Chapter 6


John Keegan’s book The Face of Battle: A Study of Agincourt, Waterloo am/ the Somme, published in 1976, popularized a new genre of modern writing about fighting. To the existing type of study which looked at battle from the perspective of the general was added another which looked at the physical and emotional experiences of the ordinary combatant. This chapter follows the latter model, looking in turn at ancient land battle, siege warfare, and naval warfare. In each section a very brief sketch of the development of techniques precedes an exploration of a specific dimension of the psychology of combat. In the final section, we look in detail at one battle; this brings together many of the themes of this book, and acts as a way into an investigation of leadership in classical war.

The hoplite

Consider a Greek hoplite (Figure 9). This one is a bronze statuette dedicated in the sanctuary of Zeus at Dodona in c. 500 bc. The amount of armour worn by a hoplite varied over time. But, with two exceptions, as we will see, this one seems typical. First, let us think about his equipment and its implications for fighting. He wears a bronze helmet, bronze body armour, and greaves. He carries a large shield. This hoplite carries a ‘Boeotian’ shield, which, like the ‘Dipylon’ shield we met in Chapter 3, may not have existed in reality.

9. Front and side views of a hoplite

Normally the hoplite carried a round, wooden shield faced with bronze, which was held by a central arm band and a hand grip at the edge. The statuette originally held a long thrusting spear in its right hand. Most hoplites would have had a sword as a secondary weapon. The equipment is heavy, hot, and tiring.

The helmet cuts down his vision and hearing. He carries nothing with which to fight at a distance. His battle will contain little manoeuvre (how well can he hear or see signals?), be fought hand to hand, and be short (before exhaustion sets in).

Now let us think about the man inside the equipment. Unless he is a Spartan or a mercenary, he is not a professional soldier.

Until the Peloponnesian War (431-404 bc), and quite often afterwards, he provides his own equipment, and serves without pay. He is relatively wealthy. Normally he is a farmer. He can be any age from late teens to about sixty. Professional instructors in weapon handling were known by the late 5th century bc, but were always the exception. Dancing and athletics were considered suitable training for combat. The man interacts with his equipment. His battle will be short, simple, and at close quarters.

Finally, let us imagine this individual’s place in the battle as a whole, using the most commonly accepted interpretation of hoplite combat. He takes his place in a phalanx: a closely packed body of men several ranks deep. He knows those around him, probably some of them are related to him. After an animal has been sacrificed, and the general has made a speech, the phalanx sets off towards the enemy. The men move at a walk, and sing the paean, or hymn, of their polls. Nearing the enemy, they break into a run, and shout a simple war cry. The aim is to crash into contact with the enemy. If their opponents are another phalanx, the othismos, or push, may result. This has been compared to a huge rugby scrum with deadly weapons added. The rear ranks physically push on the backs of those in front. Sooner or later one side will establish a forward momentum, and the losing side will break and run. Many casualties occur. Estimates put average casualties at about 5 per cent for the winners, and 14 per cent for the losers.

Some scholars see hoplite battle in a different way. They interpret the othismos as a figure of speech, arguing either that the rear ranks provided just moral support, or that hoplites actually fought in a looser formation altogether. It could be that a search for the norm is doomed to failure. Just as different Greek states adopted the hoplite phalanx at different times, so they may have developed somewhat different styles. We know that Spartans walked into contact, and that Thebans tended to fight in deeper formations than was the norm. It could be that the phalanx of a polis further varied its practice according to the circumstances of the day.

The hoplite phalanx was dominant in Greek pitched battle through the 7th to the later 4th centuries bc. Other troop types, light infantry and cavalry, were not unknown, and their use increased from the Peloponnesian War onwards. Yet there was a tendency to marginalize them, both in reality on the battlefield and ideologically by those writing history, because of the social and political control exercised by the hoplite class in most Greek states.

The phalangite

The armies with which Philip II of Macedon achieved hegemony over Greece, his son Alexander conquered the Persian empire, and the latter’s successors and the Hellenistic kingdoms they founded dominated the eastern Mediterranean were forces of combined arms. They consisted of a phalanx of heavy infantry for close combat, light infantry, and cavalry. The light infantry were equipped with javelins, slings, or bows, and operated as was conventional throughout antiquity, attempting to harass from a distance the main body of the enemy, while protecting the rest of their own side. The distinctive Macedonian cavalry were shock troops, although units of other cavalry made contributions (we will look at how cavalry functioned later on in this chapter). The core of a Macedonian-style army were the men in the phalanx. We know less about these phalangites than some modern works might lead one to assume. Our sources, literary and artistic, for Macedonian war-making tend to be about Alexander, and to focus closely on the figure of the king. As in pitched battle Alexander fought with the cavalry, we are poorly informed about the infantry. Our best source on the Macedonian phalanx, Polybius (18.28-32), was writing, in the second half of the 2nd century bc, to explain how the phalanx had been defeated by the Roman legions. Although Macedonian battle had changed, the phalanx becoming the main battle winner as the numbers and shock capacity of cavalry declined, Polybius seems to overstate his case that the phalanx was inflexible and incapable of operating in broken ground. Earlier, Alexander’s phalanx at the battle of Issus (333 bc), although it was roughly handled by an opposing phalanx of mercenary Greek hoplites, had been able to alter its depth and frontage on the battlefield and to fight across a river. As late as 197 bc, at the battle of Cynoscephalae, one wing of the Macedonian phalanx operating on a hillside was driving its Roman opponents back until attacked from behind.

