SOURCE: Wiedemann, T. E. J. “Between Men and Beasts: Barbarians in Ammianus Marcellinus.” In Past Perspectives: Studies in Greek and Roman Historical Writing, edited by I. S. Moxon, J. D. Smart, and A. J. Woodman, pp. 189-201. Cambridge, Eng.: Cambridge University Press, 1986.

[In the following essay, Wiedemann explores Ammianus's use of animal metaphors in describing individuals and groups of people.]

Dietary practices are among the more obvious ways in which one group of people can differentiate itself from another. What I eat and drink is normal and natural. A person who does not eat or drink what I do is peculiar: in structuralist jargon, I am central and he is marginal. He may be marginal geographically—simply foreign—or morally: a saint/hero (between man and god) or a sinner/heretic/revolutionary (between man and beast).¹ The ultimate dietary rule is the ban on eating the flesh of another human being. A recent study by Arens of accounts of cannibalism predictably aroused considerable discussion.² Its thesis might be summarized as follows. To eat human flesh is the mark of an animal. A human who eats human flesh thus shares in his person the characteristics of a man and of an animal: cannibalism symbolizes the mid-point between humanity and bestiality. It follows that any group satisfied that its own behaviour patterns are normal will tend to ascribe the qualities of cannibals to individuals or groups whom they consider to be hostile to their behaviour patterns or values. Thus—according to Arens—the statement, ‘x is a cannibal’ should not be taken at its face value as a piece of descriptive ethnography; it is rather a moral evaluation, labelling x as being marginal with respect to the speaker's perception of his own values or social position. Arens presses this thesis as far as it will go, and—excepting some cases of anthropophagy in dire emergencies³—denies that cannibalism has ever existed as an established social ritual in any human society; he would put reports of cannibalism among sixteenth-century Amerindians, eighteenth-century Polynesians, nineteenth-century Africans, or twentieth-century Papuans, on the same level as the atrocity stories found directed against the early Christians, medieval Jews, or the German army which occupied Belgium in 1914.⁴

We do not have to follow Arens all the way; his views are interesting enough if it is the case that, in some instances, ascriptions of cannibalism are not descriptive, but evaluative. Ethnographical passages in classical literature contain a series of stereotypes, associated with barbarians in general, to which the Arens thesis may easily be applied. One could take as a classic example Strabo's well-known description of Ireland (4.5.4):

Besides some small islands round about Britain, there is also a large island, Ierne, which stretches parallel to Britain on the north, its breadth being greater than its length. Concerning this island I have nothing certain to tell, except that its inhabitants are more savage than the Britons, since they are man-eaters as well as grass-eaters [or ‘great eaters’, depending on the reading], and since, further, they count it an honourable thing, when their fathers die, to devour them, and publicly to have intercourse, not only with the other women, but also with their mothers and sisters; but I am saying this only with the understanding that I have no trustworthy witnesses for it; and yet, as for the matter of man-eating, that is said to be a custom of the Scythians also, and in cases of necessity forced by sieges the Celts, the Iberians, and several other peoples are said to have practised it.

Strabo's insistence that he is merely repeating what others say (λεγόνται) is crucial: stories ascribing irregular or peculiar dietary or sexual habits should not, or at least need not always, be believed. The Irish behave marginally in these respects because they are, quite literally, at the edge of the human world. Indeed, they are in a sense already over the edge of the world, since they live on an island in the Ocean. Islands are distinctly marginal: they are, so to speak, both of the sea, and of the land. The seashore is another such locus of marginality, both sea and land. Caves, too, are marginal places, being both under the earth, and yet accessible to those who walk on its surface. It is only to be expected, then, that Sophocles' Philoctetes, on the margin of human society, inhabits an island-cave by the shore.⁵

The limits of the human world separate us both from beasts and from the divine. Islands, caves, and seashores are thus heavily populated by superhuman as well as sub-human beings. One could cite as examples the birthplace of Zeus, the cave of the Eumenides, the Homeric island-caves of the Cyclops and of Calypso, the cave in which Euripides wrote his plays, the caves in which the Druids are said to have done their teaching,⁶ or the unscriptural cave under the Church of the Nativity at Bethlehem; and perhaps also the cavemen postulated by nineteenth-century archaeologists who believed in Darwin and were desperate to find a missing link between men and beasts. One difference between the classical and the modern picture of the savage caveman is that, since the Enlightenment, Europeans have tended to place the savage in an evolutionary sequence; previously, the savage's marginality had been one of distance, rather than time.⁷ John Locke's celebrated phrase, ‘In the beginning, all the World was America’,⁸ combines the Enlightenment view that savagery is a social period prior to modern civilization, with an earlier idea that it is something to be found at the geographical margins of the known world.

