2:1 against: the misconception about Carolingian cereal yields

Let me try this non-aggressive corrective thing again. I posted a little while ago to the effect that I had one piece of bad history I’d like to contribute to the pile of zombie argument effigies that we would like an ideal version of the History and Policy Institute ‘Bad History’ series to be burning. That’s not very non-aggressive, I grant you, but wait, this is the new dispensation. As Magistra et Mater soberly pointed out, what the History and Policy team really want are mistakes that are misinforming public policy, and this is not one of those, so, I guess I’ll have to do it here. It’s not important to anyone now except historians, but it keeps getting repeated. Yet it’s not really sustainable, and smells suspiciously like the same contempt for the medieval that leads to stereotypes of benighted lack of hygiene or eating rotten meat. What is it? It is the idea that in the Carolingian period yields on cereal crops were scarcely twice what was sown. It would be very hard to explain how that particular super-state managed to field such large armies for so long if this were really so, and several people have disputed it, but it keeps lurching back and when I found it in a textbook I am currently evaluating (which is otherwise very good) I felt I had to set something down about why it almost certainly isn’t true.¹

Georges Duby

Where did these authors get the idea from? I could already guess the answer to this: it’s where I first met it, the work of this man, Georges Duby. And, sure enough, in Moran & Gerberding’s chapter bibliography (I told you this was a good textbook) we find Georges Duby’s Early Growth of the European Economy and if you have at that, you will find the basis for this claim detailed there, although it stemmed from his awareness of much older work that made similar points.² However, he actually set it out in rather more detail in his earlier Rural Economy and Country Life, and I’ll use that to show you where he was getting it from, since I own a copy to, er, copy:

One document only for northern Gaul provides some figures. The surveyors who visited the the royal estates attached to Annapes in the winter recorded both the quantity of the previous harvest and the amount which had just been subtracted for the sowing – they certified that the remainder was actually in the barns at the time of their visit. These figures are very baffling. Here are those for the estate of Annapes, for which the inventory gives the most complete details. It is not possible to compare seed with harvest for oats, peas or beans since the spring sowings had not been made. But of the 1,320 muids of spelt harvested 720 had to be returned to the land as seed; of 100 muids of wheat, 60; of 1,800 muids of barley, 1,100; and finally the new sowing absorbed the whole of the rye harvest, that is, 98 muids. The available surplus of the harvest did not therefore appear that year to exceed 46 per cent for spelt, 40 per cent for wheat, 38 per cent for barley, that is an output of 1·8, 1·7 and 1·6 to one respectively. There appeared to be no surplus for rye. The fragmentary evidence given for other estates agrees; an output of 2:1 for spelt and of 1·6:1 for rye at Cysoing; for barley 2·2:1 at Vitry, 1·5:1 at Cysoing, 2:1 at Somain.⁴² Taken altogether the consumable surplus is revealed as markedly less, in the year of the inventory, than the quantity which had to be reserved for sowing. Could output really have been at such a derisory level?

The text, however, is categorical. It prevents us from assuming that, apart from seed corn, grain had already been taken away between harvest time and the visit of the compilers of the inventory for domestic consumption or for despatch outside the estate…. The only reasonable hypothesis to explain the astonishingly low figures for output is to assume that the inventory was compiled after an exceptionally bad harvest. In fact, when it was drawn up grain harvested the previous year was still stored in the barns of these estates in quantities much greater than the insignificant surplus of the current year. The surveyors found at Annapes 1,081 muids of old spelt, as against 600 of new, and 1,200 muids of old barley, as against 700 of new. These important savings prove that the output of seed was clearly much higher the previous year, We can deduce from this unique document that the productivity of the fields varied enormously from one season to another and further that it could be devastatingly low.

We must not, of course, generalize from one set of figures obtained from a single source. But it is possible to find elsewhere some other traces of output, somewhat higher than that which can be derived from the Annapes inventory, but even so representing a low yield and a derisory rate of profit when compared with the value of the capital in land and seed corn….

