I just finished The Day of the Barbarians: The Battle That Led to the Fall of the Roman Empire, by Alessandro Barbero. I was slightly disappointed in it, not so much because of anything Barbero did wrong, but that there wasn’t much tactical analysis of the Battle of Adrianople (Hadrianopolis). That is not totally unexpected, however – very little has been written overall about the Battle of Adrianople. Barbero was trying to tell a story that few have even bothered to tell, let alone analyze. I’m still trying to get my own collection of accounts of this disastrous Roman defeat in 378 updated.
Adrianople is typically viewed as a seminal event – the end of the Roman Empire, the end of antiquity and the end of infantry as the dominant land fighting force, replaced by cavalry. Arguments can be made in favor of each, and, indeed, I believe that Adrianople is as good a place as any to draw the always arbitrary line to end the epoch of antiquity.
But the other two interpretations have significant holes. The Western Roman Empire continued for another century until the Emperor Romulus Augustus (Romulus Augustulus) was deposed by Odovacer in 476, though one could argue its back was broken. Furthermore, the Eastern Roman Empire continued for another thousand years, though it would become known to modern historians as the Byzantine Empire. The irony of Adrianople was that even though under this interpretation it was an event that ended the Western Roman Empire, it was the Eastern Romans who were defeated at Adrianople, yet they lived for another thousand years.
Much more problematic is the traditional interpretation that this was the end of the infantry era and the beginning of the cavalry era. In my opinion, while Adrianople was indeed a victory of cavalry, and cavalry became more prominent in medieval writings, it seems to me that the record simply does not support this interpretation. Adrianople was in many respects simply a Roman event. Nothing more, nothing less. But my interpretation requires digging back even further into history.
Before Rome, the dominant method of fighting was the phalanx. Anyone who has seen the movie 300 has a good idea of what a phalanx is, for the Greeks (particularly the Spartans) perfected this tactic, though it actually started much sooner in ancient Sumeria.
The Greeks staffed the phalanx with their famous hoplites, who were armed with a spear (usually 6-9 feet long), shield (usually circular), short sword, bronze or iron curiass, helmet with horsehair plume (or "scrubbing brushes" as I’ve always called them). The weapons these warriors had was important.
In a phalanx, you would take a formation of troops and line them up, usually 8-deep or so, but with usually as long a line as you could. The troops would then lock their shields together; that is, each hoplite would hold his shields with his left arm, covering his left side and the right side of his neighbor. There would be little, if any, space between the shields. Each solder would hold the spear in his right hand, parallel to the ground. There is considerable debate as to how the spear was held, overhand or underhand, but it would usually be held above the shield, though it could be held below if necessary. The hoplites in the front line would hunch behind their shields, leaving as little of themselves exposed as possible – remember 300.
The interlocking shields were the key to battle. If they were broken, the breach had to be filled quickly or the battle could be lost.
In battle (unless holding a good defensive position like the Spartans did in 300), the troops would lower their spears, so the front of the formation would be bristling with spearpoints, and advance in formation toward the enemy. The speed at which they would advance (fast march or slow run – remember these guys were wearing heavy armor) would depend on the situation and, more importantly, their training. Remember, each soldier depended on his neighbor to cover his right side. The guy on the far right was kinda screwed because he had no one to cover his right. So there was a tendency for that guy in particular to drift to his right so he would not be outflanked on his exposed right side. The man to his left would follow him, to keep the protection of his neighbor’s shield. This pattern would repeat down the line. Training and discipline were important to keep the phalanx going in the proper direction.
So the phalanx would advance until it met the enemy, which could be another phalanx. They would meet head on – literally. Like in 300, there could be a large crashing sound and a lot of shouting and shoving as each side tried to force the other back. The important objective for the front ranks, again, would be to hold the interlocking shield wall. Their compatriots behind them would push the lines in front and direct their spears at the enemy in front of the shield wall.