These margins need not of course point downwards; the edges of the world may also be places where a much better society can be found. Thus utopias in the classical period may be found not just at the time when the world was newly created—chronologically marginal—but also at the edge of the world, typically on islands in the Ocean like Panchaea in the Far East or the Fortunate Islands or Atlantis in the Far West; the Hyperboreans, too, live on an island.

Utopians and savages, both of them marginal to normal societies, thus share certain characteristics like living on islands. They may share social characteristics, too. Freedom from social normality may consist in sexual promiscuity, or the absence of private property. The inversion of normal rules of behaviour is a characteristic of such people. One is reminded of anthropologists' findings about the Temne in Sierra Leone, who believe that white people must normally walk upside down. This is the context of stories about the Thracians mourning a person's birth, not his death, and of a whole series of topoi about Egypt. Because Egypt is unlike Greece, it must be the opposite of Greece, and therefore—since Herodotus—they do all sorts of things the wrong way round, like writing: ‘suis litteris perverse utuntur’.⁹

One might set up an ideal type of the marginal group in terms of classical literary commonplaces. Such people, living on an island (possibly in the Ocean), would wear no clothes (Britons and Agathyrsi, like Red Indians, are noted for painting their bodies; and the Balearic Islands are ‘Gumnēsiai’ in Greek because of the nudity of the natives).¹⁰ They would not live in houses, but in caves,¹¹ wagons (Scythians), or tents (Saracens). They would be nomads rather than sedentary farmers. If not cannibals, they would at least eat curious food and drink something other than the regular Graeco-Roman wine mixed with water.¹² They might practise incest or polygamy, or make love in public;¹³ if not, then at least they might reverse the proper relationship between the sexes, and be ruled by women, like the Picts. Silius Italicus, in his list of Carthaginian allies,¹⁴ stresses that in Galicia in Spain the roles of men and women in agriculture are reversed. Otherwise he ignores non-military characteristics, except when he comes to describe the Gaetuli:

Nulla domus: plaustris habitant; migrare per arva
mos, atque errantes circumvectare penates.

Such people are not merely savage fighters,¹⁵ but fight in a peculiar way: women may share in the fighting; or the men may use bows,¹⁶ a peculiar weapon in the Graeco-Roman context, as we are reminded by the story of Philoctetes. The ultimate symbol of the barbarian's intermediate position between Greeks and beasts is when he actually turns into that quintessentially marginal animal, the wolf: the Scythian Greeks insisted to Herodotus that every single Neurian practised lycanthropy.¹⁷

If these topoi are seen as illustrations of geographical marginality, rather than objective descriptions, it becomes less surprising that we should find the same peoples variously held up as paradigms of morality, and of bestiality. The Scythians may appear as fierce, cannibalistic and primitive, but equally as sharing some of the features of an ecological paradise: ‘Cultores iustissimi, et diutius quam ulli mortalium et beatius vivunt. Quippe festo semper otio laeti non bella novere, non iurgia.’¹⁸ This positive picture of the Scythians appears elsewhere in classical literature: Horace refers to the

          Campestres melius Scythae
Quorum plaustra vagas rite trahunt domos […].(19)

The Ichthyophagi in Ethiopia are another group of savages noted for their justice.²⁰

We would expect to find these same elements in the ethnographical digressions of a classicizing historian like Ammianus Marcellinus. It may be worth noting that Ammianus' digressions are extremely formal. Almost without exception, digressions are divided off from the previous narrative by an explicit introduction, and come to an end with a formal close.²¹ Ammianus was all too conscious that he was writing a work of literature: we need only recall the many references to and quotations from Cicero, Virgil, and Homer. I suspect that this is not so much a case of a Greek-speaker who is self-conscious about writing in a foreign language, or keen to parade his literary studiousness, as that Ammianus is a soldier and official who wants to be obedient to the clear, straightforward rules of historical writing. Which is why the formal digressions contain exactly that ethnographical material which we would expect. The Saracens²² are ‘seminudi, coloratis sagulis pube tenus amicti’; they do no agricultural work, but instead ‘errant sine lare, sine sedibus fixis aut legibus’. Their sex-life is peculiar: ‘uxores mercennariae, conductae ad tempus ex pacto […] et incredibile est, quo ardore apud eos in venerem uterque solvitur sexus.’ So is their food: ‘victus […] caro ferina est, lactisque abundans copia […] et plerosque nos vidimus frumenti usum et vini penitus ignorantes.’²³ ‘Hactenus’, concludes Ammianus, ‘de natione perniciosa.’ The digressions about the nations living around the Black Sea, the eastern provinces of the Persian empire, ancient Thrace, and about the Huns²⁴ contain more of the same. Thus the Scythians ‘palantes per solitudines vastas, nec stivam aliquando nec sementem expertas […] ferarum taetro ritu vescuntur’.²⁵ Again, we find some uncertainty whether these people are savages who indulge in human sacrifice²⁶ or ‘iusti homines placiditateque cogniti’.²⁷ In Thrace, the Scordisci used to be ‘saevi et truces’; they sacrificed their war-captives and drank their blood out of human skulls.²⁸ The Odrysae went so far as to kill their own people when there was a shortage of enemies to slaughter—‘vagantes sine cultu vel legibus’.²⁹ Yet Thrace had been part of the Graeco-Roman world for seven centuries when Ammianus wrote this. The Huns³⁰ do not cook their food, avoid houses and huts ‘sed vagi montes peragrantes et silvas’; they dress in clothes made from the skins of field-mice, and wear the same clothes indoors and out until they turn to rags and fall off their backs.³¹ And day and night they sit on their horses: ‘Inconsultorum animalium ritu, quid honestum inhonestumve sit penitus ignorantes, flexiloqui et obscuri, nullius religionis vel superstitionis reverentia aliquando districti […].’³²

