The elephants who appealed to the gods: animal agency in the Roman arena and the human perception of it ARTICLES

According to our Roman sources, especially the account of Pliny the Elder, an unusual feature of the appearance of elephants in a staged hunt in Pompey’s games of 55BCE was sympathy for them shown by the crowd. With a particular emphasis on this incident and elephants, this article discusses the agency of nonhuman animals who were required to do tricks, fight with one another and with humans, and act as executioners in the Roman arena, and the perception of it by the crowd and Roman authors. The main purpose of the shows was to satisfy the potentially disruptive urban masses and more fundamentally, to emphasize the power of Rome to control forces beyond its administrative author-ity and external to its cultural mores. Animals were not able to understand the forces that brought them to the arenas or resist in any organised manner, but as sentient beings they did act in ways the trainers could not predict or control. In so doing, in all cases but refusal to attack they contributed to the excitement of events, and in the case of the elephants of the 55BCE games, even caused the normally hostile spectators to empathize with their plight. When Roman spectators or writers attributed human-like traits to animals who did extraordinary things they tacitly acknowledged animal agency, but this was not transformed into any general acceptance that animals might have any moral sense or cognitive abilities comparable in any way to humans.


Introduction
In 55BCE the Roman statesman and consul Pompey held several days of games in Rome at the dedication of the temple of Venus Victrix. On the last day some 20 elephants were pitted against javelinmen in the arena. The event was described by Pliny the Elder (23-79CE) and appears in almost every modern discussion of the Roman games because it is the only known case in which a Roman crowd openly sympathised with the plight of the animals. As an extension of the discussion of the 55BCE events I will consider how the Roman organisers, audience and writers perceived animals and their agency, if they recognized it at all, with an emphasis on the role of elephants but also reference to other nonhuman animals with whom the spectators may have identified.
From 186BCE to the fifth century CE, in Italy even after the last western Roman emperor was deposed and perhaps later in the eastern Empire, wealthy and powerful Roman statesmen tried to imitate or outdo their contemporaries or forebears by organising spectacles in which hundreds of thousands of nonhuman animals and humans were slaughtered in arenas in front of crowds of spectators. This phenomenon may represent one of the low points in the history of human-nonhuman animal interaction, but historically the killing is exceptional for its scale rather than its cruelty. Because studies have centred primarily on the human aspect of the interaction and the cultural function of the spectacles in Roman society, they emphasize their Roman-ness, which distances them from modern experience. An exception is Garrett Fagan's social-psychological study of what motivated many thousands of people to watch them and how the crowd behaved. (Fagan 2011) As he demonstrates, events involving human violence towards animals and each other and their attraction of large audiences have not been special to Rome. Ritualized hunts in the form of bullfighting and human-organized combats between animals are still widespread, albeit usually on a smaller scale and mainly involving relatively small animals like cockerels, dogs and badgers because they are illegal in many countries. While organized large-scale hunts of animals had been held in the Near East before Rome's expansion and continued to occur in Asia after Rome had fallen, some of them on royal hunting reserves, as events of this type held in enclosed purpose-built arenas the Roman ones are unparalleled in size or number.

The sources
Most of the written sources used here belong to the late Republican and early imperial period. Pliny wrote a lot about animals in his Naturalis Historia, which may have been partially founded on observation and anthropomorphic interpretation of what he , both satirizing Roman life and behaviour. Lastly, some Roman historians refer to animals in the games.  served Trajan and Hadrian as a librarian and secretary and wrote a collective biography of Julius Caesar and eleven emperors, and the historian Dio Cassius (c.155-235CE) wrote a monumental history of Rome. Not all of it has survived, but among the extant parts is his discussion of the 55BCE games. Roman historians took a moral perspective and frequently chose to cast historical personages and their actions in either a positive or negative light.
The limited number of sources, especially for the later centuries, the different genres and purposes of the authors and the time span of over four hundred years make generalisation difficult, although there is no indication of any significant change in the character of the animal spectacles in the arenas. As far as human-animal relations and animal agency in the arena is concerned the written sources are not only anthropogenic but largely those of writers or artists who accepted the mores of their society. In the period under discussion of the Roman games the most widespread belief systems, at least among the elite, were Stoicism, which flourished from the Hellenistic Era to the second century CE, to a lesser extent neo-Pythagoreanism in the early imperial period, Platonism, which became especially influential in its "Neo-Platonist" form from the third century CE onwards, Manichaeism, which began in Persia in the third century but spread rapidly before being suppressed, and Christianity, which steadily grew as a subversive religion and became the state religion in the fourth century CE. Hence beliefs that formed the background to the human destruction of other animals for reasons other than nutrition or product manufacture did change, but with very few exceptions the representatives of all of them belittled nonhuman animal cognition and excluded them from moral consideration by arguing that unlike humans they had no rationality and no moral sense. As Seneca the Younger put it, "Impulses towards useful objects and revulsion from the opposite are according to nature; without any reflection to prompt the idea, and without any advice, whatever Nature has prescribed, is done." (Seneca, n.d.) 2 Cicero and Pliny were both heavily influenced by Stoicism, and even after it declined as a distinct school many of its precepts concerning animals were retained by some Neoplatonists and especially by Christians. (Sorabji 1993, 186-88, 195-98) The Acts of the Christian Martyrs are an exception among the written sources because they were written by partisans of those condemned to die for failure to accept the mores of Roman society. The accounts include fantastical and eschatological elements as well as dreams and realistically described incidents. They give a different perspective on animal agency, but even more so than the accounts of most pagan commentators, they failed to attribute agency to animals even when believable actions are detailed. They no longer act as mere beasts, but as if humans or under compulsion from the devil or sometimes God. Persecutions occurred intermittently from 64 to 312CE, and they probably contributed to the more negative attitude towards arena activities that is shown by many Christian writers, among them Tertullian (c.155-c.240CE) and three who became bishops, Eusebius of Caesarea (260/265-339/340), Palladius of Aspuna (d. c.430) and Augustine of Hippo (354-430CE). However, negativity towards the enjoyment of the arena spectacles was not confined to them: Stoics, Platonists and Christians alike often disdained the pursuit of such pleasures in preference to dedication to selfimprovement, and among the elite there was often a snobbish contempt for those who attended the events. Seneca advised his correspondent Lucilius to avoid crowds and not to attend the games, objecting to the bloodshed inflicted on humans as a spectacle and maintaining that only the worst of people watched the executions. (Seneca,n.d.) 3 Despite the scale of the activities and the logistical effort that went into capturing and transporting the animals and organizing the events, archaeological remains offer limited information. Some arenas are still partially intact, but most useful for our purposes are the considerable number of surviving mosaics and wall paintings that depict events involving animals, which provide evidence of activities and those involved in them that is often missing from extant written descriptions.