This phase of battle would usually produce few casualties, but there would always be one. At least one. Someone on the front shield wall would either get hurt or killed, or would panic. Either way, his position on the shield wall would open up.
With his position open, his nearby compatriots on the front line would be exposed. His position would have to be filled on the line. If not, the casualties from that hole in the shield wall would mount, the hole would widen and the army would panic and start feeling. That is the point at which the casualties would really start to mount – fleeing troops with no protection from the advancing victorious phalanx. The casualties would generally be somewhat limited because Greek hoplites could not pursue a great distance due to their heavy armor, but it was at that point in the battle when the bulk of the casualties would be incurred by the losing side.
The Greeks perfected the phalanx, largely trough training and armor. The armor – not represented in 300, obviously –- was a critical advantage over the Persians, who really were armed with things like wicker shields and linen cuirasses. This is one reason why the Athenians won at Marathon and why the pitifully few Spartans were able to hold out for so long at Thermopylae.
It was a good and useful tactic, for as long as went, but it had some problems, the biggest being its inflexibility and vulnerability. The front of the phalanx was tough from the front because of the interlocking shield wall, but it was vulnerable from the sides and back where they had no shields. The shield wall would lose effectiveness if the spears were not lowered to help protect it – but once the spears were lowered, it became very difficult to change the phalanx’s direction. Everyone had their spears lowered and pointed. To get all of these thousands of troops to raise their spears to change direction in a time of battle to deal with a new threat was next to impossible – only the Spartans (naturally) came up with a technique for doing so called the countermarch, where the phalanx would reverse direction in place – the front of the phalanx would become the back and vice versa. But even this had very limited utility. Usually, when a phalanx was outflanked, they could only continue advancing and hope someone else (maybe the Olympian gods) would take care of the new threat, or they would have to drop their spears and flee.
Various techniques were tried for handling this situation. The most common was the use of cavalry to cover the flanks. Cavalry is what made Philip of Macedon and his son Alexander the Great so effective, but for them cavalry was a necessity – their Macedonian phalanx carried spears called sarissas which could be in excess of 20 feet long (whatever other issues the movie Alexander had, it represented the Macedonian phalanx very accurately). Another tactic was to have a second phalanx operate behind the first. This second phalanx would hold back as the first attacked to see if possible outflanking threats presented themselves.
In spite of these drawbacks, the phalanx, especially the Greek phalanx, was very successful – until the Romans came along.
The Romans had to deal with the phalanx as they fought their innumerable wars in Italy with their neighbors. They saw that the phalanx depended on the shield wall, that holes in the shield wall would destroy the phalanx. They sought to create that hole early on. Enter the pilum.
The pilum was a type of javelin that was the primary weapon of the Roman legionary. It came in two types – a light one and a heavy one. The heavy one could double as a spear, and did so at times, for example, at the Battle of Pharsalus.
Roman legionaries were not armed a whole lot differently than their hoplite counterparts, but what was different was significant. Their shields (the scutum) was not circular, but long and rounded, designed to protect more of the body. Their helmets allowed them to see and hear, and also protected the neck. Plus, they had the pilum.
When faced with a phalanx (or really any infantry), the Romans would throw their pila at the enemy troops. The intent was to either strike and incapacitate enemy infantry, or to strip them of their shields. For when the enemy saw these incoming javelins, they would instinctively raise their shields, which would disrupt the formation by itself. But if the pilum stuck a shield, the additional weight would make the shield unwieldy and force the hoplite to either try to remove the pilum or to discard his shield. With the formation disrupted and the shield wall pierced, the Romans would advance, still carrying their own shields, and hack at their unprotected foes with their own swords (the gladius).
