This week’s post is intended to answer a question which came up in response to the last post looking at the most common type of Mediterranean spear, which to put it simply is: what is up with the odd...
This week’s post is intended to answer a question which came up in response to the last post looking at the most common type of Mediterranean spear, which to put it simply is: what is up with the odd Roman heavy infantry kit built around a sword and two javelins (albeit two javelins of an unusually heavy type, the pilum)? How did that work and why did it work? How were the javelins, which evidently replaced the ubiquitous thrusting spear, used in combat and what did that mean for how the Romans fought?
And this is actually a pretty good question, because the Roman infantry set is, in fact, very unusual. Overwhelmingly, by far, in effectively all periods prior to the advent of gunpowder, the most common way agrarian infantry fight is with a shield and a one-handed thrusting spear. The sword figures into this system, but the sword is a backup weapon, for use if the spear breaks, not the primary weapon.
Now I do want to note something right off: this unusual Roman system is often framed as the Romans using a sword – the famed gladius Hispaniensis – over a spear. And that’s not quite right – spear infantry the Mediterranean over carried swords too, often very similar swords to the Roman gladius. The Romans haven’t replaced the spear with the gladius, they’ve replaced it with a pair of the unusual Roman heavy javelin, the pilum. So our story here is less about the gladius – we can do a post another day on the gladius – and more about the pilum.
Which of course makes it extraordinarily odd that the most military successful Mediterranean polity in antiquity – indeed, arguably ever – did not use this very standard shield-and-spear-with-backup-sword as their primary infantry kit, but instead used a sword as the primary contact weapon, supported by a pair of heavy javelins (again, called pila, sing. pilum).
Our approach to this question is going to necessarily need to proceed in stages. First, we’re going to take a brief look at what we can know about the period where the Roman tactical system we see in the Middle and Late Republic emerges (and the pilum, it seems, with it). Then we’ll discuss the pilum as a weapon, before moving to the implications that weapon has for tactics. And then finally we’ll loop back to the original question of why the Romans opted for this unusual weapon and tactical system over more typical ones.
As always, if you like what you are reading, please share it as I rely on word-of-mouth to find readers! And if you really like this and want to allow me to spend even more time studying Roman weapons (and perhaps buy some), you can support this project over at Patreon; supporters at the patres et matres conscripti level even get to vote on future topics (like our recent look at Greek and Phoenician colonization). If you want updates whenever a new post appears, you can click below for email updates or follow me on twitter (@BretDevereaux) for updates as to new posts as well as my occasional ancient history, foreign policy or military history musings, assuming there is still a Twitter by the time this post goes live. I am also on Bluesky (@bretdevereaux.bsky.social) and (less frequently) Mastodon (@bretdevereaux@historians.social).
The Emergence of the Roman Tactical System
I’ll not bury the lede here: what we can know about Roman warfare before c. 218 (264 on a good day) is frustratingly limited and the uncertainty is great. The Romans only started writing their own history relatively late – Fabius Pictor, the first Roman to do so, wrote around 210 B.C. or so – and while some of their official records reached back further than this (and while lost to us now, were available to our sources), they provided at best limited information. Archaeology can fill in more information, but as discussed before, it can only answer some questions and in any case the archaeological record of early Rome can be frustratingly patchy. And to note here at the start, I am going to lean quite heavily in this section on J. Armstrong, War and Society in Early Rome: From Warlords to Generals (2016), which I recommended in last week’s fireside.
Our information on early Roman warfare is spotty at best. We are heavily reliant on literary sources writing much later who often do not understand their own sources or lack information entirely. We have descriptions of the Roman military system under the semi-legendary king Servius Tullius (r. 578-535) – called the ‘Servian Constitution’ – by both Livy and Dionysius of Halicarnassus (Livy 1.43; Dion. Ant. Rom. 4.16), but substantial elements of that description must be wrong, given that the wealth requirements for the classes of infantry are based on a monetary standard, the sextantal as, which was only introduced in 211. Livy offers a second description of the Roman army (Livy 8.8) in c. 340 B.C., but is clearly deeply confused by his sources and the text itself appears to be at least partially corrupt (that is, the text we have doesn’t seem to perfectly reflect what Livy originally wrote).
