Arduinna

Posted by Bret Devereaux

https://acoup.blog/2025/09/05/collections-life-work-death-and-the-peasant-part-ivb-working-days/?utm_source=rss&utm_medium=rss&utm_campaign=collections-life-work-death-and-the-peasant-part-ivb-working-days

https://acoup.blog/?p=31214

This is the continuation – the first of several – of the fourth part of our series looking at the lives of pre-modern peasant farmers – a majority of all of the humans who have ever lived. Last time we discussed the survival requirements (in food and textiles) of a peasant household as well as what different levels of material comfort beyond just survival might look like.

This week, we’re going to take those figures and begin comparing them to production, modeling out our farmers and their ability to grow the food they need to survive and perhaps a bit extra to sell, trade or gift away to get other things they want. We’re going to split this into two parts: this week we’re going to model out farmers assuming they own effectively infinite land. Then next week we’re going to revise those assumptions in light of the very small farm sizes we actually see in our sources. And after that – because we’re not done – we’ll move to discussing other kinds of labor in the household, like food preparation, cleaning, textile production and so on, to get a really thorough look at household labor.

In particular, on one of the persistent myths I wanted to address in this series is the idea that modern workers work more than ancient or medieval peasants, something that we’ll see is simply not true. Finally, note that while we’re going to be modeling farming subsistence here, we’re not going to get into the gritty details of how that farming was done; if you want to read about that, we already have a series on it just for you!

But first, if you like what you are reading, please share it and if you really like it, you can support this project on Patreon! While I do teach as the academic equivalent of a tenant farmer, tilling the Big Man’s classes, this project is my little plot of freeheld land which enables me to keep working as a writers and scholar. And if you want updates whenever a new post appears, you can click below for email updates or follow me on Twitter and Bluesky and (less frequently) Mastodon (@bretdevereaux@historians.social) for updates when posts go live and my general musings; I have largely shifted over to Bluesky (I maintain some de minimis presence on Twitter), given that it has become a much better place for historical discussion than Twitter.

Works and Days

The origin point for nearly all of those “you work harder than a medieval peasant” memes and articles is Juliet Schor’s The Overworked American (1993). The argument has been debunked quite a few times, so I won’t belabor the point here. Schor bases her estimates of medieval working hours on a 1935 article by Nora Kenyon,¹ and an unpublished article by Gregory Clark,² and in both cases ignores the authors’ careful efforts to distinguish between total days worked and instead just cherry-picks the lowest number, even as the authors caution that those numbers likely don’t represent someone’s total employment. Kenyon notes a set of day-laborers working 120 days per year which makes it into Schor’s work, but Kenyon’s final suggestion that the normal annual working year was 308 days does not, for instance. I can’t get at an unpublished article, but Clark has continued to write on the topic and in his 2018 “Growth or stagnation?” presents a detailed argument for a 250-300-day work-year with no sense that this is a revision of his previous positions, leading me to suspect similar cherry-picking as with Kenyon.

In short, Schor’s works is quite shoddy and we shan’t rely on it.

Now part of the complication there is that for the European Middle Ages, across so much area, what we see is a lot of confusing evidence – statutory minimums, required labor on a lord’s land and so on – which may or may not represent a full working year. What we don’t typically get is someone just telling us how many work days were in the agricultural calendar. But as you may recall, we’re anchoring this discussion in the Roman world and in a rare instance where the ancient evidence is better, Roman agricultural writers just straight up tell us how many working days there were in a year on the Roman agricultural calendar: 290 (Columella 2.12.8-9). He allows 45 days for holidays as well as inclement weather and another 30 days for rest immediately after the crop is sown, to recover from the difficult labor of the final plowing.

The medieval work calendar is not meaningfully different. As noted above, Both Clark and Kenyon end up with similar working-day estimates from the medieval evidence as Columella’s figure. The medieval number is probably slightly lower: the medieval religious calendar might have around 45 feast days but workers might also be expected to spend Sundays in religious observance, which might pull the work-year down to around 270 total working days, plus or minus.