The Macedonian phalanx’s key difference from a hoplite phalanx was its main weapon, the sarissa. Much scholarly effort has been devoted to the sarissa, with no end in sight. At least all agree that it was a long pike that was wielded with both hands. Its length allowed the spear-points of the first four or five ranks of the phalanx to project beyond the foremost men, as opposed to the two or three of a hoplite phalanx. The idea appears to have been to increase the width of the killing zone and the number of sharp points within it. The aim was to present the enemy with an impenetrable barrier of spear-points some feet from the phalanx itself. We should not elevate the sarissa to the status of a ‘superweapon’. If it had been, it is unlikely that the Spartans would have waited for over a hundred years after its appearance on Greek battlefields before introducing it into their own armies in 227 bc. For most Macedonian phalangites, as for most Greek hoplites, the battlefield must have been a place of very limited knowledge. Only the first few ranks could have any view of what was going on. The majority - tightly packed in the body of the phalanx, their view obscured by their colleagues and the dust kicked up, and their hearing assaulted by the din - would have been unable to form any objective understanding of the course of the battle. This being so, the morale of the phalangites was especially susceptible to noises which suggested that they were in danger to the flanks or rear, or any impression that the phalanx was giving ground.

The legionary

It seems that the Romans, like their neighbours the Etruscans, adopted the hoplite phalanx, probably at some point in the 7th or 6th centuries bc. In the 4th century bc, however, they seem to have instituted the famous legion. Over time, the organization of the legion changed, most notably from a formation centred on 30 small sub-units called maniples, as described by Polybius, to one of 10 larger sub-units called cohorts, as found in the writings of Julius Caesar. Similarly, its personnel altered from a militia of Italian farmers to long-service professionals recruited mainly from the provinces under the principate. Yet, apart from the disappearance of the spear, carried by a minority in Polybius’ description, the equipment of legionaries remained remarkably the same from our earliest evidence down to the later 3rd century ad. Legionaries wore a metal helmet and, usually, body armour, and carried a large, curved shield (scutum), one or two heavy javelins (pila, singular pilum), a sword (gladius), and a dagger (pugio).

Our evidence does not allow any certainty to reconstructions of the frontage occupied by a legionary in the battle line. The most common estimates are six or three feet. It is possible that the frontage varied over time or was dependent on the specific circumstances of the battle. Although the legionary could be ordered to use his pilum as a thrusting weapon, normally against cavalry, it was primarily a throwing weapon, and its weight gave it only a short range. In Caesar’s account of the battle of Pharsalus, we find a debate over the relative merits of momentum and cohesion {Civil War 3.92). Was it better to risk a loss of cohesion but gain momentum by charging, or reverse these goals by standing to receive an attack? Caesar and most Roman generals preferred the former.

Against other infantry the legionary would expect to advance into combat, throwing his pilum on command as part of a volley. While the shower of pila disordered the enemy, the legionary should draw his sword and close to hand to hand combat. The sword could be employed to stab or slash. The shield could be used offensively. Held by a central grip, and furnished with a metal boss, it could be punched into the face of an opponent, or, with the legionary’s shoulder behind it, be used to knock the enemy off balance. Comparison of skeletons from Maiden Castle in Britain, which was stormed by the Romans, with the far more numerous ones from the medieval battle of Wisby suggests that the aim of a legionary was to get his opponent on the ground and then butcher him with numerous heavy cuts to the head from his sword (see Figure 14, page 120). That such fighting was physically exhausting - and we can estimate that some infantry battles, like Cannae, lasted for hours - has led some modern scholars to hypothesize that at times such combat reverted to a ‘default state’, where the two sides would draw back and hurl missiles and insults at each other as they got their courage up for another short burst of hand to hand fighting. In such a ‘slogging match’, the legionary would be physically helped by his training. We are told that they trained with heavier weapons than those they used in combat (Vegetius 1.11). Also, psychologically he would be aided by both the concept of discipline (disciplina), which Romans considered that they had and all other peoples lacked, and the warrior myths of Rome which stressed the ‘long haul’, such as the mythical Horatio holding the bridge against an overwhelming force.

The cavalry

As Greece consists of a collection of arid and rugged peninsulas and islands, it is unsurprising that, although the ownership of horses was a mark of elite status, cavalry was not the main striking force of classical Greek armies, with the exception of those from the broader plains of Thessaly in the north. Still further north, the wide pastures and tribal society of Macedonia allowed the development of the effective cavalry which formed a vital element in the armies with which Philip II subdued Greece and Alexander conquered the Persian empire and various Indian peoples as far as the River Indus. Under the Hellenistic kingdoms which succeeded Alexander, the numbers and effectiveness of cavalry declined. Under the Republic, the Romans appear to have had a small but effective body of citizen cavalry. This, however, disappeared in the early 1st century bc, and, although very small numbers of citizen cavalry reappeared in the principate, Roman armies henceforth relied on foreign or subject auxiliary cavalry. From earliest times, cavalry was of secondary importance to heavy infantry in Roman armies. This began to change in the second half of the 3rd century ad. Roman armies of the 4th and 5th centuries relied heavily on cavalry, and by the 6th and 7th centuries, although infantry remained numerically dominant, the cavalry were the main strike force of the empire’s armies.

The types of cavalry in the ancient world can be imagined as a spectrum. At one end were true light cavalry, such as the Numidian and Moorish horsemen of North Africa. These relied on missile weapons and tried to avoid hand to hand combat altogether, or at least until the enemy were running. They used the speed and manoeuvrability of their horses to make themselves into mobile missile platforms. At the other end of the spectrum were true ‘shock’ cavalry, such as Alexander’s Companion cavalry. These were equipped only for close combat. As horses cannot be made to run into solid objects that they are unable to see through or jump, such as close-packed bodies of infantry, ‘shock’ cavalry rely on the combined bulk of horse and man to intimidate their opponents into running, or at least breaking the frontage of their formation so that the cavalrymen can get amongst them and ride or cut them down as individuals.