    ⁴² Grierson, 247. Slicher von Bath, 31, p. 66, does not seem to interpret correctly the figures in the text.³

Well, that’s pretty damning, but how about that source? In fact it’s online, thanks to the efforts of Paul Halsall at the Internet Medieval Sourcebook all those years ago, although the extract they used doesn’t have the equally damning comparative data. Indeed, neither does the translation in Duby’s book itself (Rural Economy‘s particular staying power is at least partly because it’s half-sourcebook), but it does at least direct us to an original.⁴ So of course does the IMSB, we would expect no less, and strangely the two originals are not the same though their translation very nearly is: Duby’s original was Brevium exempla ad describendas res ecclesiasticas et fiscales, for which he cites MGH Cap. I pp. 254-255 whereas the IMSB’s source translation, Ogg’s Source Book of Mediaeval History, pp. 127-129, gives MGH LL I pp. 178-179.⁵ It’s noticeable that neither of these translations actually provide all the material that Duby cites, in particular they skip silently over any reference to last year’s barley without indicating any missing text. Hmph. So we ought to go to the Latin. Now of course these days the MGH is online, which is a thing of great joy, and makes this comparatively easy. I’ll use the MGH Cap. edition since it is newer by fifty years and as the title page proclaims, ‘denuo edito‘:

Invenimus in Asnapio fisco dominico salam regalem ex lapide factam optime, cameras III; solariis totam casam circumdatam, cum pisilibus XI; infra cellarium I; porticus II, alias casas infra curtem ex ligno factas XVII cum totidem cameris et ceteris appendiciis bene conpositis; stabolum I, coquinam I, pistrinum I, spicaria II, scuras III. Curtem tunimo strenue munitam, cum porta lapidea, et desuper solarium ad dispensandum. Curticulam similiter tunimo interclausam, ordinabiliter dispositam, diversique generis plantatam arborum. Vestimenta: lectum parandum I, drappos ad discum I parandum; toaclam I. Utensilia: concas aereas II, poculares II, calderas aereas II, ferrea I, sartaginem I, gramalium I, andedam I, farum I, secures II, dolatoriam I, terebros II, asciam I, scalprum I, runcinam I, planam I, falces II, falciculas II, palas ferro paratas II. Utensilia lignea ad ministrandum sufficienter. De conlaboratu: spelta vetus de anno praeterito corbes LXXXX, quae possunt fieri de farina pensas CCCCL, ordeum modios C. Presenti anno fuerunt speltae corbes CX: seminavit ex ipsis corbes LX, reliqua repperimus; frumenti modii C: seminavit LX, reliqua repperimus; sigilis modii LXXXXVIII, seminavit totidem; ordeo modii MDCCC: seminavit MC, reliqua repperimus. Avena modios CCCCXXX, faba modium I, pisos modios XII. De molinis V: modios DCCC ad minorem mensuram; dedito prebendariis modios CCXXXX, reliqua repperimus. De cambis IV: modios DCL ad minorem mensuram. De pontibus II: sale modios LX, et solidos II. De ortis IV: solidos XI, mel modios III. De censu: butyrum modium I; lartum de praeterito anno baccones X, novos baccones CC cum minucia et unctis; formaticos de anno presenti pensas XXXXIII. De peculio: iumenta maiora capita LI, de anno tertio V, de preterito VII, de presenti VII; poledros bimos X, annotinos VIII; emissarios III, boves XVI, asinos II, vaccas cum vitulis L, iuvencos XX, vitulos annotinos XXXVIII, tauros III, porcos maiores CCLX, porcellos C, verres V, vervices cum agnis CL, agnos annotinos CC, arietes CXX, capras cum hedis XXX, hedos annotinos XXX, hircos III, aucas XXX, pullos LXXX, pavones XXII.

