The bow (ehrmm...CRESCENT) at Cannae

[Jim Webster]

Many years ago I had "Hannibal: Enemy of Rome by Leonard Cottrell"
I wore it out and last saw it perhaps thirty five years ago.
But vague memory tells me that he had the spanish and gauls in two lines with only the front line pulled forward.

BUT having read Livy and Polybius, be damned if I can work out where he gets it from 

Jim

Thanks for that 

I still have my teenage copy of Enemy of Rome (Pan paperback with copper cover).  Cotterell does show two lines of both Romans and Carthaginians, but closer inspection shows the front lines are skirmishers.  The crescent is there with a single line behind the skirmishers - nothing behind it.

[aligern]

It would be illuminating to see the unit divisions on the diagrams.
How do the Libyans manage to lap around? What structure do tge Africans  have that allows all that bending?
How do the  Roman legions operate as units? ? There is a key part in the Polybian story where large numbers of Romans appear  to be concentrating in  the centre, in order to exploit the breakthrough and this is what exposes their flanks.  If that indeed occurs than what is your concept  of what is occurring? Are the Libyans facing Romans in acies and those Romans withdraw?  An alternative , that the Libyans are in column , hidden behind Spaniards and Gauls and that they march around  the flank  and make a half turn to face in on the flank makes sense in the shape of the battle and has the benefit of inspiring Scipio's later manoeuvre. If the Libyans are initially engaged frontally tgan how do they become a flank attack?

Whilst I think that we have to take account of the Latin and Greek of Livy and Polybius and it would be wrong to move too far away from the sources  we do have to be able to describe the  battle  in a way that fits what could  happen on the ground as long as it is  in line with the major points of the sources.
So, it is, for me, perfectly legitimate to  have either a deeper Gallic formation in the centre (because it will be fighting longer than  the flanks of the crescent) or to have a second line, if the greater point is that relatively few Romans break out and unless the Carthaginian Celts maintain a front for most of the battle then large numbers of Romans should escape. Likewise it is distinctly possible that the  fighting in the centre is protracted  for either of the  above reasons because the story of the  battle is not maintained if the Gauls break early. If they do break quickly then there has to be an explanation  to why  the Romans do not largely escape.  As I believe  that in history the biggest cause if a current decision  is what happened to the decision maker most recently then I am very willing that  prevention of another Trebia is the motivation behind the deployment Hannibal orders. 
It is important to have the best translation of the Latin or the Greek, but we should not be too beholden to the exact words if they do not give us a coherent account of the battle.  For example, if the Bayeux embroidery was our only account of Hastings then we might find someone  insisting that William only had cavalry and archers , that no Norman footsoldiers were shown, so they were not there, probably left to guard the camp. A source may  not be everything.
Roy

[RichT]

Justin

Feel free to demangle. I'm always happy to be proven wrong.

That's a lot of work. Sorry, I know that's not very helpful, but I'd just say that if you are not sure of your Latin skills your best bet is to follow the translations of those who are. If you are sure of your Latin skills, then perhaps you shouldn't be. 

It is often possible by teasing out the meaning and usage of a particular word to do better than the existing translation on technical points, but how likely is it that every translator has got the whole sense wrong, and now you have come along and got it right? The translations on Perseus (which I assume you are using) have strengths and weaknesses as all translations do, but they all render the Latin accurately (AFAIK).

Roy

Whilst I think that we have to take account of the Latin and Greek of Livy and Polybius and it would be wrong to move too far away from the sources  we do have to be able to describe the  battle  in a way that fits what could  happen on the ground as long as it is  in line with the major points of the sources.
...
It is important to have the best translation of the Latin or the Greek, but we should not be too beholden to the exact words if they do not give us a coherent account of the battle.

I don't utterly disagree, but the problem is, how do we judge what is 'a coherent account of the battle'? Your judgement, mine, and Justin's will all be very different. I think Polybius' account is coherent, though very high level. Who are we to say that we have a better idea of what was possible in an ancient battle than Polybius did? If the problem to solve is why the Romans didn't escape to their front, then I think other explanations than a second Celtic line are more likely (for example, the explanation Polybius gives). Besides, my problem with Justin's theory is not so much that he is going beyond the sources, but that he is altering the sources to fit his theory.

Edit to add:

Incidentally since I've been forum diving and looking at our old chum, the KTB thread, I notice that the start of that thread was one of my earliest posts on this forum, on a remarkably similar subject:

http://soa.org.uk/sm/index.php?topic=1508.msg16100#msg16100

I was making the same points with the same words to the same people (except Patrick now sadly), five and a half years ago. 