Instead, the first fixed point at which we get a contemporary witness to the Roman army describing it in his own day actually comes quite late: Polybius, writing probably c. 146 B.C., though he places his description of the Roman army earlier, around 216 B.C., (Polyb. 6.19-56). That said, Polybius is more or less a contemporary of this system (he would have observed it in mid-second century, but places his description in 216), seems to have primary source access and his description accords well with our archaeological evidence.1 And he describes the legion you probably know: the Romans have an initial screening line of light infantry skirmshers (velites) unarmored and armed with light javelins, then three lines of heavy infantry. The first two lines (the hastati and the principes) wield a sword (the gladius Hispaniensis), a large, curved oval shield (the scutum) and two heavy javelins (pila), while the third line (the triarii) trades the pila for a one-handed thrusting spear (the hasta). Each line has 1200 men in it except for the triarii who are 600.
This is admittedly a strange configuration. It features multiple lines of battle instead of the more typical single line and the heavy infantry use a sword as the primary contact weapon instead of a thrusting spear, which is very unusual indeed. So it is a not unreasonable to ask how the Romans came to use such an unusual system – but of course that leads us back to the problem of our evidence.
Our sources think that at first the Roman army under the kings (753-509) and the early republic (509-390) worked a lot like a Greek hoplite phalanx. That seems to be both a bit right and a bit wrong. On the one hand, Greek-style military equipment does show up in elite contexts in Italy (especially artwork) in that period, suggesting that the wealthy members of Roman society might have equipped themselves a lot like a hoplite (which is what both Livy and Dionysius think in the so-called Servian Constitution mentioned above). But the kind of warfare we associate the phalanx with in Greece is civic-centered and assumes a relatively uniform phalanx of hoplites, whereas in archaic Italy what we’re seeing is more clan-based warfare (the Roman term for clan is gens, so this warfare is often termed as gens-based or ‘gentilical’) organized around aristocratic warrior bands, where perhaps only the elite have the full hoplite kit. So hoplite gear, but not a hoplite phalanx.
At the same time, the gear isn’t entirely the same either. The circular shields we see in Latium in this period often don’t have the typical strap-grip (porpax/antilabe), but instead a single central grip (like the later scutum), which is a pretty big difference that may imply a different fighting style. And while the Greek hoplite phalanx is often defined by its lack of integrated missile weapons, in Campanian tomb paintings we see warriors wearing hoplite armor, but using javelins alongside swords and thrusting spears. Still, what we might be seeing here is still so far fairly typical spear-and-shield fighting; hard to say if it was loose order or tight order (like the phalanx).
At some point two transitions need to happen here: the first to a state-organized citizen militia under the clear command of elected generals (generally Rome’s consuls) and second the shift to the unusual Roman kit and presumably unusual Roman tactics, so that we land by 216 on the army Polybius describes. Livy’s description of the legion in 340, however garbled, strongly suggests the process is well along by that point, but we must be on guard for anachronism from Livy. The traditional solution, which still retains currency among historians (though with many caveats regarding uncertainty), is to see the Gallic sack of Rome in c. 390 B.C. as the trauma which motivates both changes, although it is possible they were already in process before or continued to develop for some time afterwards.
In particular, Livy associates the introduction of pay for soldiers with the shift to an oval shield. Now we would normally be quite on our guard here – this is the same passage of Livy (8.8) which is such a mess in other regards – but we can use archaeology to date shifts in equipment and it makes sense if the shift to the oval shield (somewhat cheaper than the aspis) is coming at the same time as the introduction of pay and the shift – probably gradual, not sudden – to a more state-run military.
In fact, our evidence does a bit more than this: it suggests that the fourth century sees a comprehensive overhaul of Roman equipment.2 The preference in sword designs shifts, with Greek-style xiphe (a straight, double-edged sword) and kopis (a forward-curving, single-edged sword) replaced by early La Tène swords, seemingly quickly adopted for local production. Likewise, the Montefortino helmet, an Italian adaption of La Tène knobbed helmets (produced in both bronze and iron in the La Tène cultural sphere, but exclusively in bronze in Italy – there is a clear divergence between the types), shows up in Italy in the fifth century and by the end of the fourth is astoundingly ubiquitous and will be by far the most common Roman helmet through the first century BC.