By all evidence, those working days were both less rigid but also longer than modern working hours. On the one hand, peasant farmers are essentially self-employed entrepreneurs, making their own hours. They can arrive in the field a bit late, sometimes leave a bit early. It was certainly common in warmer climates for workers to take a midday break (a siesta) to avoid exhausting themselves in the hottest part of the day. I will say, anyone who has done functionally any outside work in a warm climate will recognize that a midday break can allow you to work more than just pushing straight through the heat of the day because you tire more slowly.

So on the one hand the work hours are somewhat flexible. On the other hand as functionally anyone who has ever worked on a farm or spoken with someone who has will tell you, the working day in absolute terms is long, essentially starting at sunrise and running to sunset. And this is certainly the implication we get from our sources. Because of atmospheric refraction, there are actually slightly more than 12 sunlight hours per day on average (it’s around ~12.3 or so, depending on latitude), though this of course varies seasonally. The bad news for our farmers, of course, is that the shortest days are in the winter when the labor demands are lower. While festival calendars feature events throughout the year, it is not an accident that major festivals in a lot of pre-modern agrarian cultures are concentrated in late Fall, winter and early Spring. For the Christian calendar, that includes things like All Saints Day (Nov 1), Martinmas (Nov 11), the regular slew of December holidays as as the holidays of the Eastertide in early spring. For the Romans, you have major festivals like the Parentalia and Lupercalia in February, the Liberalia in March, the Cerialia in April and the Saturnalia in December.

So in practice the average maximum working day might actually be a bit longer than 12 hours, but we should account for breaks and general schedule flexibility. We might assume, for comparison, something like a ten hour work day. By that measure, our peasants probably put in somewhere between 2,500 and 3,000 working hours per year. By contrast, your average ‘overworked American’ has 260 working days a year, at eight hours a day for just 2,080 hours.³

So to answer the question: no, you do not work more than a medieval (or ancient) peasant (despite your labor buying a much higher standard of living). But there’s more complexity for us to draw out here in the structure of peasant labor, so we need to go back to our model and start working through some implications.

Farming Time

As I noted previously, we’re going to anchor our model in the Roman evidence because I know it best and here the Roman agricultural writers – the agronomists (Cato the Elder, Columella, Varro and Palladius) – equip us with a lot of useful information, especially Columella. These fellows are writing guidebooks on how to run large estates – how to be good at being the Big Man – but the information they include, especially Columella, helps illustrate a lot about the labor and subsistence structures involved. As we’ve noted above, Columella already computed a standard agricultural work-year. We could convert that back into hours, but we don’t need to, because Columella does all of his labor calculations in working days.

The temptation here is to run our model on a single crop – wheat or barley – but that is a mistake. Columella himself suggests an estate (again, he’s thinking big farms) split between a wide range of crops, with primary plantings in wheat, a variety of legumes, along with turnips (and other root vegetables), barley, and so on. Our peasants will almost certainly do the same. There are a few reasons for this variety.

First, even from antiquity, farmers were aware that some crops exhausted the soil more rapidly or differently than others. They don’t fully understand it, but even Columella notes that some improve the land and others ‘burn’ it. And he is basically correct: lupines, beans, vetch, lentils, peas and such he regards are improving the land, while things like wheat reduce its fertility and some crops (he notes flax) put substantial strain on it (Columella 2.13). Of course farmers need those demanding crops, so rotation is necessary and was practiced since antiquity.

Second these crops all have different planting and harvesting timings. (Winter) wheat is planted in late Fall, barley in early winter, chick-pea can be sown in January or February, sesame in October, and so on and so forth (Columella 2.10). That’s important because planting and harvest create huge peaks in labor demand and our farmers want to try to flatten those spikes out a bit. If you plan nothing but wheat, you’ll never have enough labor to get all of it into the ground and then harvested again during the ideal calendar windows to do so.

Third is the perspective of risk management. All of these crops are vulnerable to different problems. A dry year will savage your wheat crop but barley is less bothered by dry conditions. Crops tolerate early frosts, high heat or low heat and so on differently. Pests that afflict one kind of crop may not afflict others. So by splitting your fields between different crops, you reduce the risk that any one problem will wipe you out. Remember that our peasants are not looking to maximize profit, but to minimize risk.

