War Narratives

What emerged... was a myth - not because it was false necessarily but because it located an... event within routine structures of understanding.

(Biel 1996: 24)

There are many possible approaches to the BG as a piece of military narrative. I concentrate here on Caesar’s mythmaking: the ways in which he accommodates unfamiliar, unique experience in a distant land to patterns of speech and thought already understood and accepted by a Roman audience. Such an organization of experience into familiar, schematized categories of description and explanation does not necessarily imply that a narrative is fictionalized. It is relatively clear that Caesar did not systematically alter or propagandize events; especially where it can be confirmed by archaeology (e. g. Alesia: Le Gall 1999), the hard core of his narrative - topographical details aside - seems reliable. Nor do I intend, by this approach, to understand Caesar’s narrative technique as a means of tendentiously deforming events (as does Rambaud 1966), or even as a way of simplifying the strange for an audience who would have little idea about the geography of Gallia (Pelling 1981). Instead, I want to explore how Caesar’s decision to reduce the potentially infinite confusion and multiplied engagements of war to a few highlighted episodes harnesses a pre-existing grammar of military narrative to create a coherent, plausible literary representation of experience. In battle narratives, in particular, this grammar and its rhetoric are ‘‘not merely a machine to convert experience into words, but the very armature upon which that experience is organized and made sense of’’ (Lendon 1999: 274).

Hostes, milites, cives

Conventional military narrative divides its focus between the enemy and us, even when (e. g. in civil war narratives), in Pogo’s words, ‘‘he is us.’’ Caesar had a rich tradition of Gallic ethnography available: though our best attestations of it come from later texts, the stereotype of the fierce, loud, but ultimately unenduring Gaul can be glimpsed in the fragments of Cato the Elder and Posidonius, both of whom almost certainly served as sources for Caesar (Fantham, chapter 11, p. 151). He, in turn, combined autopsy - often filtered through the tall tales of soldiers (Aili 1995; Horsfall 1999) - with literary ethnography in creating his picture of the enemy. Remarkably, however (and despite the potential for creating fear and loathing in his readers), Caesar did not pull out all the stops in creating either his Gauls or his Germans. The opposition is factious, to be sure: this is one of the typical characteristics of barbarian nations, especially of the Gauls. It is emphasized in the BG by the ironic repetition of the phrase omni Galliapacata (and variants), a universalizing, closural statement that never in fact seems to stay closed, thanks to the endless splitting of Gallic resistance into new combinations. There is, in fact, no ‘‘Gaul in its entirety’’, much less a pacified one: the opening of BG 1 already encodes division, both of the people and of their description, as Caesar’s interwoven Gallia omnis and est divisa iconically demonstrates the impossibility of unification. Though the Gauls themselves deploy the omnis/tota Gallia rhetoric, notably when flattering the Romans (1.30.1), it is abundantly clear from the complexity of the various conflicts contained even in the brief BG 1 - the struggle with Orgetorix, the consequent bellum Helvetiorum, and the ensuing conflict with Ariovistus the German - that any hope of‘‘ruling all Gaul’’ ( totius Galliae imperio potiri 1.2.2) is as much a mirage for the Romans as it is for Orgetorix, Vercingetorix, and other chieftains. But it is also the force that drives the story: each unification leads to a further fragmentation of Gallia, which in turn requires further Roman intervention.

The seemingly endless tension between a unified Gaul and its component parts, then, becomes both a structuring and a thematic device. In other areas, too, Caesar avoids the unmotivated use of barbarian stereotypes. His northerners are not boastful, excessively ornamented, or comically oversized; the Germans at 1.39.1, ‘‘with huge-sized bodies,’’ also have ‘‘unbelievable courage (virtus)." Though the Gauls give their famous war-cry (5.37.3, 7.80.4), it is not made part of their formal ethnographic description; and both Gallic and Germanic leaders speak with nearRoman clarity and political savvy. What is more, on several occasions the enemy almost bests Caesar, who finds his army in tearful panic at the thought of the Germans (1.39), nearly routed by the Nervii (2.18-27), close to ignominious retreat from Britain (5.8-17), massacred in the forests (5.27-37), and, finally, trounced at Gergovia and surrounded at the climactic siege of Alesia (7.36-51, 81-8). Far from being all savage show and no reliable power, like the huge Gallic Goliath to Manlius’ tough little David (Claud. Quadr. 10b Peter), Caesar’s Gauls are a match even for his toughened soldiers. Most significantly, Caesar ‘‘refuses numerous self-created opportunities to draw a sharp distinction between Roman self and alien other'' (Riggsby 2006: 126). Mingling criticism of Roman self with praise of the alien other, the BG blurs any easy categorization, suggesting that in this imperialistic text, the way is already open for assimilation of the Gallic nation to the Roman state (and Caesar would soon extend citizen status to Gallia Cisalpina).

