Elements of Roman Fighting
This entry was posted on April 6, 2021.
Sword or Spear?
The generally accepted view has been that the legionary was primarily a sword fighter and that his use of the pilum was restricted to the initial charge. This view has been strongly challenged by the work of Dr. Alexander Zhmodikov (Historia, 2000). He cites plentiful ancient evidence of missile warfare continuing long after the initial charge of the armies. In developing this idea he proposes that missile warfare was the dominant form of fighting and that sword fighting, though it may have been decisive, was sporadic and brief.
Some of Dr. Zhmodikov's comments in the correspondence are quoted below:
There were two main types of battle: defensive and offensive ones.
In an offensive battle, the Romans marched at the enemy, threw their pila, and then made an attempt to attack with swords. The throwing of pila probably started when the Romans were out of range: it is a well-known fact that soldiers often open fire when the enemy is too far. Many pila, as well as enemy javelins, fell short on the ground between the two opposing lines of heavy infantry. Sometimes the hail of missiles and the first attempt to attack with swords (primus impetus) turned the enemy to flight, there are such examples in the sources. If the enemy did not turn to flight at once but showed some signs of confusion or hesitation, the Romans could attack and put them to flight after a brief and fierce hand-to-hand fighting. If the enemy showed no sign of confusion, but showered the Romans with missiles and resolutely marched forward, the Romans themselves could be turned to flight. But in many cases, both sides showed some confusion and hesitation under the hail of missiles and before the approaching enemy shield wall, and so both sides slowed down and stopped some distance apart, and sluggish missile fighting started. In this kind of combat the distance between the two opposing lines was probably from 30 to 50 yards, i.e. on average longer than the average missile range. Both sides used spent missiles and suitable stones taking them from the ground, it as also possible that missiles were supplied from the rear: either the rear ranks of the front line passed their pila on to the front ranks, or missiles were brought from the rear (Caesar once mentions that unreliable Gallic allies were ordered to supply the Roman soldiers with missiles and stones during an assault on fortifications: B.G. 3.25).
In the course of the missile fighting phase, from time to time one side or another could make attempts to attack the enemy by the whole front line or by separate units, but usually failed to engage in hand-to-hand combat, and these attempts resulted only in local intensification of missile fighting and sometimes in a slight shift of the opposing lines in one or another direction: as we know from military literature of other periods, it was very difficult to force soldiers who stopped during an attack and started to fire to renew the attack by the whole front line, and separate units had a very little chance to reach the enemy line under a concentrated hail of missiles. If one such separate unit managed to reach the enemy line and hand-to-hand fighting started, it was very brief and finished in destruction of the front ranks of the unit, its rear ranks turned to flight (this probably happened with G.Crastinus and his small group in the battle of Pharsalus: Plutarch Vit Caes. 44, Vit. Pomp. 71).
When the men in the front Roman line were exhausted, the Roman commanders replaced the front line with the second line. How exactly this was done we do not know, but I like Peter Connolly's scheme of the passage of lines: we can see very similar schemes in the 19th-century infantry drill manuals (there the maneuver was to be done by platoons). The fresh line made an attempt to attack the enemy, and so on.
In defensive battles the Romans usually took a position at a slope or at the top of a hill, met the enemy with a hail of missiles (the position at a slope was considered as very favorable for the use of throwing weapons, Caesar often writes about positions 'favorable' or 'unfavorable' in this sense), forced them to stop and to fight uphill, and if the enemy was disordered, the Romans counterattacked with swords (Caes. B.G. I.25-6).
The main web page has an extensive analysis of the elements of the charge (see especially Elements of the Charge and Model of the Charge). I believe most of what is there is valid. A few points, however, should be revisited.
How far did they run during the charge?
There is no single answer since each situation would have been different. But some general parameters can be suggested.
How fast did they run?
In my main web page I said:
The 1900 Olympic 200 meter race was run at the speed of about 10.5 seconds per hundred meters. Modern sprinters in the 100 meter dash average about 4.5 steps per second, in the longer races such as the 1500 meter, they average about 3.5 paces per second. A survey of men jogging on a treadmill shows an average running rate of 2.7 paces per second with an average of 1.07m (42") per pace. The men surveyed averaged between 5’8" and 6’1" tall.