So our farmers are likely going to rotate a number of different crops. Now as to crop rotation, we often teach a fairly simple story of technological advancement from ancient two-field rotation systems (with half the land fallow) to medieval three-field systems to early modern four-field states (with the fallow largely replaced by fodder and grazing crops). And that description is more or less true but as always complexity abounds on closer inspection. As Pliny the Elder notes, the true maxim of farming was quid quaeque regio patiatur – “whatever the region permits.” For the Romans, we find attestations of two-field, three-field and continuous cropping systems where, in the latter case, extensive manuring was used to keep land under continuous cultivation,⁴ all depending on the local conditions: the richness of the soil, the availability of water, the local value of crops (and thus the affordability of manure) and so on.

For the sake of simplicity, we can think with three crops, wheat, barley and some legumes (in this case, vicia faba, the broad bean, for instance), though we also have to remember that about a third of our fields will be fallow in any given year. The legumes here are actually pretty important (and Columella seems to think a wheat-focused farming operation would sow wheat and legumes in even quantities, even while fallowing some of the fields, Columella 2.12.7-8) because they are nitrogen-fixing (technically, they have nitrogen-fixing bacteria) and so serve to maintain the fertility of the soil.

Different crops, of course, will have different productivity, demand different amounts of labor and so on. And here, as a reminder, since I am leaning on Columella, my background calculations are taking place in Roman units: modii (8.73 dry liters) and iugera (0.623 acres).

Wheat, Columella reports, was sown 5 modii to the iugerum (that is, it takes five modii – c. 43.5 liters or c. 1.2 bushels – to provide enough seed for 1 iugerum (0.623 acres) of farmland), and requires 10.5 days of labor. Columella (2.9) has barley sown 5 modii to the iugerum but Varro (1.44.1) suggests 6 modii to the iugerum; barley being more tolerant of bad conditions requires according to Columella only 6.5 days of labor for 5 modii for one iugerum. For beans, Columella says between 4 and six modii to the iugerum with 7 or 8 days of labor. That said, Columella’s labor-time estimates have left a number of things out – particularly threshing – and has probably somewhat underestimated plowing time⁵ so we need to account for that working time too. M.S. Spurr figures the wheat figure should be 14.25 days per iugerum, while Rosenstein estimates 19.5 days when accounting for the missing tasks, though note that we have not included a lot of background maintenance like maintaining tools or structures – this is purely the work for individual crops in individual fields.⁶ If we apply a similar under-count-adjustment to the labor requirements for barley and beans, we might come to a seed-and-labor-inputs estimate that looks like this:

Now we have to think about how much labor our families have to throw at this problem. You will recall that last time we proposed three sample families, the Smalls, the Middles and the Biggs. How much farming labor do they have?

Labor patterns in these households were gendered, but not infinitely so. As Paul Erdkamp notes, in Roman artwork – and in my experience this pattern continues in medieval artwork – we do see women doing farming labor, but typically only in periods of peak labor demand (like the harvest, which has to be done relatively rapidly) or in households where some sort of misfortune has resulted in severe labor shortages.⁷ So for the sake of calculating the farming labor ‘backbone’ we may – for now – exclude the women of the household, though I want to be clear that women did farming labor when necessary and absolutely were not going to sit around starving to death if all of the men were gone. That said, as we’ll see in subsequent parts of this series, the women of the household were by no means idle: there was a ton of necessary work beyond farming required to sustain the household and they’re doing most of it.