When it comes to ‘‘us’’ (nos and nostri) in ancient military narrative, the higher up the pecking order, the more text-time one gets. So the general, ipse, is the subject of most of the third-person verbs, acting representatively for the whole community of soldiers. Legates and - especially in Caesar - centurions are named, and occasionally given snippets of speech or writing; the Tenth Legion, Caesar’s most trusted (1.42.5), is once or twice allowed a voice, even to make a joke (1.42.6); the mass of soldiers, unindividuated, supplies the big picture action (e. g. 2.27.1 ‘‘our men... renewed the battle’’; 3.6.2 ‘‘the Romans drove the terrified remainder into flight’’) and crowd noises (1.39.4 ‘‘hidden away in their tents... they bemoaned their fate; 5.32.1 ‘‘from the nighttime brouhaha the enemy perceived their departure’’).

Though this literary arrangement misrepresents the very restricted perceptions available to any given actor in the melee of war and distorts the relative importance on the field of these various groups, it economically maintains a close, heroizing focus on the ‘‘best of the Romans.’’ This in turn fits the Roman conception of significant historical action as exemplary deeds done by men who are spectatissimi, both ‘‘most visible’’ and ‘‘most viewed.’’

If the Gauls are presented as fickle and mobile, the Romans in BG are ordered and - even more important - capable of being restored to order if need be. Individuals are clearly separated into wrong (Sabinus, the over-trusting legate who leads his troops into disaster in Book 5; Welch 1998: 95-6) and right (Labienus, a tactful treatment throughout), easily frightened (Considius, despite his experience, at 1.21-2) and brave (only centurions and junior officers; Welch 1998: 89). The degree of Caesar’s literary shaping can be judged not only by his schematization of character but by his selectivity in narrating events: he presents with concision a limited number of type scenes (e. g., battles, council scenes, river crossings, sieges: see further below) in which individuals are singled out only rarely.

Because he restricts the number of his building blocks, Caesar can make textual links correspondingly more clear by emphasis and careful deployment. Selectively elaborated scenes not only draw attention to themselves but respond one to the other across books, constructing a large-scale architecture and inviting readerly comparison and judgment (both desiderata of history: Cic. De or. 2.63). Roman actors play an important role in this elaboration. So, for instance, two good centurions fight an exemplary double combat against long odds, saving each other’s lives in the process (5.44). They are put in counterpoise with two disobedient centurions - one of whom manages to compensate for the foolishness of both, again, by saving Roman milites with his life (7.50). Pullo and Vorenus (Book 5) serve as contrast and as model for Fabius and Petronius, in a complex structure that challenges our first impression of these scenes as episodic, self-contained vignettes.

The last Roman character in the BG is the populus Romanus (see also Cluett, chapter 14, pp. 202, 204), effectively the group whose preconceptions and expectations serve as the ideal receivers of this text. Caesar orients us to the campaigns in Gaul via a mention of what the Gauls are called in ‘‘our’’ language, and contrasts the far-off Belgae with the cultus atque humanitas of the Provincia (~ modern Provence; cf. 1.1.3 with 1.2.4, the Helvetii divided from provincia nostra). Linguistic, geographic, and cultural differences are established (Schadee 2008b: 159-60). Those distinctions between Romans and barbarians begin to blur, however, when Gallic ‘‘friends of the Roman people’’ are introduced (1.3.4). Caesar is playing a complicated game involving physical, mental, and cultural distance. These foreigners have a past with Rome, a past that has the effect of collapsing the differences between them. Henceforth, numerous possible configurations arise: hostile Gauls and Germans may be amici populi Romani (cf. Ariovistus at 1.43.4); allied Gauls may (and do) revolt. And indeed, the future for most of these Gauls - those who are not massacred, that is - is to become part of Rome, even, eventually, Roman citizens (Riggsby 2006: 126-32; see also above, p. 166).

