A fresh interpretation of Julius Caesar's Gallic War that focuses on Caesar's construction of national identity and his self-presentation.
Anyone who has even a passing acquaintance with Latin knows "Gallia est omnis divisa in partes tres" ("All Gaul is divided into three parts"), the opening line of De Bello Gallico, Julius Caesar's famous commentary on his campaigns against the Gauls in the 50s BC. But what did Caesar intend to accomplish by writing and publishing his commentaries, how did he go about it, and what potentially unforeseen consequences did his writing have? These are the questions that Andrew Riggsby pursues in this fresh interpretation of one of the masterworks of Latin prose.
Riggsby uses contemporary literary methods to examine the historical impact that the commentaries had on the Roman reading public. In the first part of his study, Riggsby considers how Caesar defined Roman identity and its relationship to non-Roman others. He shows how Caesar opens up a possible vision of the political future in which the distinction between Roman and non-Roman becomes less important because of their joint submission to a Caesar-like leader. In the second part, Riggsby analyzes Caesar's political self-fashioning and the potential effects of his writing and publishing the Gallic War. He reveals how Caesar presents himself as a subtly new kind of Roman general who deserves credit not only for his own virtues, but for those of his soldiers as well. Riggsby uses case studies of key topics (spatial representation, ethnography, virtus and technology, genre, and the just war), augmented by more synthetic discussions that bring in evidence from other Roman and Greek texts, to offer a broad picture of the themes of national identity and Caesar's self-presentation.
2006 AAP/PSP Award for Excellence, Classics and Ancient History
- The Social Life of Texts
- The Composition of De Bello Gallico
- Reality and Representation
- 1. Where Was the Gallic War?
- Types of Space
- Geographic Space in De Bello Gallico
- Tactical Space, Surveying, and the Possession of Gaul
- 2. The "Other" and the Other "Other"
- The Ethnographic Tradition
- Caesar's Ethnography
- 3. Technology, Virtue, Victory
- Siegecraft in De Bello Gallico
- Virtus in De Bello Gallico
- The Gallic Assimilation of Virtus
- 4. Alien Nation
- Playing the Cannibal
- Rhetorics of Empire
- What Is a Roman?
- 5. Formal Questions
- Who and What?
- To What End?
- Whose Voice?
- 6. Empire and the "Just War"
- The Theory of the Just War
- Just War Theory in the Real World
- Cicero's Textual Practice
- Caesar's Textual Practice
- 7. New and Improved, Sort Of
- Facing the Alternatives
- How Does Caesar Compare?
- Appendix A: Wars against "Barbarians"
- Apppendix B: Generals' Inscriptions
This book is a study of what is—in many senses—an already well-known historical event: Julius Caesar's De Bello Gallico, or Gallic War. To think of texts as events is certainly in line with various historicist tendencies in the field of Classics in general, but it is also an approach that has come to be seen as particularly appropriate to this work. For one thing, the direct evidence for De Bello Gallico is incomparably better than that for the Gallic War fought in the 50's B.C. We have the former actually before us (though not its prior composition nor its subsequent circulation). Slightly less obviously, however, we have much better controls for the War than for the War. After a long period in which Caesar was largely taken at his word, it became popular in the middle of the last century to try to find deceptions on the evidence of Caesar's own text. It is by now notoriously difficult to confirm or refute anything Caesar says. There are few other sources for the Gallic War, and none can be shown to be substantially independent of Caesar's account. Consequently, even disagreement with Caesar may be more a sign of invention or error in the historical tradition than of independent testimony.
For the text, on the other hand, many things can be brought to bear. Not only are there a few direct testimonia to its reception, but we also have a variety of different sources for how Romans might talk about the war and about the other topics of De Bello Gallico. Here it is enough to note the existence of contemporary texts such as Cicero's oration On the Consular Provinces, which treats Caesar's conduct of the war at some length, and Posidonius' anthropology (preserved only in fragments) of the Gauls whom Caesar was both fighting and describing.
This study has two roughly equal parts. The first, "external" part looks outward and considers the kind of Roman identity postulated by Caesar's work, particularly how it is constituted in the context of various non-Roman others. Here Caesar prefigures in small but important ways the coming Imperial order, establishing a link between empire and Empire. The second, "internal" part treats Caesar's political self-fashioning and the potential ends of his writing and publishing such a work. Of particular interest here is how (and why) Caesar persuades in a work that on its surface is so lacking in argumentation. Each part comprises case studies on key topics (spatial representation, ethnography, virtus and technology, genre, the just war) that are followed by a more broadly synthetic conclusion, filling out a broader picture of the part's topic (national identity; Caesar's self-presentation) and setting these results in the context of other current scholarly work. Thus, while I focus on the single work, I hope to provide conclusions about the interaction of that work with its cultural and political environments.
The Social Life of Texts
None of the case-study topics is entirely new, and two of them (ethnography and genre) are arguably mainstays of writing on Caesar. Substantively, however, their specific conclusions are largely new and gain by combination into the two parts just described; they are mutually illuminating in ways not possible in previous (often very insightful) studies on those topics individually. Moreover, all seven chapters share a methodological coherence, and, I think, a methodological advance. The key notion here is "intertextuality," the idea that texts ultimately and necessarily take on their meanings by comparison and contrast with other texts. (The following account of intertextuality is, I think, important to ground the study as a whole and to justify several lines of argument. However, I do not introduce any special vocabulary or the like that would be necessary to the understanding of the rest of the work. Some readers might find it more useful to skip now to the next section of the Introduction, and return to this theoretical discussion after the rest of the book.)
