Criminal Sentencing Reform: What Each Position Is Protecting

A woman in Detroit whose teenage son was shot and killed attended the sentencing hearing two years after his death and read her victim impact statement aloud. The judge sentenced the shooter to twenty-five years. She felt something she had not expected — not satisfaction, exactly, but a kind of affirmation. The community had weighed what had happened and named its cost. Her son's life had been declared worth something by the only institution that speaks on behalf of everyone. She does not want the sentence revisited.

A woman in Portland who was assaulted by a coworker was offered a choice between criminal prosecution and a restorative justice program. She chose the program. She sat in a room with him for two hours, with a trained facilitator, and described in specific detail what he had taken from her — the nights she couldn't sleep, the routes she changed to avoid walking past his apartment, the promotion she turned down because it would have required working alongside him again. He had to listen. He had to answer her questions directly. He had to make a concrete plan for accountability and repair, monitored over months. She says she received more genuine accountability from that process than she has ever seen produced in a courtroom. She does not think prosecution would have given her what she got.

A man in California has been incarcerated since he was twenty-six. He is fifty-three now. His crime was theft — $153 in merchandise — which triggered a three-strikes sentence of twenty-five years to life because of two prior burglaries he committed at nineteen. He has not been violent in three decades. His case appears in legal textbooks as an example of disproportionate sentencing. He may die in prison under a sentence designed to protect the public from a dangerous person he has not been for most of his adult life.

These three people are embedded in the same criminal sentencing debate. What each of them needs from the system is different. Only one of them is typically imagined when the debate becomes a policy question about whether sentences should be longer or shorter, harsher or more lenient. This map is about what each position in that debate is actually protecting — and why the debate is harder than it looks from any single vantage point.

This map is distinct from the criminal justice map, which focuses on the broader question of incarceration — how many people are imprisoned, for how long, under what conditions, and at what cost to communities. It is also distinct from the police reform map, which focuses on encounters before arrest. This map focuses on sentencing itself: the moment when the system decides what consequence a convicted person should face, and the four major frameworks for thinking about what that decision should accomplish.

What retributive sentencing is protecting

The retributive position holds that punishment is not a means to some further end — not deterrence, not rehabilitation, not the management of risk — but a response to what the offender deserves. When someone commits a serious wrong, they incur an obligation. The sentence discharges that obligation: it is the community's moral verdict, expressed through law, that what was done mattered and had a cost. The woman in Detroit is describing exactly this — not a utilitarian calculation, but an affirmation.

Retributivists are protecting the principle that moral accountability is real and matters. Andrew von Hirsch's landmark report, Doing Justice (1976), made the case that sentences should be proportionate to the seriousness of the offense and the culpability of the offender — neither more nor less. This is not a license for harshness. It is a constraint against it. Von Hirsch's proportionality principle was, in origin, a reform argument: the indeterminate sentences of the 1950s and 1960s, which gave parole boards enormous discretion and produced wildly inconsistent outcomes, were unjust because they treated identical offenses differently. "Desert" was the argument for consistency, not cruelty.

They are protecting the victim's claim that what happened to them was a wrong that demands a response from the community as a whole. The criminal law is unusual in that it is the state — not the victim — that prosecutes. This reflects the view that serious crimes are wrongs against the community, not merely against individuals, and that the community's response carries a meaning that no private arrangement can. When a sentence is imposed, the community is declaring: this was unacceptable. It happened to you, but it was also an offense against the shared terms of life we have agreed to maintain together. Victims who feel that restorative alternatives privatize that declaration — reducing a crime to a dispute between two parties — are protecting something real.

They are protecting the idea that wrongdoers bear genuine responsibility. A recurring critique of rehabilitative and structural arguments is that they explain crime in ways that can inadvertently dissolve moral agency — if the offender was shaped by poverty and trauma and systemic disadvantage, what remains of personal culpability? Retributivists insist that treating people as genuinely responsible for their actions, and holding them to account when they cause serious harm, is itself a form of respect for their humanity. To excuse entirely is, in some sense, to deny the full personhood of the person you are excusing. This argument is taken most seriously when the crimes are gravest.