It worked pretty well, but it too had drawbacks. It was at its best fighting with other infantry. The great Roman defeats evidence a certain pattern:
Cannae (216 BC) – Roman troops make a frontal assault on Spanish and Celtic troops under the command of the Carthaginian general Hannibal Barca. The Romans drive the Carthaginian back, past the Libyan troops Hannibal deployed on the flanks of his infantry. They turn and attack the Romans on their flanks. Meanwhile, Hannibal’s cavalry, deployed outside his Libyan infantry, had driven off their Roman opposites and appear in the rear of the Roman infantry. The Romans are encircled and eight legions are wiped out.
Carrhae (53 BC) – The Roman triumvir Marcus Licinius Crassus Dives decides to invade Parthia in his own private war to show himself the martial equal of Julius Caesar and Gnaeus Pompeius Magnus (Pompey the Great). Crassus marches into what is now Iraq with seven legions and 4,000 cavalry. Crassus wanted war, and when he found the Parthians, he got one. Crassus deployed his troops into an open square, which the Parthians proceeded to surround with medieval-style armored knights and cavalry archers. The cavalry archers poured a shower of arrows – extremely well designed arrows shot from compound bows that could pierce armor and shields – into the Roman square. The Romans were slaughtered.
Teutoburger Wald (Teutoburg Forest, 9 AD) – A staff member named Arminius plots a route for three Roman legions to take through the German forest. But Arminius is really the German Hermann, who is in contact with the German barbarian tribes. The route he picks leads the legions through narrow forest trails where they will be extremely vulnerable to ambush and where they would have no chance to deploy to defend themselves. (By “deployed,” I mean set up in a battle formation, as opposed to a traveling formation like the Romans were doing here, and expecting battle.) Again, the Romans are slaughtered.
Adrianople (Hadrianopolis, 378 AD) – Several large Gothic tribes, fleeing from the Huns, petition to be allowed to cross the Danube and enter the Empire. The Emperor Valens allows it, but seriously underestimates the number of refugees, which Barbero estimates at 100,000-200,000 people. The Romans are not prepared to house, clothe and feed the new arrivals, and what preparations they have made are nullified by corrupt local bureaucrats. The security arrangements are also inadequate, and far more Goths are able to cross than were intended, a good percentage of them without the knowledge of the Romans. The Goths are naturally not happy at the treatment they have received at Roman hands, and a failed attempt to assassinate their leaders at Marcianople – apparently without the knowledge of Valens – sparks a massive revolt. The Goths coalesce under the leadership of the chieftain Fritigern and ravage Thrace. Valens is forced to return from a campaign against Persia to deal with the crisis.
He leads an army of 15,000-20,00 legionaries (by modern estimates) out of Hadrianpolis to deal with Fritigern, who, again by modern estimates, may actually have been slightly outnumbered in terms of fighting men. Late in the afternoon of August 9, they arrive at Fritigern’s wagon laager, a large defensive formation where the Goths’ wagons are arranged in a circle, much like the Old West. Both Valens and Fritigern agree to negotiations; the historians do not agree whether this was legitimate or simply to buy time to rest their troops.
In any event, the cavalry and light troops on the Roman right begin skirmishing – basically harassing – the Gothic troops. But they get too close to the Goths, who chose to resist their tormentors, and battle starts almost accidentally. Some of the accounts claim that the Roman army was still attempting to deploy for battle, but Barbero does not share this view. In any event, the Gothic cavalry, which had been out foraging, returns – and sweep around the laager. The Roman cavalry on the right collapses. The Roman left flank cavalry holds out heroically, but eventually is defeated and flees. The Gothic infantry then sortie from the laager. The Roman infantry, with its flanks exposed and facing enemy infantry in the front, is surrounded and destroyed, both by rampaging Gothic cavalry and cavalry archers, and accompanying infantry. Valens is killed; legend has it he and his staff hid in a house which was then set afire by the Goths, who didn’t know he was inside. His body was never recovered.
From these four absolutely devastating defeats, a few patterns emerge.