Specific definitions and extensive theorization about "intertextuality" have taken many forms over time and across scholarly disciplines. Many of those understandings of the term are dependent on fairly specific linguistic and philosophical commitments, many of which are in turn debatable and are at any rate hardly shared by all theorists of intertextuality. I here rely mainly on three propositions that, though hardly novel, I hope will be largely uncontroversial.
The first proposition is that the meaning of words is constrained by the fact that they are common property; they are known through usage, and no stipulation can entirely free them from that history. To take an extreme example, some older dictionaries offer "holocaust" as a translation for Greek empura, "burnt sacrifice." Yet this is almost impossible for contemporary students who have heard much more about Nazi Germany than about animal sacrifice, even if the term is explained to them. Description of a ballistic missile defense system as "Star Wars" effectively conveyed the ambition of the program (by comparison to the movie's epic character), but also made it hard to avoid a sense of mockery (since the movie was overtly fictional).
The second proposition is that the reciprocity of language production and learning will produce a considerable degree of intersubjectivity among actual communities without requiring us to posit any ideal form of language. Different speakers of a language operate with slightly different rules (most obviously slightly different vocabularies), which linguists call their individual "ideolects." These ideolects arise in roughly similar circumstances; English speakers grow up hearing English sentences. Moreover, speakers implicitly compare and test their individual versions every time they attempt to communicate with others. Thus, although two different speakers almost necessarily have different primary linguistic experiences (in detail), they will over time tend to share more and more background, at least indirectly. Thus there are many different "Englishes," each material and at least potentially well-ordered as part of the neurophysiology of various speakers. To assert the existence of English in general is then shorthand for an essentially sociological claim about the potential for communication between these speakers.
A similar argument about acquisition could be made for many other kinds of cultural knowledge (e.g., greeting rituals), and that leads to the third proposition: constraint of language by shared history occurs not just at the level of words, but also at higher levels of organization. Minimally, speakers share fixed phrases, some with idiomatic ("hit the showers") or technical ("malice aforethought") meanings, others not ("animal instinct"). The phrase "this space intentionally left blank" not only comes as a unit, but also suggests the whole world of bureaucracy. Audiences recognize verse forms, so that, for instance, a reader of this stanza—
Thales' theory, to quickly review it,
Is that everything's made from a fluid.
How it's done, he'd not venture,
Though he'd say, facing censure:
Don't forget steam and ice somehow do it.
—is likely to view the information it contains with some suspicion, since the limerick form is typically used for jokes.
The verbal complexes with which I am most concerned in this book are particularly large ones, which I describe as "discourses" and (to a lesser extent) "genres." I intend "discourse" here in a fairly general way: a way of talking about some subject matter. That would include characteristic vocabulary, metaphors, themes or parameters, omissions, or procedures for assessing individual statements. A "genre" will be a pattern of associations between features of form, content, and context/occasion of verbal production. This includes both "literary" genres (e.g., epic, the novel) and nonliterary ones (newspaper story, conversation).
Both discourses and genres share a number of features. First, they have the same dual character as languages in general. For individuals, they are concrete and can (though need not) be stable. In the abstract, they are sociological fictions. All versions of "epic" are likely to be similar; no two will be exactly alike. Second, both genres and discourses can overlap and/or encompass other genres and discourses, respectively. For instance, a discourse on "war" might contain discourses on "tactics," "courage," or "divine favor." The latter might additionally be part of a discourse on "philosophy." That in turn might be cross-cut by "Stoicism," "Epicureanism," etc. Discourse and genre can also overlap/encompass each other in the same ways. War discourse would appear not only in military manuals and actual commands, but also in epic and history. Third, both are in principle subject to a producer's conscious control. It is well established that, say, an Ovid can carry out elaborate manipulations of his readers' generic expectations. Similarly, one could choose to discuss "war" in nonstandard terms. Such variation, however, is constrained by intelligibility. Too great a variation has the same effect as making up one's own words. So, for instance, I argue that Caesar's text is largely typical "war" discourse, but that it redefines one of its key terms, virtus (Chap. 3 below; contrast Chaps. 6 and 7). Here the modification depends on exploiting a contradiction already existing within the tradition. Conversely, Caesar pointedly avoids the discourse traditionally surrounding northern "barbarians" (Chaps. 2 and 4 below).
Though the term "intertextuality" has been in vogue for some time now, and is perhaps already losing favor, the idea remains underexploited, as argued by Fowler. (In fact, I suspect recognition of the gap between awareness and use has itself become something of a topos, though without producing much change.) Far the most common case in which classicists invoke the notion of intertextuality remains allusion in some passage of poetry (say, the opening phrase of the Aeneid) to an earlier passage in a similar context (say, the opening of the Odyssey). My plan is to exploit the broader range of the term in three ways, albeit with some overlap.
First, in probably the most common extension of the prototypical procedure, many of the relationships discussed here cross genre boundaries. Of course, it is legitimate, sometimes even necessary, to compare Caesar's work to that of other "historians" (though see Chap. 5 below on the sense of that term), but there is much to be gained from looking further afield to oratory, geography, surveying manuals, and others.
Second, since intertextuality is a general property of language rather than of literature, it can be just as important for prose as for poetry. Now, it is perhaps increasingly common to recognize prose influences on poetic texts; see, for example, Thomas 1988 on Vergil's use of agricultural texts in the Georgics. And historians are known to have responded to their predecessors in ways beyond mere collection of source materials. Nonetheless, both instances tend to be treated as isolated, artistic phenomena rather than as a normal feature of prose texts.