The potential of ancient cavalry can be illustrated by thinking about Plutarch’s account of the first, decisive day of the battle of Carrhae in 53 bc (Crassus 23-27). A Parthian army comprised entirely of cavalry, the majority light horse archers, with a smaller number of armoured cavalry equipped with lances, confronted a much larger Roman army, of legionaries supported by light infantry and cavalry, commanded by Crassus. A first charge by the Parthian cavalry was not pressed home as the legionaries presented a dense, unbroken line. This formation, however, gave the Parthian light archers an ideal target. The Romans endured the missiles believing that the Parthians would run out of ammunition. When it was seen that the horse archers were being resupplied, Crassus ordered his son, Publius, to attack with the right wing of the army. The Parthians retreated, shooting as they went, until Publius’ troops were separated from the main body. The Parthians then placed a unit of heavy cavalry to the front of Publius’ men. Publius was unable to persuade his infantry to attack, and so was bested in a cavalry melee. Falling back on his infantry, Publius drew up his remaining men on a small hill. There they were shot down by the Parthians, before finally being overrun by the heavy cavalry. The Parthians then turned back to the main body of Roman troops, the heavy cavalry attacking the front with their long lances, while the bowmen poured arrows into the flanks. When nightfall halted the battle, the Roman army had ceased to be an effective fighting force.

Motivation: only a few fight?

John Keegan’s The Face of Battle has encouraged classical historians to apply the conclusions of studies of psychology in modern combat to the ancient world. One example can illustrate the methodological care that must be employed in such work. In Men Against Fire, published in 1947, S. L. A. Marshall claimed that among American troops at the sharp end in the Second World War only one in four ever fired their weapon. A. K. Goldsworthy has applied this to the Roman army in two ways. Finding it implausible that men in close formations could have taken no action, he suggested that 75 per cent of missile troops took no aim when shooting, and the same percentage of heavy infantry would fight only defensively at close quarters. This straightforward transference must be doubted. Marshall himself did not think his figures were a universal norm. In a later work he argued that those firing had gone up to between 37 per cent and 55 per cent in the Korean War. Historians of modern war are now suspicious of Marshall’s statistics. Yet even were they correct, there are reasons to doubt their applicability to the ancient world.

Three steps should be taken when investigating whether a concept from modern military studies of combat motivation can be applied to the classical world: can any support be found in the ancient evidence; were the physical environments of battle close enough to allow the concept to ‘work’ in the ancient world; and did the underlying factors causing the phenomenon in modem times also exist in the past? On all three counts, the idea that only one in four Romans fought aggressively fails. We can use as a case study the anonymous War in Spain (Bellum Hispaniense) preserved among the works of Julius Caesar. This understudied, brief continuation of Caesar’s Commentaries gives us a rare view from someone below the higher ranks of society. Uninformed on the larger issues of strategy, the author is interested in the weather, soldier’s pay, low-level desertions, and military punishments. Long ago, Lord Macaulay guessed that he was ‘some sturdy centurion, who fought better than he wrote’. While the author does think that soldiers had different feelings waiting for battle to commence, and that the inexperienced on the other side were numbed with fear by the noise of combat, when he describes fighting, the men in units act as one. They all shoot, fight hand to hand, and refuse to close to combat, give ground, or run away. The only two exceptions do not fit the ‘one in four’ model are two individuals fighting a duel, and two centurions making the ultimate, individual sacrifice in an attempt to put heart in all the rest of the unit. The physical circumstances of American GIs and Roman troops were very different. The former tended to be prone and relatively isolated. The latter usually were standing and surrounded by comrades. Finally, apart from fear, Marshall thought that it was the Christian commandment ‘Thou shalt not kill’ that froze the GI’s trigger finger. It should go without saying that until the 4th century ad, and quite often after then, Roman soldiers were not Christians. Roman society was violent, with a legal right, even a moral imperative, to violent self-defence. Public executions drew big crowds, and gladiators were sex symbols. Although ‘Thou shalt not kill’ really was not an issue for a legionary, raising the question of the applicability of Marshall’s idea to the ancient world was not a waste of time. We learn something about both when we see the ancient world as being different from ours.

Siege warfare

Epic sieges were prominent in the classical civilizations’ views of their early history: for the Greeks the ten-year siege of Troy, and for the Romans the siege of Veii, suspiciously also thought to have lasted ten years. Despite this, Greece and Rome lagged behind the Near East in the development of siege warfare. The Persians, who invaded Greece in 480 bc, were equipped with all the techniques of siegecraft known to the ancient world, except one. That was torsion-powered artillery, which with stones as projectiles was used against walls, and with stones or bolts against personnel. The invention of this artillery is to be placed at the court of either Dionysius I the tyrant of Syracuse (405-367 bc) or Philip II of Macedon (359-336 bc). The Romans refined existing techniques, and from the principate onwards enjoyed an advantage in siege warfare over all contemporary peoples, except the Sassanid Persians.

Writing a good set piece description of a siege was one of the marks of a good historian in the ancient world. Among literary depictions, particularly important are Thucydides on Plataea (431-427 bc), Diodorus Siculus on Rhodes (305-304 bc), Julius Caesar on Alesia (52 bc), Josephus on Jerusalem (ad 70), and Ammianus Marcellinus on Amida (ad 359). From these and archaeological evidence (particularly important sites are Dura-Europos on the Euphrates and Masada in Israel), we can build up a composite picture of a siege.

To take a walled town or fortress, the attackers had to get through, over, or under its defences. Mining was employed for the latter. Subterranean tunnels were dug which could have two aims: to enable troops to emerge behind the defences, or to create a breach by undermining them. Other ways to create a breach were for the base of the walls to be weakened by sappers with pickaxes and crowbars, by battering rams, or, if available, by stone-throwing artillery. To get over the walls called for scaling ladders, movable siege towers, or earth and wood ramps. The attackers would use what missile troops and artillery they possessed to try to prevent the enemy interfering with their operations.

