Imagine a building site in London in seventeen hundred. A mason shows up, looks at the sky, and goes home. Not because he's lazy. Because there's no light to lay stone by.
And nobody writes that down as a day off. It's just... not a working day.
That's the world Daniel wants us to look at. He wrote in this week, and he's coming at it from the office. He says for anyone who's worked in one, it's hard to imagine a time when nine to five wasn't the default, and even now, outside knowledge work, it's entrenched as a norm. So he wants a general overview of how working hours and leisure were governed before the Industrial Revolution, when people were doing very different things.
He's got a specific ask in there too.
He does. He wants to know what the equivalent of the United States was back then. A country that was relatively prosperous and influential. And in whatever territory maps onto that role, he wants three things. The approximate division of labour between the major sectors of the economy. Average working hours. And what annual vacation looked like for the people doing that work.
Three questions, and the third one is the trap.
The third one is the trap. Because annual vacation, as a concept, may not have existed at all. So where do we start with this?
With the territory, because Daniel's instinct is right. England is the answer. By seventeen hundred it's the most commercialized, most urbanized, highest-wage economy in Europe, and it's about to be the birthplace of the Industrial Revolution. Broadberry, Campbell and van Leeuwen put British GDP per capita at roughly eleven hundred dollars in the fourteen fifties, in nineteen ninety international prices, rising to about two thousand by eighteen hundred. That's roughly treble bare-bones subsistence.
Treble subsistence. So not rich by our standards, but not a subsistence economy either.
Not remotely. And the Netherlands is the other candidate. Holland, Antwerp, the Dutch Golden Age. But if you want the country that's prosperous, influential, and about to become a superpower, it's England.
So we've got our territory. Now the division of labour. What were people actually doing?
This is where the first surprise lands. The old assumption was that pre-industrial England was overwhelmingly agricultural. Seventy-five, eighty percent of the workforce on the land. And the modern reconstruction says that's wrong, and it was wrong earlier than anyone thought.
How early?
Thirteen eighty-one. The Poll Tax Returns. Thirty thousand two hundred and ninety-two individuals sampled. Agriculture is about sixty percent. And here's the part that surprises people. Services are slightly ahead of industry. So you've got a service sector, in the fourteenth century, bigger than manufacturing.
Sixty percent agriculture in thirteen eighty-one. That's a lot lower than the textbook picture.
And it stays roughly there. Fifteen twenty-two, the Muster Rolls. Still about sixty percent agriculture. But now industry has edged ahead of services. Then you get into the social tables. Gregory King, around seventeen hundred. Agriculture is down to forty-three, forty-four percent of the male labour force. Under forty percent of the total labour force.
That's a big drop from sixty.
It is. And it keeps going. Seventeen fifty-nine, Joseph Massie's table. Agriculture at thirty-seven percent, industry at thirty-four. Eighteen oh one, Colquhoun. Agriculture just over thirty percent. By eighteen fifty-one, agriculture is about twenty-five percent of the adult male labour force, industry is forty-five, services twenty-four.
So over four and a half centuries, agriculture goes from sixty percent to a quarter. That's the structural transformation, and it's happening well before the factories show up.
Well before. Which is one of the reasons the Industrial Revolution story is messier than the version you get in school.
Now. You said the modern reconstruction. Which implies there's an older one, and someone disagreeing with it.
There is. Gregory Clark and co-authors, twenty twelve, using probate and wills data. They claim agriculture held steady near sixty percent right through the sixteenth and seventeenth centuries, and the big shift out of farming came late. Which directly contradicts Broadberry and Shaw-Taylor.
Same country, same centuries, two completely different pictures of what the economy looked like.
And I'll be honest, I don't think that one gets resolved by us. The wills data and the social tables are measuring different things in different ways, and both camps have defended their method. What I'd say is that the direction of travel is not in dispute. Everyone agrees agriculture's share falls. The fight is about when it falls and how fast.
Which is a pattern we're going to see again in about ten minutes, with the working days.