Third, and most important, I am interested here not in Caesar's reference to specific passages of specific works, but in his relationship to entire discourses. The existence of a discourse on some topic creates what might be called a "field of positions," a set of distinctions, contrasts, axes, and/or spectra with respect to which terms are defined and positions taken. Segments of Caesar's text (or any other) take on meaning by their locations in one or more of these fields. In one sense, such an appeal to a broader "interdiscursivity" is a common move. It is not, after all, entirely unlike appeals commonly made to "the (ancient) context" of a work. Yet this conventional formulation has unnecessary limitations. While not denying the theoretical possibility of multiple or multivocal intertexts, it tends to fix on one. Take, as an example of skillful application of the traditional method, the philosophical contextualization of the end of the Aeneid. Galinsky starts by detailing the diversity of ancient philosophical opinions on anger (and Vergil's incomplete adherence to any of them). Yet, in the end, the effect of Galinsky's interpretation is to downplay the "pessimistic," Stoicizing reading that Aeneas should not have lost control at the end of the poem. More generally, this kind of "contextualization" is usually used to limit meaning: "This passage must mean A, not B, because the latter is anachronistic." Modern theory would suggest that appeal to intertexts can open up readings, but not close them off.
More positively, a modern notion of intertextuality makes it easier to account for three phenomena that will be observed in the course of this study. One is a style of naturalization. A text (or passage) that is written according to the standard rules of some recognized form is more like to gain at least provisional acceptance, since it will be at least formally plausible. To give a non-Caesarian instance, consider Mader's recent (2000) reading of Josephus' Jewish War in the light of classical historiographical intertexts (primarily Thucydides). On this reading, Josephus is "concerned to dissociate the rebels from the traditions of Jewish piety [and] plays down, refracts, and filters out this religious dimension by applying the political and psychological categories of Greco-Roman historiography." That is, classical historiography favors (and leads readers to expect) certain explanatory gestures and categories of analysis. Hence, faction/stasis, demagoguery, and tyranny easily replace eschatology and religious traditionalism, without the need actually to argue for the former set of descriptions. Similarly, I here argue (Chap. 5) that Caesar's choice of the commentarius form and perhaps the appearance of "Gallic War" in its title make natural the exclusion of much contemporary material (politics back at Rome, Caesar's nonmilitary activities in Gaul). This allows him to omit much that would potentially have been controversial, and to focus on circumstances in which he is opposed by armed foreigners, maximizing sympathy for himself.
Another phenomenon is the possibility of feedback among several texts, absent a text/context distinction. For instance, the Latin literary letter took on a number of forms and self-definitions from the time of Cicero and Caesar to that of Pliny. Starting from Fronto, however, the genre began to see its origin and model in the collection of Cicero's letters. For some of these later letter-writers, the Ciceronian model must have been immediately useful, but, as far as we can tell, it ceased to become a choice. Even writers who were anti-Ciceronian in one sense or another had to deal with the "fact" that they were now writing in a Ciceronian genre. This fact was created by the accumulation of individual, increasingly constrained choices until non-Ciceronian readings were driven from the field. Moreover, Seneca's and Pliny's letters were retroactively read into this tradition. The meaning of these texts was shifted by the creation of subsequent intertexts. Conversely, I argue below (Chaps. 6 and 7) that De Bello Gallico and other generals' narratives gained authority by their mutual reinforcement.
Interdiscursivity is also important to De Bello Gallico because it creates covert argumentation. The narrator of De Bello Gallico does not make arguments. He rarely even offers explicit judgments of the sort, "This was treachery," or "Caesar's decision proved to be wise." Outside of a few speeches by characters, the text is narrative and (to a lesser extent) descriptive. It is punctuated only by rare and quite general sententiae ("sound bites"). But by their location in a field of positions created by other texts, descriptions and narratives can become argumentative.
Let me give non-Caesarian examples of both cases. Roman literature is very familiar with the embedded narrative form of the exemplum. For instance, Cicero asserts that suicide is normally a moral error, but can be correct or even an obligation for persons who have led a particularly rigorous life (Off. 1.112):
Since nature had given Cato an incredible gravity and he himself had fortified it with perpetual self-consistency and had always held to whatever course of action he had taken up, he had to die rather than look on the face of a tyrant.
Cicero here assumes that suicide is contrary to human nature and therefore wrong. But Cato, he claims, is an exceptional case because of his long history of self-consistency; he may and even must violate the normal rules with respect to suicide. Does his past life really explain this exception? Surely the serial killer is not ipso facto licensed to murder. Cicero has no philosophical argument how past life could create exceptions to more general rules. Rather, the conventions of exemplary discourse provide the justification. Cato did it, so it must be right. Sometimes, there is not only no argument, but also no explicit moral, as when Cato is also cited for rigor in making sure that his son was properly and personally enrolled in the army before joining combat with an enemy (Off. 1.37). The Roman reader of exempla knows not only that Cato's actions are normative, but that their salient aspect will have to do with moral punctiliousness.
The same is true of descriptions. So, for instance, Cicero frequently explains his enemy Piso's political success by reference to his eyebrows. Sometimes he explains that Piso's brow gave him a grave and serious appearance; sometimes he does not. General knowledge of Roman physiognomic discourse is what allows the omission; most of the audience could fill in the blanks. Similarly, when Cicero remarks that Catiline's followers wore "sails rather than togas," he has no need to fill out the syllogism: therefore they were effeminate, therefore they were politically and socially untrustworthy.
