Vol. 99, No. 1 Archives – Southern California Law Review

Hiring and Firing Based on Political Views

January 2026January 2026Scott Altman*

INTRODUCTION

A law professor’s Wall Street Journal op-ed urged law firms not to hire anti-Zionist students who, in his words, “advocate hate and practice discrimination.”¹ Although discriminating against employees or prospective employees based on political speech is illegal in some places, including California, where the professor works,² employers have announced similar positions,³ and private organizations have emerged to facilitate employment denial and other social and economic penalties for pro-Palestinian students, professors, and professionals.⁴

These events are hardly unique. Employers have excluded workers because of their views on sex or gender identity, including opposition to same-sex intimacy,⁵ support for same-sex marriage,⁶ and opinions that there are only two immutable sexes.⁷

Employers have fired employees for political speech or affiliations. An English company fired a bus driver after he won an election as a member of a far-right political party with racist positions.⁸ After the recent killing of right-wing activist Charlie Kirk, some conservatives successfully called for doxxing and firing people seen as celebrating his death or criticizing his views.⁹ Disney fired The Mandalorian actor Gina Carano because she compared the treatment of Trump supporters to that of Jews during the Holocaust.¹⁰ The Metropolitan Opera fired Russian-Austrian singer Anna Netrebko for refusing to condemn Vladimir Putin after he invaded Ukraine.¹¹ Seventy-five years ago, people in some U.S. industries were fired and shunned for perceived affiliation with the Communist Party.¹²

How should we assess these actions? Should the law forbid employment denial based on political or moral views, association, or speech outside the workplace? Should our assessment depend on whether the employee has reasonable views? The existence of organized boycotts targeting unpopular views? The employer’s motive (non-complicity, association, hatred, punishment, profit, job-effectiveness)? The employer’s size, the kind of job, or the procedural posture (firing versus not hiring)?

How do these examples differ, if they do, from #MeToo era efforts to hold sexual harassers accountable, including public dissemination of alleged harasser lists¹³ and the firing of accused harassers?¹⁴ Is doxxing alleged harassers different from doxxing alleged antisemites or alleged Communists?

This Article has two aims. First, it explains why ideological employment exclusion is typically wrong and appropriately prohibited. Individuals and society suffer serious setbacks from ideological exclusion: undermining personal integrity (a person’s ability to live according to their values), chilling public discourse, and exacerbating affective polarization (the tendency to dislike and distrust those with opposing views). Employers, by contrast, have less at stake. Sometimes, employers use ideological exclusion to pursue unreasonable goals, including corruption, unjust punishment, and disrespectful uses of non-complicity. At other times, employers use ideological exclusion for legitimate purposes, including protecting their associational rights and speech, as well as pursuing profitability. However, employers can usually achieve these goals without undermining employee authenticity or valuable public ends.

Second, I use these examples of ideological exclusion to examine the moral limits of social pressure, helping to sort out a longstanding puzzle. Criticizing, shunning, shaming, and boycotting are essential to our freedom. We use them to communicate, achieve political goals, punish wrongdoers, enforce norms, and exercise associational freedom. They also threaten to undermine vital interests. We use them abusively to silence dissent, punish people unjustly, and treat excluded people with unwarranted disrespect. This puzzle is often linked to John Stuart Mill. He embraced our right to shun and criticize the objects of our contempt and valorized the social pressure created as the natural consequence of misbehavior. However, he feared these same

techniques would undermine freedom, inhibiting their targets’ speech and inducing stifling conformity.¹⁵

I offer an account of social power’s limits that might address Mill’s puzzle. The account emphasizes fair norms of social cooperation. Most people want to speak freely, control associations, exert political influence, and punish wrongdoers. We also want to avoid pressures that undermine our freedom of association and speech, as well as unjust punishment. These interests sometimes conflict, and a fair resolution of conflicts requires setting reasonable terms of cooperation—terms justified by moral arguments that all reasonable people can accept. This Article uses employment denial to explore how fair norms of cooperation limit the permissible use of intentional and unintentional social pressure.

The Article includes five Parts. Part I outlines traditional justifications for banning ideological employment exclusion: it gives employers excessive control over employees’ lives, undermines democratic institutions, constitutes wrongful discrimination, and violates employee speech rights. These accounts are promising but insufficient. Part I then offers the Article’s central argument. Allowing ideological employment exclusion unfairly distributes opportunities for living with integrity and unnecessarily contributes to two social problems: self-censorship and political polarization. Part I presents the argument in general terms. Later parts fill in details and consider examples.

Part II discusses the intentional use of social pressure to change integrity-connected behavior. It argues that intentional uses of social pressure are rarely justified when the target behavior is central to integrity. They often aim to corrupt targets or impose unjust punishments. When intentional social pressure is used to impose deserved punishments, it risks dangerous expansion.

Sometimes, employers do not aim to change employee behavior. Instead, they risk public and private harm while pursuing aims connected to employer integrity, such as non-complicity, non-association, or employer speech. Non-complicity means avoiding certain connections (such as employment) with someone else’s allegedly immoral behavior. Part III argues that we should reject employer non-complicity as disrespectful and

that most employers can protect their integrity interests in speech and association in less harmful ways.

Part IV examines employer financial goals. For most employers, these too can be advanced in less harmful ways. A conclusion follows, offering final reflections that explain why the Article’s moral arguments counsel legal regulation. It revisits Mill’s puzzle and the moral limits of social pressure, returns to the examples that began the Article, and addresses the objection that we should allow employment exclusion against people who express hatred or other harmful views.

The Article does not detail how to structure legal rules, but several models are available. Throughout Europe and in several U.S. states, employment laws prohibit dismissal and sometimes refusal to hire based on philosophical or political beliefs, as well as speech and association outside of work.¹⁶ These laws sometimes exclude hate speech, speech advocating violence, or speech that is unconnected to the speaker’s worldview.¹⁷ They also exempt some employers and sometimes provide defenses for business necessity. This Article focuses on hiring and firing private-sector employees based on their associations and speech outside the workplace on topics unrelated to their employment. Questions about speech in and about the workplace (such as criticizing an employer or seeking higher wages or unionization) and in government employment are important but raise distinct issues.¹⁸

I. The Harm of Ideological Employment Denial

Like everyone else, employers have property and associational rights. In our market economy, absent pressing contrary interests, perhaps employers should be able to select employees as they wish, including based on the employees’ views and values.

Moreover, social pressures, such as shunning, shaming, and boycotting, serve valuable functions. When people do or say objectionable things, we communicate our disdain and enforce norms through social exclusion. As people often say, “Freedom of speech does not mean freedom from consequences.” Such pressure also supplements legal regulation. For example, the #MeToo movement used shaming and shunning to punish and deter illegal behavior and reinforce values. Employers have a presumptive right to participate in social exclusion that communicates disdain and enforces norms.

Nevertheless, the examples mentioned above strike people as morally and legally problematic. What is wrong with denying someone employment based on their views, speech, and associations? In the next Section, I briefly evaluate four standard accounts and explain why they need to be supplemented. I devote the following two Sections to outlining an alternative account: ideological employment exclusion unreasonably undermines employee efforts to live with integrity and exacerbates social problems of censorship and polarization.

A. Four Traditional Approaches

Writers often approach ideological hiring and firing from the perspectives of employment law or democratic theory. Progressive employment scholars situate it in the broader problem of employer power: we should abolish at-will employment and limit employers’ control over employees’ private lives.¹⁹ They lump ideological job denial with controlling employees’ diets, smoking habits, or sex lives.²⁰ Democratic theorists focus narrowly on exclusion based on party affiliation or support for specific candidates or ballot measures. Allowing employers to pressure employees about votes or political donations distorts our political process.²¹

I agree with both views, but they do not exhaust the problems with ideological hiring and firing. Many problematic job denials are unconnected to electoral politics. Although we have reason to worry about employers controlling their employees’ lives, control over their speech seems worse than control over their smoking (and should trouble people even if they do not object to employer power generally). For these reasons, we must supplement political and power concerns with a broader account.

Two other theories might better explain the range of concerns raised by ideological employment exclusion. Perhaps denying someone a job based on their views is wrongful discrimination, much like denying them a job based on their race, religion, or sex.²² Alternatively, denying them a job based on their expression might violate their free speech rights.²³ Although these claims have merit, they are only partly persuasive without supplemental accounts.

The goals of discrimination law are disputed. Various theories emphasize anti-subordination, immutability, and formal equality. Anti-subordination accounts, which focus on unjust, persistent, and systemic disadvantages suffered by some groups, seem like a poor fit because many people who are denied employment based on their views are not subordinated. Some employers refuse to hire Trump supporters.²⁴ They are not part of a subordinated group. Nor are Anna Netrebko and Gina Carano. Although pro-Palestinian protesters are advocating for oppressed people, the protesters are not (or not uniformly) part of subordinated groups. Additionally, the companies that deny them jobs believe they support anti-subordination because they are trying to combat antisemitism. Moreover, we may wonder if some ideological groups, such as avowed racists, are justly marginalized and do not deserve protection from subordination.

Immutability is generally a problematic basis for banning discrimination.²⁵ Even if we embrace immutability, it is unclear why we should regard ideas or their expression as immutable. People change their views and choose whether to speak. So, we need a theory to explain why pressuring people to change or hide their views is unreasonable.

Finally, formal accounts of equality must be supplemented to explain which features demand equal treatment and which allow unequal treatment. Formal equality advocates often quote Dr. Martin Luther King Jr.’s statement that we should judge people not “by the color of their skin but by the content of their character.”²⁶ One would think that a person’s beliefs form the core of their character’s content and that judging people by their beliefs shows them appropriate respect. Additionally, employees’ beliefs might be relevant to their job performance in ways their race and religion rarely are.

A free speech account seems more promising. Speech rights are not limited to subordinated groups or immutable characteristics. If employers can fire or refuse to hire people based on their expressed views, the prospect of lost employment might chill speech, harming individuals and society.

However, unlike discrimination, which is unlawful for many private entities,²⁷ free speech requirements in the U.S.²⁸ apply primarily to governments.²⁹ Even in countries that restrict private entities from limiting speech, the mere power to chill speech does not imply a duty to facilitate open dialogue.³⁰ For example, if young people refuse to date Trump supporters,³¹ their dating choices might chill pro-Trump speech. Indeed, some evidence suggests that Republicans systematically lie about their politics on dating apps.³² Yet, refusing to date someone for political reasons does not violate their political or speech rights.

Of course, employment differs from dating in many ways, including the centrality of associational interests in dating. I devote much of this Article to explaining how employer-created incentives differ from other social pressures that might silence unpopular views. However, merely noting that social pressure can inhibit speech does not explain when and why we should regard it as violating speech rights.³³ Explaining this requires addressing Mill’s puzzle.

B. Threats to Personal Integrity

We might address Mill’s puzzle and better understand employment denial by focusing on the value of living with integrity. Integrity (also called authenticity) means living according to the values we embrace by pursuing projects that match our values, avoiding actions that conflict with them, and, at least sometimes, revealing our values to others. Speech is critical to living with integrity in the context of social pressure, because being silenced by social pressure can render someone hypocritical.³⁴

We cannot always live according to our values. Doing so may require actions that harm others, including undermining their integrity. A core goal of social cooperation is establishing institutions that enable people to live authentically and, when conflicts arise, distribute opportunities for authenticity equitably.³⁵

These institutions sometimes place the responsibility for living authentically on individuals whose integrity is at risk. People must often be resilient in the face of pressure, protecting their capacity for authentic action and resisting incentives for hypocrisy. If I change or hide my views because I fear my friends will abandon me or people will criticize me, the fault lies with my weak character, not with my friends or critics. Resisting hypocrisy is my responsibility because criticism and friendships based on shared values are vital institutions that could not otherwise exist. This explains why those who will not date Trump supporters do not violate their speech or association rights. Nevertheless, a society in which people with unpopular views are broadly shunned and unable to find jobs does not sufficiently support authenticity. The question raised by the examples at the start of this Article (and the core of Mill’s puzzle) is how to allocate responsibility for sustaining authenticity between individual resilience and social restraint. If we need social restraint, sometimes the law can help achieve it.

Allocating opportunities for authenticity equitably demands that we sometimes constrain the use of social pressure.³⁶ Social pressure is usually wrong when used intentionally to change other people’s behavior in ways that undermine their integrity. As I will explain in Part II, it often threatens to corrupt targets or impose social punishments on people who are not culpable. Even when used for legitimate purposes, such as just punishment, the risk of expansion to illicit goals requires caution about using social pressure intentionally to alter integrity-connected behavior.

Sometimes, we inadvertently impose social pressure while pursuing other aims in ways that risk undermining other people’s integrity. Such unintended pressure is sometimes appropriate, as it is in the dating example. However, risking other people’s integrity is sometimes unjustified because our aims are illicit or can be achieved in less harmful ways. A fair opportunity to live with integrity will not favor employers pursuing illicit aims or goals that can be achieved without undermining integrity.

C. Two Public Harms: Affective Polarization and Censorship

In addition to undermining employee integrity, ideological employment denial contributes to two public harms: affective polarization and censorship.

Affective polarization (our tendency to demonize those who disagree with us) is a growing problem. We increasingly distrust those with different views, attribute bad motives to them, avoid interacting with them, live in separate places, and get information from different sources.³⁷ This problem fuels the demand for employment exclusion. As we increasingly dislike and distrust those with different views, we avoid connections with them, including workplace connections. In turn, employment denial might exacerbate polarization by increasing ideological segregation and distrust through non-exposure.³⁸ In this way, polarization and employment denial might form a vicious cycle.

Banning ideological employment denial might encourage politically integrated workplaces and support a willingness to work together despite political differences. We should not normalize the unwillingness to associate with or cooperate with people whose views we disdain. Instead, we need opportunities to build trust in depoliticized environments. An appropriately supportive workplace is a good candidate.

This argument for workplace integration mirrors Cynthia Estlund’s claim about race discrimination. She argued that laws against employment discrimination can foster more public cooperation and social trust outside the workplace.³⁹ We ban race discrimination primarily because it wrongs victims, but a side benefit might be that doing so facilitates cooperation across racial lines. The same might be true for ideological workplace integration. We should ban ideological exclusion because it undermines employee integrity for insufficient reasons. An added benefit to banning ideological exclusion might be that it reduces polarization. Some evidence suggests that workplace exposure to people with different political views fosters political tolerance.⁴⁰ As with race discrimination, ideological integration may require legal rules to prevent exclusion.

The benefits of ideological integration do not depend on workplaces fostering political discussion or expression. Indeed, a depoliticized workplace, where such discussions are discouraged, may facilitate trust-building across ideological lines, allowing people to view their political opponents as whole individuals.⁴¹ Of course, depoliticization is not appropriate for all workplaces.⁴² But for many employers, it is feasible. The argument does not depend on businesses seeking or benefiting from ideological diversity if providing diversity does not impose significant internal burdens (discussed in Part IV). Like Estlund’s, my argument offers a reason to ban exclusionary policies. The resulting workplace diversity might lead to more cooperation.

In addition to exacerbating polarization, ideological employment exclusion might undermine the public benefits of free speech. Although the actions of a few employers who fire or refuse to hire employees for their controversial speech would not affect public discourse, many employers might (in concert or merely in parallel) exclude employees with unpopular views. They might do so intentionally or based on algorithms that screen applicants’ social media. When paired with organized groups and politicians pressing employers to boycott employees with unpopular views, these exclusions could chill expression. Chilling is unnecessary because (as I argue below) most employers’ goals are either illicit or can be achieved without resorting to employment denial or speech suppression.

This speech-chilling concern may seem alarmist. After all, most employers seem to care little about employee speech outside the workplace. If, on occasion, a few law firms refuse to hire protesting students or an opera company fires a singer, those individuals might suffer an unjust penalty. However, few people will change their behavior to avoid an exceedingly unlikely job denial.

Perhaps the worry is alarmist. However, culture wars and cancel culture appear to be intensifying, as is the use of algorithms to screen job applicants, including their social media activity.⁴³ A rational person concerned about job access might err on the side of caution, avoiding any controversial statement, even if there is no current movement to deny jobs based on a specific controversy.⁴⁴

D. Summing Up

The argument in Part I outlined three harms caused by ideological employment exclusion: it threatens integrity, exacerbates polarization, and chills speech. However, these harms are not necessarily things we should ask employers to prevent. As I noted earlier, we need reasons to ask employers to help preserve free speech and employee integrity rather than impose on employees the responsibility to resist the temptations of silence.⁴⁵

Part II will outline why employers should be responsible for some of these harms. Employers often use social pressure for illicit purposes. They also impose social pressure for legitimate purposes that could be pursued in less harmful ways. Before elaborating on these ideas, I want to recall traditional worries about employer power and anti-subordination. I mentioned that power-based accounts did not fully capture the wrong of employment exclusion, which seems more problematic when it is controlling employee speech than other aspects of employees’ lives. As to anti-subordination, many excluded employees (including famous opera singers and law students who might be denied jobs at large law firms) are not necessarily members of subordinated groups.

Although I stand by this position—that employer power and anti-subordination concerns do not fully account for the wrongfulness of ideological employment exclusion—these concerns do explain why employers’ wrongful use of social pressure is especially problematic. The harm that employers inflict on integrity, polarization, and free speech would not be possible if employers did not exercise power over their employees. The need for employment as a source of basic sustenance is why employer exclusion creates intense social pressure, and this is one reason why asking employees to resist it is problematic.⁴⁶ Additionally, although some employees who are excluded based on ideology are not among the most vulnerable in society, employer power often silences vulnerable people and threatens to suppress speech supporting subordinated groups.⁴⁷ Part II will emphasize other reasons for thinking that employer exclusion results in unjust distributions of opportunities for authenticity. However, the source of employer social power and its frequent use to undermine integrity for those with little power and suppress speech advancing their rights further justify restricting employer freedom rather than imposing duties on employees to resist pressure on their integrity.

The following sections elaborate on my core claims—that employers’ interests do not justify harms to integrity, speech, and polarization because employers either aim at illicit ends or can pursue legitimate goals in less harmful ways. Part II addresses intentional efforts to change employee behavior. Part III examines employer integrity interests in non-complicity, non-association, and speech. Part IV considers employer economic interests.

II. Intending to Change Behavior through Social Pressure

Ideological boycotts sometimes aim to corrupt their targets or punish people who do not deserve punishment. To the extent that employers aim for these outcomes, we have reason to limit their power to do so, both to fairly allocate opportunities for integrity and to combat the public harms of censorship and polarization. Even if employers intend to change integrity-related behavior for legitimate reasons, such as deserved punishment, we should often restrict such efforts, because they cannot easily be confined to legitimate cases, and we have reason to fear that they will be weaponized as tools in culture wars.

A. Corrupting Aims

Organized commercial shunning is typically called a boycott.⁴⁸ Some boycotts intentionally pressure targets to change, hide, or act against their values by offering financial or other incentives. These boycotts are corrupting (they aim to corrupt their targets) because they seek to induce others to act for morally inappropriate reasons.⁴⁹ Suppose a business donates profits to charities that I think are immoral. If I organize a boycott to induce the owners to stop supporting these causes, I am treating them disrespectfully. I try to entice them to abandon or hide their moral views or to refrain from acting on them. I do not hope to persuade them that their views are wrong. Instead, I aim to pressure them into hypocrisy by offering a bribe to compromise their morality. I would reject money as a legitimate reason to change or hide my moral views. So, I should not aim to entice others to abandon their values for profit.⁵⁰

Of course, rejecting commercial ties with someone can occur outside an organized boycott and need not aim to change anyone’s behavior. Below, I address several examples, including actions based on non-association or non-complicity. Moreover, boycotts aimed at changing behavior are not always corrupting. Some boycotts target businesses to change their profit-maximizing choices, such as an anti-sweatshop boycott encouraging a manufacturer to provide better working conditions. The boycott intentionally pressures the business to change its behavior. However, the boycott does not corrupt its decision because the choice (to exploit or endanger workers) aims at profitability. Corruption arises when social or financial pressure entices someone to change their behavior for reasons the target considers morally irrelevant.⁵¹ In the anti-sweatshop boycott example, there is no corruption, because the employer was already committed to making decisions based on profitability, and the boycotters tried to make exploitation less profitable.⁵²

Another example of non-corrupting pressure intended to change behavior is shunning or shaming people for actions that are unrelated to integrity. If I shame you for failing to contribute to a group project you agreed to join, hoping to spur increased contribution, the pressure to contribute is not corrupting, assuming your non-contribution was due to mere laziness rather than a principled opposition to contributing. Such social pressure does not undermine your effort to live according to your values by tempting you to embrace irrelevant reasons. Your laziness likely does not reflect the values you embrace.

The employment boycott against pro-Palestinian student activists seems to have been aimed at corrupting them. The professor who urged law firms not to hire anti-Zionist students supported the boycott by saying, “When students face consequences for their actions, they straighten up.”⁵³ He regarded the students’ views and associational choices as hateful. So, he aimed to deter them from expressing those views or continuing their associational choices by creating financial consequences. Admittedly, the professor offered other reasons for the boycott, including a concern that law firms or their clients might not want to affiliate with allegedly antisemitic students—associational and economic concerns rather than a corrupting incentive.⁵⁴ However, the “facing consequences” and “shaping up” language suggest that creating a corrupting incentive was a central aim of his boycott plan.

B. Punitive Aims

Social pressure that is intended to influence behavior is not corrupting if it seeks to deter and punish wrongdoing.⁵⁵ Social punishments supplement state regulation and enforcement as a means of preventing harm.⁵⁶ For example, much of the #MeToo movement’s success relied on boycotts, shaming, and shunning to punish and deter sexual harassment.⁵⁷ That movement reminds us that some speakers deserve punishment for culpable expression: their speech causes harm and contributes little to debate. Workplace harassment laws appropriately restrict harassing speech that undermines workplace equality.⁵⁸

Perhaps a comparable idea applies to hate speech and employment exclusion. Although courts in the U.S. have not upheld hate speech laws, there are plausible moral arguments for punishing hate speech, which harms people, undermines equality and respect, and (if narrowly defined) contributes little to debate or knowledge.⁵⁹ We might consider embracing hate speech curtailment by non-governmental forces, because social pressure can help combat the harm of hateful ideas by punishing those who express them.⁶⁰ Job denial might deter and punish hate speech, much like shaming and shunning deter and punish sexual harassment.⁶¹

Although social punishment for hate speech has some appeal, we should approach it cautiously. Even when free speech and enforced orthodoxy are not at stake, social punishment is problematic; it circumvents due process and other legal protections, such as the requirement of proof beyond a reasonable doubt, the right to cross-examine witnesses, the right to appeal, and rules against disproportionate punishment. Punishing speech adds new dangers to these general worries.

The appeal and the dangers of this approach are illustrated by punishing pro-Palestinian protesters. The appeal lies in trying to combat antisemitism, a problem I do not mean to minimize. Many Jews viewed recent campus pro-Palestinian protests with concern, connecting them with antisemitism and antisemitic violence. They viewed refusals to condemn Hamas and chants of “from the river to the sea” as barely coded calls for violence against Jews.⁶² When paired with rising antisemitic violence,⁶³ including violence and harassment during some protests⁶⁴ and claims of discriminatory behavior by universities,⁶⁵ they saw the need to condemn and punish antisemitic rhetoric and advocacy.

The dangers of punishing speech are equally evident in protester boycotts. One danger is punishing people for non-culpable speech—speech that includes political ideas or does not embrace racial and religious hate. Boycotters cite several objectionable actions by campus activists. Student organizations would not invite speakers who supported Israel’s right to exist,⁶⁶ and protestors refused to condemn Hamas⁶⁷ and used slogans such as “from the river to the sea,” which boycotters interpret as a call for the violent removal of all Jews from Israel.

Of course, some protesters deserve punishment. They engaged in culpable behavior, including violence,⁶⁸ threats of violence, targeted harassment,⁶⁹ the use of racial and religious epithets and stereotypes,⁷⁰ and statements that Jews deserve to die.⁷¹ However, the proposed boycott did not target only protesters who were credibly charged with such wrongdoing.⁷² Nor was the violence and harassment committed by some protesters so widespread that we can presume all protesters participated.

I believe most student protestors do not deserve punishment. The speech and associational acts targeted by boycotters differed from harassment, epithets, stereotypes, and calls for violence.⁷³ Their purposes and intended meanings were ambiguous, and some reasonable interpretations of their speech included appropriately protected political ideas.⁷⁴ This ambiguity and political content argue against culpability.⁷⁵ Although political speakers sometimes use ambiguous statements to dog-whistle intentionally hateful ideas, it seems unlikely that most student protestors had that intent.

Beyond ambiguity and political content, we should protect statements made during protests that would be unacceptable in other contexts. Discourteous and violent-sounding speech has a traditional home in political protests and social movements.⁷⁶ Provocative slogans draw attention to causes and build enthusiasm; they are rarely literal statements of belief or intent. South African anti-Apartheid protestors sometimes rallied around “one settler, one bullet.”⁷⁷ Anti-police protesters in the U.S. invoked “No justice, no peace” or “All cops are bastards.” The gun rights movement repeated, “I’ll give you my gun when you take it from my cold, dead hands.”⁷⁸ Malcolm X urged “justice by any means necessary.”⁷⁹ Such slogans vary in whether they can be interpreted as non-violent or risk inciting actual violence. But they are commonly used to protest injustices.

Some people are unpersuaded by the claim that most pro-Palestinian student protesters were blameless.⁸⁰ Although I do not share this view, even if it is correct, social punishment for the protestors’ speech is problematic. One narrow reason is that punitive attitudes toward students (especially by their teachers) are inappropriate when the students’ misbehavior stems from an insufficient understanding of complex situations. Education is a more effective and morally appropriate response to uninformed young people and the radical viewpoints they often embrace. The argument is not that these students are too young to be responsible for their actions but that wrongs born of ignorance deserve different responses from wrongs born of greed, malice, and indifference to human suffering.⁸¹

Of course, specific considerations about student speech and protest slogans do not cover all employee speech. This leads to a broader set of reasons against punishing hate speech. We often slide quickly down the slope from punishing low-content hate speech to punishing speech that offends us, advocates outcomes we think deplorable, or suggests that the speaker harbors hateful beliefs. If we leave the decision about which ideas are hateful to the private sector, we may find that the category spirals to include many political ideas that people think are morally mistaken. Perhaps critics of pro-Palestinian protests are correct that protesters’ beliefs and statements are antisemitic, harmful, and based on hatred or bias. But these beliefs and statements also represent political views; they are not mere epithets. If we declare that harmful and biased expressions are culpable even though they communicate political views, we open ourselves to declaring other views culpable. Some people regard Black Lives Matter as supporting racist views,⁸² and many other people view opposition to affirmative action and Diversity, Equity, and Inclusion (“DEI”) as racist.⁸³ That they can be described this way should not lead us to punish their expression or to let the state decide which political views are hateful enough to deserve social punishment.

Punishing those whose views we find distasteful is unjust. Such punishments are also harmful, as they threaten to censor dissenting views. This worry is hardly speculative. Social and legal pressure in the U.S. and Israel has led many advocates for Palestinian rights to fear that they cannot speak freely.⁸⁴ The Trump administration’s recent visa revocations, deportations, and attacks on universities confirm that their fears are well-founded.

The worry about expansion extends from punishing speech to punishing silence. One reason for boycotting the campus protesters was their refusal to condemn Hamas. Similarly, Hollywood professionals were fired for refusing to sign loyalty oaths,⁸⁵ and the Metropolitan Opera fired Anna Netrebko for refusing to denounce Putin. It appears that corrupting Netrebko’s choices and punishing her were among the opera’s aims.⁸⁶ Given the understandable tendency to support one’s home country, reasonable fears of retaliation for denouncing a sitting leader, and the harms of coerced speech, we should not punish a refusal to condemn people or swear allegiance.

One might hope that legally regulating social punishment could prevent its excessive use to punish valuable speech and dissenting views. Laws could permit ideological employment discrimination only against those who engage in hate speech, narrowly defined. However, this approach seems likely to fail. Lawmakers often aim to suppress dissenting views by calling them hate speech, as one can see from recent hearings in the U.S. Congress.⁸⁷ European efforts to restrict ideological employment exclusion but leave hate speech unprotected have been problematic.⁸⁸ Some people think that hate speech laws in Europe have been applied in ways that produce self-censorship.⁸⁹ In the employment context, several workers were fired for declaring that only two sexes exist, and trial courts approved their dismissals.⁹⁰ Although these trial decisions were later overturned, the disputes show how easily categories can expand.

Even if we could trust lawmakers and courts to confine ideological discrimination to narrow examples of hate speech, this approach likely could not be implemented in the U.S. due to constitutional concerns about vagueness and content-based speech restrictions.⁹¹

Injustice and censorship are not the only concerns about punishing speech. Social penalties may increasingly be used to persecute ideological enemies, exacerbating our polarization.⁹² Moreover, weaponizing employment exclusion might backfire. Many current proponents of ideological boycotts could become targets of boycotts if political sentiments change. Currently, worldwide support for Israel is falling. If Zionism became sufficiently unpopular, we might see pressure on companies not to hire Zionists, an expanded version of BDS efforts. This outcome could mirror prior hiring bans that targeted Jews, including the purge of alleged communists in Hollywood.⁹³ The danger of that outcome should worry those who would wield punitive job denial against anti-Zionists.

To summarize the argument, ideological employment denial often aims to corrupt targets by inducing them to change or conceal their principled commitments through financial inducements. It also punishes people who have not acted culpably. Even when used to punish and deter culpable behavior, the practice cannot easily be contained for practical and constitutional reasons. Of course, not all ideological employment denial aims to corrupt or punish. In the following two Parts, I consider other employer purposes.

III. Inadvertent Social Pressure and Employer Integrity

Most ideological employment exclusion is unconnected to boycotts or punishment. It occurs at the firm level, is uncoordinated with other employers, and does not aim to change employees’ or prospective employees’ behavior. Instead, firms seeking to protect their interests inadvertently impose social pressure, risking, but not intending, harm to individuals and the public. In this Part, I consider employer efforts to protect their integrity by avoiding complicity and unwanted associations, as well as engaging in employer speech. I defer discussion of the employer’s economic goals to Part IV.

I will argue that exclusion based on non-complicity is usually unreasonable. Non-association and employer speech are legitimate aims. However, for most employers, non-association is not a pressing need, and employers can pursue their speech goals without undermining employee integrity. Employees’ integrity interests and the public’s interest in employee speech, as well as the public’s interest in avoiding polarization, counsel against ideological employment exclusion when employers can pursue their goals in other ways. In some instances, equity also favors employee interests because the burdens of employer exclusion fall on disadvantaged groups and suppress speech needed to defend the interests of subordinated groups.

A. Non-Complicity

Some employers might deny people jobs to avoid complicity. Complicity means materially or symbolically supporting or benefiting from someone else’s allegedly immoral acts. Avoiding complicity is part of living with integrity, much like expressing one’s views. Although people disagree on what constitutes undesirable complicity (and, of course, also disagree about what acts are immoral), living according to one’s values, including one’s views about complicity, is part of living with integrity.⁹⁴ Unlike the corrupting and punitive aims discussed above, avoiding complicity does not aim to disrupt other people’s efforts to live authentically (as a goal or an intended means of pursuing a goal).

Non-complicity is not a typical reason for ideological employment denial. Few employers believe they become complicit by hiring someone whose views or acts they regard as immoral. Because commerce connects everyone with people who spend money on things they think are illicit, most people embrace a constrained interpretation of complicity.⁹⁵

However, employers who believe that paying employees makes them complicit in employees’ speech and actions present a conflict between two integrity interests. If we protect the employer, a potential employee must remain silent or risk unemployment. If we do not, an employer must become complicit or cease employing people. The decision about who to protect depends on which person should accept the setback to protect the other’s integrity. This demands an equitable allocation of opportunities for living authentically.

How should we evaluate this conflict? We might ask who is morally right. Does the employer have a correct view of complicity? Are the employee’s actions immoral? However, the value of living with integrity (for both employer and employee) persists even for people with mistaken views about morality or complicity. Moreover, when we move from discussing moral to legal questions, it seems unlikely that a liberal state should base employment rules on the truth of employees’ or employers’ controversial moral views, or that we should trust the state to make such decisions.⁹⁶

We might be tempted to reject the employer’s view of complicity as unreasonably expansive. Many people think it is implausible that an employer becomes complicit in an employee’s actions merely by employing them, particularly if employment does not advance the employee’s goals beyond providing a salary. However, this view of complicity is no more expansive than the views commonly embraced by consumers who participate in boycotts, such as those who refused to buy sandwiches from Chick-fil-A because its owner donated to charities opposing same-sex marriage.⁹⁷

In a recent article on vendor discrimination, I argued for an approach based on reason-giving. Religious vendors sometimes claim that selling wedding-related goods to same-sex couples makes them complicit in actions the vendors regard as immoral. This creates a dilemma: If the law permits them to discriminate, same-sex couples are denied equal status in commerce. Conversely, if the law disallows discrimination, vendors become complicit or must change their businesses.

Justifications for rules should be acceptable to the people they disadvantage. The vendor offers a justification based on a seemingly reasonable aim—avoiding complicity—that any person should consider valuable and be willing to facilitate. However, this neutral description masks its underlying disdain. The same-sex couple must suffer discrimination because the vendor views actions central to their identity as immoral. Accepting this reason requires customers to treat the vendor’s desire to avoid their supposed sin as reasonable. No one should have to accept as reasonable the perspective that they are sinful, because doing so undermines their dignity. In contrast, when the same-sex couple seeks service from the vendor, they are not trying to express or enact any disdain for the vendor’s religion. I concluded that, in most cases, we should regard the vendor’s effort to live authentically by discriminating as unreasonable, because its justification asks victims to accept others’ disdain for them as a reason for rules that disadvantage them. This is especially unreasonable when vendors have other means of maintaining integrity, such as finding other jobs.⁹⁸

To be clear, my argument did not rely on condemning the vendor’s beliefs as homophobic or immoral. It would be equally disrespectful for a feminist baker not to provide a cake for a fundamentalist wedding because she thinks fundamentalist marriages are oppressive. The disrespect comes from invoking someone’s sinfulness or the evil of their views to justify

denying commercial contact. It does not lie in judging which party has morally correct views.

A comparable argument applies to most cases of complicity-avoiding ideological employment denial. For example, imagine a law firm fires or refuses to hire a lawyer whose social media post said that Palestinian violence was an understandable response to Israeli oppression. The firm believes that employing this lawyer would make it complicit in terrorism or antisemitism. Not hiring the lawyer inhibits her ability to express solidarity with Palestinians and, therefore, to live authentically. The lawyer’s desire to express her views does not rely on disdain for the law firm or its values. However, the law firm’s non-complicity efforts depend on its disdain for the lawyer’s deeply held views. By invoking non-complicity with terrorism or antisemitism, the firm asserts that the lawyer’s values are evil—that is, antisemitic or terrorist-sympathizing. The law firm’s non-complicity justification masks disdain in the same way as a feminist baker who does not want to sell wedding cakes for fundamentalist weddings.⁹⁹ Again, the problem is not that the law firm is wrong to think the lawyer’s position is antisemitic or that the feminist baker is wrong to condemn fundamentalist marriages as oppressive. The problem is expecting someone committed to the other side to accept your moral judgment as a reason to exclude them from commerce.

Of course, employment exclusion differs from vendor discrimination in several ways, including having different material consequences. Being fired or not hired often imposes financial and personal costs far greater than being turned away by a vendor. However, my point concerns the interest in non-complicity, which is similar in both examples. In any case, the more significant material stakes for employees only strengthen the argument for restricting employer action based on non-complicity compared to the vendor argument.

Some scholars argue that we should accept non-complicity as grounds for excluding people from pursuing hateful or oppressive goals.¹⁰⁰ They might forbid most employer exclusion based on ideology but allow employers to deny jobs to racists or antisemites to avoid complicity. However, their position faces problems.¹⁰¹ Like the argument for punishing hate speech, this position has the potential to expand widely. Suppose employers can exclude antisemites based on non-complicity. Can they also exclude abortion supporters or opponents because people with these views favor murder or the oppression of women? What about supporters or opponents of DEI if the employer thinks their views are racist? Many positions on controversial issues can be described as hateful and oppressive. Exempting hateful positions from a general ban on employment exclusion might expand to cover a wide range of political speech.