And just as Livy might have us suppose, our evidence supports the adoption of the scutum somewhere in the fourth century too, though perhaps not as neatly and as suddenly as Livy would like.3 Now the shield shape here may actually be Italic – particularly the curved edges of the scutum and its large size, but the metal ‘butterfly’ boss at the center and the central wooden ridge (the spina) were clear borrowings from the La Tène oval shield (which was generally flat, rather than curved, but oval in shape and large).
And while we can’t know why the Romans picked these weapons to adopt, it sure does seem remarkable that evidently in the decades immediately following a – at least, according to our sources – traumatic military defeat at the hands of some Gauls, the Romans seem to have centralized their military system, expanded recruitment down the socio-economic ladder, and done so while adopting almost a complete Gallic military panoply. This also seems, by the by, to be the period where Rome commits to a system of expansion in Italy which maximizes military power, enabling broad mobilizations of large amounts of heavy infantry.
(As an aside, those who know my research may note that nearly all of the things I am interested in documenting in the third and second centuries seem to emerge in this period in the fourth. If you are wondering then why I am so focused on the third and second centuries instead of this clearly important formative period, the answer is, for lack of a better way to phrase it, uncertainty tolerance. To work on Rome before c. 264, you have to be willing to tolerate a lot of uncertainty and I tend to want to be a lot more sure than the sources for the fourth century let us be sure.)
So what about the pilum, the last piece of our equipment puzzle? And here, the earliest evidence is profoundly frustratingly unclear.4 A number of quite early objects in Italy might be pila – we’re looking, as we’ll discuss in a minute, for distinctive long-shanked javelin tips – but the first really clear cluster of such finds are in Northern Italy, what the Romans would have called Cisalpine Gaul, in the fifth and fourth centuries. The earliest pila we can point to and say, ‘Roman’ are those from Talamoaccio, generally dated to the Battle of Telamon (225). We also see examples of the weapon showing up in Etruscan contexts in the late fourth century.
So what does that evidence suggest? The weapon probably comes from the Cisalpine Gauls; it certainly doesn’t spread out into the broader La Tène material culture sphere – when we see pila in Gaul and Spain, they seem pretty clearly to be local copies of Roman models, with other javelins (small-tipped javelins in Gaul and the soliferreum in Spain) continuing to dominate in indigenous contexts. The Romans clearly have it by 225, but of course may have adopted it much earlier, perhaps in the fourth century or perhaps in the early third. Javelins are pretty central in Italic warfare throughout this period, so just about any date makes sense. Jeremy Armstrong has made a pretty strong argument,5 that a date in the mid-fourth century makes the most sense, and I think he’s probably right. That puts the adoption of the pilum right alongside the scutum, Montefortino-helmet, and La Tène swords.
For the curious, the Romans will, probably in the very late third century, change out those La Tène swords for a Spanish sword – the famed gladius Hispaneinsis – which was itself a variant of early La Tène swords. So the Romans trade an Italian variant of the early La Tène sword for a Spanish variant of the early La Tène sword (while in the actual La Tène material culture sphere, they’ve moved on to what we call the Middle La Tène sword, a longer, parallel-edged variation of the earlier design). And the best part is, if you wait long enough (well into the imperial period) the Romans are eventually going to adopt the spatha (which had been in use among Roman auxiliaries for a long time) which seems to be yet another variant of the La Tène sword (though in this case, of the late La Tène sword), which in turn becomes the ancestor of the lion’s share of the European sword tradition. So it’s basically all La Tène swords, all the way down.
The Pilum
So now we have the Roman adoption of the pilum. By c. 200, the first two ranks of Roman heavy infantry have dropped the hasta (including, ironically, the hastati who were named for it) and instead are carrying a pair of pila. At some point before the end of the first century, the last rank of Roman infantry, the triarii, will also drop the hasta for some pila (though note that Roman light infantry, the velites, do not use the pilum, but rather seven lighter javelins called the hasta velitaris).