The Roman soldier was running in formation, over uneven ground, carrying arms, wearing armor, and in the process both throwing his pilum and avoiding oncoming missiles. For modeling purposes I used 3 paces of 40" per second for the charge, less than runners in even long distance races but faster than a jog.
After watching video clips of reenactors demonstrating a charge I think I overestimated the speed that they could run. In watching the clips I noted that the shields are held in front of the body to provide cover -- as would be necessary, of course. It was also important to maintain good order. These two factors would slow the speed, I believe. It seems more likely that their running pace would be somewhat slower than those joggers at the athletic club, not faster. The athletic club men ran, on average, about 6.44 miles per hour, I had the legionaries running 6.82 mph.
It is possible to walk faster than 4 mph without breaking into a jog. 5 mph is a very slow jog. A reasonable speed for the initial charge would be between 5.5 and 6 mph. Slower would keep the lines together better, faster would give more impetus to the charge. A speed of 5.75 mph will be used as an average.
The following chart shows some of the data.
|Paces per second||Inches per pace||Miles per hour|
|Quick March Cadence||2.0||30||3.41|
|Standard March Cadence||1.7||30||2.84|
|Athletic Club Joggers||2.7||42||6.44|
|Web site Running Pace||3.0||40||6.82|
|Revised Running Pace||2.8||36||5.75|
The following chart shows some additional data for each of the paces.
|Time in seconds to move 50 paces||Distance covered in feet||Distance in feet covered in 2 seconds||Time in seconds to run 60 feet|
|Quick March Cadence||25.0||125.0||10.0||12.0|
|Standard March Cadence||30.0||125.0||8.3||14.4|
|Athletic Club Joggers||18.5||175.0||18.9||6.3|
|Web site Running Pace||16.7||166.7||20.0||6.0|
|Revised Running Pace||17.8||150.0||16.9||7.1|
At the revised running pace of 5.75 miles per hour it would take about 17.8 seconds for the men to run 50 paces and they would cover 150 feet. If two armies were each charging, the combined distance covered would be about 100 yards. This is a reasonable estimate for the initial separation of the armies. At that distance Caesar's soldiers at Pharsalus would have gone about half way when they realized Pompey's men were not also charging. They would surely not have stopped all at once but could easily have come to a halt well outside pilum range -- about 60 feet.
At 5.75 miles per hour it takes about 7 seconds to run 60 feet. Again, if both sides are running then it only takes 3.5 seconds to close the distance. If the pila are thrown at around 60 feet of separation then the men have only 3.5 seconds to draw their swords and prepare for the clash. It would almost certainly take two seconds to draw the sword, leaving only 1 1/2 seconds to get ready to meet the enemy.
In fact, the pila could have been thrown at a bit greater distance if the enemy were also charging since they would run into the pilum volley. In the model of the charge on my original web page I showed the pilum actually being released at about 70 feet from the enemy. By the time the pilum landed the enemy had, in effect, run into it. If the men are running slower the increase in distance would be a bit less but still they could be thrown while the two sides were a bit more than 60 feet apart. This would allow a few extra half seconds for drawing the swords and getting ready for meeting each other.
Another reason to limit the speed is to maintain good order in the ranks. If each man ran at his top speed the line would quickly become scattered, the effect of the impact would be weakened, and the leading individuals would be exposed and vulnerable. Better if the line jogs at a pace that everyone can manage and keep in good order while doing it.
Did they run into battle or walk?
Goldsworthy seems to suggest that the Roman army marched silently into battle. He notes that this behavior, contrary to human nature which wants to get it over quickly, would have been terrifying to the enemy. However, Caesar quite clearly argues in favor of a loud cry and violent charge [De Bello Civili 3.92]. Most other authors accept the initial charge as a part of Roman tactics.
The two armies might be as far apart as 300 yards at the onset of hostilities. One or both sides would march forward to close the distance between them, then, at some point the front line would have begun a running charge. At what distance?
The narrative of the battle of Pharsalus offers some clues.
Sed nostri milites dato signo cum infestis pilis procucurrissent atque animum advertissent non concurri a Pompeianis, usu periti ac superioribus pugnis exercitati sua sponte cursum represserunt et ad medium fere spatium constiterunt, ne consumptis viribus appropinquarent, parvoque intermisso temporis spatio ac rursus renovato cursu pila miserunt celeriterque, ut erat praeceptum a Caesare, gladios strinxerunt. [De Bello Civili 3.93]
Caesar's men ran awhile, noticed that Pompey's men were standing still, stopped about half way there, rested a bit, resumed their run (renovato cursu), threw their pila (pila miserunt) and then drew their swords.
