Could the Persian Empire logistically support an army several million strong?

[Patrick Waterson]

This is a bit of a bizarre argument, even by normal standards.  The Greeks are convinced of their superiority over the Persians.  They think a small number of Greeks can beat any number of Persians because that is what they have grown up believing.  The idea that, on meeting the enemy, they say lets do an accurate count of that big horde just to make sure its as big as we have been led to believe, is a bit fantastic.

And is also not my point, which is rather that they did not have such elaborate fantasies derived from over-inflated historical portrayals.  Had they done so, it would have impaired their military performance.

The Greeks understood, following Salamis, Plataea and Mycale, that they were militarily superior to the Persians.  This derived from combat experience against superior numbers, not imagination ingrained by people telling them they could beat superior numbers.  The idea that they beat relatively small numbers of Persians and derived from this a superiority complex that they could beat large numbers of Persians - which seems to be the esssential corollary of any contention that Herodotus inflated his figures - is just not on.

[Erpingham]

You did read the argument further on, didn't you?  The one about a cultural artefact of the "barbarian horde" and how it affected the reporting of numbers?

If that is not imposing our preconceptions upon the past, I do not know what is.

No more so than treating (selected) ancient authors with almost religious reverence.  For the umpteenth time, we are all imposing models of interpretation on the past, from Herodotus to us here now.  Personally, I think the interpretive framework I'm offering, based on a critical approach, will get us near the truth.

2. I believe you said he spoke to Persian officials?  He's believe them because they fed his cultural expectation.

Now why would they even think of doing that?

Are you seriously asking this?  You have proposed that overawing the enemy with a display of strength is the Persian way.  The power and glory of the Persian empire etc.  Herodotus expects big numbers.  Persian officials expect to elaborate to glorify the King.

But some writers from the same cultural outlook, notably Thucydides and Plutarch, get their knives into Herodotus at every opportunity. 

But do they produce smaller numbers for the size of barbarian armies?  You've said they don't earlier. 

Then why ask?   As it happens, no, they do not - apart from Ctesias, who asserts that Xerxes had 800,000 but gives no basis for how he or his presumed source arrives at this figure.  800,000 is about half of 1,700,000 so my suspicion is that Ctesias took the figure he received from his source and halved it thinking it would have included noncombatants.  Conjectural, albeit it makes for easy reconciliation.  What does stand out is that even Ctesias does not hold with 200,000 or so, but a rather larger number.

So, despite being a member of the Persian Court, Ctesias has his figures wrong?  Other than Simonides, who does give similar figures to Herodotus? 

Furthermore, a cultural outlook which compulsively over-estimates and/or overstates enemy numbers is not going to fare as well as did the Greeks in combat against the Persian Empire.

I actually don't get what you mean here.

That is a perceptive observation.

Why, thank you

[Jim Webster]

A lot of herding cultures had numbers of slaves and semi-free males who would continue to work even when the warriors had left.

Aha ...

So there is actually no problem with mobilising the warrior population.  This is what I wished to establish.

no problem at all in mobilising the warrior population, for a short raid or two in which perhaps five percent of the male population (and perhaps a third of the warrior population) took part

You have still not provided any evidence whatsoever of a them mobilising 41% of their male population and sending them away for a couple of years.
So you've established nothing

[Erpingham]

The Greeks understood, following Salamis, Plataea and Mycale, that they were militarily superior to the Persians.  This derived from combat experience against superior numbers, not imagination ingrained by people telling them they could beat superior numbers. 

And Xenophon's comrades and Alexander were at these battles?  Of course not.  They learned about them from older men or from books.  There was a narrative of Greek military prowess and Persian vast but useless armies.  So, when they arrive on the field and see a big Persian army, they just go on and beat it, because its what their ancestors have always done.

[Jim Webster]

You have pretty much answered your own question.  What we have of Ctesias is, for one reason or another, suspect.  There is nothing to support his 800,000-man figure whereas Herodotus has the process by which the count was made and the final figure of 1.7 million arrived at, plus incidental details which correspond with a figure of this scale (e.g. trans Hellespont crossing time).

how do you know there is nothing to support his figures, we don't have his text!

As for Herodotus 'having a process'. Merely because a writer describes something doesn't mean it actually had to happen like that. He wasn't an eye witness, he was merely writing down something he'd been told, and the person who told him need not have been an eyewitness.
It may merely be recounted folk tradition, 'this is how Persians measure their armies'

If you're going to claim that Ctesias is suspect, please produce evidence. And the fact that he doesn't agree with Herodotus isn't evidence in a discussion as to whether herodotus is accurate. For some historians the fact that he differs from Herodotus might be taken as a sign that he's trustworthy.
So how is he suspect?

[Justin Swanton]

A lot of herding cultures had numbers of slaves and semi-free males who would continue to work even when the warriors had left.

Aha ...

