Why abandon the phalanx?

Problems from Rome

Did the hilly terrain of Italy force the Romans to abandon the hoplite phalanx? Did they even use the phalanx to begin with? In this article, we suggest “no” to both of those questions.

Joshua R. Hall and Roel Konijnendijk

On the AskHistorians subreddit, a user wanted to know why the Greeks kept the phalanx and the Romans dropped it. The assumption behind this question is that Rome initially used the hoplite phalanx in battle, but eventually abandoned it in favour of deployment in maniples, which were supposedly better suited for the hilly Italian countryside. The user pointed out that Greece is also full of hills and wondered why the Greeks had not done the same.

There is no easy answer to this question. Once you get below the surface of the evidence, it spirals into a very deep hole of historiographical scepticism, hoplite revisionism, and linguistic (un)clarity. First we need to determine if the Romans really adopted the manipular system because of hilly terrain when they started to conquer the Italian peninsula. The answer to that is: probably not. While it has been a regular assumption that they did, it is not a “known fact” nor is it a reason given by our ancient sources for the transition to the manipular legion.

What we read in the ancient sources paints a different picture. One version of this change is that the Romans learned to fight in loose order using scuta – oval shields in this period, not the “rectangular” shield of the Imperial era – when they started fighting against the Samnites (Ineditum Vaticanum [von Arnim 1892, p. 121]; Diod. Sic. 23.2; Athenaeus 6.106).

There is no mention of hilly terrain being the cause of the change. Instead, it is simply that the Samnite way of fighting was in some way superior to the old way of Roman fighting, which according to these sources was with bronze shields (aspides) and in phalanxes (which, incidentally, they supposedly learned from the Etruscans).

Here is what Diodorus says (Diod. Sic. 23.2):

in ancient times, when they [the Romans] were using rectangular shields, the Etruscans, who fought with round shields of bronze and in phalanx formation, impelled them to adopt similar arms and were in consequence defeated. Then again, when other peoples were using shields such as the Romans now use, and were fighting by maniples, they had imitated both and had overcome those who introduced the excellent models. From the Greeks they had learned siegecraft and the use of engines of war for demolishing walls, and had then forced the cities of their teachers to do their bidding. So now, should the Carthaginians compel them to learn naval warfare, they would soon see that the pupils had become superior to their teachers.

Athenaeus writes (Athenaeus 6.106):

for they [the Romans], maintaining their national customs, at the same time introduced from the nations whom they had subdued every relic of desirable practices which they found, leaving what was useless to them, so that they should never be able to regain what they had lost. Accordingly they learnt from the Greeks the use of all machines and engines for conducting sieges; and with those engines they subdued the very people of whom they had learnt them. And when the Phoenicians had made many discoveries in nautical science, the Romans availed themselves of these very discoveries to subdue them. And from the Tyrrhenians they derived the practice of the entire army advancing to battle in close phalanx; and from the Samnites they learnt the use of the shield, and from the Iberians the use of the javelin. And learning different things from different people, they improved upon them: and imitating in everything the constitution of the Lacedaemonians, they preserved it better than the Lacedaemonians themselves.

Modern commentators have guessed that it was hilly terrain that led to this change, but the ancients do not actually say this.

A change in tactics?

Other theories abound. Lawrence Keppie’s overview of the Roman army provides one example. He repeated a tradition that “the open-order fighting at which the Gauls excelled had shown up weaknesses in the Roman phalanx, and in the next half-century the army underwent substantial changes” (Keppie 1994, 19). There is some evidence of this, as we hear from Plutarch that Camillus trained his soldiers in 367 BC in a new way (Plut. Camillus 40.3-4):

Knowing that the prowess of the barbarians lay chiefly in their swords, which they plied in true barbaric fashion, and with no skill at all, in mere slashing blows at head and shoulders, he had helmets forged for most of his men which were all iron and smooth of surface, that the enemy’s swords might slip off from them or be shattered by them. He also had the long shields of his men rimmed round with bronze, since their wood could not of itself ward off the enemy’s blows. The soldiers themselves he trained to use their long javelins like spears, – to thrust them under the enemy’s swords and catch the downward strokes upon them.