Negotiating these complexities is ‘‘Caesar’’ as the representative of the Roman people, a relationship neatly set up in the BG’s first narrated conflict. Bracketing his initial encounter with the Helvetii, Caesar evokes the memory of earlier strife between this nation and the Romans. At the start he refuses their petition: ‘‘because he remembered {memoria tenebat) that the consul L. Cassius had been killed and his army sent under the yoke by the Helvetii, he did not think he should yield’’ {1.7.4). That memoria, we will learn, is not only Caesar’s, but also belongs to patres nostri; conversely, the avenging of Cassius' death and his army's humiliation is not only a public act but one that punishes Caesar’s private injuries {1.12.3-7):

Caesar killed a great portion of them; the rest fled and hid in the nearest forests. This pagus is called Tigurinus; for the whole Helvetian state is divided into four pagi. This one pagus, after leaving its homeland, in the memory of our fathers {patrum nostrorum memoria) had killed the consul L. Cassius and sent his army under the yoke. Thus, whether by chance or the plan of the immortal gods, the part of the Helvetian state that had inflicted a famous disaster on the Roman people was the first to pay the penalty. And in this matter Caesar avenged not only public wrongs, but also his private ones, since the Tigurini had killed his father-in-law L. Piso’s grandfather, the legate L. Piso, in the same battle with Cassius.

The underscored sentence makes an explicit ring with Caesar’s own memoria of this same event {1.7.4, quoted above), marking the episode off as a textual unity with strong links to a past event. The post-battle wrap-up, normally devoted to casualty figures {e. g. 1.53.2-4, 7.28.5), here instead reverts to the past: the significant Roman dead are those killed, not in 58 BC, but in 107, Piso’s grandfather and the consul Cassius. The repetitions punctuating the paragraph {boldface), the prominent use of correlatives {italics), the precision of Piso’s naming, the double evocation of historical precedent, and an extremely rare mention of the gods {Fantham, chapter 11, pp. 142-3) highlight this firm statement of Caesar’s right to act for the Roman people, whose historically validated interests work in tandem with his own. Named multiple times in every book, but most often by far in BG 1, the populus Romanus is the backdrop against which this action is played. Rome, too, is the end to which it all tends. So we read at 7.90.8 {the last sentence): ‘‘when the events of his year became known at Rome, a 20-day thanksgiving was decreed in return.’’

The generals

Caesar gave his cognomen to two millennia of autocrats; his personal image may be unique in its projection of solitary genius. But his self-presentation in the BG is of a commander who works within the system. I have discussed his use of the third person and his relationship with his troops above; this section will consider other aspects of Caesar the character {‘‘Caesar’’).

Perhaps the first thing to notice about ‘‘Caesar’’ is that he is not the first person we meet in the BG. That honor is reserved for Orgetorix, whose story {1.2.1-4.4) serves as a prequel not only for the narrative as a whole but most particularly for Vercinge-torix, Caesar's most formidable enemy, whom we will not meet until the last commentarius (cf. the ring structure 1.7.1 ~ 7.6.1). ‘‘Caesar’’ responds (1.7.1). He enters the narrative not in the nominative but in the dative case, in answer to a request for help against encroaching enemies. That sets a pattern for his behavior throughout most of the story, and has presented his contemporaries (and later scholars) with the nice question of how much his war-making in Gallia - and beyond - was a response to real threats, and how much was imperialistic in intent from the start (Ramsey, chapter 4, and Rosenstein, chapter 7 in this volume). As it progresses, the BG spins an ever-growing web of pax Romana: defense of an ally creates new territory taken from the enemy; that territory is in turn threatened - and the imperium goes on. In this sense, the message sent by the BG is fully in step with expected Roman performance. ‘‘The central aspect of Roman strategy was image,’’ especially the inspiring of deterrent fear and the real possibility of vengeance, and the maintenance of dignity (laus) and effectiveness (utilitas) (Mattern 1999: 81-122, quotation from 122).

Being responsive to the needs of allies makes ‘‘Caesar’’ a good proconsul; winning battles against frightening foes and being able to speak persuasively to allies, soldiers, and enemy leaders alike show that he is good both at fighting and at speaking, the two primary virtues of a hero since the Iliad. The carefully constructed opening books bring ‘‘Caesar’’ gradually into the thick of things in a crescendo of action that puts all his virtues on show (see especially Schadee 2008b: 161-5). First to face him, the Helvetii are presented in the initial geo-ethnography as ‘‘outdoing the rest of the Gauls in virtus" because they fight almost daily battles with the Germans. Their story, from Orgetorix to the decisive battle at the Arar, is a microcosm of the narrative to come. Once they have been defeated - first in part (1.12-13), then in whole (1.24-9) - the Romans encounter the Germans themselves, rumors of whom terrify even the most experienced soldiers (1.39.1). At the beginning of BG 2, the next foes ‘‘Caesar’’ meets are the Belgae, introduced as fortissimi of all the Gauls because farthest from the Provincia (1.1.3). Thus the text in quick succession pits ‘‘Caesar’’ against the strongest of all the northern barbarians. On the way, he negotiates (1.7, 2.4-5), unmasks conspiracy (1.17-20), calms and exhorts his troops (1.40), and faces off rhetorically with the Helvetii (1.13-14) and with the fearsome Ariovistus (43-5). This testing in words and deeds establishes ‘‘Caesar’’ as exemplary in both virtus and responsive intelligence.