They had to have initially run far enough to have time to notice that Pompey's men were standing still, then conclude that they should not attempt to run the whole distance. In the chart above I estimated that it would take them about 7 seconds to run 60 feet. It is hard to imagine that the whole process of realization and stopping could have taken much less than that. In fact, it seems more likely that it took longer, 10 or 15 seconds. But take 7 seconds as a minimum. If they ran 60 feet, stopped, then ran another 20 feet, then threw their pila at 60 feet from Pompey's men the total distance covered would be 140 feet. If it took 15 seconds for them to realize that Pompey's men were not running and come to a stop themselves then the total distance they covered would be just over 200 feet.
In the chart above I suggested a distance of 150 feet for each side. That would have Caesar's men running a total of 300 feet. They would have planned to run about 150 feet, as they approached that distance they would have realized the situation and come to a stop. It is impossible to imagine the entire line doing this in perfect synchronization. Some centuries might have stopped before, say at 15 seconds into the run. Everyone would have had to come to a halt within 26 seconds, leaving the foremost units about 80 feet from Pompey's lines. From the time some centuries stopped themselves to the time the last centuries followed suit could have been as long as 11 seconds. This is a somewhat reasonable time frame. Any running distance shorter than this would cut things pretty close.
The Pilum Volley
The pilum volley may be one of the more confusing parts of Roman fighting. In on-line discussion groups there are quasi-serious suggestions that the rear ranks threw their pila over the heads of those in front, and even that right during the charge the entire century came to a halt and each rank rotated to the front to throw their pila. Most descriptions of battle assume that the entire first line fought with swords after the initial charge.
There is evidence that all of these ideas are incorrect.
Certainly, the pilum was used differently in different situations. If the army stood its ground, as Pompey's did at Pharsalus, then their pila would be thrown from a stationary position. But the usual course was for the army to charge vigorously and to throw their pila while running. Caesar's narrative makes this quite clear.
Which ranks threw their pila? The usual unspoken assumption is that they all did. Can this be correct? The illustration below shows the trajectory of all 8 ranks if they all threw at exactly the same time. Each rank is separated by 7 feet to allow enough space between them for the pila to be used. Even in this idealized "best case" example it is clear that only one or two ranks could reach the enemy effectively and that the rearmost ranks would actually be a danger to their own front ranks. It thus seems pretty clear that not all of the ranks could throw their pila during the charge.
How many ranks might have actually thrown their pila? The first rank, certainly. The second, probably. The third, just possibly, but not likely. Ranks 4 through 8, almost certainly not. There would be good reason for the second rank to have thrown their pila during the initial charge. The pilum volley was an important part of Roman tactics. Surely they would have maximized the number of pila thrown at this point. If two ranks could have thrown effectively, then they likely would have. Also, once contact was made the second rank would be expected to support and replace first rank soldiers as needed. In this close-in fighting a pilum would be less use than a sword. They would not wish to drop a pilum on the ground where it could cause someone to trip or turn an ankle. Therefore it would have been good for the second rank to throw their pila during the charge and have sword in hand when close-in fighting began.
Just how effective was the pilum volley? This is where battlefield casualty statistics are helpful. Gabriel and Metz, From Sumer to Rome, p 83ff, present casualty figures for Roman battles for which somewhat reliable information exists. The casualty rates for victors run from a high of 11% to a low of 1.4% with an average of 4.6%. The defeated army often suffered casualty rates of 30% or more, but these are thought to have come after the army turned and ran. While both armies stood and fought it seems that only about 5% were killed and a similar number wounded. This places a limit on the effectiveness of all weapons, including the pilum.
Since battles could last some hours one would expect the casualties to be spread out over the entire duration of the fighting, with, perhaps, a somewhat higher frequency attributed to the initial clash. If there were a total of 5% killed and 5% wounded during the entire battle it might be reasonable that 1/5 of all casualties occurred initial clash; that is, 1%.