The defenders were seldom inactive. They might dig their own mines, either seeking to undermine a ramp, or to break into and collapse the attackers’ mine. The entombed bodies of Roman and Sassanid Persian soldiers at Dura-Europos provide eloquent testimony to the horrors of such underground fighting. Hooks or chains might be used to try to ‘catch’, and thus render ineffective, a battering ram, or ‘padding’ might be lowered in front of the wall to absorb the impact of rams or projectiles. A second, or even third, wall might be built behind where a breach threatened. The defenders would use whatever projectiles they could to hinder the attackers’ efforts. At any point the defenders might sally out from their walls to try to cause havoc among the besieging force.

Ancient stories of the wonderful, and often implausible, inventions of Archimedes defending Syracuse from the Romans (213-211 bc) show that siege warfare was at the cutting edge of ancient technology. Despite this, many defended places fell to surprise, trickery, treachery, or were starved into surrender.

Sieges were very expensive, and called for high logistic sophistication. They often lasted for a long time, during which troops usually had to be paid, and they demanded huge amounts of supplies: food and fodder for the besieging force, as well as the raw materials to construct siege works. Often these materials would have to be transported long distances. Not all supplies and transport could be requisitioned; some would have to be paid for. The siege of Potidaea (432-430 bc) absorbed two-fifths of the reserves of Athens at the height of its prosperity. Conversely, a successful siege could yield enormous amounts of booty, including slaves.

The normal rules of war did not apply in siege warfare. Various explanations can be offered of why siege fighting was so much bloodier than most land battles. One is the lack of control experienced by ancient commanders when their troops were scattered in built-up areas. Another seeks to invoke biological theories: the so-called ‘flight or fight’ mechanism released when too many aggressive humans or rats are placed in too confined a space. A third argues that the troops were attempting to reassert their control of battle, which had been subverted by engineers during the siege. Another explanation can be offered. Ancient land battles were over quickly, within a day, or at most two or three. Sieges could last for weeks, months, or even years. Throughout this time, the besiegers were in constant danger, both from sallies by the enemy, and from the clearly recognized threat of disease. The extremes of brutality employed in sacking a city may have stemmed from desires for ‘revenge’ on an enemy which was thought of as having placed the attacking troops in an extended position of fear.

Naval warfare

Consider a Greek warship (see Figure 10). This is the Olympias, a modern reconstruction of an Athenian trireme. The Olympias is a triumph of experimental archaeology, drawing on a wide variety of different disciplines, including recent developments in underwater archaeology. Yet the ship can never tell us the full potential of a trireme. The modern crews have different physiques from ancient ones, and lack the latter’s inherited skills. No one will suggest that Olympias be tested to destruction. It is not part of the project to see how nastily the crew of a trireme can drown.

Although triremes used sails, as in the picture, they were not the main form of propulsion, and were usually left ashore before a battle. A trireme was a galley with a ram, rowed by 170 men arranged in three banks. There were two approaches to fighting with triremes. The ‘light’ school centralized skilled manoeuvre, and sought to use the ram either to hole and waterlog the enemy, or to render them immobile by sheering away the oars on one side of their vessel. The distinctive tactics of the light’ school were the periplous, an outflanking manoeuvre, and the diekplous, an attempt to break through the enemy line of battle. The ‘heavy’ school sometimes reinforced the prows of their triremes and attempted ram-to-ram collisions, but more normally relied on grappling and boarding. If threatened by a fleet of superior skill, the kyklos, a close circle with rams pointing out, might be adopted.

The earliest literary mention of a trireme dates to the mid-6th century bc. From then to the late 4th century bc, triremes were the standard battleship in the Mediterranean. After which time heavier warships, concentrating on boarding over manoeuvre, dominated. These had numerical names, such as ‘fours’, ‘fives’, and so on. It is most probable that they were still rowed on three levels, but with more than one rower at all or some of the oars. The Roman and Carthaginian fleets of the Punic Wars were built around ‘fives’ (Quinquer ernes), and Roman fleets in the civil wars of the last century bc tended to employ yet bigger ships. Under the principate, with the absence of any credible naval threat, Roman fleets were largely composed of smaller vessels, including triremes. We last hear of triremes in action in a Roman civil war in ad 323 (Zosimus 2.22.2; 2.23.3-4). By the 5th century, we are told that the techniques of building triremes have been forgotten (Zosimus 5.20.3-4). When in the 6th century ad the Byzantine empire again began to build large naval forces, their standard warship, the dromon, was of different design from classical galleys.

10. A reconstruction of the trireme Olympias

The limited strategic outreach of ancient warships in part was due to their relatively poor seaworthiness, and in part to their lack of storage space. Although ancient warships could take on supplies for some days, it was normal for them to beach both at midday to take on water and for the crew to eat, and overnight for the crew to sleep ashore as well. Unlike naval sailing vessels of the 18th and 19th centuries, ancient warships could not operate for long independent of the land, and thus could not enforce a blockade on a distant shore.

Naval warfare, like siege warfare, called for vast resources and employed the most advanced technology of the ancient world, and thus it was very expensive. During the Peloponnesian War, Athens lost two large fleets in a disastrous expedition to Sicily (415-413 bc). By a supreme financial effort, Athens built a replacement battle fleet. But when this fleet was lost at Aegospotami (405 bc) Athens could not afford to rebuild again and the war was over. Their Spartan opponents were in a different position. In return for renouncing (however falsely) any intentions of liberating the Greeks of Asia Minor, the Spartans began to receive Persian financial aid (412/411 bc). The Spartans thus could suffer repeated large defeats at sea, such as Cyzicus (410 bc) and Arginusae (406 bc), yet still continue the war, as they could afford to replace the ships and men they lost.

