Four point four billion tons. That's how much Portland cement was manufactured in twenty twenty-three. It's third by mass of all manufactured materials on Earth, behind only sand and gravel and water. Concrete is the single most-manufactured material in human civilization, and I'd wager most people couldn't tell you when or why it took over from stone.
Daniel's been thinking about exactly that. Here's what he sent us — and he's been talking to Hannah Schneiderman about it, which is always a good sign for a prompt. He says: we've talked about plastic, which is everywhere and impossible to imagine a world without — but the "plastic is bad" line oversimplifies how hard it would be to run modern civilization without it. Now let's talk about another material that's everywhere: buildings, and what buildings are made from.
Mm.
Hannah tells him cement and rebar wasn't always the way people built — which, walk through the Old City of Jerusalem and look at those gigantic stones, that checks out. His question to Hannah, which he figures we can explain in more idiot-friendly terms than she did, was: when did that shift happen? And what he understood from her is there's way more to this than a simple timeline. It wasn't that everyone used stone and then one day concrete was discovered and everybody just stopped. The Romans were doing extraordinary things with concrete way before anybody else — and this is the part that sounds fascinating: people don't fully understand, even today, how their concrete was cured and how it was so strong. So before we get into the specifics, which we'll keep to future episodes, let's talk today about the historical chronology. Stone construction was the norm, but we moved over to concrete and rebar predominantly at a certain time. When was that current moment? And to presage episodes about how stone construction could be revitalized — have there been not pockets of revival, but pockets of continuity where traditions of working with stone have managed to continue through to the present, either uninterrupted or with periodic breaks?
So today we're going to build a timeline — not just of when concrete took over, but of why that takeover happened when it did, and what we lost along the way.
And Daniel's framing through Hannah is interesting here because she's an architect — she lives this. The question isn't just historical trivia. It's about understanding how we got locked into a material, and whether we can unlock ourselves.
Right. And the core tension is this: stone construction was the norm for millennia. The shift to concrete wasn't a clean break. The Romans were using sophisticated concrete from about one fifty BC onward — the Pantheon's dome, completed in one twenty-eight CE, is still the largest unreinforced concrete dome in the world. Meanwhile stone remained dominant for monumental construction right through the Middle Ages. So the timeline is messier than "stone, then concrete." It's stone, then concrete, then stone again, then concrete forever — with the real inflection point arriving surprisingly late.
So what's the arc today?
Three parts. First, the Roman concrete paradox — how a two-thousand-year-old material outperforms modern concrete in durability, and what we only just figured out about why. Second, the nineteenth-century convergence of Portland cement, rebar, and a devastating earthquake that made concrete the default. Third, the surprising survival of stone traditions and what they mean for a possible revival. The stakes are real — concrete has a massive carbon footprint, and understanding how we got locked in is the first step to asking whether we can break out.
Let's start with the mystery that Daniel flagged — Roman concrete. Because the Romans were doing things with concrete that we still can't fully explain.
And the numbers are wild. Roman concrete — opus caementicium — had compressive strength similar to modern Portland-cement concrete. About twenty megapascals, twenty-eight hundred PSI. Comparable. But modern concrete structures often start deteriorating within decades. Roman structures are still standing after two thousand years. The Pantheon's dome is the obvious poster child — completed in one twenty-eight CE, largest unreinforced concrete dome ever built, still there. But it's not just that it's still standing. It's that it's been standing in an active seismic zone for two millennia and the dome hasn't collapsed.
I want to pause on that seismic zone point for a second, because I think it's easy to hear "two thousand years" and just file it under "old things are impressive." But we're not talking about a structure in a geologically quiet area. Rome sits on a fault line. That dome has been shaken — repeatedly — by earthquakes that have leveled far newer buildings. And it's still there.
The Pantheon has survived multiple significant seismic events. The earthquake of eight forty-seven caused widespread damage across Rome — the Pantheon stood. The earthquake of thirteen forty-nine toppled parts of the Colosseum — the Pantheon stood. The nineteen fifteen Avezzano earthquake, magnitude seven, killed over thirty thousand people and damaged buildings across central Italy — the Pantheon stood. Every time the ground moved, that dome absorbed forces it was never explicitly engineered to handle, and it didn't crack in ways that propagated. And we now know why — but it took us until twenty twenty-three to figure it out.
So what did they know that we didn't?
This is where the twenty twenty-three MIT discovery comes in, and it's one of those findings that makes you rethink everything. For decades, researchers looking at Roman concrete noticed these little white lumps — lime clasts — distributed through the material. The assumption was that these were signs of sloppy mixing. Poor quality control. The Romans didn't mix their concrete properly, and these lime chunks were the evidence.
And that turned out to be completely backwards.