Animal agency
History and ethnology have long traditions of examining "otherness," but the study of animal otherness and identification of animal agency are relatively new fields that have developed in parallel with posthumanism. The interrelationship between the physical environment, humans and nonhuman animals is complicated; however, before the last century the picture was simplified, often inadvertently, by the tendency to overestimate the extent of human control over parts of the environment and its inhabitants. The development of agriculture and animal domestication has been accompanied by the development of human belief systems that represent the "human-tamed" or "human-created" environment as order and the environment unaltered by humans as "wild" or sometimes as aligned with the forces of chaos.
"Agency" has been defined in several ways. Historians and archaeologists have now ceased to exclude other animals from the narratives of humanity's history, recognizing their agency in forming society, in which they provide food, labour and raw materials for clothes and tools, companionship, and entertainment. Here I do not treat the impression made by inanimate objects, including dead animals, on their environment or animate beings as "agency", where it functions as a synonym for "influence" or "effect". Nor do I take "rational" choice (insofar as it occurs or can be defined) or intent as a prerequisite for agency. I take agency to imply action by living beings, directed or intended by the agent or enacted as an involuntary response to hunger, illness, pain or injury, independently of compulsion from other agents. Whether agency involves the body or the mind of the agent is not in question as modern cognitive and psychological science does not consider the mind and body of any living animal agent, human or nonhuman, to be separable. (Lakoff and Johnson 2003) However, Roman and Greco-Roman philosophers, and probably spectators and human participants in animal shows, did consider the soul or its intellect capable of functioning independently of the body, and even today this body-mind/soul duality is still widely assumed of living beings. Following Aristotle's example, western philosophers and theologians argued for millennia that only humans make rational decisions and devised various means to explain animal decision-making as non-rational. The word "rational" as a defining term is, however, rare in modern works on human or animal cognition; modern theorists generally argue that Greek and Roman natural philosophers underestimated nonhuman animal capabilities and overestimated human rationality by underestimating the limits imposed by, for example, cognitive capacity, the tractability of the problem and the tendency to subordinate rationality to normative and affective considerations. (Sturm 2012, 66-81;Harman 2003, 457-59) Defining the "animal" in "animal agency" has not always been straightforward either, but in the relatively confined context of this article it refers to all nonhuman sentient beings who appeared in the spectacles of the arena. Insofar as possible, the posthumanist agenda demands that we escape human constructs, but attempting to assess how animal cognition functions or what its material and sensual lived experience is difficult. Nor is distinguishing nonhuman animal agency easy, as the actions of animals as recorded in historical records have already been interpreted by the (human) writer/s and are interpreted again by the (human) reader, both having their own cultural constructs. An additional problem is identifying the cultural biases of humans who lived 1500-2000 years ago. Nowadays there are resources and methods at our disposal that can often tell us when and what part of an animal brain is active and knowledge of the brain may enable us to identify stress or emotions such as fear, but the precise "train of thought", if such there is, is unidentifiable.
Though their theories about animal behaviour and cognition are now largely discredited, many Greco-Roman natural philosophers did attempt to understand animals. Their conceptions of animal behaviour were often anthropomorphic, but this does not wholly invalidate them. Anthropomorphism is often considered an innate tendency in human psychology. With less justification, it is generally taken as a negative phenomenon. However, while the characterization of animals in tales and fables usually contributes nothing to understanding of nonhuman animal cognition or behaviour, anthropomorphism can also facilitate inferences about animal thought and feeling that provide a basis for hypotheses that can be empirically tested. (De Waal 2016) Following Steven Wagschal's example, I refer to anthropomorphism used in this way as "constructive." (Wagschall 2018, 9) Apart from mine, the human standpoints that are obviously important in this article are those of the Roman or Greco-Roman writers who wrote about the arena animals. The anthropomorphising of non-human animals that occurred because of the animals' actions, whether by the literate Greco-Roman elite or the less educated among spectators, was often an implicit recognition of animal agency.

The cultural background
At least since the advent of agriculture, humans have attempted to control the behaviour of animals and to manipulate their habits and even genes to produce animals more amenable for them to handle and more productive of food or raw materials. At the same time humans have usually represented animals that have proved largely unmanipulable as destructive forces which have to be controlled or eliminated. The Roman animal spectacles were ritualized enactments of this process. Within the human-tamed environment the perceiver often represents those outside her or his own cultural sphere, whether nonhuman or human, respectively as wild or "barbaric." The cultural sphere of the Romans was one in which their mores and laws were recognized. The Pax Romana of the early imperial era was an "enforced peace", a peace only for those who accepted it, while those slaughtered in the arena signified different groups who did not belong to the imperial system: criminals and dissidents who by their actions had refused to accept the mores of society, "barbarians" captured in war or slaves who fought as gladiators, and wild animals. The importance of the "wild" element to the Romans is emphasized by their preference for acquiring animals taken from their own habitat for use in the arena rather than breeding them in captivity, which would have been cheaper and easier. Some were nevertheless bred in this way, such as the elephants who appeared in Germanicus' show of 12CE. (Aelian 1958, 103) In modern events where animals perform for human amusement, the hope of those organizing the show and the human trainers is that the animal will perform as intended by them. The audience expects the same but is not necessarily disappointed if the animal does something unplanned that is dangerous to trainers or otherwise exciting. The same applies to most events in the Roman arenas, but the staged hunts required animal unpredictability, which means animal agency, for added interest. Where Roman shows differed from most modern ones is that the animals were often encouraged to be more violent than they would be in their natural habitat.
