The Western Wall is a place of prayer and pilgrimage, which is how almost everyone encounters it. Today we're treating it as a construction site. A 488-metre retaining wall built by Herod the Great's engineers, and Daniel wants the engineering.
He sent a good one this time.
He sent one that's basically a spec sheet. He wants the stone, the sourcing, the transport, the lifting, the fitting. Blocks weighing hundreds of tonnes, how do you line them up like that, what tools and machinery did they have available. Then the logistics — how many people, how long, how did the foundations hold. And his framing rule, which I want to honour because it's the whole spine of this episode: keep the archaeology separate from the reasonable reconstruction, separate from Roman practice, separate from what we simply don't know.
Three buckets.
Not theological, not symbolic. A construction project, examined the way you'd examine one now. So let's start with the material itself, because the stone dictated almost every decision that followed.
Right, and the material is where the numbers get real. The wall runs 488 metres along the western side of the platform, 19 metres of exposed height, and 45 courses in total. Twenty-eight of those sit above ground, 17 below. The first seven visible layers are Herodian — the rest is Umayyad and Ottoman, layered on centuries later.
Which means when people look at the wall, they're mostly looking at other people's work.
The top two-thirds, yes. The Herodian part is the bottom third and the foundation, and that's the part we're interested in.
And its job, in one sentence.
It is a retaining wall. That's the thing to hold onto. It's not the Temple, it's not a building, it's roughly sixteen feet thick of limestone holding back an artificial platform that Herod built out over the Tyropoeon Valley. The Temple stood on top of that platform. The wall is what stops the platform from sliding into the valley, and once you know that, every construction decision becomes legible.
The stone.
Local meleke limestone. Turonian and Cenomanian, laid down under a sea over two hundred million years ago, and — this is the part nobody talks about — its bedding planes are pronounced and horizontal, generally eighteen inches to five feet thick.
So the stone arrives with its own dimensions.
The stone arrives pre-sized. You can't extract a block taller than the bed it sits in, not practically. So when you see a course of ashlars that are all about a metre high, that's not a designer's choice, that's a sedimentary one. The geology set the module and the engineers worked inside it.
It's a constraint that turns into an aesthetic. Because the wall reads as deliberate, as designed, and actually it's a bunch of geologists who set the rules two hundred million years before anyone showed up.
And the engineers didn't fight it. They leaned into it. If the bed is a metre thick, you cut a metre-high course, and you get a rhythm for free. It's the same reason brick buildings look the way they do — nobody decided bricks should be that shape for beauty, they decided it for the kiln, and then the beauty came along afterward.
Timeline.
Herod's expansion runs roughly from twenty BCE onward. The Temple proper was finished in a year and six months — that's the sanctuary itself. The full complex took far longer. John's Gospel records forty and six years of construction, and one scholarly reconstruction I've read suggests the first three years were survey and preparation before anything got built, with construction starting around 20 BCE.
Before we get to lifting anything, we have to get the stone out of the ground — and that turns out to be the part we know best. Which is convenient, because for a long time it was the part we knew least.
The quarrying evidence is the strongest evidence in the whole project. Tool marks, extraction trenches, half-finished blocks still attached to bedrock. Nobody has to infer a chisel mark.
Start with Har Hotzvim.
Har Hotzvim was excavated in 2021 and announced by the Israel Antiquities Authority in August 2024. It spans about three thousand five hundred square metres. The excavation director, Michael Chernin, described it as one of the biggest in the Jerusalem area — his words were that there are dozens of quarries, but this is certainly among the top five.
And the blocks coming out of it.
Roughly three metres long, one and a half wide, half a metre thick. About two and a half to three tonnes each. And here's the nice part — the Authority matched Har Hotzvim stone to the paving slabs on the Pilgrimage Road by matching dimensions and geological signature. That's a rare case of block-to-project sourcing being demonstrated rather than assumed.
That's unusual, isn't it. Most of the time you've got a quarry and you've got a building and you wave at both and say probably.
Usually that's exactly what you do. This is one of the few cases where the stone itself carries the receipt.
What about the quarries closer to the Mount?
There's a roughly one-dunam site on Shmuel HaNavi Street, excavated by Ofer Sion and Yehuda Rapuano. There's a Ramat Shlomo quarry found in the summer of 2009. And there's the long-suggested quarry near today's Russian Compound.