Both of those qualities are fully in evidence as ‘‘Caesar’’ moves to the natural extremities of Gallia, where lie huge rivers and mysterious seas. Topography plays various roles in this text. To begin with, in a way familiar from Herodotus on, its description, especially its names, validates the historian’s authority, using ‘‘the linear progression of troops criss-crossing the landscape to map out a geographic space whose... purpose is to convince us of the narrative’s... accuracy’’ (Cobley 1993: 45). Space itself can be presented as a synthetic vision of the whole of a region, as in the opening paragraph (‘‘geographic space’’), as a network of lines along which ‘‘Caesar'' and the army move - the course of a march (‘‘strategic space'') - and, in the setting for a battle alongside a mountain, marsh, or forest, as a precisely represented, three-dimensional space with boundaries and relief (‘‘tactical space’’: Rambaud 1974). Each kind of space presents particular difficulties, all of which are met by ‘‘Caesar’s’’ forethought; here, as in every area of the BG, ‘‘Roman rationality prevails’’ (Hall 1998: 22).

This is an understated text: no Alps to break down with Herculean efforts, no quicksand-like mud to drown men in, no epic-scale storms to grapple with. Each of these motifs features in surviving Roman historiography; my examples come from, respectively, Livy 21.34-7, Tac. Ann. 1.63-5, and ibid. 2.23-4. Though each performs a complex variety of functions, at bottom is, perhaps, their important role in affirming the grand style of historical narrative and characterizing its participants accordingly. Ever expedient, Caesar prefers to sketch in the possibility of such scenes. Lists of towns and tribes suggest epic catalogue (e. g. 7.4.6), but precise terrain, and the troop movements thereon, are extensively simplified (Pelling 1981). Mountains, forest swamps, and bodies of water are natural defensive boundaries that articulate and enhance the topography of Caesarian battle and march; they also become backdrops against which to display the bravery, strength, and endurance of general and men alike. The Cevennes mountains provide a chance to surprise the enemy with a heroic show of speed (p. 161 above); marshes (paludes) repeatedly threaten deployment and lives (e. g. 7.19); deep forests threaten winter camps (5); the crossing of a vast river astounds the local inhabitants (1.13.1-2); and the expedition to Britain founders after a storm that destroys the boats, leaving the Romans exposed on the beaches were it not for Caesarian ingenuity (5.10.2). Above all, the deeds done here and the virtus demonstrated are remembered by this text; Caesarian written space ‘‘takes on a memorial function’’ (Lendon 1999: 315).

Throughout this Gallic space, we see ‘‘Caesar’’ in the distance and in close-up, as it were (for the cinematic metaphor, see Lendon 1999: 317). His speed, strategic knowledge, and endurance show him to be an ‘‘ideal general.’’ His praiseworthy skills are seen also in vignette: energy, military knowledge, and competence in his famous gerundive-laden administrative aristeia at the battle ad Sabim (2.20); care for his soldiers in his several speeches before and after battles (especially 5.52); management of his troops’ emotions (especially fear: Lendon 1999); strategic ability in his deployment of troops and, most eventfully, of himself (‘‘Caesar’’ in color: 7.88.1). One key attribute of the ideal general is his ability to awe by merely appearing on the scene; the intensity of‘‘Caesar’s’’ presence is demonstrated even in his absence, when Labienus can conjure him up in absentia to act as judge and as exhortation to the soldiers fighting on the Seine (7.62.2).

