The bow (ehrmm...CRESCENT) at Cannae

[Justin Swanton]

I've noticed that whenever somebody has no idea why a formation or a tactic was effective they affirm it worked because it was 'flexible'. This cropped up in the documentary Hannibal, Rome's Worst Nightmare, where the great general lays out his battleplan to his commanders the night before the battle, showing them the bow-shaped infantry formation in the middle of his line and affirming it will be "strong yet flexible."

I doubt flexibility had much to do with why Hannibal deployed his Gauls and Spaniards in a bow (or perhaps wedge). So why did he do it? My working hypothesis is that he wanted to immediately engage the Roman cavalry with his own thus fixing them in place, and then suck the Roman infantry forwards by creating a line that would recoil before the Roman foot but not have to engage them all at once. Only the centre of his line, where his best troops are under his personal command, initially fight the Roman foot; the flanks of his line come into the fray later as the centre falls back.

This would achieve two things: draw out the infantry engagement and separate the Roman infantry from their cavalry, enabling Hannibal to make use of his superior number of horsemen to envelope the Roman cavalry. The Romans countered by dismounting against which the Carthaginians and Gauls dismounted themselves, creating a deeper line that overwhelmed the Romans and opened the way to surround the Roman foot.

This is the only way I can make sense of bow/wedge-shaped infantry line (which otherwise seems completely bizarre). Comments?

PS: I might have already posted a thread on this but can't remember.

[Imperial Dave]

the immediate question I have is how the centre units were able to fall back in order whilst fighting to the front especially against the Romans. Was it purely design and did Hannibal have complete trust in the discipline of his own infantry to be able to pull this off or was there a luck element in it?

[Justin Swanton]

the immediate question I have is how the centre units were able to fall back in order whilst fighting to the front especially against the Romans. Was it purely design and did Hannibal have complete trust in the discipline of his own infantry to be able to pull this off or was there a luck element in it?

My reading of the sources tells me that trained infantry could fall back hundreds of yards without a problem - e.g. the Macedonian pikemen at Sellasia, Philip's pikemen at Chaeronea. I side with Patrick on the notion that by Cannae Hannibal's Spaniards and Gauls were in fact trained heavy infantry and not wild and woolly warband. If he told them ahead of time they must yield ground as part of the plan they would have done it without a problem.

[Imperial Dave]

I agree that they were well trained by Cannae by all accounts. If the plan was to retreat in order before commencing the battle and thus completed during the battle, it surely has to rank as one of the tactical manoeuvres of all time. 

[Duncan Head]

When I saw the thread title I thought for a moment you were about to reveal the untold story of archery at Cannae.

[Justin Swanton]

Thinking about it, the Roman deployment at Cannae was actually quite smart if one makes the assumption that Numidian LH always operated independently of Carthaginian/Spanish/Gallic medium cavalry. That assumption seems fairly safe as the Numidians were invariably deployed as a separate wing in all of Hannibal's battles, implying that they operated in a fashion fundamentally different to medium cavalry and hence couldn't be cobbled together with them.

At Cannae Hannibal had 4000 Numidian LH and 2000 Spanish, 4000 Gallic and 450 Carthaginian medium cav. Rome had 6400 medium cavalry. Rome would assume that the Carthaginians would deploy their Numidians on one flank and their medium cav on the other. So 4000 LH on one flank and 6450 medium cav on the other. Rome had 4800 Italian cavalry and 1600 Roman cavalry. Deploying 4800 on one flank and 1600 on the other, with the 1600 occupying a narrow space between the Roman infantry and the river ensured that the Roman cav would not be heavily outnumbered on either flank: either Hannibal's 4000 Numidians or his 6450 medium cav would confront the 4800 Italan cav on the open flank, whilst the 1600 Roman cav on the closed flank could handle any number of enemy cav since depth was of no use in mounted engagements. A good plan.

[Justin Swanton]

When I saw the thread title I thought for a moment you were about to reveal the untold story of archery at Cannae.

That comes later...

[Erpingham]

Republican Romans are something I've learned about largely by accident (i.e. by watching Patrick argue with everyone else) but I would have some questions of "the bow".  It is often shown in text books as a smooth arc or lens.  Is this supported by the sources?  making a smooth lens would be a serious piece of drill and deployment, yet the people who did it are supposed to be barbarians.  Or are we just seeing a deployment where the flanks of the centre are refused? 

