ASHFALL INSTITUTE | SUBDUCTION ZONE
Building Bridges — Episode Five
THE SHAPE OF THE BUILDINGS
P. A. Moore
Ashfall Institute | Subduction Zone
Written under the collaboration disclosed in Time Is a Wheel and We Are the Ground. Concept, argument and judgment: P. A. Moore. Research and composition: Claude Opus 5.
There is a house in Norman, Oklahoma that looks like an animal.
It was finished in 1961. Herb Greene designed it, while Bruce Goff was running the architecture school at the University of Oklahoma — and Goff’s department in those years was the strangest and best thing in American architecture. Greene had been studying at Syracuse. He saw Goff’s drawings, and he transferred.
Life magazine called the house the Prairie Chicken, and the name stuck, though it isn’t quite a chicken. It’s shingled all over in overlapping wooden scales, and it hunches. From some angles it looks like a bird about to go up. From others, like something wounded lying down in the grass.
I walked through it as a young woman. It was empty then, and the land around it had gone over to weeds, so there was nobody to ask and nothing to stop us.
I remember two things.
The light came in at angles the shape of the building invented, and it laid shadows across the inside that I had never seen in a room before. There were little rooms upstairs off a central stair, and the whole of it felt like Alice in Wonderland on the Plains.
And it was quiet.
In Oklahoma, the wind does not stop. It is not weather. It is the condition. You stop hearing it the way you stop hearing your own house settle, and then you walk into a building where it isn’t, and the silence is so strange you look around for the cause.
I was not measuring anything. I had no instruments and no training. I was a girl in an abandoned house.
And I could feel the building working.
Everything written about that house is about the shingles.
What it was actually doing was sitting in the right place, facing the right way, shaped so the wind had nothing flat to hit and nothing sharp to tear across. Sheltered from the worst of it. Open to the setting sun.
That is not decoration. That is the entire argument of this episode, and I felt it before I had a word for it.
The building answered the place it was standing in.
Nothing we are putting up in this country right now does that.
Let me tell you about the failure, because I watched that too, and I am not interested in a story where the good idea simply wins.
Earth-sheltered building was serious in Norman, and for the obvious reason. Tornadoes. You put soil against a wall and over a roof and you have thermal mass, and you have somewhere to be when the sky turns green. It is an excellent idea. I have never stopped thinking so.
Whittier Middle School was earth-bermed and open-concept inside.
Both halves of that sentence matter, and almost nobody has ever separated them.
Inside it was dungeon-like. There was not enough daylight, and you felt it within a minute of walking in. And because the interior was open-plan — no walls, one great room with the classes laid out in it like furniture — there was nothing to stop a voice traveling. It was chaotic. You could hear everything from everywhere.
And it felt damp. Not wet. Damp, in the way a building tells you its air is not moving.
Eventually they remodeled it into something conventional, and that was a shame, because they did not have to. A light well through the middle would have solved it. An interior atrium, open to the sky, throwing daylight down into the center of the plan — and they could have kept the thermal mass and kept the tornado protection and kept every good thing the building had.
They chose the box instead.
Now here is the part I want you to hold onto, because it is the shape of every argument in this series.
Two different experiments were poured into that one building.
Earth sheltering, which was structural, and which was correct.
And open-plan classrooms — which had nothing to do with soil, or wind, or tornadoes, or thermal mass. Open-plan was a teaching fashion. It came out of the open education movement of the late sixties and seventies, it was about pedagogy, and it swept through school design the way fashions do.
The acoustic disaster came entirely from the fashion. Voices carry because there are no walls. Dirt has nothing to do with it.
But they failed together, in one building, in front of a generation of parents in a college town who talked to each other.
And when the verdict came in, earth sheltering took the blame for a failure it did not cause.
The daylight problem was a daylighting design failure. The damp was a materials failure — the waterproofing membranes you could buy in 1975 were not equal to permanent ground contact, and the era’s answer to the oil crisis was to seal buildings up tight without putting any mechanical ventilation back in. Wrap a building in soil. Seal it. Don’t move the air. You will get exactly what I walked into.