My prior article concluded that vendors whose consciences preclude them from serving customers should protect their integrity by changing businesses.¹⁰² Their reason for excluding customers is disrespectful, and they have alternative ways to live with integrity. The same applies to employers who object to employing people whose values they reject. They should not employ people if they think employment makes them complicit in their employees’ speech or political acts.

This might seem unreasonably burdensome to employers. However, several considerations support it. First, disrespectful reasons should not be counted when assessing fair opportunities for living with integrity. Second, the position that employing someone makes you complicit in their activities is hard to accommodate. People who hold such views have as much interest as anyone else in living with integrity. But by embracing this view, they make outsized demands on others to help them avoid complicity. The harm they cause is not merely the undermining of one employee’s integrity. They create incentives for all potential employees to hide their views, thereby harming the public by exacerbating censorship and polarization. Unlike the values of speech and non-polarization, non-complicity serves only private ends. There is no public interest in expansive non-complicity views, parallel to the public interest in open discussion and non-polarization. Third, few people hold such broad complicity concerns, and fewer still hold them intensely. The law needs mechanisms to distinguish sincere invocations of a view strongly felt from pretextual and exaggerated versions. One sensible solution to this sorting problem is to ask people to protect their integrity, even at significant cost, rather than imposing costs on others.

B. Employer Non-Association and Speech

Employers might advance several associational interests to justify ideological exclusion. One interest is that they do not want to be near people whose views they dislike. This interest is sometimes a version of non-complicity, relying on the idea that the employer is morally tainted by proximity (discussed above). However, it might not reflect complicity; it might reflect discomfort with being around people with different views or distrust of people with different views.

We should resist exclusion claims based on employer discomfort and distrust. Discomfort often masks disdain and has been a code word to justify discrimination. More broadly, we should combat rather than indulge discomfort and distrust deriving from disagreement. Discomfort and distrust are at the core of affective polarization. The law should encourage people to become comfortable working across political and moral divides and to build trust by working together.

Other associational interests are not urgent for most businesses. They do not resemble intimate associations, in which bonds of affection and admiration are central to the relationship, or private groups devoted to causes or common interests, in which shared purposes are central to achieving the group’s goals. For this reason, courts have not accepted associational interests as pressing for most businesses.¹⁰³

Employers sometimes invoke speech interests to justify excluding employees. They hope that excluding certain workers will communicate their values to employees or potential customers. For example, to show customers and workers that the company embraces equality and diversity, it might refuse to hire workers who reject those values or fail to show adequate commitment to a specific view of equality in a diversity statement. Or, to show support for Israel, the employer might exclude those who protest Israel’s existence.¹⁰⁴

This argument has several problems. The main difficulty is that employers have many ways to signal their commitment to values without undermining others’ integrity or exacerbating social problems. They can state their values directly and demonstrate how those values shape the firm. Excluding those with different values is not necessary to achieve the end. Another problem is that such exclusion may have the effect (or even be a pretext for) objectionable discrimination or corruption. For example, an employer might want to signal its commitment to equal treatment of women by refusing to hire abortion opponents. In doing so, it might discriminate against people based on their religion or pressure potential employees to hide their views.

Although most employers do not need ideological exclusion to protect their speech and associational interests, there are exceptions. These might include organizations dedicated to political, religious, artistic, and moral goals.¹⁰⁵ Religious organizations and those with creative goals already demand the right to discriminate in some contexts.¹⁰⁶ They would likely seek similar exemptions from any law forbidding employer exclusion based on speech. Indeed, Disney tried (unsuccessfully) to defend firing Gina Carano by invoking its speech rights.¹⁰⁷

Exceptions may be warranted for mission-driven organizations, such as political groups or charities. Non-profits and political organizations may need employees who share their aims. Some for-profit employers have similar needs. For example, specialist law firms, such as labor-side employment firms, may screen for lawyers who favor their preferred clients.

The appropriate scope of such exemptions is complex. However, such exceptions must be narrowly drawn to prevent expansive or pretextual use that undermines the benefits of laws restricting ideological employment exclusion.

To summarize this Part, employers seeking to avoid complicity with employees’ views and actions should stop employing people if doing so compromises their consciences. Asking others to accept an employer’s disdain for their deeply held commitments does not rely on reasons the disadvantaged party can accept. We should reject employers’ discomfort as a justification for denying jobs and instead encourage people to become comfortable working together despite moral and political divides. Moreover, other associational interests are not pressing for most employers, and employers can advance their speech interests without ideological employment denial. That said, there are exceptions to these conclusions. Mission-driven employers, for instance, should be allowed to deny employment in narrow circumstances to protect their speech and associational interests.

IV. Inadvertent Social Pressure and Employer Economic Interests

Some employers use ideological exclusions for economic reasons. In most cases, however, employers can achieve their economic goals without undermining employee integrity, chilling speech, or exacerbating polarization. Below, I illustrate this conclusion using three economic interests: customer and employee retention, workplace efficiency, and merit.

A. Customer and Employee Retention

Businesses worry about their reputations. Employers may prefer to exclude employees with unpopular views because they do not want others to think the employer shares those views or to associate those views with their businesses. Such associations could lead to lost customers or difficulty recruiting and retaining employees.

However, this attribution argument is usually unpersuasive. For low-level employees, especially in large companies, few observers will know their views, and fewer still will assume that an employer shares them. In part, this is endogenous (that is, the inference drawn will depend on the background rule). If the law forbids ideological hiring and firing (and people know about the law), observers will be less likely to presume that an employer shares their employees’ views. Conversely, if ideological hiring is permitted and becomes typical, observers will have more reason to assume that an employer shares an employee’s views. That said, for smaller firms or high-level employees, the non-attribution interest may be weightier.

A related concern is when employees, particularly celebrities, are part of the product sold. Customers might not care about the employer’s views, but they may be so displeased by the employee’s statements that they look elsewhere for goods and services. Although this concern is legitimate, the Netrebko example highlights the dangers of protecting it. Perhaps the Metropolitan Opera fired Netrebko because it feared losing ticket sales or donations. However, unless the law requires evidence of lost income, this justification can easily mask employer dismissals based on punishment, corruption, discomfort, or non-complicity. Indeed, opera fans continue to

patronize venues that hire Netrebko,¹⁰⁸ suggesting the Metropolitan Opera’s fears may have been unfounded.

B. Workplace Efficiency

Employers may worry that employees will be uncomfortable around coworkers with unpopular views. This might lead to workplace inefficiencies or difficulty recruiting and retaining employees.

One might be tempted to dismiss this claim by pointing to studies showing the benefits of workplace diversity. However, these studies are equivocal, suggesting both benefits to creativity and drawbacks to efficiency.¹⁰⁹ If hiring employees with unpopular views leads to workplace disputes and distrust, the efficiency costs might outweigh the creativity benefits of viewpoint diversity. Moreover, employers can achieve substantial viewpoint diversity (and thus creativity) while excluding people with extreme views.

Indeed, one scholar argues that we should allow employers to discriminate based on ideology to protect workers from the harms of political polarization. According to R. George Wright, polarization and pressure to take stands on public controversies have politicized our workplaces to the detriment of worker comfort and efficiency.¹¹⁰ Although Wright acknowledges that banning ideological exclusion might support a social consensus that workplaces are depoliticized spaces—as I urge—he notes that this optimistic forecast lacks solid empirical support.¹¹¹

Wright’s skepticism is fair. My suggestion that we use workplace integration to fortify society against polarization might not work. However, reducing polarization is too important to allow the metastasis of political isolation and disdain. Even if we cannot guarantee success, we should pursue strategies to encourage integration and combat polarization.

Concerns about workplace instability stemming from political opponents working together can be addressed through alternative means. People who disagree about politics and religion have successfully shared workplaces. Managers can support such cooperation by limiting political discussions in the workplace,¹¹² modeling respectful interaction, and avoiding corporate stances on political matters that show disdain for those who disagree.¹¹³ Despite limits on political discussions, coworkers will likely learn that others hold vastly different views and may recognize that their disagreements do not impede cooperation.

Employers might seek to exclude employees or potential employees if they believe their speech outside the workplace signals a higher risk of harmful workplace behavior, such as workplace harassment. For example, chanting “from the river to the sea” in a campus protest might be protected speech, assuming it did not target Jewish students for harassment, but chanting it at a Jewish coworker during working hours could be illegal harassment.

Yet employers need not screen employees’ speech outside the workplace to prevent harassment. Absent other information about an employee, there is little reason to assume that employees will behave in the workplace as they do online or in political protests, or that they will not follow workplace rules.¹¹⁴

C. Merit

Employers might prefer employees who do not use intemperate language or express extreme or reductionist views. Such employees might be more cooperative workers or more subtle, better-informed thinkers. In this regard, screening for employee speech resembles testing for personality, knowledge, or intelligence.¹¹⁵

However, employers can assess these qualities without examining employee speech. They can instead rely on personality tests, competence tests, grades, and references.¹¹⁶ These alternatives would not undermine employee integrity, chill speech, or exacerbate polarization. There are concerns that employees may lie on personality tests to secure jobs.¹¹⁷ Even so, there is inadequate evidence to suggest that social media screening is superior to personality tests in predicting employee quality, although, admittedly, social media screening might be less costly.¹¹⁸

Employers might respond that no single metric predicts employee success equally well for all workplaces. In the absence of settled science, they should be permitted to use the screening tools that seem most suitable to them. This argument has some appeal, because employers are right when they say it is hard to predict employee quality. However, it also has the potential to be abused. Employers might cite intemperate language as a pretext (or be unconsciously biased) and exclude views they dislike by labeling them intemperate or reductionist.

That said, some controversial employee speech may justify merit-based exclusion if it reveals a candidate to be unqualified for specific jobs. Job candidates who deny established scientific consensus may be unqualified for jobs that require scientific or policy expertise. Those who embrace conspiracy theories despite contrary evidence may be unqualified for jobs as journalists or educators. However, even this exception needs to be narrowly interpreted. Science denialists or conspiracy theory enthusiasts should not be disqualified for jobs unlikely to require scientific abilities or knowledge of public affairs.

Employee speech can sometimes render an employee unfit for their job. The speech might undermine the employee’s ability to accomplish core job functions. Examples include teachers who make public remarks that demean their students, thereby undermining the trust needed between students and teachers, or police officers whose remarks demean the citizens they serve.¹¹⁹

To summarize, most employers’ economic concerns can be addressed without ideological job denial. Employee speech will rarely be attributed to employers or undermine their profits. There are some exceptions, such as celebrities and corporate leaders, who are integral to their employers’ brands. However, permitting job denial in these cases must be approached with caution lest business reasons be used as pretexts for other goals. Workplace unrest stemming from varied employee views can be addressed through effective management. Finally, screening employee speech is unlikely to be the only, or even the best, way to ensure employee quality. Other indicia of merit can be used without chilling speech and fostering polarization.

V. Final Reflections

A. Do We Need Legal Regulation?

We create social pressure through speech, association, and spending. We criticize, boycott, shame, and shun one another while pursuing aims central to our lives. However, these tools can be misused to suppress speech and diversity, undermine efforts to live according to one’s values, and exacerbate polarization. I have argued that most ideological employment exclusion falls into the second category: employers abuse social pressure either to pursue illicit ends or to achieve reasonable goals that could be achieved in alternative ways.

Sometimes, when people misuse social pressure, the best solution is counterpressure; we criticize, boycott, shame, and shun those who abuse these social tools. However, in other cases, legal regulation of social pressure is more sensible, such as when social pressure is used against the powerless, when mobs employ social pressure, or when pressure and counterpressure threaten to become a vicious cycle of revenge.

Legal tools are difficult to implement when people use social pressure for both valuable and illicit ends that are hard to distinguish.¹²⁰ However, if my arguments are correct, employment exclusion is rarely needed to pursue legitimate ends. Additionally, polarization and the targeting of unpopular views suggest that social counterpressure and self-restraint will fail as solutions. Thus, a regulatory solution likely makes sense.¹²¹ Legal rules governing social pressure can mitigate some of its downsides, helping to focus that pressure on cases where it causes less harm.

Admittedly, legal rules alone cannot prevent ideological employment denial. We must also build a social consensus against ideological job denial and in favor of free speech and cooperation across ideological lines. Employment laws are notoriously difficult to enforce, especially when

applied to hiring. Nevertheless, employment laws can provide occasional remedies and help shape this social consensus over time.

B. Moral Uses of Social Pressure

This Article began by asking what uses of social pressure are morally permissible and whether we can resolve Mill’s concern that social pressure might serve valuable ends while also censoring ideas, undermining integrity, and inducing unnecessary conformity. The tension Mill identified cannot be eliminated; social pressure, which is often a byproduct of rightly protected activities, sometimes threatens other people’s freedom. However, we can mitigate concerns about censorship, integrity, and conformity by protecting valuable uses of social pressure that cannot be pursued without risk to others and by discouraging uses in two kinds of cases: when social pressure is used for illicit ends and when it is used for legitimate goals that can be pursued in less harmful ways. Such social restraint requires institutions that fairly allocate opportunities for living authentically. They might help us avoid social pressure that needlessly curtails public debate and exacerbates polarization. They can also be alert to the possibility that social pressure will be used against people with limited power, inequitably allocating opportunities for integrity and suppressing speech necessary for emancipatory social movements.

Identifying legitimate uses of social pressure is easier when we recognize that it falls into several categories. One category of social pressure arises from its intentional use to change other people’s behavior. This pressure is permissible if it is not pursued to corrupt others or unjustly punish them. For example, we can try to change others’ behavior when they act carelessly or out of self-interest. We can shame ill-mannered strangers and shirking friends, or boycott companies that pay their workers too little. Such efforts do not corrupt their targets because the targets do not aim to exclude financial gain or personal reputation from their decisions. However, intentionally using social pressure becomes problematic when it disrespects targets by offering financial incentives to violate moral commitments, such as boycotting a merchant to deter it from donating to a charity or refusing to hire college students to deter their political speech.

Social pressure is also problematic when used to punish speech. Social penalties are often directed against those whose values we dislike, people whose speech is ambiguous, or those who offend us but have done no wrong. Even when we penalize wrongful speech, such as hate speech, we risk its further use in less straightforward cases, punishing the innocent and exacerbating tensions over culture war issues. In our pluralistic and increasingly polarized society, social penalties for speech should be avoided unless we find ways to restrict them to clearly culpable speech.

A second category of social pressure that undermines integrity arises as an unintended effect of actions taken while pursuing other, often valuable, goals. The reasons for imposing unintended social pressure can be central to the integrity of the person imposing that pressure and cannot be pursued otherwise. An example is refusing to befriend someone whose values I abhor. Friendship requires affection and admiration. So, the good of friendship cannot be pursued without risking the integrity of excluded people.

However, social pressure is inappropriate when the goals pursued are illicit and when we do not fairly distribute opportunities for living authentically. Sometimes, the interest pursued when imposing such risks is illicit. I argued that we should treat non-complicity as an illicit reason for employment denial. This is partly because it disrespectfully relies on reasons that undermine self-respect and partly because employers can avoid complicity without discriminating. Even when employers have good reasons for ideological exclusion, such as protecting speech and associational interests or advancing economic goals like profit and employee retention, most employers can adequately advance those goals through other means. We should insist they use those means because ideological job denial threatens to chill speech and exacerbate polarization. Moreover, because jobs are central to thriving in our society, social pressure created through job denial is often difficult to resist.

These ideas do not solve all the difficulties raised by Mill’s puzzle (and might not appeal to Mill). But they suggest some solutions.

C. Revisiting the Examples

The paper began with examples: job denials based on speech about Zionism, sex, and gender; the refusal to condemn Vladimir Putin; political comments about Trump and Charlie Kirk; membership in a racist political party; and perceived affiliation with the Communist Party. I asked what was wrong with these exclusions and whether they differed from doxing and shaming sexual harassers during the #MeToo era.

Boycotting anti-Zionist students is corrupting, seeking to entice them to abandon their principles for money. It also seeks to punish people who might not be wrongdoers. Even if you think they are wrongdoers, the boycott embraces norms that might spread to punish non-culpable people and undermine cooperation. Similar arguments condemn the firing of Gina Carano, Anna Netrebko, the British National Party employee, and those who celebrated Charlie Kirk’s death or criticized him after he was killed.

The #MeToo movement was different. Sexual harassers deserve to be punished, and legal rules specifying what constitutes harassment help limit the punishment to wrongful harassers. Moreover, because sexual harassment likely does not represent the deeply held views of harassers, it does not threaten integrity, speech, or polarization in the same way as ideological employment exclusion.

The arguments marshaled against these examples do not apply to employers who have good reasons for ideological employment denial that cannot be pursued in other ways. Organizations devoted to specific missions are the most obvious.

D. But What About the Nazis?

Some readers might think we should generally disallow ideological employment denial, but we should allow employers to exclude people who express hatred toward groups based on race, religion, and sex. They might contend that all my arguments lose force when applied to people who express hate. Such individuals do not deserve respect, so we do not owe them restraint from corrupt pressures or protection from insulting demands for non-complicity. They deserve punishment, so we should feel free to harm them through job denial for their culpable speech. Their presence in the workplace will likely undermine efforts to foster respect and cooperation through workplace integration. They will likely tarnish employer reputations and mistreat coworkers, which provides employers with sound moral and business reasons to exclude them.

This position, embraced in much of Europe, has evident attractions. However, we should resist it for several reasons, especially in the U.S. First, I disagree with the claim that we can withhold respectful cooperation from people with unreasonable views. We need not offer them reasons they can accept if they are unwilling to seek reasonable terms of cooperation. However, we must still respect their rights and cooperate with them on terms a reasonable person can accept.¹²² We fail to respect their rights and cooperate if we try to corrupt them or demand that they accept denigrating reasons for our actions.

Second, as noted above, we should not trust employers or the government to pick which views are so heinous that we will punish their expression through employment discrimination. The likely targets will not just be Nazis (and might not even include them). If progressives are in power, those punished might include people who oppose abortion rights or same-sex marriage or who embrace traditional roles for women. If the right is in power, targets might include people who favor abortion rights, gender-affirming care for transgender children, or anything related to DEI. If power is divided, we may see a patchwork of protections in red and blue states, creating even more incentives for residential segregation by ideology, mutual distrust, and social pressure used as a tool for retaliation. Moreover, it is unclear whether punishing purveyors of hate is the most effective way to combat its spread. It might merely fuel such movements’ feelings of isolation and persecution.

Third, as noted earlier, a law banning employment exclusion based on ideology for everyone except Nazis (or hate speech, or harmful speech, however defined) would likely be unconstitutional in the U.S. If so, we must choose between offering no protection against ideological employment denial and offering protection that extends to Nazis. Vulnerable people who hold views that we should respect far outnumber people with hate-based views. We should protect them even at the cost of helping a few Nazis.

Finally, people can work cooperatively even knowing that some of their coworkers hold views that express contempt for their values. We work with people whose religious views include predicting our eternal damnation, who believe our attitudes toward abortion condone murder or oppression, who think our views about merit show us to be racists, and who think we recently voted for a tyrant. Justices Scalia and Ginsburg worked together and formed a friendship, even though his religion might have characterized her as having spent a career supporting the murder of babies. At the same time, her feminist commitments might have portrayed him as devoted to ideals that oppress women. We should not presume that some subset of people with views we find most hateful will misbehave or undermine efforts at peaceful coexistence, or that we cannot find ways to coexist in proximity. Rather than assume cooperation is impossible, we should try working together.

99 S. Cal. L. Rev. 1

Download

Scott Altman*

*Virginia S. & Fred H. Bice Professor of Law, University of Southern California Gould School of Law. For helpful comments, I thank Catherine Fisk, Felipe Jimenez, Greg Keating, Dan Klerman, George Letsas, Erin Miller, Alain Pottage, Marcela Prieto, Daria Roithmayr, Irit Samet, Mike Selmi, Amy Sepinwall, Mike Simkovic, Danny Sokol, Raphaële Xenidis, and participants at the Sciences Po Law School Faculty Colloquium.

Systemic Absolution

January 2026January 2026Shima Baradaran Baughman*

Introduction

Ancient religious tradition forms the bedrock foundation for the prevailing approach to criminal punishment. American criminal penal statutes are steeped in retributivism, which at its core requires punishment proportional to the crime committed—in other words, the Old Testament, lex talionis, or “an eye for an eye.”¹ Modern criminal statutes and prisons are largely set up to punish offenders who harm someone proportionate to the harm they caused, as determined by the legislature, often through incapacitation rather than corporal punishment.² Traditionally, corporal punishment was inflicted on the offender in proportion to the crime the person had committed.³ Historically, however, criminal tradition also involved a community coming together periodically to wipe the slate clean of crime, typically with a ritual involving a sacrifice.⁴ It was important for the community to join together to punish a person who committed a crime, so they would publicly gather to witness corporal punishment for individuals who committed serious crimes.⁵ But equally important was a communal ritual that periodically allowed society to collectively absolve itself of crime.⁶

This concept is one referred to here as systemic absolution, where society periodically and systematically cleanses or atones itself of the burdens of crime, offers forgiveness for offenders who have been punished, and allows the community to be reunified. Robin Steinberg, a prominent legal activist and founder of The Bail Project, based a wrongful conviction defense on an ancient Jewish scapegoat ritual symbolizing atonement.⁷ She explained that in Leviticus, a book in the Old Testament, “members of a community gather[ed] their sins and load[ed] them on the back of a goat. They then sen[t] the goat out into the desert, hoping to alleviate themselves of the burden and guilt of their failures.”⁸ Steinberg analogized to this ancient practice to instruct the jury that thousands of years ago, there existed a means for people to be freed of their crimes, hoping to “undo some hardwired biases and tap into the jurors’ sense of injustice.”⁹ This ritual allowed society to sacrifice one goat and allow all debts and crimes to be forgiven. Steinberg challenged the jurors to determine as representatives of the community that they were willing to shed the weight of a woman’s former conviction and allow her defendant to go free. This discussion of atonement and ancient Jewish ritual successfully won over a criminal jury, but its application might not be limited to a closing argument at trial. The modern corollary to the scapegoat ritual is the holy Yom Kippur Day of Atonement where, once a year, Jews collectively release the sins from the past year and ask for forgiveness to wipe the slate clean.¹⁰ Other cultural traditions incorporate the idea of atonement or tabula rasa, wiping the slate clean with periodic repentance and group sacrament rituals.¹¹ While criminal punishment today is loosely based on Law of Moses principles,¹² what is missing beyond punishment is a path towards systemic absolution.

Common in the Catholic faith, absolution is a term associated with forgiveness, or removing the weight of sin or crime.¹³ A person approaches a priest for absolution from a sin, and after prayer and repentance a person receives forgiveness from God for that sin.¹⁴ The absolution considered here is not a religious one but a systemic one, where society provides a formal release from punishment, offering relief to the person receiving it. While the concept of absolution has never been used in criminal law, criminal law has largely conceptualized punishment using religious philosophy. While there are some semblances of forgiveness in existence in modern criminal law, including clemency and pardons, they are used sparingly as political tools at the end of a presidential term to benefit select individuals, not having large-scale impact on many criminal defendants.¹⁵ Historically, though, criminal punishment and accountability have typically been accompanied by the concept of societal forgiveness, or absolution.¹⁶ Indeed, many cultures have had practices for both punishing and absolving people of crimes. Like under ancient Jewish law, corporal punishment was common when a crime or sin was discovered, but the scapegoat ritual periodically eliminated sin and crime, allowing reconciliation of the community.¹⁷ While many cultural traditions have provided a way for the community to absolve former crimes, there is no longer any societal mechanism to absolve society of crime.

While forgiveness and absolution have always existed privately for citizens and have even been used in limited ways by the executive branch, these concepts have not been conceptualized in broader structural ways. Since the 1500s, common law systems of criminal justice have had a prosecutor represent the “State” or “King” and bring crimes against it to court.¹⁸ The modern criminal justice system similarly considers any crime committed against an individual as a crime against the entire state.¹⁹ As such, an individual never decides to “charge” a case.²⁰ Rather, this decision is left to the discretion of the prosecutor who enacts justice for acts against the community or state.²¹ While our system of justice is based on what is best for society, our system of forgiveness or absolution is one considered only at the individual level (apart from dealing with criminal behavior), not collectively or systemically. An individual harmed by a crime may have the State bring his case to justice, but forgiveness of the crime only includes participation of the victim and the defendant.²² Restorative justice depends on both the victim and offender being willing and responsive to a process of unification and healing, which is not always the case. But the burdens of crime and incarceration are experienced not just by the victim and defendant, but by the community and society at large. And while restorative justice principles that bring the two together to resolve their conflict have yielded promising results, there has been no consideration of how to absolve crime from society structurally.²³

To date, few legal scholars have considered the role that forgiveness, or absolution, should play in criminal law. A few notable scholars have recognized that forgiveness can play a role between victims of crime and perpetrators in civil society.²⁴ The leading legal expert on forgiveness, Martha Minow, has encouraged expanded forgiveness in the law generally, which could strengthen faith in law and offer “wrongdoers a fresh start . . . [by] wiping the slate clean.”²⁵ Minow documents forgiveness as an important tool for international crimes, in bankruptcy practice, and with amnesties and criminal pardons.²⁶ Judge Stephanos Bibas has also encouraged a role for individual forgiveness but states that “forgiveness and mercy do not square easily with a state-centered system of criminal justice.”²⁷ Forgiveness in the law has taken various forms, such as altered or reduced consequences for wrongdoing, the use of discretion by legal officials,²⁸ judges choosing leniency,²⁹ and voluntary expressions of apology and forgiveness.³⁰ Individually, restorative justice interventions like financial restitution, community service or family group conferencing,³¹ or executive pardons,³² have all played a role in individual cases to prioritize restoration over punishment. Despite some small notable successes in individual forgiveness between victims and perpetrators,³³ these examples remain few and far between. And there is no one who has provided theoretical or historical backing for a doctrine of systemic absolution—a path for society to periodically eliminate the burden of crime. Though elements of forgiveness exist in criminal law, they operate individually, focusing on the individual offender and their crimes rather than broader systemic absolution that likely involves structural change.

At the same time, the U.S. criminal justice system has a reputation for being particularly punitive, as this nation is the most carceral nation in the history of humanity.³⁴ The American system of criminal justice is infamously harsher than any other in the world.³⁵ Mass incarceration is a societal menace that many have tried to address, and without significant change, it is inevitable for the foreseeable future.³⁶ Punishment theory is vigorous in U.S. statutory and legal practice, but there has not been an exploration of a societal absolution of crime as a tool in dismantling our carceral system. It would not be foreign to borrow principles from religion to learn how to better handle crime, as crime and sin have been intertwined historically; for example, adultery and sodomy were crimes until very recently.³⁷ Our communities also carry a heavy social, familial, and racial burden as a result of mass incarceration, so the time is ripe for deeper structural inquiry into the practice of systemic absolution.³⁸

This Article is the first to consider whether systemic absolution can be considered as the foil for the dominant punishment theory of retributivism, allowing punishment to be followed by a path to societal atonement through structural change. It considers whether historical communal forgiveness practices might guide modern criminal practice to expand structural forgiveness for society at large. As considered here, systemic absolution is forgiveness of crimes on a broader scale focused on societal restoration after crime, apart from the individual forgiveness that might involve singular victims and offenders. It is focused more on fostering societal renewal rather than renewal of any individual. After considering the existing forms of individual forgiveness in the criminal justice system, this Article explores the role of systemic absolution and whether any modern mechanisms could reenvision criminal justice to allow for large-scale restoration after crime. This Article considers three avenues: automatic periodic expungement, periodic sentencing reduction, and deferred adjudication. There is historical precedent for systemic forgiveness practices, which have been applied to create bankruptcy law to allow people who go into debt to avoid prisons and be reintegrated into the business community.³⁹ With systemic absolution, it might be possible to reduce mass incarceration through periodic automatic expungement of a criminal record, similar to a release of debts in bankruptcy. Recently, federal legislation has reduced sentences broadly for certain drug offenders or incarcerated individuals above the age of sixty,⁴⁰ both structural efforts to reduce incarceration, allowing society to absolve defendants of crime broadly. And finally, deferred adjudication is happening across the country where prosecutors choose to “defer” prosecution in certain cases; if the defendant complies with certain conditions, after a certain period the state will drop the case altogether. Expanding this practice and making it a default for many crimes would allow many to change and abandon crime without maintaining a criminal record. All these efforts are examples of what I would call “systemic absolution.”

Because this is the first Article exploring systemic absolution, much of this Article is exploratory, conceptualizing the virtue of forgiveness in the criminal arena, and laying the theoretical backing and historical precedent for applying systemic absolution more broadly. As such, this Article does not provide any one prescription for how to execute absolution in modern society, but it does review the places where forgiveness plays a role in criminal justice and considers structural expansion. First, with forgiveness as the bedrock principle underlying systemic absolution, Part I of this Article provides a general consideration of the impact of forgiveness on individual outcomes. It also outlines the arguments in favor and against the adoption of broader systemic forgiveness in criminal law. Part II of this Article sets forth the history of communal absolution in ancient religion and society and its potential application to cleanse the societal slate of former crimes. Part III contextualizes a new theory of systemic absolution: the broadly accepted (though rarely adopted) forgiveness practices already existing with constitutional and statutory backing, including clemency, pardons, and expungements. The nature of this Article is exploratory and, while it does not provide a prescription for adopting systemic absolution, Part IV of this Article briefly considers how the justice system could consider systemic absolution in three areas: statutory sentencing relief, automatic expungement, and deferred adjudication, which might alleviate America’s endemic incarceration problem.

I. Conceptualizing Forgiveness in Criminal Justice

Forgiveness is best conceptualized as a conscious decision to give up a legitimate grievance against one who has caused harm.⁴¹ It is less about foregoing punishment, and more about eliminating “resentment and blame.”⁴² Forgiveness has traditionally been conceptualized as an individual matter between a victim and offender, considering the victim as the only one who can forgive. Many have analyzed the relationship between criminal offenders, their victims, and their affected communities.⁴³ But crimes can also be conceptualized as an offense against society itself—which is why the state takes on the prosecution. Under this framework, there is also an opportunity (and responsibility) for society to forgive criminal offenders through structural change. An absolution framework can provide broader structural solutions to alleviate the incarceration crisis along with communal burdens caused by crime.

Forgiveness at the victim level has real benefits both for the offenders and the victims themselves. Admitting guilt and moving on may lighten the “burden of guilt” for offenders, helping them in the next phase of their lives.⁴⁴ And victims are able to choose whether they let go of anger, grief, or sorrow. This can help them heal. And providing victims the choice to forgive or to not can empower them and help them regain a sense of agency after being harmed.⁴⁵ Policies that help the offender make amends and right the wrong and give space for the victims to choose to forgive promote this principle.

While individual forgiveness is critical, a consideration that has been missing in public dialogue is how to incorporate absolution into the structure of criminal justice. The American criminal justice system, known to be overly punitive, focuses on harshly punishing a few offenders who are caught. Policies that add to a retributive paradigm—that allow a path towards absolution—help a society relinquish a legitimate grievance and choose leniency when appropriate or allow redemption after punishment. Before shifting to collective forgiveness or absolution, Part I.A. considers more carefully the definition of forgiveness as well as the impacts of individual forgiveness, including the health and physical benefits. Part I.B considers arguments in favor of and against using the principle of forgiveness in criminal justice policy.

A. Individual Forgiveness in Criminal Justice

Providing a working definition of individual forgiveness, as opposed to societal or systemic forgiveness, lays a groundwork for later exploration of systemic absolution. Forgiveness involves “letting go of negative affect and motivations toward revenge or retaliation, despite an entitlement to such feelings.”⁴⁶ The repentance of the offender often encourages forgiveness, but is not a requirement for the victim to forgive.⁴⁷ For the victim, forgiveness requires forswearing revenge, moderating resentment, committing to letting go of lingering resentment, revisioning the wrongdoer, reframing one’s view of self, and communicating to the offender that forgiveness is granted.⁴⁸ The act of forgiveness does not necessarily make the crime go away in the victim’s eyes, but can allow them to see the perpetrator as a human and allow the victim and offender the possibility of creating a relationship.⁴⁹ While the law primarily seeks to adjudicate the past, forgiveness is forward-looking, and enlarges future possibilities.⁵⁰ Forgiveness should not be confused with acts that merely reduce animus, such as letting go, forbearance, or turning the issue over to God.⁵¹ Moreover, forgiveness is distinguishable from other responses to wrongdoing such as justification, excuse, mercy, and reconciliation.⁵² Forgiveness does not justify the conduct, does not excuse the wrongdoer as somehow not fully responsible, is more personal than mercy, and stands wholly apart from reconciliation. The distinction between systemic forgiveness and reconciliation or mercy is important. For systemic absolution of crime, there is no requirement that individual offenders are remorseful, that victim and perpetrator have reconciled, and it does not remove the role of punishment. Absolution, or forgiveness, plays a unique role in cleaning the slate for society of the burden of crime.

Studies demonstrate that individual forgiveness provides a multitude of psychological, physiological, and social benefits: reduced negative emotions, improved mental health, lower stress responses, less physical pain, and better relationships.⁵³ Evidence links forgiveness to overall better physical health.⁵⁴ Researchers believe that the physiological stress response of interpersonal conflict strains our health systems, but forgiving an offender breaks the cycle of negative thoughts and stress, thereby improving various indicators of health.⁵⁵ Perhaps most importantly, forgiveness is associated with greater happiness.⁵⁶ While there are clear health benefits to forgiveness, the adverse health consequences of unforgiveness may be more substantial.⁵⁷ But the individual harms of unforgiveness can be experienced by society, particularly within communities affected most by crime and incarceration. Moreover, while societal happiness is difficult to measure, systemic absolution might have some broader healing benefits.

B. Should Forgiveness Have a Role in Criminal Justice?

Despite the benefits of forgiveness, some scholars suggest that forgiveness may have no place in criminal law. Some believe forgiveness may be inappropriate in cases of remorseless offenders, victimless crimes, victims unwilling to forgive, or where there is disagreement among victims in multi-victim crimes.⁵⁸ Because forgiveness cannot be forced, forgiving individual offenders may also introduce dangers of inequality and discrimination, because victims and society may be less willing to forgive people of certain races or of certain socioeconomic status.⁵⁹ Moreover, some may argue that forgiveness does not serve the traditional justifications for punishment—undercutting retribution and deterrence.⁶⁰ Forgiveness of an offense can increase the likelihood of abuse and repetition of transgressions, for instance in domestic violence situations.⁶¹ Some feel that it is safer to err on the side of tough and inflexible punishment because the benefits from incarceration are immediate, certain, and concrete, while the benefits of forgiveness are “longer-term, squishier, and more speculative.”⁶²

On the other hand, scholars like Martha Minow have suggested that the U.S. should consider a more liberal use of “forgiveness mechanisms” in the criminal justice system.⁶³ Minow makes clear that she is not advocating for an elimination of punishment. In fact, the idea that wrongdoing deserves and requires punishment seems to be a universal and fundamental belief, accepted across history and human civilizations, and even evident in very young children and animals.⁶⁴ All attempts at establishing no-punishment communes have failed, or have only survived by adopting enforcement systems.⁶⁵ Communities absent of law, such as gold rush camps or concentration camps, typically adopted their own vigilante systems or practiced prisoner justice, exemplifying our natural inclination towards punishment.⁶⁶ Minow acknowledges that some wrongs may seem unforgivable, and to advocate for forgiveness for genocides and murders “can seem an insult to both victims and survivors.”⁶⁷ Indeed, one study found that offenders who commit high-severity crimes are perceived as unlikely to be rehabilitated.⁶⁸

Despite these findings, Minow argues that making greater use of forgiveness mechanisms is an important part of criminal justice and can help alleviate our country’s mass incarceration problem.⁶⁹ Minow argues that lawyers and other officials “do not adequately use the tools of forgiveness” in criminal law and argues for their expansion.⁷⁰ Forgiveness in criminal law could shift away from the narrow focus on a specific violator and victim to a wider lens that looks at broader patterns to increase fairness for all.⁷¹ She advocates for potentially expanding existing practices like apology, restitution, and forbearance from the law’s most stringent demands.⁷² It might be time to reinvent criminal law as many criminal law systems have lost credibility with their communities due to their reputation for harshness⁷³ and their uneven application of justice.⁷⁴ Introducing mechanisms of forgiveness more broadly into criminal law could help combat unfairly harsh laws or overly punitive prosecutors.⁷⁵ Forgiveness encourages the full reintegration of a convicted person into society, alleviating the collateral consequences and stigma of conviction.⁷⁶ Moreover, forgiveness may have broader benefits that allow greater healing not just for the offender or victim, but for the community in general.⁷⁷ Systemic absolution could create similar benefits as individual forgiveness if it were accepted into criminal practice. The dominant theory of punishment is retributivism, which includes little in the way of absolution of crime, rehabilitation, or a structural path towards clearing the slate of crime. The next Section provides background on the historical importance of individual and communal absolution to contextualize forgiveness as an important societal practice, which aids in creating a comprehensive criminal punishment theory that includes absolution.

