Day one of zookeeper training at any accredited facility includes a module on aggression. Not on how to avoid it, or how to punish it, or how to subdue the animal that displays it. On how to read it. The Association of Zoos and Aquariums, which accredits every reputable zoo in North America, requires that keepers learn to treat aggression as a diagnostic signal rather than a behavioural problem. A tiger pacing the fence line is not a bad tiger. A macaque biting its own arm is not a broken macaque. An elephant swaying in place for hours is not expressing a personality trait. These are organisms telling you, in the only language available to them, that something in the enclosure is wrong. The space is too small, or the social grouping is incorrect, or the enrichment is absent, or the animal cannot perform a behaviour its neurology requires. The keeper's job is not to stop the behaviour. It is to find the unmet need. This principle — that aggression and distress are symptoms of environmental failure, not evidence of individual pathology — is the foundation on which all modern animal welfare science rests. It is also the principle that human civilisation most comprehensively ignores when it turns to the question of what to do with its own aggressive members.
We cage them.
Strip the architecture away. Remove the marble columns and the oak panelling and the portrait of the monarch on the wall. Remove the robes, the wigs, the formal address, the procedural language that requires years of specialised training to parse. Look at the bare structure of what most modern justice systems actually do, and you are left with this: an organism that has harmed another organism is placed in a confined space, deprived of autonomy, social bonds, meaningful activity, and environmental enrichment, held there for a period proportional not to the organism's rehabilitation needs but to the severity of the original harm, and then released into the same environment that produced the harmful behaviour, with the additional burden of a permanent social marker that restricts its access to housing, employment, and community reintegration. The stated purpose is rehabilitation. The operational design is revenge. The outcome is predictable. How could it be anything else?
The Bureau of Justice Statistics tracks what happens to people released from state prisons across the United States. The numbers have been consistent for decades. Of prisoners released in 2005 across thirty states, sixty-eight percent were rearrested within three years. Seventy-seven percent within five years. Eighty-three percent within nine years. The 2012 release cohort showed similar patterns: seventy percent rearrested within five years. These are not outlier findings. They are the system's baseline output. The machine that claims to rehabilitate is producing, with industrial reliability, more of the behaviour it claims to prevent. A manufacturing plant with a seventy-seven percent defect rate would be shut down. A veterinary practice that returned seventy-seven percent of its patients sicker than when they arrived would lose its licence. The American prison system has been operating at approximately this failure rate for as long as anyone has been measuring it, and the policy response, with remarkable consistency, has been to build more prisons. What does that tell us about the purpose of the system?
The United States currently incarcerates approximately 1.8 million people. Until 2022, it held the highest incarceration rate of any nation on Earth — a distinction it surrendered not through reform but because El Salvador launched a mass incarceration campaign that briefly exceeded even American enthusiasm. As of 2025, the US rate stands at approximately 541 per 100,000 people, the fifth-highest in the world. For context: the rate in the Netherlands, where I live, is approximately 59 per 100,000. In Japan, 38. In India, 36. The United States incarcerates its citizens at roughly nine times the rate of the Netherlands. The country with five percent of the world's population holds approximately twenty percent of the world's prisoners. The annual cost exceeds eighty billion dollars — a median of approximately sixty-five thousand dollars per prisoner per year, rising above a hundred thousand in states like California and New York. This is more than it costs to send a person to Harvard. Let that settle for a moment. We spend more to damage a human being than to educate one.
The zoological question is simple: what are you getting for it?
The answer, by the system's own metrics, is a seventy-seven percent failure rate over five years, a population that is more damaged on exit than on entry, and a society that is not measurably safer as a result of the expenditure. A zookeeper presented with this data would conclude, without hesitation, that the enclosure design is catastrophically wrong. Not that the animals are defective. That the enclosure is failing them. And a zookeeper would be right.