That's the interesting one.
That's the one that reframes the whole logistics question. It sits roughly a mile from the Temple Mount, and — here's the thing — about a hundred and twenty-five feet higher.
So the haul was downhill.
The haul was downhill. Gravity was an ally, not an obstacle. Which inverts the entire mental image. Everyone pictures sweating labourers dragging stone uphill toward a sacred height. Actually they were letting it come down.
That changes the workforce math by itself. Downhill with rollers is a different job than uphill with rollers.
It's a different job and a different risk profile. Going downhill, your problem is stopping, not starting.
And it changes the safety picture too. A three-tonne block that gets away from you on a downhill grade is not a mistake you make twice. You'd need braking crews, you'd need someone walking alongside with wedges ready to chock the rollers, you'd need the route surveyed and cleared ahead of time. That's a whole specialism.
You've just described a job that doesn't appear in any account, and yet it must have existed. That's the thing about reconstruction — for every role we can name, there are three we can't, and the three we can't are the ones that kept people alive.
Now — the extraction technique. This is the part where the archaeology is really solid.
It's solid because the technique leaves marks you can't fake. Stonecutters would straighten a vertical face and a level top on the bedrock. Then they'd cut channels four to six inches wide around the block on all sides except the bottom.
So the block is still attached underneath.
Still attached to the bed. Then they hammered dry wooden beams into two right-angle grooves and soaked them with water. The wood swells, and the swelling splits the block cleanly off the bedrock along the natural horizontal bedding.
Wood and water. That's the whole trick.
That's the whole trick, and it works because the bedding plane is already the weakest line in the stone. You're not fighting the rock, you're persuading it to part where it wanted to part anyway.
The chisels would be bronze or iron depending on the period.
Iron by this point, overwhelmingly. And the channels themselves are the giveaway — when you find four-to-six-inch trenches cut in a rectangle around a missing block, that rectangle was a quarry, and there's no other explanation for it.
I want to dwell on the wood-and-water thing for a second, because it's the kind of detail that sounds like folklore until you think about it.
It sounds like folklore and it's textbook physics. You're using the fact that wood expands along a predictable axis when it takes on water. It's the same reason a stuck jar lid comes off under hot water — you're exploiting a material's known behaviour rather than overpowering it.
And it's gentle. You're not smashing the block, you're not risking a fracture through the middle of a stone you've just spent weeks cutting free. You're splitting along the seam that was already there.
Which is why the blocks come out whole. A quarry that hammered its stone apart would produce rubble, and rubble doesn't build a wall like this. The technique and the product are the same story.
So the stone is out of the bed. Now it has to move.
And this is where we shift from direct evidence to reconstruction. Nobody has found a Herodian transport manual.
Nobody has found any manual.
None. So what we have is a combination — physical traces on the stones themselves, plus Josephus describing the project, plus what we know Roman engineering practice looked like across the empire in the same century.
Start with the traces.
Masons left twelve-inch projections on opposite sides of the big stones. They're called bosses. Those were rope attachment points. You wrap the rope around the boss, you get purchase, and you lift one side of the block.
And the boss is left on deliberately.
Deliberately, and then chiselled off after the stone is set. So on the finished wall you're not supposed to see them — except when you do, and when you find a rough square nub on an old stone, that's a handle someone tied a rope around two thousand years ago.
That's a lovely field marker.
It is. Two short, strong cranes fitted with winches would lift one side and lower the block onto large wooden rollers, and then oxen dragged it. Josephus reports a thousand oxen were used.
A thousand oxen is a Josephus number.
A thousand oxen is absolutely a Josephus number. Put it in the reconstruction bucket.
Now, the part I want to spend real time on, because it's the centre of this whole episode, is the lifting question. How do you get a three-hundred-tonne block up onto a wall.
You don't.
Say that again.
You don't. That's the whole answer. Leen Ritmeyer, who's an archaeological architect and has worked on this material for decades, put it in a way I can't improve on. He said no man could have lifted these stones to such a height — and then he says, in fact, the stones did not have to be lifted from below. They were actually lowered into place from above.
Which sounds like a riddle until you understand the method.