II. Individual Forgiveness and Communal Absolution in Religious Tradition

Religious principles have always influenced criminal punishment theory.⁷⁸ Covenant principles⁷⁹ form the roots of the leading criminal punishment theory (retributivism) and dominate criminal philosophy, criminal statutory principles, and even prison administration.⁸⁰ In simple terms, retributivism requires that a person be punished proportionately given the exact nature or harm of their crimes—a figurative eye for an eye, at least in value.⁸¹ It is not until after an offender has been punished appropriately for their crime that they should be released.⁸² Criminal statutory practice and punishment are dominated by desert and punishment without an avenue for societal forgiveness or growth. The absolution theory proposed here has its roots in religious traditions that create a more unified community through cleansing society of wrongs.⁸³ Restorative justice has its place and can be extremely effective when both parties are inclined towards a resolution, but this often is not the case. In many cases, one side would like to make amends, but the other side is unwilling. But since crimes are committed against society, could there be systemic absolution to restore society after harm is caused?

Forgiveness traces its roots back to influential world religions. Forgiveness considers restoring relationships after sin or crime and is a fundamental part of all major religions. These individual traditions of forgiveness provide a foundation for considering communal forgiveness or absolution as discussed in Part II.A. Then, Part II.B considers absolution in religious contexts and considers its application to criminal justice.

A. Individual Forgiveness in World Religions

There are foundational teachings about forgiveness in all the world religions, but this Section provides a very brief overview of these teachings in Christianity, Islam, Judaism, and Hinduism as they pertain to forgiving crime. This Section focuses on individual forgiveness, or receiving forgiveness from God for wrongs committed.

The concept of forgiveness might be most central to Christianity.⁸⁴ Instances of forgiveness are plentiful in the New Testament, from the forgiveness of the adulterous woman, the prodigal son parable, the Sermon on the Mount, the Lord’s Prayer, and Jesus’s forgiveness of those who crucified Him.⁸⁵ As it relates to forgiving criminal offenses, Jesus Christ specifically says to the criminal defendant hanging next to him on the cross, “[t]o day shalt thou be with me in paradise,” insinuating that forgiveness is possible even for criminals and potentially even beyond the grave.⁸⁶ Christian rituals, such as baptism and Holy Communion, are fundamentally rooted in individual forgiveness.⁸⁷ Christianity often urges that forgiveness must be offered unconditionally, as an act of grace, and without regard to the punishment or remorse of the offender, which may seem irreconcilable with conceptions of retributivism.⁸⁸ Christian forgiveness is predicated on the unconditionality of God’s love, and this kind of unconditional forgiveness serves as an ideal in the life of the believer.⁸⁹ Although Jesus preached that people should repent, Christian forgiveness requires forgiving even the unrepentant.⁹⁰ While some argue that the New Testament presents forgiveness as a desirable virtue but not a mandatory duty,⁹¹ forgiveness, including of crimes, is a key part of an individual’s journey to becoming more like Jesus Christ.

Islam views forgiveness as an important virtue practiced by the Prophet Mohammed that Muslims should emulate.⁹² The Prophet Mohammad teaches by example to seek God’s forgiveness three times after completing his prayer,⁹³ and he specifically asks for forgiveness for his enemies.⁹⁴ However, Islam does not advocate for unconditional or absolute forgiveness, and does not require forgiveness before justice or punitive measures are implemented.⁹⁵ The Quran explains that divine forgiveness is possible even for serious offenses, but an offender should never take forgiveness for granted even if he expresses repentance because it is a decision for Allah.⁹⁶ However, Islamic texts emphasize that interpersonal forgiveness is more virtuous than vengeance, even if retaliation is permissible under the circumstances.⁹⁷ A study of the motivations to forgive among Moroccan Muslims found that their religious values were their primary motivation to forgive others.⁹⁸ In Islam, tawba is the repentance process,⁹⁹ in which an individual feels guilt about past sins, engages in righteous acts that the sin prevented them from doing, recompenses those that they wronged, asks forgiveness of those that they wronged, resolves to avoid the sin in future, and continues in obedience to God.¹⁰⁰ The tawba is reminiscent of restorative justice principles currently in place in criminal policy, and the focus on the Prophet’s forgiveness of enemies is reminiscent of the Christian dictates to forgive all. A foundational virtue in Islam is forgiveness—though not mandated, it is highly admired if a person can forgive someone for a crime committed against them.¹⁰¹

Judaism places emphasis on individual forgiveness, repentance, and atonement, and is particularly influential when it comes to retributivism, which continues to dominate criminal punishment theory.¹⁰² Every year, ten days are set aside between Rosh Hashanah and Yom Kippur, or Day of Atonement, to repent.¹⁰³ This period is called teshuva (meaning return).¹⁰⁴ During this period, Jews ask forgiveness of others and engage in acts of service.¹⁰⁵ On the Day of Atonement, Jews ask God for forgiveness in prayer, sometimes in synagogue.¹⁰⁶ The Mishnah, or oral Torah, recognizes that an injury to another person results in two harms: a material harm, which is rectified through monetary compensation; and a more intangible injury, which can be remedied only through seeking forgiveness.¹⁰⁷ In Judaism, teshuva is the process of making amends; it includes a full return to the right path, and restoration of good standing with the community and with God.¹⁰⁸ The concept of return means that the sinner abandons his sin, removes it from his thoughts, resolves to not do it anymore, expresses remorse for the transgression, and makes an oral confession and apology.¹⁰⁹ After the offender undergoes the process of return, the victim is required to forgive, even when the wrong can never be fully righted.¹¹⁰ A victim may choose to forgive even without the offender’s repentance but is not obligated to do so in Jewish tradition.¹¹¹ Judaism does not definitively resolve the issue of whether certain individuals do not deserve forgiveness, but it does emphasize the belief that people can change for the better and must be given the opportunity to do so.¹¹² The Jewish faith also includes rituals of communal absolution discussed in the next Section which will allow us to consider these principles in how they might create a more comprehensive criminal justice policy that includes not just retribution but also atonement.

Other religious traditions, Hinduism and Buddhism, include foundational teachings in forgiveness. Buddhism handles forgiveness through ritual confession.¹¹³ Because the end goal of Buddhists is the cultivation of inner peace and enlightenment,¹¹⁴ the purpose of confession and forgiveness is to develop personal virtues such as responsibility and patience.¹¹⁵ Buddhism also upholds a model of forgiveness in its call to relinquish bitterness and resentments.¹¹⁶ One of the six perfections of Buddhism, forbearance or patience, corresponds closely to forgiveness and represents “a virtuous response to harm brought upon oneself by the conduct of another.”¹¹⁷ In the Dhammapada, the Buddha warns against hostility and advises that “[h]atred is never appeased by hatred in this world. By non-hatred alone is hatred appeased.”¹¹⁸ Similarly, Hinduism’s sacred texts describe multiple incidents of divine forgiveness.¹¹⁹ The Mahabharata, one of two Sanskrit epics revered in Hinduism, even features an important “hymn of forgiveness.”¹²⁰ The scripture further recommends forgiveness (ksama in Sanskrit) as one of several “divine qualities” to which a spiritually-minded person should aspire.¹²¹ Both Buddhism and Hinduism focus on the importance of forgiveness as a divine virtue to cultivate, even for those who harm you.

Christianity places the strongest emphasis on forgiveness, teaching unconditional forgiveness even in the absence of remorse and reparation.¹²² Judaism commends forgiveness, but focuses on the conduct of the wrongdoer who must make amends to earn forgiveness.¹²³ Islam commends unconditional forgiveness as virtuous, but does not require it.¹²⁴ Buddhism and Hinduism both elevate forgiveness as a divine trait. While principles of forgiveness are important in many key world religions, the concepts of individual absolution through a neutral person, like a priest, or community absolution are found in Christianity as described below.

B. Christian Individual Absolution

To better theorize criminal punishment and absolution, a greater understanding of the religious doctrines of absolution is instructive. The word “absolution” comes from the Latin absolutus, which means to “set free,”¹²⁵ and the concept of absolution is fundamental to various Christian faiths. Absolution in the religious context involves a neutral party offering forgiveness, which is not focused between a victim and offender, or a victim and God like in many other faiths. In Catholicism, priests grant absolution, or forgiveness, to the penitent as outside observers providing forgiveness for a wrong.¹²⁶ This concept is called the “sacrament of Penance,” one of the seven sacraments of the Catholic faith.¹²⁷ Catholics believe that, upon receiving this sacrament, they are free from mortal guilt and eternal punishment.¹²⁸ In this process, church members confess their sins to the priest, and the priest then absolves the person from their sins “in the name of the Father, and of the Son, and of the Holy Spirit.”¹²⁹ In an absolution rite, a priest is a proxy for God allowing a person to be forgiven for a wrong.¹³⁰ The first step in a system of absolution involves allowing a neutral party to provide absolution—a person unaffected by the crime who assists in the atonement process.

Absolution, in the religious context, allows an individual to have wrongs removed from their life so that they can move forward, reconciled with God and their community. Part II.C defines and discusses the concept of communal absolution of sin or crime in world religions.

C. Communal Absolution in World Religions

Many world religions rely on forgiveness and mercy towards transgressors as a foundational value.¹³¹ Communal absolution describes when a group of people are unburdened of crime or sin and has its roots in the world’s largest religions.¹³² This Section defines a new concept, “communal absolution,” which considers cleansing and reunification with society after a defendant is held accountable for a crime. Communal absolution—as defined here—considers crime as an act that afflicts a community and considers a societal forgiveness and letting go of the crimes that have afflicted some of its members. This is a precursor to “systemic absolution,” discussed in Parts III and IV, which is applicable to a society that has grown more complex and could serve as the modern corollary to communal absolution.¹³³ This Section considers communal absolution practices important to understanding the religio-cultural backgrounds existing for crime.

While this emphasis on individual repentance and forgiveness is highly influential in many faiths, the communal cleansing and atonement rituals that have existed in many faiths historically provide parallels for communal absolution. For instance, in the Catholic and Lutheran faiths, priests are permitted to perform “general absolution” to groups of people in emergency circumstances in which individual, private confession would be impossible.¹³⁴ These, along with ancient cleansing rituals, provide the groundwork for considering communal absolution of crime today. Understanding communal absolution opens the door for creating a comprehensive criminal theory that considers reunification, or cleansing after crime.

Communal Cleansing Rituals

Various ancient Near Eastern societies participated in community cleansing rituals in which freeing animals into the wilderness marked the absolution of crimes or purification of a community, as precursors for the Jewish scapegoat ritual.¹³⁵ The exact origins of these rituals are contested among Ancient Near Eastern scholars, but the earliest of this type of rite was likely performed in Ebla (now Syria) in the 24th century BCE.¹³⁶ The Eblaite ritual took place before a wedding, when the house was purified by putting the imperfections of the home on a goat and sending it away: “We purify the mausoleum before the entrance of (the Gods) Kura and Barama. A goat, a silver bracelet (hanging from) its neck, towards the steppe of Alini we let it go.”¹³⁷ Similarly, the Hittite people performed a ritual that involved sending away a ram to purify a community.¹³⁸ Like the goat that was sent away in the Eblaite ritual, here a ram was “sent away to the land of the enemies and offered to the hostile deity who caused the pestilence.”¹³⁹ There are several other Ancient Near Eastern societies who also performed similar rituals with animals being sent away to cleanse a community.¹⁴⁰

Likely drawing from earlier societies, the ancient Israelites started performing a ritual that is described in the Bible and has come to be more commonly known as the “scapegoat” ritual. Performed in preparation for Yom Kippur, this ritual is explained in the book of Leviticus, where Aaron cast lots over two goats: one that will be sacrificed, and one that carries the sins of the community into the wilderness.¹⁴¹ The name Azazel is given to the goat that is sent away.¹⁴² To transfer the sins, Aaron laid his hands on the goat’s head and “transfer[red] the transgressions of the Israelites to it,” so the Israelites could “openly and honestly hav[e] the goat remove their[] [sins].”¹⁴³ The Azazel goat was then led to a desolate part of the wilderness, which marked the absolution of “crimes and sins.”¹⁴⁴ The goat served as a “medium of atonement . . . and forgiveness” when performed by the Israelites.¹⁴⁵ While important for many years, the scapegoat rituals lost significance in Jewish practice in 70 CE.¹⁴⁶ However, these absolution concepts transformed into individual forgiveness through prayer and personal repentance associated with Yom Kippur.¹⁴⁷

Other societies also adopted and modified forms of the Hittite, Eblaite, and Jewish rituals, but sent away humans instead. For example, the Ancient Greeks adopted a similar form of the Hittite ritual using human sacrifices called pharmakos rather than animals.¹⁴⁸ For the Greeks, the “scape-person” represented the “evil that is being expelled.”¹⁴⁹ However, the Greek pharmakos ritual was performed for the purpose of stopping plagues, not absolving sins.¹⁵⁰ The scapegoat rituals also influenced the ancient Romans, who had a ritual in which army “commander[s] could offer a common soldier to die for all.”¹⁵¹ Like the Greek pharmakos ritual, the Roman ritual sacrificed humans instead of animals. While for different purposes, the Jewish ritual demonstrates the far-reaching, long-lasting effect of these rites in which communities can be cleansed by letting something go.¹⁵² Communal absolution practices found in Jewish, Greek, and other ancient societies provide important examples for how a society can cleanse itself from crime. The next Section discusses the change from ritual absolution to a year of absolution in which all wrongs, crimes, and debts were eliminated every fifty years.

Periodic Absolution Practices

Another influential practice, jubilee, is interesting to consider for modern criminal purposes. Serving as another communal absolution ritual, it underscores the importance of community cleansing of crime and sin in a periodic manner.

The Jubilee year represents how the forgiveness and cleansing of past wrongs permeate community ideals. The Jubilee year, which is outlined in Leviticus 25, demonstrates the role for which forgiveness began with Israelites and has since been adopted by Christians. Leviticus 25 instructed the Israelites to observe practices that marked a time when injustice was corrected and justice reestablished.¹⁵³ Jubilee consisted of slave release, debt release, restoration of land to original owners, and other restoring practices; it took place every fifty years.¹⁵⁴ In releasing slaves and debtors and restoring land, the Jubilee “sought to be an institution of sweeping social reform” for the Israelites.¹⁵⁵ It is unclear whether the Ancient Israelites actually ever put the ideas of the Jubilee year into practice, but the message of forgiveness the Jubilee year carries nonetheless impacted their society,¹⁵⁶ and has spread in Christianity as a practice still common among Christians.¹⁵⁷ The Jubilee year serves as a model in considering periodic release of injustices and public wrongs to create a more just society. As a criminal corollary, considering a release of criminal records or adjustment of criminal sentences provides a modern application to these principles that have had import in various ancient contexts.

The idea of communal absolution played a fundamental role in ancient Israel and provided a bedrock foundation for Greek and Roman civilizations. These ancient practices also influenced early American history, providing further support for considering systemic absolution in modern criminal justice.

Communal Absolution in American History

Communal absolution, or atonement rituals like the scapegoat and pharmakos ritual, have corollaries in American history. Group absolution rituals manifest in early colonial culture in several ways. Scholars have stipulated that prosecution of colonial “witches” in the eighteenth century “function[ed] . . . as a scapegoat ritual that served to rid the community of undesirable elements.”¹⁵⁸ Like the goat that was exiled from the Israelite community to purify and remove sin, here witches were also removed from early American colonial communities with the hope that the community would be purified. The scapegoat ritual has also served as an archetype in early American literature, with many prominent authors, including Mark Twain and William Faulkner “indicat[ing] varying attitudes toward and understanding of sacrificing a victim to banish guilt.”¹⁵⁹ Even the Thanksgiving turkey has been argued to serve as America’s symbol of a scapegoat.¹⁶⁰ Dr. Karen Davis, president of United Poultry Concerns, explains how when we kill and eat turkeys on Thanksgiving, they “function[] as . . . bearer[s] of impious sentiments,” just like the scapegoat bore the Israelites’ impurities and sins.¹⁶¹ Davis and others have connected the yearly presidential pardoning of a Thanksgiving turkey to the Israelite tradition, calling the pardon an “inverted scapegoat ritual,”¹⁶² in which “[t]he pardon therefore performs the same basic function as the scapegoating ritual . . . although instead of one special victim being scapegoated, every animal except for one special non-victim is scapegoated.”¹⁶³ Communal absolution, like the scapegoat ritual, provides the framework for modern criminal justice to embrace forgiveness of crimes.

Criminal absolution considers a communal forgiveness for offenders who collectively, as society determines, are set free from their past crimes and are permitted to rejoin society with similar rights as other citizens (after they have been properly held accountable). Communal absolution, or group cleansing of past wrongs, has corollaries in U.S. and world history and has the potential to transform our criminal justice system. Before introducing the concept of systemic absolution in criminal practice, the next Section considers the underutilized forms of absolution already existing in a punitive U.S. criminal justice system, though not recognized as such.

III. Forms of Absolution Existing in Criminal Justice

The American criminal justice system is overly punitive with little opportunity to remove the permanent stain of crime.¹⁶⁴ The United States incarceration rate is about “five times higher than the rest of the world.”¹⁶⁵ The roughly 2.3 million people incarcerated in American jails and prisons cost the state about $80 billion a year.¹⁶⁶ Scholars have increasingly argued that the “main driver of [this] mass incarceration” is not the imprisonment of drug and other nonviolent offenders, but the “over-sentencing of people for serious and violent crimes.”¹⁶⁷ Indeed, much of the “prison boom” of the latter 20th century can be traced back to increasingly punitive sentencing.¹⁶⁸ There is not one culprit or crime, but instead a “broad shift toward more punitive treatment for offenders in every major crime category,” both state and federal.¹⁶⁹ And while Americans’ societal attitudes toward crime are more punitive than some Western democracies, the country still has significantly longer sentencing than countries that are culturally similar.¹⁷⁰

There are existing forms of forgiveness in criminal justice, though they have not been articulated as such or relied upon extensively. In creating a more comprehensive punishment theory, there should be consideration of absolution after retribution, restorative justice, or incapacitation. This Section lays the groundwork for what already exists. The most common forms of absolution in criminal justice include sentencing and charging reduction efforts, expungement, and clemency. Part III.A discusses sentencing and charging reduction efforts including specific charging measures, deferred prosecution in Part III.A.1 and III.A.2, and safeguarding judicial discretion. Part III.B.1 and III.B.2 discuss automatic expungement through clean slate legislation. Finally, Part III.C discusses clemency and pardons. While these latter types of absolution have not been implemented broadly or institutionally, they demonstrate constitutional and statutory precedent for forgiveness in criminal justice and provide a place for broader implementation of systemic absolution in our system. Each type of absolution is discussed below as it currently exists with some considerations on how it can be expanded to systemic absolution.

A. Sentencing and Charging Revision

Sentencing and charging revision provide the largest-scale impact for systemic absolution. Regularly reassessing punitive measures enables society to move past transgressions, fostering unity and healing. In reevaluating sentencing and charging practices, it is important to recognize that punitive sentencing is a recent development in federal and state legislation over the past forty years.¹⁷¹

The criminal justice system currently uses sentencing and charging revisions to offer some measure of absolution to convicted individuals, though they are not articulated as such. Three programs—second look resentencing, deferred adjudication, and restoring judicial discretion—are discussed in greater detail below. Second look laws resentence offenders based on age, type of crime, or behavior in prison. Deferred adjudication is a common practice in some jurisdictions that allows prosecutors to delay or suspend a criminal case—usually for first-time or low-risk offenders—with the understanding that the charges will be dismissed if the individual avoids reoffending. In essence, it serves as a form of conditional forgiveness for those whose prosecution would not meaningfully serve justice. When used systematically, deferred adjudication enables prosecutors to postpone cases where formal charges are unnecessary to achieve a fair or effective outcome. Finally, restoring judicial discretion by eliminating mandatory minimum sentencing laws or determinate sentencing could reduce automated punitiveness and allow individuals to receive individualized sentences and allow for rehabilitation or change in prison. These three approaches provide systemic shifts towards absolution in criminal justice and are discussed in order below.

Second Look Resentencing

Expanding second look resentencing is one approach to expanding systemic absolution. Current criminal practice limits a second look at sentencing to cases with extraordinary and compelling reasons, but a true categorical second look provision would allow automatic periodic review after a certain period—such as after five years of imprisonment—where all prisoners would receive an opportunity to evaluate their sentences. Expanding second look resentencing requires an indeterminate sentencing system, discussed in Part III.A.3, in which sentences could be reduced based on rehabilitation and progress made by the inmate behind bars. Shon Hopwood urges that a second look provision should include all offenders, including those convicted of violent crimes and sex offenses because the best incentive to participate in recidivism-reducing programs¹⁷² such as behavioral therapy, is the incentive of early release, in addition to the fact that such programs have been very successful with low recidivism rates.

In many cases, it is difficult to identify which defendants have the capacity for rehabilitation and redemption at the sentencing stage, so the justice system needs an avenue for decisionmakers to take a second look at individual cases.¹⁷³ Second look laws provide a way to reduce the sentences of those who received excessively punitive punishments or who have demonstrated rehabilitation.¹⁷⁴ Moreover, second look resentencing can lower the chances of recidivism, ameliorate race-based sentencing disparities, and alleviate harm to family members.¹⁷⁵ It can also provide a systemic way to reduce prison populations and the financial costs of incarceration.¹⁷⁶ For the offender, resentencing also provides a sense of redemption and a recognition of their efforts towards rehabilitation.¹⁷⁷ Second look resentencing recognizes, values, and encourages absolution and

transformation, and has vast support from faith communities—who can be integral in motivating change from inside prisons.¹⁷⁸

Second look resentencing can provide systemic absolution, without harm to public safety, when conducted with consideration of empirical data on crime and punishment. First, studies show that long sentences do not deter crime, but rather it is the certainty of punishment which most effectively deters crime.¹⁷⁹ Moreover, offenders tend to age out of crime, as research indicates that even among chronic offenders, most will cease committing crimes by their forties.¹⁸⁰ This remains true for those convicted of violent crimes, including homicide; in fact, violent offenders who are older at release tend to have lower recidivism rates relative to other offenders.¹⁸¹ For these reasons, those serving long prison sentences, older offenders, and even violent offenders should be considered for periodic resentencing.¹⁸² Excluding violent offenders from second look sentencing programs eliminates the majority of the prison population from access to these programs. Currently, nine states—California, Colorado, Connecticut, Illinois, Louisiana, Minnesota, New Mexico, Oregon, and Washington—and D.C. have enacted second look laws.¹⁸³ Second look laws authorize judges to revisit long sentences after substantial time served and reduce them when the person no longer poses a risk or when the punishment is out of step with current sentencing standards.¹⁸⁴ At the moment, thirty-seven states have pending legislation to reduce overly long sentences.¹⁸⁵ Many states have passed legislation that sentences between 40 and 120 years are unconstitutional and thus the trend is moving towards change in this area.¹⁸⁶ Broader adoption of state laws that prohibit sentences beyond forty years and adjusting state sentencing schemes towards indeterminate sentences that allow for adjustment after five or ten years provide incentives for individual rehabilitation and allow for periodic adjustment of sentences in systemic absolution. Federal legislation, like the First Step Act of 2018, has also worked to reduce federal prison populations through reducing sentences for those with good behavior in prison.¹⁸⁷ Also, analysis on the First Step Act, six years after its implementation, has shown that inmates released under the Act showed lower recidivism rates and served less time than those released before the Act was passed.¹⁸⁸ Other legislation has allowed federal judges to reduce drug sentences when appropriate and depart from federal guidelines.¹⁸⁹

Reconsidering the terms of sentences periodically after their initial imposition and eliminating indeterminate sentencing practices are both ways to allow for systemic absolution of crime. While recognizing the importance

of criminal accountability and punishment, these changes recognize that people can change and that cultural shifts might allow many sentences to be reduced to remove the societal burden of incarceration.

Deferred Adjudication

Deferred adjudication provides another avenue whereby defendants may have their sentences diverted and thus receive absolution from the criminal justice system.¹⁹⁰ Deferred prosecution agreements (“DPAs”) enable a defendant to avoid prosecution, maintain a clean record, and move towards rehabilitation.¹⁹¹ When making a DPA, a prosecutor delays criminal charges and instead sets conditions for the defendant to follow under the government’s supervision.¹⁹² Defendants are not prosecuted unless they fail to comply with the conditions set forth. Similar to DPAs, a deferred sentence offers a defendant probation instead of conviction if they complete certain steps toward rehabilitation.¹⁹³ However, the defendant must first plead guilty, and that plea will remain on their record.¹⁹⁴ These alternatives to prosecution shift the focus from punishment to rehabilitation and spare defendants the often-debilitating collateral consequences of conviction.¹⁹⁵ Indeed, many individuals who have been convicted experience unemployment, exclusion from society, and a tarnished reputation and communities suffer with families separated.¹⁹⁶

Not only does broadening deferred adjudication help create systemic absolution, but it may also benefit the public by lowering recidivism rates. Some studies have shown that offenders receiving suspended sentences or probation are less likely, or at least not more likely, to reoffend than those who are sentenced to prison.¹⁹⁷ Additionally, DPAs and deferred sentences are far more cost-effective than confinement.¹⁹⁸ All fifty states have codified DPAs in some capacity, although most states place limits on the type of defendant and offense to which they can be applied.¹⁹⁹ Despite the many benefits of DPAs, research demonstrates that they are still used in less than a quarter of cases and could be a useful way to increase systemic absolution.²⁰⁰ Currently, prosecutors must justify declining to prosecute a case rather than bringing charges. This assumption could be reversed,²⁰¹ and deferred prosecution could be employed as the default for crimes that do not pose a threat to public safety, or where other options exist for holding criminal defendants accountable, such as mental health or drug rehabilitation. With broader use, deferred adjudication could serve a crucial role in curbing mass incarceration and offering absolution with a second chance.

Safeguarding Judicial Discretion

In the last fifty years, sentencing changes were made intending to make punishments both more severe and less determinate, and less subject to judicial discretion.²⁰² This provided judges and parole boards with less control and fewer opportunities for rehabilitation to reduce sentences.²⁰³ A major consideration in creating room for absolution is a shift toward indeterminate sentencing, allowing judges and public executives the opportunity to adjust sentences where appropriate. Another approach could be eliminating mandatory minimum sentences.

Some criminal actions have mandatory minimum sentences that judges must impose on convicted individuals. For example, Aguilera-Mederos was a truck driver who lost control of his truck on a Colorado highway, resulting in the death of four people.²⁰⁴ While some of his actions may have been reckless, the crash was unintentional and Aguilera-Mederos was repentant, begged for forgiveness, and complied fully with police.²⁰⁵ Because of the mandatory minimums required by law, the judge reluctantly handed down the required sentence of 110 years in prison.²⁰⁶ In Aguilera-Mederos’ case, the governor commuted his sentence to ten years, recognizing the injustice of the long sentence.²⁰⁷ However, convicted persons cannot rely on a rare pardon. The ABA has long opposed mandatory minimum sentencing and has urged Congress, as well as state and territorial legislatures, to repeal laws requiring minimum sentences.²⁰⁸ Although mandatory minimums were an attempt to address inequities, they now seem to be the source of punitiveness and an unnecessary enhancer of mass incarceration.²⁰⁹ Eliminating mandatory minimum sentences would improve fairness in sentencing as well as systemic forgiveness for many who are punished harshly in the criminal justice system. Federally, the Smarter Sentencing Act has been proposed several times, including as recently as 2023, to eliminate mandatory minimums for federal drug sentences.²¹⁰ Allowing judicial discretion and removing legislative hurdles and determinate sentencing would increase the potential for systemic forgiveness in the criminal justice system.

B. Automatic Expungement

Nearly one out of every three American adults has a criminal record, and nearly half of American children have a parent with a criminal record.²¹¹ The overwhelming majority of these individuals have never been imprisoned.²¹² Yet, they often suffer the consequences of carrying a criminal record for the entirety of their lives.²¹³ The traditional option for relief, expungement or sealing of records, is far from accessible.²¹⁴ Studies show that expungement both has positive real, measurable effects and is heavily underutilized.²¹⁵ In fact, the authors of one recent study suggest that only 6.5% of individuals receive expungements within five years of becoming eligible.²¹⁶ Many factors contribute to this low usage, chief among them perceived ineligibility, administrative hassle, monetary costs, and lack of information.²¹⁷ Thus, even individuals that legislatures have already deemed worthy of forgiveness often still do not obtain clean records.

Expungement allows a clearing of a criminal slate, and with analogy to bankruptcy can parallel ancient scapegoat rituals in clearing the societal criminal slate of debt and crime. A broader pact of expungement among states or a federal expungement scheme might be considered to expand systemic absolution. As it stands, once an individual commits a crime, they continue to be punished with collateral consequences, including social stigma of conviction, difficulty in finding employment, lost voting and jury rights, and inability to hold office.²¹⁸ Considering bankruptcy as a parallel provides some insight on how to expunge crimes from individuals as debts are discharged in a bankruptcy. A criminal conviction permanently labels and segregates an offender from society.²¹⁹ This creates a growing class of internal exiles who, despite serving their punishment, have no way to successfully reintegrate into society.²²⁰

Expungement is a process where the court destroys or seals an individual’s criminal conviction, removing it from a defendant’s criminal record, and ideally also from the public record.²²¹ This allows restoration of civil rights, limits the harm of inquiries by employers about criminal histories, and seeks to prohibit exclusion based on a prior conviction.²²² While expungement does not mean that the victims of the crime have forgiven the offender, it does signify a second chance and redemption from society. In other words, the community absolves the offender, allowing them to change and shed their crime and become someone new.²²³ At the same time, expungement does not eliminate responsibility for crimes as it is currently only available to people who have served their sentences and often to those who have gone beyond that by demonstrating rehabilitation.²²⁴

Traditionally, the legislature provided rules for expungement in state law, and under the applicable statute, individuals could petition the court for an expungement.²²⁵ But a unique characteristic of clean slate legislation is their automatic clearance of former crimes, which makes it a worthy form of systemic absolution. Several states have recently enacted clean slate legislation providing for automatic expungements, making the process much less burdensome for those with convictions.²²⁶ National momentum began with Pennsylvania first passing a clean slate law in 2018 and then twelve other states following their lead.²²⁷ These laws function by automatically erasing criminal records after a specified amount of time.²²⁸ This automated process has dramatically shifted the landscape of expungements in those states. For example, one year after Pennsylvania’s law went into effect, forty-seven million criminal charges were sealed, just over half the total charges in the criminal database.²²⁹ Between the twelve states, some fourteen million individuals are now eligible to have their records automatically cleared.²³⁰ At least five more states have pending clean slate bills,²³¹ and two Federal Clean Slate laws have also been proposed.²³² These new laws hopefully reflect a real shift in attitudes towards criminal records. For example, Utah’s Clean Slate Law seals eligible crimes so that they do not appear on records, and individuals will legally be able to respond to questions about criminal history as if the crime never happened.²³³ Utah’s law allows for some misdemeanor crimes to be expunged automatically if there is not another conviction for five to seven years.²³⁴ Indiana has an expansive law making expungement and sealing of records mandatory for records where no conviction has occurred.²³⁵ This includes serious felonies, and expungement is mandatory for misdemeanors if eligible.²³⁶ The law requires in all cases that an expunged conviction shall be treated “as if the person had never been convicted of the offense.”²³⁷ California also forgives old convictions for marijuana possessions, following the state’s legalization of recreational use of marijuana.²³⁸ At the end of 2023, New York became the twelfth state to pass a “clean slate” law.²³⁹

Many of these clean slate laws are a form of systemic absolution in which, based on certain qualifications, offenders can have their convictions automatically expunged without having to inquire into the specific facts of each case. Crimes automatically disappear from criminal records in the way ancient societies were healed ritually through a symbolic ritual. And although not directly reducing the number of people incarcerated, clean records have been shown to correlate with reduced recidivism rates,²⁴⁰ as well as arrest rates lower than the general population of the state and higher employment outcomes.²⁴¹ Clean slate laws also exemplify a forgiveness-centered legislative approach to criminal justice reform. They go beyond simply softening punishments, as the public through their elected representatives demonstrates a possible systemic approach to forgiveness, or a cleansing of the public slate.²⁴²

It is possible to expand automatic expungement to all states and consider this for the federal system. A historical corollary is bankruptcy law which allows those who are unable to repay their debts to have them wiped clean under federal law, as the Bankruptcy Act of 1800 facilitated the release of debt prisoners.²⁴³ Before this act, people who could not pay their debts were punished and put into prisons with poor conditions and punished as criminals. Bankruptcy law in the United States requires a person with debts to report all of their assets. Then, they are released from the burden of having to pay the debts they cannot afford to pay, and the debt is eventually discharged.²⁴⁴ After a legal bankruptcy, the individual who owed a debt and their company are free from any legal claims, though they may remain “a poor credit risk for future loans until they develop a record of making timely payments of bills.”²⁴⁵ Automatic removal of a bankruptcy occurs after seven or ten years, depending on the type of bankruptcy.²⁴⁶ The federal system could consider expungement of a criminal record after a similar period of time.

Expungement is an existing procedure that could be expanded so that more people who have been convicted of a crime and served their time could have their records expunged automatically and periodically. Federal expungement could mimic the bankruptcy system, with the goal of improving systemic absolution. Federal legislation called the Redeem Act has been proposed to tackle some of these issues.²⁴⁷ States who have not done so could legislate for an automatic expungement procedure for people who have been arrested and then remained crime-free for a certain number of years to have arrests removed from their record. Implementing an automatic expungement procedure where all people who have not been rearrested after a certain number of years could be expunged is one means by which to bring about societal absolution in our criminal justice system. Like the scapegoat ritual of old, this could be one way for society to understand that people can change and there should be periodic room to create a clean slate.

C. Clemency

While clemency is a broad term, American law primarily recognizes five forms of clemency: pardon, amnesty, commutation, remission of fines, and reprieve.²⁴⁸ Clemency traces its roots back to the Code of Hammurabi through the Greeks and Romans to the English legal system, and is enmeshed in principles of forgiveness.²⁴⁹ Because it is based on an executive considering an individual’s case and providing them forgiveness for their crimes, it does not have the potential applicability for systemic absolution. However, because clemency provides constitutional and historical authority for the importance of forgiveness in criminal justice, it is discussed below, though such approaches are not recommended for expanding into broader systemic absolution.