1. Enemy cavalry was a decisive factor against the Romans.
2. Roman cavalry was comparatively ineffective.
The obvious exception here in this sense is Teutoburger Wald, which was not a proper battle by ancient standards but is more properly styled as a series of ambushes enabled by treason. Barbero even points out the ineffectiveness of the Romans against cavalry, and uses Cannae and Carrhae as examples.
The weakness of the Roman legions to cavalry, it seems, was due to the legionaries’ lack of a proper spear. Horses can be devastating charging infantry, unless those infantry are somehow protected by large, pointy things. Large, pointy things can have fatal effects on horses.
But the early Romans faced mostly infantry in their initial expansion phase in central Italy. In the Punic Wars, the Carthaginians seemed wedded to the primacy of the elephant, even though the Carthaginian unit the Romans most feared was their Numidian cavalry. Among the early legionaries, only the veteran third-line infantry (
triarii) carried spears, while the first (
hastati) and second (
principes) carried the pilum.
After Marius’ reforms, in which the distinctions between the hastati, principes and triarii were abolished, all legionaries carried the pilum. The heavy version of the pilum could be used as a spear – and, again, was used as such by Julius Caesar’s so-called forth line against Pompey at Pharsalus – but was not really large enough to be consistently effective as a large, pointy thing.
Roman cavalry had its own issues. Cavalry was usually not a priority for a power such as early Rome, due in large part to its impracticality. Throughout history, possession of horses was more of a social status than anything else, because horses were (and are) expensive both to acquire and to maintain. Cavalry, for which horses have to be specially bred, trained, maintained and equipped, was much more so. It is much easier for a nomadic tribe living off the resources of others (say, the Huns or the Mongols) to maintain cavalry than it is for a civilized state to do so. A landlocked power like early Rome, which didn’t face a lot of cavalry, didn’t see much of a need for it. And the Romans never seemed to like the idea of cavalry much either, treating it as a necessary evil, much like a navy, rather than taking real pride in it.
Maybe that explains why the Romans were never really good at the cavalry thing. Notice that the major example of the pilum being effective as a spear was against Pompey’s Roman cavalry. In the big Roman victory involving cavalry, against Hannibal at Zama in 202 BC, much of the cavalry on the Roman side was not actually Roman, but was Numidian, under the command of the talented Numidian prince Masinissa, who had formerly fought on the Carthaginian side but had changed for political reasons.
Yet, with all their battle experience, you would think the Romans would have learned their lesson by 378 AD. And they did, perhaps too much so.
The Roman army that faced Fritigern at Adrianople was not the same as the legions of Caesar, just as the legions of Caesar had not been the same as those of Scipio. Roman infantry in 378 were basically divided into two types of units – the
limitanei, whose job it was to guard the border and serve as a sort of trip wire; and the
comitatenses, which served in the Empire’s mobile field armies. The limitanei was to slow down invaders until the comitatenses could come in and finish them off. This had the additional benefit of keeping Roman troops scattered, except for the mobile field armies under the Emperor’s direct control, which could inhibit the declaration of every General Maximus, Illiterate Thracian Giant (with apologies to Larry Gonick) as
Imperator to challenge for the throne.
The Roman infantry was armed pretty much the same – as their predecessors in the Greek phalanx. The comitatenses had newer chain mail armor (cheaper and more comfortable) and retained something of the distinctive Roman helmet, but they had circular shields and spears, not the pilum of old. At least they could deal with cavalry now. Yet in some sense it could be argued this was a regression.
Their cavalry had improved, too, or so they thought. Their heavy cavalry was based on the Parthian model, and looked largely like medieval knights, no mean feat without stirrups. They also had the heavily armored cataphract horse archers.
Their training and discipline was up to the standards of any previous Roman army. They were the best equipped army in the world.
And yet it all fell apart at Adrianople. Why?