So there is actually no problem with mobilising the warrior population.  This is what I wished to establish.

no problem at all in mobilising the warrior population, for a short raid or two in which perhaps five percent of the male population (and perhaps a third of the warrior population) took part

You have still not provided any evidence whatsoever of a them mobilising 41% of their male population and sending them away for a couple of years.
So you've established nothing

Sorry, where does the 41% come from? If 5 million is 41 percent of all males in the empire, that gives a male population of 12.2 million and a total population of 24.4 million. Most estimates of the population range between 40 and 50 million, which means 20% to 25% of males took part in the campaign.

[Duncan Head]

If you're going to claim that Ctesias is suspect, please produce evidence.

Ctesias, even if he has put into his work a perfect farrago of extravagant and incredible tales ...

Mind you, Plutarch was even harder on Herodotos. The only moral seems to be that even ancient historians thought that ancient historians were unreliable.

[Justin Swanton]

If you're going to claim that Ctesias is suspect, please produce evidence.

Ctesias, even if he has put into his work a perfect farrago of extravagant and incredible tales ...

Mind you, Plutarch was even harder on Herodotos. The only moral seems to be that even ancient historians thought that ancient historians were unreliable.

Which also shows they set as high a store by reliability as we do. If you knew you were going to be attacked by your colleagues for any mistakes you checked your sources carefully. Just like contemporary academia.

[Duncan Head]

Which also shows they set as high a store by reliability as we do. If you knew you were going to be attacked by your colleagues you checked your sources carefully.

Herodotos, being more or less the first historian, cannot have known that he'd be attacked by Plutarch 500 years later (or even by Ctesias 70 years or so later), so I don't think that is a very strong argument at all.

[Justin Swanton]

Which also shows they set as high a store by reliability as we do. If you knew you were going to be attacked by your colleagues you checked your sources carefully.

Herodotos, being more or less the first historian, cannot have known that he'd be attacked by Plutarch 500 years later (or even by Ctesias 70 years or so later), so I don't think that is a very strong argument at all.

What it does show, at least from Ctesias onwards, is a critical approach to works like Herodotus's. They didn't suck it up with kind of reverent gullibility (which implies they wouldn't necessarily have sucked it up when he published it). Herodotus himself wanted to be accurate - he says so - which suggests that repeating old wives' tales was no longer an acceptable option for an intelligent writer like him.

It certainly suggests that later writers who give similarly enormous figures for the Persian army knew they needed to be on their toes. There was no captive audience.

[Jim Webster]

Sorry, where does the 41% come from? If 5 million is 41 percent of all males in the empire, that gives a male population of 12.2 million and a total population of 24.4 million. Most estimates of the population range between 40 and 50 million, which means 20% to 25% of males took part in the campaign.

Population of males of military age. There was some feeling that I'd been too generous and the 41% figure should have been higher
Military age in the ancient world corresponds pretty closely to useful working age in agriculture as well

[Justin Swanton]

Sorry, where does the 41% come from? If 5 million is 41 percent of all males in the empire, that gives a male population of 12.2 million and a total population of 24.4 million. Most estimates of the population range between 40 and 50 million, which means 20% to 25% of males took part in the campaign.

Population of males of military age. There was some feeling that I'd been too generous and the 41% figure should have been higher
Military age in the ancient world corresponds pretty closely to useful working age in agriculture as well

Oh I see. So the agricultural workforce will be at 59% strength whilst the army is away, which will vary in time depending on whether one started out nearer to Greece or further away. What we need then is a economic system that could grow enough food to offset that drop in labour for a 1 - 2 year period.

I'd like to return for a moment to the subject of subsistence agriculture. It is assumed the Persian Empire had a subsistence agricultural economy, in other words that the 80% or so of the population that worked the land could grow just enough in an average year to feed themselves, their families, and the 20% who didn't live on farms. In a good year they would have a small surplus which would feed everyone in a bad year.

Using Roman land measurements as a guide, 2 iugera (1.3 acres) was enough to feed a man for a year. Presuming a man had a wife who helped sometimes and several small children who couldn't help all that much, and he would probably have to cultivate 6 iugera (4 acres) to feed his family. Once the children grew older he would have help and he could probably count on the services of unmarried labourers, at least sometimes. But let's assume that 4 acres is the limit. Tack on the 20% he had to grow for non-farmers and he is stuck with 5 acres. This is the worst-case scenario.

Roman farms came in several sizes. Small farms ranged from 18–108 iugera (12 - 70 acres). These were family-owned. A family of 6 people would need to cultivate 12 iugera (8 acres) of land to meet minimum food requirements (without animals).