I have chosen the above examples of "covert" argumentation from overtly argumentative contexts: philosophy and oratory. We know from the broader context that these descriptions and narratives should have an argumentative point. To recover what that point is, we need to refer to other texts. (The texts we have today are, of course, just a small fraction of the oral and written discourses available to the original audience, but the mechanism is the same.) Hence we can be sure that such interdiscursivity was one means for Roman authors to generate argument. But nothing prevents the same mechanisms from working in less overtly argumentative texts.
Given that such arguments only come into being via the contact of at least two texts (or discourses), it is probably easiest to think of intertextuality in general as an element of the reading process, as part of interpretation. There are, however, two caveats that should be offered. First, the author (here Caesar) is himself a reader. Presumably, one part of deciding what is to be written is weighing, at various levels of consciousness, how it might play against various intertexts. Second, as I suggested above, knowing the language means sharing intertexts with the author to a significant extent. And to the extent that a reader is closer culturally to the author (as Roman aristocrats would have been to Caesar), they will share more. Still, perfect unanimity of readings is unlikely. It hardly needs pointing out that no two readers will have precisely the same previous experience. On the other hand, actual intention may cause potentially salient intertexts to escape an author's notice.
Caesar's famed celeritas may provide a quick example. I argue in Chapter 1 that Caesar tends to depict Gaul as a series of unconnected spatial "islands," and that this has certain consequences in the context of the more general Roman spatial imagination. One textual feature that emphasizes this depiction is the lack of detail in narrating Caesar's journeys from one point to another. This feature of the text is perfectly well motivated as an advertisement of Caesar's swiftness and decisiveness, and one could easily imagine that it was so written for precisely that reason. Yet it also plays into the scheme of division, which is otherwise visible in the text. For at least some readers, then, that lack of detail is likely to emphasize that spatial scheme even if such emphasis played no part in the author's intention.
On the whole, I am inclined to suspect that most of the effects I argue for in De Bello Gallico were part of Caesar's intention (whether that amounts to elaborate planning or just an intuition of what "sounded right"), but I will not generally be arguing in those terms. In part, this is because I do not find "intention" a useful explanatory term. Even if it were, however, my intentions in this book are ultimately historical. I am concerned with the likely or possible effects of De Bello Gallico rather than its meaning in some potentially pure sense. Thus a reader-oriented focus, though not required on general theoretical grounds, is appropriate here.
The Composition of De Bello Gallico
Many basic questions about the composition of De Bello Gallico have remained open despite extended scholarly discussion. As for the time of composition of De Bello Gallico, one camp has maintained that it was composed and circulated book by book, that is, year by year, presumably having been written in the midst of Caesar's other administrative duties in the winter after each campaigning season. The other camp holds that it was written all at once (though probably incorporating earlier material, such as dispatches to the Senate). Various specific times have been suggested, but a date between very late 52 and sometime in 50 is generally accepted. Though a few core arguments have been advanced on either side, interpretation of the evidence continues to be problematic.
Purported anachronisms and self-contradictions have been used to demonstrate unitary and serial composition, respectively. A small number of clear examples might be decisive, but surprisingly few candidates have been proposed in either direction, and none is obviously dispositive. On the one hand, for instance, Caesar refers to the near-total destruction of the Nervii in 57 (2.28.1: prope ad internecionem), but three years later he had to confront a force of (allegedly) 60,000 men led by the Nervii in their own territory (5.49.1; cf. 5.39.3, 7.75.3). There is certainly a prima facie contradiction, but, even if we ignore the question of how many actual Nervii were in the force they led, we cannot be sure that the earlier claim was not meant as hyperbole for political or literary reasons. On the other hand, some have pointed to 1.28.5, "[The Boii] to whom [the Aedui] give lands and whom they afterward accepted into the same state legal equality and freedom as themselves." The "afterward," it is alleged, refers to a time outside 58, and therefore suggests composition after the fact. Some have questioned the authenticity of the "and whom . . ." clause (see below on interpolation), but, perhaps more important, the temporal reference of that clause is still unclear. The Boii are still dependents of the Aedui early in the last book of De Bello Gallico (7.10.1), and it is only plausible that they are liberated at some point later in 52; the later references are not explicit (7.17.2, 7.75.4). Anachronism in the first book is certainly a possibility, but it could also be that the legal equality of the Boii came early, yet did not give them practical political equality.
Most readers of De Bello Gallico sense stylistic and substantive development over the course of the work. Most objectively, the quantity and importance of direct discourse grow over the second half. Additionally, Görler has argued that the narration becomes less character-focalized and more "Olympian." It is often claimed that what starts out as a commentarius ("commentary"?) becomes more like "true" history (see Chap. 5 below on questions of genre). Such stylistic development is taken to support serial composition.
Those who argue for unitary composition have two responses. One is to quibble with the premise of development. It is often possible to show that Caesar's changes in style are not entirely smooth or regular, but little has been done to shake the basic claim that many parameters vary in fairly predictable ways as the work goes on. Alternatively (and this strikes me as the more powerful argument), the developments can be seen as having literary aims. Thus, if von Albrecht (1997, 332-333) is right that the changes in direct discourse in De Bello Gallico are replayed over the course of Caesar's later Civil War commentary, it is hard to attribute either development individually to mere change over time. Mutschler has made the same argument for a number of other features. Similarly, I argue below (Chap. 3) that some kinds of "progression" in Caesar's knowledge of the Gauls are thematically significant. Moreover, since the reader of a continuous text would learn and change by virtue of the reading, Caesar might well have made adjustments to the texture of his text accordingly.