But a key element of Caesar’s self-presentation in the BG is that he is not, in fact, alone. Two other leaders share his stage: the German Ariovistus (BG 1) and the Arvernian Vercingetorix (BG 7). We see less of Ariovistus; but his big meeting with ‘‘Caesar’’ is instrumental both in introducing Caesar’s use of significant landscape and in giving voice to barbarian libertas (Griffin 2008). It seems as if the colloquy will be evenly balanced between the Roman and the German commanders: they meet in a central space equidistant from each camp and from their escorts, who are themselves matched in number and rank - though ‘‘Caesar’’ has to provide his infantry with cavalry props in order to maintain the balance (1.42.4-6). Despite this preparation, however, their speeches are not balanced: Ariovistus’ is by far the longer (‘‘Caesar’’ is summarized: 1.45) and the arguments he makes are precisely those, one imagines, that Caesar himself would make were he in the same position. Ariovistus, then, can match, even outdo, ‘‘Caesar’’ in speaking (Christ 1974); unfortunately, despite the advance billing, he and his Germans are not a match for the Romans in fighting. His escape, in a little boat (1.53.3 navicula), has an element of the miraculous about it, presaging Jugurtha’s solo escape from the killing fields at Sall. BJ 101.9. It miniaturizes this great German, whose death Caesar ignores: his unnarrated demise is, by 5.29.3, a past matter of‘‘great grief’’ to the Germans alone.

Vercingetorix has a more elaborated history. He is, indeed, the best-defined of all the Gallo-Germanic characters, ‘‘the most Caesarian of all Caesar’s antagonists’’ (Adcock 1956: 54). Unlike Ariovistus, he is definitely a military match for ‘‘Caesar’’ - a match that is brought out spatially in the first half of BG 7, as the two generals track each other along first the Loire, then the Allier (for the campaigns see Wiseman 1994: 408-12). He has ‘‘Caesar’s’’ ability to speak persuasively (7.4, 7.20), and his speed (7.4.6, 29.5, 64.1), and is nearly his equal in commanding loyalty (7.30.1). He receives similar narrative treatment, as well: the wide sweep and the close-up vignette, especially with his army (7.20-21.1). Among other things, this parallelism with ‘‘Caesar’’ establishes the climax of the narrative, with the worst opponent positioned, and elaborated, last: in a familiar move, Caesar raises the enemy to the level of the Romans, enhancing his victory. That victory, like those in 57 and 55, was celebrated with a thanksgiving of unprecedented length (cf. 2.35.4, the last words of the book: quod ante id tempus accidit nulli). It effectively caps Caesar’s portrayal of this war as one that, above all, brought the Roman people where no Roman had gone before.

The work of war

Ancient military narrative, both poetry and prose, had well-established conventions of image (the gleam of weapons, the piles of corpses), character (the loyal sidekick, the gifted opponent), and scene (the flight of frightened rustics, the cohortatio, noncombatants on the walls). Caesar’s narratives repeatedly appeal, through their privileging of information gathered through first-hand knowledge, to the autoptic authority of the historian/commentator (cf. 1.22.4, a scout ‘‘had reported as seen something he had not seen’’). Yet, as is well known, Caesar describes with a wealth of circumstantial detail events he did not see (e. g. the battles in BG 5) and those he could not possibly have seen (e. g. the council inside Alesia in 7). It is easy to posit - though this still does not account for the precision of detail - that in those cases he was working from notes given him by his legates, or from information garnered later (cf. BC 3.57). Recent analyses, however, show that, like his portrayal of character (above), his often precise descriptions of military activities are shaped also by the literary conventions of war narrative, which tends toward type scenes. Since these are drawn from Roman readers’ narrative experience, Caesar can vary his treatment of a given type without significantly compromising the narrative’s impact. He can prune a march plus attack, for example, down to a single sentence: so 4.4.6 ‘‘having finished this whole march with the cavalry in a single night they overwhelmed the Menapii.’’ Or, he can give it a more leisurely treatment (5.28, the march to the Thames and defeat of Cassivellaunus), even a full-blown narrative with speeches, as he does at 7.57-62, describing Labienus’ march to and battle near the Seine. Again, the literary technique tends to distort the real action on the field; but it serves the wider purpose of presenting that confusion in patterns which allow readers to understand and interpret the represented experience. In what follows I briefly consider the siege and capture of cities, the pitched battle, and the engineering opus.

The urbs capta is familiar both from epic and from rhetorical theory: it is one of Quintilian’s primary examples of an embellished description that can make an orator’s speech emotionally persuasive. His account claims, moreover, that even the single word eversio can summon up in imagination a picture of the details attendant in literary descriptions of the event. Like other narrative elements, sieges and city-captures in the BG run the gamut from a one-word mention (1.5.4) to a brief narrative (2.12.2) to a full-scale description, as in the sieges of Gergovia and Alesia.