Cato the Elder describes a farm of 100 iugera. He claimed such a farm should have "a foreman, a foreman's wife, ten laborers, one ox driver, one donkey driver, one man in charge of the willow grove, one swineherd, in all sixteen persons; two oxen, two donkeys for wagon work, one donkey for the mill work." Exclude the ox and donkey driver and the wife and you have 13 men working 100 iugera = 65 acres. The farm can feed 65 adult or near-adults for a year. Not all the 16 adults mentioned have families so adding another 16 individuals (wives and children) seems reasonable. That brings it up to 32 people. This means the farm produces twice as much as the the needs of the people cultivating it. Notice that only 10 men actually work the land, each handling about 6 acres. This is not subsistence agriculture.

Presuming then that a typical family farm could produce a surplus of 33% to 50%, Persian farms, during the 4 year run-up period, would have to produce enough to offset the loss of labour during the campaign - and farm enough to produce a 20% surplus only. This means that for each of the 4 years, a labourer would have to cultivate 20% more than he normally would. If 5 acres produces the necessary surplus in normal times, then 6 acres will produce enough to offset the deficit of labour in the two years' campaign. This pretty much matches what an average Roman farm could handle in any event, so there is no reason to assume the Persian agricultural system would buckle under the strain.

[Jim Webster]

Sorry, where does the 41% come from? If 5 million is 41 percent of all males in the empire, that gives a male population of 12.2 million and a total population of 24.4 million. Most estimates of the population range between 40 and 50 million, which means 20% to 25% of males took part in the campaign.

Population of males of military age. There was some feeling that I'd been too generous and the 41% figure should have been higher
Military age in the ancient world corresponds pretty closely to useful working age in agriculture as well

Oh I see. So the agricultural workforce will be at 59% strength whilst the army is away, which will vary in time depending on whether one started out nearer to Greece or further away. What we need then is a economic system that could grow enough food to offset that drop in labour for a 1 - 2 year period.

I'd like to return for a moment to the subject of subsistence agriculture. It is assumed the Persian Empire had a subsistence agricultural economy, in other words that the 80% or so of the population that worked the land could grow just enough in an average year to feed themselves,

the standard figure is 90%.

[Justin Swanton]

Sorry, where does the 41% come from? If 5 million is 41 percent of all males in the empire, that gives a male population of 12.2 million and a total population of 24.4 million. Most estimates of the population range between 40 and 50 million, which means 20% to 25% of males took part in the campaign.

Population of males of military age. There was some feeling that I'd been too generous and the 41% figure should have been higher
Military age in the ancient world corresponds pretty closely to useful working age in agriculture as well

Oh I see. So the agricultural workforce will be at 59% strength whilst the army is away, which will vary in time depending on whether one started out nearer to Greece or further away. What we need then is a economic system that could grow enough food to offset that drop in labour for a 1 - 2 year period.

I'd like to return for a moment to the subject of subsistence agriculture. It is assumed the Persian Empire had a subsistence agricultural economy, in other words that the 80% or so of the population that worked the land could grow just enough in an average year to feed themselves,

the standard figure is 90%.

Fine, that gives more leeway.

[Erpingham]

Which also shows they set as high a store by reliability as we do. If you knew you were going to be attacked by your colleagues you checked your sources carefully.

Herodotos, being more or less the first historian, cannot have known that he'd be attacked by Plutarch 500 years later (or even by Ctesias 70 years or so later), so I don't think that is a very strong argument at all.

What it does show, at least from Ctesias onwards, is a critical approach to works like Herodotus's. They didn't suck it up with kind of reverent gullibility (which implies they wouldn't necessarily have sucked it up when he published it). Herodotus himself wanted to be accurate - he says so - which suggests that repeating old wives' tales was no longer an acceptable option for an intelligent writer like him.

Which is why Aristotle called Herodotus a great storyteller?  Why Cicero, his greatest fan, thought he was too prone to including legends?

It certainly suggests that later writers who give similarly enormous figures for the Persian army knew they needed to be on their toes. There was no captive audience.

This doesn't follow.  I think it is reasonable that Greeks thought that Persian armies were huge and therefore all histories, to be taken seriously, would have huge Persian armies in them.  This does not mean that they were good at quantifying what really huge was. 

I'm no means up on all this scholarship (Duncan is the only legitimate scholar of the Persian army here I think - Christopher Tuplin thinks Duncan's book on the subject is "one of the most useful publications on Achaemenid warfare" ) but, contra Patrick , it seems most ancient authors place their Persian hordes in the 100,000s, not the millions.  Most of these were not effectives - they were poor quality infantry or support elements who were just estimated in line with expected literary convention.  As I've already said, when it comes to specific corps of elites or critical troops, numbers are probably more accurate. 

Incidentally, if we took Ctesias' figures and assumed he also believed there were as many hangers on as fighters, we would get 400,000 soldiers and a similar cast of supporters.  His ship figures are similar to Herodotus (he has 1000 triremes) so around 200,000 men in the fleet.  About 2% of the population. He leaves only 120,000 in Greece with Mardonius.

Addendum : Talking of Christopher Tuplin, I don't know if this article on Achaemenid cavalry is of any use in the numbers game?