Forty-three maybe-pound-weights of the present year’s cheese! Ahem! I’m sorry. I won’t give a new translation when there’s a perfectly good one out there, other than to confirm as far as I can see, that translation is in fact perfectly good. So, beyond basic Carolingian-puffing, why do I have any basis to think this isn’t correctly interpreted? Well, because of a man called Peter Reynolds whom I’ve mentioned before, and who by his very appearance seems to doom blogposts to a lack of commentary but, dammit, his work was important. He ran an experimental Iron Age farm in the UK, growing historical crops with historical methods, but he also participated in a parallel set of experiments about medieval farming there and in Catalonia. This, unfortunately, means that his most important work on the subject of crop yields, an article called “Medieval Cereal Yields in Catalonia & England: an empirical challenge”, came out in a Barcelona journal that is very difficult to get hold of anywhere else: I can’t find anywhere in the UK that has the relevant volume.⁶ But, the same team minus the late Dr Reynolds has done further work with this stuff in the paper I first blogged about Reynolds because of, which was in Vegetation History and Archaeobotany and which is therefore online through SpringerLink and from that (fair use!) I can scrounge this table of their crop yield results, 1992 to 1995, on two different field-rotation systems.⁷

Harvest year	Cereals	Weight (tons)	Production ratio
3-year, autumn sowing
1992	Triticum dicoccum	2.40	1:34
	Hordeum vulgare	3.67	1:53
1993	Triticum dicoccum	3.58	1:51
	Hordeum vulgare	2.18	1:31
1994	Triticum dicoccum	1.56	1:22
	Hordeum vulgare	1.15	1:16
1995	Triticum dicoccum	1.35	1:19
	Hordeum vulgare	0.36	1:0.5
3-year, spring sowing
1992	Panicum miliaceum	3.82	1:54
	Hordeum vulgare	2.47	1:35
1993	Triticum dicoccum	2.45	1:35
	Hordeum vulgare	2.40	1:34
1994	Triticum dicoccum	–	–
	Hordeum vulgare	–	–
1995	Triticum dicoccum	1.53	1:21
	Hordeum vulgare	0.82	1:12
2-year, autumn sowing
1993	Triticum dicoccum	2.66	1:38
	Hordeum vulgare	2.06	1:29
1994	Triticum dicoccum	1.18	1:17
	Hordeum vulgare	1.05	1:15
1995	Triticum dicoccum	0.11	1:1.5
	Hordeum vulgare	1.11	1:1.5

My goodness I’m glad I didn’t have to code that myself. But you get the point. There are some gaps and drop-outs, which are down to weather. They say: “A dry spring (as happened in 1994, 2005 and 2006) causes a total crop failure, and creates many problems for the following season’s seed corn.”⁸ Otherwise, though, their typical yields, for crop varieties and with techniques at least notionally similar to our peasants at Annapes, or at least peasants in Catalonia who were presumably not massively advanced compared to the Carolingians’ big estates, were in the range of twelve to fifty times seed sown, that is, a whole order of magnitude higher than the Annapes record. Now, if it were just two or three times higher I might wonder, and indeed I do, whether the genotype of these crops could really have remained totally unchanged over those centuries and whether the guys at l’Esquerda may not just have been sowing better crops (for all that they probably aren’t as dedicated farmers as the ninth-century guys whose lives depended on it, which might counter-balance such trends). But this is more than that. This is the point at which we have to ask if we’ve really read the source right.

The test fields at l'Esquerda, Osona, Catalunya, where the trials here cited were carried out

I should say before I go any further that I’m pretty sure I’m not the first person to quarrel with these figures. I don’t remember originating the following counter-argument. It’s in one of my old lectures, too, so I’ve had it for at least two years, but it doesn’t seem to be in anything that I know I read in order to write that lecture.⁹ So if you’ve seen this before, would you like to let us know where? Anyway, all that I am doing here that is original, as far as I know, is linking the documentary scholarship up with the experimental archæology to show that there is good reason to think this old reading of the source is bunk.