The high costs of a fleet can be illustrated by thinking about the Athenian institution of the Trierarchy. This was a sort of wealth tax to pay for the running of warships. In theory, while Athens provided the hull, rigging, and pay and provisions for a trireme, a rich Athenian individual was appointed as Trierarch to act as captain and pay for repairs. In practice, Athens provided the pay and provisions retrospectively, only half being up front in the 5th century bc, and nothing in the 4th. Also Trierarchs, for their own glory and both physical and financial security (as they both served on the ship and would be responsible for its loss), tended to buy their own rigging and to hire skilled crews out of their own pocket. As the financial position of Athens deteriorated, it became harder to find enough individuals to be Trierarchs. During the Peloponnesian War, two men (Syntrierarchs) were appointed to share a Trierarchy, and in the 350s bc the 1,200 richest men in Athens were organized in 20 groups (Symmoriai) to bear the total trierarchic costs.

The financial costs of serving as a Trierarch are illuminated wonderfully by a speech from the 4th century bc preserved among the works of Demosthenes (Or. 50), although not composed by him. This speech was written for (and possibly by) the Athenian banker Apollodorus. As a rich man and as a new citizen wishing to make his mark, Apollodorus had lavishly equipped his trireme. Apollodorus’ patriotism was abused. Apollodorus’ trireme, as one of the best equipped, was sent by his commander on extra missions. These included a trip back to Piraeus, the port of Athens, where Apollodorus’ crew demanded more money to re-embark. In the speech Apollodorus is suing for the additional expenses incurred because his successor avoided taking over the ship for over five months. Evocative as this speech is, when reading it we should not take it as unbiased reportage, but remember that it is a forensic oration designed to persuade a jury.

It is striking that in ancient naval battle morale tended to operate at the level not of the individual rowers but the ship. As an example we can take Thucydides’ account of the Spartan defeat at Naupactus in 430 bc.

the deck-hands were shouting and taking evasive action, and abusing one another, so that they were not listening for the words of command or for the boatswains, and as the ill-trained oarsmen, being unable to clear their oars in the choppy water, made the ships unresponsive to the helm, at that precise moment... the Athenians moved in to ram.

In the confusion none of the Spartan ships fought, hut all fled.

(Thucydides 2.84, slightly abridged)

For the Spartan ships to flee, their rowers, despite being ill-trained and in physical difficulties, had not given way to individual panic, but were still operating as a team. The collapse in morale rested with the Trierarchs. We can suggest why that might have been by considering some comparative history. In an enthralling study of the British navy of the 18th century, N. A. M. Rodger found that in battle, while some individual sailors might attempt to leave their posts and hide in the hold out of immediate harm’s way, the vast majority, having a limited appreciation of the overall course of the combat, were kept too busy to give way to fear. The opposite was the case for the ships’ commanders. They had a good view of wider events, were in an exposed position, and, not being called on to perform continuous physical activity, had plenty of time to dwell on the risks that they were running, and decide that enough was enough. We can imagine that much the same applied to the commanders and crew of ancient warships. For the latter, there was not even a hold in which to try to hide.

Winter quarters: exploring a battle and leadership

A shortage of grain throughout Gaul, caused by a drought the previous summer (54 bc), led Julius Caesar to break from his normal practice and divide his army when it went into winter quarters. At Atuatuca, somewhere in the Ardennes, one division, comprised of a legion recently raised in northern Italy, another five cohorts (equivalent to half a legion), and some Spanish auxiliary cavalry, were snug in thatched wooden huts, well supplied with provisions, and protected by a ditch and rampart sited in a strong position. The local tribe, the Eburones, seemed docile. Their leaders, Ambiorix and Catuvolcus, had met the Romans at their borders, and had brought corn into the cantonment. After about 15 days, all appeared peaceful. A detachment of the troops was out gathering wood, and the rest were unarmed in the camp, when the attack came. The detachment was overrun. In the camp the legionaries ran to arm themselves, and, taking their places on the rampart, tried to fight off the assault. When the Spanish horses sallied out and got the advantage of the Gallic cavalry, the tribesmen drew back.

The Eburones asked for a parley. Ambiorix, through Roman intermediaries, offered on oath a safe conduct to the besieged force. The Gallic chief said that it had not been his wish to attack. His hand had been forced by his own people, who had been encouraged to act by a general rising throughout Gaul; all Caesar’s camps were under attack, and a large force of Germans had crossed the Rhine and would be at Atuatuca in two days. At the officers’ council of war held on receipt of this message, a deep division emerged between the two Roman commanders. Cotta argued that they should stay put. Sabinus, raising his voice so that the soldiers could hear, demanded that they should go. At midnight Cotta yielded, and it was announced that they would march at dawn. The soldiers passed the rest of the night talking and going through their possessions to decide what they would abandon in the camp.

At first light it was a straggling column of tired soldiers impeded by much baggage that marched out into the heavily wooded landscape. After about two miles the column entered a steep defile. As the vanguard tried to climb the ascent out, the ambush was sprung. Ambiorix had lied. There was no general revolt in Gaul, and no Germans rushing to aid it. He had led his tribe into the attack, and he had no intention of keeping to the safe conduct. The noise in the Roman camp had given the Eburones warning that the Romans would march that morning.

With the army assailed from all sides, the Roman generals reacted very differently. Sabinus ran about issuing ineffectual orders to post the cohorts here and there. Cotta, however, addressed the troops to encourage them, and fought like a soldier. As the length of the column made effective supervision impossible, the generals ordered their men to form a square. This manoeuvre, smacking of despair, merely encouraged the enemy, and lowered Roman morale. In a confusion of shouting and weeping, Roman soldiers were abandoning their posts in attempts to retrieve their treasured possessions from the baggage train. The Eburones obeyed their leaders and kept in rank, rather than attempting to loot the Roman baggage. Seeing that the Romans were still causing many casualties in hand-to-hand combat, Ambiorix ordered his men to fight with missiles from a distance. The cohorts that charged could not catch the Gauls, whose mobility was enhanced by the lightness of their equipment and their daily training. The charging Romans were pelted with missiles on their unshielded right side, and when they attempted to retire were surrounded and cut down. For nearly eight hours those who remained close-packed in the square endured a rain of missiles. In response to a request from Sabinus, Ambiorix said that if the general wished for a parley he would guarantee his life and would try to prevail on the tribesmen to spare the rest of the Romans. Cotta, who had been wounded in the face by a slingshot, refused to negotiate with an enemy under arms. Sabinus ordered those officers nearby to accompany him. Commanding the Roman party to cast away their arms, Ambiorix span out the negotiations until they were surrounded and killed. On the death of Sabinus, the barbarians, raising their customary shouts, charged back into close-quarter combat. Cotta died fighting.