These collections of commonplaces going back to Herodotus are exactly what we would expect in literary digressions. Much more interesting is their absence from the digression on Gaul and its people.³³ The Gauls do have a peculiar drink, beer, which Ammianus claims to have tasted himself, and there is a hint of sexual role-reversal in the statement that a Gaul's wife is ‘multo fortiore et glauca’, with a punch like that of a Roman catapult (Ammianus omits to say whether he had had personal experience of that too). It looks as though Ammianus was quite capable of avoiding topoi if he wished; and it might therefore be worth examining how stereotyped his treatment of non-Romans is in the rest of his narrative.

One is immediately struck by the frequency of adjectives like ‘efferati’.³⁴ Isaurian bandits are ‘flagrans vesania’;³⁵ they become bolder as a consequence of madness—‘rabie saeviore amplificatis viribus’.³⁶ The German tribes who invaded Gaul are ‘superbae gentes’;³⁷ theirs is a ‘barbarica rabies’.³⁸ The Alamanni may act with even more savagery than usual, ‘saevientes ultra solitum’,³⁹ and we are told about their horrible war-cries—‘ululantes […] lugubre’.⁴⁰ At the battle of Strasburg, Julian in his speech to the Romans talks of their ‘rabies et immodicus furor’; Ammianus himself uses the terms ‘barbara feritate’, ‘frendentes immania, ultra solitum saevientium comae fluentes horrebant, et elucebat quidam ex oculis furor’; ‘violentia iraque incompositi […] in modum exarsere flammarum’; they attack ‘velut quodam furoris afflatu’.⁴¹ And everywhere else too, the qualities repeatedly ascribed to the Germans are ‘furor, amentia, feritas’.⁴² The same qualities appear in connexion with the barbarians who cause havoc in Britain,⁴³ as well as the Sarmatians.⁴⁴ The Moors too suffer from ‘barbarica rabies’.⁴⁵ The Saxons may be merely ‘superbi’;⁴⁶ the Goths, on their first appearance, are ‘saepe fallaces et perfidos’.⁴⁷ At one point—after the death of Julian—we have a regular list of what Ammianus considers to be ‘gentes saevissimae’ endangering the empire: the Alamanni, Sarmatians, Picts, Saxons, Moors, Goths and Persians.⁴⁸

But none of this is sufficiently specific to deserve the status of ethnography. All Ammianus is saying is that the enemies of Rome are undomesticated (‘ferae’), violent, insolent, or mad when, and in so far as, they attack the Roman empire; we should not interpret these epithets as evidence that he thought he was describing a peculiarity of the tribe in question. In the case of the Persians, there may be a slight tendency for Ammianus to associate them with a particular characteristic, viz. duplicity; that of course is a traditional element in Roman attitudes to Persia, a state which could not so easily be visualized as hostile to order and good government per se. Interestingly, the Persians alone of non-Romans are themselves said to be engaged in fighting ‘ferocissimas gentes’, or ‘ferarum gentium’.⁴⁹ But Ammianus does not use this theme often. More usually, it is the Persian king, Sapor, individually who is described as enraged (‘effrenata regis cupiditate’), as ‘truculentus rex ille Persarum, ardore obtinendae Mesopotamiae flagrans’, or as ‘turgidus’. At one point, he is ‘ultra hominem efferatus’. His ‘ira’, ‘efferata vesania’, and ‘rabies’ are mentioned⁵⁰—but then, so are those of some Roman emperors. The Persian king's ‘perfidia’ is mentioned only once, when he attacks Armenia ‘per artes fallendo diversas’ in a.d. 368.⁵¹ As for the Persians as a nation, they are ‘fallacissima gens’ and called ‘gentem asperrimam per sexaginta ferme annos inussisse Orienti caedum et direptionum monumenta saevissima’. Julian refers to their treachery in a set speech: ‘nihil enim praeter dolos et insidias hostium vereor, nimium callidorum.’⁵²