Consider the enclosure itself. A standard cell in a United States federal or state prison measures approximately six by nine feet — fifty-four square feet. The American Veterinary Medical Association guidelines specify that a single dog weighing over forty pounds requires a minimum of twenty-four square feet of floor space, with the expectation of regular exercise outside the kennel. The European Convention for the Protection of Animals kept for Farming Purposes specifies that a pig must have a minimum lying area equal to or greater than the pig's body size, with space to turn around freely. The animal held in a US prison cell has, in absolute terms, more space than a kennel dog. It has less freedom of movement, less access to the outdoors, less social contact, and less environmental enrichment than the minimum standards applied to livestock in most European countries. Would we accept this enclosure specification for any other captive species? We would not. We would call it what it is.
The comparison is not rhetorical. It is structural. When we design enclosures for non-human animals, we begin with the species' behavioural needs: social grouping, locomotion patterns, cognitive stimulation, territorial requirements, sensory environment. We design the space to accommodate the animal. When we design prisons, we begin with the institution's needs: security, containment, processing efficiency, cost minimisation. We design the animal to fit the space. The result is an environment that systematically denies the human organism access to the conditions required for psychological stability, and then interprets the resulting psychological deterioration as evidence that the organism is dangerous and requires further containment. Can you see the loop? The cage creates the behaviour that justifies the cage.
This feedback loop is most visible in solitary confinement — the practice of isolating a human being in a cell for twenty-two to twenty-four hours a day, sometimes for months or years. The neuroscience is unambiguous. Research published in the Proceedings of the National Academy of Sciences has documented that social isolation produces measurable structural changes in the mammalian brain: shrinkage of the hippocampus, the region responsible for memory and spatial orientation; hyperactivation of the amygdala, which mediates fear and anxiety; reduction in the formation of new neurons. A study on mice found that after one month of isolation, neurons in sensory and motor areas of the brain shrank by twenty percent — a reduction that remained stable after three months in isolation. In humans, the American Association for the Advancement of Science has stated plainly that solitary confinement "fundamentally alters the brain." Former prisoners of solitary report memory impairment, loss of spatial navigation, hypersensitivity to sound and light, severe cognitive dysfunction, paranoia, hallucinations, and self-mutilation. Many of these effects appear to be permanent. We know this. The data has been published for decades. And we do it anyway.
Harry Harlow's isolation experiments at the University of Wisconsin in the 1960s — conducted on rhesus macaques, which are, we should note, less neurologically complex than humans — demonstrated that monkeys placed in total social isolation for twelve months emerged incapable of normal social interaction, unable to mate, unable to parent, and prone to self-destructive behaviour. Two of the subjects refused to eat and starved to death. Monkeys isolated for six months showed partial recovery when subsequently exposed to younger, non-threatening companions. Those isolated for a year never recovered. The critical period, in Harlow's framework, was somewhere between six months and twelve months. After that, the damage was irreversible. What happens, then, to a human being isolated for two years? For five? For a decade?
In the United States, as of the most recent comprehensive data, approximately eighty thousand people are held in solitary confinement at any given time. Some for periods exceeding a decade. The system knows what isolation does. The neuroscience has been published. The animal studies have been published. Harlow's data has been available for sixty years. The damage is documented, predicted, and then inflicted anyway, on a population that has fewer institutional protections than the macaques in Harlow's laboratory. Harlow's experiments, at least, had ethics committee oversight. American solitary confinement does not.
I introduced Kalief Browder in Chapter 1. He requires fuller attention here, because his case is not an aberration. It is the system functioning as designed. In May 2010, a sixteen-year-old boy was arrested in the Bronx for allegedly stealing a backpack. He could not afford bail — three thousand dollars. He was sent to Rikers Island to await trial. The trial did not come. For three years, Kalief Browder was held in a facility in which he was beaten by guards and by other inmates. He spent more than seven hundred days — nearly two years — in solitary confinement. A seven-by-twelve-foot cell. Twenty-three hours a day. He attempted suicide multiple times while incarcerated. In 2013, the charges were dropped. The accuser never came to court. There was no trial, no conviction, no finding of guilt. Browder was released at the age of twenty. He had entered Rikers as a teenager and emerged as a young man whose brain had been subjected to conditions that the neuroscience literature predicts will produce irreversible damage. On June 6, 2015, less than two years after his release, Kalief Browder hanged himself from an air conditioning unit at his mother's home. He was twenty-two years old.