Some of the Romans managed to fight their way back to their camp.

The legionary eagle was saved by the self-sacrifice of the standard-bearer, and the remnant held out until nightfall. Realizing that their position was hopeless, that night they slew one another to the last man. Out of the whole army only a handful survived, slipping away through the woods.

We are well informed about the disaster at Atuatuca in the winter of 54-53 bc. Several ancient sources tell of it. Yet all are derived from one text: Julius Caesar’s Commentaries (Gallic War 5.24-37). These can be described as propaganda, in the sense that Caesar seeks to convince a Roman readership of the rightness and the greatness of his actions. We cannot be sure how many legionaries, and thus Roman citizens, plus auxiliaries, died at Atuatuca. A legion at this time probably contained about 5,000 men. The main legion in this force was recently raised, so probably had not had time to reach the levels of undermanning which seem to have been common. On any estimate thousands of citizens under the ultimate command of Caesar had been massacred. Caesar had some explaining to do, and it is fascinating to see how he seeks to free himself of blame.

First, he is at pains to show how accurate is his account; he knows about the event from both sides. The few Romans who got away escaped to the camp of another deputy, Labienus, who sent Caesar a letter telling their story. Later, when Caesar captured some of the Eburones, he learned more details, and he lets us know that he has subsequently been to Atuatuca. Second, Caesar attempts to show that he had taken every possible precaution. The need for supplies had forced him to divide his army. Yet he had made sure that all the winter quarters were at no great distance from one another, and he was waiting in Gaul until he heard that all the camps were fortified and supplied. Third, Caesar shifts almost all the blame on to Sabinus (we will look below at how Caesar portrays the actions of the soldiers). The behaviour of Sabinus is marked by stupidity; if it was foolish to trust a Gaul once, it was doubly so to trust him a second time after he had broken his oath. In council Sabinus betrays signs of demagoguery in raising his voice so that the ordinary soldiers can hear him. He panics when they are ambushed. There is even a suggestion of duplicitous cowardice; asking Cotta to join him in negotiating with Ambiorix, Sabinus suppresses the fact that the Gaul has already promised to spare Sabinus’ life, but not those of the other Romans. Later in the text (5.52), Caesar tells his troops not to be downhearted because of this disaster, it was entirely the fault of Sabinus and his temeritas (rashness or perfidy). Sabinus’ dreadful performance is made evident in the text as Caesar goes on to write up at length (5.39-52) the excellent behaviour of Quintus Cicero (the famous orator’s brother) under almost identical circumstances: when his camp is attacked he turns down the overtures of the Gauls and conducts a vigorous defence until rescued by Caesar.

Apart from the drought of 54 bc being evident in the tree-ring record, archaeology provides no direct evidence for the battle. Atuatuca is often identified as Tongres in Belgium, but the scanty and imprecise topographical details furnished by Caesar allow no certainty. It was in Caesar’s interests to emphasize the natural strength of the site of the winter quarters, and the Ardennes are full of wooded defiles. In contrast, the site of a similar disaster of even greater magnitude which befell a Roman force under Varus in ad 9 in the Teutoburg Forest in Germany, outside the modern town of Kalkriese, has recently yielded a mass of evidence. Among the many items of Roman equipment found, one is especially poignant; the skeleton of a baggage mule, the large bronze bell around its neck stuffed with straw in an attempt to deaden its sound and not give away the movements of the army. If we consider what archaeologists call the formation of an archaeological site (in this case, broadly what has happened to the site between the event and its discovery/excavation), the contrasting fate of the two battlefields in the archaeological record becomes clear. Although a later Roman expedition reached the site of Varus’ defeat and buried some of the bodies, the battlefield remained outside the empire in the territory of free German tribes. It seems that the Germans left the bodies of the Romans and some of their equipment on the battlefield as a dedication to the gods. For the Germans, the site became a sacred, lasting memorial to their triumph. The aftermath was very different at Atuatuca. The Romans recaptured the area within weeks, and for centuries it remained within the province of Gallia Belgica. The Romans had reasons beyond piety for burying their dead and tidying up the battlefield.

The events at Atuatuca forcibly remind us of the importance of logistics in ancient strategy and tactics. The need to find supplies forced Caesar to divide his army, and the overlong baggage train contributed to Sabinus’ defeat. The beating back of two Gallic assaults on the camp at Atuatuca and Quintus Cicero’s successful defence of his camp under days of continuous attack by huge numbers of barbarians point to the superiority in siege warfare enjoyed from the 4th century bc onwards by the Greeks and Romans over almost all other contemporary cultures. One of the very few meaningful constants over centuries of warfare appears to be the ability of well-supplied regular troops in fortifications to defy seemingly overwhelming numbers of irregular warriors. Atuatuca shows the potential of terrain in ancient battle. On steep, wooded slopes the heavily armed Romans could not catch the more lightly equipped Gauls, who were accustomed to those conditions. Deployment on the small area of open, flat ground at the bottom of the defile packed the Romans so close together that they could not fight effectively and made them an easy target for missiles. Although equipment matters, in Caesar’s account it is morale that determines success or failure. He indicates fatigue as an important factor in undermining the will to combat. We are told that the Gauls were the equals of the Romans in courage (virtus), and it is the effect on the morale of both sides of the death of Sabinus and the officers with him that is the turning point. This testifies to the relatively ‘low-technology’ nature of ancient land battle.