Exactly the same pattern. Contested numbers, agreed direction.
So that's who was working. Now, how was the year structured? And this is where it stops looking like anything we'd recognize.
Work was profoundly seasonal. Judy Stephenson's work on the St Paul's Cathedral records, seventeen hundred to seventeen ten, is the best window into this. One mason contractor's team worked about forty-five percent fewer days in January to March than they did in June to August. The cathedral's own labourers worked half the days in January that they did in July.
Half.
Half. And the reason is mundane. Lack of light. You can't lay stone in the dark, and candles cost money. Little happened in Britain in January and February. Full teams were back on building sites by early March.
What about other trades?
Shipbuilding crews swelled by a third or more across the year. Dock work followed the shipping seasons. And this is the thing that makes the whole picture cohere. Most manufacturing and service workers were paid by task or by piece, not by the day. So employers kept no day-records. There was no reason to.
Because you weren't buying a day of a man's time. You were buying a finished job.
You were buying the job. And Stephenson says the early modern economy was very much like a gig-economy. Her words. Zero-hours contracts, site-to-site employment, paid by the task. That's not an analogy she's reaching for. That's her description of what the records show.
So the modern gig worker and the seventeenth-century mason have more in common with each other than either has with a nineteen fifties factory hand.
Considerably more. And it explains why the hours debate is so hard to settle. If nobody's keeping day-records, you can't just go and count the days.
Which brings us to time off. Before there's formal leave, what governs it?
Two things. Holy days, and Saint Monday. And the holy days story is striking. Allen and Weisdorf worked out that the removal of forty-nine holy days in England in fifteen thirty-six, the Protestant Reformation, coincides with a steep rise in required working days. From about one hundred and eighty to about three hundred and five, between fifteen thirty-six and sixteen sixteen. That's a seventy percent expansion in eighty years.
They abolished the holidays and the working year nearly doubled.
Not doubled, but expanded by seventy percent. And the Netherlands follows in fifteen seventy-four, France in sixteen sixty-six, Austria in seventeen fifty-four. This is a Europe-wide pattern, and it tracks the Reformation almost exactly.
So the Reformation is, among other things, a labour supply intervention.
It's hard to look at that timing and not see it that way. You strip out the saints' days, you've just added a large number of working days to the year, and you've done it by decree.
And Saint Monday?
Taking Monday off. Common among craft workers since at least the seventeenth century. The logic is straightforward. Pay day was typically Saturday, so on Monday you had money in your pocket and no particular reason to be at the bench. E.P. Thompson's nineteen sixty-seven essay called it universally observed until the beginning of the nineteenth century.
And Benjamin Franklin.
Franklin, in his Autobiography, writing about his London printing-house days. "My constant attendance, I never making a St. Monday, recommended me to the master." He's telling you he got ahead by being the man who showed up on Monday.
Which only works as a boast if everyone else wasn't showing up.
That's the whole joke. He's singling himself out for a virtue that only reads as a virtue against a background where Monday off was normal.
Now. Douglas Reid traced the decline of Saint Monday.
Seventeen sixty-six to eighteen seventy-six. And it survived longest among better-off, self-employed workers who controlled their own hours. Which tells you something about who gets to keep their customs. Reid's line is that when Black Country workers submitted to the norms of industrial capitalism by giving up their Saint Monday, the notion of a proper balance between work and leisure was lost.
The custom wasn't replaced by a better custom. It was replaced by nothing, and then people spent a century trying to invent a replacement.
That's roughly how Reid reads it.
So we've got the shape of the economy. We've got the seasonal structure. Now we get to the fight. How many days a year did these people actually work?
And this is a full-on scholarly brawl. Stephenson, writing for the Economic History Society, lays out the current estimates. People worked about two hundred and seventy days a year by seventeen hundred, rising to about three hundred after seventeen fifty.
Two hundred and seventy. That's a five and a half day week, fifty weeks a year.