Quintilian identifies a package of elements found in the typical urbs capta description, including flames running through homes and temples, the crash of falling buildings, a sound made up of many shouts, people embracing their loved ones, the wailing of women and children, looting and pillaging, bound prisoners, and mothers defending their children (8.1.68-9). At Gergovia, for example, Caesar includes the shouting (7.47.4), women and children begging for mercy (7.47.5), the plundering enemy (7.47.7), mothers holding their children (7.48.3). The narrative is not, ultimately, pathetic (because these are Gauls, not Romans?), but the urbs capt(and)a scene is unmistakable.

Once again, the literary convention need not mean that the description is fiction: sieges are formulaic by nature, and their descriptions will be so as well (Roth 2006). But that Caesar deploys such scenes for more than purposes either of ‘‘simply’’ recording what happened or of ‘‘propaganda’’ is manifest from his inclusion of not one but two sieges in which the Romans are effectively beaten: the Gallic massacre of Sabinus and Cotta in Book 5, and the Roman attack on Gergovia, an extensively narrated siege from which ‘‘Caesar’’ has to withdraw, being driven almost to defeat by his own troops’ disobedience. Both are elaborated beyond what is even probable, and include details of the characters’ feelings that would have been inaccessible to Caesar. And though one can read these stories in ways that show ‘‘Caesar’’ to advantage (e. g., Rambaud 1966: 220-1), he was under no obligation even to include them (Pelling 1981: 741-2). Literarily, however, these episodes build tension and suspense; show the Romans, as represented by their commanders and centurions, at their most stubborn and strong-willed; and contribute to the overall structure of the grand narrative’s trajectory from victory to defeat and back to victory. All of these effects are economically illuminated by the delay that a siege necessarily introduces into a narrative. Further, the static setting of a besieged camp or town offers a place to stage significant events, from a teichoskopia to a fight before the gates, from the pathetic display of non-combatants to the virtus of Romans and enemy alike.

Pitched battles offer less chance for spatial concentration. The action can range over a potentially vast area, especially if one includes the pursuit of the routed, and can incorporate large spatial features, especially rivers and forests. That potential spread is typically played off against snapshots of individual action, such as a commander’s aristeia. Despite its multiplicity of parts and capacity to cover terrain, however, a battle ‘‘must obey the dramatic unities of time, place, and action’’ (Keegan 1991 [1976]: 16). Caesar achieves this by wrapping a battle narrative around a central, ‘‘Homeric’’ hero - usually ‘‘Caesar’’ - who focuses the action and imposes control on it (R. D. Brown 1999: 339). He can exploit the dramatic device of a peripeteia (e. g. 1.52, 5.48); or he may track separate contingents until they turn and, by constituting a physical center of action, bring the narrative to a climax (e. g. 6.34-40).

The last element to discuss here is Caesar’s war works: the descriptions of war engines, walls, camps, and bridges that punctuate the narrative. They are remarkable for their detail and frequency - no other Latin historian includes so many. The wealth and precision of detail in the descriptions has led many to try to reconstruct the objects. The bridge over the Rhine (4.17) has been the most popular, with simulations and actual reconstructions ongoing: I single out here the BBC program, ‘‘Secrets of the Ancients,’’ which in November 1999 tried, and failed, to reproduce Caesar’s feat. But attempts even to draw structures such as the tower at Massilia (BC 2.9) or the Gallic wall (7.23) have met with mixed results (Holmes 1911: 711-24, 746-8). One specialized study concludes: ‘‘there is substantial evidence that throughout the Commentaries the sections on engineering are not accurate and complete descriptions of the structures’’ (Dodington 1980: 5). As objects in the narrative, however, these opera serve a variety of functions. In one remarkable passage, describing the fieldworks at Alesia, Caesar employs bits of the sermo castren-sis, explaining for his non-soldierly audience the names of his defensive structures (7.73). Like the occasional moments in Homer where we hear the language of the gods (e. g. Il. 14.291), these instances of ‘‘soldier talk’’ emphasize our separation from the world of the BG at the moment of connecting us closer to it. The other opera, too, offer tantalizing glimpses of physical objectivity that turn out, on closer inspection, to be less useful as mimetic descriptions than they at first appear. But they do a tremendous job of representing physically the superiority of the Romans. In addition, like centerpieces, their ecphrastic detail highlights the writer’s ars (Scarola 1987). They raise questions of narrative and historiographical authority (Kraus 2007). Finally, like other triumphal narratives (cf. Plin. Ep. 8.4 on the bellum Dacicum), they offer future writers material for panegyric.