What is noteworthy about this is not that Ammianus should describe Rome's enemies in these terms—he himself, after all, had personally been involved in the business of fighting most of them—but how infrequently he applies these critical and negative concepts, and how unsystematically. Furor or ira are qualities of people Ammianus does not like, whether they are Romans or not: they are not national characteristics. This is the common-sense approach of a soldier whose job it is to hurt the enemy without getting hurt himself; and it is striking how many examples of fallacia or saevitia on the Roman side Ammianus describes, without moral comment—occasionally even with approval. There are at least a dozen such instances: the Romans massacre some Alamanni ‘promiscue virile et muliebre secus, sine aetatis ullo discrimine’; Julian's duplicity in attacking the Franks after having pretended to come to an agreement with them is praised; twice peaceful Alamanni are massacred ‘sine ulla parsimonia’ (in both cases by Roman soldiers under the influence of ‘ira’ or ‘incitante fervore certaminum’). During Julian's campaign against Persia, Ammianus describes without comment how the Romans butcher women and children: ‘qua [sc. civitate] incensa, caesisque mulieribus paucis quae repertae sunt’; at Maiozamalcha, the massacre is ‘sine sexus discrimine vel aetatis’. On the other hand, a non-Roman chieftain who acts thus is a brigand: ‘Malechus Podosax, famosi nominis latro, omni saevitia per nostras limites grassatus.’ The Romans may be praised for using ‘dolos occultiores’ against the Saxons; Ammianus approves of the massacre of a group of Syrian brigands, the Maratocupreni, and shows no concern at the killing of all their young children. Count Theodosius tortures and burns alive Moorish rebels. Valentinian executes a group of Quadi ‘iugulata aetate promiscua’.⁵³ What these examples show is that Ammianus does not intend his readers to understand his ascription of savagery and duplicity to the enemies of Rome as in any sense ethnographic. He is well aware that Roman armies behave no differently.

Instead of looking at straightforward moral characteristics for evidence of how Ammianus saw particular ethnic groups, it might be worth examining one particular group of literary similes, namely comparison with animals, to see how it is applied to barbarians. After all, if what makes barbarians ‘efferatae gentes’ is the fact that they are at the margins of the human world, not just geographically but also so to speak zoologically, by sharing certain social characteristics with sub-human animals, it is not surprising that, in more or less literary passages, Ammianus should use a series of animal metaphors to depict the behaviour of such groups. To describe Isaurian bandits, he quotes from Cicero's pro Cluentio:⁵⁴ ‘Atque—ut Tullius ait—ut etiam bestiae, fame monitae, plerumque ad eum locum ubi aliquando pastae sunt revertuntur […].’ The transrhenane Germans continued to attack Gaul even when Julian, as Caesar, had shown that he was prepared to fight them:⁵⁵ ‘Utque bestiae custodum neglegentia raptu vivere solitae, ne his quidem remotis, appositisque fortioribus abscesserunt, sed tumescentes inedia, sine respectu salutis, armenta vel greges incursant.’ In the digression about the lands to the north of the Black Sea he says that the Scythians—or at least most of them—know no agriculture, but ‘ferarum taetro ritu vescuntur’.⁵⁶ And the Goths, on the day after they had destroyed Valens' army at Adrianople, attack the city itself ‘ut bestiae sanguinis irritamento atrocius efferatae’.⁵⁷ That Isaurian bandits, Germans, Scythians, and Goths should be described as wild beasts fits into the rhetorical pattern. What is surprising is that Ammianus should use this metaphor for barbarians so rarely. When he describes the Gothic attack on the Romans in a.d. 378, and wants to use epic literary metaphors to stress the scale and importance of that invasion, he likens the Goths to burning fire-darts (‘ut incensi malleoli’) or to a torrent (‘ut amnis immani pulsu undarum obicibus ruptis emissus […]’).⁵⁸ The only application of the beast-metaphor to the Persians appears to be at 24.8.1, where Julian—now aware that his army is in grave danger of being cut off and destroyed—parades a number of starving Persian prisoners in order to persuade the Romans that they are not formidable as foes, and calls them ‘deformes illuvie capellas et taetras’ (ugly she-goats disfigured with filth). This should not lead us to think that Ammianus considered Persians to be animals; it is Julian who says this, and it is rhetoric. Indeed, there is a parallel passage where the Romans are, by implication, compared to wild beasts. It is in the Persian king Sapor's letter to Constantius, advising Roman withdrawal from the disputed provinces of Mesopotamia, and inviting him to reflect that ‘hocque bestias factitare: quae cum advertant cur maxime opere capiantur, illud propria sponte amittunt, ut vivere deinde possint impavidae’.⁵⁹