Which is more than a lot of modern workers put in. And then Stephenson turns around and argues those estimates are too high. Her own study of the St Paul's masons' actual working records found men worked fewer than one hundred and fifty days a year for that employer. And she argues two hundred to two hundred and twenty days a year is the plausible ceiling.
One hundred and fifty versus two hundred and seventy. That's not a rounding error.
It's nearly a factor of two. And the average days worked per week in her records was five point two. Not because of Saint Mondays, which she found almost indiscernible in the St Paul's data, but because of what she calls idiosyncratic breaks.
Hold on. Saint Monday is supposed to be the famous one, and it doesn't show up?
It doesn't show up in those records. Which is a genuine problem for the Thompson picture. Though I'd flag that a single employer's site records from one decade in one city is a narrow window. It's possible Saint Monday was real and simply wasn't how masons at St Paul's behaved.
And Stephenson quotes John Hatcher on the consensus figure.
She does. Hatcher's view is that the idea anyone had two hundred and seventy days of paid work a year before the Industrial Revolution is fanciful.
Fanciful. That's the word. So we've got one camp saying two hundred and seventy to three hundred, another saying one hundred and fifty to two hundred and twenty. And then there's Voth.
Voth is the counter-evidence, and his method is completely different. He used Old Bailey and Northern Assize witness testimony. Court records, where people state what time of day it was and what they were doing. And from that he reconstructs working hours.
Court testimony as labour data. That's clever.
It's very clever, because it's the one place ordinary people had a reason to state the hour precisely. And what he finds is that Londoners' annual working hours rose by at least one fifth, with some estimates closer to forty-eight percent, between seventeen fifty and eighteen hundred.
What changed?
Not the length of the working day. The average working day started at seven in the morning and ended about twelve hours later, so roughly eleven hours of work after meals. That's stable. What changed is that Monday became a regular workday, and the religious and political holidays disappeared.
So the day was already long. They just added more of them.
They added more of them. And Voth's annual hours estimates are worth hearing, because they're big numbers. London, about two thousand and sixty-nine hours in seventeen sixty, rising to about two thousand seven hundred and thirty-eight by eighteen hundred. The broader English sample, about two thousand six hundred and six in seventeen sixty, up to about three thousand two hundred and twenty-seven by eighteen hundred, and three thousand three hundred and fifty-one by eighteen thirty.
Three thousand three hundred hours. That's sixty-four hours a week, every week, with no holiday.
That's the shape of it. And I should say, Voth's numbers and Stephenson's numbers are not measuring the same population. Voth's London is a big commercial city with a lot of casual and service labour. Stephenson's masons are a skilled building trade. Both can be right about their own people.
Which is the honest answer, and also the frustrating one.
It's the frustrating one. There's no single number.
Now, Allen and Weisdorf come at this from a completely different angle. Not how many days people worked, but how many days they needed to work.
That's the subsistence calculation, and it's the cleanest way into the whole problem. Late-medieval farm labourers needed less than two hundred days of work per year to provide for an entire family.
Two hundred days to feed a family for a year.
By the mid-fifteenth century it's better than that. Less than half a labourer's full capacity. Around one hundred and fifty working days per year.
One hundred and fifty days. That's a three-day week for most of the year. That's less than most people work now.
Considerably less. And then by eighteen eighteen the requirement spikes to about three hundred and ninety-one days. More than a year.
More days than exist.
More days than exist in the year. Which means one person's labour cannot cover the family. Which is why women and children go into the workforce. Allen and Weisdorf estimate women and children supplied about nineteen point three percent of low-income agricultural household earnings by eighteen hundred. Horrell and Humphries get eighteen point four percent for seventeen eighty-seven to eighteen fifteen. Two independent estimates landing in the same place.
So the family wage collapses, and the household patches it with more workers rather than more hours.
That's the mechanism. You can't add hours to a day that's already full. So you add people.
And that's the fifteen hundreds to the eighteen hundreds arc. One hundred and fifty days needed, then three hundred and ninety-one.