The zoological assessment, as I noted earlier, is straightforward. A juvenile social mammal was removed from its group, placed in prolonged isolation, subjected to chronic physical stress, deprived of social contact and environmental enrichment for a period exceeding one thousand days, and then returned to an environment without any modification of the conditions that led to the original incarceration — which, it bears repeating, was the inability to pay three thousand dollars. Every element of this trajectory is predicted by the animal welfare literature. Isolation produces neurological damage. Neurological damage produces behavioural deterioration. Behavioural deterioration, in the absence of social support and environmental modification, produces further crisis. The cycle terminates in the way that untreated welfare emergencies typically terminate: the organism dies. The system did not fail to predict this. The data existed. The system predicted it and proceeded anyway. That is a different diagnosis. That is not ignorance. What is it?
There is a prison in Louisiana that sits on eighteen thousand acres of land. It is called Angola. The name comes from the country in southern Africa from which many of the enslaved people who worked the original plantation were abducted. Before the Civil War, the land was owned by Isaac Franklin, a slave trader, and it was one of the most productive cotton plantations in the American South, producing over three thousand bales per year through the forced labour of enslaved Africans. After the war, the Thirteenth Amendment to the United States Constitution abolished slavery — with an exception. The text reads: "Neither slavery nor involuntary servitude, except as a punishment for crime whereof the party shall have been duly convicted, shall exist within the United States." Except as punishment for crime. Read that again. The exception is the system. In 1869, a former Confederate major named Samuel Lawrence James leased the Angola property from the state of Louisiana and used convict labour — overwhelmingly Black men, arrested under laws designed specifically to criminalise Black existence — to resume cotton production. The prisoners were worked in the same fields, by the same methods, under the same conditions as the enslaved people who had preceded them. The legal classification had changed. The function had not.
In 1901, the state purchased the land and established the Louisiana State Penitentiary on its grounds. Today, Angola's predominantly Black prisoner population still works the fields. The prison operates a working farm. Inmates who refuse to work face punishment including solitary confinement and revocation of family visitation. The pay ranges from four to twenty cents per hour.
I include this not as an editorial but as a data point. The system's defenders describe it as rehabilitation through labour. The zoological assessment is simpler: an organism is confined, compelled to perform physical work under threat of further confinement, paid a fraction of subsistence wages, and held on land where the identical work was performed by the identical demographic under a legal framework that differed only in the name attached to their bondage. The Thirteenth Amendment did not abolish forced labour. It relocated it. The plantation became a penitentiary. The classification changed. The function persisted. Do we not see the continuity? Or do we see it and look away?
This matters for the zoological argument because it exposes something the modern justice system works very hard to obscure: punishment is not a side effect of the design. It is the design. Rehabilitation requires investment in the organism — social support, education, psychological care, gradual reintegration, and modification of the conditions that produced the original behaviour. Punishment requires only a cage. One of these is expensive and difficult to measure. The other is cheap and politically gratifying. The American system chose. It chose a long time ago. And the choice, once made, generated its own institutional momentum: the private prison industry, the prison labour economy, the political utility of appearing "tough on crime," the bureaucratic inertia of a system that employs over 400,000 corrections officers. The cage, once built, develops constituents who depend on its continuation. The cage acquires a lobby. And then the cage is no longer a response to crime. It is an industry.
Halden Prison sits in the forests of southeastern Norway. It opened in 2010 and houses approximately 250 inmates, including those convicted of murder, assault, and sexual violence. The facility looks nothing like a prison. Each cell has a flat-screen television, a small refrigerator, a private shower and toilet. The walls are decorated. The windows face trees. The corridors are painted in warm colours. There is a recording studio, a library with over ten thousand books, a woodworking shop, classrooms for education, and a kitchen where inmates prepare their own meals. Inmates move freely between buildings during the day. They interact with guards who are unarmed and who eat meals alongside them. Conjugal visits with partners are permitted. The maximum sentence in Norway is twenty-one years — a provision based on the explicit principle that the purpose of imprisonment is not punishment but the preparation of the inmate for return to society.