[Justin Swanton]

It would be illuminating to see the unit divisions on the diagrams.

How do the Libyans manage to lap around? What structure do tge Africans  have that allows all that bending?

How do the  Roman legions operate as units? ? There is a key part in the Polybian story where large numbers of Romans appear  to be concentrating in  the centre, in order to exploit the breakthrough and this is what exposes their flanks.  If that indeed occurs than what is your concept  of what is occurring? Are the Libyans facing Romans in acies and those Romans withdraw?  An alternative , that the Libyans are in column , hidden behind Spaniards and Gauls and that they march around  the flank  and make a half turn to face in on the flank makes sense in the shape of the battle and has the benefit of inspiring Scipio's later manoeuvre. If the Libyans are initially engaged frontally tgan how do they become a flank attack?

I suspect that my last diagram is a bit inaccurate, in that the Libyans don't actually contact the ends of the Roman infantry, but just move in battle column around to the rear. Here's a plausible reconstruction of the wrap-around using the kinds of formations and manoeuvring described by the tacticians. The Libyans are organised into pseudo-centuries on the Roman model (they are equipped in Roman gear which suggests they fight in the Roman manner). Each Libyan line has 20 centuries, each century 10 x 10 men. The width to depth ratio in the diagrams below is accurate. Click on them to get the zoomed version.



As the Romans contact the first Libyan line, the centuries of each second line wheel outwards to face away from the centre, forming a battle column 10 men wide.



Each column, commencing with the lead century, wheels 90 degrees and marches past the ends of the Roman line.



The lead centuries then wheel again and march into the Roman rear.



When it reaches the edge of the Roman line, the lead century wheels to face the Romans and advances 10 yards. This permits the following centuries to march past and wheel into line adjacent to it.



The remaining centuries wheel into line next to the others and the reconstituted Libyan line charges the Romans.



This does leave gaps at the ends through which some Romans could and did escape. As you can see the Gauls really do have to hold the line otherwise a lot of Romans will get away.

[Jim Webster]

A nice parallel account of livy and polybius

https://www.johndclare.net/AncientHistory/Hannibal_Sources6.html

What struck me, from the text I'd say that there is no reason for the legionaries to meet the African infantry frontally.

Has anybody worked out relative frontage of legionaries and celts?

[aligern]

I tend to agree Jim, The nature of the terrain at Cannae is that it is very flat.  As I said before the Roman commanders cannot see over the line of Celts and Spaniards. Hannibal can keep his Africans hidden until the Romans are fully engaged and thus pinned.
The Celts and Iberians in their crescentiform line are designed to be pushed back slowly.  Whatever Polybius might seem to say I see Hannibal as managing time brilliantly, holding the Romans in the centre until their flanks are exposed and then  releasing the Africans who have not fought until deployed to the flank. The joy of Hannibal's manoeuvre is that  the Romans are funnelled towards the centre, which is much less less likely if they are engaged with the Libyans frontally because it would be very difficult to converge centrally if your formation was already engaged to the front with such tough fighters as the Libyans.
Rereading Jim's excellently cted parallel account it seems clear that the Libyans are nit engaged til they are turned n to the flanks and that there are reserve units in the centre or at the very least the Celts and Spaniards are sufficiently revovered to return to the fray in rder to box the Romans in. Either  way it does not argue for a severe fracturing if the Celtic linebecause it is unlikely that troops who have been thus broken can be brought back.  Actually its rather like Magnesia and Argentoratum  and indeed Mons Lactarius    where terms are used that are translated as severe fragmentation, but then the line comes back into action. Perhaps we could reconcile this with a very localised penetration of the line  awe might then ask why the troops that greak through do not turn and take their opponents in the rear.?
Roy

[Justin Swanton]

A nice parallel account of livy and polybius

https://www.johndclare.net/AncientHistory/Hannibal_Sources6.html

What struck me, from the text I'd say that there is no reason for the legionaries to meet the African infantry frontally.

We then still need to explain how "the Romans were caught between two hostile lines of Libyans" - Polybius: 3.115. Livy makes it clear that the Libyans moved around to the Romans' rear: "the Africans had already begun a flanking movement on either side, and as the Romans rushed incautiously in between, they enveloped them, and presently, extending their wings, crescent-wise, even closed in on their rear." - Livy: 22.47. Combined with Polybius that means that there were some Libyans on the Romans' front.

Has anybody worked out relative frontage of legionaries and celts?

I had a crack at it earlier in the thread.

[aligern]

Wouldn't the description still stand if  the Libyans were on the left and right flanks of the Roman mass? They would still be between two hostile lines of Libyans.