That said the pilum is an unusual javelin. The most common form of javelin we see in the ancient Mediterranean is what I’ve taken to calling the ‘small tip, wooden haft’ javelin. It has a light wooden haft; these don’t generally survive so it is hard to be exact on how much they’d weigh, but javelin-tips have thinner sockets than spears, suggesting a narrower haft (and probably a shorter one). That’s tipped by a small iron spear-tip usually a miniature version of the local thrusting-spear shape (though often with a less pronounced mid-ridge). The whole assembly probably going to around half a kilogram, though again there’s probably a fair bit of range here (the Roman hasta velitaris, that lighter javelin for the velites, probably had a total mass of around 200-250g).
The pilum is…not like that. Instead of a narrow wooden haft, it has a thick, heavy wooden haft with estimated masses typically around 1kg instead of maybe 200-400g. And that’s topped not with a c. 50g javelin-tip, but with a tip that has a long iron shank connecting the point (which can be ‘arrowhead’ shaped or ‘bodkin’ (square-sectioned pyramidal) shaped) to the haft and massing 250-350g. Thus a pilum altogether has about the same mass as a thrusting spear, but it is very much not a thrusting spear. The whole thing tends to be around 1.25-1.5kg; this is a hefty weapon.
Via Wikipedia, Roman soldiers on the Tropaeum Traiani carrying pila. In this case, these seem to be the ‘flat tanged’ variants, as that’s the type with that distinctive triangular bulge at the top of the haft, which is where the flat tang is inserted and then secured.
The Romans end up using a few different systems to connect the iron shank to the wooden haft. The most common was actually the ‘socketed’ pilum, in which the shank (which is square in cross-section) terminates into a socket (round in cross-section) that fits on to the haft. This type is what we see in Cisalpine Gaul before the Romans, its the version of the pilum that ends up being locally copied in Spain and Transalpine Gaul in response to the Romans and it occurs in basically all periods. It’s a very simple, straight-forward sort of design, which makes it a bit funny that it is almost never how pila are shown.
Instead by far the most common depiction of pila is of the ‘flat tanged’ type, where the square-sectioned shank is hammered flat at the base, with two holes for large rivets. That is then inserted into a cut in the wooden haft and secured by the rivets. Through the first century AD, these pila are very common, so this is not some minority type, even if it is less common overall than socketed pila. It’s also distinctively Roman: this type basically only shows up in Roman contexts and so appears to be a Roman or at least Italian innovation on the design. Finally, there’s also the later spike-tanged pilum you can see up there; this comes in well after the period of adoption, so we can leave it aside for today.
In all cases, the pilum tip features a long iron shank, although length varies a lot, with some types being as long as 100cm, but others being around half this length. The shank is frequently square-sectioned, regardless of the tip or method of joining (square sections become more common over time, but are common from the get-go). As noted, there are a range of point-shapes seen. Perhaps the most striking are the square-pyramidal ‘bodkin’ points, which may have been intended to have more shield or armor penetration capability than the more traditional ‘arrowhead’ points.
How would this perform in combat? Older testing by Peter Connolly and newer testing by Tod Todeschini both seem to confirm a maximum effective range of around c. 25-30m or so; as ranged weapons go, these are relatively close-in ones. The shield and even armor penetrating potential of the weapon is considerable: once the point has punched through a shield (or armor), the long shank is thinner than the hole the point has created, allowing the weapon to continue moving with the momentum (and if thrown in a relatively high arc, the weight) of the heavy wooden haft pushing it forward. With long shanks anywhere from 50 to 100cm, that’s enough length for the weapon to potentially punch through a shield and then keep going to strike the man behind it and experiments with reconstructed versions generally seek to back up this application.
One thing of note is the question of bending; it is often asserted that the purpose of the long shank was to bend on impact, making the weapon impossible to throw back. Which is tricky because our sources don’t quite actually say this. Caesar (BGall. 1.25) comes closest to saying this, noting bent pila fixed into Gallic shields, but his point is not that the pila bent, but that they pierced multiple shields. We’ve discussed Plutarch’s probably-made-up Marian pilum before. Meanwhile, Polybius’ comment about bending javelin heads (6.22.4) is often taken to mean the pilum, but he’s clearly referring to the lighter and thinner hasta velitaris. Testing has produced variable results. Peter Connolly6 in a series of tests wasn’t able to produce a bending action. On the other hand, the recent tests by Tod Todeschini did produce bending actions on ground-impacts.
