Latest / Soft Ratios Research / A theory of place turned against people (on broken windows)

Transcript

A theory of place turned against people (on broken windows) · Soft Ratios Research

• 0:09
  Figure: Soft Ratios Research.
• 0:12
  Ground: The voices are synthetic [pause] no act, just the thinking.
• 0:19
  Figure: Hoyd Breton, a designer at Soft Ratios Studio, teaching himself in the open [pause] which works better with a reply. [pause] If this is interesting, write a review with your thoughts.
• 0:31
  Ground: One subject at a time, twice a week [pause] Tuesday and Thursday. [pause] Today, the broken window.
• 0:38
  Figure: Nineteen sixty-nine. A psychologist, Philip Zimbardo, abandons two cars — hoods up, plates stripped off. One in a poor part of the Bronx, one in Palo Alto, near Stanford. Then he waits and watches.
• 0:53
  Ground: And the Bronx car gets taken apart.
• 0:56
  Figure: In about ten minutes. The first people to reach it are a family — a father, a mother, a boy. They lift out the battery and the radiator. Inside a day it's a carcass. Children end up climbing through it.
• 1:10
  Ground: The one in Palo Alto.
• 1:12
  Figure: Untouched for more than a week. Nobody so much as leans on it. So Zimbardo walks over with a sledgehammer and takes the first swing himself. Once that first blow lands, passersby line up for a turn.
• 1:25
  Ground: That's the image that runs the whole theory. A place looks abandoned, so it gets treated as abandoned. And it's almost too good — because you feel it in your body before anyone explains it to you. [pause] You walk past a stairwell with one smashed light, and something in you lowers its voice. You know, without deciding to know, that nobody's minding this. That whatever you did in here, no one would come. [pause] And that feeling is real — I want to say that up front. The theory grips because the feeling underneath it is true.
• 1:57
  Figure: Real isn't the same as evidence, though — and that's the whole question. Look at what Zimbardo actually watched. The people stripping the Bronx car, in both cities, he described as well-dressed. Respectable-looking. And in Palo Alto nothing spontaneous happened at all. The trigger wasn't a broken window. It was a professor with a sledgehammer — an authority figure publicly granting permission.
• 2:20
  Ground: So the founding scene isn't a window breaking on its own. It's someone in charge showing you it's allowed.
• 2:27
  Figure: A different claim entirely. Thirteen years later, two men build a doctrine on the softer version. Nineteen eighty-two, in The Atlantic — James Q. Wilson and George Kelling, 'Broken Windows.' One line carries it: a single unrepaired broken window is a signal that no one cares, and so breaking the rest costs nothing.
• 2:49
  Ground: That's a beautiful sentence, honestly.
• 2:52
  Figure: And a careful essay — more careful than its reputation. They say outright it is not inevitable that serious crime will follow. The sequence they draw is disorder, then fear, then people withdraw, the informal control of a street weakens, and only then is the place vulnerable. Vulnerable. Not doomed.
• 3:14
  Ground: Where's the study underneath it, though? What were they actually standing on?
• 3:18
  Figure: A foot-patrol experiment in Newark, in the seventies. Kelling worked on it. And here's the part almost nobody carries forward — the foot patrols did not reduce crime. Crime didn't budge. What changed was that residents felt safer. They believed crime had dropped, they took fewer steps to protect themselves, they thought better of the police.
• 3:38
  Ground: So the founding evidence is about fear. Not crime.
• 3:43
  Figure: About feeling, more than fear. What fell was fear — not offending. And that distinction gets buried for the next forty years — burying it is where all the trouble starts. [pause] For a while, the evidence seemed to catch up to the intuition. Wesley Skogan, nineteen ninety, in a book called Disorder and Decline — he lines up neighborhoods and finds disorder and crime moving together. Kelling said Skogan had verified the whole thing. Established the causal link.
• 4:13
  Ground: And then somebody reran the numbers.
• 4:16
  Figure: Bernard Harcourt. He goes back into Skogan's own data. And he finds the relationship is being carried by a handful of neighborhoods in Newark. Pull those out, control for poverty, for how stable the place is, for race — and the link basically dissolves. It survives for one category. Robbery. That's the whole of it.
• 4:39
  Ground: One reanalysis is a thin thing to hang a reversal on.