Then we have the withdrawal part.  What does this tell us about the fight?  We know in other battles of the Republic, barbarians could fight fiercely, become worn out and fall back in a defensive formation.  Is Hannibal exploiting that here?   Does the fact that the Carthaginian centre can fall back tell us anything about the intensity of fighting?  If the entire Roman front executed a "chuck and charge", would a withdrawal be possible?  Or are the Romans more in "ebb and flow" mode, at least to begin with, perhaps commiting more as the enemy centre seems to be giving way?

[Nick Harbud]

When I saw the thread title I thought for a moment you were about to reveal the untold story of archery at Cannae.

...or 'How to do Creçy during the 2^nd Punic War'   

[RichT]

OK I'll bite.

Justin

So why did he do it? My working hypothesis is that he wanted to immediately engage the Roman cavalry with his own thus fixing them in place, and then suck the Roman infantry forwards by creating a line that would recoil before the Roman foot but not have to engage them all at once. Only the centre of his line, where his best troops are under his personal command, initially fight the Roman foot; the flanks of his line come into the fray later as the centre falls back.

First of all I don't think that 'bow' is the ideal word - it causes the confusion with archery (which was my first thought too). 'Wedge' is worse. Polybius uses 'menoeides', usually translated 'crescent' or 'crescent-shaped', so best stick to that.

Then, your working hypothesis seems OK, as after all it is broadly what Polybius says (except for the stuff about cavalry, which I'll return to). We can't know what was in Hannibal's mind, and neither did Polybius, but Polybius probably had a better chance than us of guessing, or of hearing something close to the truth from others, so the best bet seems to be to follow what he says, which is:

"his object [in forming the crescent] being to have his Libyans as a reserve in the battle, and to commence the action with his Iberians and Celts" (Pol.3.113.9)

"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." (Pol.3.115.11)

So the point of the formation was that the Romans would fight the Celts and Iberians first, and would be outflanked by the previously unengaged Libyans. Of course there are all sorts of practical difficulties and 'whats' and 'whys' and 'hang on a minutes' around all this, which we have probably all been over many times - but the basic plan, as Polybius understands it, was this.

Cavalry - I don't see any relevance of the infantry formation to the cavalry. The cavalry will do their own thing on the flanks, and Carthaginian cavalry superiority was no doubt always part of Hannibal's plan, but the crescent shaped infantry formation doesn't seem to contribuute to it at all.

Controlled withdrawal of the trained and disciplined Celts and Iberians - perhaps, but it's not what Polybius describes:

"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." Pol. 3.115.5-6

This doesn't sound like a controlled withdrawal but a defeat, as the heavy Celtic losses would indicate:

"On the side of Hannibal there fell four thousand Celts, fifteen hundred Iberians and Libyans, and about two hundred horse." Pol.3.117.6

So it appears that Hannibal's plan (or at least the way things worked out, which he could later claim was his plan) was to advance the Celts and Iberians to engage the Romans first; when they were defeated the Romans in pursuit would close up toward the centre, where the fighting was, and advance between the flanking Libyan forces; then the Libyans would take the Romans in the flank, assisted by the cavalry which had in the meantime won their battle against inferior Roman opposition.

It appears that this plan could only work if the Romans did what they did, which is form up deeper than usual, with smaller gaps between maniples than usual, and then close further toward the centre as they pursued the Celts (and Iberians), all of which seem to be things it would have difficult for Hannibal to foresee. But that is just one the many mysteries about Cannae.

[RichT]

Anthony:

Republican Romans are something I've learned about largely by accident (i.e. by watching Patrick argue with everyone else) but I would have some questions of "the bow".  It is often shown in text books as a smooth arc or lens.  Is this supported by the sources?  making a smooth lens would be a serious piece of drill and deployment, yet the people who did it are supposed to be barbarians.  Or are we just seeing a deployment where the flanks of the centre are refused?

That is indeed the question. The neat little battle diagrams that wargamers and those of a wargamery disposition always draw show a lovely curve, like a young moon, of Celts/Iberians, flanked by two columns of Libyans, and probably bear as little relation to what really happened as these things generally do.