My own view is that bending was possible, but was not the purpose of the long-shank design and that most pila would not have bent dramatically on impact. Instead, the long-shank design is both shield penetrating (it can strike a man behind a shield) and shield disabling (the weapon is really heavy and quite hard to get out of a shield easily even if it doesn’t bend). In short, this is a weapon designed to disrupt shield-using infantry; with the addition of the square-sectioned point, it may also have had armor – especially mail – penetration value too.
Via Wikipedia, some modern reenactment pila, showing the long shank and heavy metal haft.
But of course the consequence of this was that this is a heavy weapon. At 1.25kg, it is probably about three times as heavy as a normal javelin and twice as heavy as the formidable Spanish soliferreum. That weight is part of its striking power against shielded targets, but also limits how many of these you can carry and what else you can carry. In short, if you are carrying a pair of pila, you really don’t have the hands (remember you have a shield too) or weight capacity to also carry a traditional thrusting spear, especially if your shield of choice is the large and relatively heavy Roman scutum. That means ammunition for this weapon is limited: you get two shots. And that has implications as we go to:
Roman Infantry Tactics
Our actual topic, at last!
There has actually been, for the past two decades, something of a debate among experts as to exactly how the pilum fits in to Roman infantry tactics and what that means for how we should understand a legion engaging.
There are a few things our sources are relatively clear on. The legion is organized into three primary heavy infantry lines (hastati, principes and triarii), with the last line at half-numbers compared to the other two. These lines are broken into units called maniples (manipuli, ‘handfuls’) of 120 men with intervals between them. Those maniples in turn are split into centuries (centuriae, ‘group of 100’) consisting of sixty men,7 one in front and one behind. That leaves an interval between the maniples, creating a kind of checkerboard formation we call a quincunx after the symbol for ‘five’ on a die. Out in front of all of this there is a diffuse screen of light infantry, the velites, who are differently equipped (no armor, small shield, seven light javelins and a sword for close-in defense). The standard way this army fights then is that each line attacks in sequence, falling back through the gaps of the line behind it if it cannot defeat the enemy. Formations other than the three-line quincunx (called the triplex acies, ‘three battle lines’) are used sometimes, but the triplex acies is standard.
The question is what does it look like when the first two lines – the hastati and the principes (who make up the main force of the army) – engages.
The traditional view is what we might term ‘volley-and-charge.’ Here the assumption is that when the hastati engage, the rear centuries move to close (or at least shrink) the gaps and then the whole line advances (six men deep) and at c. 25m or so, hurls their pila before drawing swords and charging. This was the axiomatically accepted position of basically everyone.
Then in 2000 in an article in Historia – the standard journal for arguments about the Roman army – Alexander Zhmodikov tipped over the apple cart by arguing that if we looked closely at battle narratives from the fourth to the second century what we see were much longer missile exchanges.8 Zhmodikov rejects the notion of infantry in melee contact being able to withdraw through the lines of the infantry behind them, and so assumes these ‘change outs’ must be happening during prolonged missile exchanges or not at all. He also notes that Roman battles could be long – a short engagement might run two hours and a long one more than four – and doubts, not unreasonably, that a shock action could last this long. The image that emerges, though he never quite sets out a clear ‘model battle,’ is one of long missile exchanges, punctuated by short shock contacts.
In my experience, Zhmodikov’s article continues to bounce around a fair bit among enthusiasts and reenactors, and so continues to have a lot of impact in that community, which is odd to me, because on the balance I do not think Zhmodikov has held the field.9 His core body of evidence, which are descriptions of Roman generals getting hit by javelins, is particularly weak, because of course those are, conspicuously, not Roman javelins hitting them. And it would be quite odd for Roman generals to get hit by anything that wasn’t a ranged weapon, given that Roman commanders did not fight in the front lines!
To me, the larger problem with Zhmodikov’s thesis is that the pilum is a terrible weapon for the combat he describes. It is very heavy, remarkably heavier than basically any other available model of javelins. And if we take a moment to think about what it would take to use that weapon, we can fairly quickly, I think, both explain why the Romans dropped the thrusting spear and also why Zhmodikov’s revision to their tactics is probably not right.