• 4:44
  Figure: It isn't one. Robert Sampson and Stephen Raudenbush do something almost absurdly thorough in Chicago. They drive the streets with video cameras and code the disorder off the tape — more than twenty-three thousand street segments, across a couple hundred neighborhoods. Litter, graffiti, loitering, all of it. Then they ask what actually predicts crime.
• 5:05
  Ground: And it isn't the litter.
• 5:07
  Figure: It's what they call collective efficacy. Two things braided together — whether neighbors trust each other, and whether they'll act. Step in when kids are skipping school, or spraying a wall, or when a fight breaks out in front of the house. They surveyed almost nine thousand people to measure it. Where that's high, crime is low and disorder is low. The disorder isn't causing the crime. They're both downstream of whether a place can act together.
• 5:35
  Ground: Two symptoms of one condition.
• 5:37
  Figure: And then, two thousand four, they publish the finding that should have ended the simple version for good. They take what people say about how disorderly a block is, and set it beside what's objectively there — their own coded footage. And what's visibly there still matters — but race and money matter even more. Once you account for what was actually on the ground, people of every race saw more disorder in the poorer, more heavily minority blocks.
• 6:07
  Ground: So 'disorder' isn't a thing you measure. It's partly a thing you project. [pause] But hold on — because there's a study that cuts the other way, and it's a good one. Groningen, the Netherlands. That's not a professor with a sledgehammer. That's out on a real street, with real strangers who don't know they're in an experiment.
• 6:29
  Figure: Keizer, two thousand eight, in Science. Field experiments — which matters, because it's not a lab. They prop an envelope in a mailbox with a five-euro note showing through the slot. Clean wall behind it: thirteen in a hundred passersby take the envelope. Then they cover that wall in graffiti and run it again.
• 6:49
  Ground: And it jumps.
• 6:50
  Figure: Roughly doubles. Same envelope, same money, one changed thing — the state of the wall behind it.
• 6:57
  Ground: That's the intuition, alive, out in the world. A visible sign that the rules are slack, and more people break a different rule. You can't just wave that off with collective efficacy.
• 7:11
  Figure: I'm not waving it off. It's the strongest single pillar under the whole edifice, and it's a field result, which puts it above nearly everything else in this area. But watch what it actually shows, and what it doesn't. It shows a visible cue nudging petty behavior — littering, cutting through a fence, pocketing an envelope — right there, in the moment. It does not show disorder breeding robbery and assault down the line. That stronger, popular policy claim goes past what Keizer tested — and past Wilson and Kelling's hedged argument about vulnerability.
• 7:42
  Ground: Fair. Small wrongs, in the moment. Not the murder rate.
• 7:47
  Figure: And then the effect thins under scrutiny. A large replication and extension followed — over twelve thousand people, dozens of field experiments, across thirty-odd neighborhoods. The cue still bends behavior, but less than Keizer found, and it leans hard on the neighborhood and the person. It's not a lever you pull. It's a whisper some people hear and some don't.
• 8:09
  Ground: Why does that shape feel familiar.
• 8:12
  Figure: Because it's the story of a whole field. This kind of research — a subtle cue quietly reshaping behavior — had a terrible decade. The famous one: prime people with words about the elderly, and supposedly they walk more slowly out of the room. That collapsed when others tried to repeat it — the slow walking only appeared when the experimenter expected it. Daniel Kahneman had featured this stuff in his book, and he ended up publicly taking the endorsement back. So the honest shape of the evidence isn't a wall. It's a slope — Keizer's field work near the solid end, the lab priming near the part that's crumbling away.
• 8:52
  Ground: So what actually survives all of it. In one breath.
• 8:57
  Figure: This much. A visible sign that a place is untended changes how safe people feel, and whether they use the street — and, a little, sometimes, whether they'll commit a small violation right there. That's the durable core. Fear, feeling, petty behavior. Everything grander than that gets thinner the harder you look.
• 9:21
  Ground: Then here's what I keep snagging on. If the finding that holds is just 'it changes how people feel,' and the big one — it stops real crime — never survived... what did we actually go and do with it? Because it did not stay an essay.
• 9:36
  Figure: No. It became a way of running a city. New York, nineteen ninety-four. William Bratton takes over — first the transit police, then the whole department under Giuliani. He turns the essay into a program. Go after the fare-beaters, the panhandlers, the public drinking, the squeegee men at the intersections. Wilson and Kelling had a line ready-made: the unchecked panhandler is the first broken window. Bratton built a machine on that line. Crime mapped block by block, commanders held to the numbers every week.