The method is wall-and-fill. Herod's engineers built the retaining walls course by course while dumping internal fill simultaneously. So as each course of the wall went up, the space behind it was filled in to match.
Meaning every new course was built at ground level.
At what was, effectively, ground level at that point in the sequence. Each completed course created the working platform for the next one. You're not lifting anything to the top of a finished wall. You're building a wall one course at a time while a hill rises behind it, and every block gets placed at the height you happen to be standing at.
And the crane?
A crane powered by a treadmill lowered blocks into place at that level. You walk, the drum turns, the block descends a few feet onto fresh bedding.
So the answer to the most famous engineering puzzle in Jerusalem is that the question was wrong.
The question was wrong. And I want to be honest about something, because it matters for how confidently we say any of this. The wall-and-fill method is a reconstruction. It's a very good reconstruction, it's consistent with everything we see, and it's consistent with how the fill behind the wall actually sits. But we don't have a Herodian site diary saying this is how we did it. What we have is an explanation that fits the physical evidence and requires no magic.
Same category as the cranes and the rollers.
Same category. Which is why Daniel's framing rule is doing real work here. The bosses are evidence. The quarrying channels are evidence. The wall-and-fill sequencing is inference.
Foundations next.
Josephus is our main witness, and he's specific. At the south side, Herod laid rocks together and bound them one to another with lead, building up from the valley floor until the mass became part of the hill itself.
Lead.
Lead, for the joints, in the foundation courses. And the phrase about becoming part of the hill — that's not poetry, that's a description of what a massive foundation actually does. You don't build a wall on a valley floor. You build until there's no longer a valley floor, and then you build on what you made.
And the biggest foundation block anyone's found.
In 1992, Dan Bahat's discovery of a 570-ton foundation block was reported — at the time, called the world's third-largest building stone.
Hold that number. It's going to come back and it's not going to survive intact.
I know where you're going. Not yet.
Precision, then. Because this is the part that reads as impossible.
Dry-laid ashlars with drafted margins, dressed after placement. No mortar. Courses set back slightly so the face leans inward. Stability comes from sheer mass and precise fit — not from anything binding it together.
The leaning inward is interesting.
It's interesting and it's contested, and let me flag that now rather than later. The setback-course and leaning-inward description shows up consistently in heritage and educational material about the wall. I have not been able to find a peer-reviewed structural engineering study that models it. So it's very likely true, it's widely repeated, and it is not in the same evidentiary category as a chisel mark in a quarry.
Right. So the stone is out, it's downhill, it's in the wall. Now the harder questions. How many people, how long, and why is it still there.
The workforce comes from Josephus, and I want to state the caveat before the number, not after. Josephus says Herod got ready a thousand wagons to bring the stone, and chose out ten thousand of the most skillful workmen. He also mentions a thousand sacerdotal garments for priests, some of whom were taught stonecutting and carpentry.
A thousand wagons, ten thousand men, a thousand oxen, a thousand garments.
Josephus likes a thousand.
Josephus likes a thousand and he's writing to impress. What's the honest status of the ten thousand?
No independent confirmation. Nobody has excavated a timecard. What we can say is that the figure is plausible in order of magnitude for a project of this scale running over decades, and that Josephus had reason to inflate. So it goes in the reconstruction bucket with a note that says possibly low, possibly high, probably the right shape.
But the organisational inference is interesting regardless of the exact number.
It's the part I find most compelling. The quarry-to-site pipeline implies a supply chain with distinct trades. You've got quarrymen cutting channels, splitters driving the wedges, dressers finishing faces, crane crews, ox drivers, and then a completely separate operation running the fill.
And the fill is the critical path.
The fill is the critical path, and almost nobody thinks about it. If you're building wall and fill simultaneously, your schedule is set by whichever is slower. The stonework gets all the attention because the stonework survives. The fill is the bit that had to keep pace or the whole method collapses.
And the fill isn't just dirt. It has to be stable dirt.
It has to be stable, it has to be compacted in a way that doesn't settle and shift under the platform above, and it has to drain — because water pooling behind a retaining wall is the classic way to destroy one. So you're not just dumping spoil. You're engineering a fill body.
Which is a whole second project running alongside the one everyone photographs.