In the Roman Empire hunting and capturing animals was a major and continuous operation that involved regional administrations, the Roman army and private contractors. By the fifth century CE animals suitable for arena spectacles were becoming harder to find in the empire and bordering areas. Safely caging and transporting animals by land or sea from remote provinces to the capital was extremely complex -they were of no use for show or the arena if they arrived sick, emaciated or dead. Documents such as the correspondence between Cicero, as governor of Cilicia, and Marcus Caelius Rufus, and the letters of Symmachus repeatedly mention the difficulties and delays of acquiring animals, although in Cicero's case his lack of enthusiasm for the task was a contributory factor. However, the logistics of the games is relevant here only insofar as it affected the actions of the animals in the shows.

Types of event and the animals involved
Animals participated in various types of events in the Roman arenas. They might perform tricks alien to their nature, as they have in modern circuses. Ancient sources refer to some remarkable acts; Aelian says elephants danced and "dined" in Germanicus' games of 12CE, and Suetonius and Martial, the latter a witness, mention elephants walking along tightropes and dancing, during the reign of Tiberius and in 80CE respectively. (Aelian 1958, 105-09;Suetonius 1997, 200;Martial 1919, 94) In his Life of Apollonius Philostratus refers to the trained elephant allowing its master to place his head in its mouth. (Philostratus 2012, n.d.) 4 The second function of animals was to fight with one another. As in the gladiatorial combats, in which differently armed and armoured opponents were pitted against each other to provide the most interesting spectacle, different species of animal usually opposed one another, such as bears and lions. The spectacle might be a "duel" involv-ing just two animals, which were sometimes chained together like the bull and bear depicted in the Zliten mosaic (Dunbabin 1999, 119-20), but as in the same mosaic, often many different animals were released into the arena at once, in the hope that they would attack one another. In his Natural History Pliny mentions that elephants fought one another and against bulls. (Pliny 1906, n.d., 8.7.;Martial, 1919, 15) 5 The third function was to take part in the spectacle of a ritualized hunt, usually in which trained venatores ("hunters") fought and killed them, and occasionally were killed by them. Examples of all three types of event are depicted in mosaics from various regions.
Animals also took part in plays, frequently re-enactments, sometimes with altered plot, of mythological scenes. The most infamous of these is the sexual act between a bull and a woman substituting for Pasiphae, the mythological Cretan queen, which took place at the opening of the Colosseum in 80CE. (Martial 1919, 6) Lastly, animals functioned as instruments of execution for criminals and enemies of the state who suffered damnatio ad bestias, condemnation to death by beast. This spectacle, regarded as a degrading death for the victims as well as a suitably gory spectacle for the crowd, was sometimes combined with the venatio. The Romans distinguished most "ferocious" animals as bestiae dentatae, "toothed (or tusked) animals", thus including animals such as wild boars as well as predators. There are no extant records of execution by elephant in the imperial period, but a historical anecdote of the first-century CE writer Valerius Maximus records that they were used to trample army deserters to death in public in 168 and 146BCE (Valerius Maximus 2000, n.d.). 6 The trainers had to "persuade" animals to behave unnaturally if they were to perform tricks, but there were also common activities that required the animals to be aggressive if they were to satisfy the spectators and which were either unnatural or enacted in unnatural circumstances. Even predators are not habitually aggressive by nature; they attack other animals that threaten their young, themselves or their social group, which often has a defined territory, or to obtain food. But humans did and often still do define them by their supposed savagery. Other animals who are not predators but can kill other large animals or humans, usually when defending against a perceived threat to themselves or their young, including hippos, rhinos, elephants, wild boars, buffalo and bulls, all of whom appeared in the arenas, may also be known to humans as vicious. Wild animals were often encountered by humans when at bay during a hunt or trapped in nets and pits, when it is not surprising that they reacted fiercely. Shows in which animals behaved savagely and attacked people or each other demonstrated their (supposed) otherness and lawlessness, and if defeated by humans, the power of humans and Rome to control and suppress such activities, whereas shows in which animals behave contrary to nature, either by performing tricks or, if predators, by showing gentleness to potential victims, functioned as proof of human ability to dominate them and to "civilize" the wild, to train the apparently untrainable and alter nature to fulfil human requirements. As one nineteenth-century reporter put it when referring to elephant circus acts, they seemed to "reverse the order of things." (Nance 2013, 65) The requirement for savagery explains why many of the animals appearing in the arenas were predators or other animals capable of "anger" when provoked, but ferocity was not the only prerequisite. Size, strength, appearance and exoticism were important, as the organisers of the events wanted to impress the audience. Elephants appeared regularly, but by 80BCE they were not particularly exotic, the Romans having met them in battle since the third century BCE and captured many of them. Roman impressions of them were coloured by their use by enemies in warfare, most famously Pyrrhus (319/18-272BCE) and Hannibal (247-183BCE), although they were last instrumental in defeating a Roman army in 255BCE. In the pre-games era, in 251BCE, in what was probably