And the fifteen thirty-six holy days abolition sits right in the middle of that curve.
Which brings us to the thing I think Daniel's actually circling. If people are working more days after seventeen hundred, but real wages are stagnating, what happened to living standards?
This is the industrious revolution paradox, and it's the most interesting question in the whole literature. Jan de Vries named the phenomenon. People working more, consuming more, but the consumption is being bought with time rather than with higher pay.
So the extra consumption isn't a raise. It's a purchase.
And Voth's conclusion is blunt. Moderate gains in per capita consumption during the Industrial Revolution have to be balanced against this decline in leisure. And when he adjusts consumption per capita for lost leisure, it's essentially unchanged between seventeen sixty and eighteen thirty.
Seventy years of industrialization, and the average person's standard of living, once you count the hours they gave up, is flat.
That's what his numbers say. And it reframes the whole Industrial Revolution story. The gains were real in aggregate. Output went up. But a large part of what looks like rising living standards is people working more, not people being paid more for the same work.
The factory didn't just change where you worked. It bought your leisure from you.
It bought your leisure from you, and it paid for it in goods you hadn't had before. Whether that's a good trade is a question the numbers can't answer.
Now, Daniel asked specifically about annual vacation. So let's answer it.
The honest answer is that there's no evidence of formal paid annual leave. None. What existed instead was four things. Holy days, before fifteen thirty-six. Saint Mondays, where they were observed. Seasonal lulls, particularly January and February. And idiosyncratic breaks, which is Stephenson's phrase for the gaps that don't fit any pattern.
So vacation existed as a category of time, but not as a thing you were granted.
That's the cleanest way to say it. It wasn't employer-granted leave. It was customary holiday and seasonal downtime. The idea that you accumulate days and then take them is a much later invention, and it comes with the factory, because the factory is the first workplace where the employer needs to know exactly which days you were there.
Because the machine runs whether you're at it or not.
That's the whole shift. When you're paid by the task, your absence is your own business. When you're paid by the day at a machine, your absence is a line in a ledger.
There's one more thing in the research I want to get to, because it complicates the picture in a way I like. The fifteenth century.
The golden age of labour.
It gets described as a golden age. High real wages, short working years, around one hundred and fifty days. And it also gets described as an era of idle labour, of surplus workers, of underemployment.
Both descriptions are in the literature, and they're not compatible. Christopher Dyer's reconstruction is the one I find most convincing. He says a plausible reading of workers' attitudes between thirteen forty-nine and fifteen twenty is that they set themselves goals in cash or consumption needs, and worked until they'd achieved their aims. Then they ceased to work.
They stopped when they had enough.
Which is a completely different relationship to work than anything we'd recognize. It's not a forty-hour week. It's a target.
And the disagreement is whether that's prosperity or whether it's a labour market with nowhere to put people.
It's whether short hours are a choice or a symptom. If you can earn your year's needs in a hundred and fifty days, that's prosperity. If you can only find a hundred and fifty days of work because nobody's hiring, that's underemployment, and it looks identical in the wage data.
Same number, opposite meaning.
And I don't think the sources settle it.
So let's take stock, because Daniel asked three questions and we should be honest about what we can answer. Division of labour. Agriculture falls from about sixty percent in thirteen eighty-one to about twenty-five percent by eighteen fifty-one, with a serious dissenting view that says the fall came much later.
Working hours. Estimates range from about one hundred and fifty days a year to over three hundred, and the disagreement isn't a gap in the data. It's two different methods measuring two different populations.
Annual vacation. No formal leave. Customary holidays, Saint Mondays, seasonal lulls, and idiosyncratic breaks.
And the thing that ties it together is that the working year wasn't governed by a clock. It was governed by light, weather, saints, and custom.
Which is a very different world from the one Daniel's writing from.
Hilbert: A Casio F-91W. Eleven pounds ninety-nine.
Go on.