On the battlefield, the Roman soldier has one c. perhaps 7kg shield (the scutum), two pila (c. 1.25kg each, perhaps one somewhat heavier than the other), and a gladius sheathed at his side. He also, as a rule, has two arms, each with one hand. Evidently, his shield is going to occupy one hand and because Roman shields are horizontal-center-grip (that is, you hold the shield by a single metal bar at its center-of-mass which runs parallel to the ground), that hand is pretty busy. That said, it is possible to hold the scutum and a pilum in one hand, pinching the latter with your thumb. It’s not comfortable and you wouldn’t want to do it for very long (if for no other reason than that’s a lot of weight on that hand), but you can do it.
And then you have the other hand. On the march, that hand is going to be occupied by your two pila, an entrenching stake (if it wasn’t carried on the mule), any large tools (like the dolabra, the ubiquitous Roman military pick-axe), and probably the furca, the forked stick used to hold the Roman marching pack (though the introduction date of this device is unclear; there is no reason to associate it with Marius though). And that’s not too hard to carry either all together in your hand.10 Presumably in a battle, you can get rid of all of the non-combat gear and at least advance initially with both pila held the same way (hold at the base, balance on the shoulder, like a musket).
What you do not have space for is a thrusting spear! With a lighter, thinner javelin, you might well carry the spear and the javelin both in the right hand on the approach, shift the spear to the shield hand to throw the javelin, and then switch back to the spear. And it seems no accident that archaeologically and in artwork, we tend to see these lighter javelins, when carried by shield-and-spear infantry, show up in a one-spear, one-javelin formation. But with two pila, each probably slightly heavier than your average thrusting spear (c. 1kg or so) and with a haft every bit as thick, it would be really hard and quite awkward to add a long spear.11 I’m not saying carrying both would be impossible, but certainly that it would be awkward and awkward is not what you want in a fight.
Now we come to the problem of ‘battle pulses’ punctuated by pila exchanges: once you draw that gladius, you are down a hand, because the sword completely occupies the hand carrying it (the design of the guard and pommel of the gladius, which hug the hand to provide good leverage, essentially ensures you can’t carry anything else, even awkwardly). What that means, to me at least, is that if you are going into a close-combat fight, you are going to drop any remaining pila: you need one hand for your gladius and the other to use your scutum and you sure don’t want any extra weight there when in close-combat.
Instead, it seems to me that the only way that works, mechanically, given a Roman soldier possessed of exactly two arms, is that he advances with both pila in the right hand, and the scutum in the left. As he approaches throwing range, he transfers one pilum (we might suppose the ‘thicker’ or ‘heavier’ type reported by Polybius, Polyb. 6.23.9) to his shield arm for just a moment and then throws the first pilum. Then a quick shift to get the second pilum in his right hand – so it has only spent a few moments in the left hand with the scutum (because, again, you can do it, but its heavy and awkward) – to be discharged at close range just before the lines come together.
Which is to say, ‘volley-and-charge.’ J.F. Slavik has argued,12 that what probably covered for the change-out in battle lines was the presence of the velites. I think that makes a fair bit of sense. At the very least, an enemy would be very foolish to launch into a pell-mell rush to chase hastati who were falling back, given not only the velites but also the presence of a second line of Romans (the principes) who are waiting in good order for just such an opportunity. Meanwhile, Michael Taylor,13 has shown that some intervals between the maniples would have remained even with the ‘rear’ centuries shifted over the gaps. That seems like a weakness, but small gaps of that sort would be quite hard to exploit: if you charged into them, you’d just be putting yourself between two large bodies of Romans (with yet more Romans in front of you). But that means that the velites would always have some space to operate, throwing their lighter javelins, the hasta velitaris. And notably, where the heavy infantry has just the two heavy pila, the velites carry seven light javelins, so they could potentially keep peppering enemies throughout even a longer engagement.
So that brings us back to volley-and-charge, albeit volley-and-charge screened and perhaps supported by screening and harassment ‘fire’ from the velites. So in that context, what does the pilum do for you?