Light. Sound. Moisture. Siting.
Every one of those is fixable. Not one of them is a fact about putting earth against a wall.
The idea has been dead for forty years on the strength of a verdict it never earned.
The successes were out there. They were just quieter, because a building that works does not make the paper.
What I saw work was straw bale, in Texas.
And the part of it I have never forgotten is where the material came from. There were students at OU who went home afterwards — to family farms, to ranches — and took their ideas with them, and started baling and selling straw as a building material. Not a company. Not a program. Young people who had been taught to think differently about buildings, going back to the places they came from and turning what was already in the field into walls.
Now. I know what you are thinking, because everybody thinks it.
A house made of straw. We all know how that story ends.
So let me give you the number, because it is the most surprising fact in this episode and it is not close.
A plastered straw bale wall carries a two-hour fire-resistance rating under ASTM E119 — the standard assembly fire test, the one commercial fire-rated construction has to pass. An inch of cement stucco over wheat bales: two hours. Earthen plaster at an inch and a half: two hours. The assemblies hold at around eighteen hundred degrees for the full two hours and keep the room separated.
The reason is the thing nobody guesses.
Densely packed straw does not have enough oxygen inside it to burn. It smolders. It will not flame. Flame penetration runs forty to sixty percent slower than through a conventional stud wall.
The material everyone assumes is a tinderbox outperforms the wooden frame we build almost every house in America out of.
That is not a clever trick. That is what happens when somebody actually tests a thing instead of knowing about it.
I want to put one more building next to those, and it is not a house.
The western half of the transcontinental railroad was built by Chinese labor. The Sierra grades. The tunnels driven through granite by hand and black powder. The retaining walls and the rock work, dry-laid, holding a mountainside back.
It was finished in 1869.
The Chinese Exclusion Act passed in 1882.
Thirteen years between building the spine of the country and being barred from it by statute.
And a great deal of that stonework is still standing. Still holding. You can go and put your hand on it.
I am not going to make a speech about that. I only want it in the record, in an episode about what lasts, that we have done this before — taken the work, kept the work, and legislated against the people who did it. The wall outlived the welcome by a hundred and fifty years and counting.
So why aren’t we building this way?
Not because we can’t. That is what I want to establish, and I am going to be careful about it, because the fastest way to lose an argument like this is to overclaim by one sentence.
Here is what exists, in order of how ready it is. I will tell you the stage of each.
Built, tested, and in the code: straw bale. You can point at houses. It went into the International Residential Code as Appendix S in 2015. It took decades it should not have taken, but it got there.
In pilot, with real projects behind it: self-healing concrete. This one caught my attention for underground housing and basements, and once you understand it you will see why.
Dormant bacterial spores are mixed into the concrete along with a food source — usually calcium lactate. They sit there inert, potentially for years. Then a crack opens, and water gets in, and the water wakes them up. They eat, and they precipitate limestone, and the limestone fills the crack from the inside.
Water arriving is not the failure. Water arriving is the trigger.
Think about what that means for the building I walked through in Norman. The thing that destroyed a buried wall in 1975 is the thing that repairs this one. It cannot work without the moisture. The literature knows it, too — the applications named again and again are tunnels, basements, underground rail stations. Structures where you cannot prevent cracking and you cannot tolerate water.
The honest limits: it heals fine cracks, up to about eight tenths of a millimeter, a little more with granular healing agents. It does not repair structural failure. It costs more. And nobody knows yet whether the spores are still viable at fifty years, because the material is not fifty years old.
In pilot, with a research program: transparent wood. Take wood, remove the lignin, infiltrate it with a polymer, and you get a material that passes light. Not clear like a window — it diffuses, so the light comes through soft and even, the way north light does. It insulates far better than glass. It carries load. And it does not fail in shards.
There is a European project running right now, using machine learning across scales, aimed specifically at developing it to replace glass and plastic in construction.