Hilbert: That's what the crew wore on the second site I worked, the one out past the bypass. The ganger handed them out. Nobody asked what time it was. You knew what time it was. What you didn't know was whether the pour was going to go off before the rain came, and that's what the watch was for. It had a stopwatch and a light and it cost less than a round.
So it wasn't telling you when to start.
Hilbert: It was telling you how long the mix had been down. That's the only number anybody cared about. You'd get there when the lorry got there. If the lorry was late you sat in the cabin and had a sandwich and nobody wrote anything down.
That's the task-orientation Stephenson describes. The clock is measuring the material, not the man.
Hilbert: The man's time was never the unit. I did a winter on that site where we had maybe forty days on. Not forty days each. Forty days where there was anything to do. The rest of it you found something else, or you didn't.
And nobody called that a holiday.
Hilbert: Nobody called it anything. You weren't on holiday. You were waiting. There's a difference and it matters, because a holiday is something you've got and waiting is something you haven't. My uncle did the same thing on the boats out of Grimsby and he'd say he was on the slack. Never said he was off. On the slack. There was a book, at the first place, the one by the river. The wages book. And the only thing written in it was what you'd done. So many yards. So many courses. If you did nothing there was nothing written, and a blank isn't a nought, it's just a blank. Nobody was counting the blanks.
That's the record-keeping problem in one image. The absence leaves no trace, so you can't count it.
Hilbert: You can't count it. And the men who kept that book are the same men who'd tell you the trade was going to the dogs, which they'd been saying since before I got there. I've still got the watch. My mother's got it, actually. It's in a drawer at hers.
And the point about the forty days. That's not leisure. That's just no work.
Hilbert: That's no work. January you're not resting, you're cold and you're eating into whatever you put by. The lads who talked about the good old days were talking about the good old days when there was work in them.
So the seasonal lull isn't vacation. It's the absence of a wage.
Hilbert: That's why I'd be careful with the word leisure. Leisure is when you've got the money and the time. On the slack you've got the time and no money, and that's a different thing entirely.
That reframes the whole one hundred and fifty days figure, actually. If the short year is partly involuntary, then the golden age reading gets a lot weaker.
It does. A hundred and fifty days of work because you only needed a hundred and fifty is one thing. A hundred and fifty days because that's all anyone would give you is another.
Hilbert's point about the blank in the book is the methodological version of the same problem. The records can't distinguish between the two, because the man who chose not to work and the man who couldn't find work leave the same blank.
Which means the historians aren't just missing data. They're missing the thing that would let them tell the two cases apart.
They're missing the intention. The wage book records the work and nothing about the reason.
Here's where I land on Daniel's question. The nine-to-five isn't the natural state of human labour that we drifted away from. It's a specific arrangement that came out of the factory, and the factory needed it because the machine runs whether you're at it or not.
The pre-industrial world wasn't a golden age of leisure either. It was a world where work was seasonal, task-based, and paid by the piece, where time off was customary rather than granted, and where a short year might mean prosperity or might mean hunger and nobody wrote down which.
The most common wrong belief here is that people before industrialization worked less because life was simpler.
The correction is that they worked in a completely different shape. Long days when there was work, no days when there wasn't, and no concept of leave at all. The total might be lower or higher than yours. It depends entirely on which workers you count and which decade you pick, and the honest answer is that the scholars don't agree.
One thing I keep coming back to. Voth's finding that consumption adjusted for lost leisure was essentially flat between seventeen sixty and eighteen thirty. Seventy years of the most consequential economic transformation in history, and the average person's material life, once you count the hours, didn't move.
It's a hard number to sit with.
It is. And it makes you wonder what our own version of that looks like. We've got the always-on work, the gig platforms, the notifications at ten at night. Whether we're being paid for that in money or in something else is a question someone will be asking in a hundred years, with better records than the masons kept.
If they keep records at all.
That's My Weird Prompts. Thanks to our producer, Hilbert Flumingtop. If you want to send us a prompt, email us at show at my weird prompts dot com. We'll be back soon.
See you tomorrow.