Grasping Reality on TypePad, by Brad DeLong

Jonathan Jarrett:

I should have read this the moment I bought it, II « A Corner of Tenth-Century Europe: The second article in this book I’m reading, not counting the introduction and the mises-au-point that Michael McCormick supplies for each section, is as I said by Joachim Henning and it’s a bit less limpid. The argument is basically that finds of slave-chains match up with settlement patterns to suggest that:

1. The Romans didn’t allow Germanic-style village farmstead sort of affairs with individual enclosures, but the barbarians on their borders farmed like that normally

2. Once the Germans are on the inside Merovingian Francia is full of that sort of settlement, agricultural slavery inside the old imperial limes is basically over and slavery becomes something you only see on the military frontier

3. Under the Carolingians however some of the Roman unification of settlements into grand estates resumes and this is bad for the economy

What we have here is basically two patterns, of slave-chain finds and of settlement nucleation, on which is being hung an awful lot, as well as some problematic stuff about the spread of the heavy plough that I thought we’d got over by now. An awful lot is hanging from those chains but they self-evidently don’t reliably index the whole slavery complex; they only show that someone had to be constrained. These could be military captives, their absence still wouldn’t prove that the people who worked an estate weren’t bound to it by fear, threat and the law.

Then, an awful lot hangs from the belief that the evidence shows that bipartite estates were less efficient agriculturally than decentralised farming because of the need for coercion. Philosophically this seems likely to be right, and one could cite the collective farms of the Soviet Union as well as studies of plantation slavery in the USA to show that people work harder when they work for themselves (and Henning has a nice example to counter Chris Wickham’s belief that where there are no lords the peasants eat more and work less).2 All the same, there is an issue here: we are asked to believe that the Carolingian nobility, or the US planters, must have been status-hungry megalomaniacs, otherwise they’d have realised that ’slavery does not pay’.3 But think about it: the question is not about how much their estates produced, at least in the Frankish case, it’s about how much of that production they could appropriate. If labour on your estate is 10% less effective for being combined and worked as a demesne compared to hutted coloni working their own plots, but you can get 5% more labour out of the slaves for having them right there and that also means you can impose renders with 7% greater efficiency (or other made-up figures that would work, if these don’t), then you as lord are in profit and it does make economic sense. But Henning has a basic ’slavery is bad mkay’ assumption here that makes it difficult for him to see this. I mean, I agree with that statement, but the estate managers in any of those periods obviously didn’t hold that conviction, that doesn’t make them stupid. And then it doesn’t help that, as Angeliki Laiou points out in the response at the end of the section, that the slave-chain finds for the Carolingian period don’t occur in the areas where there were bipartite estates.4 That is, unlike the Roman slave estates apparently the Carolingian ones didn’t use chain-gangs but got their labour more willingly. So, er, hang on, where did that paradigm go? There are some points here but they aren’t all the ones the author feels that he’s made.

On the whole what I take from this is that being part of the Roman fiscal complex tended to produce a different kind of estate organisation, and that the Carolingians also did some of that. I take with more salt the idea that the Carolingian period might have been less well off than the Merovingian period, and if I accept it I would again want to blame the weather for most of that and wonder if the conversion to bipartite estates in places where that can be done isn’t more of a response, both to diminishing yields and also to newly huge scales of estate ownership.5 (In other words, Chris’s ‘aristocrats make complexity’ argument tied up with my own macro-economic ones.) I don’t think slave-chains prove what Henning thinks they do, though the distribution is interesting (if, as he admits, potentially faulty).6 But most of all I wonder whether the horse of cause is not before the cart of effect here. Even if we accept the correlations Henning proposes, correlation is not causation, and the causation is still to be sorted out I think.