Historians have debated this question for almost two thousand years. A myriad of contributing factors could be cited, many of which have significant weaknesses:
1. Valens, unpopular at home as an Arian Christian, needed a military victory to shore up his popular support, and so chose not to wait for reinforcements from the Western Roman Emperor Gratian, yet Valens’ army on hand should have been enough to do the job;
2. The Romans had marched all day from Hadrianopolis, and by the start of the battle late in the afternoon they were tired, hungry and thirsty, yet their morale remained very high;
3. The Romans had underestimated the size and nature of the Gothic army, because the cavalry was away foraging, but they should have had an idea of the size of Fritigern’s army and the types of units they would be facing because of information gathered form their previous engagements with his forces;
4. The Roman troops were angry about the devastation the Goths had wrought across Thrace, which caused them to be overly aggressive in their skirmishing, yet skirmishing had always been a part of ancient warfare; and
5. The Gothic cavalry counterattacked before Valens’ army was fully deployed, yet Barbero refutes this argument.
As one might expect, it was probably a combination of all of these factors. Barbero’s apparent belief that the Romans were fully deployed at Adrianople is curious. Infantry armed with spears, even when surrounded by cavalry, can beat them off if properly deployed. Barbero states that the Romans were ineffective in resisting the Gothic cavalry because most of their spears were broken by that point in the battle, which suggests to me that the army was not yet fully deployed when they were engaged with the cavalry. But given the amount of time it took for negotiations to
almost get under way, it’s also hard to believe that they were not yet fully deployed.
It’s hard to determine precisely what happened at Adrianople because, like Zama, the exact site of the battle has never been identified. Unlike Zama, there are not enough accounts of the battle to help fill in the gaps.
In any event, history has been correct in determining its effect on the Roman Empire, even in the rather ironic twist that this defeat of an eastern army led to the fall of the western empire. As Barbero relates it, Valens’ successor Throdosius ultimately ended up buying off the Goths and other barbarians and adding them to the imperial armies. But the barbarians, though normally loyal to Rome, were neither popular nor trusted in Constantinople, and the Eastern Roman government gradually sought to push them further away – that is, west. The Western Roman Empire became dependent on them for defense, but politics and a lack of money eroded the loyalty of the barbarians further. It should be noted that Alaric, the barbarian who sacked Rome, was actually a general who answered to Rome. Odovacer wasn’t much different.
But the traditional interpretation of Adrianople as the emergence of cavalry as the dominant weapon is at best a stretch and, I would argue, inaccurate. A plethora of factors made this particular event unique to the Romans, not the least of which was their historic inability to really get a grip on this cavalry thing. The rampages of the cavalry-based Huns and, later, the Mongols certainly adds credibility to the cavalry dominance argument. But their success was based mostly on sheer numbers of cavalry, not the nature of cavalry. Remember, it is much easier and cheaper to develop and maintain cavalry when you’re using someone else’s resources, as nomadic peoples tend to do.
Cavalry did not magically become more effective with Adrianople. They had always had their effectiveness, most exemplified by the Achaemenid Persians, the Macedonians of Philip and Alexander the Great and Hannibal’s (and Masinissa’s Numidian) cavalry. And after Adrianople, cavalry would suffer major defeats and serious bouts of ineffectiveness as they always had – Charles Martel’s troops, though surrounded by Muslim cavalry, were able to defeat them at Tours (Poitiers) in 732; the Norman cavalry of William the conqueror was actually ineffective against Saxon infantry at the Battle of Hastings in 1066; and heavily-armored French cavalry were completely unable to deal with a few English long bowmen at Agincourt in 1415 (then again, the French always had cavalry issues, if the Battle of Nicopolis in 1396 is any indication). They would continue to have their good points and bad points, but the cost of cavalry meant that infantry would remain the dominant land force.
So don’t make more of Adrianople than it was – a catastrophe that destroyed the mystique of the Roman army and sent the Roman Empire on its death spiral that resulted in the end of the western empire. It represented the end of antiquity.
But it was nothing more than that.