There are two other groups of persons to whom Ammianus applies the beast metaphor. He expresses surprise that Eutherius, although a eunuch, was a good man (eunuchs being, of course, the marginal group par excellence): ‘sed inter vepres rosae nascuntur, et inter feras non nullae mitescunt.’⁶⁰ And there is the notorious reference to the violent enmity among Christians: ‘nullas infestas hominibus bestias, ut sunt sibi ferales plerique Christianorum’.⁶¹

Much more frequently, however, Ammianus uses animal metaphors to illustrate specific behaviour, or the behaviour of specific individuals, rather than to typify groups. People in an extreme state of degradation or passivity are said to be, or to be treated like, animals: the supporters of Magnentius, when imprisoned by Constantius, are dragged about like beasts (‘in modum beluae trahebatur’). During Julian's campaigns against the Germans, innocent women and children are massacred ‘ut pecudes’. When some Roman soldiers lay an unsuccessful ambush for the Armenian king Papas, they are said to expect him to behave like a hunted beast running into a trap: ‘quasi venaticiam praedam’.⁶²

Usually it is some aspect of ferocity that Ammianus wants to stress. The parallel is frequently taken from the Roman arena.⁶³ Ammianus himself, sent to Cologne in the entourage of Ursicinus to deal with Silvanus' revolt, feels himself ‘ut bestiarii, obiceremur intractabilibus feris’. At the siege of Amida in a.d. 359 Ammianus noted that some Gallic soldiers threatened their tribunes for not allowing them to sally forth against the Persians: ‘utque dentatae in caveis bestiae, taetro †paedore acerbius efferatae, evadendi spe repagulis versabilibus illiduntur […].’ Presumably Ammianus himself was one of these tribunes, and the wild-beast metaphor represents his feelings at the time. Before the battle of Adrianople, the Goths are described as ransacking Thrace ‘velut diffractis caveis bestiae’. And in two places an individual is compared to the animals in the amphitheatre: Maximinus, as prefect of Rome, ‘effudit genuinam ferociam, pectori crudo affixam, ut saepe faciunt amphitheatrales ferae, diffractis tandem solutae posticis’. And there is the emperor Valens, anxious that all who might be involved in Theodorus' treason be punished: ‘totus enim devius ab aequitate dilapsus, iamque eruditior ad laedendum, in modum harenariae ferae, si admotus quisquam fabricae diffugisset, ad ultimam rabiem saeviebat.’⁶⁴ Some other individuals appear as unspecified beasts. These comparisons are usually decidedly negative; an apparent exception is that of the circumstances of Procopius, hiding out of fear of Jovian, to the life-style of a wild beast, ‘ferinae vitae’; he lived ‘in locis squalentibus, stringebatur, hominumque egebat colloquiis’. Later it becomes apparent that, far from evoking our sympathy, the animal metaphor shows how Ammianus dislikes Procopius: at c. 10 he likens Procopius in hiding (‘latenter’) to a beast of prey in its lair: ‘subsidebat ut praedatrix bestia: viso, quod capi potuerit, protinus eruptura’. Elsewhere the criticism is obvious. Rusticus Julianus, a potential candidate for the imperial office in a.d. 367, is described as ‘quasi afflatu quodam furoris [sc. smitten by] bestiarum more humani sanguinis avidus’. Valentinus, an opponent of Count Theodosius, is irked at being exiled to Britain, ‘quietis impatiens ut malefica bestia’.⁶⁵

