is another part of the 1 ~ 7 ring (above, n. 36).

(52) The triple repetition of murum closely centres our attention; see Konstan (1986) and Jacobs (2009), ch. 4 on thematically significant walls in historiography and epic.

(53) In their case, Caesar points us directly towards the previous vignette with 48.3 ‘[matronae] who had just now been holding their hands out to the Romans’; this shift from defensive to offensive action is picked up instantly in the narrative, which applies to the besieged adjectives proper to the besieger (48.4): ‘For the Romans it was an unfair contest, in terms of both ground and numbers. Tired by their climb and the duration of the fighting, it was not easy to withstand [Gauls] who were fresh and unscathed [recentes atque integros]’, cf. BG 5.16.4, Liv. 26.45.6, and cf. 6.4.10 with Oakley's n.

(54) For the attacker's body thrown over the walls, see B. Hisp. 15.6 (where it is identified as a barbarian custom), Liv. 41.11.5, Val. Max. 9.2.4, Jos. BJ 7.3.23.

(55) He leads a small group that sustains the charge of the enemy, allowing the main Roman army to escape (Cato, Orig., HRR F 83).

(56) I owe this point to Chris Pelling. These two centurions also form a pendant with the duelling pair in Book 5; on them see Brown (2004).

(57) The ultra‐Ciceronian rhetoric occurs at 1.5 communemfortunam, 2.1 communis salutis causa, 4.4 communis libertatis causa arma capiant. c. fortuna occurs as often in Cicero as in the rest of extant Latin; communis salus, lacking in Sall. and Tac. (and only twice in Liv., at 32.20.5, 22.6), is an emotive phrase (47 times in Cicero, 39 in speeches); with BG 7.4.4 cf. Cic. Rab. Per. 27 qui tum arma pro communi libertate ceperunt, Phil. 10.8 non suntinnumerabiles qui pro communi libertate arma capiant.

(58) This kind of self‐reflection on Roman imperialism is found repeatedly in Latin historiography: a famous example is Calgacus' speech in Tacitus' Agricola, but it dates well back, to Cato's Origines and his own speech on behalf of the Rhodians included therein (HRR 95). See Fuchs (1938), Adler (2005‐6), and Levene (Ch. 17 in this volume).