What We Get Wrong About Violent Crime
Late on a Sunday night in June of 2023, a woman named Carlishia Hood and her fourteen-year-old son, an honor student, pulled into Maxwell Street Express, a fast-food joint in West Pullman, on the far South Side of Chicago. Her son stayed in the car. Hood went inside. Maxwell is a no-frills place—takeout-style, no indoor seating. It’s open twenty-four hours a day. Hood asked for a special order—without realizing that at Maxwell, a busy place, special orders are frowned upon. The man behind her in line got upset; she was slowing things down. His name was Jeremy Brown. On the street, they called him the Knock-Out King. Brown began to gesticulate, his arms rising and falling in exasperation. He argued with Hood, growing more agitated. Then he cocked his fist, leaned back to bring the full weight of his body into the motion, and punched her in the head.
When the argument had started, Hood texted her son, asking him to come inside. Now he was at the door, slight and tentative in a white hoodie. He saw Brown punch his mother a second time. The boy pulled out a revolver and shot Brown in the back. Brown ran from the restaurant. The boy pursued him, still firing. Brown died on the street—one of a dozen men killed by gunfire in Chicago that weekend.
In the remarkable new book “Unforgiving Places” (Chicago), Jens Ludwig breaks down the Brown killing, moment by moment. Ludwig is the director of the University of Chicago Crime Lab, and he uses as a heuristic the psychologist Daniel Kahneman’s version of the distinction between System 1 and System 2 thinking. According to Kahneman, these are the two cognitive modes that all human beings toggle between. The first is fast, automatic, and intuitive. The second is slow, effortful, and analytical. Ludwig’s innovation is to apply the dichotomy to criminal acts. A System 2 crime might be a carefully planned robbery, in which the assailant stalks and assesses his victims before attacking them. This is what criminologists call instrumental violence: acts, Ludwig writes, “committed in order to achieve some tangible or ‘instrumental’ goal (getting someone’s cash or phone or watch or drug turf), where violence is a means to some other, larger end.” A System 1 crime, by contrast, is an act of what Ludwig calls “expressive violence”—aimed not at gaining something tangible but at hurting someone, often in a sudden burst of frustration or anger.
The central argument of “Unforgiving Places” is that Americans, in their attempts to curb crime, have made a fundamental conceptual error. We’ve assumed that the problem is instrumental violence—and have fashioned our criminal-justice system around that assumption. But the real problem is expressive violence. The ongoing bloodshed in America’s streets is just Maxwell Street Express, over and over again.
For the better part of a generation, the study of American crime has been in a state of confusion. The first destabilizing event came in the nineteen-nineties, with a sudden and sustained drop in urban crime across the United States, most notably in New York City. At the time, the prevailing view was that gun violence was deeply rooted—a product of entrenched racism, poverty, and despair. But, if that were true, how did New York’s homicide rate fall by more than half in the span of a single decade? Deeply rooted problems aren’t supposed to resolve themselves so swiftly.
The conventional wisdom adapted. Attention turned to shifts in policing—specifically, the rise of proactive tactics in the nineties. The N.Y.P.D.’s stop-and-frisk strategy, aimed at getting guns off the street, was credited with driving the crime decline. But then, in 2013, a federal judge ruled that the police’s stop-and-frisk practices violated constitutional rights. And what happened? Crime continued to fall. New York got safer even though the police stopped doing the things that we thought were making the city safer. It made no sense.
Then there were those who argued that violent crime was a matter of individual pathology: stunted development, childhood trauma, antisocial tendencies. Look closely at the criminal, we were told. But research—from criminologists like David Weisburd and Lawrence W. Sherman—showed that, in city after city, crime was hyperconcentrated. A handful of blocks accounted for a disproportionate share of violence, and those blocks stayed violent, year after year. In other words, the problem wasn’t people. It was place.
Last summer, I was given a tour of a low-income neighborhood in Philadelphia by the Pennsylvania Horticultural Society. Its program Transforming Vacant Lots has led a concerted effort to clean up thousands of vacant lots scattered across the city. The approach is simple: clear the weeds, pick up the trash, plant a lawn, put up a post-and-rail fence. The initiative works on over twelve thousand lots, and the results are striking. What once looked like a struggling neighborhood now resembles, at a glance, a middle-class one.
What’s remarkable, though, isn’t just the aesthetics. It’s that the neighborhoods where these lots have been turned into green spaces have seen a twenty-nine-per-cent drop in gun violence. Twenty-nine per cent! The people haven’t changed. The pathologies haven’t changed. The same police force still patrols the neighborhood. The only new variable is that someone comes by to mow the lawn once or twice a month. As economists like to say: How do you model that?
This is the puzzle that Ludwig sets out to solve in “Unforgiving Places.” His answer is that these episodes confound us only because we haven’t appreciated how utterly different System 1 criminality is from that of System 2. System 1 thinking is egocentric: it involves, Ludwig writes, interpreting “everything through the lens of ‘What does this have to do with me?’ ” It depends on stark binaries—reducing a range of possibilities to a simple yes or no—and, as he notes, it “focuses more on negative over positive information.” In short, it’s wired for threats. System 1 catastrophizes. It imagines the worst.
Brown’s encounter with Carlishia Hood pushed him into System 1 mode. He made an immediate egocentric assumption: if he knew that special orders were a norm violation, then Hood must know, too. “Given that System 1 assumption,” Ludwig explains, “from there it is natural that Brown believed the person in front of him was deliberately holding things up.”
Hood, meanwhile, didn’t know about the special-order taboo, so she was operating under her own egocentric assumptions. She “knew she wasn’t being disrespectful and deliberately trying to hold up everyone else in line, so the curse of knowledge led her System 1 to assume that Brown surely also knew that,” Ludwig writes. “So why was he getting so bent out of shape? She didn’t mean to be inconsiderate to the people behind her in line; she just wanted the Maxwell Street Express people to change whatever it was that she wanted changed on the burger.” Neither had the cognitive space to consider that they were caught in a misunderstanding. They were in binary mode: I’m right, so you must be wrong. From there, things escalated:
At which point he punches her—hard.
Hood’s son is standing in the doorway, watching the assault of his mother. Had he been in System 2 mode, he might have paused. He might have asked for help. He might have called 911. He could have weighed the trade-offs and thought, Yes, it’s unbearable to watch my mother being beaten. But, if I kill this man, I could spend years in prison. But he’s filled with adrenaline. He shifts into catastrophizing mode: There is nothing worse than seeing my mother get pummelled by a stranger. Brown punches her again—and again. The boy shoots him in the back. Brown runs. Hood tells her son to follow him. There is nothing worse than letting him get away. Still in System 1, the boy fires again. Brown collapses in the street.
Ludwig argues that this is what most homicide looks like. Much of what gets labelled gang violence, he says, is really just conflict between individuals who happen to be in gangs. We misread these events because we insist on naming the affiliations of the combatants. Imagine, he suggests, if we did this for everyone: “ ‘This morning by Buckingham Fountain, a financial analyst at Morningstar killed a mechanic for United Airlines.’ Naturally you’d think the place of employment must be relevant to understanding the shooting, otherwise why mention it at all?”