Assume, as a starting point, that all of these initial casualties were either directly or indirectly attributable to the pilum. Consider a legion of 4,800 men arranged in a triplex acies formation with 4 cohorts in the first line and each century arranged in 10 files and 8 ranks. The legion has 240 men in the front rank of the front line. A 1% casualty rate is 48 men out of the legion's total of 4,800: 96 total casualties (killed and wounded). If two ranks threw their pila then the 96 total casualties were the result of 480 pila being thrown, a 20% effectiveness ratio. 8 out of 10 pila thrown did not cause an injury.
Further assume that 2/3 of the casualties occur to men in the front rank, the remainder to those behind them. In the opening phase of the battle, then, 64 (2/3 of 192) of the 240 men in the first rank were killed or wounded. The casualty rate for first rankers would be 27%. This is a casualty rate for that rank to have regularly suffered. 1% is probably too high a number to assign to the initial pilum volley. But the analysis does provide a starting point for trying to understand the weapon's effectiveness and to visualize how this attack phase might have looked. If the real number were 1/2% or even 1/4% the numbers can be correspondingly halved or quartered.
At the other extreme we have some evidence that a pilum volley might not have been very effective at all. Caesar claims to have lost only 300 men at Pharsalus. Pompey fielded 11 legions and probably had about 1,800 men in the first rank of the first line. If two of his ranks threw their pila then those 3,600 pila could hardly have caused more than 60 deaths and, perhaps, a similar number of injuries. 97% of the pila thrown would not have been effective at causing injury or death.
Furthermore, assuming that Pompey used a triplex acies and fully committed at least his first two lines to the fighting, then during the course of the battle at least two of Pompey's lines would have thrown all of their pila at Caesar's men. each other. Pompey's two lines would have held at least 30,000 men. Caesar's dead were listed at 300. Since he was the victor let us say that many of the wounded were saved so that overall his casualties were 1,000 killed and wounded. Some casualties were certainly caused by sword fighting; one should think at least half, leaving maybe 500 to have been caused by pila. We really do not know, but, no matter what casualty figure one projects, it is a relatively small number compared to the 30,000 pila thrown -- 1.5% to 3%.
The pilum would appear to have had less than 3 chances in 100 of actually doing any damage.
The Initial Clash
The standard way of starting a battle was for one or both sides to rush the other. Many ancient battles ended right there. Once side broke and ran away. But when that did not happen, what should we imagine did happen? Did the two armies run together and smash into each other, shield to shield? Did they intermingle, as Hollywood movies show? Or did they come to a sudden halt a few feet away and then begin cautious sword-fighting?
It seems pretty clear that they did not intermingle. But did they run into each other, banging shields and trying to knock the other person backwards or even to the ground? This would have been a extremely aggressive and high risk action. With a keener awareness of the desire for self-preservation felt by the common soldier, it would seem a bit unlikely. And, unless everyone did it, it would be nearly suicidal for just one or several men to try. It probably did not happen that way on a regular basis.
That leaves us imagining the two lines running at each other, but slowing as they come together. The more aggressive might continue forward to bang a shield into the other man. The more timid may actually have stopped a couple of feet away and only closed slowly and carefully.
Roman sword fighting is hard to imagine because no one has ever seen it. The dominant images are those Hollywood has created. Battle is depicted as a wild melee affair, men from both sides intermixed, wild swinging of blades, horsemen riding through the jumble, men being killed on all sides, and the whole affair over in a matter of minutes. No part of which would have been likely.
The Romans placed great emphasis on maintaining the ranks. They preserved their lines and formations and fought within that framework. They used their large scuta for protection, many probably hiding behind them. Their form of attack would be more like a cautious jabbing just beyond the edge of the shield than a wild overhead swing. Horsemen hardly ever rode through a group of fighting men; they might pursue and overtake infantry that was running away but would never have ridden into the middle of fighting between infantry groups. Battles were likely either over at the first rush or lasted hours. And the killing happened slowly. If as few as 1% casualties occurred at the initial charge then there were about 4% casualties spread out over a couple of hours. For a cohort of 480 men that would translate into 20 men killed over a couple of hours, about one man killed and one man wounded every six minutes.
Real fighting was a different affair. While we lack the kind of reporting from Roman authors we would like, we do have information from later ages. Sabin cites one such that probably reflects Roman fighting as well. The weapons and tactics may be different but human psychology and the desire to stay alive and in one piece are the same.:
Eye-witness accounts of real seventeenth-century pike duels do indeed suggest that they involved cautious and prolonged fencing with remarkably few direct casualties.