This is not mentioned by Livy, however. It appears that Plutarch’s source is talking about the introduction of something that looks like a manipular style of fighting. At least, this is the case if we equate “long shields” (i.e. scuta) with this type of warfare.

We also hear something similar from Dionysius of Halicarnassus (Ant. Rom. 14.9), though in his version the Roman soldiers are given javelins rather than thrusting spears. Perhaps we should imagine something like the manipular legion – perhaps an intermediate between it and a “phalanx”? The problem is, of course, that the primary sources were written centuries after the fact.

The emergence of a professional army

The historiographical problem is made worse by the existence of yet a third version of when the scutum – and thus the manipular legion – was introduced. According to Livy (8.8), this was when the Romans introduced army pay. He dates this to the Siege of Veii (Livy 4.59.11-60.8), in ca. 396 BC, which is corroborated by Diodorus Siculus (14.16.5).

Whether or not this was when pay for the soldiers actually came in, this is a third possible point of introduction for the manipular formation, if we see the scutum as indicative of that transition.

There is a problem with this third line of thinking, though. Livy (8.8) says that the transition was from a Macedonian style phalanx (phalangites armed with sarissae, i.e. long pikes), rather than the more ancient Hellenic phalanx implied by the other traditions (and partially supported by the material evidence).

Regardless of whether Livy’s sources were right that the scutum came into play when the army began to draw a salary, there is clearly some sort of corruption in the tradition. The Macedonian phalanx, the mainstay of Hellenistic armies, did not exist in ca. 396 BC, the traditional date of the fall of Veii.

The emergence of maniples

What does all of this tell us about the adoption of maniples? The Romans probably didn’t know the true history behind the transition. As Nathan Rosenstein has remarked about this confused tradition, “it suggests that what little they did know offered a number of developments that could plausibly be identified as the transition point” (Rosenstein 2010, p. 299).

All the examples above show that there was no single tradition amongst the Romans, which makes any modern theory problematic. Even the seemingly acceptable explanations in the Ineditum Vaticanum, Diodorus Siculus, and Athenaeus are made difficult because of their use in establishing the topos of Roman military adaptability – the idea that the Romans would always quickly adopt the best way to fight, and then use it to crush its original inventors.

How much trust can we put in something that was used to “prove” the point of the cultural supremacy of the Romans? Through all of this we reach a problem when we try to use the Roman case as a starting point, or even a comparative point, to the question of “why did the Greeks maintain the phalanx for so long?”

We cannot say without a shadow of a doubt that the hilly nature of Italy had anything to do with a transition from a Roman phalanx to a Roman manipular army. The sources we have only imply that whatever style of combat the Samnites were using was superior to whatever the Romans were using in that earlier period. Therefore it does not really make sense to ask why the hilly terrain of Greece didn’t impact the Greeks the same way.

While we can look at the battles Rome fought during its conquest of Greece and say that the manipular legion was superior to the phalanx, this wasn’t necessarily the case in every encounter between the two systems. For example, Rome suffered defeats against Pyrrhus’ phalanxes in the third century. The latter was not an inherently worse system.

Did the Romans ever use the phalanx?

At this point, we should ask if the Romans ever fought in a phalanx to begin with (see Armstrong 2016, pp. 111-26). As we have seen, the Romans did think that this was the case. But what phalanx are they talking about?