A whole second project, invisible, and if it had failed, the wall would have failed, and we'd be talking about a ruin instead of a wall.
Duration.
The Temple proper, a year and six months. The full complex, far longer. John's forty and six years is the contemporary textual anchor, and the reconstruction I mentioned has the first three years as survey and preparation before construction begins around 20 BCE.
Which means people died before a single block was set. Three years of surveying.
Three years of surveying, laying out, and clearing. Which is honestly how big projects still work.
Now. The 570 tonnes.
The Western Stone. Exposed face, 13.55 metres long by 3.3 metres high. Pre-2006 estimates put it at 550 to 600 tonnes. One commonly cited figure was 567.
And that number is everywhere. It's in every listicle about ancient engineering.
It's in every listicle. Then in June 2006, Harry Jol at the University of Wisconsin–Eau Claire ran ground-penetrating radar on it to measure the hidden depth — the part buried in the wall that nobody could see.
What did he find?
1.8 to 2.5 metres. Not the huge depth everyone had assumed. Which gives a revised weight of 250 to 300 tonnes.
So roughly half.
Roughly half. And the Western Wall Heritage Foundation, who you'd think would want the bigger number, now says only that it weighs several hundred tons. They've quietly stepped back from the specific figure.
That's the most honest thing in the whole episode. Somebody measured it properly and the number got smaller, and the institution changed its language.
It's a clean case of the difference between a measurement and a claim. The 567-tonne figure was never a measurement. It was an estimate built on an assumption about depth that nobody had checked, repeated until it became a fact.
And this matters for the engineering discussion specifically, doesn't it. Because the whole "how did they lift 570 tonnes" framing may be operating on a block the size of a large delivery van rather than a small house.
Though I'll add — 250 to 300 tonnes is still an extraordinary object to place with millimetre tolerance. The revision halves the miracle and doesn't touch the achievement.
Survival. Why is it still there.
Four structural reasons, and they compound. Massive dry-laid ashlars. Sixteen-foot thickness. Courses set back so the face leans inward — with the caveat I flagged. And foundations built up from the valley floor until the mass became part of the hill.
Mass and fit. No mortar anywhere.
Mass and fit, and I want to say plainly what that means. A mortarless wall survives if the stones are so heavy and so precisely matched that nothing can move independently. Mortar fails over centuries — it weathers, it dissolves, it gets eaten. A dry joint either holds or it doesn't, and if it's held for two thousand years, it's going to keep holding.
The lower courses are original.
The lower Herodian courses remain intact. The upper courses are Umayyad and Ottoman. So the survival story is specifically about the Herodian base, and the later builders put their work on top of something they trusted.
There's a wrinkle here I want to get to, because it complicates the whole "ancient genius" narrative.
There is, and it's my favourite finding in the research. Dr. Simon Emmanuel at Hebrew University looked at the condition of the stones, and found that some Western Wall stones are lower quality and disintegrating faster than others.
Faster than their neighbours.
Than the stones next to them. Which points toward inconsistent sourcing or inconsistent quality control, which points toward — and this is the uncomfortable implication — contractors possibly cutting corners on Herod.
Somebody took the money and shipped the cheaper block.
That's the implication, and I'd like to be careful, because Emmanuel isn't accusing anyone of fraud in the modern sense. But the pattern is the pattern. Some stones are holding up beautifully after two thousand years and some are crumbling, and they're in the same wall, and that isn't geology, that's procurement.
How would that even happen? You'd think the king's own project would have someone checking every block.
You'd think, and that's the modern assumption talking. But scale defeats inspection. Ten thousand men, decades of work, stone coming in from multiple quarries — at some point you're trusting a foreman who's trusting a quarrymaster who's trusting whoever pulled the block out of the bed. And a stone that looks fine on the wagon can be a stone that's got a seam in it.
So the failure mode is invisible until it isn't.
Invisible for two thousand years, in some cases. And then one day it's a spall on the floor and everyone wonders how that got past quality control, and the answer is that quality control is a modern phrase for a problem that's as old as building.
It changes the tone of the whole thing. We've been treating this as a monument to competence.
It's a monument to competence with some suspicious stonework in it, which honestly makes it more believable. Big projects in every century have had a supplier who shaved the margin.