Here's what Polybius says:

"Having now got them all into line he advanced with the central companies [tagmata] 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." Pol 3.113.7-9

This is Shuckburgh's translation, or if you prefer Paton:

"After thus drawing up his whole army in a straight line, he took the central companies of the Spaniards and Celts and advanced with them, keeping rest of them in contact with these companies, but gradually falling off, so as to produce a crescent-shaped formation, the line of the flanking companies growing thinner as it was prolonged, his object being to employ the Africans as a reserve force and to begin the action with the Spaniards and Celts."

Which is clear enough - advancing tagmata by varied amounts to produce a crescent (on the grand scale), which would have (on the smaller scale) have been formed from stepped units, so not a smooth curve (Peter Connolly illustrates this well IIRC).

The problem is with "diminishing in depth towards its extremities" which doesn't make sense if the crescent is formed of stepped units (which would have been facing forward, and have had the same frontage, and therefore depth, as if they were in a standard line). So something is amiss with the formation, or with Polybius' understanding of the formation, or with our understanding of Polybius (or all three).

[Justin Swanton]

OK I'll bite.

Justin

So why did he do it? My working hypothesis is that he wanted to immediately engage the Roman cavalry with his own thus fixing them in place, and then suck the Roman infantry forwards by creating a line that would recoil before the Roman foot but not have to engage them all at once. Only the centre of his line, where his best troops are under his personal command, initially fight the Roman foot; the flanks of his line come into the fray later as the centre falls back.

First of all I don't think that 'bow' is the ideal word - it causes the confusion with archery (which was my first thought too). 'Wedge' is worse. Polybius uses 'menoeides', usually translated 'crescent' or 'crescent-shaped', so best stick to that.

Fair enough.

Then, your working hypothesis seems OK, as after all it is broadly what Polybius says (except for the stuff about cavalry, which I'll return to). We can't know what was in Hannibal's mind, and neither did Polybius, but Polybius probably had a better chance than us of guessing, or of hearing something close to the truth from others, so the best bet seems to be to follow what he says, which is:

"his object [in forming the crescent] being to have his Libyans as a reserve in the battle, and to commence the action with his Iberians and Celts" (Pol.3.113.9)

Makes sense to an extent. The Libyans weren't just a reserve in the sense of a second line ready to buff up / take over from the first line, but were divided into two contingents that outflanked the Roman line. But if they were deployed just behind the crescent, then they wouldn't have to fall back with the retreating front crescent as the flanks of that cresecent didn't move since they hadn't yet engaged the Romans. So the Libyans were free to commence their outflanking manoeuvre when Hannibal judged the time right.

"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." (Pol.3.115.11)[/color]

So the point of the formation was that the Romans would fight the Celts and Iberians first, and would be outflanked by the previously unengaged Libyans. Of course there are all sorts of practical difficulties and 'whats' and 'whys' and 'hang on a minutes' around all this, which we have probably all been over many times - but the basic plan, as Polybius understands it, was this.

Fine.

Cavalry - I don't see any relevance of the infantry formation to the cavalry. The cavalry will do their own thing on the flanks, and Carthaginian cavalry superiority was no doubt always part of Hannibal's plan, but the crescent shaped infantry formation doesn't seem to contribuute to it at all.

I'm musing on the fly, but the question does arise about the Roman cavalry deployment. Unlike infantry, cavalry don't get any advantage from depth, and the Roman left flank cavalry wing of 1600 horse could have kept any amount of enemy cavalry at bay provided its flanks were secured by the river and the Roman infantry. Once the infantry advanced that was no longer the case. I argue that Hannibal intended it that way which was one reason at least for using a crescent - it would encourage a Roman advance in the centre.

Controlled withdrawal of the trained and disciplined Celts and Iberians - perhaps, but it's not what Polybius describes:

"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." Pol. 3.115.5-6

This doesn't sound like a controlled withdrawal but a defeat, as the heavy Celtic losses would indicate:

"On the side of Hannibal there fell four thousand Celts, fifteen hundred Iberians and Libyans, and about two hundred horse." Pol.3.117.6

So it appears that Hannibal's plan (or at least the way things worked out, which he could later claim was his plan) was to advance the Celts and Iberians to engage the Romans first; when they were defeated the Romans in pursuit would close up toward the centre, where the fighting was, and advance between the flanking Libyan forces; then the Libyans would take the Romans in the flank, assisted by the cavalry which had in the meantime won their battle against inferior Roman opposition.