Pardon Power

Clemency can be used to improve justice by adjusting individual sentences to avoid undeserved punishment.²⁵⁰ In the United States, clemency power lies in the Executive branch in the Pardon Clause which vests the president with “[p]ower to grant Reprieves and Pardons for Offences against the United States, except in Cases of Impeachment.”²⁵¹ The president’s pardon power is expansively broad, and they may exercise their discretion to pardon for any reason, with no requirement to base their decision on objective criteria.²⁵² The president can use clemency to correct decisions of prior administrations with which they disagree and to make a statement about broader policy views.²⁵³ The majority of state constitutions also provide for a similar clemency power.²⁵⁴ The Founders understood the importance of establishing an avenue for mercy within the criminal justice system as a necessary check on a system that may allow for retribution.²⁵⁵ Moreover, it allows for more uniformity of treatment because it allows the president to consider national patterns, whereas prosecutors may focus on local interests.²⁵⁶ Some scholars argue that the Framers intended clemency to serve as an important check on overreach in punishing criminals, and if it goes unused, the justice system becomes unbalanced.²⁵⁷ Furthermore, while states may provide alternative mechanisms such as expungement, in the federal system, clemency is the only avenue for offenders to clear their records without further legislation.²⁵⁸

In the past, pardons were used more expansively to address policy issues, such as President Carter issuing amnesties for draft evaders,²⁵⁹ or George Washington pardoning participants in the Whiskey Rebellion.²⁶⁰ In more recent years, Presidents Reagan, Clinton, and Bush used the pardon power only sporadically, and did not use it to promote policy initiatives.²⁶¹ Modern presidents have differed in their approaches to the use of clemency. For example, President Obama granted 1,927 clemency petitions in his time as president,²⁶² with many recipients being offenders serving time for drug-related crimes.²⁶³ President Trump granted far fewer petitions,²⁶⁴ and many of the petitions he did grant were for people he knew personally.²⁶⁵ Clemency and commutation rates have declined drastically, most likely driven by tough-on-crime political forces.²⁶⁶ For example, while President Kennedy granted 33% of clemency requests, President George W. Bush granted only 2%.²⁶⁷ Similarly, from 1909 to 1913, President Taft granted 39% of requests, but President Obama granted only 5% during his tenure.²⁶⁸ Recent presidents have also used commutation to grant relief to criminal offenders. In 2024, President Biden commuted sentences of thirty-seven prisoners who were facing the death penalty with mixed reactions, some believing these were a “justifiable mercy,”²⁶⁹ and others claiming that victims were not shown appropriate justice.²⁷⁰ Overall, Biden pardoned over 4,000 individuals and Donald Trump has pardoned over 1,500.²⁷¹

The executive branch could consider data in its clemency decisions, potentially releasing a larger percentage of those incarcerated or considering data to determine which defendants are serving unfairly harsh sentences where others have received lesser punishments. Using clemency to reduce incarceration terms or inequity between those serving are two possibilities of expanding clemency beyond individual cases but considering it more broadly as a criminal justice reform measure. A potential problem to consider with the executive pardon power is its unlimited and unreviewable nature, raising the possibility for corruption and unequal treatment.²⁷² Another problem is that the agency in charge of submitting cases for review to the president is the Department of Justice (DOJ), which creates “tension because all clemency decisions are, in effect, reviews of prosecutorial decisions already made” by the same executive branch.²⁷³ The DOJ does not see its role as a policymaker but rather as an enforcer of the law, so clemency will be unlikely to be popular in that department.²⁷⁴ Rachel Barkow and Mark Osler suggest that using data-driven clemency might systematically help reduce sentences and check prosecutors who are charging offenders too aggressively.²⁷⁵ The data collected can also help to identify and correct areas where prosecutors are prone to error.²⁷⁶ Currently, clemency is not used in any meaningful way to address mass incarceration either on the federal or state level; but with the right considerations, it could be.²⁷⁷

Amnesty

Amnesty is a type of pardon, typically granted by a government to a group of people who have committed crimes, sometimes political in nature. President Andrew Johnson famously provided amnesty to all Confederate soldiers who decided to rejoin the Union.²⁷⁸ More recently, some scholars have proposed “Second Chance” programs to lawmakers that would grant first offenders amnesty after completing their sentence and a fixed term of probation.²⁷⁹ In 2018, the St. Louis Municipal Court established a warrant amnesty program, allowing anyone with an outstanding warrant for violating certain city ordinances to pay their original fines without late penalties.²⁸⁰ A potential use of amnesties is to stop mistakes in applications of the law. Martha Minow recommends the use of amnesties to “check overly zealous or biased prosecutions, excessively punitive rules, and simply mistaken policies.”²⁸¹ But thinking more broadly, amnesties could be applied extensively in cities and states to forgive certain types of criminal offenders, or those who have served a certain period or have undergone particular educational or rehabilitation programs while incarcerated. Amnesties can be used broadly and systematically by executive branch leaders to allow those serving long or punitive terms to have reduced sentences, or even to forgive those serving time due to mistaken policies.

While amnesty and pardons are not the best approaches to systemic absolution, as they are best handled on an individual basis, there are ways to

expand both practices that are already grounded in constitutional and legislative provisions to create larger avenues for forgiveness of crime.

IV. The Role of Systemic Absolution in Criminal Justice

Systemic absolution can create a more comprehensive theory of punishment and has the potential to transform criminal justice. While forgiveness or absolution might appear as only relevant individually, systemic absolution is already used in second look sentencing, automatic expungements of criminal records, and executive clemency as discussed above. It can also be expanded by increasing indeterminate sentencing, automating deferred adjudication, and by removing barriers for judges or parole boards to reduce sentences. Establishing a structural framework for these existing practices under the label of “systemic absolution” could help place them in proper context, promoting a more comprehensive approach to justice—one that allows for broad-scale forgiveness after punishment. There is a historic basis for societal forgiveness in Judaic, Roman, and indigenous laws. Understanding the religious roots behind our criminal justice policy allows us to question the guiding principles of retributivism and incapacitation that dominate our criminal justice policy. While it will always remain important to punish people proportionately and hold people accountable for crimes, these retributive aims are not complete without considering absolution after accountability. Current criminal justice theory is missing absolution, or a way to clean the slate after crime. Given that crimes by their nature are offenses against a state or community, a systemic absolution framework allows legislative, judicial, and executive branches to sort out the American mire of criminality.

The U.S. criminal justice system includes many people imprisoned for crimes, often because of a mistake made years ago or for those lacking resources or desire to change. Some have no interest in rehabilitation and remain a danger to society, but this is not the vast majority of the almost two million incarcerated. Many have potential to change but lack opportunities to rehabilitate and reintegrate into society. As such, the system hardly grants opportunities for individuals to start afresh, leaving many who remain in the cycle of criminality. This final Section introduces three considerations in adopting a systemic absolution framework. First, it considers criminal offenders potentially avoiding adequate accountability for crime. Second, it considers societal shifts on absolution that could allow forgiveness principles affecting individuals more universally to expand to macro-level systemic forgiveness. Third, it considers the perception of society on punishment and the age effect on crime and forgiveness. These principles contextualize systemic absolution with current accountability rates for crime, increased social interest in forgiveness, and empirical data demonstrating that society views current punishment trends as unduly harsh.

A. Perpetrators of Crime Avoiding Adequate Accountability

The criminal justice system, which currently fails to punish most offenders, could benefit from systemic absolution to deal with the twin challenges of mass incarceration and low criminal accountability. Scholars and policy advocates are aware of the American incarceration problem; America is number one in incarceration rates and will be for the foreseeable future. What is less understood is the amount of unresolved crime in America, or the low criminal accountability rate.²⁸² About twenty million serious crimes occur in America per year,²⁸³ but only a small fraction of these crimes are solved by police and then punished by the criminal justice system. Only 50% of serious crimes are reported to the police, 20% are solved, and 10% lead to a conviction.²⁸⁴ Thus, 90% of offenders go free, many without even being arrested. Arguably, America’s criminal justice system is technically absolving many serious crimes already, by failing to arrest most individuals. Many criminal justice scholars have overlooked this problem, but any argument that we should deal with criminal justice issues in a way that excludes these considerations is missing a big piece of the puzzle. Currently, society is both accepting high incarceration rates while ignoring the victims of serious crimes and failing to find a meaningful solution to the lack of criminal accountability.

Very few would argue that increased severity in incarceration is the best long-term approach to this twin problem of mass incarceration and low accountability.²⁸⁵ Even though there is a low probability of apprehension for offenders, the number and types of felonies that we punish have increased.²⁸⁶ The combination of low apprehension and high punishment increases the risk of recidivism, so the severity of punishment is not effective as a deterrent, and longer sentences are not the solution to recidivism.²⁸⁷ Incarceration is also an inadequate way to restore victims; because few offenders even get to the incarceration stage, most victims of crime are left without meaningful resolution. Increased certainty of punishment could help increase deterrence of crime. The few experiments with decriminalization demonstrate that this is not a viable solution.²⁸⁸

Punishment does not necessarily have to involve long-term incarceration and could include alternative approaches to criminal accountability. For instance, one approach discussed above that might institute broader criminal absolution would be to expand prosecutorial use of deferred prosecution as the norm for first-time offenders of crime, or those who might be especially well suited for mental health or drug rehabilitation. Prosecutors might charge crimes and delay trial or a plea agreement and sentencing rather than immediately convict offenders.²⁸⁹ Offenders must comply with certain standards that the prosecutor imposes as part of a “deferred prosecution agreement.”²⁹⁰ These requirements could include drug court, mental health court, drug testing, rehabilitation, mentoring and counseling, or other programs. If they meet the requirements and are not arrested again within a certain period, the prosecutor can drop the charges against them.²⁹¹ Like absolution in religion, criminal absolution might involve an offender recognizing her crime and taking measures that show her penitence. In return, the prosecutor, the person situated to impose justice on behalf of society, absolves her of punishment. However, this can be done automatically in a prosecutorial office where all individuals charged with certain crimes can have their charges dropped automatically if they remain law-abiding after a certain period of time. Deferred prosecution is one approach that might allow for a regular absolution of crime in a way that still maintains adequate deterrence for crime while allowing the principles of forgiveness to influence criminal justice.

B. Societal Shifts on Forgiveness

In recent years, forgiveness and restorative justice have been more prominent as individuals recognize the individual harm of holding on to resentment and hate. Empirical scholars like Megan Stevenson have argued that even well-designed criminal justice studies rarely produce meaningful change.²⁹² Stevenson points out that systemic change is a potential solution that involves changing the “minds and hearts” of many people “as well as changing the concrete structural factors.”²⁹³ The possibility of systemic change in the criminal justice system could be for communities to advocate for more formal adoption of systemic absolution, or broadscale forgiveness measures.

International charities are embracing forgiveness concepts,²⁹⁴ and the Templeton World Charity Foundation aims to reach 100 million people through evidence-based approaches to forgiveness in a worldwide “Discover Forgiveness” campaign.²⁹⁵ Similarly, in 2004, UK journalist Marina Cantacuzino founded The Forgiveness Project, which “provides resources and experiences to help people examine and overcome their own unresolved grievances.”²⁹⁶ The Forgiveness Project provides resources including shame training for people working in restorative justice, a group-based prison reform project “RESTORE,” and forgiveness-based lesson plans for schools.²⁹⁷ Other international initiatives, Beyond Conflict and the Institute for Historical Justice and Reconciliation, aim to rectify conflicts and reconcile contentious histories by employing practices of individual forgiveness.²⁹⁸ In one of the areas where societal norms have led before the legal academy, media and social science thinkers have explored important ground in considering the benefits of forgiveness to individuals.²⁹⁹

The legal academy has considered that the harm of punishment to an individual does not make the victim whole.³⁰⁰ Indeed, research demonstrates that forgiveness is more likely to help victims than punishment is. Michael Wenzel and Tyler Okimoto found that “a sense of justice restored through restorative responses promotes forgiveness, but justice restored through retributive responses does not.”³⁰¹ According to research, victims of robbery, burglary, and assault are twenty-three times more likely to feel they have received a sincere apology and four times less likely to desire revenge after restorative justice conferences as opposed to ordinary legal proceedings.³⁰² Victims expressed more satisfaction with restorative than adversarial processes and expressed excitement to help offenders turn their lives around, even though they rarely articulated the concept of forgiveness.³⁰³

There are many physical, mental, and emotional benefits for victims who choose to forgive their offenders, and negative effects when people fail to forgive.³⁰⁴ Positive psychology experts explain that when people forgive, their emotional reactivity decreases and their level of happiness goes up.³⁰⁵ However, the benefits of forgiveness for a victim are not as great as the “adverse health consequences” of failing to forgive.³⁰⁶ Unforgiveness “is associated with stress-related poor mental health, rumination, and depression.”³⁰⁷ Dr. John Clabby found physical harm for holding on to emotional injury and notes, in particular, adverse effects on the heart.³⁰⁸ Having each individual perpetrator or victim enter through the criminal justice system is not necessarily vital, as some benefits for both sides can be found through restorative measures.³⁰⁹ These individual benefits can extend to society if systemic absolution were adopted in a comprehensive crime framework.

C. Punishment Levels Need Adjusting

Both empirical research and societal surveys show that punishing for as long as we do might not benefit society. First, criminology research has long supported the crime-age curve that demonstrates how criminality rates peak in the late teens and drop steadily as age increases.³¹⁰ In 2019, individuals between the ages of fifteen and twenty-nine made up nearly forty-two percent of all arrests in the United States.³¹¹ Individuals aged fifty and above accounted for only thirteen percent of arrests.³¹² Other research has shown that there is a dramatic drop-off in criminality after the age of forty.³¹³ Indeed, we understand that older individuals are unlikely to recidivate. Congress has passed the First Step Act to allow release for those serving after the age of sixty and not charged with violent crimes. We also know that people typically age out of violent crime by fifty.³¹⁴ There is room for sentence revisions based on this information. However, older inmates charged with violent crimes are also very unlikely to recidivate.³¹⁵ And given that the bulk of state inmates are serving long sentences of violent crimes, and two out of three are over the age of fifty-five, reconsidering sentences for defendants between forty and sixty could dramatically decrease incarceration rates.³¹⁶ For example, society would not face increased safety threats when second look laws include individuals in their fifties and forties who have already served lengthy sentences for violent crime.³¹⁷

Empirical work by Paul Robinson demonstrates that society’s sense of justice is to punish much less than we currently do.³¹⁸ Robinson argues that most criminal law systems have already lost the moral credibility with their communities because they are far too harsh.³¹⁹ Robinson contends that there is more room for mercy in criminal justice and has advocated for a codification of rules that could be applied consistently and would not become dependent on the judge or some other extraneous factor.³²⁰ Robinson advocates for a “Mercy Code” with guidelines for what kind of behavior begets mercy and to what extent mercy will change the ultimate outcome of the sentencing for the offender.³²¹

Reevaluating sentencing through the lens of forgiveness is not to say that punishment should be abandoned altogether, but that second look laws should be expanded. Robinson argues that to encourage mercy without any kind of limits would create mayhem,³²² and I would argue that expanding forgiveness to avoid accountability would create a lack of order and improper incentives when it comes to crime. The research demonstrates that the “impulse to punish wrongdoing” is so deeply ingrained that most people would resist the abolition of all punishment.³²³ Additionally, the possibility of punishment serves to deter delinquency and a total lack of consequences could, and has, led to chaos in places that have experimented in this regard.³²⁴ As suggested above, to maintain accountability but also allow systemic absolution, offenders would be given the opportunity to enter into deferred prosecution as a default measure which allows a defendant to change or face prosecution under the law.

Conclusion

Communal absolution has been a fundamental part of societal practice throughout history, particularly in the criminal context. While some might be concerned that systemic absolution might undercut retributive justice that dominates modern criminal justice, its introduction could ameliorate the effects of an overly punitive system and make a meaningful dent in the United States’s mass incarceration problem. Forgiveness has an expansive history, particularly with ancient cultures that form the bedrock of criminal punishment theory. Many religions, including Christianity, Islam, Judaism, and Buddhism, employ some concept of absolution as a central part of their religious practice. Forgiveness also plays an integral role in individual mental health, while unforgiveness negatively impacts people’s overall well-being. While individual forgiveness is already part of our criminal justice system as a part of restorative justice, there has been no recognition of systemic absolution in modern criminal justice. Individual forgiveness in criminal justice currently includes clemency, expungement, apologies, the right of the victim to forgive, and restorative justice.

These are all important developments. However, without a move towards systemic forgiveness in criminal justice, the community continues to be burdened by the crimes of offenders long after they have been convicted and sentenced.³²⁵ Systemic absolution, discussed above, might include expansion of sentencing reduction legislation, defaulting to deferred adjudication by prosecutors, and automatic expungement efforts after a certain period of time of avoiding crime. Currently, communities carry a heavy social burden resulting from mass incarceration. Families and communities are broken apart as fathers, mothers, and siblings are removed from their support systems. Disproportionate harm towards minority groups results from higher rates of crime and incarceration and resulting collateral consequences inflicting communities after prison.³²⁶ Ultimately, distrust and hopelessness dominate criminal justice when punishment systems lack a system of absolution. Our current punishment structure envisions people as two-dimensional archetypes rather than complex human beings with inherent dignity. A person’s crime defines them indefinitely, limiting opportunities for change and holding them captive with societal stigma. Without a systemic mechanism recognizing that all of us collectively should not be punished perpetually for the worst thing we have done, society remains burdened by the weight of mass incarceration and punishment.

By incorporating systemic absolution into the criminal justice system, we can heal the many individuals who never receive justice—and even those who do—since the system has not proven to be effective at solving the underlying problems. Incorporating systemic forgiveness into our criminal justice system will also help society be more compassionate towards offenders and see “the full picture of their humanity,” unobscured by “their worst moment.”³²⁷ As Robin Steinberg observes that “[e]ach one of us has done something that could be defined as criminal—driving home after a few too many drinks . . . or being less than truthful on a tax return . . . . Whether or not our crimes have landed us in the criminal justice system . . . none of us wants to be defined by our most shameful moment.”³²⁸ Just as we do not want to be defined by these regrettable mistakes, neither do individuals who are incarcerated. Systemic absolution could expand healing for society as former offenders are given hope to live a normal life. Moving towards absolution in criminal justice provides helpful nomenclature for what is already happening when crime goes unpunished but allows for accountability in areas where societal priorities do not align with carceral outcomes. As past civilizations and religions have championed communal restitution,³²⁹ systemic absolution has the power to transform criminal justice today.

99 S. Cal. L. Rev. 41

Download

Shima Baradaran Baughman*

*Woodruff Deem Professor of law at BYU Law and Distinguished Fellow at Wheatley Institute. This piece benefited from an earlier less developed presentation at the Stanford/Yale/Harvard Criminal faculty workshop. I am grateful for Benjamin Hill, Bonnie Stewart, Lane Gibbons, and Mason Spedding for research assistance as well as Kory Staheli and Erica Larsen for institutional support. I am grateful for the excellent editing by the Southern California Law Review staff particularly Christopher Shahverdian and Kevin Zhang.

Sleep Deprivation in Prison

January 2026January 2026Sharon Dolovich*

Introduction

Human beings need sleep. We all know this. When we do not get enough sleep or when the sleep we get is repeatedly interrupted, our ability to function—to focus, to treat others appropriately, to respond effectively to life’s daily challenges—will be severely compromised. And these are just the effects we notice. Over the past decades, advances in sleep science have made clear just how necessary adequate sleep is to every dimension of human health. A chronic insufficiency of sleep increases the risk of, among other medical conditions, “obesity, diabetes, cardiovascular disease, and infections.”¹ As noted sleep scientist Matthew Walker puts it, “Every major system, tissue, and organ of your body suffers when sleep becomes short. . . . [T]he shorter your sleep, the shorter your life.”² And persistent inadequate sleep takes an equally great toll on mental health, exacerbating “all major psychiatric conditions, including depression, anxiety, and suicidality.”³

For adults, the consensus recommendation is a minimum of seven hours of uninterrupted sleep per night; otherwise, the processes operating at each stage of the sleep cycle will be unable to perform their restorative work.⁴ It is no secret that many Americans from all walks of life routinely fail to meet this target. Judging from the size of the sleep aid market—an estimated $65 billion is spent annually in the United States alone⁵—access to financial resources is no guarantee of adequate sleep. Yet, as Keisha Ray observes, the more limited one’s access to the social determinants of health—including, among other things, “food security, stable housing,” and “the freedom to choose where and how we sleep”—the less well-slept a person is likely to be.⁶ Access to adequate sleep, in other words, exists along a continuum, on which those lacking socioeconomic power and control over their physical environment are likely to experience a greater measure of sleep deprivation than society’s more privileged members.

This Article focuses on a group at the extreme end of this spectrum of disempowerment: those incarcerated in American prisons. Drawing on original data collected through in-depth, semi-structured interviews with people who served time in prisons all over the country, this Article maps the myriad overlapping conditions that systematically prevent people in prison from getting adequate sleep. Whether through luck, privilege, or their own affirmative choices, some people sometimes manage to get reasonably adequate sleep inside.⁷ But as my interviews show, most people in prison routinely sleep far less per night than the minimum seven hours that public health experts recommend.⁸ And for just about everyone living in prison, the quality of the sleep they get is consistently poor. In this Article, I offer a detailed account of why this is so.

To my knowledge, this study represents the first systematic effort to understand a destructive and dehumanizing experience that, although an inherent feature of prison life, has gone almost entirely unnoticed even by those academics, advocates, and policymakers invested in ensuring humane carceral conditions.⁹ Its conclusions are unambiguous: for a host of reasons explored in these pages, the incarcerated are chronically, perpetually sleep deprived—a condition they endure every single day, for years and sometimes even for decades.

The consequences of this situation are dire. Sleep deprivation compromises prisoners’¹⁰ physical and mental health, making them age faster and die younger.¹¹ It leaves people brittle, prone to frustration and anger, and without the cognitive resources to make good decisions, thus promoting conflict and violence in already volatile environments. And it strips them of the ability to think deeply and to exercise reasoned deliberation, capacities that are necessary for personal growth and healthy interpersonal bonds. Sleep is a basic human need, as vital to human survival and proper functioning as food, water, shelter, and personal safety. Yet persistent sleep deprivation—the product of what might be termed systematic sleep interference by the actors and institutions charged with administering carceral penalties—is a constitutive feature of life in custody. It heaps hardship, injury, and all manner of suffering on top of the formal punishment. And it represents a key, underappreciated reason why American carceral facilities are so noxious and inhumane.

Recognizing chronic sleep deprivation as a standard feature of prison life sheds new light on aspects of the carceral experience that have long seemed fixed and inescapable, including the high level of interpersonal tension, the pervasiveness of mental illness, and even the seemingly illogical willingness of people inside to put themselves into debt—a dangerous posture to adopt in prison—for the sake of a few candy bars or a bag of chips.¹² In prison, moreover, residents are not the only ones unable to get enough sleep. Thanks to shiftwork and overtime (whether mandatory or voluntary), correctional officers (“COs”) too are chronically tired.¹³ In short, in the often tense and high-octane environment of the prison, many if not most people in a facility at a given time, staff included, are wrestling with the debilitating effects of persistent sleep deprivation. This shared state of impairment cannot help but impact the functioning of the institution and poison interpersonal interactions between and among residents and staff.¹⁴

In future work, I will explore at greater length the implications of pervasive chronic tiredness on the character and operation of the modern American prison. I will also map the occupational health and safety effects on COs of a work schedule that often precludes anything close to sufficient sleep.¹⁵ Here, I focus on the basic phenomenon of sleep deprivation as a constitutive feature of the prison experience for those we incarcerate—a feature that, I argue, is as central to the punitive character of a prison sentence as is grossly inadequate medical and mental health care, the use of solitary confinement, the risk of physical and sexual assault, and the threat of excessive force by COs.

Two additional introductory points: First, this project focuses primarily on prisons, where time served is measured in years and decades. Yet sleep deprivation is also a problem in jail.¹⁶ If anything, it is worse. All my interview subjects did jail time before winding up in prison. And they universally reported that, however hard it was to sleep in prison, it was exponentially harder to sleep in jail. In jail, people may be detoxing. They are on average younger, more impulsive, and generally lack familiarity with the code of respect that can prevail in prison among people with long sentences. The constant churn means that people are always surrounded by strangers, a situation certain to exacerbate fear. The majority of those in jail at any given time are pretrial and understandably anxious about their cases. There is no programming to speak of, and mental illnesses will mostly go untreated. The combined result of these dynamics is a cacophony of noise and disruption in the housing units that, I am told, never stops.¹⁷ Among the troubling implications of this situation is that people detained pretrial routinely go to court exhausted, even to the point of falling asleep at counsel table during their trials.¹⁸ For most people, a stint in jail is relatively short. But even short periods of insufficient sleep can take a toll on physical and mental health. And for many people, a jail stay can last months or even years.¹⁹ For these reasons, much of what I say in this Article about sleep in prison should be taken to apply equally, if not more, to the context of jail.²⁰

Second, that COs are also chronically sleep deprived equally demands our attention. As a basic ethical matter, we ought to care about the experiences of COs, fellow human beings tasked with doing an impossible and often traumatizing job under difficult and at times dangerous circumstances.²¹ The least we can do is ensure that they have the opportunity to get adequate sleep, and thus avoid the harms that may come to them and their families and communities if they do not. That staff routinely get insufficient sleep is also extremely salient for those living behind bars. When staff are sleep deprived, prisoners suffer.²² As noted, in future work, I will investigate sleep deprivation among staff both as a matter of occupational health and safety and for its impact on the prison environment. I flag this issue now so that, by focusing here on the experience of sleep deprivation among the incarcerated, I do not create the impression that I imagine the problem to be exclusive to prisoners.

The argument of this Article proceeds as follows. Part I offers a short overview of the sleep science, focusing on the physiological and psychological harms and cognitive impairments traceable to insufficient sleep. Part II briefly describes the study parameters and methodology.²³ Parts III and IV represent the sociological heart of the Article, offering a thick description of the experience of trying to sleep while incarcerated. Part III describes the impact of a set of concrete conditions that systematically prevent people in prison from getting adequate sleep: uncomfortable beds, hunger, extremes of temperature, noise, and excessive light. Part IV identifies a set of what I call “meta-conditions”—specifically, fear of violence, trauma, poverty, overly-intrusive rules enforcement, and daily humiliations. These conditions, which likewise routinely compromise sleep, are also products of institutional choices, but they have been so naturalized to the life of the prison as to leave us blind both to their destructive effects and to their ultimately contingent character. Part V begins to explore the normative implications of the phenomenon identified here, including what it means for our understanding of prisons and of carceral punishment, the prospects for Eighth Amendment conditions claims grounded in sleep deprivation (and, very briefly, for claims of torture under international human rights law), and the policy challenges likely to confront efforts to address the problem. On each of these topics, there is much more that could be said. The goal here is to start the conversation.

I. A Brief Overview of the Sleep Science²⁴

Over the past twenty-five years, the field of sleep science has exploded.²⁵ Although the discipline is still developing, more than two decades of research strongly attests to the profound negative health effects—physical, psychological, and cognitive—of less-than-optimal sleep. Total sleep deprivation—enforced wakefulness over days—is “understood to be particularly detrimental.”²⁶ But interference with sleep need not be total to compromise physical and psychological health. Even people who regularly sleep five to six hours per night—considered “short sleepers” in the literature²⁷—exhibit notable adverse health effects, in some cases after only a few nights of restricted sleep.

Perhaps the most jarring finding to date has been confirmed in three separate meta-analyses. Encompassing studies comprising 1.32 million, 3.58 million, and 5.17 million subjects, respectively, all three investigations found short sleep duration²⁸ to be associated with an increased mortality risk of 12–13%.²⁹ No single driver of this mortality effect has been identified. Instead, the data shows significant associations between short sleep and a host of serious medical conditions, including heart attack,³⁰ hypertension,³¹ cardiovascular disease, coronary heart disease, and stroke.³² Short sleep and sleep disturbance have been associated with an increased risk of developing Alzheimer’s disease,³³ as well as a host of metabolic disorders, including a heightened risk of insulin resistance,³⁴ obesity,³⁵ metabolic syndrome,³⁶ and Type 2 diabetes.³⁷ Insufficient sleep has also been found to cause imbalances in the autonomic nervous system, the indicia of which include increased heart rate,³⁸ elevated cortisol levels,³⁹ decreased heart rate variability,⁴⁰ and increased levels of norepinephrine.⁴¹ Among other adverse physiological effects, this collection of symptoms increases the risk of cardiovascular disease and represents a predictive marker for “sudden cardiac death.”⁴²

Sleep deprivation has also been found to generate adverse psychological effects. For example, sympathetic bias in the operation of the autonomic nervous system has been linked with conditions indicating “emotional dysregulation, such as anxiety, depression, and rigid attentional processing of threat.”⁴³ Persistent sleep disturbance is consistently associated with a heightened risk of developing major psychiatric disorders, including depression, anxiety, and suicidality.⁴⁴ Concerning depression, one major meta-analysis, incorporating research from twenty-one qualifying studies, found that “nondepressed subjects” who experienced “difficulties in initiating/maintaining sleep or non-restorative sleep”⁴⁵ have “a twofold risk to develop depression, compared to people with no sleep difficulties.”⁴⁶ Research also indicates a greater likelihood of mental health issues among those who experience sleep disturbance over time. One early study found that respondents with persistent sleep disturbance “had significantly higher rates of new cases of both major depression and anxiety disorders compared with those whose insomnia resolved.”⁴⁷ Other studies have begun to map the link between sleep disturbance and the symptoms of bipolar disorder.⁴⁸ There is even some evidence suggesting that sleep disturbance increases proneness to substance use disorder.⁴⁹

Inadequate sleep also compromises cognitive functioning—and the more sleep deprived a person is, the greater the cognitive effect.⁵⁰ Multiple well-controlled studies have demonstrated the impact of sleep deprivation on the capacity for vigilance, working memory, and sustained attention, with one experimental study finding that “chronic restriction of sleep to [six] h[ours] or less per night produced cognitive performance deficits equivalent to up to two nights of total sleep deprivation.”⁵¹ Sleep deprivation also slows reaction time and compromises performance in the completion of cognitive tasks.⁵² When sleep restriction is chronic, these cognitive defaults are cumulative,⁵³ and when people routinely sleep less than seven hours per night, the resulting cognitive impairments are as marked as in those experiencing “severe acute total sleep deprivation.”⁵⁴

Finally, sleep deprivation impairs one’s ability to successfully navigate stressful or frustrating situations. These effects are evident after just one night of inadequate sleep. As Andrea Goldstein and Matthew Walker put it in their survey of the field, even “one night of experimentally controlled sleep loss increases subjective reports of stress, anxiety and anger in response to low-stress situations and increases impulsivity towards negative stimuli.”⁵⁵ Insufficient sleep is also “associated with a lack of flexibility and capacity to respond to emotional challenges,” and thus with a decrease in emotional resilience.⁵⁶ Together, these emotional regulatory failures undermine the capacities necessary if one is to get along with others. In one study, researchers showed that sleep deprivation was “associated with increased outward expression of aggressive responses and a greater tendency to assign blame to others when confronted with frustrating situations.”⁵⁷ Subjects tended towards “antagonistic, self-serving, and uncooperative behavioral responses, such as increased blame and hostility, reduced willingness to make concessions and restitutions, and a general increase in atypical or impulsive responses, all of which are antithetical to harmonious interpersonal relations.”⁵⁸

The aim here is not to provide a comprehensive account of all relevant findings related to the negative health effects of sleep deprivation. The field is too vast to make such a goal workable. The point is rather to make clear that interference with sleep is not simply a matter of losing access to the “creature comforts” widely considered inconsistent with carceral penalties. Chronic sleep deprivation carries an increased risk of physiological harm, early mortality, and a range of psychopathological effects. It takes a toll on cognitive functioning and on a person’s capacity for prosocial interpersonal interaction. To the extent that prison systematically impedes a person’s ability to sleep over months, years, and even decades, this feature of a prison sentence can inflict substantial long-term damage. This brings us to the central question of this Article: How much and how well do people sleep in prison?

II. The Study⁵⁹

In prison, it is virtually impossible for people to get the recommended minimum seven hours of uninterrupted sleep. On most nights, most people inside will get only four to six hours⁶⁰ of light sleep punctuated by frequent interruptions—a pattern that persists night after night, for years.

This is the unmistakable conclusion that emerged from over seventy‑four hours of interviews I conducted in 2023 with thirty-nine people⁶¹ formerly incarcerated in American prisons nationwide.⁶² Of these thirty-nine interviews, thirty-five were formal, semi-structured interviews based around a questionnaire with roughly 200 questions,⁶³ with the average interview lasting just under two hours.⁶⁴ Subjects were asked about their experiences of sleeping/trying to sleep in prison, the factors that may have prevented them from getting adequate sleep inside, and the ways their own persistent sleep deprivation—or that of staff or other residents—may have affected their daily lives while incarcerated. In addition, I conducted four somewhat more freeform on-the-record conversations (labeled “ad hoc” or “AH” interviews to distinguish them from the more formal interviews) with subjects who were outside the study parameters for the formal interviews, but whose insights and experiences promised to provide valuable perspective and context. During these conversations, I asked many of the same questions I asked during the formal interviews. In the end, much of what I heard in all my interviews, ad hoc interviews included, proved remarkably uniform. As a result, I wound up drawing on all thirty-nine interviews when analyzing the data.⁶⁵

With one exception, each interview subject had spent at least four years in prison.⁶⁶ The average length of incarceration among my interviewees was thirteen and a half years,⁶⁷ with fourteen individuals in the sample having served more than twenty years and eight serving more than twenty-five years. In terms of jurisdictional diversity, eighteen interviewees served time in four states: California (five), New Jersey (five), New York (four), and Louisiana (four). Between them, the remaining twenty-one participants had experience of prisons in fourteen other geographically diverse states plus the federal Bureau of Prisons.⁶⁸

Everyone I interviewed was incarcerated in multiple prisons over the course of their confinement. This broad experience proved valuable in two ways. First, it allowed me to learn about conditions in many more facilities than I had interview subjects. Collectively, my interview subjects did time in at least 185 separate institutions,⁶⁹ including 126 state prisons, 18 federal prisons, 33 jails,⁷⁰ and 8 youth detention centers.⁷¹ By speaking to thirty-nine people, I was thus able to hear firsthand about conditions in 8.5% of all prisons in the United States, including 8% of state prisons and 18% of federal prisons.⁷² Second, it meant that virtually everyone was able to compare conditions among institutions. Although the interviews revealed some notable regional variation, perhaps most striking was just how nationally uniform were experiences bearing on sleep—with any jurisdictional differences for the most part proving more a matter of degree than kind.

In terms of race, my interview sample self-identified as follows: 15 Black (38%), 13 White (33%), 7 Hispanic (18%), 1 Asian (3%), 1 Native American (3%), and 2 mixed race⁷³ (6%). The goal was to replicate as nearly as possible the racial distribution of the American prison population, and in the end, I came close.⁷⁴ As for gender, four of my thirty-nine interview subjects were housed in women’s prisons, comprising 10% of the total,⁷⁵ with the remainder doing their time in men’s prisons.⁷⁶ These numbers also came close to the gender breakdown of the American prison population overall.⁷⁷

It would be impossible to reproduce in full the richness of the narratives the interviews yielded. Instead, in what follows, I distill the key points that emerged around the experience of sleeping in prison and the conditions that impede sleep. In part, the power of the methodology employed—the long-form, semi-structured interview—is the degree of confidence it can afford in the accuracy of what one hears when the same thing is reported by multiple subjects in a diverse sample.⁷⁸ On this measure, confidence is warranted as to many of the topics covered in my interviews, which yielded a high degree of consistency despite the diversity of the sample across, among other factors, race, gender, jurisdiction, housing configuration, and security level.⁷⁹ Where my interviews revealed variation as to a particular issue or experience, I indicate as much in the text.

To my knowledge, this is the first study to attempt to capture what it is like to try to sleep in prison in the United States. The goal here is to begin to excavate a phenomenon that has to this point eluded scrutiny, despite the centrality of sleep—and, it appears, sleep deprivation—to the American carceral experience. Inevitably, some relevant aspects will remain unaddressed. One central contribution of this work is to provide an initial account that may be built out through future research.

III. Findings I: Carceral Conditions Impeding Sleep in Prison

In this Part, I describe specific, discrete conditions persistently impeding the ability of people to sleep in prison. Representing features of the physical plant (e.g., lighting, HVAC, beds, and bedding) and institutional operations (e.g., housing configurations, methods of conducting count and security checks, mealtimes), they are of the sort typically considered appropriate targets for more conventional policy reform. Of these sleep-compromising conditions, two categories—beds and bedding, and food and hunger—appear from my interviews to be experienced across the board by pretty much everyone, regardless of security level, jurisdiction, or any other factors. A further set of conditions—those salient during periods of extreme heat or cold—impact people differently depending on region and time of year. Finally, two other issues—noise and excessive light—are experienced differently depending on a range of factors, including housing configuration, security level, staff behavior, and the happenstance of the unit in which one is housed and where one’s bunk is located.