Why the Pilum?
Naturally, this part has to be a bit more speculative; no Roman source tells us why the pilum was preferred over the spear; the shift for most Roman heavy infantry in any case happens earlier than our sources provide that kind of granularity.
But I think that the pilum fits into a Roman system of combat which expects to win with infantry in the center and also which aims to disorder and disorient the enemy army.
We can start with what we might call a Roman tactical ‘theory of victory,’ that is how a Roman army expects to win a battle. In short we need to ask, “what is the infantry’s role in producing a victory here.” In Achaemenid and even earlier Assyrian armies, in open field battles, the infantry’s job seems to have been acting as a pinning force as well as creating a supporting base of fire for decisive cavalry actions and you see infantry well-suited for that role: neo-Assyrian shield-bearer and archer pairs eventually merged into the Achaemenid bow-and-spear troops. Hellenistic armies generally expect to pin an enemy with the phalanx in a shock action to enable decisive maneuvers by cavalry, exploiting either a gap or a flank; consequently the Macedonian phalanx is deep-set, with heavy pikes that allow it to push forward to create that pinning pressure without taking overwhelming casualties. Of course for hoplite phalanxes, there is the famous rightward drift that creates a proclivity towards flanking actions.
While Roman armies are often shown in popular culture employing hammer-and-anvil tactics like Hellenistic armies – and they certainly sometimes did so – that wasn’t the dominant mode of engagement. Instead, the Romans concentrated their theoretically best troops – the citizen legions – in the center of the line, flanked on either side by the alae of socii. Roman generals did not hang out with the equites on the flanks, but ‘drove’ one of the two legions in the center, which is also suggestive of where they expected the decisive action to be. In short, the Romans seem to have concluded that the shortest way to an enemy’s vulnerable rear was through their front.
And the legion is then built for that approach, hammering the enemy’s front-center with successive waves of attritional attacks in order to weak it to collapse. Compare for a moment that the ‘normal’ depth of a hoplite phalanx was usually something like 6-8 men. The famously deep Macedonian phalanx is in 16 ranks. Now the Roman fighting line is just 6 men deep, but they have three of them (the last being at half-depth), plus a light infantry screen, so each enemy file is actually facing 15 (6+6+3) heavy infantrymen; if we distribute the velites evenly, there will be six more of them for every heavy infantry file of frontage too, bringing the total to twenty-one. That’s a lot of concentrated combat power! And unlike other deep formations where perhaps only the first few ranks engage, the Roman system of successive attacks by these lines means you really will need to fight most of those guys.
The Roman legion thus aims to wear down its opponent to a point of ‘rupture’ where the line fails. And the pilum seems ideal for that, offering a lot more attritional utility than a thrusting spear would in its place.
In particular, as noted above the heavy weight and narrow point of the pilum give it shield-disabling ability, a point which has been demonstrated with modern reconstructions and is attested in the sources (Caes. BGall. 1.25). But key to that function is the greater weight of the weapon, which gives it the momentum to punch decisively through a shield and then weighs down the shield once it has done so. The long shank, expensive to make (that’s a lot of iron and a fair bit of labor), also lets the weapon potentially offend a target behind their shield. That wound might not be lethal, but it could be disabling or at least hindering. Latin has this phrase, ‘exhausted’ or ‘worn out by wounds’ (confectus vulneribus, e.g. Liv. 24.26.14, 31.17.11, Caes. BGall. 5.45); it can sometimes mean the fatal accumulation of multiple wounds, but more often describes soldiers still very much alive but incapable of effective resistance and I think the phrasing suggests less one debilitating injury so much as a steady accumulation of smaller cuts, bruises, etc., until the combatant is combat ineffective (keep in mind that javelins are not firearms; shields, armor and even clothing (e.g. Polyb. 2.30.1) can meaningfully limit the damage of an impact). Those wounds and disabled shields can then progressively add up as a single enemy battle line is forced to confront a series of Roman battle lines in sequence.