What rehabilitative sentencing is protecting

The rehabilitative position holds that the purpose of a criminal sentence is to reduce future harm — to change the person who caused harm in ways that make them less likely to do so again. If a sentence accomplishes that, it is successful. If it does not, it has failed regardless of how deserved it felt. This is a consequentialist framework, and its strength is that it is measurable: we can ask whether programs reduce recidivism, and the answer, in many cases, is yes.

Rehabilitationists are protecting the people who can change. The brain science on adolescent and young adult development — including work documented in the sentencing context by developmental psychologist Laurence Steinberg — shows that impulse control and long-term planning continue to develop into the mid-twenties. People who commit serious crimes at eighteen or twenty-two are not the same people at forty. A system organized around desert and proportionality can sentence someone to what they deserved at the moment of the offense; a system organized around rehabilitation asks what the person needs to become in order to live without causing harm, and how the sentence can serve that process.

They are protecting the evidence that some things actually work. The dominant intellectual move of the "tough on crime" era was the 1974 article by Robert Martinson, "What Works?", which reviewed 231 rehabilitation studies and concluded that almost none produced measurable reductions in recidivism. "Nothing works" became the political rationale for mandatory minimums and the abandonment of parole. But the conclusion was contested almost immediately and was later partially recanted. Subsequent decades of research, synthesized by Francis Cullen, Paul Gendreau, and others, showed that well-designed, targeted programs — especially cognitive behavioral therapy, substance treatment, and educational programs in prison — do reduce reoffending, some by 15 to 22 percent. Protecting this evidence against the "nothing works" default is not academic. It is a claim about what a sentence should actually be designed to do.

They are protecting the public safety that comes from successful reintegration. People released from prison re-enter communities. Whether they reoffend depends, in substantial part, on what the sentence was designed to accomplish while they were incarcerated and what conditions they encounter on release. A sentence that warehouses a person and releases them with no job skills, a felony record, and no community ties has not produced safety. It has deferred it, often briefly. Rehabilitationists are protecting the public-safety case for treating incarceration as preparation for life outside, not just as punishment for life inside.

What restorative sentencing is protecting

The restorative justice position holds that crime is a violation of people and relationships, not merely a violation of law. The appropriate response is therefore not suffering imposed on the offender but repair — repair to the victim, repair to the community, repair to the social bonds the offense damaged. Repair requires the offender's genuine participation: they must acknowledge what they did, hear its impact, and make a concrete plan for accountability. Howard Zehr, who formalized the framework in Changing Lenses (1990), described the shift as moving from the question "who broke the law and how should they be punished?" to "who was harmed, what do they need, and who is responsible for meeting those needs?"

Restorative advocates are protecting what victims actually say they want. Danielle Sered, who directs Common Justice — a program working with survivors of violent crime in New York — has documented extensively that victims of serious violence often want things that prosecution does not deliver: acknowledgment, an honest account of what happened, a genuine apology, assurance that it won't happen again, and something practical to address the harm caused. What they least often name as their primary need is the number of years their attacker spends in prison. This does not mean victims are indifferent to accountability; it means that the formal system's definition of accountability often does not match what victims describe as what would help them heal. The woman in Portland is not an outlier. She is representative of a pattern in the victim satisfaction research.

They are protecting accountability that actually reaches the offender. A striking feature of the conventional sentencing hearing is that the person who caused harm almost never has to face, in a sustained and direct way, the person they harmed. They face a judge. They may hear a victim impact statement read aloud. But the kind of encounter the Portland woman described — sitting with someone for two hours, answering their specific questions, being unable to leave or redirect — rarely happens in criminal court. Restorative advocates argue that this encounter, uncomfortable as it is, is precisely what genuine accountability looks like, and that a system that replaces it with state power acting on behalf of the victim actually delivers something more institutional and less real.