That the last Romans choose mass suicide rather than try to surrender makes manifest the barbaric nature of their opponents. This is a clash of cultures between those who fight in the Western

Way of War’ and those who do not. Much in Caesar’s account of the Gauls in this battle fits the conventional template of the barbarian. They lie and, being irreligious, break their oaths. In battle they are noisy and prefer ambush and long-distance combat to open, hand-to-hand fighting. Yet Caesar also plays with his audience’s expectations. Although he stresses the collective courage of the Romans and gives us vignettes of individual heroism, here and there they behave rather like barbarians. The legionaries lack discipline, first when they chatter all night, tiring themselves and alerting the enemy, and second when, shouting and weeping, they desert their posts in battle to rescue their personal possessions. Conversely, the barbarians at times behave almost like Romans. They have been training daily, obey orders, and keep in their ranks rather than individually loot the Roman baggage. Caesar’s partial inversion of classical cultural expectations is made more acceptable to his audience by the information that the main Roman legion is composed of raw recruits; it implies that they had not yet achieved Roman disciplina. Yet in his final analysis, blame lay not with the troops but with their commander Sabinus.

One of the most striking features of Caesar’s account of Atuatuca is his judgement of the actions of the two Roman generals. Sabinus, the general of whom he disapproves, ‘ran to and fro arranging the cohorts; but even this he did nervously and in a way which showed that he was at his wits’ end - as generally happens to those who are compelled to make decisions when a battle has actually begun’. While Cotta, the general who is held to have taken the correct course, ‘did everything possible to save the army - calling upon the men and encouraging them as their commander-in-chief [Caesar himself] might have done, and fighting in the ranks like a soldier’ (5.33, tr. H. J. Edwards). A modern reader might consider that if one’s army was caught in an ambush while out of formation in a straggling column, trying to achieve some tactical order was as important as words of encouragement and adding one man to the fighting line, if not rather more so. But that is not Caesar’s judgement. This reminds us that, contrary to much that has been written on the subject, generalship is not a universally constant activity. What generals do, and are expected to do, in battle are products of their culture.

In Homer’s Iliad the leaders are heroes. They can make tactical decisions both before and during battle. Nestor gives his fellow Greek leaders advice on the drawing up of the battle line (2.362-8), and is described getting his own men in order (4.293-309). In combat Hector organizes the Trojans into five units to attack the enemy camp (12.81-7). Yet these are exceptional passages in a long poem. Usually the heroes, who are to be understood as being continuously accompanied by a personal retinue, move around the battlefield, or even completely away from it, at their discretion. They lead by example, attempting with their words to encourage the fighters on their side and depress the spirits of their foes, while seeking glory in personal combat. They belong to the category of commanders who always fight in battle, although not all through the day.

Less changed than we might have expected with the coming of the hoplite phalanx. The hoplite general was busy just before combat. He would draw up his battle line. He took part in an animal sacrifice to gauge the attitude and/or secure the favour of the gods. We emphatically should not dismiss this as a charade, either because it seems very alien to us, or because in some examples the ‘result’ was rigged. He made a speech to his troops. The reality of this practice should not be dismissed either on the grounds that the pre-battle speeches we have in literary texts are the composition of the author not the general, or because of the practical difficulties in a general making himself heard by many in a large army. The general gave the signal for the phalanx to advance. The general had been brought up on Homer, and thus the perceived need to encourage the fighters, and the desire to prove himself among the aristos (the best) caused him to fight in the front rank. It is commonly believed that this, with the absence of a clear command structure and the troops’ lack of training, in all except Spartan armies, served to preclude almost any tactical manoeuvre in hoplite battle during its first incarnation, c. 725-431 bc. The two phalanxes advanced straight towards each other. It seems strange to us that with the change from Homeric to hoplite battle, which could be thought a change from ‘primitive’ to ‘civilized’ battle, the symbolic element of the general’s role remained constant, while the practical side somewhat declined, as the ability to reorganize troops once combat had been joined lessened.

The period from roughly the Peloponnesian War, 431-404 bc, into the 4th century bc saw an increasing complexity on the Greek battlefield. Hoplites remained the mainstay of Greek polis armies, but greater roles were given to cavalry and light infantry, and more sophisticated tactics were attempted. For example, at the battle of Mantinaea in 362 bc the Thebans opened with a charge of combined cavalry and light infantry, then attacked with one wing of their phalanx, while ‘refusing’ the other. In the 4th century bc literature which explored the theory of military command appeared, with such works as Xenophon’s On the Cavalry Commander and The Education of Cyrus, and ideas were expressed which seemed to undermine the role of the general as a physical combatant. The Athenian general Iphicrates said that he was not a cavalryman, hoplite, archer, or peltast (light infantryman), but someone who knew how to direct all of them (Plutarch, Sayings of Kings and Commanders 187B). Yet attempts at manoeuvre after the initial setting out of the battle line remained rare, and prone to confusion and disaster. One reason for the Spartan defeat at the battle of Leuctra in 371 bc was the disruption brought about by trying to change formation by increasing the depth of their phalanx at the moment of contact. An expectation remained that the general would take his place at the front, and many generals seemed to have lived up to it. At Leuctra the great Theban general Epaminondas appears to have been in the front line when during the othismos, the ‘push’, he shouted ‘give me one step [forward] and we shall have the victory” (Polyaenus 2.3.3). The hoplite general even in this period seems to have been a commander who usually fought in battle, and if he did probably fought for the duration.