What I think is the answer is quite simply that these were not stores for consumption, at Annapes or in the other documents that Duby went on to cite, but reserves. Duby said, you’ll recall: “The text, however, is categorical. It prevents us from assuming that, apart from seed corn, grain had already been taken away between harvest time and the visit of the compilers of the inventory for domestic consumption or for despatch outside the estate.” But if you look at the Latin, actually, there’s no basis for this, the text isn’t categorical at all. It tells us that seed corn had been taken away, but it certainly doesn’t tell us that nothing else had and, even more, it doesn’t tell us the unspoken assumption, that everything harvested from this estate wound up in these barns. And in fact, if we go further into the text, we can see that that’s false. There are five mills listed there, you notice? And it says:

From the five mills: 80 modios of the lesser measure; there had been 240 modios given to the prebendaries, the rest we found.

That’s more grain that’s not in the stores, right there. Categorical my foot: this harvest had already been divided, mostly milled and handed out. Then, following up the reference that Duby gives to Philip Grierson’s work adds still another possible qualification, as Grierson said: “… the account of the stock and produce given by the Brevium exempla applies only to the lord’s demesne, the mansus dominicatus ; it leaves us quite in the dark as to the land in each villa which lay outside the demesne and was held by the serfs”.¹⁰ Now, you could argue that the stuff at the mills would have been brought in from those estates, but if so, it surely shouldn’t count in the survey totals and even if, cruelly, all the serfs’ produce was assumed to be the king’s (in which case, what did they eat?), it still knocks the seedcorn ratio sideways. So, I honestly think that has too many holes in to float.

If you think about the text, and what it actually is, the explanation that we’re counting reserves makes sense. Although as we have it it is preserved as an example of how to do an estate survey (exempla, you see) what this bit clearly was was a stock-take of what was in store at the estate centre. Now, why did Carolingian royal estates collect food? Partly to feed the army, we now suspect, partly for emergency famine relief, and apparently also for a seed bank, but mainly to feed the court if it should come there. If this figure was to be worth reporting, then, it was kind of inherent that the stuff would be there if the king come a-calling, modulo perishability of baccones and so on. If it had all been eaten that wouldn’t be much use. And if it was there to be eaten, it would be hard to explain why there was so much left from the previous year as well. So, what were people eating? Presumably, the produce straight out of the fields as divided at harvest. Now admittedly there’s no evidence for that that I know of, but there’s also no evidence for queues of peasants arriving at the estate centre for a dole of grain from the store every week either, except in time of famine. Why would you organise storage like that? You’d lose so much labour. As long as the renders are correctly coming in, you’d leave the rest to the peasants to sort out, wouldn’t you?

Catalan peasants at work, from the Biblia de Ripoll

So I think that’s how we reconcile these estate survey figures and the radical difference in yields from the archæological experiments. At Annapes and elsewhere, what was being inventoried was what was available to the king and his men after the usual provision for feeding the estates’ inhabitants had been taken. (Otherwise, these surveys would need to contain a population census and a dietary allowance for each person. Which would be fabulous data! But they don’t.) This in turn means, of course, that we have no idea how much of the crop the seedcorn represents, or even how much of the estate it was expected to be seedcorn for. It would probably be good to store seedcorn centrally, not least because sowing the same plants’ seed in the same ground year on year causes crop deterioration so this would mix it up a bit, but also because it means there would be some help available if someone’s fields were washed out and so on. But I don’t see this storage as being for food at all.