They are protecting a community's role in responding to harm. The state's monopoly on criminal justice means that communities most directly affected by a crime — neighbors, family networks, local institutions — are largely excluded from the formal response to it. Community circles and restorative conferences bring those affected into the process of determining what accountability and repair should look like. This is not a soft option; the research suggests participants often find the process more demanding than conventional prosecution. It is a different theory of whose response to a crime legitimately constitutes justice.

What abolitionist and decarceral positions are protecting

The abolitionist position holds that the sentencing framework itself — the question of what punishment a person deserves — is the wrong place to start. The vast majority of people who cycle through the criminal justice system come from communities that have been systematically deprived of housing, employment, mental health services, substance treatment, and educational opportunity. A system organized around sentencing those individuals is not solving a problem; it is managing the symptoms of one while the disease continues. As Ruth Wilson Gilmore writes, abolition "means creating the conditions that make imprisonment unnecessary."

Abolitionists are protecting the communities that incarceration itself damages. Long sentences do not produce safety in the abstract. They remove people — mostly men, mostly young, mostly from specific neighborhoods — and return them years or decades later without the skills, credentials, social connections, or mental health support needed for reintegration. The communities left behind are not made safer. Research suggests that high levels of incarceration in a neighborhood can actually increase crime rates by disrupting community social structure. The children who grow up without fathers, the women who lose partners, the local economies that lose workers and gain former felons who cannot be hired — these are not side effects. They are predictable consequences of a policy choice that the sentencing debate rarely centers.

They are protecting the people sentenced for what they could not prevent. The man in California serving twenty-five to life for theft is the paradigm case. Mandatory minimums, truth-in-sentencing laws, and three-strikes statutes were designed to remove judicial discretion and enforce the community's moral verdict consistently. What they produced, in practice, was outcomes that the judges imposing them often considered unjust, in cases where the formal rules stripped them of the capacity to exercise any judgment at all. The argument for prosecutorial and judicial discretion — that the person in front of the court is a specific individual, not a statistical category — is an argument that the sentencing system, as currently constructed, too often sacrifices this. Nils Christie, in Limits to Pain (1981), asked whether a society that imposes intentional suffering on human beings in the name of justice has ever seriously interrogated whether it should, and how much.

They are protecting the possibility of investing elsewhere. American state corrections budgets average around $45,000 per incarcerated person per year — more than the cost of a year of college. The abolitionist argument is not primarily emotional. It is a resource allocation claim: the billions spent warehousing people who pose little public safety risk could be spent on the community investments that actually prevent crime from occurring in the first place. This argument does not require believing that incarceration is never necessary. It requires only asking what the same money would accomplish if spent differently.

Where the real disagreement lives

All four positions are responding to real failures of the current system. The fault lines run beneath the surface agreement.

What is a sentence actually for? This is the deepest disagreement, and it is rarely named explicitly because naming it would require each position to acknowledge how much it is assuming. Retributivists say sentences are for justice — for declaring what the offense was worth and what it cost. Rehabilitationists say sentences are for safety — for reducing the probability of future harm. Restorativists say sentences are for repair — for addressing what was actually damaged and involving the people most affected. Abolitionists say the question is wrong — the system is for control, and the alternative to asking what someone deserves is asking what would prevent harm from arising. These are not policy positions. They are different theories of what criminal law is. The debate about sentencing length, mandatory minimums, parole, and prison programming is, underneath, a debate about which of these theories should organize the system. Almost no policy argument engages at that level.

What does "accountability" mean? Every position in this debate claims the word, and each uses it to mean something different. For retributivists, accountability means bearing the legal consequences proportionate to the offense. For rehabilitationists, it means completing the work of change — demonstrating, over time, that the person is different. For restorativists, it means facing the person you harmed, acknowledging what you did, and actively making repair. For abolitionists, it means community transformation that addresses the conditions producing harm. None of these definitions is obviously wrong. All of them are genuine claims about what it means for someone who has caused harm to reckon with it. The political debate almost never surfaces this: the competing definitions slide past each other, and each side accuses the other of being "soft on accountability" while meaning entirely different things.