their first appearance before a crowd in Rome, captured elephants had appeared in the circus being driven round the circuit by slaves carrying blunted spears, very likely in part to belittle what was then a feared animal weapon of the Carthaginians during the First Punic War. As Jennison noted, it would have been possible to manipulate the elephants into running before the slaves by placing mahouts in the crowd to call them or some such trick. (Jennison 2005, 44) In Pliny's Natural History his account of each animal species usually includes its first appearance in the arena, which reflects the importance of the games to Roman perceptions of animals. According to him, an elephant first appeared in the circus in 99BCE. Pliny took a special interest in elephants: The elephant is the largest off these [land animals] and is closest to man in perception, for it understands the language of its country and obeys commands. It remembers the duties that were taught to it, the enjoyment of love and fame, indeed, to an extent that is rare even among men, honesty, prudence and equity; it also has a religious respect for the stars and venerates the sun and moon. They are also believed to understand the different religious beliefs, [...] and when they are about to cross the sea they will not board the ship until their keeper has sworn an oath that they will return, and because even these enormous creatures are threat-ened by disease, they have been seen, when tired from illness, lying on their back throwing vegetation up towards heaven just as if they were offering prayers asking the earth to act on their behalf. (Pliny 1906, n.d.) 7 The view of Pliny and Aelian that their ability to learn tricks is evidence of their intelligence is still prevalent, but so too is the view of Augustine, who saw trick-learning as clear evidence of animal subordination to humans. (Nance 2013, 65;Pliny 1906, n.d., 8.1.;Aelian 1958, 103;Augustine 1982, 43-44) 8 As Ingvild Gilhus notes, trick-learning tends to turn animals into abortive humans. (Gilhus 2006, 73) Apollonius of Tyana (c.15-c.100CE), according to Philostratus, also had a high opinion of elephants, partly because they were allegedly instinctively obedient to humans: "This animal," he [Apollonius] answered, "is docile beyond all others; and when he has once been broken in to serve man, he will put up with anything at the hands of man, and he makes it his business to be tractable and obedient to him, and he loves to eat out of his hands, in the way little dogs do; and when his master approaches he fondles him with his trunk, and he will allow him to thrust his head into his jaws… But of a night the elephant is said to lament his state of slavery, yes by heaven, not by trumpeting in his ordinary way, but by wailing mournfully and piteously. And if a man comes upon him when he is lamenting in this way, the elephant stops his dirge at once as if he were ashamed. Such control, O Damis, has he over himself, and it is his instinctive obedience which actuates him rather than the man who sits upon and directs him. (Philostratus 2012, n.d.) 9 The unusual aspect of the above account is its description of service to humans as slavery that elephants dislike, although the description of their actions as involving human emotions and behaviours such as mourning, self-control and perhaps shame is obviously anthropomorphic. Like Pliny, Philostratus' Apollonius respects the elephant as having certain Greco-Roman attributes, but here esteemed as a valued human slave might be, one who recognises his station and duty as subservient but nevertheless regrets that he or she has ended up in that situation. Cicero too understood that animals suffered in captivity, but his recognition of their desire to wander functions as a reinforcement for his point that all humans have an urge to mental or physical stimulation. (Cicero 1931, 456-57) 6 Capture and conditioning of animals for the arenas Philostratus also mentions "breaking in" of elephants, their reduction to servile status. For their task in the arena in 55BCE the elephants would not have required any special preparation having travelled to Rome. Pliny explains that when captured, elephants were chased into pits by men on horseback and then left without food or water until they were physically weak enough for relatively easy transport. Aelian gives a more detailed explanation of this "breaking-in" process, adding that having been captured the elephants were kept apart with ropes and then tied to large trees so that their movement was limited. Then they were drained of their strength by denial of food, to "reduce their spirit and their inflexible determination, so that they forget their hitherto indomitable fierceness and abandon their former temper." Then they were offered food from the hand, "when they already wore a mild and fatigued expression." But some extremely powerful and full-grown elephants burst their bonds and tore up and broke trees, and had to be tamed by alternate feeding and starvation and with goads. (Aelian 1959, 296-99) The Roman animal handlers presumably had a reasonable knowledge of how to get the animals they transported and trained to kill or to do tricks, whether they had any respect for them or not. We know from modern studies that animals react badly to being removed from their natural habitat, and often worse still from being held in very restricted spaces. An example of what might go wrong appears in one of Symmachus' letters, where he mentions that crocodiles he had acquired at great expense "refused to eat" after one appearance in public and became so weak that they had to be killed. (Symmachus, n.d.) Probably they had an illness that affected their digestion and which Symmachus' handlers failed to recognise, such as adenoviral hepatitis, salmonellosis or a nutritional disease, which can easily be transmitted from meat fed to them or other animals. (Huchzermeyer 2002, 265-70) It was important to the patrons of the games that the animals could attack their opponents when they appeared in front of the crowd, but this did not always happen.