So much for beasts in general. Of particular beasts, there are seven comparisons of people to vipers. These include groups of people: the Isaurians again, five years after the police actions against them in a.d. 354 (recorded at 14.2.2), are described as ‘paulatim reviviscentes, ut solent verno tempore foveis exsilire serpentes […]’. The Roman mob, demonstrating against the urban prefect Leontius in a.d. 355, is likened by Ammianus to serpents: ‘Insidens itaque vehiculo, cum speciosa fiducia contuebatur acribus oculis tumultuantium undique cuneorum, veluti serpentium, vultus […].’ Among individuals, there is the notary Paulus, responsible for various administrative disorders, ‘ortus in Hispania coluber quidam’ (according to Bentley's restoration of the text). The Caesar, Gallus, aware that Constantius plans to remove him, behaves ‘ut serpens appetitus telo vel saxo’; this refers to a fit of anger leading to the murder of various officials loyal to Constantius. During the treason trials which followed Gallus' downfall, Arbitio's attack on Ammianus' commander Ursicinus is described as follows: ‘ut enim subterraneus serpens, foramen subsidens occultum, adsultu subito singulas transitores observans incessit: ita ille odio alienae sortis […]’. After describing the judicial crimes of Maximinus, vice-prefect of Rome, Ammianus says that, when appointed praetorian prefect, he continued to be dangerous at a distance, ‘ut basilisci serpentes’.⁶⁶ It may be worth noting in this context that Ammianus did have alternatives to animal- or snake-metaphors which he chose to use when it came to describing brutal administrators: Simplicius of Emona is compared, not to a beast, but to the literary tyrant-king Busiris.⁶⁷ In a.d. 373, Valens sent two officials to re-arrest the Armenian king Papas, who had escaped from imprisonment by the Romans at Tarsus. The two officials Danielus and Barzimerus failed in the attempt—and ‘probrosis lacerati conviciis, ac si inertes et desidentes, ut hebetatae primo appetitu venenatae serpentes, ora exacuere letalia, cum primum potuissent, lapso, pro virium copia nocituri’. In fact they accused Papas of practising the magic arts, thus persuading Valens to have him assassinated by the general Trajanus, an act which Ammianus deplores as totally contrary to the rules of war.⁶⁸

There are some other straightforwardly negative animal similes. Picking up a Ciceronian comparison of two associates of Verres to hunting-dogs, he says that the informers who took advantage of the overthrow of Gallus are ‘honorum vertices ipsos ferinis morsibus appetentes […] non ut Cibyratae illi Verrini, tribunal unius legati lambentes, sed rei publicae membra totius per incidentia mala vexantes’. He goes on to describe their leader Mercurius ‘ut clam mordax canis interna saevitia, summissius agitans caudam’. The comparison with hunting-dogs appears in another much later reference to informers, in Ammianus' celebrated digression on lawyers and lawsuits; it is possible, he says, to see powerful and rapacious men ‘ut Spartanos canes aut Cretas, vestigia sagacius colligendo, ad ipsa cubilia pervenire causarum’.⁶⁹ There are two comparisons to birds of prey. One refers to the Saracens, ‘milvorum rapacium similes’, the other to a North African tribe which attacked the cities of Tripolis, Lepcis and Oea, the Austoriani: ‘Austoriani successu gemino insolentes, ut rapaces alites, advolarunt, irritamento sanguinis atrocius efferatae.’ But if a tribe of barbarians can be likened to vultures, an equally unflattering comparison can be applied to the senators of Rome in a famous passage of satire: ‘ex his quidam cum salutari pectoribus oppositis coeperint, osculanda capita, in modum taurorum minacium, obliquantes, adulatoribus offerunt genua savianda vel manus […]’. Indeed there appears to be only one instance of an animal comparison which is at all positive, and that is of Ammianus' general Ursicinus to a lion; but in fact Ursicinus is compared to a disabled lion, to explain why he was unable to assist the beleaguered Amida: ‘ut leo magnitudine corporis et torvitate terribilis, inclusos intra retia catulos periculo ereptum ire non audens, unguibus ademptis et dentibus’. Of course, Ammianus is here exploiting the association between lions and kingliness, which goes back to Aesop and beyond, in order to flatter his hero. Ammianus is well aware that lions are traditionally royal, and positive: this is clear from the account of the omen of the dead lion found by Julian shortly before his own death. There are of course old and powerless lions in Aesopian fable. But it is interesting that Ammianus should choose to use the figure of a lion to illustrate Ursicinus' weakness, not his strength. And there is another negative comparison of someone to a lion. Gallus indulged in a string of judicial executions; ‘post quorum necem, nihilo lenius ferociens, Gallus, ut leo cadaveribus pastus, multa huius modi scrutabatur.’ So in the last analysis all animals illustrate negative qualities for Ammianus—ferocious like the beasts in the arena, keen on human blood, insidious like snakes, or insufferably proud like bulls.⁷⁰

On the other hand, Ammianus does not particularly associate animal qualities with marginal groups. There are occasions where he does that, but he also has a wide range of alternative literary metaphors to choose from, and is as happy (for instance) to liken the Goths in Book 31—where the description of the battle of Adrianople allows him to wallow in epic similes—to rivers, firebrands, collapsing walls, warships, or forest fires. Animal metaphors are applied to any person or group he disapproves of as falling short of the standards of civilized human behaviour, whether emperors or usurpers, officials or informers, Roman Christians, the Roman mob, or Roman senators. The last of these at least can scarcely be labelled marginal. Bestiality is not a special mark of non-Roman barbarians.