As to the rear I have a sneaking feeling that the diagrams thatyou so helpfully created show the Roman  line as too shallow. Crudely, if the Carthaginian line is 16000 men eight deep then its 16 yards deep and 2000 yards wide. Rome has  64,000 men on this frontage so its force is 64 yards deep.
Please correct my maths  if necessary .
The Africans have 4000 men a unit ( that's a legion isn't it) 500 men wide and  thus 500 yards wide, plenty of frontage to hit the flank and lap around.
Roy

[Justin Swanton]

Wouldn't the description still stand if  the Libyans were on the left and right flanks of the Roman mass? They would still be between two hostile lines of Libyans.

There's an interesting idea there. The Libyans all advance in column down the flanks of the Romans and then form line facing inwards. The two short thick lines then squeeze the Romans from the ends inwards, breaking up their structure and compressing them.







Problem is that the Libyans then don't wrap around the Roman rear as described by Livy.

As to the rear I have a sneaking feeling that the diagrams thatyou so helpfully created show the Roman  line as too shallow. Crudely, if the Carthaginian line is 16000 men eight deep then its 16 yards deep and 2000 yards wide. Rome has  64,000 men on this frontage so its force is 64 yards deep.
Please correct my maths  if necessary .
The Africans have 4000 men a unit ( that's a legion isn't it) 500 men wide and  thus 500 yards wide, plenty of frontage to hit the flank and lap around.
Roy

My reconstruction* is based on the notion that a Roman legion was normally about 200 yards wide, each hastatus and princeps line consisting of 10 maniples which themselves consisted of 2 centuries that deployed 10 men wide and 8 men deep once the velites were withdrawn (the velites forming the last 2 ranks of the hastati, principes and triarii). There are no maniple-sized gaps between the maniples, the gaps being the file gaps between files deployed in open order and doubling to intermediate order once the line in front withdrew through them. On that assumption (no way of going into the arguments for it here) to create a speira many times deeper than wide according to Polybius, the simplest way is to deploy the centuries 5 men wide and 16 men deep. Each legion will now occupy a frontage 100 yards wide. This is a double-double consular army of 16 legions in total, so its frontage will be 1,600 yards. The hastati will be 16 men deep, the principes 16 and the triarii 10 men deep. Total of 42 yards. Add some distance between the lines and you get 50 - 60 yards or more. The legions were approaching a square shape.

Infantry didn't deploy too wide. The manuals fix the width of an infantry phalanx at about 1000 yards. Vegetius does not recommend a width greater than 1,650 yards: "One thousand paces contain a single rank of one thousand six hundred and fifty-six foot soldiers, each man being allowed three feet." Any wider is problematic: "Six ranks drawn up on the same extent of ground will require nine thousand nine hundred and ninety-six men. To form only three ranks of the same number will take up two thousand paces, but it is much better to increase the number of ranks than to make your front too extensive." So 1,600 yards frontage for the Cannae legions looks about right.

Working on the Spanish being about 4,000 strong and the Gauls 16,000 strong, we need a means of having a front line of alternating Gallic and Spanish speira, which means the Gauls and Spanish are in equal numbers in the crescent line. That means 8,000 men. If they spread across the entire 1,600 yards they are 5 men deep. You then have to account for the remaining Gauls - 12,000. Simplest way is to put them in a second line (which Livy mentions several times) which will be 7-8 men deep. However I worked on the Libyans being deployed in line and facing the Roman infantry since there is no mention of them engaging the Roman cavalry. If they deploy each in a single line they can't 'extend the wings' and encircle the Romans - impossible to execute such a manoeuvre whilst fighting to your front. If they deploy in 2 lines each 10 men deep (could be a little less)  each wing will front 200 yards of the Roman line, leaving the Gauls and Spanish with 1,200 yards. That makes the crescent 6-7 men deep and the support line 10 men deep. It works.

*largely Patrick's actually

[Imperial Dave]

if the Libyans were in column, then they would also be partially (or even wholly) hidden from Roman eyes in the lead up to the first clash of the infantry lines

[Jim Webster]

I don't think that the Africans need to be in column
I suspect that because of drawing the centre forward and the screens of light infantry both sides deployed, the Romans might have struggled to see what was going on with the Africans anyway.

I apologise for the low grade picture

Take Polybius
"he stationed the Iberian and Celtic horse opposite the Roman cavalry; and next to them half the Libyan heavy-armed foot; and next to them the Iberian and Celtic foot; next, the other half of the Libyans, and, on the right wing, the Numidian horse. Having now got them all into line he advanced with the central companies of the Iberians and Celts; and so arranged the other companies next these in regular gradations, that the whole line became crescent-shaped, diminishing in depth towards its extremities: his object being to have his Libyans as a reserve in the battle, and to commence the action with his Iberians and Celts."