References to the training of animals for the Roman games are rare. Not all animals went straight to the arena. Since the acquisition of exotic creatures was expensive, they would often be sent to holding centres, menageries or zoological gardens around Rome or other places with arenas to be tamed and trained for public entertainment before they reached the games, where they would usually die. Like the one outside the Porta Maggiore at Rome, most holding places were presumably in the less densely populated districts, but animals sometimes escaped and residents were forced to shut themselves in their homes while they roamed the streets before being captured. In his Historia Miscella, Landolfus Sagax (late 10th c.) refers to an elephant that escaped and killed or injured many people in Constantinople in year 23 of Justinian's reign (527-565). (Landolfus Sagax 1851, n.d.) 10 The behaviour of this elephant is analogous to the destructive behaviour of nineteenth-century escaped circus elephants that had been denied the opportunity for species-specific activity and subjected to training involving violence, and the so-called "rogue elephants" of India, driven by destruction of their habitat and social structure to invade and damage areas of human habitation. (Nance 2013, 104-18) In elephant social structure a herd's matriarch serves as the ultimate arbitrator in social relations, and older bull elephants teach younger males how to act. While in male-only groups the hierarchy is adjustable to assist reproduction and survival, but conflict may occur, especially during musth when bull elephants become very aggressive. (Sukumar 2003, 180-86) Possibly this was the condition of the elephants Aelian refers to as tearing up trees, but the disruption that must have occurred with capture by humans removed the social controls that existed in the wild and was likely to make elephant behaviour more unpredictable. Ancient Indian mahouts certainly knew a lot about elephant behaviour, but we do not know how much of this knowledge passed to Roman trainers. Probably their level of knowledge was comparable to that of the nineteenth-century American circus trainers described by Susan Nance. The accounts of Aelian and others strongly suggest that the Roman trainers would have had a similar desire to achieve absolute dominance, a situation unfamiliar to elephants that had experienced "negotiable dominance" in their own social groups in the wild.
In the Roman period considerable time and effort went into the training, which was certainly not always done in a humane and painless way. The Roman training methods were probably not unlike those used in later circuses. Animals resist doing acts that are not required in their natural surroundings or unfamiliar to them. In most circumstances animals are aware when actions may be dangerous to their wellbeing and have no natural purpose; hence they are given a purpose, either a reward such as a treat or avoidance of painful punishment. In certain circumstances this apparently considered attitude may be overridden by powerful impulses such as panic, to which humans are prone just as many other vertebrates are. The goad, known as a hawkus, hendoo or elephant hook in English, was used to subdue and intimidate elephants in circuses during the last two centuries. The same instrument was known in ancient India and often carried by mahouts in ancient battles. In the Roman world even predatory animals such as big cats or canines had to be persuaded to attack humans when required to, and herbivores lacked the incentive of flesh as a meal.
The trainers attempted to achieve a balance between getting the animals to do as they wished but not making them too weak to act. Human-nonhuman animal interaction of this sort raises potential ambivalences between animal intent, training and instinct. Apart from the starvation and beatings used to make them tractable in training, or in the case of predators to make them hungry, we know that whips, burning brands and chaining to victims were used in the arena to make them act as demanded while discouraging them from attacking the trainers. Elephants must have recognised and feared the pain the hook could cause, as they did in modern circuses, where simply the sight of it would usually produce obedience. All animals also had to be encouraged to perform in front of a large and noisy crowd. Assessment of arena capacity is dependent on estimated seat area per spectator. The largest Roman arena that regularly held combats and venationes was the Colosseum, built in 79CE and designed to hold over 30,000 people, but venationes also took place at regular intervals in the Circus Maximus in Rome, whose primary purpose was to hold chariot races, and probably equivalent arenas elsewhere. The Circus Maximus held between 115,000 and 220,000 people (Meijer 2010, 32-39;Rose 2005, 118) References to animals refusing to enter the arena or otherwise act as required despite the brutal "encouragement" they received show that many were traumatized by their capture, training and the arena crowd.

The animals in the arenas
There is an obvious problem with the methods detailed by Aelian if elephants were intended for a venatio or combat with other beasts: they would have been too docile for the crowd's liking. The trainers had to aggravate them again, at some risk to themselves, but alternate taming and aggravating increased the already high risk of the animal not acting as required in the arena. The Passion of Perpetua and Felicitas gives a plausible account of the death of a Christian called Saturus in which several animals did not act as planned. First he was tied to a wild boar, but having dragged him for a short distance the boar turned on its handler and gored him so that he later died. The task passed to a bear, but it refused to leave its cage, presumably despite pressure to do so. In the martyrologies this sort of behaviour was interpreted as the work of God. A leopard finally mauled Saturus, producing so much blood that spectators mockingly shouted the bath-house phrase "Well-washed!" (Salvum lotum!). But still a human executioner had to finish him off. (Musurillo 1972, 126-31) We cannot assess the precise cognitive processes of boars and bears, but there were many factors that might have contributed to their disobedience. Nor do we know how the crowd reacted to the death of the boar's handler, but handlers were held in low regard. Martial refers to a lion that wounds its trainer as "perfidious", "with ungrateful jaws", as if it had been trained with kindness, and describes it as justly punished at the hands of the venatores. (Martial 1919, 8) We should bear in mind Martial's sardonic style, but if we were to adopt his anthropomorphism and assume that nonhuman animals should be able to show gratitude and loyalty and be adjudged guilty if they did not, we might attribute to both the lion and the boar another human emotion, vengefulness for captivity and probable rough treatment. However, given that the boar's trainer was provoking when he was attacked it was not necessary for the boar to recognise him as an earlier tormentor. The tendency of humans when animal action excites or interests them, especially if it challenges human dominance, is to adopt an anthropomorphic theory of mind rather than the "Morgan's Canon" principle that "in no case may we interpret an action as the outcome of the exercise of a higher psychical faculty if it can be interpreted as the outcome of the exercise of one which stands lower in the psychological scale." (Morgan 1903, 53).
A similar problem arises with the question of whether animals were distracted by unarmed intended victims during a combined damnatio and venatio; can animals distinguish humans that are primary threats from those that are not, in this case armed and mobile humans from unarmed ones that may have been tied to one spot, a question that is related to an animal's "learning ability"? Virtually every known vertebrate species reacts more to another moving animal than a stationary one, especially if approached, whether by attacking it or seeking safety in flight. In his Church History, Eusebius of Caesaria, who witnessed persecutions at Tyre, says the "man-eating beasts" ignored naked victims tied to stakes to attack the armed men who were goading them forward. (Eusebius 1890, n.d.) 11 The event organisers had no way of knowing exactly how any animal would act in the arena. The damnatio victims, if not bound, may have attempted to evade the beasts or fight them with bare hands, but many Christians remained stationary to await their fate, as one of their objectives seems to have been to show that they were prepared to die for their faith and to disappoint, not to say infuriate, the crowd.