Columella’s labor calculations are for large estates utilizing slaves or paid workmen and so assume fully fit adult males, but our actual peasant households are more varied than that. We ought to assume that each adult male (none of our model families has any very old men, so we don’t need to factor for that) is supplying a full unit of labor, 290 working days per year. Children under 6 or 7 or so are not going to be performing the main labor tasks, but we might figure that males in their late teens (16 and older) are providing something like three-quarters the labor-power of a fully grown adult and younger sons (10-15 or so) perhaps half as much. Based on those assumptions, our labor ‘backbone’ (which, again, would be supplemented by the women and girls of the household when needs be) looks like this:

Assuming then that land is no object (which it obviously is, but that’s next time’s problem) and a roughly even split between wheat, barley and beans, we might suppose totals for land under cultivation for each family very roughly like this (trying to get reasonably close to maximum labor employment without going over):

Now, I see you there in the back, your hand already shot straight up because these farming areas are way, wildly bigger than what we’ve said typical peasant farms look like and yes, that is true. We’ll see how land scarcity factors in the next part, which is why I want to reiterate that this week’s analysis is not complete in itself for the obvious reason that very few peasants have unencumbered ownership of anything close to farms this large. Even a farm of 49 iugera would mark a household out to be very rich peasants. Nevertheless, we have to establish a baseline somewhere and this is a reasonable spot to do it.

Our next question has to be what these farmers might expect to get out of all of that work.

Farming Yields

As a rule, farming yields in the pre-modern are discussed not in terms of productivity per-land-area but rather in seed yields: for a given dry measure of seeds planted, how many of the same dry measure of seeds (because those are the tasty, edible parts of these plants) do you get back? So they are expressed in figures like 4:1 which means for every one modius/dry liter/bushel sown, four are harvested.

Yields were extremely variable, both season to season and region to region and our evidence for historic yields is often frustratingly limited or difficult. This is complicated by the fact that we cannot use modern farming yields to estimate, because hundreds or thousands of years of selective breeding have come to mean that modern crops are not identical to their ancient forebears and often have substantially higher yields, even if you used ancient farming techniques. As Theophrastus notes, ἒτος φέρει, οὒτι ἂρουρα, “the year bears [the harvest], not the field” (Theophr. Caus. pl. 3.23.5) by which he means seasonal variation was greater than regional variation: a bad year on excellent farmland was often worse than a good year on marginal land. That extreme level of variability makes charting an ‘average’ difficult. That problem is further intensified by the fact that our sources for antiquity often distort reported yields for rhetorical purposes – suggesting unreasonably low yields for crops they do not favor, or reporting absurdly high yields to simply the richness of specific regions.⁸

The best compilation of the evidence for ancient yields, which includes some comparative evidence for early modern and medieval crop yields, is in P. Erdkamp, The Grain Market in the Roman Empire (2005), 34-54.⁹ Yields on grains (like wheat and barley) might vary a lot, from as low as 3:1 on poor land in bad years to as high as 12:1 or more on good land in good conditions. Regional variability here is substantial: Spurr notes that in medieval Italy, hilly, marginal lands often yielded 3:1 or 4:1, while more typical flatter farmland might yield 5:1 or 6:1, but Sicily – with unusually good farmland – seems to have often yielded between 8:1 and 10:1. The general range for these yields is fairly consistent through the pre-modern evidence, improving modestly over time (so we might expect significantly, but not radically, higher yields for a peasant in 1500AD as compared to 1500 BC).

For our farmers, we probably ought to pick a pretty modest yield: our peasants probably don’t have the very best land available (the Big Man will have tried to get control over that) and also don’t have unlimited access to things like manure to really push yields at the upper end. On the flipside, our peasants are probably not on entirely marginal land (rocky ground, hills and so on). So we might propose something like a range of 4:1 to 8:1 to get a sense of the range from a bad year (4:1 yield) to a good year (8:1). For what it is worth, regions with very high productivity don’t tend to necessarily have richer peasants – they tend instead to have higher taxes.

Now of course some seed must be held back from each harvest to provide the seed for the next planting, but our yield ratios neatly contain this information. So while at a 4:1 yield, four modii/liters/bushels are harvested, one of those goes straight back into the ground, so the net yield is 3 units of whatever dry measure we’re using; at 8:1, the net yield is 7 units. In this case that works out to the following productivity per iugerum:

With that in hand, we can loop back to our chart above and calculate the range of net harvest after removing seed for next year that our model families might expect from their farming listed above.