In the laboratory: nano lattices. Carbon structures designed by AI, reported at roughly the strength of steel at the density of foam.
Early research: living and regenerative materials. Mycelium composites and their relatives. Real, interesting, and nowhere near a wall you could get signed off.
That is the honest ladder. I have put the weakest claims at the top of the stairs where they belong.
And I want to say plainly where AI sits in this, because it is not decoration and it is not enthusiasm.
The search space for a new material is astronomically large. Human intuition has been working on cement chemistry for a century and is exhausted. The bottleneck was never imagination. It was the rate at which candidates could be evaluated.
That is a search problem. It is the same shape as the problem in the last episode about finding the three or four mutations in a tumour that the immune system can actually see.
Different substrate. Identical bottleneck.
Possibilities were never abandoned. Materials design was lagging too far behind.
It isn’t lagging now.
So we come to it.
We are rebuilding, right now, in wildfire country and in floodplain, out of materials we know will burn and we know will rot. We are doing it on purpose, at scale, with the ashes still warm.
Why?
It is not ignorance. The builders know. The architects certainly know.
The answer is underwriting.
An insurance company prices risk from loss history. That is the whole trade — you look at how a thing has failed before, and you put a number on it. A novel wall assembly has no claims record. So it is not rated as risky. It is rated as unknown, and unknown prices worse than bad.
A material can be better in every measurable physical property and still be uninsurable, because superiority without a claims record is indistinguishable from ignorance.
And insurance gates everything downstream. No policy, no mortgage. No mortgage, no house.
So it is the underwriter who decides what gets built. Not the architect. Not the building official. Not you.
The known loss is priceable, and a priceable loss is a business. The unknown one cannot be sold.
Except the known losses have stopped being priceable, and this is where the argument breaks open.
California’s insurer of last resort — the FAIR Plan — went from 242,440 residential policies in September 2021 to 642,010 in September 2025. A hundred and sixty-five percent increase. Enrollment rose forty-three percent in the fifteen months to December 2025 alone.
Private carriers are pulling out of 46 of California’s 58 counties — including neighborhoods that are not at high fire risk at all.
The 2025 Los Angeles fire destroyed about twelve thousand homes. Roughly forty billion dollars.
The FAIR fund had three hundred and seventy-seven million dollars in it when those fires started. The commissioner had to levy a billion-dollar special assessment on the member insurers — the first such assessment in more than thirty years.
So the cautious path — build the familiar box, price the familiar loss, refuse the unfamiliar one — has produced a failing market and a public backstop swallowing what the private one will not touch.
The safe option is not safe. It is only familiar. And the bill arrives either way.
Now I have to say something harder, and I have had to correct myself to get to it.
It is tempting to tell this as a story about caution. Prudent institutions, moving slowly, sometimes too slowly. And there are famous cases that seem to support caution — thalidomide, leaded petrol, radium paint. Look what happens, the argument goes, when we rush.
But that is not what happened in any of them.
Thalidomide was sold over the counter from 1957, including to pregnant women. From 1959, doctors were reporting nerve damage. The manufacturer disputed the reports, fought to keep the drug non-prescription, and resisted withdrawal even after the birth defects were identified in late 1961. The warning existed. The profits were enormous. The warning lost.
The dangers of leaded petrol were demonstrated in 1924, when workers at a refinery in New Jersey hallucinated and died. It stayed on the market for another fifty years, defended by research the industry paid for. When a scientist finally proved the contamination in the 1960s, he lost his funding and his seat on the advisory panels.
At the radium dial factories, the executives and the chemists used shielding. The women were told to point the brushes with their lips. When they began to die, the company disputed the diagnoses.
In every one of those cases the warning already existed, and was overridden by the party making money.
Not one of them is a story about innovation failing honestly. They are all the same story, and it is a different story.
Which changes the question.
The variable was never new against old. It is whether the party holding the information has a stake in the answer.
And once you see it that way, you cannot let the underwriters stand as the cautious ones either.