If it is true that 75% of men fought to not be hurt and only 25% fought to actually inflict harm on the other person, then most soldiers were not very aggressive and did not take risks. Many would have hung back just out of harm's way, taking a quick poke at the other fellow but mostly staying behind the relative safety of the scutum. It is difficult to try to model this but here are some images that may come somewhat close to representing the idea.
Most of the men are keeping a distance from the enemy, staying close behind their scuta, and sticking close together for mutual support. A couple of men are shield to shield. The red soldier in the middle is thrusting at the enemy to his right, not the one in front, as he may present a more accessible target.
Some fighting may have been more open and aggressive as depicted in this image but the safer, more conservative style in the first picture was probably more common. The second image also shows some possible participation by men of the second rank (brown).
At this point it may be well to ask what the 3rd - 8th ranks were doing all this time. If only the first two ranks had thrown their pila, then these ranks would still be armed with a pilum instead of with a sword. Were the entire formation to be shown it should look something like this:
What are these ranks doing during the fighting? Should we imagine that ranks 3-8 were present just behind the second rank or might they, perhaps, have hung back just a little? The pilum would not be a particularly useful weapon in these circumstances. It would make a poor thrusting weapon. It could not really be thrown at such short range. A little toss to get it just up over the heads of those in front would not have much force and, quite likely, the pilum itself would not rotate in the air sufficiently for the tip to point down. The men of the 3-8 ranks would not want to just drop their pila on the ground because they would be a hazard underfoot. It is hard to imagine them not closing up behind the first two ranks, but it is also hard to imagine just what they would have been doing while the first rankers were fighting.
Though not depicted in any of the illustration above, the scutum itself should be considered an offensive as well as a defensive weapon. It could be used to bang at the enemy, possibly unbalance him, even knock him down.
Whatever form it took, hand-to-hand sword fighting could never have lasted very long. Time estimates vary from less than 5 minutes to as long as 15 minutes. The actual time would depend on a lot of factors. But at the end of a relatively few minutes active hand-to-hand combat would simply come to a halt due to exhaustion on both sides.:
It is almost impossible to envisage opposing front-rankers fighting at arms length or closer for well over an hour without far greater mutual casualties. . . . It is hard to see a non-suicidal way for one side to break off from such combat in order to be relieved. [Sabin, The Mechanics of Battle in the Second Punic War]
At that point there are two things that could have happened. The fresher men of the second rank could somehow slip in-between the front rankers and take up the fight. Or, fighting stopped and a lull developed. Modern writers seem to be favoring the second scenario. The lull in the fighting is a core element to the new way of thinking about ancient fighting and provides new windows of opportunity for maneuvers such as line replacement. The lull will be discussed in the next section.
Before leaving the topic of hand-to-hand fighting I would like to offer a quotation from a correspondent who is a centurion in a reenactment club and also participates as a Roman in SCA combat. Most authorities tend to dismiss the experience of SCA and reenactors. SCA bears but little resemblance to real fighting and reenactors cannot duplicate the real thing either. However, even though their experiences have very limited validity, they are the only people around who have actually tried fighting with Roman equipment. Their experiences, if read carefully, can offer some insights. In this spirit his observations are offered.
We typically field 15 to 20 Roman soldiers at SCA wars, wars which have 150 to 1200 "fighters" on the field at a time. I have through practical experience learned a great deal about face to face combat with a determined enemy.
When we fight "wars" it is extremely competitive. We like to win. SCA wars are very physical events. Shields smashing into shields at full runs. Individuals trying to smash shield walls. Full contact blows with simulated weapons. Our gear is as authentic as practical. The Lori segmental is a superior form of armor. Easy to wear for hours and hours and easily repaired with leather thongs on the field when internal straps break due to the rigors of battle.
What I have learned from my experience is that most historians/scholars have no idea how Roman soldiers fought. Same with most reenactors. Until you put the gear on and go at it "full tilt bozo" you have no idea what it's like.