• 10:11
  Ground: And crime fell. That's the part everyone remembers.
• 10:15
  Figure: Crime fell — enormously. Something like eighty percent over nineteen years, the largest drop on record. But that's exactly where it turns slippery, because everything fell at once, and the city insisted on crediting the arrests. [pause] Franklin Zimring wrote the careful book on it. His reading: the decline is real, but the cause is mostly more police, placed in the right spots, plus shutting down the open-air drug markets. And he says flatly that zero-tolerance was never even a consistent part of the strategy. [pause] And here's the detail that undoes the tidy story. While crime was falling, New York's incarceration rate went down twenty-eight percent. The rest of the country's went up sixty-five. The city locked up fewer people and got safer at the same time.
• 11:05
  Ground: So the crackdown and the safety just happened in the same years, and everyone assumed one was pulling the other. [pause] But the crackdown had its own product. Separate from the crime rate entirely. It did something to people.
• 11:21
  Figure: It did. Stop-and-frisk. Between two thousand four and the middle of two thousand twelve, the NYPD made about four point four million stops. In two thousand thirteen a federal judge, Shira Scheindlin, ruled it unconstitutional — she called it indirect racial profiling. Of those millions of stops, more than half were of Black residents, roughly a third Hispanic — about eighty-three percent were Black or Hispanic. And in nearly nine out of ten stops, the officer found nothing at all. No summons. No arrest. Just the stop.
• 11:57
  Ground: Sit with what that is as an experience. Millions of people — hands on a wall, pockets emptied, and almost always for nothing. Over and over. [pause] And which people. The same ones from that other study — the block that reads as more broken because of who's standing on it. That's the loop closing on itself. You perceive disorder where there are Black and poor people, whether or not the litter's actually there. Then you police the perception. Then the policing becomes the daily texture of the place. The theory told you to fix the window. In practice it stopped and searched the person standing near it. [pause] And the tell is what came after. Once that ruling landed, the stops fell off a cliff — and crime kept dropping anyway. If all that stopping had been holding the city down, it should have sprung back. It did not spring back; crime continued to fall.
• 13:07
  Figure: And it wasn't only New York, and it wasn't only the stops. Ferguson, Missouri. After the town became a national flashpoint, the Justice Department went through its own records. What they found wasn't really about crime — it was revenue. The city ran its municipal court as a money source. Black residents were two-thirds of the town and the overwhelming share of the stops, the citations, the arrests. Fines stacked on fines, warrants issued for missed payments. The disorder being managed was, mostly, poverty.
• 13:40
  Ground: The broken window as a billing opportunity.
• 13:43
  Figure: And the human face of it, the same summer — Eric Garner, on a Staten Island sidewalk, stopped for selling loose cigarettes. Loosies. About as pure a quality-of-life offense as exists. He dies in a chokehold, saying he can't breathe.
• 13:58
  Ground: What did Kelling make of all this? He lived to see it.
• 14:02
  Figure: He did — and this is the part that stays with me. [pause] A year after Garner, Kelling writes that the theory had been, his word, misunderstood. It was never meant to be a high-arrest program. He'd wanted the least intrusive fix — police working alongside social workers and clergy and code inspectors. He told FRONTLINE that when he heard chiefs rolling it out as blanket policy, his reaction was, 'Oh s—t.' He wrote that a lot of sins had been committed in its name. A colleague of his joked that when the war-crimes trials began, the two of them would be first in the dock.
• 14:41
  Ground: The man who built the metaphor spent his last years explaining what it wasn't.
• 14:46
  Figure: And here's the reversal that should have been the headline. Because if the intuition has a true form, somebody finally tested it properly — a randomized trial, the gold standard. Philadelphia. Charles Branas and colleagues take hundreds of vacant lots — the literal broken windows of a city, trash-strewn, overgrown — and they randomly assign some to be cleaned and greened. Haul the garbage, grade the soil, plant grass, put up a low fence. Then they measure crime around them for a year and a half.
• 15:16
  Ground: Fix the actual window.
• 15:18
  Figure: Crime around the treated lots dropped — about nine percent overall, and in the poorest neighborhoods the local drop reached twenty-nine. Gun violence down. And the people nearby felt it — they reported far less fear about stepping outside their own front doors.
• 15:34
  Ground: So the metaphor was right the whole time.