The kings of Macedon were the heirs to both ancient and modem Greek thinking on generalship. Alexander was said to sleep with a copy of the Iliad under his pillow, and Philip II had been a hostage in the Thebes of Epaminondas. They were also the inheritors of indigenous Macedonian thinking about warriors. It was a cultural ideal that a Macedonian should not wear the belt of an adult until he had killed a man in battle. Given the confluence of these two types of thinking, it is unsurprising that the kings took their place in the forefront of battle. They endured the consequences. Among many other wounds, Philip II lost an eye, and Alexander suffered a punctured lung. Despite leading from the front, they seem to have been more able than earlier Greek commanders to implement tactical changes during battle. At the battle of Chaeronea in 338 bc Philip II, while standing in the front line (Polyaenus 4.2.2), managed to get his phalanx to make a tactical withdrawal while actually in contact with the enemy (Diodoms 16.86.4). That such manoeuvres were possible was due to three factors. First, the Macedonian army did have an effective chain of command. Second, its troops were sufficiently well trained to put into effect new orders. Third, the kings must have had some respite from physically fighting. Either it was possible for them to temporarily withdraw through the ranks, or, as was suggested above, battle at times reverted to a ‘default state’, when the combatants temporarily drew apart.

In the 3rd century bc, towards the start of what we call the Hellenistic period in Greece and the Middle Republic in Rome, for reasons that would repay further study, a change took place in the theory and practice of generalship. Generals were no longer expected in the ordinary run of things to take their place in the front ranks, and few did so. Instead, they began to act as what are often described as “battle managers’. They moved around, usually on horseback, behind the front line, observing developments, encouraging their men, and issuing tactical orders. Only as a last resort would they now actually fight, and then in such circumstances they often turned to suicide. The changed thinking emerges clearly in the advice given by Philo of Byzantium (c. 200 bc) to a general besieging a city: keeping yourself out of range of missiles, or moving along the lines without exposing yourself, exhort the soldiers, distribute praise and honours to those who prove their courage and berate and punish the cowards: in this way all your soldiers will confront danger as well as possible.


This style of generalship continued through the late Roman Republic and into the empire. Its underlying rationale was well expressed in the 1st century ad by Onasander: ‘he can aid his army far less by fighting than he can harm it if he should be killed, since the knowledge of a general is far more important than his physical strength’ (33.1). Modern scholarship has hailed Julius Caesar as the epitome of the general as ‘battle manager’.

The modern emphasis on ‘battle management’ might be considered to give a slightly anachronistic slant to our appreciation of generalship in this period. The role of the classical general on the battlefield can be analysed via three categories: the ‘physical’, actually fighting; the ‘practical’, issuing tactical orders; and the ‘symbolic’, actions aimed at altering morale, such as the general riding calmly in the no man’s land between the armies, sending away his horse, taking off his helmet, picking up a standard, calling for weapons, and pushing men back into line, or even cutting down a man trying to flee, as well as communicating with the gods, and making formal or off-the-cuff speeches.

We have already seen that the ‘physical’ was rare, being usually an option of last resort, and then sometimes suicide was preferred. The ‘practical’, of course, was of undeniable importance, especially in the build-up to fighting. Generals were expected to make great efforts to give battle at the time and place of their choosing. The nature of the terrain was of serious import. Finding a Gallic army drawn up on a hill which was almost completely surrounded by a wide marsh, Caesar overruled the impatience of his troops and refused to give battle (Gallic War 7.18-9). The key aspect of the practical side of generalship was drawing up the battle line and issuing orders for pre-planned moves to be made when battle was joined. Seeing the enemy’s dispositions on the morning of Pharsalus, Caesar altered his own line of battle and issued instructions on which elements were to charge on what signal (Civil War 3.89). Successful impromptu tactical manoeuvres could be ordered during the fighting. At the siege of Alesia, Caesar found a good vantage point from which he sent out a body of reinforcements under Labienus to one point in the fighting, before moving to another spot from which he ordered out two successive bodies to another place in the combat, before leading a third group there in person. That crisis passed, he then ordered some troops to accompany him and others to take a different route to where he had sent Labienus (Gallic War 7.85-8). We should not automatically assume that spur of the moment tactical moves were always initiated by the general. His knowledge of the battle was limited to what he could see and hear, and what he was told. At times, others took charge. When outflanked fighting the Helvetii, the third line of the Roman army wheeled to face the new threat. It is notable that Caesar, who is not noted for playing down his own involvement, does not say that he gave the order (Gallic War 1.25). We should not take it for granted that all battles contained spontaneous tactical orders. Caesar sometimes Vanishes’ for large portions of his narratives of his own battles (e.g. Civil War 3.67-71). It was recognized that a plethora of orders could merely serve to confuse the army, as did Caesar’s at the battle of Zela (Alexandrian War 75). In some battles, the general’s role of ‘battle manager’ escaped him altogether when the troops ignored them and took matters into their own hands, as Caesar found at Thapsus in Africa (African War 80-6).

The ‘symbolic’ aspect of the general’s role was always appropriate, whether in theory or practice, and if things were going well or badly.

If we look again at the passage of Philo of Byzantium quoted above, we see that he highlights the need for the general to raise morale by moving about and speaking to his men. In his presentation of his activities in the latter stages of his victory at Pharsalus, Caesar gives one set of tactical orders but twice makes speeches to encourage the men in what they are already doing (Civil War 3.95-7). When things were going very badly indeed at the battle of the Sambre:

Caesar grabbed a shield from a soldier in the rear ranks (he had come without his own), moved into the front line, spoke to the centurions by name and cheered on the soldiers, ordering the standards to advance and the units to spread out so that they could use their swords more easily. His arrival brought hope to the soldiers and lifted their spirits, each man wanting to do his best under the eyes of his general even in such a desperate situation. So the enemies advance was checked for a while.

(Gallic War 2.25)

Future research may show that a concentration on the ‘practical’ in the role of the ancient general should not lead us to marginalize the ‘symbolic’; that the “battle management’ of the classical commander was as much, if not more, about motivating his men as about tactical finesse.

If you find an error please notify us in the comments. Thank you!