I can feel the objection forming. It formed in me, the first time I read about Halden. This is soft. This is naive. These are murderers. The objection is worth sitting with, because it reveals something about what the cage has taught us to expect from justice. We have been trained to equate severity with seriousness. The harsher the punishment, the more seriously we are taking the crime. But what if severity and seriousness point in opposite directions? What if the serious response — the one that actually reduces future harm — looks nothing like what we have been taught to expect?
The principle that governs this system is called normalisation. It holds three positions. First: apart from the restriction of freedom, prisoners retain the same rights as every other citizen. Second: no more security measures should be imposed than are strictly necessary. Third: all aspects of life in prison should approximate life in free society as closely as possible. The logic is that the prisoner will eventually return to society, and that the prison's job is to prepare the prisoner for that return — not to damage the prisoner so thoroughly that return becomes impossible. This is not softness. It is engineering. It is designing the enclosure for the outcome you want.
Norway's recidivism rate is approximately twenty percent. The United States' is approximately seventy-seven percent over five years. Norway spends significantly more per prisoner — roughly three times the American average. The cost is higher. The output is a human being who is dramatically less likely to harm another human being after release. The American system spends less per prisoner and produces, with seventy-seven percent reliability, a human being who will return to the cage. Which system is the expensive one?
The objection is immediate and deserves engagement. Norway is smaller, more homogeneous, wealthier, culturally different. The comparison is unfair. These factors are real. They are also insufficient to explain a recidivism differential of fifty-seven percentage points. The Scandinavian countries did not stumble into low recidivism by accident. They designed for it. They studied the animal. They asked what conditions produce the desired outcome — successful reintegration into the community — and they built the enclosure around the answer. The American system studied the political landscape, asked what conditions produce the desired outcome — the appearance of punishing wrongdoing — and built the enclosure around that answer instead. Both systems got what they designed for. The difference is in what they were designing for. And that difference — twenty percent versus seventy-seven percent — is the difference between a system that serves the animal and a system that serves the institution.
Bastoy Island, another Norwegian facility, operates as an open prison on an island in the Oslo Fjord. There are no walls, no fences, no physical barriers to escape. Inmates live in houses. They farm, fish, maintain the island, and participate in anti-violence counselling and drug rehabilitation programmes. In the facility's history, there has been one escape attempt. The recidivism rate among Bastoy alumni is among the lowest of any prison population in Europe. The critics say this works because the inmates are carefully selected. The data says the selection criteria are not dramatically different from those applied at other Norwegian facilities. What differs is the enclosure.
A zookeeper reading these two systems — the American and the Norwegian — would recognise the pattern instantly. One enclosure is designed around the institution's needs: containment, control, political optics. The animals display high rates of distress, aggression, and relapse. The other enclosure is designed around the animal's needs: social contact, meaningful activity, gradual autonomy, environmental complexity. The animals display dramatically lower rates of distress, aggression, and relapse. The zoological conclusion is not controversial. It is the same conclusion the profession reached about tiger enclosures in the 1970s: if the animal is failing, the problem is the cage. We learned this lesson for tigers fifty years ago. Why have we not learned it for ourselves?
Every system in this book began as something reasonable. Money began as cooperation at scale. The cage began as protection.
The impulse is real and it deserves respect. A community has a member whose behaviour threatens the safety of others. The community must respond. The child who is being harmed by a parent requires immediate protection. The person who has committed violence presents a real risk. The elderly woman whose home was broken into has a legitimate claim to safety. The need for protection is not an invention of the punitive state. It is a feature of every social species. Chimpanzees ostracise members who violate social norms. Wolves expel individuals who threaten the pack's cohesion. Meerkats mob predators to protect the group. The impulse to protect the community from internal threats is as old as cooperative living itself. Our impulse is sound. Our execution is the catastrophe.