Completely. The MIT team, led by Linda Seymour, published in Science Advances in January twenty twenty-three — they showed that those lime clasts were intentional. The Romans used a hot mixing technique with quicklime instead of slaked lime. When you mix quicklime directly into the concrete, it creates these reactive calcium deposits. And here's the genius part: when cracks form in the concrete and water seeps in, those lime clasts react with the water to produce calcium carbonate crystals that reseal the cracks. The concrete heals itself.
Wait. The concrete fixes its own cracks.
It fixes its own cracks. For two thousand years. And we spent decades looking at the mechanism that makes this possible and thought it was a mistake.
That is... the most elegant engineering insult I've ever heard. We called their quality control bad because we didn't understand what quality control meant.
And it gets better. Roman marine concrete — the stuff they used for harbors and seawalls — is even more extraordinary. Seawater percolating through tiny cracks reacts with volcanic rock in the mix to form aluminous tobermorite crystals. These crystals actually strengthen the concrete over time. The Washington Post in twenty seventeen called it "the most durable building material in human history." Modern concrete exposed to saltwater? Deteriorates within decades. Roman seawalls? Still there, still getting stronger.
So they had self-healing concrete that gets tougher in seawater, and we have... the stuff that crumbles under highway overpasses after forty years.
That's not entirely fair to modern concrete, but — yeah, the comparison isn't flattering. And the Pantheon dome demonstrates another layer of lost sophistication. The aggregate in the concrete isn't uniform. At the base, they used heavy travertine — density about twenty-two hundred kilograms per cubic meter. As the dome rises, the aggregate transitions to lighter materials — tuff, then pumice at the very top, density around thirteen fifty. They engineered a graded-density design that reduces weight while maintaining strength. That kind of material optimization was lost for centuries.
I want to understand what that actually means in practice. If they'd used the same heavy aggregate all the way up, what would have happened?
The dome would have been substantially heavier at the top, which means the walls supporting it would have needed to be much thicker to handle the outward thrust. You'd lose the soaring interior space. Or worse — the dome collapses under its own weight. By lightening the material as they went up, they reduced the structural demands on the lower sections. It's the same principle behind modern skyscraper design — lighter materials as you go up — but they figured it out with empirical knowledge two thousand years before anyone wrote a structural engineering textbook.
So how much was actually lost? Because the narrative is usually "Rome fell, concrete knowledge vanished, rediscovered in the eighteenth century." But that can't be the whole story.
It's not. The "lost art" narrative is overstated. Concrete quality definitely degraded in many regions after Rome fell — Britain had low kiln temperatures, no access to pozzolana, poor mixing. But concrete never fully disappeared. The Anglo-Saxon crypt at Hexham Abbey, built in the seventh century, used concrete — not as sophisticated as Roman work, but recognizably the same tradition. Quality started improving again from the eleventh century. And just last year, in twenty twenty-five, a construction site from seventy-nine CE was uncovered at Pompeii — published in Nature Communications — containing unmixed Roman concrete materials. Direct evidence of their mixing techniques, frozen in time by Vesuvius.
A construction site buried mid-pour.
Mid-prep, really. The materials were laid out, ready to be mixed. You can see the piles of lime, the volcanic aggregate, the proportions they were about to combine. It's the closest thing we have to a Roman concrete recipe card.
So the knowledge degraded but didn't vanish, and we're still piecing together exactly what they were doing. Which makes the next part of the story weirder — if the Romans had this incredible concrete technology, why didn't it stick? Why did it take until the late nineteenth century for concrete to become dominant?
The answer involves three innovations that converged in about a thirty-year window, and none of them was Roman.
Walk me through them.
Innovation one: Portland cement. Joseph Aspdin patents it in eighteen twenty-four — names it because it resembles Portland stone, the fancy building material of the day. But the patent was for a relatively weak product. His son William Aspdin is the one who really perfected modern Portland cement in the eighteen forties. By the eighteen fifties, Portland cement had largely replaced the earlier Roman cement formulations. And then Frederick Ransome patents the rotary kiln in eighteen eighty-five to eighteen eighty-six — that's what enables mass production of consistent, strong cement. First US production at Coplay Cement Company in Pennsylvania, eighteen seventy-five.
So you've got cheap, consistent cement. That's step one.
Step one. Innovation two: reinforced concrete. And this is where the history gets messy because multiple people invented it independently. François Coignet built the first iron-reinforced concrete structure — a four-story house in Paris, eighteen fifty-three to eighteen fifty-five. William Wilkinson reinforced a concrete roof and floors in England in eighteen fifty-four, demonstrating he understood tensile stresses. Joseph Monier patented wire-reinforced flowerpots in eighteen sixty-seven — that's often cited as the starting point, though he wasn't first. He later patented columns and girders using iron rods.