• 15:36
  Figure: The benign half of it was. Repairing the place works. And that's the almost cruel part — the same body of research that gives you 'fix the broken window' gives you no support for 'arrest the person near it.' The big meta-analyses land in the same place: problem-solving and cleanup at specific corners reduce crime; aggressive misdemeanor sweeps don't. We spent forty years and four million stops doing the half that doesn't work.
• 16:04
  Ground: And meanwhile the metaphor jumped the fence. It stopped being about streets at all.
• 16:10
  Figure: It went everywhere order can rot. Software first. Nineteen ninety-nine, a programming book — The Pragmatic Programmer — imports it wholesale. One bad piece of code, one shortcut the team's stuck with, and the standard slips: the rest is already a mess, so why not add to it. They called it software rot. Here's the irony — they cited Zimbardo, one offhand remark about New York, and nothing else. They borrowed criminology's authority for what was really a hunch. And they did it in ninety-nine — after the criminologists had already started to doubt.
• 16:44
  Ground: A vibe in a lab coat.
• 16:47
  Figure: There's a more honest cousin — technical debt. Ward Cunningham, ninety-two. Shipping quick-and-dirty code is like borrowing money; fine, so long as you pay it back promptly with a rewrite; every minute spent with not-quite-right code accrues interest. But even that drifted. Cunningham meant a deliberate choice you repay. It became an excuse-word for any mess anyone ever left.
• 17:10
  Ground: Underneath the vibe, though, there's something that isn't a vibe. Things fall apart if no one tends them. That's not criminology — in software, Lehman's laws give that intuition a more formal, thermodynamic-flavored form. Anything slides toward disorder unless someone spends effort holding its shape. And the frightening version of that isn't messy code. It's the code nobody is tending at all.
• 17:37
  Figure: The commons. Say the ones that hurt.
• 17:40
  Ground: There's a small piece of software called xz — a compression tool used across Linux. And it was being kept alive by essentially one burnt-out volunteer. In twenty twenty-four it came out that an attacker had spent about two years — two years — being friendly, being helpful, earning that exhausted person's trust, purely to slip a hidden backdoor inside it. [pause] That's the unrepaired window that actually matters. Not graffiti on a wall. A crucial thing everyone leans on, and nobody is paid to watch.
• 18:16
  Figure: And there's a slow, deliberate version too — Cory Doctorow's word is enshittification. A platform is good to you to get you hooked, then good to its business customers at your expense, then it squeezes them too, until it's a husk. What you feel as a user is exactly the broken-window read: the little degradations pile up, the dark patterns creep in, and you conclude the place has been abandoned, and your trust drains away.
• 18:43
  Ground: The window's still intact. The building behind it is being hollowed out.
• 18:48
  Figure: Which is the sharpest form of the whole problem. Because there is one place the intuition is genuinely well-evidenced — and it isn't criminology, it's safety science. Diane Vaughan, studying the Challenger disaster, named it: normalization of deviance. NASA watched the O-ring seals eroding, and because nothing blew up the first time, they reclassified the erosion as an acceptable risk. Then a little more erosion. Acceptable. Each tolerated deviation becomes the new normal — until the thing you've been tolerating kills seven people.
• 19:24
  Ground: That's the broken-windows idea finally coming true — because it's built on wreckage instead of correlations.
• 19:31
  Figure: And industry's answer to it is the opposite of a police sweep. Toyota's andon cord — any worker on the line can pull it and stop everything over a single loose bolt. You fix the small deviation the instant it appears, before it can become the norm. That's 'repair the window now,' turned into a discipline. Nobody gets arrested. The line just stops until the thing is right.
• 19:54
  Ground: But there's a trap in all of this, and I want to name it before we go praising tidiness. You can manage the appearance of order and let the substance rot underneath. Bruce Schneier calls it security theater — the pat-downs that make you feel safer without making you safer. There's an order theater to match it. A gleaming lobby, a spotless screen, a swept plaza — and behind it, the same extraction, the same neglect. A clean surface can be the most honest signal of care, or the best possible disguise for its absence. From the outside they look identical.
• 20:36
  Figure: So how do you tell them apart. That's the real question hiding under forty years of argument. When is a well-kept surface care, and when is it a lie.