Does "desert" survive what we know about the causes of crime? The strongest version of the retributive argument depends on moral agency: people make choices, some choices are seriously wrong, wrong choices create obligations. But the rehabilitative and structural traditions have accumulated substantial evidence that crime is correlated with poverty, trauma, neglect, brain development, and structural disadvantage in ways that complicate the clean desert framework. This does not mean that people who commit serious wrongs bear no responsibility — very few people in this debate actually believe that. It means that "deserved punishment" cannot do as much theoretical work as retributivists need it to do without accounting for the conditions that shaped the person being sentenced. The honest retributivist answer to this challenge — that treating people as morally responsible agents is itself a form of respect — is real. But so is the counter: desert is a demanding standard that the society imposing punishment has rarely been willing to apply to the conditions it created.

Who are victims, and what do they get to say? All four positions claim to speak for victims. Retributivists argue that victims are owed the community's moral declaration that what happened was wrong. Rehabilitationists argue that victims are protected by a system that makes future offending less likely. Restorativists argue that victims are best served by genuine encounter and repair, not punishment that occurs largely without their input. Abolitionists note that people in the communities most affected by both crime and incarceration are the same people, and that a system that protects them would look different from the one we have. The disagreement about what victims deserve maps exactly onto the disagreement about what sentences are for — because victims, more than anyone, reveal what the system's purpose actually is.

What sensemaking surfaces

Sentencing is where the philosophy of punishment meets the specific human being standing in front of a judge. The abstract debates about desert, rehabilitation, repair, and abolition collapse into a single question: what should happen to this person, who did this thing, in this community, with this history? No theory settles that question, and every theory contributes something to it.

The retributive tradition's greatest contribution is proportionality — the insistence that sentences must be commensurate with the seriousness of what was done, and that a system that loses sight of this becomes arbitrary, cruel, and corrosive to its own legitimacy. The mandatory minimum era violated this principle systematically. Three-strikes laws violated it dramatically. The retributive case for proportionality is, in fact, one of the strongest arguments against the sentencing policies of the past forty years — even though those policies were often defended in retributive terms.

The rehabilitative tradition's greatest contribution is measurement — the insistence that the question "does this work?" be asked seriously and answered empirically. Warehousing people in conditions that increase their propensity to reoffend is not just cruel; it is counterproductive by the system's own safety logic. The evidence for specific rehabilitative programs is strong enough that refusing to use them is not a neutral position. It is a choice to accept preventable harm in the name of a punitive logic that is not producing the safety it claims to protect.

The restorative tradition's greatest contribution is the victim. Formal criminal justice has treated victims primarily as witnesses — people whose experience provides the evidence for a case that the state then prosecutes on their behalf, often without consulting them about what outcome they actually want. The restorative evidence that victims frequently want something different from long sentences — genuine acknowledgment, direct accountability, practical repair — is not a reason to remove victim input from the system. It is a reason to take that input seriously, rather than assuming that the state's verdict represents what victims need.

The abolitionist tradition's greatest contribution is the question it insists on asking: compared to what? The default — incarceration — is not evaluated against alternatives that do not yet exist at scale. It is evaluated against what it was like before, or what it would be like with less enforcement, imagining that the choices are between the current system and nothing. What the abolitionists have pushed into the debate, with some empirical support, is the possibility that the same resources invested in the conditions that produce crime would produce more safety than the same resources invested in the punishment of crime after it occurs. This is a claim that can be tested. It has not been, at scale, which is precisely their point.

What runs under all four positions is a question the sentencing hearing is structurally ill-equipped to answer: what does this person, who did this thing, actually need in order not to cause harm again — and what does the person they harmed actually need in order to heal? Courts are built to determine guilt and impose sentence. They are not built to hold the full complexity of what repair, change, and accountability actually require. Every serious position on sentencing reform is, at some level, an argument about building something that can hold that complexity — whether by designing sentences more carefully, by investing in programs that work, by creating processes that center victims and communities, or by asking whether the sentencing apparatus is the right place to start at all.