Slavery and early medieval economic growth

I am very slowly working my way through last term’s IHR Earlier Medieval Europe seminars, and the next one I want to blog is Marek Jankowiak talking about “Dirhams for slaves: investigating the Slavic slave trade in the tenth century”. Jon Jarrett has already written a report on an earlier version of this paper, so I can summarise the argument fairly quickly.

Marek’s basic thesis is that in the ninth and tenth centuries, a massive trade in slaves underpinned Europe’s economic revival. As he points out, this was a view popular among nineteenth-century historians and more recently eastern European historians, who saw the slave trade as the motor of the economy. More recently Michael McCormick has argued for a large-scale Mediterranean slave trade.

Marek’s looking at evidence from the eastern European network, or rather two networks: one going to Umayyad Spain, with Prague as the key centre, plus another network from shifting areas in Gotland, Poland and Estonia to the Caliphate, on which Marek was focusing. One of his key points was trying to show the scale of this trade, deducing this from hoards of silver in the north. As far as I could note it down, his calculations were as follows:

1) He estimated there was around 1 tonne of silver dirhams in the known hoards.

2) Of this silver, around 2.5-5% were imitations: coins produced by the Khazars and the Volga Bulgars, who acted as middlemen. Marek argued that these coins were produced solely for this trade, as a way of evening out trade flows: that is, when slaves were brought and there wasn’t an immediate buyer, the Khazars/Bulgars minted silver to pay off the northern traders and then waited for the Muslim buyers to come.

3) Marek then used what I think was an argument from die chains in the imitations to argue for the total size of that coinage. He then used this to scale up the total inflow of silver into the north in the tenth-century via this route to 25-50 tons. (This is where my hazy notes meet my limited knowledge of numismatics, so if I misunderstood him, apologies).

4) He assumed that 75% of this inflow was for slaves, arguing essentially, that there wasn’t much else worthwhile that the parts of the north he was interested in had to sell. Large fur production sites are visible in the far north, but there’s nothing much in Poland, for example.

5) He therefore argued that in the tenth century around 5-10 million dirhams were paid for slaves, and given that around 100 dirhams was the price of a slave in the Bulgar markets, that would mean around 50,000-100,000 slaves sold in the tenth century.

As Marek said, that was probably a figure that erred on the cautious side: even if you added in slaves in the western trade, you’re looking at around 1000 slaves a year. He then went onto argue for some of the large number of big hill-forts that appear in tenth-century Poland as being slave camps, places where caravans of slaves being transported could be kept for a night. He also suggested that signs of depopulation between the ninth and tenth centuries in e.g. the western part of Greater Poland might be entire tribes disappearing because they’d been captured and sold.

Jon’s discussed some of the archaeological and numismatic problems with this hypothesis. I want to think about a more general question: is a slave trade on this scale likely and could it act as a motor for economic growth? One of the things that the Making of Charlemagne’s Europe project is doing is providing a prosopography of everyone in charters from 768-814 and I can already say that there were a hell of a lot of unfree. I’m inputting a charter currently that lists 60 mancipia by name given to Wissembourg and though that’s a big list, it’s not the biggest there is and the donor is at most from the regional elite. Elipandus of Toledo claimed that Alcuin, as abbot of Tours, controlled 20,000 servi and that doesn’t sound implausible.

The problem with the argument for slave-trading as the motor of the Carolingian economy, however, is there isn’t much evidence for a large-scale slave trade in the Carolingian world. Even Joachim Henning, talking about a rise in Carolingian period slave shackles, admits that they’re not found within the Carolingian empire. In addition, if slave-trading (i.e. selling on of captives to Muslim markets) was the best way to get rich in eighth and ninth-century Francia, you’d expect more discussion of it in the Carolingian sources. (Probably to justify it morally, given the way that Carolingian authors could justify practically anything rich noblemen did).

There is some evidence both for the resettling of “foreign” captives within the bounds of the Carolingian empire and of the normal taking captive of women and children in warfare that John Gillingham has discussed. But as Henning points out, the selling abroad of the western European rural peasant population by the Frankish nobility was exceptional. If we believe Marek, however, it wasn’t in Eastern Europe at the same time. Why the difference?

Henning, in “Slavery or freedom? The causes of early medieval Europe’s economic advancement”, Early Medieval Europe 12 (2003), 269-277 comments on McCormick’s thesis (p. 273):

it is difficult to imagine that landlords of the west would have systematically destroyed their most effective rural production network – the manorial system – by selling their own peasantry into slavery only to buy silk and drugs in southern markets.

.

In a settled society, the agricultural use of captured slaves and the unfree who might have been “bred” on estates normally makes more economic sense than selling them for luxuries. But there are a few situations where such economic logic may not hold. One is if there is a shortage of agricultural land, so that having extra mouths to feed isn’t compensated for by extra productivity. That’s possibly the case for at least some parts of Scandinavia.

Secondly is when a short-term influx of money allows you to gain a major competitive edge. North-Eastern Europe at this point is having proto-states form (one of Marek’s incidental points was that the Piast state in Poland appears later than previously thought). Is it possible that you’ve got a situation something like Steve Bassett’s FA cup model of small kingdoms violently clashing and being absorbed? In that case it might be worth hoovering up a small tribe or two and selling them off for cash, because one of the things we know that is being bought by people in the eastern borderlands from the Carolingian empire is arms and armour. If you can get money to pay and equip warriors now, that may be a better bet waiting for better harvests six months or even longer down the line from an increased labour supply. Marek’s argument that slavery provided the capital for creating northern states may be correct, but that wouldn’t necessarily make slave-trading the motor for expansion in the whole of ninth-century Europe.