Flowerpots.
Flowerpots. The material that would build skyscrapers started with a gardener trying to make stronger pots.
That's somehow perfect. And step three?
The nineteen-oh-six San Francisco earthquake. This was the watershed. Julia Morgan's reinforced concrete bell tower at Mills College survived undamaged while stone and brick buildings collapsed around it. That single structure dramatically shifted public and regulatory acceptance. San Francisco changed its building codes in nineteen-oh-eight to allow wider use of reinforced concrete. The Ingalls Building in Cincinnati had already demonstrated the potential — sixteen stories, one of the first reinforced concrete skyscrapers, built in nineteen-oh-four. But the earthquake proved the material wasn't just novel. It was safer.
I want to sit with the Mills College bell tower for a moment because it's the kind of detail that sounds minor but actually changed everything. Julia Morgan was the first woman licensed as an architect in California. She designed that tower, it survived a catastrophe that killed thousands, and suddenly everyone wanted to know what she'd built it from. It's not just an engineering milestone — it's a reputational earthquake in its own right.
And Morgan understood exactly what she'd done. She went on to design over seven hundred buildings, many in reinforced concrete, including Hearst Castle. The Mills tower wasn't an accident — she'd been advocating for reinforced concrete and the earthquake gave her the proof she needed. Sometimes a disaster doesn't just change building codes. It changes who gets listened to.
So the convergence is: cheap cement from the rotary kiln, rebar solving the tension problem, and an earthquake proving it all worked. That's what tipped the world from stone to concrete.
And the reasons it tipped so completely are worth unpacking. Fire resistance was huge — early reinforced concrete pioneers emphasized that it was fireproof compared to iron-framed buildings, and after events like the Great Chicago Fire of eighteen seventy-one, that argument carried enormous weight with insurers and city governments. Cost and speed: Portland cement used abundant limestone and shale, dirt cheap, and you could cast it into any shape. No skilled stone cutting needed. Tensile strength: stone is strong in compression but weak in tension — rebar solved that, enabling longer spans, thinner structures, skyscrapers. And industrialization meant the rotary kiln made cement available everywhere, consistent quality.
Concrete won because it was cheaper, faster, safer from fire, and you didn't need a master mason to pour it.
Once the system was in place — the kilns, the rebar mills, the building codes, the engineering curricula — it locked in. Path dependence. You don't switch back from concrete when every structural engineer is trained in it, every building code assumes it, and every supply chain delivers it.
Which brings us to Daniel's second question. Stone never really went away, did it?
No, and this is the part that surprises people. There are genuine pockets of continuity — not revival, continuity — where stone construction traditions kept going, adapted, and in some cases are now cutting-edge again.
Give me the best example.
Fernand Pouillon. French architect, post-World War Two. He developed a method called pierre de taille — massive precut stone. Precisely machine-cut stone blocks assembled by crane. This wasn't decorative stone cladding over a concrete frame. This was load-bearing stone construction, adapted for modern speed and cost. He built housing projects across France and Algeria this way — thousands of units, all in load-bearing stone. Genuine continuity of the stone tradition, industrialized.
He skipped the concrete step entirely.
For the structure, yes. And the tradition kept evolving. Modern tensioned stone uses steel tendons threaded through drilled stone blocks — post-tensioning, like we do with concrete, but with stone. The Southwark Gateway Needle in London is an example — a slender stone spire that wouldn't be possible without internal tensioning. Digital stereotomy — using CAD and computer modeling to design complex stone vaults and arches with precision cutting. Designers like Giuseppe Fallacara and Philippe Block are doing things with stone that medieval master masons would recognize immediately, but with tolerances they couldn't have dreamed of.
People are actually building with this now?
Fifteen Clerkenwell Close in London. Contemporary building using a trabeated stone exoskeleton — post-and-lintel, load-bearing stone on the outside. It's a real, occupied building, not a concept project. At least two more high-rise stone buildings are underway in London and Bristol. And there's a quieter development that might matter more: quarries are now cutting stone to standard brick sizing, so bricklayers can use stone in the same workflows they use for clay brick. Cost is similar to clay bricks, but the carbon emissions are dramatically lower.
Because stone doesn't need to be fired.
Stone uses much less energy to produce and emits less carbon dioxide than either brick or concrete. A clay brick has to sit in a kiln at over a thousand degrees Celsius for days. A stone brick is just... cut. The environmental argument is strong. The challenge is economic and skills-based — stone requires skilled masons, while concrete can be poured by unskilled labor. And we've spent a century training people in concrete, not stone.
The bottleneck isn't the material. It's the humans.