The question is not whether to respond. It is how. And the answer the modern justice system provides — isolate, deprive, damage, mark permanently, release without support — is so catastrophically misaligned with the stated goal that it requires either extraordinary ignorance or extraordinary dishonesty to maintain. I do not believe most people involved in the system are dishonest. I believe they are operating within an institutional structure that has made punishment feel like justice for so long that the distinction has become invisible. The wig is real. The robes are real. The language is real. The architecture — the raised bench, the dock, the gallery, the formality — all of it serves to locate the proceedings in a register of gravity and legitimacy. It feels like justice. The recidivism data says it is not. So what is it?
The zoological correction is not complicated. Protection is sometimes necessary. Separation is sometimes necessary. A keeper who has an animal that is a danger to conspecifics does not release that animal into the group and hope for the best. The animal is separated. But the separation is accompanied by assessment: what need is unmet? What environmental condition is producing the behaviour? Is the animal sick? Is it in pain? Is it isolated? Is there resource competition? Is the enclosure too small? The separation is diagnostic, not punitive. And the goal — always, without exception — is to modify the conditions so that the animal can be returned to the group. The separation is a tool for rehabilitation, not a substitute for it. We know how to do this. We do it every day in our zoos. The question is whether we are willing to do it for our own species.
The idea that justice might mean something other than punishment is not new. It is, in fact, older than the punitive model. Among the Maori of Aotearoa New Zealand, the concept of tikanga describes a justice framework based not on punishment but on the restoration of relationships damaged by harmful behaviour. The underlying principles — tika (what is right), pono (integrity), aroha (compassion) — orient the process toward the community rather than the offender. Conferencing, a practice adapted from Maori customary processes, brings together the person who caused harm, the person harmed, and the wider community in a structured dialogue aimed at understanding what happened, who was affected, and what must be done to repair the damage. The practice was formally incorporated into New Zealand's youth justice system in 1989 and has since been adopted, in various forms, across North America, Europe, and southern Africa.
South Africa's Truth and Reconciliation Commission, established in 1996 after the end of apartheid and chaired by Archbishop Desmond Tutu, represented the most ambitious application of restorative principles to systematic violence in modern history. Witnesses identified as victims of gross human rights violations gave testimony about their experiences. Perpetrators could also testify and request amnesty from prosecution. The commission's mandate was explicitly restorative: not to punish, but to establish a complete record of what had happened and to create the conditions for national reconciliation. The outcomes were mixed. A survey by the Centre for the Study of Violence and Reconciliation found that most victims felt the process had been weighted in favour of perpetrators, and that justice — in the sense of material accountability — had not been achieved. The commission demonstrated both the power and the limitation of restorative approaches: the process can hold an account of harm with extraordinary care, but it cannot, by itself, undo the structural conditions that made the harm possible.
Neither system is presented here as a solution. Both are presented as evidence that our species has repeatedly generated alternatives to the punitive model — alternatives that start from the same good impulse (community protection) and arrive at radically different designs. The Maori model has been operating for centuries. The punitive model has been operating for centuries. One has a track record of community reintegration. The other has a track record of seventy-seven percent failure. The question is not whether alternatives exist. They do. The question is why the dominant system persists despite its own data. Why do we keep building the enclosure we know does not work?
The answer, predictably, is the same answer that recurs throughout this book: the system was not designed for the animal. It was designed for the institution. Courts require cases. Prisons require inmates. Police require arrests. Prosecutors require convictions. The political apparatus requires the appearance of toughness. The media apparatus requires drama. Every node in the network has incentives that point toward punishment and away from prevention. The individual actors within the system may be compassionate, intelligent, and well-intentioned — and many are. But the structure they operate within rewards the cage and defunds the alternative. The enclosure, once again, is not maintained by villains. It is maintained by a set of institutional incentives that are aligned with the enclosure's continuation rather than the animal's welfare. We are all inside it. Most of us never look at the walls.