I would conclude not only that we should be wary of believing that the material contained in Ammianus' ethnographic digressions is descriptive, but also that Ammianus himself did not intend these statements about barbarian tribes as descriptive ethnography. Some barbarians, like some Romans, threatened imperial order and stability; Ammianus, as a soldier and official, described such persons in negative terms. But as a historian, he felt that the rules of his genre of literary writing required regular digressions, some of them ethnographic; and he knew that his audience required such digressions to contain the centuries-old stereotypes about marginal groups. A historian writing in the classical tradition could not disappoint his audience. Perhaps his audience was rather disappointed since there is only one reference to cannibalism in the surviving books.⁷¹

Notes

1. Food taboos: Douglas (1966), Bremmer (1980) 33 n. 9.

2. Arens (1979).

3. Some ancient instances: Hdt. 3.25.6, Thuc. 2.70.1, Plb. 1.85.1, Diod. 33/34.2.20, Caesar BG 7.77.12, Jos. BJ 6.3.4; but not Ammianus 25.8.15.

4. Cf. Sall. BC 22 (conspirators drink human blood).

5. Buxton (1982), Avery (1965).

6. Pomponius Mela 3.20.

7. Meek (1976).

8. John Locke Two Treatises of Government 2.49 in edition by Laslett (1967) 343.

9. Pomponius Mela 1.75. My informant on the beliefs of the Temne is Dr I. Hamnett (personal communication).

10. Diod. 5.17.

11. Like Diodorus' Balearic islanders. For Trogodytae [sic] cf. Pliny NH 6.189f., Aelian 6.10.1, 9.55, Virgil G.3.376f.

12. Food and drink: Bremmer (1980) lists references to Scythians and Indians as drinkers of either unmixed wine, or water. Cf. also Str. 16.4.17 on the Trōglodutai; Hdt. 4.104 remarks that the nomadic Scythian Boudinoi are ϕθειροτραγέουsι (eaters of pine-cones, rather than lice, perhaps?). Further fruit-eaters occur near the Caspian Sea (Hdt. 4.203).

13. Scythians: Hdt. 4.104 (where wife-sharing is said to eliminate jealousy); Trōglodutai: Str. 16.4.17. Unashamedly public sex: Hdt. 1.203 (Caspian Sea), Xen. Anab. 5.4.33 and Ap. Rhod. 2.1023ff. (the Mossynoeci). The comment of How and Wells (1912) 1.153 on this ‘lowest stage of degradation’ is illuminating: ‘But among modern savages it is, to say the least, very rare.’

14. Punica 3.349ff.

15. On savagery: Scyths: Hdt. 4.62 (human sacrifice), 64-66 (scalping, blood-drinking). Violence is of course appropriate to barbarians, as persuasion is to adult male Greeks: Isoc. Philip 16; Buxton (1982), esp. on Aeschylus' Suppliants.

16. Cf. Sophocles Philoctetes; Scythians, Amazons, Macrobian Aethiopians (Hdt. 3.21 etc.).

17. Hdt. 4.105.2.

18. Pomponius Mela 3.5.

19. Hor Odes 3.24.9f.

20. Hdt. 3.20f.

21. The exception is where the end of a digression corresponds to the end of a book, as with the excursus on the Persian provinces at 23.6: cf. Wiedemann (1979) 13f. On Ammianus' digressions generally, Rosen (1982) 79-86.

22. 14.4.

23. Note the ‘nos’: the claim to autopsy appears in other such accounts, e.g. Cassius Dio 29.36.4, on the Pannonians. But whether in fact the writer had seen what he describes is irrelevant; the point is that he reports what his audience expects him to report.

24. 22.8, 23.6, 27.4, 31.2.

25. 22.8.42.

26. 22.8.34: skulls fixed to the temple of Orsiloche/Diana.

27. The Aremphaei, 22.8.38.

28. 27.4.4.

29. 27.4.9f.

30. 31.2.

31. Cf. Hippocrates Airs, Waters, Places 19.

32. 31.2.11.

33. 15.12.

34. In the digression on ancient Thrace: 27.4.9.

35. 14.2.15.

36. 14.2.14.

37. 15.8.7.

38. 16.5.16f.

39. 16.11.3.

40. 16.11.8.

41. 16.12.31, 16.12.2, 16.12.36, 16.12.44, 16.12.46.

42. E.g. in the necrology of Julian, ‘regna furentium Germanorum exscindens’ (25.4.10).

43. ‘gentium ferarum excursus’: 20.1.1.