Fernando Echeverría has masterfully shown that the definition of “phalanx” from a loose concept of a military unit to a specific type of heavy infantry formation took place between Homer and Xenophon.After Xenophon’s death, it also came to be applied to the dense Macedonian pike formation.So, if a Roman writing in the third century BC – the era of Fabius Pictor, the first Roman to write a “history” – thought that his ancestors fought in a phalanx, what was he envisioning? Was it the “epic phalanx” that Echeverría has described, which was simply one band of warriors in a battle line, or was it the formalized, deep, formation of Xenophon? Or was it in fact the Hellenistic phalanx that the Romans had been fighting against since Pyrrhus crossed the Adriatic in the early third century?

We cannot be sure of the answer. Even if whoever originated this tradition in the Roman zeitgeist found a phalanx in some early record in Rome, would he be able to understand it in its fourth, fifth, or sixth century BC context? Knowing the nature of Roman historiography, the answer is probably ‘no’.

Some people will look at the archaeological evidence of aspides – the large, round, shield of the hoplite – and say that this indicates that the Romans fought in phalanxes. After all, this big shield was useless in single combat, and could only be used effectively in a tight formation. But this is an old-fashioned view that took hold in the 1940s (see e.g. Lorimer 1947; Snodgrass 1965).

As scholars such as Hans van Wees (2000) and Louis Rawlings (2000) have shown, there is no need to relegate soldiers bearing aspides to the rigidness of the Xenophontic phalanx. The soldiers or warriors who bore these in early Rome could just as easily have participated mostly in small-scale raids, or even been horsemen or mounted infantry (Brouwers 2007). Fernando Echeverría is again helpful here, in his piece on technological determinism (2010).

Theoretical claims

This point about the hoplite and his shield brings us to the other assumption at the root of the original question. Just like the idea that the Romans picked up manipular tactics when they were fighting in hilly Samnium, the idea that the Greek hoplite phalanx can only work on flat ground is a theoretical claim made by modern historians. It goes back to George Grundy’s book on Thucydides (1911), in which he argued that hoplites are simply too cumbersome to function anywhere except in the open field, formed up tightly together in unbroken ranks; obstacles would break up their formations and leave them fatally vulnerable.

At first glance, there seems to be some evidence to support this idea that only the flattest ground was suitable for hoplites. For example, we have this passage where Aristotle speaks about the causes of civil strife (Aristotle, Politics 1303b.12):

For just as in wars the fording of watercourses, even quite small ones, causes the phalanxes to break up, so every difference seems to cause division.

A passage in the Hellenistic historian Polybius also suggests that even the smallest irregularity can tear a phalanx apart (18.31.5-6):

No one denies that for its employment it is indispensable to have a country flat, bare, and without such impediments as ditches, cavities, depressions, steep banks, or beds of rivers: for all such obstacles are sufficient to hinder and dislocate this particular formation.

But these passages both date to the time after Philip of Macedon introduced the pike phalanx in the mid-fourth century BC. This is a very different beast from the older hoplite phalanx: a tighter formation, with longer pikes (sarisai) and smaller shields, making individual men much more dependent on the cohesion of the whole line.

It is very likely that this is the phalanx Aristotle was talking about. It is also definitely, explicitly the phalanx that Polybius is talking about. No claims like this survive from earlier times. There are no indications that a hoplite formation was as vulnerable in broken ground as a pike formation. While earlier Greeks cared a lot about good order in the ranks at the beginning of a battle, this order was much less tight and rigid than that of the later pike phalanx. Nothing suggests that it was harder to establish where the ground was not level.

The earliest evidence that is usually trotted out to show that Greek phalanxes always fought in the plain (and therefore, presumably, couldn’t do anything else) is a speech which Herodotus puts into the mouth of the Persian noble Mardonius (Hdt. 7.9b.1):

The Greeks wage war senselessly, in their foolishness and stupidity. When they have declared war against each other, they come down to the fairest and most level ground that they can find, and fight there, so that the victors come off with great harm; I will say nothing of the defeated, for they are utterly destroyed.