It appears that this plan could only work if the Romans did what they did, which is form up deeper than usual, with smaller gaps between maniples than usual, and then close further toward the centre as they pursued the Celts (and Iberians), all of which seem to be things it would have difficult for Hannibal to foresee. But that is just one the many mysteries about Cannae.

There is a real problem with Polybius' passage, in that taken at face value it makes it inevitable that the Roman foot would have escaped the trap by bursting through the Carthaginian centre - just as they had done at the Trebia. Hannibal had not forgotten the Trebia and neither had the Romans. How did the Roman infantry get boxed in?

Polybius states that the Spanish and Gauls were organised into alternating speira. This implies there were as many Spanish as Gauls, but this wasn't the case. I'm doing this at work so don't have all the sources in front of me but the Wikipedia article on the battle affirms there were 8,000 Spanish and 16,000 Gauls. If 8,000 Gauls shared the crescent with the Spanish, where were the rest? My take is that there must have been a second line of Gauls behind the crescent, which acted as a classical relief line a la Romain. When the Spanish/Gauls "gave way and retired to the rear" they retired to rear of what? Retiring to the rear doesn't sound like breaking and running; it sounds more like falling back behind some reserves.

Putting it all together, Polybius gives a picture of Hannibal using the natural impetuosity of the Gauls and Spanish to convince the Romans that this was a classic barbarian bash, and that when they fell back - a planned fall-back - the Roman infantry would believe they had won the battle and advance against them, which is exactly what Hannibal wanted.

I'm inclined to think that part of the Libyans were in the second reserve line and part formed the outflanking contingents, which would better suit what Polybius says about them.

[Jim Webster]

When I saw the thread title I thought for a moment you were about to reveal the untold story of archery at Cannae.

thank the lord it wasn't just me 

[Jim Webster]

Seriously are we looking at it from the wrong side.

I have no problem with the Carthaginian army, including Spanish and Gauls, being decent regular heavy infantry who'd been constantly under arms for well over a year

The Romans on the other hand?

How about the Romans were a pretty poorly trained and motivated militia, a lot of whom had no combat experience, which is why their commanders decided to keep them in deep columns and just send them forwards.
Hannibal knew enough about the quality of the army that he faced to know that if he pinned them, they were probably competent to roll forward but would probably be a nightmare to stop and even worse to turn or expand

When we had the discussion about Cannae on the DBMM list I suggested that rather with creating Gauls as Regular auxiliaries you ought to start with legionaries being Horde superior

And no, I'm not bitter

[Justin Swanton]

Anthony:

Republican Romans are something I've learned about largely by accident (i.e. by watching Patrick argue with everyone else) but I would have some questions of "the bow".  It is often shown in text books as a smooth arc or lens.  Is this supported by the sources?  making a smooth lens would be a serious piece of drill and deployment, yet the people who did it are supposed to be barbarians.  Or are we just seeing a deployment where the flanks of the centre are refused?

That is indeed the question. The neat little battle diagrams that wargamers and those of a wargamery disposition always draw show a lovely curve, like a young moon, of Celts/Iberians, flanked by two columns of Libyans, and probably bear as little relation to what really happened as these things generally do.

Here's what Polybius says:

"Having now got them all into line he advanced with the central companies [tagmata] 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." Pol 3.113.7-9

This is Shuckburgh's translation, or if you prefer Paton:

"After thus drawing up his whole army in a straight line, he took the central companies of the Spaniards and Celts and advanced with them, keeping rest of them in contact with these companies, but gradually falling off, so as to produce a crescent-shaped formation, the line of the flanking companies growing thinner as it was prolonged, his object being to employ the Africans as a reserve force and to begin the action with the Spaniards and Celts."

Which is clear enough - advancing tagmata by varied amounts to produce a crescent (on the grand scale), which would have (on the smaller scale) have been formed from stepped units, so not a smooth curve (Peter Connolly illustrates this well IIRC).

The problem is with "diminishing in depth towards its extremities" which doesn't make sense if the crescent is formed of stepped units (which would have been facing forward, and have had the same frontage, and therefore depth, as if they were in a standard line). So something is amiss with the formation, or with Polybius' understanding of the formation, or with our understanding of Polybius (or all three).

"Diminishing in depth towards its extremities" actually makes perfect sense if the centre of the crescent is expected to fight for longer than its wings since the centre is in contact with the Romans for longer and hence needs deeper speira to better weather the fighting. It would be stepped speira as you say, not a true curved crescent.