That’s a lot of modii. But of coruse now we have another problem to account for: the modius is a dry measure. Pre-modern farmers mostly reckoned in dry measures because it was easy to measure but it is awkward for us because these crops, once harvested and put in sacks for storage, do not have the same density (that is, mass per unit volume) or calorie density (that is, calories per unit mass) as each other. So we need some way to convert these figures to our subsistence measure we developed last time which was kilograms-of-wheat-equivalent.

For wheat that’s relatively easy: threshed wheat has a density of roughly 6.72kg per modius (about 770 kg/m³), so a modius of wheat is 6.72kg of wheat equivalent. For the other two, we need to convert from a dry measure to a density to a calorie value to convert back to wheat equivalent. Barley is a little less dense than wheat, roughly 6.465kg per modius (740 kg/m³) but much less calorie dense – just ~2,160 calories per kilogram compared to wheat’s 3,340.¹⁰ So a modius of barley has roughly 13,960 calories in it, making a modius of barley just 4.17kg of wheat equivalent. A modius of beans (vicia faba) is about 5.43kg and contains about 18,842 calories, making that modius 5.64kg of wheat equivalent.

That neat exercise should also tell us something about farming strategies. A single iugerum, planted with wheat, yields (net after seed) between 100 and 235kg of wheat equivalent. Planted with barley, it takes much less labor and is more tolerant of bad (particularly dry) weather, but yields only between 75 and 175kg of wheat equivalent. Planted with beans, it consumes an intermediate amount of labor, helps the soil recover and provides unique and necessary nutrition – man cannot, as a matter of biology, live on bread alone – but provides only 68 to 158kg of wheat equivalent. A farm that finds itself strained – especially if the limit is land and not labor -might focus more and more heavily on barley (if it is very dry) or especially wheat at the expense other crops in order to maximize the yield per land area (which in turn means employing more labor). Keep that in mind for next time when we start factoring in land scarcity.

However for now, let’s head back to our tables and now factor our yield ranges into wheat equivalents to a get sense of how they stack up against our subsistence requirements (I’m rounding some of these figures off, so there may be some imprecision in the table):

If we compare to the subsistence and respectability needs of our households, we can make a few observations. First, given maximum labor employment and no land scarcity, even in modestly bad years each family clears its subsistence needs (though only the Smalls clear their respectability needs). In something like an ‘average’ year, the Smalls produce around 187% of their respectability needs, the Middles 155% and the Biggs 150%.

If labor and not land was the limiting factor in peasant agriculture, we ought to expect our peasants to live quite well. Of course even a casual glance at the first post in this series will warn against jumping to that conclusion. After all, by ignoring – so far – land scarcity, we’ve put our families on enormous farms by pre-modern standards, between 30 and 60 acres, more or less. But we know from the evidence that while our families might have the ability to farm 30-60 acres, the typical size of an actual smallholder farm was closer to 3-6 acres than 30-60; a farm of even something like 15-25 acres might mark a family out as ‘rich’ peasants. And above we can see why: a family on 20 or 30 acres probably has enough land to get close to or even reach its respectability basket without engaging in any kind of tenant or wage labor. Instead, that family may have so much land it can afford to rent out what it does not farm itself.

What we have done here so far is essentially simulated very rich peasants, which is well enough but as we’ve seen, rich peasants represented only a fairly tiny minority of the peasantry. In practice, households with as much land as above would be likely to begin repurposing some of it for things like livestock, vineyards or orchards – things with a lower per-acre calorie yield but which might provide greater food variety or market value. As you can tell from looking at the relative balance of labor- and land-intensity for crops, the “mostly grains” strategy is going to be a direct response to land scarcity rather than abundance.

Rather, as we’ll see, most families will have nowhere near enough land to match either their labor or their subsistence demands, which in turn will provide some of the wedges that the Big Men and the broader society will use to try to turn that ‘surplus’ labor to their own ends.

And that, of course sets up where we must go to next: how this model changes – and goodness, does it change – once we start thinking about land scarcity and tenant farming.

https://acoup.blog/2025/09/05/collections-life-work-death-and-the-peasant-part-ivb-working-days/?utm_source=rss&utm_medium=rss&utm_campaign=collections-life-work-death-and-the-peasant-part-ivb-working-days

https://acoup.blog/?p=31214