The illustration [of a soldier fighting] is seriously flawed. Any soldier opening himself up to thrust as illustrated would be open to attack from four angles (oblique left, front, oblique right and from the right) plus three weapon forms (missiles, thrusting spear and sword). By opening himself up like that he would be dead very quickly. Roman soldiers did not fence. The action is completely unnecessary. Any thrust that puts the arm beyond the edge of the scutum would risk the sword arm unnecessarily. If an opponent is just out of reach it is far better to step forward and remain protected by one's scutum. By keeping his scutum in front he optimizes his protection. Soldiers can fight with scuta touching even slightly over lapping. In the push and shove of battle all it takes is a thin opening to thrust at the exposed thigh, groin or lower abdomen of an an opponent pressing hard against the scutum of the soldier to your right. An additional thrust at an upward angle would contact an opponent's throat or face. Also, a thrust does not have to be hard or even fast if an enemy's legs and lower abdomen are un-armored. Very rarely would you attack the man in front of you. He has a shield too. You would more often attack the man obliquely to your right. He is most likely right handed and open on his left! Soldiers fought from a balanced legs bent position. Left leg forward. Right leg back. Sometimes the left knee is pressing against the scutum, the scutum often planted against the ground especially if the enemy is pressing hard. If he stands straight he is very likely to be bowled over. This is not good.
What followed the period of active close-in fighting was a period of inactivity. Sabin:
Most writers already accept that there were pauses and lulls in infantry combat during which exhausted men could rest and line replacement could occur. I would take this further, an argue that such stand-offs were the natural state of these confrontations, with disciplined troops periodically surging forward in unison at various points while warriors would launch repeated individual charges led by the bolder spirits among them. After a flurry of stabbing, fencing and shoving, one side or the other would back off again to the 'safety distance', and stability would be reimposed. A succession of such local recoils all along the line could produce the gradual pushing back to which the sources attest, and there would also be scope for the Romans to back away further to exchange lines. [Sabin: The Mechanics of Battle in the Second Punic War]
I suspect that this is the real explanation for the long drawn-out but not initially bloody nature of ancient infantry confrontations, and that a lot more time was spent throwing insults and missiles from a distance of several yards than was spent in what were actually only brief flurries of sword-to-sword or shield-to-shield contact. [Sabin: The Mechanics of Battle in the Second Punic War]
Why would parts of each line sporadically surge forward into contact? The key individuals would surely be the 'natural fighters' and junior leaders, who would encourage a concerted lunge forward to overcome the understandable reluctance among their comrades to be the first to advance into the wall of enemy blades. [Sabin: The Face of Roman Battle]
The challenge is to develop a visual model of just how this all might have looked. To do so it is helpful to break the process down into smaller elements.
After the flurry of stabbing, fencing and shoving, as Sabin describes it, had gone on for some minutes the individuals, one by one, would slow down and eventually stop fighting altogether out of sheer exhaustion. Neither front rank is much of a danger to the other, the men are simply too physically exhausted to continue.
It is at this point that one asks what would happen. Would they simply stand there nearly touching but hardly moving, panting for breath, nearly unable to lift a sword arm? Would they try to step back a few paces to create some space between them and the man they are tyring to kill? Would they try to get out of the first rank so that a fresher man from the second rank could come forward? And while whatever is happening with the men in the front rank, what should we imagine the men in the second rank doing. They have already, we suppose, thrown their pila and have swords in hand. They could enter the fray, but do they? The third rankers have not yet thrown their pila. At such close range they would have difficulty using them. The two ranks in front of them would make a direct throw risky. Lofting one lightly over the heads of the first two ranks so that it came down on the enemy just three or four yards away would be ineffective -- the pila would have no velocity behind it, have no penetrating power even if it did not land flat. The third ranker is standing less than ten feet from the enemy, he has a pilum in his hand that he cannot use effectively, what is he doing? What will he eventually do with that pilum?
The model Sabin proposes is that a lull in the actual fighting takes place. Is a lull necessary? It is certain that an individual cannot engage in hand-to-hand swordfighting for more than a few minutes. The only way for fighting to continue would be for each rank to move, in turn, to the front. One of the older authors proposed just such a scenario -- constant orderly rotation of the ranks, men moving forward to fight, back to rest in a kind of conveyor line. I am not aware of anyone today who would support such a notion. It is far too complicated, too improbable.
The alternative is the lull in the fighting. During the lull the two sides pull back from each other a bit, perhaps just a few steps at first, but enough to clearly break contact and put each other out of the immediate danger zone.
Psychological Safe Distance
At this point I would go beyond Sabin to draw the men back even further. Once it became clear that both sides were willing to break off the engagement, then I believe the leaders on one side or the other would take an additional step to pull back to what I would call the psychological safe distance. That is, a distance that allows the man to feel that he is out of immediate and imminent danger.