In what follows, I explore each factor in turn, describing how each impacts the ability of people to sleep in prison. But this approach carries a risk, as it makes it seem as if the culprit is exclusively poor system design. To some extent this is so, and for this reason, implementing changes on the ground is one obvious policy response to the findings reported here. At the same time, this focus on concrete institutional arrangements risks obscuring a deeper, more pernicious truth, one that I address in more detail in Part IV. There, I explore a second set of carceral conditions—termed “meta-conditions”—which more directly reflect the prison’s toxic moral foundations and COs’ consequent blindness to the humanity of those in their custody. If prison officials cannot see the people in the beds at night as human beings who need sleep to survive and properly function, they will not act, or run the prisons, in ways that protect and promote adequate sleep. And sure enough, in American prisons, they do not.⁸⁰

A. Uncomfortable Beds/Bedding/Mattresses

In prison, the construction and nature of the beds and bedding alone would be enough to keep most people from getting decent sleep. The beds are rock hard—typically just a metal or concrete slab—and so narrow and short that people who are taller or heavier than average often will not fit. “Mattresses” are nothing more than thin pads covered with a rubbery plastic that is sticky in the heat and cracks in the cold. But equally if not more disruptive of sleep is the construction of the bunkbeds, the main sleeping configuration for people in prison. Bunkmates feel and are disturbed by every slight movement, which, depending on how people get along, can make sleeping a stressful experience for those afraid to spark conflict simply by turning over. For this reason, or simply out of respect for a bunkmate’s sleep, those on top bunks may hesitate to leave their beds at night even to urinate. Not only do those who try risk waking frustrated bunkmates who are themselves desperate for sleep, but the high likelihood of disrupting their neighbor’s personal property as they try to get in and out of their bunks in the dark while half asleep generates further stress and potential conflict. None of this is conducive to enabling sleep.

In prison, the beds are “rock hard.”⁸¹ For the most part, the design is the same: metal frames with a solid concrete or metal pallet—a “large . . . piece of sheet metal with four legs bolted to it.”⁸² They are also narrow, so “[y]ou[’ve] got to train yourself to sleep in this small space and not roll over . . . onto the floor.”⁸³ People who are bigger or taller than average have an especially hard time. For those who are overweight, it can be “hard to turn over . . . because the beds are small.”⁸⁴ One person, who at one point weighed 330 pounds, described how he needed to “lift up and flop” to turn over “[b]ecause you can’t roll. If I roll, I’m going [to wind up on] the floor.”⁸⁵ Those who are tall have a different problem. One especially tall person “had to sleep on the bottom bunk with his feet on the toilet.”⁸⁶ Another “big guy,” the cellmate of an interviewee, “had to sleep [in] reverse [of] the normal way so his feet could hang off the edge of the bed.”⁸⁷ To do this, “he had to tuck his head underneath the shelf” which would ordinarily be at the foot of the bed.⁸⁸

Were the mattresses sufficiently thick and springy, the concrete or metal pallets might not cause problems. This, however, is not the case. Each time I asked about mattresses, I was treated to the same visual representation,⁸⁹ with people using thumb and index finger to indicate roughly three to four inches, or five inches at most. And this was when mattresses were new. As one person explained, “at one point they were about four or five inches thick, but then . . . as time[] went on, the middle part bec[ame] smaller . . . [and] at some points, it can become nothing more than two pieces of fabric.”⁹⁰ The thinness of the mattresses was an issue for everyone: “you’re sleeping pretty much on metal,”⁹¹ and “you’re gonna feel that iron underneath you every frickin night.”⁹² Many people reported experiencing perpetual discomfort, with each night an ordeal of tossing and turning. The hardness of the mattresses was “difficult on your back, so you wind up with back problems”⁹³ or “sciatic nerve damage.”⁹⁴ In some cases, people reported such problems persisting after release.⁹⁵

The mattresses’ material also causes problems. The covers are a plastic, rubbery substance to which bodies stick, especially on hot nights.⁹⁶ One person described preferring to “sleep on the middle piece of steel,” because on the mattress he “would feel the heat and the sweat.”⁹⁷ Even on temperate nights, the rubbery material was unpleasant to the touch. One could cover the mattress with a sheet, but the “vinylized rubber” material “will [still] make you very hot and sweaty.”⁹⁸ And as the mattresses age, “the plastic [gets] all cracked and ripped,” and “the person who slept in it before you [was] sweating [so] it gets moldy—the foam holds the mold, and you can smell it.”⁹⁹ Replacing bedding, including mattresses, is at the COs’ discretion. As one person put it, “[y]ou could request a new mattress, and depending upon the officer . . . on duty at the time, he’ll put your request through, or he’ll just toss it into the garbage can or run it through the shredder.”¹⁰⁰

Blankets are generally thin: “[S]ummer blanket[s]” are “the lowest quality of . . . cotton”¹⁰¹—“more like fish nets, . . . maybe crocheted or something but . . . they weren’t solid, so air can go through it.”¹⁰² Whatever the ambient temperature, people tend to be allocated one blanket only.¹⁰³ As for pillows, most prisons appear not to supply them. In many places, the best one might get is a pillowcase, which allows for handmade pillows stuffed with “hoodies and sweatpants.”¹⁰⁴ This approach does not afford a comfortable sleep. If pillows are supplied, they “could be hard as a rock or . . . flat as a pancake.”¹⁰⁵ As one person with a gift for metaphor put it: “[T]he pillows you’re sleeping on . . . might as well stack a couple sheets of paper together and put your head on that.”¹⁰⁶ In some prisons, the mattresses came with “a little bump at one end that was supposed to be a pillow. . . . [A]n inch high or something.”¹⁰⁷ When I asked if that did the trick, people just scoffed.

Then there is the fact that people largely sleep on double bunks that share a metal frame.¹⁰⁸ It is hard to overstate how much this arrangement compromises sleep. When the beds are connected this way, you feel every movement the other person makes. Being “crummy old metal bunks,” they “move and squeak a lot,” which means “you’re going to wake up when your bunkmate gets up or even just moves around.”¹⁰⁹ When people put their weight on the bed, it is not unusual for the flat metal of the pallet to dip slightly and then make a loud pop when the weight is lifted.¹¹⁰ People described strategizing ways to shift position in bed that would not make noise, and even staying put rather than, say, getting up to urinate, out of a desire not to disturb their cellmate. Even the ordinary movement of a restless sleep could prevent a bunkmate from sleeping. One person described an instance when he was in terrible pain from a toothache, “I would wake up in the middle of [the] night, I’m like [makes a face like in pain, moving head around], but I would make sure I wouldn’t toss and turn too much [out of] respect for my cell[mate] . . . so he gets his sleep.”¹¹¹ Disturbing a cellmate or neighbors in the dorms at night can be a serious source of friction “[s]o it was very important to be quiet.”¹¹² The burden on sleep here is twofold: people hesitate to make adjustments that would help them fall (back) asleep, and they experience stress at the thought of disturbing a frustrated, sleep-deprived cellmate. The physiological effects of this stress alone can cut against achieving the relaxed state that sleep requires.

In some prisons, bunks are not attached to one another but only to the wall. Although these bunks “weren’t as sensitive, they definitely still connected through the ladder and everything. So . . . one hundred percent you could feel it” when the other person moved in bed.¹¹³ And with beds connected to the wall, the bed of “the person in the cell next to you” may be connected through the same wall, “so when that person moved, your bed moves.”¹¹⁴ If that person happened to be “a very deep and robust snorer . . . you could actually feel the wall and your bed vibrate when they snore[d].”¹¹⁵

Whatever the structure of the bunks, the person on the top bunk is typically forced to find some creative way to get in and out of bed. There are small ladders attached to the beds—really just “a couple of rungs attached to the bunk”¹¹⁶—but if you use them, “your bottom bunkmate is going to feel the bed move.”¹¹⁷ So people descending from top bunks may instead step on the desk, commode, or other available surface, which might lead them to jostle or even displace the property of their cellmate. The risk this operation entails is particularly elevated in the middle of the night. For this reason too, those on the top bunks might forbear from getting out of bed at night to relieve themselves, even if the need were pressing, simply to avoid provoking conflict.¹¹⁸ Bottom bunks are widely preferred, although people on bottom bunks also need to take care that in moving, they don’t disturb their bunkmates. One person reported that “in the middle of the night, when [he would] go to turn over, if [he] was on the bottom bunk, [he would] make sure to keep [his] hands down low, so [he] won’t accidentally hit [the top bunk].”¹¹⁹ And bottom bunks present their own issues. More than one person described the jarring impression of waking from sleep to find a dark shadow looming over them. By the time they realized it was just their cellmate using the commode, they were wide awake, the effect of an adrenaline rush sparked by fear.¹²⁰

Just about everyone tries in some way to increase the comfort level of their sleeping arrangements. The most common approach is to angle for a second mattress, blanket, or pillow, usually by being the first to snag one left behind by someone being transferred or released. But in every facility, having anything beyond what was issued by the prison—generally one mattress, one blanket, two sheets, and possibly a pillow or pillowcase—is a clear rules violation and could result in a write-up.¹²¹ In many cases, COs won’t bother writing the ticket, but they will likely confiscate the extra items—“if you get caught with doubles, they’ll take it.”¹²² Those who manage to procure an extra mattress might pursue a more effective (if more audacious) approach: stuffing one mattress inside the other. As one person explained, “[T]here’s a seam on the mattress, so I unstitched it and slid the other one in it and then sewed it back up.”¹²³ These measures could make a difference on the margins: after a while, one’s “mattress would be like pancakes. . . . [T]here’s a difference between two mats and one mat for sleeping.”¹²⁴ Still, the difference is only a matter of degree, and even doubled up, “it’s still bad.”¹²⁵ As one mattress-doubler put it, “I was limiting the amount of damage that the regular mattress was doing to me by a small percentage. But when you’re gasping for air, you’ll take any air you can get.”¹²⁶

B. Hunger

Three basic aspects of food in prison contribute to the hunger that, according to my interviews, universally impedes sleep: quality, quantity, and timing. The quality of the food is so poor that people often have to force it down and will often opt not even to try. The quantity is frequently insufficient to induce satiety. And most significantly, the timing of the meals is such that there is typically a gap of at least five or six hours without food between “dinner” and lights out. Every single person I spoke with affirmed that if you have no money on your books for food from commissary, you will go to bed hungry, a condition that makes it difficult if not impossible to sleep. Those with family support will usually have food to eat before they go to sleep. The rest scramble to make money with side hustles (e.g., cleaning cells, making and selling food, doing legal work), or resort to less salutary strategies that could put them at risk (e.g., stealing, gambling, selling sex). The desperate desire not to go to bed hungry appears to be a driver of some of the most dangerous and pathological behaviors adopted by people in prison.

The other obstacle to sleep universally attested to in my interviews relates to food. Together, three food-related issues—quality, quantity, and timing—ensure that those able to eat only what the prison serves will routinely go to bed hungry, a condition that can make it hard to fall or stay asleep. First, the quality of the food served in prison is notoriously poor and often inedible.¹²⁷ One person I spoke to, capturing the general sentiment, described prison food as “garbage, . . . like what they feed a pig. If you put [slop] in a bucket and throw it in a pig’s face . . . [t]hat’s what it looks like.”¹²⁸ So unappetizing is the food served that those with other options will choose to forgo the prison fare altogether. The quantity is also frequently insufficient, and, as a result, even people who eat everything served to them still may not get enough to feel satiated.¹²⁹ One person reported losing “thirty or forty pounds just on the food. . . . It was just very, very, very small portions.”¹³⁰

But in terms of impact on sleep, the most consequential issue is the timing of meals. At some point, I began asking when meals were served. As to breakfast, answers varied from as early as 3:00 a.m. to as late as 8:00 a.m., with the majority reporting breakfast times between 4:00 a.m. and 7:30 a.m.¹³¹ Of still greater significance in terms of sleep is the timing of dinner. Here, answers ranged from 2:00 p.m. to 6:30 p.m., with the overwhelming majority reporting a dinner time between 4:00 p.m. and 5:00 p.m., or earlier.¹³² In most prisons, people appear to be locked in for the night around 9:00 p.m. or 9:30 p.m., with the lights dimmed for sleep around 10:00 p.m. Even those who go immediately to sleep at 10:00 p.m. will have gone at least five hours without eating, and for those who stay up later, the delay without food will be correspondingly longer—what one person called “intermittent fasting before it became fashionable.”¹³³

Those with the good fortune to work in the kitchen have “access to extra food left over.”¹³⁴ Otherwise, thanks to the timing of dinner, “[a] whole lot of people went to sleep hungry.”¹³⁵ To avoid this situation, people try to have food on hand to eat before going to sleep. In facilities with a central dining area, although “you’re not supposed to,” people would “bring the food out [of the chow hall] . . . . Might be rice and beans. I’ll bag that up and bring that back.”¹³⁶ In facilities where food is served in cells or day rooms, people might keep back food from their trays. But in many prisons, being caught with prison-issued food in one’s cell is a rules violation and “sometimes, you’re going to get caught.”¹³⁷ When that happens, even if you are not written up, the food will be confiscated, leaving the problem of nighttime hunger unaddressed.

There is, however, one thing that allows people to stave off hunger: access to commissary (also known as “canteen” or “store”). In every prison, people can purchase a range of items from what is in effect a virtual prison shop, which typically sells a mix of food, personal hygiene products, and other miscellaneous items.¹³⁸ Each facility has an approved vendor, and people fill out forms indicating what they want to buy, with the funds coming out of the purchaser’s prison account. As in society in general, there are all kinds of reasons to want money in prison.¹³⁹ But judging from my conversations—which again, involved people from all regions of the country, diverse as to race, gender, jurisdiction, housing configuration, and security level¹⁴⁰—the single biggest reason is that having money or its equivalent¹⁴¹ allows you to get the food you need so that you do not go to bed hungry. As one person recalled, “When I first [got to prison], I didn’t have as much money coming to me. So there were nights where I would go to bed starving. And you just couldn’t sleep because of how hungry you were, or [you would] wake up early waiting for breakfast . . . just because you’re so hungry.”¹⁴² People described “nights that you would be tossing and turning because your stomach was hurting” from hunger;¹⁴³ “nights of not having food,” when “you [were] so hungry, you[’d] take your hands and push in on your stomach to make [it] tighter, and try[] to fall asleep like that”;¹⁴⁴ and nights of “being so hungry to the point where you can’t even sleep.”¹⁴⁵

In prison, commissaries typically price goods close to or even higher than market rates,¹⁴⁶ yet people who work in prison will earn well below $1 an hour.¹⁴⁷ A key determinant of a person’s ability to buy items from the store is thus whether someone on the outside is putting money on their books or whether they have some other way to make money while inside. I spoke to one person who had a relatively well-paying job with a private company run with prison labor.¹⁴⁸ But in most cases, people who received no funds from family found some side hustle in the prison.¹⁴⁹ People would “make and sell wine,”¹⁵⁰ “homemade . . . cards,”¹⁵¹ or “prison burritos”¹⁵² crafted from ingredients bought from commissary or kitchen workers and sold for a markup.¹⁵³ Or they might “wash other people’s clothes” or “bowls [as a] kind of dishwasher,” “sell their phone time,” “do legal work for people,” or “type papers” for those in school.¹⁵⁴ One person I spoke to “fixed all electronics like headphones, fan, Walkman, TV, radio, anything. . . . Even the officers would bring [him] broken fans and watches and stuff.”¹⁵⁵ Like others with side hustles, he took his payment “[i]n commissary, in food.”¹⁵⁶

Those without family support or some effective way to earn money found other ways to get food or the money to buy it. One person I spoke to, who “didn’t have [any money] for a long time, . . . started gambling and stealing . . . to provide for [her]self.”¹⁵⁷ People “sold drugs to make ends meet”¹⁵⁸ or exchanged “sex for food”¹⁵⁹ or “g[o]t jobs in the kitchen so they could steal.”¹⁶⁰ Others would “panhandle, ask your neighbors, your buddies,”¹⁶¹ or they might resort to a more general appeal:

[T]here were people who had no money and they would stand out by where we get our trays. And they’re asking people, you know, “Oh, do you want the rest of that?” and they’re collecting it, and putting it into a bowl to try to have what they need for that night.¹⁶²

However one does it, the main goal is to have provisions available so that, when 10:00 p.m. rolls around, a person has something to eat before they try to sleep.

Even people with the resources for commissary may sometimes go to bed hungry. People may run out of food before the next canteen day. Or they might have had extra expenses that ate into their stash for the month. COs might have tossed their cell, and in the process rendered whatever food they had inedible—tearing open packages, contaminating their provisions, and so on.¹⁶³ A rules violation sometimes carries the penalty of loss of canteen privileges. People are often transferred between facilities, and it can take time for the finances to catch up, which can mean missing several cycles of commissary while you wait.¹⁶⁴ One person reported being transferred to “a very restrictive facility,” where “you come in [and] you can’t get commissary for the first thirty days.”¹⁶⁵ And perhaps most notably, access to property is contingent on remaining in general population. When people are sent to solitary (a.k.a. administrative segregation or “ad seg”), whether for discipline, protective custody, or suicide watch, they are not permitted to bring any property with them and are highly restricted in what they can get while they are there. Generally, in ad seg, people “don’t have any access to commissary,”¹⁶⁶ so people are completely dependent on whatever food is served. As one person explained, people in ad seg will “get fed [dinner] around 3:30 or 4:00 [p.m.], and they don’t get anything [else] [un]til about 6:30 [a.m.].”¹⁶⁷ To make matters worse, in ad seg, you can’t even “save food from your evening meal to try to hold it [t]o have something to eat in the morning or in the middle of the night”¹⁶⁸ because doing so is a rules violation—“if they find you with that, then they add extra time to the time that you stay in seg.”¹⁶⁹ Whatever the obstacles to accessing commissary, when people are unable to buy food to supplement prison-issue meals, they go to bed hungry, which makes it hard to sleep—even assuming a sleeping environment that is temperate, quiet, and dark, which, as will be seen, prison environments frequently are not.

C. Extremes of Heat and Cold

To get adequate sleep, the body needs to be at ease. Yet for months every year, people in prisons nationwide are forced to try to sleep in sweltering heat or freezing cold. Predictably, this leads to conflict. Men in dorms fight over where to direct the few available fans—battles generally won by those whose threats of violence are most credible. People struggling to get comfortable in their bunks toss and turn, drawing the ire of bunkmates whose own efforts at sleep are thereby disturbed. And cellmates with different sleeping preferences clash over whether to close windows or block vents. It would be too simple to put this down to changing climate. The ambient temperature in carceral facilities is an operational choice. Although average annual temperatures continue to climb, most prisons have no air conditioning in the housing units (even while providing air conditioning in the spaces where COs congregate, including staff lounges and officers’ booths). And in the coldest months, poorly maintained physical plants—broken windows, ineffective or broken heating systems, and so on—leave people too cold to sleep.

In prison, the ability to sleep is greatly impacted by the ambient temperature. In society in general, people who have control over their sleeping environments will adjust their thermostats to a comfortable setting. When two or more people share living space, there may be competing preferences as to the best temperature for sleeping. But it is rare that, given the choice, people would opt to sleep in the freezing cold or in sweltering heat. The reason is obvious: to get adequate sleep, the body needs to be at ease. In conditions of extreme heat or cold, the physical relaxation necessary to fall and stay asleep becomes unattainable.

Apart from those few fortunates who did their time in facilities with air conditioning, functional heating systems, and well-maintained physical plants,¹⁷⁰ my interviewees consistently reported long stretches each year when extremes of heat or cold greatly impeded sleep. Consider first excessive heat. Multiple interviewees from across the country described enduring many months every year when it was too hot to sleep, when it got “so hot and humid that the walls are sweating [and] [t]he floors constantly have that dewiness on them,”¹⁷¹ when you “sweat so much, [i]t wears you out,”¹⁷² “your shirts would be dripping wet. You actually could wring your shirts out.”¹⁷³ As one person succinctly put it: “In the summer, it’s so hot, it’s like you’re in a blowtorch.”¹⁷⁴

Because heat rises, those on the top tiers suffer most. One person described “working in the infirmary,” and “hav[ing] to respond to guys that were living up there. And as you walked up five flights of stairs, it was like . . . [you were] in a sauna all of a sudden, . . . like [a] forty to fifty degree temperature change sometimes. It was ridiculously hot.”¹⁷⁵ In the summer months, the dorms can also get “exceedingly hot”: “If it’s 100 degrees outside, it’s gonna be 110 degrees in the dormitories.”¹⁷⁶ When it was this hot, no one was sleeping. “[All that] body heat. . . . Everybody’s tossing and turning.”¹⁷⁷ One person I spoke to described being “up all times a night.”¹⁷⁸ He would “get up soaking wet. Go to the bathroom, get a cold towel, wipe my skin off. Go lay back down. Wake up an hour and a half later [and] can’t sleep.”¹⁷⁹ Even outside the South, this situation was hardly rare; one person in a federal facility in the Northeast described being “super-duper hot and miserable” during the height of summer.¹⁸⁰ In the South, heat this intense could last six months or more each year.¹⁸¹

Predictably, this degree of discomfort sparks frustration and even threats of violence. In the dorms, prison officials sometimes set up industrial-strength fans, but their placement and direction can provoke anger and even lead to fights as people desperate for sleep struggle to get or retain access to the relief the fans provide. As one person described it, “[e]verybody’s fighting for the fan. The fan has been cocked this way, cocked that way. You got guys waking up in the middle of the night, ‘Man, don’t turn that fan no more,’ which caus[es] a fight [and] now everybody waking up. ‘Man, listen, leave the fucking fan alone.’ ”¹⁸² If tensions did not run as high in the cells, periods of extreme heat still sparked issues. In the heat, people who are unable to sleep find themselves shifting in their bunks trying to get comfortable. The mattresses, as we have seen, are typically wrapped in a kind of rubber¹⁸³ to which bodies would stick in the heat. And as we have also seen, when people are double-bunked, every move one person makes can be heard and felt by the other, especially when the bunks share a single frame.¹⁸⁴ Out of respect, people try to minimize movement as much as possible. But the cost of this courtesy is being forced to remain immobile even amidst the great discomfort excessive heat creates, making sleep even more elusive.

It is not that prison officials are unaware of the heat. More than one person I spoke to described housing units so hot that staff avoided coming inside. One interviewee noted that COs would “sit outside in an air conditioned booth all night, and they would come inside of the dormitory, make their rounds, . . . and then go back into an air conditioned booth while we stayed in the dormitory.”¹⁸⁵ The lack of air conditioning in Southern prisons in particular has already become the subject of constitutional litigation, with Eighth Amendment challenges to excessive heat grounded in the substantial risk of serious harm posed to people with certain medical conditions, the symptoms of which are exacerbated by heat exposure.¹⁸⁶ In Ball v. LeBlanc, a case concerning excessive heat on death row at the Louisiana State Penitentiary (a.k.a. “Angola”), the district court appointed outside experts to monitor the temperature over three weeks in July and August.¹⁸⁷ During that time, the heat index in the facility rose as high as 107.78 degrees Fahrenheit and remained over 100 degrees for a full week.¹⁸⁸ The district court ordered the Louisiana Department of Corrections (“DOC”) to install air conditioning, but the Fifth Circuit vacated that part of the order.¹⁸⁹ In 2018, the case settled, with the DOC committing to providing a minimum of fifteen minutes of shower time each day, ice machines and ice containers, fans, water, icy breeze units, and diversion of cool air from the guard’s pod to the area of the plaintiffs’ cells on any day when the heat index in the plaintiffs’ cells exceeds 88 degrees.¹⁹⁰ Judging from what I heard, whatever else these half-measures might accomplish, they are unlikely to do much to improve the sleep of people in Louisiana prisons during the hot summer months.

Being too cold can also obstruct sleep. Virtually all facilities have heating systems for cold winter months, yet judging from my interviews, excessive cold remains a pervasive issue. In many cases, the problem stems from a poorly maintained physical plant. One person incarcerated in the Northeast told me that, in his prison, the windows had no glass, so “[w]hen it snowed, it snowed [directly into the] room.”¹⁹¹ It was like “sleep[ing] on a bench in Central Park in the wintertime.”¹⁹² Likewise, in one Midwestern prison, thanks to broken windows in the cells, people would wake up “with snow on [their] blankets.”¹⁹³ In a prison in New England, “a lot of windows were broken out” and people resorted to using newspapers to cover windows and block the draft coming in under the doors.¹⁹⁴ In one Southern prison, “the place was so old and rundown” that the “window closing mechanism . . . was just broken.”¹⁹⁵ This meant that “in January and February, the windows were literally stuck open and it was against the rules to stuff blankets or towels in the windows . . . so it was freezing in there all the time.”¹⁹⁶ In a facility on the West Coast, “you could almost see your breath in the air,” and in some cells, “you could even see ice forming on the wall.”¹⁹⁷ One person in the Northeast reported having actual “icicles . . . forming inside the room.”¹⁹⁸

Plainly, under such extremity, it is very difficult to sleep. I asked one person how he slept in the cold he was describing. His answer: “You didn’t. . . . [W]e were bundled up—we put our sweatsuits on. You put [on] your full tans [i.e, the prison-issued uniform] and you’d have your jacket on with your hat and your gloves, and you’d sit there and try to get as warm as you could.”¹⁹⁹ Others painted a similar picture: “[B]asically you’re laying there with this wind that’s blowing constantly through into your cube, and you’re just trying to stay warm. . . . Not very good [quality of sleep]. Because you’d wake up in the morning . . . frozen.”²⁰⁰

As with extreme heat, conditions of extreme cold can spark tensions among people frustrated by their inability to sleep. One person who lived in a dorm in a Southern prison described

power struggles with the older generation [who] wanted the heater on [while] the younger generation didn’t. Sometimes . . . the older generation would win [and] the heater would stay on, . . . which would make it warm in there. So some guys would open their window, but the gust from the cold is coming through the window and [now] you’re cold. . . . [and you’re] trying to convince him, “Hey, man, I’m cold, could you close the window?”,²⁰¹ which could lead to a fistfight.²⁰²

Excessive cold was frequently a problem for people in ad seg. One person explained that, in ad seg, “you don’t get [any] blankets,” even though “it was freezing”—so cold, in fact, that he “had to cut [his] mat open and . . . go to sleep inside [the] mat.”²⁰³ This move did not keep him warm enough to sleep, but “it kept [him] a lot warmer than it would have been otherwise.”²⁰⁴ For those on suicide watch, the discomfort is intensified by the fact that they are typically forced to wear nothing but tear-resistant gowns known as “suicide smocks.” While these garments are made of durable fabric, they are entirely lacking in warmth. People on suicide watch are also denied standard blankets, sheets, and mattresses (although they may receive so-called suicide blankets). The justification for these deprivations is the need to prevent people contemplating suicide from accessing anything they might use to hang themselves. But the combined impact of these conditions leaves people too cold to sleep—a counterproductive effect, given that sleep deprivation appears to exacerbate suicidality.²⁰⁵

D. Noise

Prisons are extremely noisy places. Although the nature and extent of the noise varies by housing configuration, having one’s sleep disrupted by noise is a standard part of the carceral experience. Sources vary widely. The regular nighttime soundtrack may include loud voices, music, televisions, toilets flushing, the alarm clocks of night workers, or people “screaming and hollering.”²⁰⁶ Dorms, housing as many as eighty or a hundred people, are especially loud and chaotic. Some housing units have an ethos of nighttime quiet, maintained by mutual respect (as in honor dorms) or threats of violent reprisal (as in some maximum-security cellblocks). But some pathological sources of noise are not so easily quelled. People with mental illness can be loud at all hours. Some screams may come from victims of violence. And by far the most persistent and resented source of nighttime noise comes from staff who move through housing units at regular intervals doing count or security checks. Some COs try to do their checks without waking people up. But too often, COs on the night shift conduct themselves as if completely oblivious to the fact that they are surrounded by fellow humans desperately trying to sleep.

Prisons are extremely noisy places. In the daytime, the cacophony can be deafening. Although the decibel level drops considerably once people are locked in for the night, the nighttime soundtrack is still far from peaceful. This is especially so in dorms, where as many as eighty or a hundred people may live together in one cavernous room. One person compared it to “being outside at a parade” or “on [a] construction site,” with “a lot of guys having a nightmare, screaming and hollering.”²⁰⁷ In a dorm “people are moving around, people are going to work even in the middle of the night,” and “almost never would there be a night with nobody shouting or nobody making noise.”²⁰⁸ There were times when “people would stay up all night, and they’[d] leave their radio [or TV] on kind of loud, . . . or somebody [would be] partying, drinking, and making a lot of noise, up all night with the music playing.”²⁰⁹ Although the rule—and the strong norm—is that headsets must be worn by anyone watching TV or listening to music, in practice, this rule is often ignored. If only one or two people in a dorm flout the norm, everyone’s sleep is disrupted. And even when people wear headsets, those close by may hear enough of the sound—“like a fly buzzing in your ear all night”²¹⁰—to make it hard for them to fall or stay asleep. One person mentioned sometimes using earplugs to try to cut the noise. But as he explained, “[E]arplugs are tricky in prison. It’s a little easier if you’re in a cell [and] especially fine if you’re in your own cell. But in a dorm, it’s not that comfortable to not know what’s going on around you.”²¹¹

In some dorms, the bathroom is connected to the living space, so people whose beds are nearby may be disturbed all night by others going back and forth, or by flushing toilets and running water. As one person put it, “imagine eighty-seven people going to the restroom at all different times of the night. And sometimes it’[s] just one person going to the restroom starts a chain reaction, because they hear the water flow and everything. And so if you’re by the restroom, it’s not fun.”²¹²

In more orderly dorm communities, there may be less of a problem with other people making noise at night. In one dorm setting I heard about, “there was this common respect that everyone [had] to go to work call in the morning and if you make[] noise in [the] dormitory, somebody’s going to get up and say something and tell you something.”²¹³ This ethos is especially evident in the honor dorms,²¹⁴ where people do their best to avoid disturbing others, creating an environment that can be more “peaceful” and “laid back” with “less drama.”²¹⁵ Yet even still, people invariably move around at night, especially older residents, who tend to be overrepresented in honor dorms and who may need to visit the restroom multiple times a night, creating frequent disruption.

A similar ethos of keeping noise creation to a minimum also appears to prevail in at least some high-security cellblocks housing people serving long sentences. In these units, there is a particular premium placed on displays of mutual respect, and making noise while others are sleeping is considered a strong show of disrespect.²¹⁶ In such units, people may be quieter at night “because they don’t want to have to deal with the [violent] consequences.”²¹⁷ Certainly, even in high-security facilities, extant norms will vary. And outside maximum security, no such collective enforcement is likely, especially when people may have “a release date . . . within five years” and are “more likely to be on [their] [best] behavior because [they] can see the door.”²¹⁸ If others are disruptive at night, “somebody might say something to them [and] at worst it would be a fight. . . . But more likely, it would[] be nothing. . . . People don’t want to risk it by stabbing some guy for being noisy.”²¹⁹

Viewed normatively, threats of violence against noisemakers are hardly desirable. But where no anti-disruption norm operates, a cellblock at night can be extremely loud. One person described trying to sleep in a housing unit with five galleries—a hundred men to a tier—where “individuals [would be] yelling from one gallery to [another] all night.”²²⁰ Another spoke of “people . . . banging on the door, yelling through the doors, talking to each other.”²²¹ Or “[t]here are guys who will have radio battles all night long.”²²² And in “open tiers, . . . you take all that [exposure to people’s sound] and you multiply it, so there’s probably a small window of time at night where there is no noise.”²²³ Other sources of noise may add to the din. One person described “big industrial fans” in his housing unit that “would just blow and make so much noise.”²²⁴ Another spoke of air blowing through the vents so loudly that it sounded like “a jet engine going off.”²²⁵ Some nighttime noise in prison mirrors the ordinary soundtrack of life in any setting. You might hear “different things fall[ing] in the middle of the night” (which “is gonna ring”), “other people’s alarm clocks going off,”²²⁶ “their TVs when they’[re] up at night,” or—in facilities with “stronger flushes”—people flushing their toilets.²²⁷ This is only to be expected in places where people live. But multiplied by hundreds of people, the effect is constant, pervasive interference with sleep.²²⁸

Other sources of nighttime noise reflect the deepest pathologies of American prisons. For example, even where some norm of respectful nighttime silence obtains, people with untreated mental illness may be unable to comply with or even process expectations around noise. As a result, those trying to sleep may find themselves disturbed by screams or other vocalizations coming from those unable to remain quiet. One person described being housed in her first months in prison with someone who was “schizophrenic and [who] was up all night, having . . . arguments with herself . . . like really heated arguments.” As a result, she “could not sleep.”²²⁹ And people with mental health issues need not scream or yell to disturb the sleep of those around them. Another woman spoke of how, in her dorm of 150 people, “there were the people who would get up super early and make noise . . . cook in the microwave, or get their breakfast ready, or be the first ones to shower.”²³⁰ These, as she explained it, were “people with mental illness” who “didn’t want to be bothered at 6:00 a.m. [when the lights came on]” and were rigid about “doing everything before [others woke up].”²³¹ Although this was “an issue of inconvenience” to others, people “just learn[ed] to tolerate that stuff,” because “that’s how they are.”²³² This realization may have alleviated resentment, but it did not help reduce the disturbance at a time when everyone else was trying to sleep.

Mental illness among residents is also one reason why some solitary confinement units can be so unremittingly loud. Judging from my interviews, ad seg can vary from extremely noisy to extremely quiet. Those who experienced the former reported that people would be “screaming [and] . . . kicking the doors,”²³³ or “yelling from door to door, talking to the homeboys down the hall, [and if] somebody thinks they can sing—they wanna sing.”²³⁴ “Everybody’s making noise, can’t nobody sleep.”²³⁵ If it was not the other residents, it was the doors: “[T]he [Secure Housing Unit] was always loud because the doors . . . are metal [and] they slam. Anytime anybody comes through—nurse, counselors, whatever—it’s always a bunch of noise.”²³⁶ As one person vividly described it, “There were these automated doors that would make this loud ‘ahhhh’ slam.”²³⁷ Depending on the location of your cell or who was working, “all night long, you[’re hearing] ahhhh clang, ahhhh clang, ahhhh clang.”²³⁸

Then there are the sounds—the yells, the cries, the screams—produced by victims of violence. Based on my narrow sample, it is impossible to know how frequently this occurs. I did, however, hear it enough to credit that it happens. When asked about nighttime noise, people mentioned “cellies fighting each other,”²³⁹ or “[s]omebody might be attacking somebody [or] somebody might be screaming and hollering in the shower.”²⁴⁰ One person described being awakened one night by the harrowing sounds of a man who, it transpired, was being killed in a nearby cell.²⁴¹

But by far the most persistent, intrusive, and resented source of nighttime noise is that made by staff. In every prison, whatever the housing configuration, COs are expected to come through each unit multiple times a night—as frequently as every fifteen or thirty minutes—to scan each cell and bed and see that all is well. The intended purpose of these rounds (known as “security checks”) is to prevent assaults and suicides, to break up fights, to have staff available for anyone who may need medical attention, and so on.²⁴² In addition, at certain set intervals,²⁴³ COs fan out through the facility to conduct “count,” an event occurring several times over the course of the day—and night. At these designated times, COs count every person in the facility to make sure the number of people inside corresponds to the official tally.