The cage does not end at the prison gate. In the United States, a criminal record follows the organism for life. Devah Pager, a sociologist at Princeton, conducted audit studies in Milwaukee in which matched pairs of job applicants — identical in qualifications, education, and experience — applied for the same entry-level positions. The only difference was that one applicant in each pair disclosed a criminal record for a nonviolent drug offence. White applicants without a record received callbacks thirty-four percent of the time. White applicants with a record: seventeen percent. Black applicants without a record: fourteen percent. Black applicants with a record: five percent. A Black man with a clean record received callbacks at roughly the same rate as a white man with a felony conviction. Read those numbers again. What do they tell us about what the system is actually sorting for?
More than seventy million Americans — roughly one in three adults — have some form of arrest or conviction record. Two-thirds of formerly incarcerated people remain unemployed one year after release. Sixty percent are rearrested within three years. The system labels the organism, releases it into an environment where the label prevents it from accessing employment, housing, and social networks, and then re-incarcerates it when the predictable consequences of that exclusion manifest as further illegal behaviour. A zookeeper who marked an animal as aggressive, released it into a habitat with reduced food access and no social support, and then re-confined it when it showed aggression would not be practising animal management. They would be manufacturing the conditions for the outcome they claimed to prevent. We would fire that zookeeper. We would revoke the zoo's accreditation. And yet this is precisely what our justice system does, at industrial scale, to our own species.
This is not an analogy. It is a description of the mechanism. The permanent criminal record is the human equivalent of a territorial mark that signals danger to conspecifics: it tells every employer, every landlord, every community that this organism is not to be trusted. In the wild, such marks serve a protective function — they warn vulnerable individuals of genuine threat. In the human system, they are applied indiscriminately to individuals whose offences range from violent assault to marijuana possession, and they persist regardless of whether the individual has changed, regardless of whether the conditions that produced the behaviour have been addressed, and regardless of whether the mark itself is producing precisely the social exclusion that drives reoffending. The mark is not a protective measure. It is a mechanism for permanent exclusion from the conditions required for rehabilitation. And we apply it to one in three of our adult population.
A doctor who waits for cancer to metastasise before acting is not practising medicine. A doctor who waits until the patient is symptomatic, performs surgery to remove the tumour, sends the patient home without addressing the conditions that produced the malignancy — diet, environment, stress, exposure — and then expresses surprise when the cancer returns, is not a doctor in any meaningful sense. The doctor is a technician performing episodic intervention on a chronic condition that requires systemic response.
The analogy to criminal justice is exact. A system that waits for crime to occur, removes the individual from the community, holds them in conditions that worsen every risk factor for further offending, returns them to the same environment without modification, and then re-incarcerates them when the crime recurs is not practising justice. It is practising cleanup. Expensive, ineffective, perpetual cleanup. Is this really the best our species can manage?
The evidence for what actually works is not obscure. It has been published, replicated, and ignored. Stable housing reduces reoffending. Employment reduces reoffending. Education reduces reoffending. Treatment for substance abuse and mental health conditions reduces reoffending. Social connection — relationships with people who are not themselves involved in criminal behaviour — reduces reoffending. These are not ideological claims. They are empirical findings, published in the criminology literature by researchers at every major university in the Western world. The factors that prevent crime are the same factors that the Eight Life Areas framework identifies as requirements for human flourishing: a body that is cared for (the Vehicle), meaningful activity (the Master, the Slave), social connection (the Herd Member), agency and purpose (the Monk, the God), and a stable environment (the Zookeeper). The conditions for a flourishing human and the conditions for a non-offending human are the same conditions. This is not a coincidence. It is the central insight of the zoological approach: an animal in environmental surplus does not exhibit distress behaviours. Remove the deficit and you remove the symptom. We know this. We have always known it.