The animals were at a considerable disadvantage in most venationes. Even if they were sufficiently aggressive and attacked, they were in an unfamiliar situation.
The arena would usually be flat, open and restricted in size, surrounded by a large and noisy crowd that they were probably not accustomed to. In the show of 55BCE the elephants faced an unspecified number of Gaetulian javelinmen. The status of these men and how they ended up in Rome is unknown: they obviously had skills acquired during their upbringing, but they were probably not professional venatores like those depicted in some later mosaics. The professionals would have been trained and knowledgeable about what to expect from the animals they faced, and if they appeared as a team they supported one another. A well-known depiction of them in action is the so-called Magerius mosaic from Smirat (Tunisia), on which four men from a company called the Telegenii dispatch leopards or cheetahs with spears. (Beschaouch 1966, 137) They are evidently practised in killing animals and they receive bags of money as a reward for the show, 500 denarii for each animal killed. In theory the money was given by the crowd, supposedly the "lords" (domini) of the proceedings, but the patron provided it. (Fagan 2011, 129-30) In the arena, faced with the venatores in daylight and in a restricted and usually open and level space, animals were denied many of the advantages they might have in their natural habitat, especially predators with heightened senses, particularly hearing and smell and good night vision, such as tigers, leopards or crocodiles. Many, like leopards and tigers, were solitary hunters by nature and ambush predators. In the wild they use grass and bush as well as undulations in the ground as cover to approach prey. Lions and canines such as wolves and hyenas hunt in their family groups with a degree of coordination, also using cover to approach prey. For the elephants in their combat of 55BCE, the main problem would have been the restricted space; as we shall see, an attempt to escape the confines of the arena may have played a major part in the peculiarity of the crowd reaction. In all cases, some of the methods of preparing animals for the contests must have weakened them, at least by comparison with their state in the wild. Changes in behaviour that demonstrate stress have been identified in animals in relatively modern zoos, but capture and transportation to a Western zoo nowadays, albeit traumatic for the animal, is likely to be quicker and more comfortable than it was in the Roman era.
The venatio was thus heavily weighted against the animal when the humans were venatores, but there is no evidence of a deliberate effort to protect human contestants' lives. The crowd no doubt admired the skill of venatores and generally favoured them rather than the beasts. As in the case of the gladiators, they probably had their heroes, such as Carpophorus, observed by Martial to have killed over 20 lions and other animals, but both groups had a low social status and the venatores the lower status as combatants. (Martial 1919, 12) The risk that they would be seriously injured or die was built into the event and one of its attractions. The nature of the arena was primarily determined by its purpose: to provide entertainment visible to all the spectators within the structure and without risk to them. (Scobie 1988;Rose 2005) Moreover, not all the animals' opponents in venationes were professional or experienced hunters. Some were likely slaves, criminals or prisoners of war like the Jews taken prisoner by Titus during the Jewish Revolt (66-73CE) and sent into the arenas to face animals in 70CE. (Josephus 2009, n.d.) 12 Against unskilled and unarmed or inappropriately armed human opponents, animals with the capacity to kill would have had the advantage.

The slaughter of the elephants
In his description of the event of 55BCE, Pliny says the crowd were delighted with the elephants' actions, especially a wounded one, which, though brought to its knees, seized the shields of its tormentors with its trunk and hurled them up in the air so that they fell to the ground in a great arc around it. But this excitement does not signify any form of empathy. The wounded elephant was performing an act resembling a trick, but one not taught by humans. We can only speculate as to whether the crowd perceived this particular "trick" in the same way as those taught to the animals or whether they recognized it as the elephant's own agency. If the latter, the shield-hurling neither made the elephant an abortive human nor functioned as evidence that humans could manipulate and dominate wild animals.
Pliny also says the mob were delighted when an elephant was killed outright when a javelin penetrated its brain under the eye. Evidence from other sources suggests that this attitude was the norm. Most disturbing to modern people unaccustomed to watching animals die in agony is the mosaic panel from Torre Nuova, which shows the death of three leopards pierced by barbed spears. The writhing poses are realistic and one is bleeding copiously as it is being transfixed by two venatores. Toynbee wonders that householders could wish to see such a scene every day in their homes. (Toynbee 2013, 83;Dorigo 1971, pl. 115) But they obviously did want to, which merely emphasizes that once a living being, whether human or nonhuman, becomes "the other" and is perceived as a threat to the existing social structure it does not attract affective empathy but its death is seen as just and maintaining or restorative of the correct order, and may indeed cause people to rejoice in its suffering. (Brown 1992, 184, 200) Similarly, there is no trace of empathy for animals in the writings of arena witnesses such as Martial. Christians critical of the arena events such as Tertullian or Augustine do not mention the fate of the animals at all, and Augustine was adamant that animals were without reason and outside moral consideration. The sources that object to the slaughter often cite it as wasteful rather than cruel. Referring to the events of 55BCE, Cicero wrote, "But what pleasure can it possibly be to a man of culture when either a most feeble man is mangled by a powerful beast or a splendid beast is transfixed by a hunting-spear?" (Cicero 1957, 7) However, elsewhere he expressed his Stoic beliefs "that other things were born for the sake of mankind and the gods, but that humans were born for their own fellowship and society, so that humans can use them for their own purposes without wrongdoing." (Cicero 1931, 288-89) Cicero's critique of the arena spectacles was made in a letter in which he tried to console a sick friend who missed them and insofar as it registers disgust, he objects to the destruction of fine objects as he might to smashing a quality vase. Over four