But this is not a statement of fact. Nothing in the history of the wars of this time suggests that this is how the Greeks fought. Instead of taking his words at face value, we should look at what Mardonius is doing when he says this (Konijnendijk 2016). In Herodotus’ story, he is trying to persuade Xerxes to invade Greece. He is trying to make it sound easy by lying about the Greeks, calling them poor, weak and stupid. He is also trying to convince Xerxes that he will have a nice big battle in the open, which the Persians, with their good archers and cavalry, are most likely to win. This is not evidence for Greek hoplites avoiding the hills.

The only other thing that can be cited here – and undoubtedly the reason behind the theory of the emergence of Roman manipular tactics – is the unusual tactical innovations of the Greek mercenaries of the Ten Thousand in their attempt to make it out of the rugged interior of Cappadocia and Armenia in 401/400 BC.

Xenophon, who was one of the commanders of these mercenaries, tells us that their marching column kept falling into chaos every time they had to cross a bridge or go through a defile, and that they struggled to fight enemies in high places without their line disintegrating. Their answer was to split the hoplite line into smaller units called orthioi lochoi, which can be translated as simply “units in column” (as deep as they were wide) but also as “uphill units”.

Instead of locking these units together in a phalanx, they would use them to move and attack independently, so that it was easier to find ways up hills without losing cohesion, and so that they could support each other in combat. Many scholars have noted that this looks an awful lot like manipular tactics. And it was explicitly created to deal with a mountainous environment.

The story of the uphill units has been been used to generalize in both directions. On the one hand, people have used it to argue that the hoplite phalanx simply wasn’t suited to fighting in broken ground. On the other hand, people have seen it as proof that it is the challenge of hilly terrain that inspired manipular tactics.

We have already explained why the forward generalisation doesn’t work. Nothing in the Roman sources connects maniples to hilly ground. But the backward generalisation doesn’t work either, because hoplites clearly didn’t have any trouble fighting in broken ground (see especially the chapter by Louis Rawlings cited above). They typically encamped there, frequently drew up battle lines there to have a stronger defensive position, and also liked to use it to channel the enemy into bottlenecks (most famously at Thermopylai).

Classical Greek history is full of hoplites fighting in hills and mountains. In fact, most battles we hear about did not take place on a level plain. It may have been harder for hoplites to fight in the hills – as Xenophon evocatively shows, with his anecdote of him straining to run uphill in full armour (Anabasis 3.4.48) – but the high ground also brought advantages. Especially if enemy cavalry was near, hoplites preferred to stick to the hills.

Why, then, the orthioi lochoi? Well, just because something works moderately well, doesn’t mean it can’t be improved. The Ten Thousand developed various unique tactics that are never used again in Greek history. They could do this because they were a highly trained professional army fighting together for many months in exceptionally difficult circumstances. They became better organised and more tactically capable than any hoplite army before or since. If these men were able to perfect hoplite mountain fighting, it should not surprise us. But no one seems to have been able to bring those tactics back to home and teach them to normal hoplite militias.

In other words, the only clear evidence that hoplites were considered inadequate for mountain fighting applies to an exceptionally skilled and capable force. Most hoplites did not feel that hills were a problem for them. On the contrary, it was where they preferred to fight. It is better to take the many practical examples from Greek history as a guide than to allow ourselves to be distracted by false generalisations.


The question of “why did the Greeks keep using the phalanx” through to the period of Roman conquest is a good one. But it is worth remembering that the phalanxes of Cynoscephalae and Pydna were not the same as those of Leuctra or Mantinea.

In general, the armies of Greece did change the way they fought over the long stretch of time between Homer’s earliest mentions of phalanxes (as individual units) and the time of Philip V and Perseus. It is unfair to say that there was no tactical innovation here. We can actually flip the original question on its head a bit and ask “if the Romans abandoned the hoplite phalanx for the manipular legion, why did the Greeks move towards a tighter and more dependent pike phalanx?”

And the answer to that is simple: military changes are not just determined by terrain alone, and they do not follow a simple evolutionary trajectory, unless we are talking about periods of major technological shifts.