• 20:44
  Ground: I think the answer's in who's doing the keeping — and in whether the wear is allowed to show. [pause] Start with what Western modernism often gets wrong. We read wear as failure. A scuff, a crack, a stain — that's decay, that's the broken window, hide it or fix it. But there's an entire aesthetic that reads wear as history. Kintsugi — a Japanese repair where you mend a broken bowl with lacquer and gold, so the crack isn't hidden, it's gilded. Made into the most beautiful line on the object. The break becomes the part you're meant to see. [pause] There's a story of a tea master shown a mended bowl whose seams didn't quite align, and the verdict was: leave it — that's exactly what makes it interesting. [pause] And then there's Ise — a shrine in Japan that's the strangest maintenance idea I know. It's Shinto, more than a thousand years old, and every twenty years they build an identical shrine on the empty plot right beside it, then dismantle the old one. They've done that for around thirteen hundred years. The building is never old and never new. [pause] Which is a completely different theory of the broken window. In our version, the window is a thing you keep from breaking. In theirs, the shrine stays alive precisely by being broken down and remade, on purpose, forever. Permanence through renewal instead of permanence through resisting decay. Maintenance as a rite, not a chore you resent.
• 22:28
  Figure: That's the warm reading, and I want to stay honest about it, because care has a shadow — built from the very same instinct. Walk any city and you'll find benches you can't lie down on. The Camden bench — a lump of contoured concrete with no flat surface, engineered so no one can sleep on it, skate it, or leave anything on it. Spikes in doorways. Ridges on ledges so a board can't grind. All of it sold as keeping a place orderly. Cared-for. And it's the opposite of care — design whose entire function is to exclude a person. Somebody looked at a sleeping human and filed them under disorder.
• 23:05
  Ground: The hostile version of 'no one's minding this place.' Someone's minding it, all right. They're minding it against you.
• 23:13
  Figure: And that's where 'Seeing Disorder' comes all the way back around. If race and money powerfully shape how disorderly a place reads — even after you account for what's actually present — then every impulse to clean it up carries a hidden verdict about whose presence counts as mess. The spikes and the four million stops are the same sentence in two grammars. Both answer 'this place feels disorderly' with 'remove the people' — not 'repair the place.'
• 23:42
  Ground: So I don't get to keep the clean, warm version of care. That's fair. But I don't think the answer is to stop caring for places. It's — who counts as part of the place. The vacant lot in Philadelphia and the Camden bench are both 'doing something about disorder.' One repaired the lot and improved nearby safety and how residents felt about it. The other made a corner less usable for the people it was designed to exclude. The tell was never tidiness. It's whether the fix takes people in or pushes them out.
• 24:15
  Figure: There's a whole body of writing trying to say exactly that — Shannon Mattern, and the Maintainers, Andrew Russell and Lee Vinsel. Their argument is that we built a civilization that worships innovation — the new thing, the launch — and starves the upkeep. And the upkeep — the cleaning and the repair and the tending — is the actual ground everything stands on, and its conditions are shaped by race, gender, class, and other forces. The unrepaired window, scaled all the way up, isn't a policing story at all. It's a whole society deferring its maintenance and calling that progress.
• 24:54
  Ground: The broken window at the scale of everything.
• 24:57
  Figure: And the volunteer keeping the compression library alive. And whoever empties the bins in the spotless lobby. The same invisible job. We only ever notice it in the negative — when the thing finally breaks and we go looking for who was supposed to be watching.
• 25:12
  Ground: So I keep coming back to where we started. The stripped car, the lit stairwell where your voice drops. That reflex was never wrong. A place does tell you whether it's tended. You feel it in a second, and you may be picking up real maintenance signals — but also the social meanings you've been taught to attach to a place.
• 25:34
  Figure: The reflex is fine. It's the sentence we kept finishing it with that did the damage. We read 'no one is caring for this' and answered 'someone should be punished.' The whole forty years lives in the distance between those two responses.
• 25:49
  Ground: And the evidence, in the end, is almost gentle about it. The thing that works is the thing the word actually says. Not the crackdown. The repair. Someone showing up, over and over, to tend a place and the people in it — and being seen doing it, so the care becomes the thing the place says about itself.
• 26:12
  Figure: The gold in the crack instead of the spike in the doorway.
• 26:16
  Ground: Which is smaller and harder than any crackdown. No arrest in it, no law — just the slow work of tending, kept up long past the point anyone's watching.
• 26:29
  Ground: When a place makes you drop your voice — when it reads as untended — how much of that is the place itself, and how much is who you've been taught to see as disorder?