Caesar's continuator, Aulus Hirtius, makes two remarks that are sometimes taken as evidence for the composition of De Bello Gallico. Unfortunately, neither is very explicit, and the inferences conventionally drawn from them are opposed. In the prefatory letter to Book 8, he remarks to Balbus (8.pr.6) that only they really know how great a writer Caesar was: "Others know how well and correctly he wrote; we also know how easily and swiftly he finished those books." This has suggested to some that all of Books 1-7 were composed together in a fairly short period of time in 51, but obviously Hirtius could just as well be referring to the composition of individual books in, say, just a few days per year each. Later he justifies the inclusion of two different years in one book (8.48.10): "I know that Caesar produced individual commentarii of individual years. I thought I ought not to do that because nothing important happened in the next year." Now, it is possible that the individual years (singulorum annorum) to which he refers here are actual years, and so Hirtius is alluding to serial composition. However, both the construction and the general sense of the passage require that he be talking about years primarily as units of composition—separate campaigning seasons. Not only is neither passage decisive, but it is far from clear that either bears on the question at all.
Finally, advocates of unitary composition point to the absence of testimonia to the existence of the commentarii before Cicero's Brutus, from the year 46. In particular, neither Cicero's speech On the Consular Provinces (in large part about the war) nor his letters to members of Caesar's staff make explicit reference to books of De Bello Gallico appearing during the war. The facts here seem not to be in question, but the interpretation is more problematic. We have substantial evidence of regular dispatches from Gaul to keep Caesar in the public eye. Does it matter that we are not told explicitly that commentarii were involved?
To my mind, then, the question of the timing of composition of De Bello Gallico is one where we are left assessing comparative probability, not established fact. Moreover, we must keep in mind that at least some of the purported evidence (for instance, Hirtius' two statements) can only be shown to support a particular position if we presuppose that it is salient to the debate at all (unlikely in those cases). This kind of information should probably not be allowed at all in the weighing. Nonetheless, I am inclined to accept the theory of serial composition, simply because of the obvious value to Caesar in keeping the public aware of his deeds throughout the war. This historical consideration seems to me to tip the scales where the philological arguments are roughly equal and quite weak on both sides. This weakness is also the subject of one further observation. I argue in Chapter 5 that, for strategic and generic reasons, Caesar wants to give the impression (whether true or not) of writing as he goes, not just year by year, but almost line by line. (More precisely, I will suggest that the choice of genre is in part a way of advancing those strategic aims.) If this is the case, then Caesar may be deliberately writing in a fashion that would (perhaps less deliberately) neutralize internal evidence for a distinction between serial and unitary composition.
Another of the much-debated traditional questions about the composition of De Bello Gallico has to do with interpolation. All classical texts handed down in manuscript traditions are vulnerable to addition to some extent. "Corrections" can be made deliberately, or, perhaps more commonly, marginal and interlinear notes can be incorporated into the main text by an incautious scribe. In the case of De Bello Gallico, however, it has been suggested that long passages were deliberately composed and inserted in the text in (perhaps) late antiquity. There is considerable variation in just which passages are suspect, but what is essentially at issue is the authenticity of the various geographic/ethnographic excursuses throughout the work.
In part, the question arose from a hyper-skepticism common to textual critics in general in the late nineteenth and early twentieth centuries and should perhaps simply be ignored today. Moreover, even on its own terms, the interpolation theory was never fully worked out. How and why were additional passages composed and inserted in the tradition? Take the "double" description of Gaul in 1.1 (to be treated at length in Chap. 1 below). The versions are similar enough that it is not clear why an interpolator would feel the need to insert material, but too far apart to suggest the intrusion of a marginal gloss. Contrast the case of interpolations in the legal texts that make up the Digest of Justinian. Mechanisms (including a known redaction) and a motivation (keeping law current) are clear; yet even so, scholars are much more cautious today in identifying interpolations in the Digest than they once were. Nonetheless, the idea of interpolation in De Bello Gallico may not be quite dead yet, and I make considerable use of some of the allegedly interpolated passages (especially in Chaps. 2 and 3), so it may be worthwhile to say a few words about their authenticity.
There have been both stylistic and structural arguments for interpolation. The geographic and ethnographic passages are unquestionably different from the rest of De Bello Gallico in, say, vocabulary, but these differences are simply those of content or genre, not of authorship. Any author switching from narrating battles to describing giant, kneeless elk will show changes in style. The structural argument is that these passages break up the narrative. Not only are they "interruptions" in the general sense of being descriptions in the midst of a mainly narrative text, but, it is alleged, they would leave behind a seamless narrative if they were removed. For instance, in Book 5 we read a description of Britain and its inhabitants (5.12-14) just after being told that Cassivellaunus was put in charge of the entire war effort against the Romans (5.10.9). Immediately after the excursus, we are on the battlefield, getting a description of the disposition of the British forces (5.15.1). On the one hand, it is probably true that the excursus would not be missed if it were removed. On the other hand, it is well motivated at the opening, and not hard to follow at the close. Where there is a potential problem, as at the opening of the Gallic ethnography (6.11.1), the reader is warned about what is going on. If the "digressions" are arranged so as to minimize narrative disruptions, it is not clear why that should be attributed to interpolators rather than to Caesar himself. Recall, moreover, the write-as-you-go style I suggested that Caesar had adopted in De Bello Gallico. Given that choice, Caesar would have been forced either to incorporate visible narrative breaks to provide these descriptions or to forego them altogether. The latter solution would not have been surprising, but the former is hardly so incredible that we should suspect tampering with the text instead.