So, in summary. These figures give us some idea of what was felt like a good quantity of food to keep in a royal store on a big royal estate. They also tell us that one use of those stores, apparently, was to act as a sort of seed-bank, though we don’t know for how much of the estate they were supposed to provide seedcorn. They do not tell us how much was originally harvested in either year, or how much of the harvest was dispersed before storage and inventory, though they do tell us that there was some so dispersed. And they certainly don’t tell us that the yield of the crops of the time is computable from these figures, and experiments elsewhere suggest that such computations are probably out by a factor of ten. So, OK. Let’s give Georges Duby his due for doing this kind of work at all, but admit that the texts don’t say what he read in them, and had been taught to read of course. Most of early medieval historiography for the last fifty years could be understood as arguments with George Duby in one way or another, after all, which is some tribute to his greatness. But the peasants of Carolingian Francia are unlikely, by any means we can test, to have been that awful as farmers, and the history of that empire makes a good deal more sense if there was a substantial surplus available and worth controlling. The old 2:1 figure can, I think, be safely disposed of.

---

1. Jo Ann Hoeppner Moran Cruz & Richard Gerberding, Medieval Worlds: an introduction to European history, 300-1492 (Boston 2004), pp. 162 (yields of between two to four times seed sown) & 223 (yield ratios of “1·5 to 2:1”).

2. Georges Duby, Guerriers et paysans, VII-XII^e siècle : premier essor de l’économie européenne (Paris 1973), transl. Howard B. Clarke as The early growth of the European economy: warriors and peasants from the seventh to the twelfth century (London 1974), pp. 26-29.

3. Georges Duby, L’économie rurale et la vie des campagnes dans l’occident médiéval (France, Angleterre, Empire, IX-XV siècles) (Paris 1964), 2 vols, transl. Cynthia Postan as Rural economy and country life in the medieval West (London 1968), pp. 25-26 of the which quoted here. His footnote references these works: “GRIERSON, P., ‘The Identity of the Unnamed Fiscs in the Brevium exempla ad describendas res ecclesiasticas et fiscales‘ in Revue belge de philologie et d’histoire [Vol. 18], 1939[, pp. 437-461]” and “SLICHER VON BATH, B. H., The Agrarian History of Western Europe (A.D. 500-1850) (trans. O. Ordish), London, 1963.” The Grierson article, which is characteristically good, finds one or two of the estates described in later charters, attempts to map the areas concerned, and in doing so discusses the yields and gives references to six older works also engaging with these figures, with similarly dismal and debated conclusions (Grierson, ‘The Identity of the Unnamed Fiscs’, yields discussion and references at pp. 452-456).

4. Duby, Rural Economy, ap. 2 (p. 364).

5. Alfred Boretius (ed.), Capitularia regum francorum, Monumenta Germaniae Historica (Legum sectio II: Capitularia regum francorum) I (Hannover 1883, repr. Berlin 1984), no. 128. The IMSB’s text is Frederic Austin Ogg (ed./transl.), A Source Book of Mediæval History (New York City 1908), pp. 127-129, spelt with the ash contra IMSB and citing Georg Heinrich Pertz (ed.), Monumenta Germaniae Historica (Legum) I (Hannover 1835, repr. Berlin 1991), pp. 176-181.

6. Peter J. Reynolds, “Mediaeval Cereal Yields in Catalonia & England: An Empirical Challenge” in Acta Mediaevalia Vol. 18 (Barcelona 1997), pp. 467-507.

7. Carmen Cubero i Corpas, Imma Ollich i Castanyer, Montserrat de Rocafiguera i Espona & Maria Ocaña i Subirana, “From the granary to the field; archaeobotany and experimental archaeology at l’Esquerda (Catalonia, Spain)” in Vegetation History and Archaeobotany Vol. 17 (New York 2008), pp. 85-92, online at http://www.springerlink.com/content/j418g4qt35038806/fulltext.html, last modified 19 June 2007 as of 4 January 2009, following table on p. 90 (Table 3).

8. Ibid. p. 90.

9. I assumed it was in Norman Pounds, An Economic History of Medieval Europe, 2nd edn. (London 1994), for which I have a lot of time as a textbook, but it turns out that he also repeats the Duby figures, p. 198.

10. Grierson, ‘The Identity of the Unnamed Fiscs’, p. 455, citing L. Halphen, Études critiques sur la règne de Charlemagne (Paris 1921), p. 252.