centuries later, we find a similar objection in the Christian ascetic On the races of India and the Brahmans, commonly attributed to Palladius. The ascetic "Brahman" Dandamis defends vegetarianism and criticizes contemporary Romans for taking their "fellow creatures" and "works of God" to the cities to be abused and annihilated rather than to breed and produce something useful. (Palladius 1665, 45-46) However, both Pliny's and Cicero's comments indicate that the attitude of the crowd changed during Pompey's elephant venatio. Pliny lived a century after the 55BCE games and knew of the events from earlier sources, but we have a brief reference to the elephant show in a letter by the witness Cicero, who describes "a great wonder arising in the common throng, but no pleasure. Indeed, a certain pity arose, a sort of belief that the huge beast had something in common with humans." (Cicero 1957, 7) Pliny explains that at its end the elephants, having failed to escape, ceased to fight and merely walked around "lamenting", raising their trunks to heaven and wailing. Some were presumably in considerable pain, as it would often have taken many javelins to kill one. Nothing in Pliny's account of the behaviour of the elephants suggests that they had been made docile by their transfer and incarceration in Rome. According to him, the spectators interpreted their noise and actions as appeals for mercy, and "all" were moved to tears and cursed Pompey. If Pliny was right and not excessively influenced by his own beliefs about elephants, the crowd exercised an anthropomorphic theory of mind, in this case one that enabled affective empathy for the dying elephants. In the terms he presents in his Natural History, Pliny might have understood the elephants as "speaking", not just by uttering sounds but by their gestures. He, and probably some among the crowd that witnessed the event, were unwittingly anticipating modern theories of language that transcend the logos (reason and human speech) and allow for nonhuman animal communication. Having become susceptible to empathy with the elephants, they became receptive to their language of gestures and sounds. There need be no doubt that the elephants were communicating something, almost certainly relating to their distress, whether to one another, to the crowd, or, if we were to adopt the Morgan's Canon principle, simply pain. It seems that many Romans thought they cried out to the spectators.
Cicero, Pliny and Dio Cassius were intrigued by the pity for animals shown by the crowd, which we are safe in assuming was exceptional by their reactions. Dio is not concerned for the animals who died in the arenas; he is highly critical of the emperor Commodus for making personal appearances in the Colosseum, where he "hunted" (avoiding danger to himself) and killed hundreds of animals, but the root of this criticism was Commodus' persecution of senators. (Dio Cassius 1927, 107-13) Regarding the 55BCE events, Dio adds one piece of information to Pliny's, he or his source/s' account making use of another supposed elephant attribute mentioned by Pliny: "the rumour spread that they did not do it [lament] by chance, but cried out against the oaths in which they trusted when they crossed over from Libya, and they were calling on the gods for retribution." (Dio Cassius 1914, 363) He then mentions the tale that they would not board the ship without guarantees of their safety. Dio Cassius is more circumspect than Pliny: referring to the elephant "lament", he writes, "I do not know whether this happened or not, for some have also said that elephants can understand the language of their own countries and have knowledge of the heavens..." Perhaps the tone of the wailing was reminiscent of that made by injured humans, but why would this alone induce affective empathy in a crowd used to seeing and hearing the death of animals and humans they identified as enemies, or turn them against their benefactor, Pompey?
Animal herd behaviour is still not fully understood, but it is characteristic of herbivores and non-predators. One motive to bunch together is for protection against predators. While this functions well against attacks at close quarters, it is counterproductive against humans armed with missile weapons. Elephants, like other herd species, are prone to stampede as a group when one runs from a threat or simply something that startles it. This is what the elephants did during Pompey's spectacle, in this case a response to the attacks by the Gaetuli, Pliny tells us that they threatened to break down the iron barriers protecting the crowd and caused alarm amongst them. The propensity for elephants to "panic" and stampede in battle, causing mayhem in their own army, was well-known to all Roman generals who faced them. Elephant use for warfare was declining by the late Republic because of this tendency.
The barriers held during the arena stampede and the slaughter of the elephants continued, but Fagan may be right that it was at this point that the shift of mood to hostility to the patron occurred in the crowd, since it was an unwritten rule that the spectators were not endangered by the arena activities. (Scobie 1988, 197;Fagan 2010, 250) In turn, such a shift would have enabled affective empathy for the animals Pompey had sent to their deaths, particularly when they seemed to appeal for mercy, and empathy may enable a feeling that the nonhuman animal is more than an unfeeling destructive force. We can probably rely on Cicero's testimony that empathy developed, but how much of the crowd displeasure was sympathy for the elephants and how much irritation at Pompey we cannot know. We should also note two other factors. Dio Cassius mentions that there was popular hostility to Pompey's military alliance with Crassus at this time, although in general they were satisfied with the events he organised. (Dio Cassius 1914, 362-63) Secondly, the elephants' opponents in the arena, the Gaetuli, were a Berber people who lived beyond the frontier of Roman-controlled Mauretania, described by Sallust (86-c. 35BCE) as a "rude and uncivilized folk, who fed like beasts on the flesh of wild animals and the fruits of the earth. They were governed neither by institutions nor law, nor were they subject to anyone's rule": in other words, barbarians who sometimes fought against the Romans and with whom the crowd were unlikely to identify strongly. (Sallust 1931, 172-73) Throughout history social groups comprising humans who identify psychologically with each other have likened humans who have a different culture or do not follow the mores or customs of their group to beasts. Though Sallust admitted "I could not easily give an account based upon certain information", his view of "barbarians" was probably the "default view" of almost every Roman. (Sallust 1931, 170-71) 9 Animal heroes?