The third compositional question, and a more recent one at least as a matter of overt debate, has to do with audience. Most scholars seem to have assumed an elite audience for De Bello Gallico, perhaps senators and equites for the most part. Wiseman has recently challenged this view, suggesting a popular audience not only in addition to but instead of the elite one. We would then have to imagine large-scale oral performances of De Bello Gallico. Although Wiseman has not proven his case, the question deserves consideration. The crucial, if general, passage purportedly illustrating popular interest in such works comes from Cicero (Fin. 5.52): "What about the fact that men of the lowest station, with no hope of a public career, even craftsmen take delight in history?" Wiseman reasonably argues that Cicero's off-hand tone shows that he expected the proposition to be uncontroversial, but problems remain.
First, while Cicero may be sincere, we do not know whether he was expert in popular culture. Moreover, although Cicero's historia is less ambiguous than the "history" I have used to translate it, it is not entirely clear whether in this passage it refers to the literary genre or to any knowledge of the past. This raises two questions. First, how did ordinary fans of history like their history delivered? Readings of literary historians? Plays? Art? Stories told round the campfire or hearth? All of these forms are substantially better attested than full-scale historical narrative, whether read to oneself or to a group. Second, and more fundamentally, would the audience envisioned by Cicero, even if they did have a taste for Roman history, have found De Bello Gallico satisfying, given that for them it was a narrative on current affairs? In modern American terms, are Caesar's potential audience members Civil War buffs or CNN junkies (or something in between)? The (again limited) evidence appears to argue for the former. This is not necessarily to deny political interests to the masses. Some significant (if not necessarily representative) fraction must have been keenly interested. Rather, the point is that there is no evidence that popular taste in literature and popular interest in politics had any connection.
Wiseman also finds internal evidence for a popular audience, both in the respectful treatment of soldiers in the ranks (as opposed, in some cases, to aristocratic staff officers) and in the "talismanic" repetition of the phrase populus Romanus. Yet, as Welch's careful study (1998) has shown, the officer class is on the whole treated reasonably well. As she puts it, "No legate, with the exception of the dead Sabinus, could complain about what Caesar says about him." Even the soldiers could be questioned (2.8.1-2), or publicly reprimanded if necessary (7.52). If officers are slightly more often criticized, it may well be for internal reasons. As I argue in Chapters 3 and 4 below, Roman generals traditionally distanced themselves from the worth of their troops (mostly to deny responsibility for their failures). Caesar, however, presents a picture of battle in which the soldiers are dependent on their general for their prowess. Since one of Caesar's implicit claims to distinction lies (unusually) in the virtue of the ranks, he has good reason to praise them, whoever the audience may be.
The repetition of the phrase "Roman people" raises some interesting questions. It might be suggested that the crucial word here is "Roman" rather than "people," and that Caesar is merely aligning himself with national as opposed to personal interests (cf. Chap. 5 on how many of his narrative choices are directed at a similar effect). It would probably be incautious to throw out the other word entirely, but we might inquire further as to the meaning of the phrase as a whole. If the whole nation comprises the "Senate and People of Rome," then the people are in principle everyone but the Senate, that is, a group that extends across all meaningful class (if not juridical) boundaries in Roman society. So, for instance, even on the most conservative estimates of literacy, the vast majority of the reading population of Italy would have been part of the "Roman people" in this sense. This group (literate nonsenators) could easily have had more political clout than the urban proletariat audience Wiseman imagines. They would have been better spread out throughout the voting units of the Roman assemblies. They would also have been largely well-to-do, and thus in a better position than most to go to Rome to vote. Moreover, this group may well have been not just part of the Roman populus, but the whole of what many members of the elite meant by that group. Mouritsen points out that the term could be used of any number of official and semi-official gatherings of potentially highly unrepresentative fractions of the populace. We might reasonably accept Hall's view of a substantially "middle-brow" "middle Ital[ian]" audience if we keep in mind that, due to extreme inequalities in social condition, a group in the "middle" in terms of juridical strata (senators, knights, . . . slaves) was part of a cluster of very small absolute numbers at the top of the pyramid. It is possible that Caesar had in mind a universal, or even a strictly non-elite, audience, but the evidence is not compelling. For the time being, it seems best to retain the assumption that Caesar is aiming at the top of Roman society, though perhaps at a slightly larger segment of that top than is sometimes implied.
In addition to their social location, we should also consider the level of sympathy of Caesar's target audience. This can only be inferred from internal evidence, and in particular by working back and forth between interpretation and potential audience. Thus I will be developing my answers in the course of the book. However, it may be helpful if I briefly sketch some positions here. Let me distinguish two parameters: strong commitment for or against Caesar vs. more malleable attitudes, and sympathy for Caesar vs. that for the general project of a Gallic War. Regarding the first distinction, Hall is surely right to point out that there is nothing Caesar could have written that would have won over a committed enemy like Cato. We might add that there would be just as little reason to target firm allies. I argue in Chapter 7 that Caesar has in mind a public that is not automatically committed to him, but would like to be given a reason to believe he was acting well and correctly. As for the second point, Chapter 5 below will show that Caesar was taking steps to emphasize the Gallic War (even beyond his own actions) over his personal career (beyond this war). This suggests (probably not surprisingly) a larger audience that wants to root for the Romans against the Gauls than for Caesar against his political rivals. This is not, of course, to suggest that Caesar is not trying to sell himself, only that patriotism is a crucial part of his strategy for doing so.