If this distance were not established then the front ranks would remain under incredible stress throughout the duration of the battle, several hours, perhaps. I think this would be untenable.
How far might this draw back have been? I think there could have been several physical and psychological factors that would have determined it.
Ten feet, for example, would not be enough. At that distance the other side could launch an attack in, literally, a second. A pilum could be thrown across that distance in less time. No man in the front rank could relax his vigilance for even one second without risking death.
The distance I think makes the most sense is at the edge of ordinary pilum range, about 60 feet. In the first place I would ask myself why anyone would want to take a position closer than that. If I'm a soldier and intend to break contact with the enemy for awhile I would not want to stand close enough that he could easily hit me with a thrown spear or pilum. Instead, I would take the few additional steps to be just out the the immediate danger zone.
At 60 feet the two sides would be somewhat safe from surprise attack. Pila or spears could be seen coming and ducked or caught on a shield. Charges could be anticipated and prepared for.
At this safe distance the men could relax a bit. The wounded could be tended, equipment adjusted, order re-established. Men from the rear ranks could be moved forward to fill in any holes in the front rank. Water could be distributed. The centurion could exchange information with other centurions or with his tribunes. Almost none of these activities could happen at a closer range. One can hardly imagine a soldier some ten feet from his mortal enemy calmly sheathing his gladius and setting his scutum on the ground so he can lift a skin of water to his mouth, tilt his head up and take a drink.
At 60 feet these things are possible. Furthermore, at 60 feet it would still be possible to harass the other side with missiles.
After the initial clash ranks 3 through 8 still have their pila. The lull presents an opportunity to use them. If the two sides were right at the far edge of pila range then it would be necessary for the soldier to move through the ranks to the space between armies to make his throw. This would be much in the patter of primitive warfare where the bolder fighters advance beyond the body of men to threaten and attack the enemy. If the two sides were a bit within pilum range then it might be possible to throw from just behind the first rank. This would have the advantage of shielding the thrower from the view of the enemy and of preserving the integrity of the front rank formation.
It is at this point that the ideas of Zhmodikov, Sabin and Goldsworthy all come nicely together (Zhmodikov, I should note, does not agree with the notion of lulls and stand-offs, however I believe some of his ideas are compatible with the model). These lulls or stand-offs account for most of the time spent in battle. There could be a lot of missile warfare during these periods. Looked at from that perspective, one could say that most of the fighting consisted of missile throwing. However, since missiles were probably not very effective, one could also say that the charge of the infantry with swords was the deciding element.
Re-dress the Ranks
An important part of what might have happened during the lulls would be re-dressing the ranks and files. This would surely have meant restoring order to a formation that would have become disorganized during the previous phase of close-in fighting. But it may also have meant a regular and systematic replacement of men in the front ranks with fresher men from the ranks behind. Since lulls and safe distances would give them the opportunity to do this, it is hard to see why they would not have. If this did happen, it would provide a useful function for the depth of the formation in that all 8 ranks could be directly involved in the fighting.
The quote from Sabin above describes how local charges might be mounted at the instigation of the natural fighters. The only change I would suggest is that the initiative be thought of as originating from the centurions. A lot of emphasis was placed on maintaining order. It would be more in the Roman character, I think, that the centurion should instigate and organize local attacks than that a private soldier should take this leadership role upon himself.
I would see local actions happening at the level of the century, perhaps the cohort level. Several centurions could pretty easily have coordinated their attacks. A kind of post hoc coordination would be even easier. Centurions would be aware of what the centuries on either side were doing. As soon as one started a charge it could easily trigger charges by centuries on either side. In this way local action beginning at one point could spread fairly quickly along the line.
A local charge could be mounted by the centurion simply shouting the command and the first rank following him to the attack. It could also have been a miniature version of the initial charge. This would be especially effective if there were regular replacement of the ranks during the lulls. Men still armed with their pila could be moved to the front to mount the charge. They could then incorporate another pilum volley as a part of the charge.
Re-adjustment of the Front
When one side initiates a new phase of sword fighting the other side can withdraw, hold its ground or meet the charge half way. After the interval of active sword fighting the withdrawal to new safe positions could be mutual or one sided. In this way the battle line could slowly shift in favor of the more vigorous, more confident, more aggressive side.