COs’ performance of these rituals means constantly interrupted sleep. Virtually everyone I spoke with recalled being woken up repeatedly at regular intervals by those staff members doing rounds who chose to conduct this task by banging on the doors or bars of their cells, kicking their bunks, or shining flashlights in their eyes, seeking (and sometimes loudly demanding) some indication from the sleeping person that they are still alive.²⁴⁴ People told of COs who would “deliberately pound on the door or flash the light in our face until we move,”²⁴⁵ or who might “take their stick, or . . . the chirp thing for their rounds [i.e., the wand some COs must use to electronically record the completion of mandated security checks], and bang it on the bar or bang it on the door until you move.”²⁴⁶ As one person described it,

[E]very two hours or three hours . . . the officers do security check[s]. . . . They can just kick your door, boom, boom, boom, say, “Move your legs! Move your legs! Let me see you alive!” . . . So any given day, the officers . . . will wake you up . . . . [And] if you sleep with a cover over your head or something, the police will hit your window, boom, boom, boom, “let me see something.”²⁴⁷

Often, people are woken up long before the COs reach their cells—whether because the officers “just slam the doors” when they walk through,²⁴⁸ “walk around with their walkie-talkies or radios turned all the way up,”²⁴⁹ or “have a bunch of different big, humongous keys on their belt, on a keychain that jangle[s] all the time when they come by.”²⁵⁰ In prisons not generally designed for effective sound absorption, when COs “knock on someone else’s door . . . you can hear the echoes from the whole hallway. . . . So you have to cover your ears because there’s always [constant] noise.”²⁵¹

In the dorms, in addition to flashlights in eyes, the kicking of bunks, and the shouts of “let me see you move,” several people described CO behavior reflecting seeming obliviousness to the sleeping people around them. They might “stand by your bunk and just start talking in the radio . . . extra loud,”²⁵² or “bang[] [keys] on that corner of the bed [to wake someone up], . . . [which] pretty much wakes the whole dorm up.”²⁵³ Or a pair of COs “may just come to the dormitory talking to one another. One would be on one side of the dormitory and the other one would be on the other side. And they’re screaming across people while they’re making rounds.”²⁵⁴ One person reported that, rather than counting quietly, some staff would yell out the numbers: “[yells] 12345678!”²⁵⁵ As a result, “you could never get a deep sleep there . . . because there was always noise.”²⁵⁶ And it is not only during rounds that staff on the night shift conduct themselves in ways that interfere with sleep. People also described COs in the officers’ booth laughing and joking without making any effort to keep the noise down. They would be in “their bubble, and they’d have a [fellow officer] come over, and then they [would] talk . . . so loud that you can hear them inside the dorms, you know, and they’re telling the funniest stories [and] [a]ll you hear is them laughing.”²⁵⁷

In fairness, when officers conducting security checks insist on evidence that a person is present and alive, they are only following policy. As one CO I spoke to²⁵⁸ explained,

For me to be able to do my job confidently, I have to see flesh or body movement. So . . . that requires me . . . to knock on the door and wake them up. “Hey, are you alright?” “Yeah, man, can’t you see me?” “No, your bed sheets are up. I have to be able to see you, man. It’s nothing personal. I’m just trying to do my job.”²⁵⁹

And notably, COs’ intrusive behavior on the night shift may also arise in part from their own struggles with chronic exhaustion arising from sleep deprivation.²⁶⁰ As my research has shown, COs too are often sleep deprived, and this is especially true of those on the graveyard shift (typically 10:00 p.m. to 6:00 a.m.).²⁶¹ The resulting fatigue appears to drive at least some of the behaviors that disturb prisoners’ sleep. For example, “most [COs] are constantly talking to [co-workers] to get them through the night.”²⁶²

Even still, the degree of disruption repeatedly described by my interviewees appears excessive, even accounting for institutional requirements and COs’ own exhaustion. Some considerate COs prove this point nightly by making affirmative efforts not to wake people up while doing rounds (with the proportions varying widely depending on the facility),²⁶³ thus showing it is possible to do so. These courteous COs do not yell or bang the bars or kick the bunks or shine their flashlights directly into people’s eyes. They instead “hold their keys as they’re doing count and . . . put the light up toward the ceiling”—using the ambient light to establish that people are present and breathing—“[s]o it’s not really affecting our sleep. And they kind of walk on by, . . . come by real quick, [look in], and take off”²⁶⁴ or they might “just tap real gently enough to make you move subconsciously.”²⁶⁵ Such efforts were not always successful, but they were always appreciated. At least some people reported that, when COs approached count in this way, they were able to sleep right through it. And respectful behavior by COs conducting count had a second sleep-promoting effect: when a CO makes an obvious effort not to disturb, those who are nonetheless woken up are at least spared the feelings of resentment, frustration, and anger so often provoked when COs seemed to be gratuitously inconsiderate—feelings that would in turn impede efforts to fall back asleep.²⁶⁶

E. Light

In some units in some prisons, the lights stay on 24/7. In such environments, deep, restorative sleep can be impossible. In most cases, the lights are dimmed at night, but they never fully go off, leaving it “dark, but . . . not . . . sleep dark.”²⁶⁷ Those who suffer most are people on top bunks, especially those stuck directly below the lights. But the perpetual glow makes it hard for everyone to sleep, leading many to attempt creative strategies for blocking the light. Unfortunately, most such strategies are only marginally successful and quickly run up against several pathological features of the prison environment. In this way, exploring the phenomenon of excessive nighttime light effectively illustrates the way that, in prison, the causes of sleep deprivation are never only about a single isolated factor.

Excessive light also impedes sleep in prison, although, as with noise, the effects vary depending on housing configuration. In the worst cases, people are housed in units where the lights are on full blast 24/7. This is so in some ad seg units, where “very bright lights are on twenty-four hours a day.”²⁶⁸ In general population, in some instances, prison administrators may decide that certain units are too dark at night. One person told me that, in one facility he was in, “they used to cut the lights out” at night, but then there was “a fight and [the COs] really couldn’t pinpoint who did what, . . . so the sergeant actually made us keep the lights on . . . in the dorm. . . . Full strength. All night.”²⁶⁹ Whatever the reason, when the lights stay bright all night, sleep eludes pretty much everyone.

In most housing units in most prisons, the standard practice is to dim the lights, usually an hour or so after everyone is locked in for the night. If this situation is far preferred to round-the-clock brightness, for many people it still remains too light to sleep. This is especially an issue in dorms, in which several lights stay on all night and “it [is] never anything other than like twilight in that big space”²⁷⁰—“it’s dark, but it’s not . . . sleep dark.”²⁷¹ It is worse for people with bunks close to lighted areas of the dorm, typically those near “the guard booth and bathrooms.”²⁷² Cells generally have light switches that turn off the cell lights completely. But the hallway lights, while dimmed, remain on. In at least some places, rather than all the lights being dimmed to the same degree, the lights are calibrated, with some portion remaining at full strength. It might be “half a tube that would still be on, and that could be directly in front of your cell”; understandably, those cells “are the least coveted.”²⁷³ “[E]ven if they turned off the big light, the small light was still there. And so it was never ever dark at all,”²⁷⁴ still “[t]oo bright to sleep.”²⁷⁵ Multiple people reported cellblock night lighting “bright enough to read by.”²⁷⁶ Hallway night lighting was especially a problem for those with cell bars rather than a solid door, because with this setup “[there’s] nothing filtering the light.”²⁷⁷ And even those with solid doors could find their sleep disturbed, because the “lights that stay[] on at night . . . still have a tendency of coming in through the window and shining something in your room.”²⁷⁸

In both dorms and cells, those on bottom bunks have an advantage when it comes to light. Depending on the angle of the bed, “if you’re on the bottom bunk, the top bunk would shed some of the light off,” while “if you’re on the top bunk, you get all the light.”²⁷⁹ People on bottom bunks have an additional advantage: the configuration allows them to “put a sheet under the [top bunk] mattress” to make a tent.²⁸⁰ And assuming the COs on duty are not sticklers,²⁸¹ this move may afford “some type of shade, to shade your face from the actual light.”²⁸² But putting up such tents is against the rules²⁸³ and if the COs are unwilling to look the other way, the tents must come down, letting in light even for those on the bottom bunks.

Those on top bunks suffer most from excessive nighttime light. This is especially so in the dorms, in which some have the bad luck to be in beds directly under lights that remain on all night—so-called “bad bunk[s].”²⁸⁴ People likened the experience of a bad bunk to feeling “like you’re [a] cockroach [under] the light . . . and [you] can’t get away,”²⁸⁵ or like a “rotisserie [chicken]” sleeping under “rotisserie lights.”²⁸⁶ In terms of lighting, the worst moment—mentioned by several people I spoke to—is when all the lights come on first thing in the morning, often as early as 5:00

a.m. or 6:00 a.m. Especially for those on top bunks, once that happens, further sleep becomes impossible—“like trying to sleep with the sun on your forehead.”²⁸⁷

Although the intensity of the intrusion varies, pretty much everyone must find a way to block the light in order to sleep. In the cells, people have somewhat more control over their environment and may employ hacks that involve tweaking the physical plant. When the lights themselves are easily reachable, the most common move is to make paper light covers and place them directly over the lights to cut the brightness.²⁸⁸ Sometimes these efforts could be very involved. One interviewee explained that he would use

a cardboard box—like a Ritz cracker box—you open it and flatten it out. [Y]ou would get three or four of those [boxes . . . and use] tape. Or . . . if you don’t have tape, you could get toothpaste [as an adhesive] and get paper, and that will dim [the light] out a lot. . . . I used to order art supplies . . . like construction paper [to] make a light block at night.²⁸⁹

Another person described an even more audacious approach: “In max facilities, if the light is directly in front of the cell, people have gone so far as to actually paint that side panel [of the light itself] to block out as much [light] as they can.”²⁹⁰

These strategies are not open to people in dorms, where the lights will generally be too high to cover or paint. Even if someone succeeded in getting at the lights to cover them, the COs would spot it immediately. As one person put it, his frustration apparent, “[T]here were nights [in the dorm] whe[n] I literally had to put my coat over my face, just to get some darkness.”²⁹¹ Sleeping with something covering the face appears to be the primary way people try to block the light. Various strategies are employed to this end, but none is without downsides. If you use a T-shirt or a towel, it is likely to fall off in the night. When this happens, the sudden light exposure will wake you up. As one person described it, using a towel “would block the light, but when you’re sleeping [and] moving around, the towel comes off, and you are dreaming about light and then you open your eyes to see just why [and find] it’s not a dream.”²⁹² And if one’s T-shirt or towel falls from a top bunk, there is no way to retrieve it without disturbing the occupant of the bottom bunk, who will feel the bed shifting as you get down and get back up.²⁹³ As we have seen, such maneuvers can create conflict that is best avoided. So, unless you have something else to hand, you are out of luck. If instead you try to cover your head with a blanket, you are likely to be woken up the next time a CO does rounds and insists on seeing your face: “If you slept with a sheet or a blanket [motions with hands over his face/head] . . . to cover your face, [makes knocking motion and imitates CO saying,] ‘I can’t see you. I can’t see you.’ ”²⁹⁴ Some people try to forestall this disruption by sleeping with one bare foot outside the blanket, but during the winter this approach may leave one too cold to sleep. And even should a person find an effective way to cover their face, it won’t necessarily do the trick, because you “could still feel the light.”²⁹⁵

In some facilities, enterprising craftspeople make eye masks—a homemade version of the type people get when flying first class. They may even sell eye masks to others. Using an eye mask appears to be the best way for people on top bunks to reduce the intrusive effects of the lights when they are trying to sleep. But this strategy too has its challenges. One problem is that, in many cases, the masks are made from materials taken from items issued by the prison: fabric from T-shirts, cotton from pillows, elastic from boxer shorts. In every facility, destruction of state property is a rules violation, which means that the mask you may rely on if you are to have any hope of sleeping may also be the basis for a write-up.²⁹⁶ Then there is the fact that, in prison, fully covering one’s eyes (or blocking one’s ears) while asleep can impede a person’s ability to quickly recognize danger. As one person put it,

Sleep is great, but [in prison,] what you don’t want to be is unaware. So . . . wearing earplugs or putting an eye mask on—and actually I did use an eye mask . . . made out of socks—[is not a good idea]. [P]rison’s a place where you want to be aware of your surroundings, and especially if you’re in a dorm environment.²⁹⁷

All this goes to show how, as prisons currently operate, something as simple as trying to block the light to enable sleep quickly collides with many pathological aspects of prison life, including the delicate dance of conflict avoidance, overly intrusive nighttime CO rounds, the perpetual fear of violence, and the enforcement of what prisoner-turned-criminologist John Irwin labeled “chickenshit” rules.²⁹⁸ As with all the causes of sleep deprivation explored in this Part, the sleep-compromising effects of excessive light in prison go well beyond those experienced by nonincarcerated people who may face this issue in their own lives.

IV. Findings II: Meta-Conditions Impeding Sleep in Prison

The conditions discussed in Part III are specific and concrete, of the sort typically considered appropriate targets for more conventional policy reform. But it is also possible to identify a number of what might be called “meta-conditions”²⁹⁹ that also tend to impede sleep in prison. By “meta-conditions,” I mean aspects of the carceral experience that are institutionally pervasive and highly constitutive of life inside, yet so deeply embedded in the life of the prison—so wholly naturalized—that it can be hard to recognize either their destructive impact or the institutional role in their production. In this section, I highlight the sleep-compromising effects of five such conditions: fear of violence, trauma, poverty, overly intrusive rules enforcement, and daily humiliation. As will be seen, there is some thematic overlap here with the discussion in Part III, a function of how, in prison, all aspects of the experience are fundamentally interconnected.

A. Fear of Violence

In prison, the fear of violence is endemic. It compromises sleep in many ways, most profoundly when people are forced into close quarters with others not of their choosing. Whether in a cell or a dorm, it takes time to feel comfortable enough around strangers to fall asleep. People in dorms, who live in company with scores of others, are especially at risk and thus especially apprehensive when arriving in a new environment or when new people enter their unit. But rarely is anyone—even a long-time cellmate with whom one has a good relationship—ever fully trusted, and people pretty much always sleep “with one eye open.”³⁰⁰

It might be imagined that, in prison, a perpetual fear of assault would keep people awake 24/7. This is not the case. At the same time, in all but the most functional lowest-security facilities,³⁰¹ the fear of violence is real, and undermines sleep in numerous ways.³⁰² In my interviews, I asked: Did you ever have difficulty sleeping because you were in close proximity to someone you didn’t know or trust? In response, everyone said pretty much the same thing. In double cells, people sleep little when first encountering a new cellmate; it takes time to be comfortable enough to be able to sleep in close quarters with someone who is basically a stranger. “If you’re just entering a cell with somebody new that you don’t know, if this person moved, or you heard the mattress twitch or anything—any slight movement at all—your eyes are wide open.”³⁰³ But eventually, “you kind of get comfortable with the person you’re with.”³⁰⁴ As one person described it, when assigned to a cell with “a roommate that [he] did not know,” at first all he could think was, “I do not want this guy to go crazy and try to kill me in my sleep [or] to take my stuff in my sleep.”³⁰⁵ It took “a few days to a week to really feel like okay [signals with hand lowering]; the tension is okay and I can sleep.”³⁰⁶ It is not unusual for people to spend years in the same cell with the same bunkmate. In these instances, this initial interference with sleep is brief enough not to be especially significant. But it is also not uncommon for people to be frequently transferred between cells, units or institutions, or to have a rotating series of cellmates. For those in this situation, the disruption to sleep from being in close quarters with others neither known nor trusted can be considerable.

In dorm settings, any influx of new people generates an even greater sense of insecurity.³⁰⁷ Because of the open setup, “you had to watch your back”³⁰⁸ because “most people know that when a person is asleep, that’s his most vulnerable time because he can’t see the attack coming. And there have been a lot of guys who have been attacked in their sleep.”³⁰⁹ New arrivals are thus always viewed with suspicion, and people will be hesitant to sleep deeply until they can be sure of their own safety:

You’re sleeping with one eye open and one eye closed because you don’t know who you’re in a dormitory with . . . . [I]f a new guy is coming to [the] dorm, everybody’s kind of leery of who this cat is. Everybody’s watching him [to] see how he’s moving, what he’s about. And then it doesn’t happen until you begin to talk to him. You realize, okay, he’s cool. I can go to sleep.³¹⁰

Depending on the character of the dorm, people may never feel secure enough to give in to deep sleep. Or they may band together with trusted associates and sleep in shifts, taking turns to watch over each other to ensure that all is well. I heard of this strategy from two people housed in dorms in two very different jurisdictions, one in the Northeast and one in the South. As one person who did his time in the South described it:

[I]n a dormitory setting, so you have a lot of chaos going on. So you might have four guys [who are] really close. And they’ll be like, “Man, look, we’re gonna go take a nap [so] keep an eye out on things,” [which] means, watch over us. So they do for a certain amount of time. . . [a]nd then . . . the guys [who were watching] are going to go to sleep, and [the] other guys [will] stay up.³¹¹

The fear of violence also compromises sleep in other ways. In prison, particularly in high-security facilities, people who are perceived as weak are especially vulnerable to being victimized.³¹² In dorm settings, where people must sleep out in the open, the stress of feeling oneself to be at risk from innumerable potential assailants can make it very hard to sleep. I asked people this question: Did you ever have trouble sleeping because you felt physically vulnerable or were afraid for your safety? In response, several told the same story: they had gone into prison young or physically diminutive and were only able to gain a sense of security by showing they could fight. As one person described their experience, “I went into prison at 18 years old, 185 pounds, and a very effeminate male. So . . . I was beaten up pretty badly a couple of times [but then] I learned how to very quickly overcome that by becoming the person that nobody wanted to mess with, whether I could back it up or not.”³¹³ For those who did not manage to build a reputation of this sort, the ongoing sense of vulnerability would make it very hard to sleep.

Then there are the sleep-disrupting effects of knowing that others around you may be targeted for violence. One man I interviewed lived for twelve years in a dorm setting in a Midwestern prison. Although he did not feel himself to be at risk,³¹⁴ he described being assigned a bunk adjacent to that of a “young kid” who was repeatedly sexually assaulted after lights out.³¹⁵ Even if one were not inclined to come to the defense of someone being attacked (and in prison, there are many good reasons not to intervene, including the desire to avoid becoming a target oneself), people need to be able to achieve some measure of physical relaxation in order to sleep. In environments where such violations are possible, the fear of victimization will be a perennially disruptive force, whether the potential victim is you or someone in your vicinity.

For those in two-person cells, even when you know someone well, trust in a cellmate can only ever be provisional. “In a double cell, even if you become friends with a person, there is still that unsurety of safety. So even the heaviest sleeper [is] sleeping with one eye open.”³¹⁶ This is because “no matter how high of a level of trust you may have for somebody inside prison, you don’t ever [fully] trust them. So if you hear somebody moving in the middle of the night, you’re paying attention, whether it’s consciously or subconsciously.”³¹⁷

Although most of the time, most people in prison will not be assaulted while they sleep, the fear of violence, including sexual violence, remains pervasive and never wholly disappears.³¹⁸ Living with such insecurity, in an environment where violence is an ever-present possibility, means that when people sleep, they sleep “light”: “In prison, there is no way in hell you should ever sleep [so] hard that a man can come in the cell and leave and you not know.”³¹⁹

This is especially the case for those who are gang-involved or who have affiliations with groups that may require them to engage in collective violence at a moment’s notice. When people in prison talk about “sleeping with their shoes on,” this is what they mean. For example, in the federal system, your state of origin may dictate who you “run with” and thus on whose behalf you may be expected to fight. As one person who spent years in a federal penitentiary explained, for those in this situation, there is no sleeping until the cell doors lock for the night: “[Y]our homeboys, everybody’s on duty. You gotta keep your boots on and just sit there until the CO says it’s time to go lock the doors.”³²⁰ And in the morning, the moment “the doors pop, you have to be awake because somebody might run in the cell [and attack you] because something happened in another unit with somebody from the same state you’re [from].”³²¹ Those committed to this arrangement learn not to let themselves sleep deeply so they can always be ready to act quickly if need be.

Those who lack such commitment but who are nonetheless expected to participate in collective violence can lose sleep for different reasons: the stress of being unsure whether to respond when the moment comes, and the fear of retaliatory violence if they opt not to join in. One of my interviewees described exactly this conundrum:

I had some friends in gangs. . . . They would get into situations where it’s like, “Yo, we’re about to move on this dude. . . . [Y]ou a Blood just like we Blood[s] [so] you gotta move with us.” And I see my friends sleep uncomfortable. Because it’s like, man, I’m going to school. I got something good going on. I’m trying to leave that [and] change my life. But if I don’t move with them, . . . I become one of the enemies. There [are] so many different factors . . . to this life in [prison] that sleep is not something that comes by easy.³²²

As this person explained it, given all the stresses of life in prison—including the pressure from the gangs and the dangers people may face if they try to remain unaffiliated—“If you’re able to sleep in [prison], you’re almost looked at as [an] extra-terrestrial. You must be [an] alien. No seriously, something is wrong with you if you’re able to sleep [inside].”³²³

B. Trauma

People in prison are exposed to countless traumatic events, whether experiencing them directly or as witnesses to them. In my interviews, people described seeing people stabbed or beaten to death, flayed open, burned alive, or committing violent acts of self-harm. Experiences like these can leave people terrified for their own safety and may also interfere with sleep as people struggle to process psychologically and emotionally what they have witnessed. Not everyone has this reaction to witnessing senseless violence; many reported becoming inured to the brutality around them. This desensitization may serve a protective function, allowing people to survive impossible circumstances—and perhaps to sleep a little more easily. At the same time, it reveals a core mechanism by which the carceral experience alienates people from their own humanity.

Judging from my interviews, the experience of trauma further compromises the quality and quantity of sleep that people get inside. This dynamic, moreover, is self-reinforcing, as those who sleep poorly after traumatic experiences appear to be more likely to develop post-traumatic stress disorder (“PTSD”) and other mental health conditions³²⁴—conditions that, among other unwelcome effects, may in turn impede sleep. True, poor sleep is a diagnostic component of many mental health disorders,³²⁵ sometimes raising the question of which came first. But in the case of PTSD, the evidence suggests that sleep disruption itself—in particular, disrupted REM sleep—may provoke the development of PTSD following trauma exposure.³²⁶ This possibility is especially concerning given that both persistent trauma exposure and systematically disrupted sleep are hallmark features of the prison experience.

In my interviews, I heard countless stories from people who witnessed deeply disturbing events while incarcerated. These included seeing “somebody get beat to the point where they’re laying in a puddle of blood and they have to be carried away because they can’t walk”;³²⁷ seeing “a kid get slammed on his neck and, then you see him getting a white sheet over his body and he’s getting carted off the unit. Or somebody leaking, somebody’s white T-shirt is now red”;³²⁸ seeing “one man that was literally cut open from the bottom of his chin to the top of his hips, cut and flayed wide open because they were looking for something they assume that he had swallowed”;³²⁹ and seeing “another guy running down the tier with his eyeball bouncing off of his cheek.”³³⁰ At some point, trying to process the weight of all these accounts, I began asking some version of this question: Did you ever find it hard to sleep because you were traumatized by things that you had seen?

For those who responded in the affirmative, the answers suggested two ways that witnessing traumatic events can interfere with sleep. First, there was the sleep-disrupting effect of the fear engendered by exposure to terrible violence. Seeing the awful things that happened to others made it hard to sleep, “knowing that any given time somebody can choose me as their victim.”³³¹ This effect is of a piece with the fear of violence previously explored. But there was also a second theme that emerged in the answers: the sleep-disruptive effects of the psychological or emotional processing triggered by exposure to trauma, as when “you just keep seeing that same image over and over and over.”³³² One person, recounting a particularly vivid and distressing event, effectively captured how the brain’s need to process can make it hard to sleep:

[S]ee[ing] guys have . . . padlocks put on their doors, flammables thrown in on [them], and one guy had been burnt alive. I witnessed this man, that’s gone from someone 6’1”, 6’2” . . . [to] not even eighteen inches long, you know, charcoal, just burnt alive. . . . [W]hen you’re seeing all of these things . . . [s]leep is hard. . . . [Y]ou’re trying to sleep, but you have all these things going on in your head, and any little sound that you hear, you[‘re] coming out of this sleep.³³³

Not everyone connected the trauma they experienced to difficulty sleeping. But it is hard to imagine that the quality of one’s sleep would remain unaffected in an environment where such exposure is common.

At the same time, many people reported that the distressing things they saw while they were in prison did not affect their ability to sleep. Over time, they instead became inured to the violence and abuse to which they were exposed:

[While in prison] I saw a lot of violence. But it became, when you see it, you just see and you don’t see, you just sort of get out of the way . . . [and] you keep it moving. Prison teaches you to literally stay in your lane. . . . If it doesn’t affect you, oh well.³³⁴

Whatever the example may be—seeing somebody cut, seeing somebody stabbed, seeing an officer beat on somebody, break their arm, or whatever it may be, you get so desensitized that this is natural.³³⁵

Eventually you get numb to all the stuff that happens.³³⁶

It may be that these individuals remained genuinely unaffected emotionally by what they saw and experienced. But the terms they used to describe their state of mind, which included becoming “numb” and “desensitized,” suggested their recognition of the defense mechanisms they were relying on to get through. To experience a state of equanimity in such a traumatizing environment, the level of denial would have to be considerable. If one’s sleep is thereby protected, it may also indicate an alienation from one’s own moral core, a feature of prison’s dehumanizing effect that carries its own psychic and emotional costs.

C. Poverty

Having money in prison does not guarantee a good night’s sleep. But it helps. There are innumerable items that those with money can procure, many directly from the commissary, that may enhance sleep inside: fans to drown out noise, extra clothes for cold nights, high-quality eye masks, softer sheets, and warmer blankets, not to mention sleep-inducing drugs available on the prison’s black market. And those with money can stock up on commissary items, ensuring they will not go to bed hungry. We have already seen that people who lack family support or are otherwise unable to earn what they need may opt to engage in survival strategies that put them in danger, including stealing, gambling, or selling sex or drugs. The official practice of paying incarcerated workers nominal wages—in those jurisdictions that pay them at all—can leave the least well-off people in prison forced to choose between forgoing items that might improve sleep or adopting strategies that could put them at risk of serious physical or sexual harm.

Most people enter prison without financial means.³³⁷ Although most institutions require prisoners to work, the hourly wage—when paid at all³³⁸—is a pittance: on average, between $0.14 and $0.63 an hour.³³⁹ Depending on the prison, some people may secure positions with the private industries that operate behind the walls. But even in such cases, although the hourly wage may be considerably higher than the standard prison wage, it will remain far lower than the federal minimum wage³⁴⁰—typically somewhere between $0.33 and $1.41 an hour.³⁴¹ Some fortunate few can make even more: one person I spoke to trained call center agents for a private company. He reported making $285 per month (roughly $1.80 an hour), in contrast to the kitchen workers in his facility, who made around $30 per month (roughly $0.19 an hour).³⁴² But these relatively well-paid opportunities are rare,³⁴³ and as we have seen,³⁴⁴ if people are to have access to money while in prison, they must either get it from loved ones outside³⁴⁵ or find off-the-books ways to earn it.

Although everyone in prison, regardless of financial status, faces serious obstacles to getting adequate sleep, access to money can help improve one’s odds. Many people mentioned buying fans from the commissary, “the whole point [of which is] to drown out the noise from the tier.”³⁴⁶ As one person explained, echoing an experience shared by many, “I could not sleep without my fan. I pretty much set it up right by my head . . . to help drown out the other ambient noise . . . outside the cell.”³⁴⁷ But fans that do the trick do not come cheap in prison—the most durable of them can cost up to $40.³⁴⁸ Given the extremely low wages paid to incarcerated workers, a fan could cost almost a month’s earnings.³⁴⁹ And if a person still chooses to buy one—which they well might, if it means the difference between sleeping and not sleeping—they would have little left over for other necessities like food or basic hygiene products.³⁵⁰

Money, or the lack thereof, impacts sleep in other ways. Those with resources to spare can buy extra clothes—sweatshirts, warm socks, etc.—to help keep themselves warm on cold nights.³⁵¹ They can buy softer sheets³⁵² and warmer blankets. They can buy marijuana, heroin, Seroquel, or other sleep-inducing substances on the prison’s black market,³⁵³ as several of my interviewees reported doing.³⁵⁴ If they are not themselves skilled with a needle, they can pay someone to make them a top-shelf eye mask.³⁵⁵ They can buy a second fan to cool their bodies on hot nights. And if they use a CPAP machine, assuming power strips are allowed, they can buy one³⁵⁶ to guarantee themselves access to a plug without risking getting into a fight over control of the wall sockets with a cellmate or others in the dorm. These advantages may seem minor. But for people in prison, they may mean the difference between eking out a few solid hours of sleep or being consigned to sleepwalking through another day.

Still, when it comes to being able to sleep, the primary reason to want money or its equivalent is to ensure you have ready access to food. As we have seen, “if you don’t have the money, you can’t eat, and . . . you can’t

sleep if you’re hungry.”³⁵⁷ As one person explained, if he didn’t have much money on his books,

I might not have been able to buy as much canteen [as I needed], so I would go through it. . . . When you’re hungry it’s hard to budget it and make it through. Like one night . . . I’m trying to go to sleep and I’d be hungry so that I would go in and eat some [food items that had been] budgeted for later on [in] the week. But I just I had to eat it then. . . . Eventually, though, when I started getting more money, [hunger interfering with sleep] wasn’t a factor.³⁵⁸

The lengths to which people will go to get money for food are a testament to the urgency of the need. One person I spoke to, who had no family support, “started gambling and stealing while [she] was in there to provide for [herself],” because “[o]nce you[’re] full, you can sleep better.”³⁵⁹ It is hard to overstate the desperation that would drive a person in prison to engage in either of these strategies. Someone caught stealing may earn a violent reprisal from the victim of the theft.³⁶⁰ And gambling is equally risky. It is a sure-fire way to get into debt, and in prison, the penalty for nonpayment of debts is often physical violence or forced sexual servitude.³⁶¹

Another risky, last-resort strategy is “two-for-one,” the prison loan-sharking practice whereby someone with a load of commissary will provide you with something you want—say, a candy bar, a bag of chips, or a soup. The catch is that, when the next day for canteen rolls around, you have to pay back double what you borrowed—two candy bars, two bags of chips, or two soups. If you can’t, the quantity owed doubles again. This is another way people in prison can get into debt, which, again, can result in violence or pressure for sex. “[Y]ou have dudes who are wealthier than others who max out every canteen . . . , and they sell stuff to guys. ‘Well I’ll give you one for two back.’ ”³⁶² Or “[t]hey won’t tell you it’s two-for-one. [T]hey’ll say, ‘Here’s a box of [cookies].’ Then . . . [the] next week you owe double, . . . they’d say, ‘[W]ell, we could exchange it for sexual favors.’ That’s how they were trapping these young kids coming in.”³⁶³ According to my interviews, the desire to avoid going to sleep hungry is a major reason why people risk putting themselves in this position. Note that at least some of this felt pressure to access commissary, which can drive people to act in ways that could put them in serious danger, would be alleviated if prisons simply made decent food available to everyone in the evenings before lights-out.

These dynamics mean that those without money—the prison’s poorest, who lack outside support or the ability to accrue funds while inside—experience appreciably worse conditions than those with means. To some, this situation may seem unproblematic, just how the world is. But as the example of sleep makes clear, what is at stake is not simply differential access to luxury items. Sleep is a basic human need, and insufficient sleep causes considerable harm to physiological, psychological, emotional, and cognitive health. To the extent that poverty leaves some people less able to mitigate the sleep-disrupting aspects of the prison environment, poor people are experiencing a harsher, more destructive punishment than that experienced by those with access to funds.

D. Overly Intrusive Rules Enforcement

In prison, virtually every aspect of a person’s life is governed by rules enforced by COs. Many of the strategies people adopt to try to improve the quality of their sleep run afoul of these rules. This situation leaves people forced to choose between breaking the rules—and possibly getting written up, but perhaps eking out slightly improved sleep—or following the rules and forgoing small comforts (an extra mattress, a light cover, some food held back from a tray to eat before lights out) that might help them sleep. The stress involved in making this choice itself compromises sleep, especially for those who opt to roll the dice and break the rules. Compounding the problem is the fact that penalties for minor rule violations of the sort at issue here often include loss of commissary or yard privileges, thus depriving people of access to food and exercise—two pathways for improved sleep available to those in prison.

Pretty much every aspect of a prisoner’s life is governed by rules.³⁶⁴ Enforcement authority belongs to line officers, who, when they witness a rules violation, may issue a ticket (also known as a “shot”). The most extreme violations, which generally involve physical violence, may earn the perpetrator delayed release and even additional charges. In cases involving less serious offenses, penalties are less severe but still deeply undesirable. They may include time in ad seg, lockdown in one’s own cell, or temporary loss of yard time, commissary, phone privileges, or even shower access. As with law enforcement on the outside, COs have considerable discretion when enforcing prison regulations. As one person put it, discussing light covers, “[M]aybe when you go to sleep at night, the CO doesn’t care [so] they let you do it. But then the next shift comes on, and then they’ll bang on the door, [and] you’ll take that down.”³⁶⁵ If the violation involves possession of contraband—which generally means any item not directly issued by the prison or purchased directly from the commissary—a CO may choose only to confiscate the offending item, or they may confiscate it and issue a write-up, potentially exposing the offender to some form of penalty.

Many rules enforced through this system serve valid institutional purposes; few people would quarrel with prison prohibitions on physical violence or the distribution of narcotics. But innumerable other “chickenshit rules” penalize conduct that is less obviously necessary to maintaining institutional order.³⁶⁶ As Irwin has observed, “[M]any of [the facility’s] rules intrude into prisoners’ ordinary practices and significantly interfere with their attempts to carry on their already excessively reduced life routines.”³⁶⁷ In other words, people inside may find themselves at risk of punishment simply for trying to make intolerable conditions slightly more bearable.

This effect is certainly evident in the context of sleep. One person I spoke to clearly captured this dilemma. As he explained, “[I]f the light was right in front of my window, I would cover it up. And then that’s usually when the CO would bang on the window, . . . but then I would just take it down and put it right back up. . . . I was basically being forced to break the rules to be able to sleep.”³⁶⁸ The use of light covers is only one example of a

standard strategy for improving sleep that violates prison rules and can expose people to sanctions should they get caught.³⁶⁹ Among other things, it is generally against the rules to:

Drape a sheet from the top bunk to curtain off one’s bottom bunk (a.k.a. make a “tent”);³⁷⁰
Cover the vents³⁷¹ or windows;³⁷²
Shower outside designated hours (especially on cellblocks);³⁷³
Possess an extra mattress;³⁷⁴
Possess an extra blanket or sheet;³⁷⁵
Destroy state property (including the shorts, T-shirts, or stuffing from mattresses or pillows used for making eye masks);³⁷⁶
Remove food from the dining hall;³⁷⁷
Store food from the dining hall in one’s cell;
Hold back food from trays.

Rules thus prohibit precisely the strategies that many people use to try to improve the quality and quantity of their sleep in the face of excessive light, excessive heat or cold, uncomfortable beds or inadequate bedding, and insufficient food. To compound the problem, the penalties for such violations may include loss of commissary privileges and loss of yard time. Yet, as we have seen, without access to food, people will often be forced to go to bed hungry, a state that considerably impedes sleep. And for many people, vigorous exercise is a way to “exhaust themselves, [so it’s] that much easier to go to sleep at night.”³⁷⁸ Without access to the yard, this strategy is more difficult to operationalize, and sleep becomes that much harder.³⁷⁹

The perverse structure of this arrangement tells us everything we need to know about the moral orientation of the carceral system toward those inside. First, the prison operates in innumerable ways to undermine prisoners’ ability to sleep. Then, COs police the housing units to ferret out any items beyond each person’s allotment—items they may have expended considerable effort to assemble to try to improve their sleeping environment and perhaps increase their chances of getting a little more sleep. These items are then confiscated, a move that strips the possessor of the benefits, however slight, the seized items may have offered. COs are then empowered to impose penalties likely only to compromise still further a person’s ability to sleep. It is tempting to call this system Kafkaesque, except that, for Kafka, the bureaucracies that outrage and dehumanize are so infuriating in part because their procedures have no moral valence and are simply manifesting their own internal imperatives. In the prison, the cruelty of the process just sketched is of a piece with the callous indifference with which the institution and its COs seem to regard the daily hardships faced by those in their custody.

For those on the receiving end, COs’ interference with their efforts to improve the sleep they get may seem spiteful and even sadistic. And in some instances, this impression may be apt. It does, however, bear noting that COs’ decisions as to how to exercise their considerable discretion vis-à-vis rule violations may also be shaped to some degree by the institutions’ own pathological impact on the COs themselves—including, ironically, the chronic sleep deprivation that is a standard comorbidity experienced by those in the role.³⁸⁰ When COs are sleep deprived—a condition almost certainly exacerbated by the host of other comorbidities to which COs are disproportionately prone, including depression, alcohol overuse, PTSD, and suicidality³⁸¹—they become less able to treat the incarcerated with consideration and sympathy.³⁸² To the extent that prisoners’ sleep is compromised by exposure to rules unevenly and unpredictably enforced by people who at times seem incapable of basic human sympathy and understanding, the resulting frustration may be less a product of individual spite than the broader operational logic of a system that is blind to the humanity of all parties, incarcerated and COs alike.