Reality and Representation
In the first words of this chapter I appealed to the notion of the De Bello Gallico as a historical event. Some readers will, quite reasonably, be reminded of the distinction commonly drawn today between the study of real events and the study of representations. Even if we take that distinction at face value, it is hard to apply here. On the one hand, this is a study of reality, as the production and circulation of objects—including representations—are real events. On the other, much of that history will be inferred from the character of this and other representations. Given that complication, it may be worthwhile to take a closer look at the original distinction through examination of a case whose themes will recur in the chapters below.
Diodorus reports that the Celts were very fond of wine. They drank it unmixed and in excessive quantities (5.26.3). The Greeks and Romans generally drank their wine mixed with water, so to drink it straight could be seen as just another, more indirect way of expressing barbarian excess above and beyond explicit reference to sheer quantity. Posidonius (preserved in Athenaeus) similarly claims that Celts normally drink their wine unmixed (4.152c). Both authors also stress that the wine had to be imported from Italy, or at least from Marseilles. One might expect under these circumstances that wine would be relatively more expensive in Gaul than in Greece or Italy, and this is precisely what Diodorus claims. Delighted Italian wine merchants, he says, can trade a slave for a single keramion of wine. Nor is distance necessarily the only factor in pricing. While it is probably meaningless to try to establish a "normal" ratio of wine to water in the Greco-Roman world, recorded figures tend to fall between 1:3 and 1:1. Drinking unmixed wine in full quantities would be proportionately more expensive. The combination could make wine a vastly more expensive beverage for the Gauls, and this in turn might give it the status of a luxury item, not the ordinary drink it was along the Mediterranean. In fact, Posidonius says it was served in the houses of the rich, while the poor drank wheat-beer. He also describes "gifts" of wine (along with gold and silver), apparently a means by which the elite secured the allegiance of followers (Ath. 4.154c). On these grounds, Celticists have seen wine as a "prestige good," that is, one of a set of items that "[do] not move so much along commercial lines, but along those determined by social and political relations," and especially serve as "status indicators among the elite."
Wine plays a somewhat different role in Caesar's ethnography. The Suebi reject its importation entirely. So do the Nervii, fierce among the Belgae, already the nearest of the Gauls to the Germans (BG 2.15.4, 1.1.3). In both cases this is part of a broader rejection of commerce (4.2.1, 2.15.4), and in both cases the reasoning is similar:
[The Nervii] allowed no wine or other luxury goods to be brought into their territory, because they judged that by these things their spirits were relaxed and their virtue (virtutem) lessened; they, on the other hand, were fierce folk of great virtue (virtutis). (2.15.4)
[The Suebi] do not permit wine to be imported to themselves at all, because they think that by this thing people grow soft and effeminate in putting up with labor. (4.2.6)
Drinking unmixed wine is merely barbaric; it means you are doing civilization wrong. To reject wine entirely is to reject human civilization altogether. Prohibiting wine is thus emblematic of the supposed "nomad" attempt to maintain primitive (manly) virtue intact by avoiding all civilization.
That Caesar is cribbing from himself in writing these passages, that he is talking about an item (wine) with such clear symbolic value in the ancient world, and that he is doing so in a way that lends support to his much broader argument might suggest that we read these little stories about wine as purely literary devices. Similar arguments might suggest that the accounts of wine-drinking in Diodorus and Posidonius as well are simply a (different) ethnographic trope. This is certainly possible, but let me suggest that there is at least one other plausible interpretation. The details of unmixed wine, transport difficulties, and elite consumption in Diodorus and Posidonius cohere very neatly, without either account clearly being derived from the other. Furthermore, the latter two points do not seem to have nearly the motivation in terms of broad ethnographic thematics as the first. Finally, we should keep in mind that at least Posidonius had direct experience of some of these Celtic peoples. Perhaps the story we can reconstruct from the Greek accounts is essentially true. It may then either have been completely unaffected by the tradition, or perhaps it was particularly interesting to Posidonius (and others) because of its partial implication of categories that were of standing importance to Greek culture.
What, then, of Caesar's account? Let us imagine that it, too, contains at least a kernel of truth, and that at least some of these tribes, perhaps the more remote ones (where wine was the most expensive?), actually prohibited or restricted the importation of wine. Understood, with Caesar, as the rejection of the "virtues" of civilization, this account is rightly suspect; let us instead begin with the Posidonian/Diodoran understanding of wine. If it was in fact an expensive luxury good, and perhaps one particularly symbolic for being imported, then such prohibition or restriction makes sense as a sumptuary law, a type common in Roman and other societies. Consider what we know of the contemporary Celtic political context. It is widely (though not universally) held that the century or so before the Roman conquest saw a move in Gaul from political organization based largely on "tribes" or "chiefdoms" to more state-like arrangements. Among the purported causes of this shift is increased trade with the Mediterranean area (including wine). These transformations would have
opened a new channel of access to obtaining the means necessary to operate in one of the important traditional arenas for status competition: that is, the "commensal politics" of feasting. . . . An escalation of competition played out in feasting activities was likely to have occurred between those with privileged access to the traditional forms of drink . . . and those with newly advantageous access to the exotic form.