The above incident is the only recorded one in which a Roman crowd sympathised with animals in distress and "listened to them", but, as noted, there are other known cases of identification with animals, in these cases as victors. In many of the surviving mosaics and friezes of venationes the animals are named as well as the venatores, although these rarely feature elephants. In the Magerius mosaic there are four leopards called Victor, Crispinus, Luxurius and Romanus who have laurel wreaths round their bodies. The messages conveyed by some of the names are ambiguous, as they could be family names of people but also have other meanings. For example, Crispinus may be a combination of the words for "hair" and "pine" ("Curly"?). Other names, such as Rapidus for a leopard on a frieze from Lepcis Magna (Tripolitania), are simply descriptive. (Perkins, Toynbee and Fraser 1949, 181) Naming the animals gave them a more human-like status as opponents, but it was human agency, not animal. Nevertheless, it also allowed the crowd to recognise an animal that performed repeatedly or survived to fight again, even raising the possibility of identifying with one as they might with a human gladiator or venator. In such a case the animal's agency contributed to a shift in human attitude, just as in 55BCE.
The presence of an "animal hero" in a venatio raised the stakes and the excitement. There are records of crowds demanding that an animal be freed or rewarded for repeatedly evading injury from venatores, or for a bestia dentata "winning" in venationes, much as an unfree gladiator might be freed eventually. (Martial 1920, 427) The death of a lion who had obviously been victorious in previous encounters with other animals is related by Statius in his poem Leo Mansuetus, perhaps best described as a "lament without empathy." (Statius, n.d.;Jennison 2005, 79-80) 13 Even an animal that killed large numbers of condemned criminals or "deviants" may have become a "personality", kept alive because of its efficiency in acting as the handlers wished. It may have developed a taste for human flesh, as man-eating tigers, lions, leopards and crocodilians have occasionally done in Africa and India. The crowd's theory of mind attributed thought to these animals insofar as they achieved a status equivalent to a human competitor's, and therefore were due rewards for success. Leaving aside the probable anthropomorphic "reasoning" of the Roman crowd, an animal that survived the venationes or even killed venatores might be lucky initially, but as the animal trainers knew, animals do learn from experience, and not all those who faced animals were experienced hunters like the Gaetuli or professional venatores.
for human entertainment enabled them to survive capture and incarceration and to respond to the demands humans put on them, but also enabled them to respond in ways unplanned by their human trainers. The animals' frequent refusal to carry out executions as required may sometimes have been disappointing, but their unpredictability and ability to act, even to die, in surprising ways made the venationes more of a spectacle than unopposed killing would have been. The executions of criminals that usually occurred after the venationes and before the gladiatorial combats in the afternoons were often poorly attended, presumably because they were not contests and relatively monotonous.
Social historians have sometimes extrapolated from the experiences of famous human individuals to make points about famous animals. In a sense the "famous animals" in this article are the few individual animals mentioned in our sources and known by their recorded actions, who stand as representatives of their kind who found themselves in the same roles in the arena. Fame, as the state of being known of by many people, is a human experiential phenomenon. A nonhuman animal cannot experience fame but may become famous among humans, in which case it achieves in some measure a human status in their eyes. In the Roman world the slaughter of animals destined to become food or commodities was not hidden from most humans as it is now, but these animals died anonymously and without the opportunity to display agency that might make them memorable to humans. An animal sent to the arena was virtually certain to die there, but it was the centre of attention in an event important to both the wealthy organisers and the Roman public, and if it achieved something remarkable like killing a human in combat or repeatedly evading hunters' spears, or, as in the case of the elephants of Pompey's games, carrying out a series of actions that activated affective empathy, it might achieve semi-human status in human eyes. Many horses in the hugely popular Roman chariot races were also "animal heroes" known by name. This phenomenon is still prevalent: companion animals who are often the centre of attention to their "owners" are frequently treated as humans, animals that rescue human accident victims are rewarded as if they were humans, animals that repeatedly win in races are attributed characteristics possessed by human sportspeople, and even animals destined for slaughter that escape thereby acquire "personality" and are often spared the slaughterhouse.
The elephant show of 55BCE may be the only recorded example of affective empathy with animals in the Roman arena, but there are other records of spectators applying theory of mind, identifying an animal who did something notable as humanlike, which was a tacit recognition of agency. This form of anthropomorphism was an implicit challenge to the dominant human belief systems that denied other animals selfgenerated intent. However, nonhuman animals could not form any understanding of the human culture that placed them in the arena or make any organised or persistent challenge to their captors. In 55BCE the elephants were complaining as a meaningful reaction to their plight, but the meaning to them and what they felt is beyond us. Pliny was right that humans cannot understand nonhuman communication, at least not fully, but the understanding of human culture attributed to elephants by him and perhaps by spectators in Pompey's show was human, not elephant. By contrast with the arena animals, the human Christian victims of the damnatio understood their fellow-humans' intent which enabled a resistance that ultimately benefitted their co-religionists. They pondered their response before they entered the arena and it was subsequently recorded as systematically defiant, motivated by beliefs that other humans could understand to inspire more devotion to their faith.
Up to the fourth century CE there were a few Sceptics, Platonists and others who allowed nonhuman animals abilities that humans normally denied to them, but they were isolated voices within the elite. Whereas anthropomorphism may be constructive in modern natural science, it was rarely so in the Roman era because of the non-empirical methods of natural philosophy and belief systems that hindered any general empathy for nonhuman animals. Any implications that the arena public's brief attribution of fame to animals might have for their attitudes to nonhuman animals in general are hardly likely to have preoccupied them for long, if at all. Nevertheless, how and what nonhuman animals learnt was debated throughout the classical and medieval worlds because even many Stoic, Neoplatonist and Christian natural philosophers who denied them rationality had to recognize that many could and did acquire skills that were not innate and occasionally challenged human dominance. Throughout history human actions in relations with animals, including those in the Roman arenas, have often involved a human awareness of nonhuman animal agency, albeit unacknowledged as such.