Yet however incapacitated COs may be by the hours they must keep and the conditions in which they work, it is the prisoners who are forced to daily navigate a challenging context in which any efforts they make to alleviate conditions of extreme discomfort are liable at any moment to be nixed by those COs who for whatever reason are not inclined to look the other way.³⁸³ As a consequence, not only do people in prison expend considerable energy strategizing how they might improve their sleeping environment, but a good part of this effort turns out to involve a constant process of weighing the benefit to be gained from a rules violation against the odds of getting caught and incurring a penalty. As I wrote in my field notes after one interview, “even if you weren’t going to get written up, even if the staff weren’t going to give you a ticket for an extra blanket or an extra mattress or whatever, there’s always the stress of knowing they could do it.”³⁸⁴ In this situation, “you were always trading the ability to be slightly more relaxed and not able to sleep”—because, in a bid to reduce stress, you opted not to avail yourself of the strategies that might help make sleep possible—against “put[ting] yourself in a position where you were stressed about whether you were going get into trouble just to improve [your] sleep a little bit.”³⁸⁵

The stress this situation occasions can be both considerable and relentless. For example, as one interviewee reported, there were days when, as frequently happens, the people on his tier were not able to shower:

And you don’t want to go to sleep sweaty and dirty [because] that really impedes the ability to sleep. [So] you would have to find a way to sneak out of your cell to get a shower just so you would be able to sleep. So how would you do that? Well, if there was a time when tier 1 was supposed to be on the floor and the mezzanine, [and] tier 2 is supposed to be in their cells, you might sneak out of your cell even when you were supposed to be in it. Or you would try to get them to open your cell door for some reason and then grab a shower.

In other words, “it was hard to sleep when you weren’t clean and [there was] also stress involved in trying to get a shower when there is no time available for you to do it.”³⁸⁶

Note that the goals driving this stressful negotiation of prison rules—sleeping, eating, keeping warm, showering—are at once basic requisites of human functioning and integral to maintaining one’s own humanity and sense of self. And, of course, the frustration occasioned by the enforcement of what are widely experienced as petty rules is also likely to interfere with the ability to get restful sleep.

E. Daily Humiliations

Every day, people in prison receive multiple reminders that the system regards them as undeserving of consideration and respect. At night, as we have seen, staff members conduct count and security rounds often without seeming to care whether they wake up the people sleeping around them. And similar demonstrations of callous disregard persist all day long: the incarcerated find their health problems minimized, their reasonable requests peremptorily denied, and their valid grievances ignored by prison officials at all levels. These routine humiliations understandably leave people frustrated, humiliated, resentful, and outraged. These feelings do not simply dissipate once an interaction is over. They persist and fester, making it even harder for people already coping with countless sleep-impeding obstacles to fall or stay asleep at night.

On a daily basis, people in prison are treated in myriad ways as if they are morally worthless. One manifestation of this attitude recurs nightly, when many COs appear to make no effort to do their rounds quietly. Instead, many behave as if the beds around them are empty, as if they are not surrounded by human beings who are trying to sleep—as if, as one person put it, “[T]hey don’t see you [or] your head right by the pillow that [they]’re standing next to.”³⁸⁷

Nighttime noise by staff doing rounds wakes people up, thereby directly interfering with prisoners’ sleep. But there is a second, more subtle sleep-disrupting effect of this intrusive conduct: it forces the incarcerated to perpetually confront the disdain in which they are held, an experience that can spark feelings of helplessness, frustration, and rage that themselves impede sleep. Imagine how hard it would be to fall asleep if, every night, thoughts like these were swirling around in your head:

So on top of already fighting to get some type of sleep, those little couple of hours that you may get, then it’s interrupted by [the COs] . . . . [A] lot of them . . . have no type of empathy at all. . . . [Y]ou’re asleep [and] they don’t care. You’re an animal. This is the[ir] house [so] you got to do what they say, when they say, how they say.³⁸⁸

[F]rom the officer’s perspective, they were . . . told like, “[H]ey, look, your job, and the way this job is supposed to be done, is more important than any of these inmates’ sleep.” W]e were . . . pieces of trash. We were manipulators, and we weren’t to be trusted so who cared what we thought and . . . that’s what [we] get for being in prison

It’s a power issue. You ask them to keep it down or something like that, they’re gonna look at you like, . . . ‘[W]ho the hell are you? Yeah. Right.’ And they’ll spite you [by being] even louder. . . . ³⁸⁹

As one person I spoke to observed, “[S]leep is more than just laying down and resting your body. [To] sleep, your mind needs rest as well—being able to actually go to sleep and not have to worry about anything, [to] just be comfortable in that moment and let your body and . . . mind rest.”³⁹⁰ Far from promoting a restful mind, daily humiliation seems far more likely to

achieve the opposite. Compounding this effect is the way those in this position must swallow these feelings—however justified—because to express them would be to court retaliation that among other hardships would only guarantee still further sleep disruption.³⁹¹

The ethos of contempt and callous disregard reflected in the behavior of many COs on the night shift also manifests in multiple ways during the day. Staff ignore the needs of those in their custody, refuse their reasonable requests, and target them for abusive treatment, whether out of malevolence, caprice, or poor judgment sparked by their own intense sleep deprivation.³⁹² Medical personnel dismiss prisoners’ symptoms, deny requests for x-rays or specialist access, and pass them off with ibuprofen and pamphlets. Access to loved ones, both by phone and in person, is highly restricted, often for no reason other than official caprice. Life is constantly governed by “chickenshit rules,” which can seem designed more to demean than to address genuine institutional needs.³⁹³ Grievance procedures are generally byzantine, and complaints, however meritorious, are almost always denied.³⁹⁴ When disciplinary infractions require a hearing, these hearings often feel like shams, with those officials serving as adjudicators virtually always siding with the CO who wrote the ticket.³⁹⁵

The relentlessness of these humiliations can take a deep emotional toll, leaving people feeling frustrated, resentful, disrespected, and outraged. Such feelings are not conducive to sleep, to say the least. In my interviews, I asked people what they thought would have to change to enable people in prison to get adequate sleep. Multiple people spoke about the dehumanizing treatment that they experienced daily, the constant reminders of their own powerlessness, and the pleasure staff seemed to take in ignoring their needs. One interviewee described what this treatment looks like in real time:

[For example,] I gotta go to school. Now I gotta ring the doorbell on the cell one thousand times for this individual to let me out. When you coming back from school [you might] want to use the restroom, but the officer got me out here waiting forty-five minutes. When we’re going to yard, we[’re] delay[ed] forty minutes and now when we get to the yard, instead of getting an hour and a half, we get thirty minutes in the yard. . . . It’s just little components of them treating us like [we]’re not human beings.³⁹⁶

Or consider this account of a cell search:

When [COs] come in to search your cell, they don’t just nicely pick things up and put them back down. It looks like a hurricane ran through your room. They literally flip your bed over, your pillow is on the ground, your sheets are on the ground, your stuff gets mixed in with your bunkie’s stuff. Your clothes are all over the floor, your food’s mixed everywhere. Some [COs] will take the jelly and just squeeze it on your bed for no reason. . . . [T]hey come in and they destroy. That’s what they do.³⁹⁷

When this is your daily experience, when nighttime rolls around, the state of relaxation requisite for sleep remains elusive. One woman I spoke to was especially eloquent on this point. What she said is worth quoting at length:

[I]f we’re constantly having these bad interactions, these bad experiences, and the culture isn’t conducive, or it always . . . undermine[s] you and . . . makes you less than and makes you feel worthless and all these other things—you’re not going to sleep like oh, I conquered the world today. No, you’re going to sleep feeling like shit, feeling even worse than you’ve felt. . . . I go to the freaking clinic to get medication for . . . a toothache and I get treated like a fucking drug addict. And [I] don’t even get any Motrin, so I’m not going to sleep feeling too good. . . . And then when you’re constantly looking at me and talking down to me as if I’m less than you, [as] if I’m not a person, . . . all those things don’t make for a good night’s sleep. You don’t feel good at the end of the night. . . . [A]nd not only that, but the setting in general . . . doesn’t make it conducive for family reunification. They make it even harder for you to have relationships [and] more so [for] women. [T]he women’s prison, geographically where it’s placed, is in the middle of nowhere. So all the people that often are coming into the prison settings live so freaking far away. . . . [T]hese inner cities are so far away from this rural fucking place in [location of prison] in the middle of nowhere, that it makes it hard for you to have visits, to have a constant contact with your family or the . . . how much is it? Fifty cents a minute a phone call? When we’re making one dollar a day on our freakin’ wages? Like, seriously? So all these things just work to produce even more stressors and, as we all know, when I’m super stressed throughout the day, all that’s happening at the end of the night when it comes time for me to sleep is these things [are] replaying in my head . . . and making me feel even more incompetent and worthless.³⁹⁸

In short, sleep will not come easily to people who are constantly humiliated.

Like those conditions canvassed in Part III, the conditions discussed in this Part contribute significantly to the inability of people in prison to get adequate sleep. If there is a difference to draw, it lies in the pervasiveness of these “meta-conditions,” the extent to which each is woven into the carceral project. While the more concrete and specific conditions described in Part III appear tied to policy decisions concerning prison administration and operations, the conditions described in this Part seem to arise directly from the culture and ethos of the American prison. Yet, as far as moral valence goes, the distinction I have drawn between the more concrete sleep-impeding conditions and what I have labelled “meta-conditions” may not much signify. Whatever the moral orientation of individual COs, all the conditions that systematically compromise the potential for sleep in prison reflect a system shaped at its core by a callous indifference to the well-being of the incarcerated, a blindness to their humanity, and a refusal to accord those in custody the basic consideration and respect owed to any fellow human. This moral injury is part of the daily experience of imprisonment. And as the foregoing has shown, among the innumerable practical manifestations of this systematic institutional disdain is the persistent interference with the fundamental human need for sleep.

V. Implications: Prisons, Punishment, Law, Policy

This Part begins to explore the normative implications of chronic sleep deprivation in prison. Section A considers, as a descriptive matter, what this phenomenon adds to our understanding of the prison experience. Section B flags several ways the inevitability of chronic sleep deprivation deepens the punitive character of the carceral penalty. Section C offers a first cut at the core constitutional question—whether the conditions mapped in this Article violate the Eighth Amendment’s prohibition on cruel and unusual punishment. It also touches on a question that frequently arises when I discuss this work: whether the conditions chronicled here constitute torture. Finally, Section D offers some thoughts on the prospects for policy change and the challenges that await efforts to tangibly improve sleep in prison. For reasons of space, my contributions on these topics will necessarily be brief. The aim here is twofold: to begin mapping the normative implications of the problem, and to offer preliminary thoughts on the four dimensions identified here in the hope of sparking a broader conversation.

A. Prisons: Sleep Deprivation and Carceral Life

Prior to the interviews, many of my subjects had given little thought to their experiences of sleep inside, or to how those experiences shaped their time in prison.³⁹⁹ Yet what I learned during those conversations brought into focus certain basic realities of prison life that have, until now, been largely unrecognized, not only among many of my interviewees, but also among those who study incarceration. Below, in no particular order, I identify four such implications.⁴⁰⁰

First, chronic sleep deprivation is a constitutive feature of life in prison. Regularly getting insufficient sleep—night after night, for months, years, and even decades—is as central to the daily experience of incarceration as lousy food, crushing boredom, grossly inadequate medical and mental health care, solitary confinement, and the perpetual fear of physical or sexual violence from staff or fellow prisoners. Being sleep deprived is an intrinsic part of what it means to be inside.

True, people in prison sometimes manage to get decent sleep. In some cases, housing configuration makes the difference. Virtually everyone I spoke to agreed that people in single cells sleep much better than anyone else.⁴⁰¹ As one person memorably put it, “[T]he best cellie is no cellie.”⁴⁰² In other instances, people can benefit from their own wise choices. One person I spoke to spent thirty-seven years in prison. He reported that, especially in later years, he was able to get sufficient sleep and wake feeling rested. He credited several aspects of the life he built in prison for this atypical experience, including staying out of prison politics (“not trying to be king of the jail”), making peace with not being able to control the behavior of loved ones on the outside (not “trying to live one foot in prison and one foot in the world”), and building a supportive religious community of fellow Orthodox Muslims who bunked together and provided one another with “comfort and security.”⁴⁰³ He was also an accomplished artist, which compounded his sense of purpose and seemed to enhance the personal equanimity that helped him sleep. Yet he also described multi-year stretches over his long incarceration during which he slept far less.⁴⁰⁴ And even in single cells—the equivalent of temporarily winning the prison sleep lottery⁴⁰⁵—people struggle with noise, light, extremes of heat and cold, and many other obstacles to sleep chronicled here. In other words, in prison, even the outliers only sleep so well.⁴⁰⁶

In any case, the fact that some people may sometimes beat the odds scarcely goes to controvert the core finding that sleep deprivation is endemic in prison. To take a parallel example, even though some people may sometimes receive reasonably acceptable and even effective medical or mental health treatment while incarcerated, there is no question that a systematic failure to provide adequate health care constitutes a definitive feature of the prison experience. A central claim of this Article is that chronic sleep deprivation, produced by persistent, systematic interference with sleep, is likewise intrinsic to the carceral experience—even if some people in prison occasionally manage to sleep reasonably well.

Second, at any given time, most people living in prison will be operating from a sleep deficit, possibly a considerable one, and this situation will have profound effects on what happens inside. In future work, I will take up the question of how sleep deprivation in prison impacts interpersonal dynamics and institutional functioning. For now, it is enough to say that, judging from my interviews (and as we would expect), the fact that virtually everyone in prison is chronically tired greatly heightens institutional tension and volatility, along with the likelihood of interpersonal conflict. And because COs too are frequently sleep deprived, the heightened potential for conflict and all that goes with it also implicates COs—who, it bears noting, are legally authorized to use force. At this point, it is hard to know just how much the instability and threat of violence that shapes life in the contemporary American prison is attributable to the poor judgment and short tempers emblematic of fatigue. Yet I am certain that the impact is far greater than heretofore recognized.

Third, the overwhelming weight of the sleep science, combined with the striking unidirectionality of the findings, strongly suggests that, in the aggregate, people in prison will suffer considerable physiological and psychological harm traceable to the long-term systematic interference with their ability to get adequate sleep.⁴⁰⁷ Again, at this stage, it is impossible to say with any precision what shape this harm takes.⁴⁰⁸ However, in light of what is known about the health-compromising effects of even short periods of sleep deprivation,⁴⁰⁹ it is hard to imagine that no such effects occur. In addition, the health-compromising effects of sleep deprivation suggest several other troubling possibilities, each of which merits serious consideration:

The inability to sleep properly in prison suggests a vicious circle as to the health of the incarcerated: thanks to chronic sleep deprivation, people inside are vulnerable to illness and disease that well-rested people might avoid, and when they get sick, the constant sleep deprivation and disruption that define the carceral experience are likely to undermine the body’s natural healing processes.
Sleep deprivation may help to explain why people age faster in prison. Research suggests that the biological age of incarcerated individuals is as much as ten to fifteen years greater than their chronological age.⁴¹⁰ This effect is traceable in part to “the high prevalence of risk factors for poor health” among the incarcerated, including “a history of substance abuse, head trauma, poor health care, and low educational attainment and socioeconomic status.”⁴¹¹ Yet given the breadth of the harms generated by insufficient sleep, it seems hard to imagine that chronic sleep deprivation does not also play a central role in the accelerated aging process prisoners experience.
If, as some studies suggest, sleep deprivation constitutes a risk factor for—or exacerbates the symptoms of—mental illness, the systematic interference with sleep that people routinely experience inside could help to explain the strikingly high incidence of mental illness among those in custody.⁴¹² These psychopathological effects mean that those concerned with the epidemic of mental illness behind bars should pay particular attention to the quality and quantity of the sleep people get inside.

Fourth and finally, the causes of sleep deprivation in prison are inextricably bound up with the normative design and operation of the modern American prison. Every person alive, whether inside prison or out, knows what it is like to try to sleep in the face of obstacles like noise, excessive light, extremes of temperature, and so on. For this reason, it may at first seem that people struggling to sleep in prison are on a continuum with others across society who fight for sleep in unconducive situations. But if sleep deprivation is not unique to prisoners, the impediments to sleep behind bars have a particular normative cast, reflecting the distinctive animus towards the incarcerated that shapes virtually all aspects of prison life.⁴¹³

For example: Prison beds are not fiercely uncomfortable by happenstance. People are forced to sleep on metal slabs with thin, cracked, moldy, sticky mattress pads because those who decide what form the beds will take view the resulting discomfort as appropriate for people serving time as criminal punishment and would begrudge as undeserved the cost of anything more substantial. People sleep on flimsy, creaky bunkbeds for the same reason, and fear moving in their sleep in case they provoke frustrated, angry bunkmates who, like themselves, are forced to live in moral ecosystems where they are routinely humiliated and dehumanized and who may lash out at small indignities simply to feel some small measure of control. Every night, people are woken up by COs who, “hav[ing] zero respect for prisoners,”⁴¹⁴ make no effort to do their rounds quietly. During the height of summer and the depths of winter, people labor to sleep in intense heat or extreme cold, a hardship that could be largely addressed were state corrections agencies to invest in air conditioning and routine winter maintenance in the housing units where prisoners sleep, as they already do as a matter of course in the parts of the prison frequented by staff. And so on. Obstacles to sleep that may at first seem of a piece with what people routinely experience outside prison turn out on further examination to constitute specific, sleep-compromising manifestations of the callous indifference and even hostility towards the incarcerated that in countless ways shape the American carceral system.

B. Punishment: Sleep Deprivation and the Carceral Penalty

Recognizing the fact of sleep deprivation in prison and the shape it takes also deepens our understanding of the nature of the punishment inflicted on people sentenced to prison time. In this section, I identify four related dimensions of one key insight: the way the inability to get adequate sleep heightens the punitive character of the carceral penalty.

First, it is already well understood that in practice, a prison sentence carries with it a raft of noxious conditions beyond simply the deprivation of liberty. This Article strongly suggests that, among these conditions, we must include the experience of being chronically sleep deprived for the duration of the stipulated term. Not merely a side effect of being in prison, chronic sleep deprivation is a constitutive feature of the punishment itself, a part of the penalty imposed when the judge pronounces sentence.

Second, the punishment incarceration represents also includes being subjected to the harms that arise from chronic sleep deprivation. To put the point more concretely, when we sentence people to prison time, we are sentencing them to insufficient sleep, which is very likely to (1) trigger a degradation of their body’s proper functioning and defenses against illness and disease,⁴¹⁵ (2) expose them to an increased risk of early mortality,⁴¹⁶ and (3) ignite psychological harms, activating or exacerbating the symptoms of a wide range of psychopathologies.⁴¹⁷ We are also sentencing them to perpetually live with the cognitive deficits that daily plague people who are poorly slept—the irritability, ready frustration, fuzzy-headedness, impaired judgment, proneness to conflict, and general incapacity that collectively make it difficult to get through the day, much less make good decisions, achieve one’s goals, or build healthy relationships. To the extent that sleep deprivation is an in-built feature of the prison experience, all these harmful effects become part of the punishment prison time represents.

Third, the punitive character of the carceral penalty is intensified by the effects of pervasive sleep deprivation on the prison’s social ecosystem. Although sleep-impeding conditions are experienced individually, they are simultaneously endured by everyone. In other words, people in prison are routinely forced into close quarters with scores and perhaps hundreds of others who are themselves struggling with the psychological challenges and cognitive deficits produced by chronically inadequate sleep. People in prison are already likely to be bored, frustrated, resentful, and sometimes distraught, a situation that helps to explain the constant undercurrent of tension, conflict, and volatility that defines the carceral environment. These dynamics will only be exacerbated when everyone is denied access to restorative sleep.⁴¹⁸ Being forced to live in this environment, with all its stress, instability, and incipient danger,⁴¹⁹ is yet another essential component of the carceral penalty traceable to collective sleep deprivation—part of what the punishment of prison time entails.⁴²⁰

Fourth, being subjected to chronic sleep deprivation is dehumanizing, which adds a further dimension to the punitive character of prison. Dehumanization is the process of “stripping people of human qualities” so that “they are no longer viewed as persons with feelings, hopes[,] and concerns but as sub-human objects.”⁴²¹ The mechanisms of dehumanization experienced by people in prison do not only lead others to deny their moral worth. They also turn the process inward, undermining a person’s most elemental capacities, including the internal resources necessary for self-reflection, reasoned judgment, moral fortitude, and personal growth.

Sleep deprivation is not generally recognized as among the core dehumanizing aspects of the carceral experience. But it should be. As we all know from personal experience, when we do not get enough sleep, our most fundamental faculties are impaired. It can be hard to think, to reason, and to read situations and other people. Even things that might otherwise feel urgent or meaningful can seem flat or insufficiently important to motivate action.⁴²² These effects are regularly experienced by people in prison, where the stakes may be especially high. Those who are sleep deprived may react with irritation or hostility even to those they love the most—including the friends and family on the outside who are their lifeline.⁴²³ Like people everywhere, when they are tired, people in prison make poor choices. They opt not to participate in activities they would otherwise value⁴²⁴—even those that might improve their chances of successful reentry.⁴²⁵ Thanks to diminished self-restraint, they sabotage their prospects.⁴²⁶

In short—again, like people everywhere—prisoners who are chronically tired may find themselves sleepwalking through life. Obviously, by virtue of being incarcerated, the human potential and personal agency of people in prison are necessarily drastically curtailed. Yet it is also true that people make meaning in custody. Within the constraints of their incarceration, to a greater or lesser degree, people can be agents of their own situation. But the less sleep they get—and the more they are thereby deprived of the restorative effects of adequate sleep—the less equipped they will be to engage in meaningful, productive interactions,⁴²⁷ and the more readily they will be dragged into the dark, negative, hostile space that so frequently traps those lacking the resources to escape it.⁴²⁸ In all these ways, being sleep-deprived severs people from central aspects of their own moral characters and those features that define their humanity. This severing too is a part of the punishment traceable to—and inflicted by—systematic sleep deprivation in prison.

C. Law: Is Systematic Interference with Sleep Unconstitutional? Is it Torture?

The conditions described in Parts III and IV raise two obvious legal questions: (1) Do these conditions violate prisoners’ constitutional rights? and (2) Do they amount to torture under international human rights law? This Section considers these questions in turn. It focuses primarily on the constitutional question and the likely prospects for successful constitutional claims in the courts, before briefly touching on the matter of torture, which frequently arises in conversations about this research.

The primary provision for assessing the constitutionality of prison conditions is the Eighth Amendment prohibition on “cruel and unusual punishments.”⁴²⁹ At its most basic, the question is whether plaintiffs challenging the conditions catalogued here could satisfy current Eighth Amendment standards—and the answer, I argue, is yes. To see why requires a short overview of the governing law.

Under existing doctrine, to prevail on a prison conditions challenge, plaintiffs must satisfy two components, styled by the Supreme Court as “objective” and “subjective.”⁴³⁰ The objective component considers the challenged treatment, asking whether “the deprivation [was] sufficiently serious.”⁴³¹ The subjective component focuses on the responsible officials, asking whether they acted “with a sufficiently culpable state of mind.”⁴³² Of the two, the objective component may require somewhat more doctrinal explication to make the case. However, as currently defined,⁴³³ neither component should pose insurmountable obstacles for plaintiffs challenging the conditions that impede prisoners’ sleep.

As the Court made clear in Wilson v. Seiter, the objective component requires a showing that the challenged conditions “have a mutually enforcing effect that produces the deprivation of a single, identifiable human need.”⁴³⁴ Writing for the Wilson majority, Justice Scalia identified “food, warmth, or exercise” as examples of “human need[s]” that might ground such claims.⁴³⁵ A complaint, he observed, might allege “a low cell temperature at night combined with a failure to issue blankets,” thus implicating multiple “mutually enforcing” conditions depriving the plaintiffs of the “single, identifiable human need” for warmth.⁴³⁶

Sleep deprivation seems a textbook case of the sort of objective component Wilson explicitly approved. It is beyond peradventure that sleep is a basic human need, as intrinsic to human survival as “food, warmth [and] exercise.”⁴³⁷ Even should plaintiffs choose to focus their complaint only on the more concrete factors explored in Part III—uncomfortable beds, hunger, extreme heat and cold, noise, and excessive light—these conditions together make getting adequate sleep close to impossible and would thus seem to readily satisfy the Wilson standard.⁴³⁸

Yet incarcerated plaintiffs will not typically claim the complete denial of sleep but rather the inability, over months and years, to get even close to the recommended seven hours per night. And as Justice Powell observed in Rhodes v. Chapman, “the Constitution does not mandate comfortable prisons.”⁴³⁹ This means that, when incarcerated plaintiffs bring constitutional claims grounded in sleep deprivation, there is a need to determine the point at which insufficient access to adequate sleep ceases being simply the denial of those “creature comforts” of the sort incarceration necessarily entails, and instead qualifies as the “deprivation of a single, identifiable human need” that satisfies the objective component of Eighth Amendment conditions challenges.⁴⁴⁰

In Garrett v. Lumpkin, an ongoing case out of Texas, the district court offered one answer.⁴⁴¹ It held that “because Garrett failed to show that his sleep deprivation—which was undisputed—actually caused his health issues, he had not satisfied the objective element of cruel and unusual punishment.”⁴⁴² Seeing why this answer fails as a matter of basic Eighth Amendment doctrine helps make clear why sleep deprivation need not be total to satisfy this hurdle. When Michael Garrett filed his claim, he had been in prison for more than thirty years. At the time he drafted his complaint, he was living in the McConnell Unit and was later transferred to the Estelle Unit.⁴⁴³ In his complaint, Garrett challenged a set of conditions that, he alleged, interfered with both the quantity and quality of his sleep and left him chronically sleep deprived. As to quantity, the schedule operating in the Estelle Unit afforded residents only three and a half hours of sleep a night, with bedtime set for 10:30 p.m. and breakfast starting “around 2:00 a.m.”⁴⁴⁴ (The hours in the McConnell Unit were only slightly better, with bedtime at 10:30 p.m. and breakfast commencing around 2:30 a.m.⁴⁴⁵) And “even during this three and a half hour window, sleep is not continuous,” because residents must “be awake for a 1:00 a.m. bed-book count,” which meant that “the most continuous sleep Garrett can theoretically receive is two and a half hours” (assuming he falls asleep instantly at 10:30 and sleeps undisturbed until the bed-book count).⁴⁴⁶ As to sleep quality, the “nighttime prison conditions—namely, the hallway lighting, heavy doors slamming, and prisoners yelling—further imperil” the sleep residents are able to get during the brief period the prison’s compressed schedule allows.⁴⁴⁷

In his complaint, Garrett challenged this raft of conditions on Eighth Amendment grounds and sought an injunction “that would mandate a prison schedule with six hours per night designated for sleep.”⁴⁴⁸ Yet the district court found that Garrett had failed to make out “the objective element of the Eighth Amendment inquiry.”⁴⁴⁹ The reason? Having failed to present “ ‘any expert testimony establishing that a lack of sufficient uninterrupted sleep has, within reasonable medical probability,’ caused his health conditions,” Garrett could not establish a cause-and-effect relationship between his sleep schedule and any medical complaint.”⁴⁵⁰

But this notion—that a plaintiff in Garrett’s position can satisfy the objective component of an Eighth Amendment conditions claim only by showing actual harm—is directly at odds with governing doctrine, specifically the Supreme Court’s clear holding in Helling v. McKinney. Helling was decided just two years after Wilson. While in prison in Nevada, William McKinney was “assigned to a . . . [cellmate] who smoked five packs of cigarettes a day.”⁴⁵¹ McKinney filed suit, arguing that exposure to this degree of cigarette smoke would unconstitutionally “jeopardiz[e] his health.”⁴⁵² At trial, the magistrate judge granted the state’s motion for a directed verdict in part on the ground that McKinney “had failed to present evidence showing . . . medical problems that were traceable to [exposure to] cigarette smoke.”⁴⁵³

Helling wound up in the Supreme Court. The state, following the magistrate, argued that “unless McKinney can prove that he is currently suffering serious medical problems caused by exposure to [environmental tobacco smoke (“ETS”)], there can be no violation of the Eighth Amendment.”⁴⁵⁴ The Court, however, rejected this suggestion,⁴⁵⁵ with its implication that “prison authorities . . . may ignore a condition of confinement that is sure or very likely to cause serious illness and needless suffering the next week or month or year.”⁴⁵⁶ As Justice White put it, writing for the majority, “[A] remedy for unsafe conditions need not await a tragic event.”⁴⁵⁷ Instead, the Court held, it is enough to show that McKinney was exposed “to levels of ETS that pose an unreasonable risk of serious damage to his future health.”⁴⁵⁸ The following year, the Court decided Farmer v. Brennan, which phrased the general inquiry slightly differently: whether the plaintiffs faced a “substantial risk of serious harm.”⁴⁵⁹ In the three decades since Farmer was decided, it is this latter formulation that has governed.

When Garrett came up on appeal, the Fifth Circuit made short work of the district court’s reasoning.⁴⁶⁰ In her opinion on behalf of a unanimous panel,⁴⁶¹ Judge Clement drew directly on Helling, which declared it “ ‘cruel and unusual punishment to hold convicted criminals in unsafe conditions,’ regardless of whether those conditions actually cause injury.”⁴⁶² Because “[i]t would be odd to deny an injunction to inmates who plainly proved an unsafe, life-threatening condition in their prison on the ground that nothing yet had happened to them,” the Fifth Circuit had previously held that, to satisfy the objective component of an Eighth Amendment conditions claim, the plaintiffs “need not show that death or serious injury has already occurred.”⁴⁶³ Instead—in keeping with Farmer—they “need only show that there is a substantial risk of serious harm.”⁴⁶⁴ In short, when sleep in prison is sufficiently restricted as to expose people to a substantial risk of serious harm, plaintiffs will have satisfied the objective component, whether or not the danger has yet manifested and even if they still manage to get some sleep each night.⁴⁶⁵

What of the subjective component? Here too, were a court to faithfully follow governing precedent, there is a clear pathway for plaintiffs alleging unconstitutional sleep deprivation to meet their burden. The requisite state of mind for Eighth Amendment conditions claims is “deliberate indifference.”⁴⁶⁶ In Farmer v. Brennan,⁴⁶⁷ the Court held that to satisfy this standard—and thus the subjective component of an Eighth Amendment claim—plaintiffs must show that defendants “kn[ew] of and disregard[ed] an excessive risk to inmate health or safety.”⁴⁶⁸ As the Farmer court explained, Eighth Amendment deliberate indifference is equivalent to criminal recklessness as defined by the Model Penal Code: “[T]o act recklessly . . . a person must ‘consciously disregar[d]’ a substantial risk of serious harm.”⁴⁶⁹

To satisfy this standard, plaintiffs bringing an Eighth Amendment sleep deprivation claim must show that prison officials subjectively realized the substantial possibility that, because of the challenged conditions, those incarcerated in their facility were unable to get adequate sleep. As always, plaintiffs must take care to name defendants who will have some knowledge of the challenged conditions. But given the pervasiveness of the conditions described here and the many operational decisions that daily compromise the quality of prisoners’ sleep—the nightly counts, the double-bunking, the crowded dorms, the early dinner hour, the lack of air conditioning, and so on—it would not be hard to identify prison officials with full knowledge of the conditions that deprive the incarcerated of the basic human need for sleep.

Defendants in such cases might try to argue that, although they knew of the challenged conditions, they were not aware that these conditions posed “a substantial risk of serious harm” to prisoners. Yet as the Farmer Court made clear, “[w]hether a prison official had the requisite knowledge of a substantial risk is a question of fact subject to demonstration in the usual ways, including . . . the very fact that the risk was obvious.”⁴⁷⁰ At a minimum, prison officials who work the night shift are well acquainted with the conditions under which those living in their facilities are forced to try to sleep. And because they are human themselves, and thus equally in need of the full complement of sleep necessary for human functioning (not to mention equally likely to have wrestled with environmental conditions impeding sleep), it would be reasonable to infer that they realized that people subjected to this set of conditions would be unable to get adequate sleep. Given the way the Court defined the Eighth Amendment deliberate indifference standard in Farmer, the plaintiffs should be readily able to identify defendants as to whom they can satisfy the subjective component of an Eighth Amendment sleep deprivation claim.

I am aware that, in this discussion, I have glossed over innumerable issues that plaintiffs may run up against in litigating these two aspects of an Eighth Amendment conditions challenge. For example, defendants might argue that plaintiffs failed to exhaust administrative remedies.⁴⁷¹ They might invoke qualified immunity⁴⁷² or “Turner deference,”⁴⁷³ or deny that plaintiffs suffered actual injury.⁴⁷⁴ They might claim relief is foreclosed absent a showing of physical injury.⁴⁷⁵ Or they might claim—as prison officials repeatedly did during COVID—that, despite taking no meaningful steps to reduce the risk of harm, they reasonably responded to the risk prisoners faced from lack of sleep and could thus not be found deliberately indifferent.⁴⁷⁶ At least as to some challenged conditions—hunger in particular comes to mind—they might insist that they lacked knowledge of the condition itself and thus of the risk of harm it might have posed. Litigators will need to address these issues as they arise.⁴⁷⁷ To some extent, their force will depend on the particulars of individual cases. But my point here is a more general one: assuming a well-litigated case, so long as courts apply the governing Eighth Amendment doctrine in an evenhanded and straightforward way, plaintiffs alleging unconstitutional sleep deprivation should have a clear doctrinal pathway to prevailing on the merits.

All this means that in theory—assuming the ability to overcome the many procedural hurdles to getting into court and getting a hearing on the merits—an affected class that meets its burden under governing Eighth Amendment standards should be entitled to injunctive relief designed to remedy the many sleep-impeding conditions producing harm or risk of harm. However, in practice, those seeking relief in the courts from the conditions undermining sleep are still likely to face a steep uphill climb.

One notable challenge would arise from the nature of the relevant evidence. On the one hand, the considerable body of sleep science demonstrating unequivocally the myriad harms—physiological, psychological, and cognitive—arising from sleep deprivation will help incarcerated plaintiffs make their case.⁴⁷⁸ On the other hand, making such a showing demands expertise and resources that pro se plaintiffs are likely to lack, which means that, to succeed, plaintiffs will require outside counsel with deep pockets and a commitment to building the strongest possible case.⁴⁷⁹ And even still, such efforts may run up against what Margo Schlanger has characterized as “a general hardening of attitudes about causation” on the federal bench, producing demands by courts for “more rigorous proof on harm and causation.”⁴⁸⁰

Then there is the fact that today’s federal courts are extremely reluctant to order meaningful changes to how prisons are run. Except for one brief period in the 1970s and 1980s,⁴⁸¹ federal courts have been notoriously unwilling to issue broad injunctive orders in prison cases.⁴⁸² And as we have seen, the concern with sleep deprivation implicates wide swaths of prison operations—the design of housing units, maintenance of the physical plant, a prison’s ability to accommodate large populations in relatively small spaces, the way staff shifts are organized, how COs do their jobs, and even the fundamental matter of how staff perceive and treat the incarcerated. In today’s judicial climate,⁴⁸³ the breadth of any sufficient remedy is likely to incline courts to try to find some way to dispose of the case long before the remedial stage—however strongly supported the constitutional claim.⁴⁸⁴

For these and other reasons, litigation will not be a cure-all.⁴⁸⁵ Yet even granting the impediments, these cases are still very much worth bringing. For one thing, with the right case, dedicated and adept advocates with sufficient resources to mount the litigation may well find a way to win.⁴⁸⁶ And partial successes are still worth pursuing. Even if a court were unwilling to order sweeping relief encompassing all the issues explored here, it would still make for better sleep inside if they were at least to require, say, adequate air conditioning during hot summer months or that prisoners be provided real beds with higher quality mattresses. To this, some may object that, given the cruelty of forcing people to try to sleep in the dehumanizing and inhumane conditions mapped here, we ought to accept nothing less than complete institutional overhaul.⁴⁸⁷ But when conditions are this dire, advocates do prisoners no favors by letting the best be the enemy of the good.

Moreover, we should not discount the possible virtuous circle that might arise were courts to order—and prison officials to implement—even partial relief. Of the obstacles to sleep catalogued here, those I have labeled meta-conditions are perhaps least likely to form the basis for injunctive relief. Yet if a carceral facility took steps to improve the more concrete conditions impeding sleep, some of the meta-conditions may at least partly take care of themselves. If a prison provided decent beds, served a meal before bed, reduced the frequency of nighttime counts, or ensured an ambient temperature conducive to sleep, these efforts might mitigate the daily humiliations that can themselves compromise sleep. Depending on the nature of the policy changes, people in such facilities might also feel less compelled to break the rules to try to improve the quality of their sleep, or to engage in risky behavior for the sake of having something to eat before bed. This possibility of felicitous knock-on effects is another reason Eighth Amendment claims grounded in sleep deprivation are worth bringing, even if the results may fall short of total institutional transformation.