Sumptuary legislation, whether or not literally effective, delegitimizes the "luxuries" at which it is directed by appeal to morality and/or tradition. This, in the present context, would be a sensible strategy for the old Celtic aristocracy in the face of opponents who were attempting to establish novel authoritative symbols. Such a conservative strategy seems particularly plausible among the Suebi and Nervii, who seem, archeologically, far from Roman influence.
One can then imagine Caesar grasping the truth (if that is what it is), but rejecting it so as to make the bare facts serve his immediate purposes better. Or, perhaps just as likely, it might genuinely never have occurred to Caesar that impoverished "barbarians" such as these would have felt any need for sumptuary laws as he knew them. Especially if he knew nothing but the bare fact of the regulations, the explanation he offers in De Bello Gallico may well have been the one he felt was most reasonable.
My interpretation of the Greek tradition is, of course, speculative; my reading of Caesar, doubly so. I would not claim that I was myself totally persuaded by either. Nonetheless, I think the attempt makes two important methodological points. First, recent years have seen an increased awareness of the rhetoric of ancient historians. On the whole, this must be taken as a positive development, but it has had some unfortunate side effects. In particular, it is now easy to find claims that a particular passage, because it invokes an identifiable topos, has no evidentiary value, except perhaps for the attitude of the author (particularly the case for invective or for "tyrannical" stereotypes directed at "bad emperors"). The present example suggests that, even when discourse is shaped by common topoi, those topoi may signify something other than their own presence. Second, in Chapter 1 I consider the claim that Caesar's account of Gaul was distinctive because of autopsy; he had been there, so his information was allegedly better than earlier, quasi-mythical stories. That claim can be rejected, at least as a complete explanation, on several grounds. Here we see another case where the other sources may well be closer to the truth (perhaps, admittedly, with the aid of autopsy themselves). We may even have a case where Caesar has had direct experience but has still refused (or been unable) to arrive at an interpretation that has a claim to being the better one.
It has been fashionable in the last couple of decades to speak of the "textuality of history," that is, the fact that our access to history is always textually mediated. Brunt has responded to essentially this challenge by arguing that the problem is not a new one, and furthermore, that we negotiate it successfully every day:
There is no distinction in principle between the propositions that Caesar was killed by Brutus and Cassius and that Smith has just been injured by a golf ball. . . . In ascertaining what we need to know (in the loose sense of that term applicable in daily life), we draw not only on our own perceptions but on those of others. We accept what they tell us, provided that experience has not seemed to show that they are generally mendacious, at any rate in the kind of report that is our immediate concern, or faulty in their recollections, or incapable of giving accurate accounts of their own perceptions.
It is tempting simply to say that Brunt is right. It is an old problem, and the method he suggests has long been (and must be?) the response. Yet, the question of the wine shows how many things could go wrong in a fairly ordinary exchange of understanding (whether or not they have actually gone wrong in this particular case), especially if we are translating across many centuries and between multiple cultures. Even though the processes of day-to-day and historical reconstruction are not different in principle, there are huge differences in practice; in matters near to us we have vastly more background information about the habits of our informants (and their informants and their informants . . .), as well as the opportunity to cross-check information in various ways.
The situation is actually worse than this. The hypothesis that I have advanced so far claims that in some sense the Greek version is better than Caesar's. Suppose that Caesar is "right" as well. Sumptuary legislation often lends itself to the kind of moralizing, macho language in which Caesar casts the barbarians' refusal of wine. Perhaps he has represented their original claims accurately in this respect. It might be objected that he still erred out of naivete, that he (and conceivably the Nervii and Suebi themselves) did not "really" understand what was going on. We have then (at least) two incompatible representations—one "sociological," one "moralizing"—neither of which is clearly false. The distinction between the two is not so much one of truth value, but of appropriateness to different circumstances. Actually, Brunt almost admits that this is the historian's usual dilemma, though he cannot quite bring himself to do so. Note his phrases "what we need to know" and "the kind of report that is our immediate concern." These are not epistemological categories, but pragmatic ones; the criterion here is utility, not truth. I adopt here a viewpoint that (like skepticism) questions whether language (and thought more generally) can even in principle be said to "represent" reality, but that nonetheless (unlike skepticism) does not object to the existence of a real world that could have some kind of causal relationship to language (and thought). As a result, the utility of narratives is constrained (though never absolutely) by factors other than coherence with other narratives. Thus Brunt's dismissal of "Pyrrhonists" on pragmatic grounds is valid, but not salient to the point he claims to be making. The position that his arguments seem to me actually to support (and that I myself favor) is more like Rorty's neopragmatism or Lakoff's "experiential realism." This leaves little left of the distinction between the history "of representations" and the history of anything else.
The point here is not, of course, to condemn any particular traditional historical practice (and certainly not to do so categorically); historians are good at telling stories that meet their shared needs, and there is no generalized reason to stop doing so now. In fact, historians have also been fairly good about allowing new types of story when new needs demand them—the rise of family history is an example. Rather, the point is to keep in mind that just as we ourselves work on shifting sands, so do our sources. But in the latter case, we rarely see the motion itself, only traces of its results. Hence, a more or less deductive mode of historiography cannot be uniquely valuable. Room should be made for other modes in addition, such as model-building or quasi-fictionalization. Responsible theorizing, then, calls for more histories, not fewer.