No mentions of column, just an army drawn up in line as you'd expect

then the legionaries took the place of the light-armed and closed with the enemy. For a short time the Iberian and Celtic lines stood their ground and fought gallantly; but; presently overpowered by the weight of the heavy-armed lines, they gave way and retired to the rear, thus breaking up the crescent. The Roman maniples followed with spirit, and easily cut their way through the enemy's line; since the Celts had been drawn up in a thin line, while the Romans had closed up from the wings towards the centre and the point of danger. For the two wings did not come into action at the same time as the centre: but the centre was first engaged, because the Gauls, having been stationed on the arc of the crescent, had come into contact with the enemy long before the wings, the convex of the crescent being towards the enemy.
The Romans outflanked by the cavalry.
The Romans, however, going in pursuit of these troops, and hastily closing in towards the centre and the part of the enemy which was giving ground, advanced so far, that the Libyan heavy-armed troops on either wing got on their flanks. Those on the right, facing to the left, charged from the right upon the Roman flank; while those who were on the left wing faced to the right, and, dressing by the left, charged their right flank,1 the exigency of the moment suggesting to them what they ought to do. Thus it came about, as Hannibal had planned, that the Romans were caught between two hostile lines of Libyans—thanks to their impetuous pursuit of the Celts. Still they fought, though no longer in line, yet singly, or in maniples, which faced about to meet those who charged them on the flanks.

1 ἐξ ἀσπίδος ἐπιπαρενέβαλλον. The ordinary word for "forming line" or "taking dressing" is παρεμβάλλειν. In the other two passages where ἐπιπαρενβάλλειν is used, ἐπί has a distinct (though different) force. I think here it must mean "against," "so as to attack." And this seems to be Casaubon's interpretation.

My conclusion is that the Romans never hit the Africans, the Africans hit the Romans and not from the front.

[Justin Swanton]

The Romans, however, going in pursuit of these troops, and hastily closing in towards the centre and the part of the enemy which was giving ground, advanced so far, that the Libyan heavy-armed troops on either wing got on their flanks. Those on the right, facing to the left, charged from the right upon the Roman flank; while those who were on the left wing faced to the right, and, dressing by the left, charged their right flank,1 the exigency of the moment suggesting to them what they ought to do. Thus it came about, as Hannibal had planned, that the Romans were caught between two hostile lines of Libyans—thanks to their impetuous pursuit of the Celts. Still they fought, though no longer in line, yet singly, or in maniples, which faced about to meet those who charged them on the flanks.

One interesting word there: 'faced about' - from στρέφω strepho - with the meaning in a military context of turn/twist about, i.e. turn 180 degrees: http://www.perseus.tufts.edu/hopper/text?doc=Perseus:abo:tlg,0032,010:11:9&lang=original

Which implies the Libyans were in the Roman rear rather than strictly speaking on their flanks.

[Erpingham]

Reading the two source passages, I was struck by Livy's account of the survivors. Most flee to the two Roman camps behind the army.  Only 3000 push forward to Cannae.  So it seems most of the survivors didn't break out forwards Trasimene style, they broke out from the cavalry encirclement at the back.

[Duncan Head]

Working on the Spanish being about 4,000 strong and the Gauls 16,000 strong, we need a means of having a front line of alternating Gallic and Spanish speira, which means the Gauls and Spanish are in equal numbers in the crescent line.

If we can take Polybios' "alternately" (enallax) absolutely literally (G:S:G:S ...), rather than meaning just "intermixed"; and as applying to the whole line and not just, for instance, the central portion.

And if Gallic and Spanish speirai were of the same strength.

And if Gallic and Spanish speirai formed up in the same number of ranks.

I submit that none of the above are certain.

[Justin Swanton]

Working on the Spanish being about 4,000 strong and the Gauls 16,000 strong, we need a means of having a front line of alternating Gallic and Spanish speira, which means the Gauls and Spanish are in equal numbers in the crescent line.

If we can take Polybios' "alternately" (enallax) absolutely literally (G:S:G:S ...), rather than meaning just "intermixed"; and as applying to the whole line and not just, for instance, the central portion.

And if Gallic and Spanish speirai were of the same strength.

And if Gallic and Spanish speirai formed up in the same number of ranks.

I submit that none of the above are certain.

OK, but we still need a thin crescent and that second line behind it (unless, as Richard says, I got Livy's Latin completely wrong).