This does not mean that every crime is a product of deprivation. It does not mean that individuals bear no responsibility for their actions. It means that a system designed to prevent crime would look nothing like the current system. It would invest in housing, not prisons. In education, not courts. In mental health services, not police. In community structures, not surveillance. The Norwegian system demonstrates that this approach works even after the crime has occurred — even when the organism is already in the cage. The zoological argument goes further: a competent system would make the cage unnecessary for most of its current inhabitants, because the conditions that produce the behaviour would have been addressed before the behaviour occurred. The question is not whether this is possible. Norway, New Zealand, and Finland have demonstrated that it is. The question is whether we want it enough to build it.
I have two sons. They are seven and five. They hit each other, as siblings do. When the older one hits the younger, I do not hit him back and call it justice. I do not isolate him in his room for a period proportional to the force of the blow. I do not put a mark on his record that follows him to school the next day. I ask what happened. I ask what he needed. I ask what the younger one did that preceded the hit. I hold both of them accountable for their role in the sequence of events, and I modify the environmental conditions — the toy that was being fought over, the hunger that made them irritable, the boredom that led to provocation — so that the sequence is less likely to recur. This is not permissiveness. It is not the absence of consequence. It is the only approach that actually reduces the frequency of hitting in the long term, because it addresses the cause rather than performing retribution for the effect. Every parent reading this knows it. We do it instinctively with our children. Why do we abandon the instinct the moment the child becomes an adult?
This is not moral sophistication. It is zoology. Any parent who hits a child for hitting another child has demonstrated, in a single act, the central incoherence of the punitive model: the use of the prohibited behaviour as its own remedy. The child learns nothing about why hitting is wrong. The child learns that hitting is wrong when you do it and right when the larger organism does it. The lesson is about power, not ethics. And the lesson produces — as the data consistently shows — more of the behaviour it claims to prevent.
I am aware of my position. I write from a country — the Netherlands — with an incarceration rate of fifty-nine per hundred thousand, a system that is functional by most comparative metrics, and a society that is conspicuously less violent than many of the systems I have described. My sons have never seen the inside of a police station. My neighbourhood is safe. My experience of the justice system is entirely theoretical — a privilege that is itself a product of the enclosure I inhabit: white, European, educated, housed, employed. The cage is not for me. It has never been for me. And any argument I make about its redesign must begin with that acknowledgement, because the zoological lens, if it is honest, must include the observer. I am inside the enclosure too. I just happen to be in the part of it where the lighting is better and the bars are further apart.
But the lens does not require personal experience of the cage to recognise its failure. It requires only the data. And the data is this: a system that spends eighty billion dollars a year to achieve a seventy-seven percent failure rate, that damages every organism it processes, that falls most heavily on the poorest and most marginalised members of the population, that creates the conditions for the very behaviour it claims to prevent, and that has been producing these results for decades without meaningful structural reform, is not a justice system. It is a recycling plant for human suffering. The input is a damaged organism. The process adds damage. The output is a more damaged organism. The cycle repeats. We pay for it. We vote for it. We call it justice. And then we wonder why our society is not safer.
The good impulse — protection — remains valid. Communities need safety. Individuals who have been harmed need accountability. Children need to be kept from those who would hurt them. None of this requires the cage. It requires separation when necessary, assessment of unmet needs, modification of environmental conditions, structured reintegration, and a system that measures its own success not by the number of organisms it processes but by the number who never enter it. Every one of these elements exists. Norway has implemented them. New Zealand has implemented elements of them. The evidence is available in every criminology department in the world. The system continues as it is not because the alternative is unknown but because the enclosure, as always, has constituents. Our cage persists not because we lack the knowledge to build something better, but because the cage itself has become someone's livelihood.
The cage is where the enclosure puts the animal that broke the rules. The question that follows naturally — that I could feel forming as I researched this chapter, pulling at me with the particular gravity that personal implication creates — is this: what about where the enclosure puts the animal to learn the rules? What about the other institution that takes the young organism, places it in a confined space, regulates its behaviour through a system of rewards and punishments, and claims that the process is for the organism's own good?
The cage has bars. The school has bells. The difference is less structural than we would like to believe.