It's always the humans. Vernacular traditions — dry stone walling, rubble masonry — have continued uninterrupted in rural regions worldwide, but they're a diminishing craft. The knowledge is retiring, literally. In the UK, the Dry Stone Walling Association has been tracking membership for decades, and the average age of a certified waller keeps climbing. You can still find people who can build a field boundary that'll stand for two hundred years without mortar, but you might have to drive quite far to find them.
Let me pull on the thing that's been lurking under this whole conversation. We've got Roman concrete that heals itself and gets stronger in seawater. We've got modern concrete that crumbles in decades and has a massive carbon footprint. We've got stone traditions that never died but shrank to a niche. And we've got twenty twenty-three MIT research that finally explains what the Romans were doing. So... why aren't we building with Roman concrete right now?
The honest answer is that knowing the mechanism and scaling it to four point four billion tons a year are two completely different problems. The Romans used volcanic ash from specific deposits — pozzolana from the Bay of Naples. That's not available everywhere. Their hot mixing technique with quicklime is exothermic and harder to control at industrial scale. And modern building codes aren't written for materials that change properties over decades as they absorb carbon dioxide and heal themselves. The regulatory pathway doesn't exist.
Can we just pause on that regulatory point? Because I think it's easy to hear "building codes" and think it's just paperwork. But what you're actually saying is that a material that gets stronger over time is harder to certify than one that predictably degrades.
A building inspector can look at modern concrete and say: I know what this will do in five years, ten years, fifty years. The failure modes are documented. The liability is understood. Roman concrete's self-healing properties mean its behavior changes over time in ways we can model but can't yet codify into a standard that an inspector can check against a checklist. And if you can't codify it, you can't insure it. And if you can't insure it, you can't finance it. And if you can't finance it, you can't build it.
It's not that we can't. It's that our entire system — codes, supply chains, training, liability — is built around concrete that behaves predictably, even if that predictable behavior includes predictable failure.
That's the path dependence problem in one sentence. We optimized for predictability and speed, not durability. A Roman engineer would look at a modern concrete bridge and say "why did you build it to die?"
Probably while gesturing at a two-thousand-year-old aqueduct behind him.
Still carrying water.
Alright, let's land this. What's the steelman against everything we've just said? The best argument that concrete won for good reasons and we shouldn't romanticize stone?
The steelman is this: concrete can be cast into any shape, poured by unskilled labor, reinforced to handle tension, and produced anywhere from abundant local materials. Stone requires specific geology, skilled masons, and shapes that are limited by what you can cut and lift. Roman concrete was durable, yes — but it also used slave labor to mix and place it, and it took months to cure to full strength. Modern concrete sets in hours. The speed difference alone explains most of the shift. And the carbon footprint argument, while real, has to contend with the fact that we need to house billions of people quickly and affordably. Stone can't do that at scale.
That's fair on speed and scale. The carbon point is harder to dodge — concrete production accounts for something like eight percent of global CO2 emissions. But the Roman comparison isn't really the steelman's strongest ground. The stronger objection is: even if we wanted to revive stone or Roman concrete, we've built a world where every structural engineer, every building code, every construction loan assumes Portland cement. Changing that isn't a materials science problem. It's a rewrite-civilization problem.
Yet. The stone bricks going into standard bricklaying workflows — that's the kind of quiet change that actually works. You don't rewrite civilization. You make stone fit the system that already exists.
The system absorbs what fits it.
Ignores what doesn't. Which is why Pouillon's precut stone mattered — he made stone fit the crane-and-speed logic of modern construction. Digital stereotomy does the same thing for complex forms. None of this requires tearing down the concrete industry. It just... sidesteps it.
The open question we're left with — and this is where I think Daniel and Hannah are pointing us — is whether we're on the cusp of something. Not a stone revolution. A stone re-entry. Quiet, niche, driven by carbon math and digital fabrication. And the Roman concrete discovery suggests we might eventually engineer self-healing concrete at scale — but that's a different path, fixing concrete rather than replacing it.
Both paths are worth watching. The episode we didn't do today — the one about how concrete cures, the chemistry of it, and what a stone revival would actually cost — that's where the practical questions live.
We'll get there. For now, the timeline is the thing: stone dominated for millennia, Rome cracked concrete two thousand years ago, the knowledge degraded but never vanished, and the real shift to modern concrete-and-rebar happened in a thirty-year sprint between eighteen seventy-five and nineteen-oh-six. Stone never died — it just got sidelined by a system that valued speed and simplicity over durability. And now the carbon math is making durability interesting again.
Thanks to our producer Hilbert Flumingtop for making this episode happen.
This has been My Weird Prompts. If you found this episode interesting, please leave a review — it helps other curious people find the show.
We'll be back soon.