And the volley of pila before impact is disorienting. Both in a very literal sense – your enemies are being knocked around quite hard by a bunch of quite heavy javelins slamming into shields, breastplates, helmets and of course also bodies – and in the sense of disordering the formation. Combatants whose shields have suddenly been disabled need to stop to pull the pilum out, or drop the shield; in the later case, those fellows probably would rather not be in front anymore and might try to fall back. The sudden Roman onset might then exploit that disorder to inflict casualties and possibly rout the enemy line, akin to later gunpowder volley-and-charge tactics like the ‘Swedish volley‘ or the ‘highland charge‘ – a single volley of gunfire followed by a determined immediate charge.
And those various effects become quite important if you are expecting the infantry not to just hold their ground but to produce the decisive effect in the center. There’s a mentality contrast with Greek hoplites that seems notable; our sources for hoplites value most a man’s ability to stand steadfastly in place in line. But that’s not good enough for a Roman legion, which needs not just to hold its position but to advance and punch through the enemy; Roman martial values match, with the tension of virtus (aggressiveness, courage, valor) and disciplina (discipline) matching that more decisive, offensive spirit.14
Finally, I think the pilum also spoke directly to the combat environment emerging in Italy in the fourth and third century. As I’ve argued elsewhere, compared to the rest of the Mediterranean, Italy seems to have been becoming a relatively ‘high armor’ environment. We see evidence for relatively greater investment in body protection there than in Greece (indeed, Greek-style armors in Italian contexts tend to be heavier than when they show up in mainland Greece) and much more than among the typically very lightly armored warriors of Gaul or Spain. Meanwhile, the Italian shield tradition in that period is, of course, moving towards the scutum, marrying the structure of the La Tène oval shield with a generally somewhat larger, curved body-shield that offered a lot of protection. A heavy javelin with armor- and shield-penetration ability due to its narrow point, long shank and heavy weight (and thus higher momentum in impact) seems well-suited for that environment.
And all of that, I think, also explains why the Roman choice to drop the hasta for the pilum also remained unusual. Relatively few pre-gunpowder armies treated their infantry as the decisive offensive arm of the army, for both tactical and socio-cultural reasons (elites like to believe that elites – on horses – win battles), especially with such an attritional focus as the Romans seem to have.15 And of course it comes in Italy, where there is both a robust tradition of heavier armor and shields (for the pilum to disable) and a robust tradition of javelin-armed close-combat warriors going back quite a ways, as Jeremy Armstrong16 I think demonstrates well.
And the Romans don’t seem to have been the only ones. By the time we can see them, it seems fairly clear that the ‘Roman’ style of fighting is also the socii-style of fighting and we can’t be sure that these weapons and tactics were only adopted after these communities were subordinated by Rome.17 Either by imitation of the Romans or simply a response to the same pressures, it seems like other peoples in Italy responded to the same pressures Rome faced with the same shifts in weaponry and tactics, such that by the time we can see them clearly, the alae of the socii are tactically interchangeable with Roman legions.
And this attritional, multi-stage way of battle into which the pilum fit well could clearly be very effective, when paired with experienced troops in heavy armor (who could thus come off better in the extreme close-combat – sword-reach being less than spear-reach – after the volley and the charge). But it did seem to require a bit more experience and training to make this work; Michael Taylor has noted how clearly the edge seems to have come off the Roman gladius during the late second century as military demands slackened and so Romans are spending less time in the army and thus Roman forces have fewer battle-hardened veterans.
It’s also an expensive way to fight: you need pretty heavy armor (for the Romans post 225 or so, this will be mail) to stand up in the close-combat and the pilum itself requires a lot of labor to make, a lot of expensive iron and can’t necessarily be reused (because it can bend on impact, even if that isn’t a design intention), making it a very expensive ‘disposable’ weapon to equip your troops with. That cost may also explain why this style of fighting doesn’t ‘catch on’ elsewhere: the Romans seem to have been unusually tolerant – indeed, exceptionally so – of deploying lots of expensive metal equipment on their infantry.18
But with solidly experienced troops, the Romans from 264 to 101 BC bat way above average. They’re not unbeatable, of course – Hannibal proves that handily enough – but Roman armies with this system sure tend to win a lot more than they lose, while the exceptional Roman mobilization system19 ensured that the Romans had the strategic resilience to bounce back from the odd defeat here or there.