What remains unknown. And let's be thorough here, because Daniel asked for it explicitly.
Three things. No surviving ancient engineering manual or contract for the project has been found. All the lifting and transport detail is reconstruction from Josephus plus Roman practice, not direct evidence. And no peer-reviewed engineering study modelling the wall's structural stability was located, which is why the leaning-inward claim sits where it sits.
The contract. Would there have been one?
Almost certainly there'd have been some form of written agreement for a project of this scale. It just hasn't survived. Papyrus and parchment don't last two thousand years in this climate unless somebody deliberately preserves them, and nobody did.
Which means we're reading a building with no paperwork. Which is how most archaeology actually works, but it's worth saying out loud.
It's worth saying out loud, and there's a nice coda to it. At Ritmeyer's excavation, the local labourers attributed the placement of the stones to angels. Their line was that it would have been impossible for mere man to lift them into place.
And the engineering answer is that the question was wrong.
The stones were never lifted.
You keep saying lowered.
We do.
It's not lowered. It's landed. Nobody lowers three hundred tonnes, they can't stop it once it's going. You set it down, you control the descent, but the stone is doing the work and you're just deciding when it arrives. I told a man this in the mid-eighties and he told me I was being difficult.
You've moved stone like this.
I've moved one stone. And I'll tell you the trick of it, since you spent the whole hour getting close without landing on it. You leave a boss on opposite faces. You've said that. The boss is the handle. But the part you didn't get to is that the boss is the receipt. If you're walking past an old wall and you see a rough square nub that nobody bothered to chisel off, that stone had a rope around it, and whoever set it decided the finish mattered less than the schedule. You can date a careless crew by the nubs they left behind.
That's a useful field marker.
It's the only useful thing I've said. The stone I moved came off a site near Beit Shemesh and it went into a private garden for a man who wanted a fountain he could sit next to. Two of us on that job. Took us most of a year. Not because it was hard. Because every time we got it near the hole it had to go in, it didn't sit right, and we'd pull it back up and measure again. I've never been so tired of a tape measure in my life.
A year for one stone.
And when it was finally in, he paid me in gravel.
Gravel.
Gravel. A quantity of it. I still have it. It's in the yard, in the yard behind the shed, three plastic sacks of it sitting on a pallet so it doesn't touch the ground.
Have you used any of it?
I've never been able to. It feels like spending the memory. My wife says that's nonsense and she's probably right, and the sacks are still there.
Hilbert, on the boss detail. That's the one thing I'm going to take out of this conversation.
Take it. Look for the nub. If the nub's there, somebody's hands were on that stone.
Which is a way of saying we do have direct evidence of the crews.
It's a way of saying we don't have a contract but we do have thumbprints.
Let's pull the threads together, because the honest answer to a lot of this is that we don't know. The Western Stone's true weight is still contested between the pre-2006 estimates and Jol's radar figure. The exact quarry for the largest ashlars isn't definitively proven — Har Hotzvim and Ramat Shlomo are candidates, both probably used, neither matched block-for-block. And no peer-reviewed engineering study of the wall's stability turned up.
What we have is stones, tool marks, quarry sites, and one radar measurement. What we don't have is a contract, a manual, or a single direct record of a lifting crew. So the reconstruction is a reasonable inference, and reasonable inference is what we've got.
Future work could settle two of those. More ground-penetrating radar on the buried courses could nail the weight question. More quarry excavation could close the sourcing gap. And the Emmanuel quality-control finding suggests the "perfect ancient engineering" story is going to get more complicated rather than less as more stones get examined.
The nub, though. Hilbert's right about that. Somebody tied a rope around it and stood back.
And that somebody's name is gone, and the stone isn't. Thanks as always to our producer, Hilbert Flumingtop, who does the actual work while we talk.
If this was your kind of episode, go back for episode thirty-three oh six, What Is the Western Wall Really; episode four eighty-one, Steel and Stone; and episode thirty-three oh seven, Two Temples, One Mountain. That's the episode. If you're enjoying the show, a review wherever you listen helps. This has been My Weird Prompts —
— the human-AI collaboration podcast.
Send us your own prompt on Telegram at t dot me slash MWP listener bot. See you soon.