Refusing to insure a better wall is not disinterested prudence. It is a business protecting a model that depends on the losses being the ones it already knows how to price. They have a position. They benefit from the answer.
Their caution and the drug company’s recklessness are, structurally, the same act: a commercial interest deciding what counts as known.
So the question is not how much caution.
The question is who is permitted to declare a thing safe or unsafe, and what they stand to gain by the answer.
There is one person I want to name, because this series has had enough mechanisms in it and not enough people.
Frances Kelsey was a medical officer at the FDA. The thalidomide application landed on her desk in 1960, and she refused it. The company came back, and she refused it again. They went over her head, and complained about her, and called her a bureaucratic obstacle, and she refused it six times in all — while they distributed two and a half million tablets to American physicians as “investigational samples,” which required nobody’s consent.
She was not cautious by disposition. She had no stake in the answer, and she asked for the data.
That is the whole difference, and it is available to us at any time. We simply have to decide that it is worth paying for someone whose salary does not depend on the verdict.
So here is the ask for this episode, and it is small, and it is boring, and it would change what gets built in this country.
The obstacle is not risk. The obstacle is the absence of data about risk. So build the mechanism that produces the data.
One. Standardized testing and evaluation pathways for innovative wall assemblies — a route by which a new material can be proven to a published standard rather than waiting decades for an appendix.
Two. Monitored pilot builds that generate real performance and claims data, published, where anyone can use them.
Three. A public backstop for early adopters — so the first few hundred houses can be insured while the claims history accumulates that makes the next ten thousand insurable on their own merits.
That is it. Nothing is banned. Nobody is forced. We simply stop pretending that the absence of a number is the same thing as the presence of a danger — and we go and get the number.
We have done it before. Straw bale is in the code. It took far longer than it should have, and the path exists.
You know the story about the three little pigs.
Straw, and sticks, and brick. The wolf blows down the first two, and the third pig is safe, and that is the end of it. That is the whole moral: be afraid, and build heavy.
And it worked. Nobody disputes that it worked.
But think about what happened afterward, because the story never does.
The third pig spent the rest of his life inside a poorly insulated, environmentally expensive box. He was cold in the winter and hot in the summer and he paid for it every month for the rest of his life. He never questioned it, because it had worked once, on the worst day, and after a day like that you do not go back and re-examine your assumptions.
He bought survival with permanent discomfort and he called it wisdom.
We have been living in that pig’s house for a hundred years.
And there is a fourth pig.
He could have thick walls that hold their temperature and will not carry a flame. Light coming through a wall that will not shatter and does not weigh what glass weighs. A buried room that heals its own cracks when the water comes, because the water is what wakes it up.
He could stay inside while the fire goes past.
He is not in the story.
He wasn’t there when the first three settled the building code.
Next time: the grid — which was already failing before anyone said the word “AI,” and what gets built for the data center that stays for the town.
Sources: Herb Greene, Prairie House (Prairie Chicken House), Norman OK, 1961 — ArchDaily; Architectural Record. Whittier Middle School renovation program, Norman Public Schools; the renovation architects identified lighting and acoustical privacy among four core problems in an open classroom layout with little natural light. Straw bale fire testing: ASTM E119 two-hour ratings for plastered assemblies, Ecological Building Network; straw bale entered the IRC as Appendix S in 2015. Self-healing concrete: RILEM Technical Letters, full-scale demonstrator projects; ScienceDirect, microbial self-healing concrete in underground sidewall application. Transparent wood: AI-TRANSPWOOD (EU); Nature Reviews Materials. California FAIR Plan policy counts, September 2021 and September 2025; Milliman, California homeowners insurance and the Los Angeles wildfires. Frances Kelsey and the thalidomide application: FDA historical record. Chinese labor on the western transcontinental grade, completed 1869; Chinese Exclusion Act, 1882.
Building Bridges is a fourteen-part series. Previous: Episode Four — The Loop. Next: Episode Six — The Grid We’d Have to Build Anyway.
P. A. Moore is the pen name of Pamela King, philosopher and artist. Available through the Ashfall Institute.