44. ‘latrocinandi peritissimum genus’: 16.10.20.

45. 27.9.1.

46. 28.5.3.

47. 22.7.8.

48. 26.4.5.

49. 14.3.1, 18.4.1.

50. 17.5.15, 20.6.1, 27.12.11, 20.7.3-11.

51. 27.12.2-4.

52. 21.13.4, 22.12.1, 23.5.21.

53. 16.11.9, 17.8.4, 17.10.6, 17.13.13, 24.2.3, 24.4.25, 24.2.4, 27.8.9, 28.2.14, 29.5.49f., 30.5.14.

54. 14.2.2; Cic. pro Cluentio 25.67.

55. 16.5.17.

56. 22.8.42.

57. 31.15.2.

58. 31.7.7, 31.8.5.

59. 17.5.7.

60. 16.7.4; Patterson (1982).

61. 22.5.4.

62. 14.5.3, 16.11.9, 30.1.15.

63. It is interesting that Ammianus chooses parallels taken from his own experience as an official from an urban, curial background, rather than metaphors involving agricultural beasts as was usual in classical literature.

64. 15.5.23, 19.6.4, 31.8.9, 28.1.10, 29.1.27.

65. 26.6.4, 26.6.10, 27.6.1, 28.3.4.

66. Vipers: 19.13.1, 15.7.4, 14.5.6, 14.7.13, 15.2.4, 28.1.41.

67. 28.1.46.

68. 30.1.16.

69. 15.3.3, cf. Cic. II Verrines 4.21.47 and 4.13.30; 15.3.5, 30.4.8.

70. 14.4.1, 28.6.13, 28.4.10, 19.3.3, 23.5.8, 14.9.9.

71. 31.2.15.

Abbreviations and Bibliography

Note. List A is the key to the abbreviations of standard works adopted in the book. List B, which is a cumulation of all works cited by author and date in the text and notes, provides the essential bibliographical information. Only abbreviations of periodical titles attested from L'Année philologique have been employed.

A. Abbreviations

ANRW: Aufstieg und Niedergang der römischen Welt, ed. Temporini, H.-Haase, W. Part I. 1-4 (1972-73), Part II. 1- (1974-84). Berlin/New York

FGrHist: Jacoby, F. Die Fragmente der griechischen Historiker, 3 parts in 15 vols. Berlin/Leyden, 1923-58

HCT: Gomme, A. W.-Andrewes, A.-Dover, K. J. A Historical Commentary on Thucydides, 5 vols. Oxford, 1945-81

ILLRP: Degrassi, A. Inscriptiones Latinae Liberae Rei Publicae, vol. 1 (2nd edition, 1965), vol. 2 (1963). Florence

LSJ: Liddell, H. G.-Scott, R.-Stuart Jones, H.-McKenzie, R. A Greek-English Lexicon, 9th edition. Oxford, 1940

ML: Meiggs, R.-Lewis, D. A Selection of Greek Historical Inscriptions to the End of the Fifth Century b.c. Oxford, 1969

PIR²: Prosopographia Imperii Romani saec. I. II. III, Parts 1-3 (2nd edition, 1933), ed. E. Groag and A. Stein; Part 4 (2nd edition, 1952-66), ed. A. Stein and L. Petersen. Berlin/Leipzig

RE: Paulys Real-Encyclopädie der classischen Altertumswissenschaft, ed. G. Wissowa et al. Stuttgart, 1893-

RRC: Crawford, M. H. Roman Republican Coinage. Cambridge, 1974

B. Bibliography

Arens, W. (1979). The Man-Eating Myth. Chicago

Avery, H. C. (1965). ‘Heracles, Philoctetes, Neoptolemus’, Hermes 93.279-97

Bremmer, J. (1980). ‘Marginalia Manichaica’, ZPE 39.29-34

Buxton, R. G. A. (1982). Peitho. Cambridge

Douglas, M. (1966). Purity and Danger: an analysis of concepts of pollution and taboo. London

How, W. W.-Wells, J. (1912). A Commentary on Herodotus, vol. 1. Oxford

Laslett, P. (1967). John Locke: Two Treatises of Government. A critical edition with an introduction and apparatus criticus, 2nd edition. Cambridge

Meek, R. (1976). Social Science and the Ignoble Savage. Cambridge

Patterson, O. (1982). Slavery and Social Death: a comparative study. Cambridge, Mass./London

Rosen, K. (1982). Ammianus Marcellinus. Erträge der Forschung 183. Darmstadt

Wiedemann, T. (1979). ‘Nunc ad inceptum redeo’, LCM 4.13-16