Sleep deprivation may also ground claims beyond class actions seeking broad structural change. As with other pathological features of the carceral experience—medical neglect, inadequate mental health care, the risk of physical or sexual assault, excessive force, and so on—sleep deprivation is both a shared experience produced by macro-level institutional design and a personal experience the precise shape of which will vary according to the individual and their context. Depending on the circumstances, people in custody may face particular obstacles to adequate sleep for which individual prison officials may be liable. For example, prison officials may persistently refuse to provide testing and treatment for people exhibiting obvious signs of sleep apnea, or to authorize a cell move for someone whose cellmate is a kitchen worker with a 2:00 a.m. wake-up time, or to provide a longer or wider bunk for someone of greater than average height or weight. Such situations too deprive people of the basic human need for sleep and thus raise potential constitutional claims, along with claims under other legal frameworks—perhaps most notably the Rehabilitation Act and the Americans with Disabilities Act, the “two principal federal disability anti-discrimination statutes.”⁴⁸⁸

Again, I do not wish to overstate the likelihood of success. Given the dispositional favoritism federal courts routinely exhibit toward defendant prison officials in prison law cases, incarcerated plaintiffs will always face long odds.⁴⁸⁹ Yet advocates should still bring these claims—and not only because they may sometimes win. They should also do so because, just by filing cases, they can put the issue of perpetual sleep deprivation in prison on the public radar.⁴⁹⁰ If change is to be possible, the public first needs to be aware of what goes on behind the walls. And litigation, with its built-in drama and familiar script, offers a way to expose the realities of prison life to public view. Legal cases can become the focus of media attention; as reporters cover the various stages of the case, they also educate the public about the underlying issues. Judicial proceedings offer inflection points for organizing, around which advocates can build political campaigns. And lawsuits themselves can open pathways for sympathetic engagement by legislators, who may in turn hold hearings and push for policy change.⁴⁹¹ Eighth Amendment litigation alone cannot achieve these aims. It can, however, play a crucial supporting role in a multi-pronged, multi-stakeholder strategy.⁴⁹²

One other legal issue, implicating not the United States Constitution but international human rights law, bears addressing here. That is, some readers may be inclined to regard the systematic interference with sleep in American prisons as tantamount to torture. Whether such a claim could succeed under international law is an open question. There is a high bar to a finding of torture under the United Nations Convention Against Torture (“CAT”),⁴⁹³ which explicitly excludes from the definition “pain or suffering arising only from, inherent in or incidental to lawful sanctions”—a provision that would greatly complicate efforts to apply CAT to sleep deprivation in American prisons (if not for cases focusing on pretrial or administrative detention).⁴⁹⁴ For this and other reasons,⁴⁹⁵ advocates wanting to argue that the conditions mapped here violate international law bans on torture may also face a steep uphill climb.⁴⁹⁶

Those pursuing this legal pathway might find it more promising to try framing sleep deprivation in prison, not as torture, but as “cruel, inhuman or degrading treatment” (“CIDT”)—a category of treatment also prohibited under international human rights law.⁴⁹⁷ CIDT falls somewhat below torture on the schedule of human rights abuses, and includes “forms of punishments . . . that cause ‘serious mental and physical suffering,’ ”⁴⁹⁸ which persistent interference with sleep of the sort described here arguably does.

I leave it to experts in international human rights law to determine how best to categorize the conditions described here within CAT and other international human rights frameworks. Here, I would simply note that, at this stage, how this doctrinal debate (torture or only CIDT?) would come out under international law may matter less⁴⁹⁹ than the fact that, for many, this is the question that most immediately comes to mind when confronting the fact of systematic sleep deprivation in American prisons. That we find ourselves in a moral universe where these are the stakes provides the strongest possible evidence that the conditions identified here are among the features of American prison life most worthy of condemnation and challenge.

D. Policy: Challenges and Complications

The conditions chronicled in this Article are the product of multiple failures: of institutional design, of moral obligation, of constitutional protection and enforcement. In these ways, the phenomenon of sleep deprivation is no different from many other toxicities constitutive of the American carceral experience. All the unconscionable conditions that shape prison life, sleep deprivation included, have the same moral driver: an inability (or unwillingness) to recognize the humanity of the incarcerated, and a consequent callous indifference to their health and well-being. They also have considerable mutual interaction effects, each reinforcing and amplifying the harms prisoners endure.

This interconnection means that none of prison’s noxious aspects can be fully resolved while the others persist. Yet, however much we might wish it, we will not witness the wholesale transformation of American carceral practice any time soon. Of course, we ought still to push for sweeping change. But this hard truth means that, for now, the best we may realistically be able to achieve for those currently living behind bars is to ameliorate their suffering as much as possible.

When it comes to improving the conditions in which prisoners sleep, there are some obvious policy fixes. To name just a few: people should have real beds, decent mattresses and pillows, and sufficient blankets to keep them warm at night. Meals of reasonable quality and quantity should be served at appropriate times, and food should be available in the evening for people who are still hungry. Facilities should be air-conditioned in summer, well-heated in winter, and generally well maintained to ensure a temperate environment. Nighttime counts should be reduced and staff trained to respect prisoners’ sleep. Approaches to lighting should be found that do not interfere with sleep and leave some people feeling like “rotisserie chicken[s].”⁵⁰⁰ And ideally, people would have their own rooms, so that their sleep will not be disrupted by the movements and noises of a cellmate⁵⁰¹—or ninety-nine other dormmates. Even if policymakers balk at implementing the policy reforms just itemized, these are changes advocates should push for on every available front.⁵⁰²

No doubt, these proposals and others in a similar vein would be condemned by abolitionists (and other advocates of sweeping change) as woefully inadequate. But given the actual distribution of power in this space, what is of more practical moment is that such proposals are also likely to be fiercely resisted by prison officials and policymakers. Such changes would not come cheap, and prisoners, largely viewed as undeserving across society at large, are not a constituency in which legislators are generally eager to invest. Although the desire of politicians to seem “tough on crime” has waned somewhat since its peak in the 1990s, there is still a palpable reluctance among policymakers to address even the most pressing needs of the incarcerated.⁵⁰³ This is not the place to fully address these troubling dynamics. Here, I would simply note that the state’s decision to incarcerate carries with it the duty to provide for the basic needs of people in custody. This is the state’s carceral burden.⁵⁰⁴ If efforts at economy produce conditions that deprive people of what Justice Powell once labeled “the minimal civilized measure of life’s necessities,”⁵⁰⁵ the state is obliged to reorient its priorities. If it finds the costs too great to bear, the appropriate response—both morally and constitutionally—is not to subject the incarcerated to dehumanizing and affirmatively harmful conditions, but instead to reduce the carceral footprint.⁵⁰⁶

It is well beyond the scope of this Article to fully explore the policy challenges facing those seeking to ameliorate the sleep-disrupting conditions prisoners face. I will, however, flag one additional complication that advocates will invariably need to confront: efforts to improve the conditions enabling sleep in prison will at times run up against policy initiatives intended to promote unrelated dimensions of prisoners’ interests and needs. For example, as we have seen, one central source of nighttime disruption comes from COs conducting count and performing regular security checks. But in the view of many prisoners’ advocates, these rounds are vital to ensuring the health and well-being of those in custody, by preventing or disrupting suicide attempts and intra-prisoner violence and allowing timely intervention in the event of medical crises.⁵⁰⁷ In my view, this justification is questionable.⁵⁰⁸ Yet even assuming its validity, it is also the case that these nighttime checks exact a considerable cost to the health and well-being of the very people they are supposed to help—a cost measured in the disrupted sleep, every single night, of hundreds of thousands of people incarcerated in locked facilities around the country. It may be that nighttime rounds should

be maintained, but at the very least, their considerable downside effects on sleep should be taken seriously and addressed.⁵⁰⁹

A similar tension will also confront efforts to improve sleep in prison by increasing access to single cells. Judging from my interviews, it is impossible to overstate the sleep-enhancing effects of having one’s own room. Given the choice, virtually everyone I spoke with would opt for a single.⁵¹⁰ Yet since the 1980s, when the American prison population exploded, double-celling—that is, housing two people in cells designed for one—has been the norm.⁵¹¹ This practice allowed states to cram many more people into each facility, and solo housing became relatively rare. But although the size of the incarcerated population remains high, it peaked nationwide in the early 2000s and continues to exhibit appreciable declines.⁵¹² Among other advantages, this drop has somewhat eased pressures on housing and, in some jurisdictions at least, could allow for a greater number of people to be housed alone. However, for some advocates, the preferred response when jurisdictions reduce their numbers is not to reduce population density but instead to close facilities, thus shrinking the carceral capacity of the system. If overall decarceration were the only goal, seeking prison closures would make sense. Yet viewed through the lens of sleep, shuttering prisons carries a significant downside for those who remain behind: the missed opportunity to expand the number of people able to live—and sleep—alone. This too is a serious tension advocates must navigate when considering how best to fight for change in this arena.

This brief policy discussion has focused primarily on challenges likely to attend efforts to make concrete changes to the conditions affecting prisoners’ sleep. But as I have repeatedly emphasized, the phenomenon of sleep deprivation in prison cannot be fully understood independently of the moral character of the carceral enterprise, specifically, the normative hostility and callous indifference that drive institutional choices and shape official behavior. If those with the power to shape the carceral environment are blind to or unmoved by the humanity of people in custody, they will easily dismiss demands for the concrete changes that would help those inside get adequate sleep. Any strategy for achieving practical change on the ground must therefore include efforts to generate a broad cultural recognition of the shared humanity of the people society chooses to incarcerate.

Conclusion

It may be tempting for some to dismiss concerns with the quality or quantity of the sleep people get in custody as simply part of the loss of those so-called creature comforts that necessarily comes with a carceral penalty.⁵¹³ But this notion misapprehends the stakes. Getting adequate sleep is not a matter of personal indulgence or a luxury only to be enjoyed by those at liberty to chart their own course. It is a basic human need, as fundamental to human survival and adequate human functioning as is access to food, water, and shelter from the elements.

For those who live in prisons, being denied the ability to sleep night after night takes a substantial toll. Most immediately, the experience is frustrating, infuriating, and even humiliating. Chronically insufficient sleep also inflicts substantial physiological, psychological, and cognitive harm. Yet when aspects of the American prison are singled out for condemnation, sleep deprivation is virtually never mentioned. In this Article, I have argued that chronic sleep deprivation should be counted as among the fundamental harms of incarceration, explicitly acknowledged as a core part of what makes prison so unlivable. And the claim might be put still more forcefully: the persistent inability of people in custody to get adequate sleep is not only a product of some of the most damaging and degrading features of the prison experience, but also in turn helps to produce them. Not to reckon with this endemic aspect of prison life is to miss a key driver of the toxicity of the prison environment.

Appendix A

Research Methodology

The findings presented in this Article are based on thirty-nine interviews conducted primarily over four months in 2023⁵¹⁴ with people formerly incarcerated in prisons across the United States. These interviews were of two types. Thirty-five were semi-structured interviews conducted on Zoom using a 200-item questionnaire developed for the purpose. These interviews were video-recorded after obtaining consent. Each research subject had spent at least four years in prison,⁵¹⁵ and all but three had been released within five years of the interview.⁵¹⁶ I established these parameters to ensure clearer memories and a depth of experience with the issues.⁵¹⁷

The remaining four interviews—which I came to label “ad hoc interviews”—were more free-form. Participants in these interviews were people I connected with who seemed to have insight and experience to contribute, but who did not meet the study parameters at the time or who otherwise seemed unsuited for formal enrollment.⁵¹⁸ These interviews were audio-recorded after obtaining consent. To get context and perspective, I asked these individuals many of the same questions as appeared in the questionnaire, and their answers were entirely consistent with what I was hearing in the formal interviews. As a result, I wound up drawing on all thirty-nine interviews when analyzing the data.

The primary focus of the research, as well as the main source of the data, was the thirty-five formal interviews conducted using the 200-item questionnaire. What follows is a description of the protocol I followed in undertaking this core aspect of the research.

I began by compiling a draft questionnaire, initially shaped by many conversations with people with firsthand carceral experience as well as by my own knowledge of prison conditions accrued over years of study. I then enlisted three volunteers, each of whom had spent time behind bars (two in California and one in Louisiana), to do individual mock interviews based on my draft questionnaire. After each mock interview, the volunteer shared feedback on which aspects warranted revision. I also incorporated feedback from Joe Doherty, former Director of the Empirical Research Group at the UCLA School of Law, who advised me throughout this project on matters of research methodology.

After receiving IRB approval, I began recruiting participants. I used two main channels. First, I put out a call for participants through the networks I have developed as part of my long-standing research into prisons and prison conditions. Targeted recipients included the Prisoners’ Rights listserv, a national list populated primarily by advocates for the incarcerated. This call briefly described the study and the format of the interviews and asked recipients to share the information with anyone they knew who might fit the parameters. I also posted a notice on Twitter,⁵¹⁹ advertising the study and inviting participants. In each notice, I included a link to a Microsoft form asking for basic demographic data and contact information. I also indicated that study participants would receive a $30 Amazon gift card. The idea here was to provide some compensation for the time people spent doing the interview, while keeping the amount low enough that it would be unlikely to draw people who would not otherwise be inclined to share personal details or revisit prior experiences likely to have been traumatic. Unfortunately, I did not anticipate that the up-front promise of a gift card might incline some people without prior carceral experience to pretend to have been incarcerated in an effort to receive this benefit. The first few iterations of the sign-up form proved to be a mix of genuinely qualifying individuals and fraudsters, and it took some time to figure out how to sort them out. Eventually, I began scheduling “pre-calls” with those who filled out the form. During these brief calls, I described the study and then explained that I had been hearing from people who were only pretending to be formerly incarcerated and that, before we could schedule an interview, I needed to hear enough to confirm that the person was in fact formerly incarcerated. In each instance, it took only a few sentences of the person recounting their carceral history⁵²⁰ for me to know they were in earnest.⁵²¹

Once I got through this rough patch, it did not take long to receive expressions of interest from more potential participants than I had time to interview. My aim was to build a research sample diverse as to race and roughly proportionate as to gender.⁵²² The reason for seeking a diverse sample was to ensure that my findings would not reflect only the experiences of one group of people. I also sought a sample that had broad jurisdictional/geographic representation. I did so because, as is common knowledge among people who study prisons or who work in the field, carceral systems vary considerably across the country on a variety of metrics, including facility size, general conditions of confinement, and—most importantly—moral orientation toward the health, safety, and well-being of those in custody.

I started by interviewing people in the order in which they reached out, but as my sample took shape, I began to pick and choose among those who had filled out the initial form, seeking people who would broaden the sample on the metrics I sought. Some states were better represented than others in my participant pool, and at some point, I realized that a certain degree of jurisdictional depth would enhance the data. I therefore set out to interview four or five people from four states well-represented in my pool: California, Louisiana, New Jersey, and New York. This approach helped strengthen confidence in my findings, as there proved to be considerable overlap in what I was hearing regarding the conditions impacting sleep from these four very different prison systems.⁵²³ I then did my best to speak to people from as many other state systems as I could to build out the picture more broadly.⁵²⁴

Each interview proceeded as follows:

First, prior to beginning to record, I briefly described the study and what types of questions the interview would cover. I explained that the person was free to decline to answer any question they would prefer to skip over.⁵²⁵ I also explained that if, at the end of the interview, they decided they did not want their data to be used in the study after all, they would be given an opportunity to retroactively withdraw their consent. Finally, I explained that the content of our exchange would be, not only confidential, but also fully anonymous. To that end, with the permission of each participant, I changed the name in their Zoom box to the random code I had previously selected for their interview, each taking the form FI.# (with FI standing for “Formerly Incarcerated”). Doing so ensured that even the person who transcribed the interview would not know the subject’s name. And once the interviews were transcribed, the (anonymized) transcript would be the only record of the interview. I then solicited consent to participate in the study and for the session to be recorded. Only after consent was granted on both fronts did we start the interview.

The interviews were semi-structured. On this approach, the questionnaire becomes the vehicle for introducing issues, any one of which may invite deeper inquiry depending on how the conversation goes. Each interview began with the interviewee providing their carceral history, including which facilities they were in, how long they spent in each, and the housing configuration of each. After asking them to estimate how much time they slept in each housing context, I asked one series of questions about issues that may have impaired sleep, and another about how if at all being chronically tired (or being surrounded by other chronically tired people) may have impacted their carceral experience or the functioning of the prison. A final set of questions addressed demographics and related issues.

Over the course of the research, I began to get a feel for which questions were most fruitful, and which yielded little of interest. Consequently, I stopped asking some questions altogether.⁵²⁶ I also started asking additional questions as salient issues emerged. This meant that, as to some issues—those that only emerged over time—I received input from fewer people than on those issues I had been asking about from the beginning. Among the questions that emerged in this way were:

Did you ever have trouble sleeping due to the presence of insects, rodents, or other vermin in your cell or dorm? If so, tell me about it.
Did you ever find it hard to sleep because of untreated physical pain? If so, tell me about it.
Did you have any difficulty sleeping in the bunks because of above-average weight or height, or did you ever see others in this situation? If so, how did you/they handle it?
Do you think staff were also tired/sleep deprived? If so, what makes you say so? If so, how did staff being tired affect your daily life or that of others in the prison?⁵²⁷

During the initial process of drafting the questionnaire, an issue came up that I thought might be fruitful to pursue more systematically: the experience of people diagnosed with sleep apnea who used/tried to use CPAP machines in custody. I therefore designed a second, shorter questionnaire intended for use with this population. It contained targeted questions about sleep apnea and CPAP use,⁵²⁸ and also asked what I thought were likely to be the core questions for all interview subjects: those concerning how long and how well people slept, the specific obstacles to sleep, and how being sleep deprived or being surrounded by people who were sleep deprived affected their carceral experience. After conducting two such interviews, I realized that I would not get sufficiently consistent data on the sleep apnea/CPAP piece to be able to say anything reliable on these issues. I therefore folded these two interviews into my larger data set and ceased soliciting participation from those in this category.⁵²⁹

A question arose early on about whether doing the interviews over Zoom might disadvantage some participants. But it quickly became clear that, in a post-COVID world, virtually everyone was comfortable with the Zoom platform. Occasionally the start of an interview would be delayed by technical difficulties, but these were readily overcome.

At the close of each interview, after I ceased recording, I gave each person the opportunity to offer any thoughts they might have on the experience we had just shared. Once or twice, the subject began talking about an entirely new issue, one we had not discussed during the interview. Each time this happened, I hesitated to risk interrupting the flow by asking to restart the recording. I therefore took as detailed notes as I could and immediately after ending the call recorded my recollection of the exchange as part of the field notes for that interview.

At the close of each interview, I also gave each person the opportunity to withdraw their initial consent to participate and not to have their data included in the study. No one took me up on this offer. After each interview, I followed up by email to share my contact information and the contact information for the UCLA Office of the Human Research Protection Program, in case they had any questions or concerns about the research that they preferred not to go through me. I also shared the unique code for the $30 Amazon gift card that had been promised to each participant as a token of thanks for their participation.

Once the interviews were completed, the recordings were shared with a team of research assistants (“RAs”) for transcription. There were two steps to the transcription process. First, each interview was handed off to an RA who ran it through Otter.ai, which produced an initial transcript that caught roughly ninety to ninety-five percent of the content. That same RA then went simultaneously through the recording and the initial transcript and filled in any missing pieces. Second, the interview was handed off to another RA on the team, who went through the recording and transcript to catch any remaining errors.

The coding process proceeded as follows: I created an initial, lengthy list of coding categories, each taking the form #issue (e.g., #noiseinadseg, #extremeheat, #mattresses). My RAs then went through each transcript, identifying any passages related to each category. Whenever they identified a relevant passage, they dropped the code in the interview text (producing a set of coded interviews searchable by issue) and also added the text to a master coding document that grouped interview excerpts according to categories. To ensure quality control, I reviewed the first set of coded transcripts and an early version of the master coding document. As they went, my RAs sometimes added to the initial list of coding categories. We also created tables, populated with any data amenable to being preserved in this way (e.g., race, gender, state of origin, length of incarceration, age at initial incarceration, age on release, weight on entry, weight on release, and so on). I then went through the master coding document to identify relevant issues and representative passages. I also frequently went back to the coded transcripts to read for context and confirm impressions. In this Article, I only quoted passages that were representative of what I heard more generally.⁵³⁰ Where the experiences varied, for example by security level or housing configuration, I indicated as much in the text.

Appendix B: Descriptive Statistics of the Sample

In this Appendix, I provide tabular representation of several aspects of the research sample amenable to quantification. Unless otherwise indicated, each table includes pertinent data for the thirty-five subjects formally enrolled in the study and the four individuals with whom I conducted ad hoc interviews.⁵³¹

Table 1. Racial Distribution (n=39)
Identified as	# in sample	% of sample	Overall % in U.S. Prison and Jails⁵³²
Black	15	38%	42%
White	13	33%	36%
Hispanic	7	18%	20%
Asian	1	3%	N/A
Native American	1	3%	3%⁵³³
Mixed Race⁵³⁴	2	5%	N/A

Table 2. Gender Distribution (n=39)
Housed in facilities for	# in sample	% of sample	Overall % in U.S. Prisons⁵³⁵
Men	35⁵³⁶	90%	93%
Women	4	10%	7%

Table 3. Jurisdictional Breakdown (n=39)⁵³⁷
As this Table shows, the data on which this Article is based derived from interviews with people with firsthand experience of incarceration in seventeen state systems, as well as the D.C. Department of Corrections and the Federal Bureau of Prisons. In addition, six members of my sample served time in private prisons located in five different states.
Alabama	2
California	5
Colorado	2
Illinois	2
Louisiana	4
Michigan	1
Missouri	1
New Jersey	5
New York	4
Ohio	1
Oregon	2
Pennsylvania	2
Rhode Island	1
Tennessee	1
Virginia	3
Washington, D.C.⁵³⁸	2
Wisconsin	1
Wyoming	1
Federal Bureau of Prisons	4
Private prisons⁵³⁹	6 (located in Arizona, California, Colorado, Ohio, Oklahoma, and Virginia)

Table 4. Housing Configuration (n=39)
The figures in this table represent the total number of study subjects who experienced each type of housing configuration. If, as was common, a subject experienced multiple housing configurations over their incarceration, then all these configurations were counted in this table. It was not always possible from the interview transcripts to determine with precision the full range of a person’s experience, so these numbers are most likely undercounts.
Housing Type	Total # Interview Subjects Reporting Experience
Solitary confinement	29
Single cell	29
Double cell	30
3+ cell (range: 3–10 occupants)	10
Cubicle in dorm	12
Open dorm	25

Table 6. Misc. Data Points
	Minimum Value	Q1	Q2	Q3	Maximum Value
Hours of sleep per night (n=39)	3 (n=3)	4.5	5.1	6	11.5
Total years incarcerated (n=39)	3	9.5	13.5	23.5	45
Interview length (n=35) (hours: mins)	1:03	1:44	1:58	2:14	2:48
Years between release and interview	0	1	2	4	12 (n=2)

Table 6 features several data points as to which the answers reflected a range. To capture the breadth of responses, I used the standard “five number summary,” indicating as to each category the minimum value, Q₁ (25th percentile/lower quartile), Q₂ (50th percentile/median value), Q₃ (75th percentile/upper quartile), and the maximum value. As to each data point, some further explication is in order, as follows:

Hours of sleep per night: The effort to calculate the hours of sleep per night my subjects received was complicated by many factors, including the long duration of the sentences served by many of my subjects and the wide range of experiences each had across housing configurations, not to mention inevitable variability night to night. I asked people to give me their best estimate of how many hours per night they slept on average in each housing configuration they experienced. When respondents reported variation across their incarceration, I took an average of the reported hours slept. If someone gave a range (e.g., four to six hours per night), but characterized the sleep quality as very poor or described themselves as waking frequently during the night, I took the low end of the stated range as more accurately reflecting the actual amount of sleep they likely received. These limitations mean that the numbers reported should be considered rough estimates only, although the considerable sleep deficit they indicate seems entirely consistent with the narrative accounts of the extremely limited and poor-quality sleep reported in the interviews.

The maximum value for hours of sleep per night is 11.5. The person who reported this outlier value served time in two prisons and reported sleeping eight hours a night in the first facility and fourteen to sixteen hours per night in the second. To calculate his hours slept for purposes of this chart, we took the average of eight hours (from prison #1) and the middle of the range he named for prison #2.⁵⁴⁰

Total years incarcerated: In building my sample, I sought to enroll only people who had spent at least six years in prison. I wound up enrolling two individuals who had done four years each, because each offered diversity as to other metrics. The minimum value noted in the table (three years in custody) reflects the experience of one of the individuals with whom, in part for this reason, I opted to conduct an ad hoc interview rather than formally enrolling them in the study.

Interview length: For this data point, I included only the thirty-five interviews conducted with those individuals formally enrolled as subjects in the study. I did so because the four ad hoc interviews took a different form and did not follow the path shaped by the questionnaire. For reference, the four ad hoc interviews ranged in duration from 1:05 to 2:09.

Years between release and interview: The interviews were conducted in 2023. Two members of my sample were released that same year and were coded as having had zero years between release and interview. Two members of my sample were released in 2011, twelve years prior to the interview. This gap was considerably greater than my intended parameter of five years. Each of these individuals was interviewed very early in the process, when I was still unsure whether I would find enough participants to assemble an adequate sample. Once things progressed and I had more people expressing interest in participating than I had time to interview, I reverted to requiring a release date within five years of the interview. One other enrolled participant was outside this stipulated limit. He had entered the wrong release date when he filled out the initial form expressing interest in participating, substituting 2018 for 2015. When I enrolled him in the study, I believed he was five years

out, and was surprised when, at the end of the interview, he named 2015 as his year of release. Having already conducted the interview, I opted to retain his data as part of the full data set.

99 S. Cal. L. Rev. 95

Download

Sharon Dolovich*

*Professor of Law, UCLA School of Law. Thanks to Joe Doherty, Phil Goodman, Keramet Reiter, and Ashley Rubin for generously sharing their methodological advice and good counsel; Sasha Natapoff for many fruitful conversations; Ahilan Arulanantham, M. Forrest Behne, Chesa Boudin, Beth Colgan, Ethan Corey, Phil Goodman, Emma Kaufman, Sasha Natapoff, and Keramet Reiter, as well as participants at the Criminal Justice Roundtable at Harvard Law School and the UCLA Faculty Workshop, for helpful feedback; Muna Ali, Cecilia Bain, Zeke Bolden, Regina Campbell, Eugine Choo, Daniel Dulaney, Michaela Firmage, Jay Foss, Kyle Groves, Leslie Hernandez, Brennan Hickson, Ankita Katukota, Hadley Levenson, Emma Maynard, Kate Mitchell, Eireann O’Grady, Ginny Oshiro, Taylor Robinson, Natalie Sorrentino, Jack Stephens, Allie Tackett, Claudine Ushana, and Sorcha Whitley for their tremendous research assistance; Omer Sharon for guidance on the sleep science; Harvey Ellis, Freddy Guzman, Ken Hartman, Haller Jackson, Bart Lanni, and Ginny Oshiro for helpful input on the content of my questionnaires; and the Vital Projects Fund, the UCLA Academic Senate, and the UCLA ORCA Acceleration Grants Program (OAG) for financial support of this research. My deepest gratitude goes to the many individuals (who per IRB requirements must remain anonymous), who generously contributed their time to this study in the hope that, once people understand this dimension of the carceral experience, conditions on the ground might change for the better.

Controlling the Narrative: Government Speech and Book Bans in the Public Library

January 2026February 2026Kelcey Sholl*

INTRODUCTION

Books bring personal joy and potential for individual growth. “The right book put in the right hands at the right time, could change the course of a life or many lives.” ¹ But access to books is not just a matter of personal enrichment: books are essential sources of knowledge, and free access to that knowledge is fundamental to a successful, thriving democracy. Michael J. Barsanti, the former director of the Library Company, the first American library, pointed out that the Founders “knew that democracies were inherently fragile and that the only way you could sustain a democracy was by having an educated populace.”²

Access to literature thus empowers individuals to engage thoughtfully in public life and fosters critical thinking and informed decision-making. Sociological studies have identified that literature introduces students to current cultural issues, which can help them develop global awareness.³ Further, when readers vicariously experience the lives of others through literature, they gain the ability to look critically at the world by engaging with other points of view.⁴ The saying “knowledge itself is power,” first published in 1597,⁵ has endured over the centuries because its truth is undeniable—and it is precisely this truth that drives some individuals in governments and communities to fight to keep certain books, and the power books hold, out of the public’s hands in an effort to control public discourse.

Libraries have become battlegrounds for partisan agendas. The concerted effort by State legislatures and citizen “activist” groups to ban certain books in schools⁶ is no longer confined to school classrooms but is spilling over into public libraries.⁷ This expansion of book bans from schools to libraries reflects an increasing desire to control public discourse and limit the availability of ideas that challenge the status quo. And by limiting access to certain books, these groups aim to control the public narrative and prevent the spread of ideas they perceive as threatening.

Efforts to ban books are rooted in fear—fear of ideas that challenge the status quo, of diverse perspectives, and of the power knowledge gives individuals to question public discourse, think critically about societal issues, and advocate for change.⁸ They are at the center of an ongoing culture war around “woke” ideology, and it is a serious constitutional concern.⁹ J.B. Pritzker, Governor of Illinois, has stated, “There are few perils to our democracy as dangerous as book bans” because “[t]hey threaten the very freedom of thought and speech that underpin our republic.”¹⁰ Book bans and censorship, therefore, threaten more than a personal freedom to read: they undermine the foundation of our democratic society by stifling an exchange of ideas that stem from the freedom to engage with diverse selections of ideas held within books.

This freedom of thought and speech is guaranteed by the First Amendment of the Constitution, which not only enshrines the rights to speak and publish, but also protects a wide range of mediums, including books, newspapers, art, music, clothing, and online content from political interference.¹¹ The U.S. Supreme Court and other lower courts have also established that the right to receive information is a fundamental extension of the right to free speech.¹² Book bans, which directly restrict access to ideas and information, therefore warrant close examination under the First Amendment.

This Note will focus on cases that have addressed book bans, including the 2025 Fifth Circuit case of Little v. Llano County and the 2024 Eighth Circuit case of GLBT Youth in Iowa Schools Task Force v. Reynolds. These cases are critically important because they are two of the first cases to make their way to federal courts of appeal amid a recent surge in litigation surrounding the proliferation of book bans. In addition, the defendants in both cases argued that library book curation constitutes government speech, which would render the government immune from First Amendment claims.

Little is particularly important due to its potential to severely limit First Amendment protections in the context of public libraries—spaces that historically play a vital role in providing access to diverse ideas and expression. The case is notable because after the Fifth Circuit panel originally ruled in favor of the plaintiffs and granted them an injunction due to the likelihood that they would prevail on the merits of a First Amendment violation, the decision was vacated to be heard en banc,¹³ with a plurality of the panel ruling on May 23, 2025, that library book curation is indeed government speech.¹⁴ The Little ruling created a circuit split on the issue, as the Eighth Circuit held in GLBT Youth that such an action is not government speech.¹⁵ The Fifth Circuit is widely viewed as one of the most conservative circuits in the nation that is willing to overturn longtime precedents,¹⁶ and in reaching its decision in Little, it has overturned Campbell v. St. Tammany Parish School Board, a thirty-year precedent that held books cannot be removed from libraries solely due to disagreement with the ideas they contain.¹⁷

With Little poised to make its way to the Supreme Court after plaintiffs petitioned the case for review, this Note will argue why the Supreme Court should grant certiorari and hold that library book curation is not government speech. The Supreme Court has not heard a book banning case since its 1982 non-binding, plurality decision in Board of Education v. Pico.¹⁸ This Note explores how the Pico decision could be strengthened. Specifically, it will advocate for distinguishing the important difference between public libraries and public-school libraries, confirming that book removal cannot be motivated by viewpoint discrimination, and providing guidance for determining when an action is motivated by viewpoint. This Note will then examine Little, arguing that the actions at issue in Little are, in fact, motivated by viewpoint discrimination. Finally, this Note will address the serious concerns and dangerous consequences of holding that library book curation is government speech—a new issue not presented in the Pico case—and will argue that the Court should clearly assert that it is not government speech.

This Note will refer to terminology including “bans” and “censors.” A book being “banned” traditionally and commonly refers to a book as being “removed from school curriculums [or] public libraries.”¹⁹ Censor, used as a verb, means to “examine in order to suppress” or “delete anything considered objectionable”; further, censor, used as a noun, is “a person who supervises conduct and morals.”²⁰ There are various methods in which a censor can censor or ban a book: (1) redaction, (2) restriction, and (3) relocation.²¹ Redaction refers to removing information within a book but keeping the book available for patrons. Restriction refers to removing the book altogether. Relocation refers to requiring a book to be shelved in a separate area that often requires permission to access.²² Overall, this Note will use the term “ban” to refer to a combination of restriction and relocation but does not focus on instances in which books have been redacted.

I. Board of education v. PICO

Although the Supreme Court has heard numerous First Amendment challenges, those addressing the restriction of access to books are limited.²³ The Court’s 1982 decision in Pico is the only opinion in which the Court has directly addressed the issue of book banning in libraries—particularly a school library.

The case revolves around a decision by the Island Trees school board to remove several books from its middle and high school libraries after obtaining a list of “objectionable” books from an organization called Parents of New York United (“PONYU”).²⁴ PONYU was a conservative organization of parents focused on education legislation in New York State.²⁵ The “objectionable” books included popular titles such as Slaughterhouse-Five by Kurt Vonnegut and Go Ask Alice.²⁶ The school board characterized these books as “anti-American, anti-Christian, anti-[semitic], and just plain filthy,” and concluded that it was the board’s “moral obligation” to “protect the children in [their] schools from this moral danger as surely as from physical and medical dangers.”²⁷ Several students in the district, led by Steven Pico, challenged the removal of these books, arguing that the school board violated students’ First Amendment rights. They contended that the desire to suppress ideas that were politically and socially controversial was a form of viewpoint discrimination.

In its 5-4 plurality decision, the Court held that a school board’s removal of books could not be based on a desire to suppress ideas.²⁸ The Court did reason, however, that school boards have significant discretion in regulating the content of materials in their schools,²⁹ although this discretion was not unlimited.³⁰ The Pico decision was narrow in several respects: notably, it lacks the establishment of a standard to determine when book removals are in violation of the First Amendment because of viewpoint-based restriction. The Court left open the vague possibility that books could be removed for reasons such as vulgarity or lack of educational value, without offering guidance on how those reasons could be distinguished from pretextual justifications. While Pico was important for affirming that students have the constitutional right to receive information, and that school boards do not have unlimited authority to remove books, it leaves unanswered questions about when book removals violate the First Amendment.

Additionally, the case leaves open questions about whether the Pico school board’s actions—which closely resemble the actions Llano County took in its public library in Little—did in fact violate the First Amendment, given that the case was remanded and then settled with no holding based on the facts. Despite these limitations, Pico is an important precedent in the ongoing debate over book banning. Lower courts that have decided on similar issues have closely followed the Pico holding.³¹ For example, in the Fifth Circuit case Campbell v. St. Tammany Parish School Board, in which a school board removed books such as Kurt Vonnegut’s Slaughterhouse-Five, the Court held that public schools cannot censor books merely because they contain controversial or offensive content.³²

Considering the increasing number of book bans in the United States, there is a growing reason for strengthening the Pico decision. This Note will argue, as detailed above, that to strengthen the holding of Pico, the Court must distinguish the important difference between public libraries and public-school libraries. Additionally, given that Pico was a plurality opinion, the Court must reaffirm that book removal cannot be motivated by viewpoint discrimination in the context of libraries, and it must more clearly define the types of actions that constitute a violation of the First Amendment principle on viewpoint-based discrimination.

II. THE FIRST AMENDMENT

The First Amendment has been greatly revered over the course of American history and is considered by many to be the most influential and important amendment in the Bill of Rights.³³ The First Amendment guarantees “the people” freedom from government intervention in their right to free speech, which ensures the exchange of free ideas.³⁴ As with all rights, the First Amendment is not absolute; however, its central purpose lies in protecting the expression of unpopular ideas.³⁵

One of the key rationales behind vigorously protecting various forms of expression, including unpopular ideas, is the “marketplace of ideas.”³⁶ The phrase was first popularized by Justice Oliver Wendell Holmes in his 1919 dissent in Abrams v. United States.³⁷ The concept is grounded in the belief that for society to progress, “truth” will emerge only from the free exchange of ideas, which necessarily includes unpopular ones.³⁸ Ensuring that society can critically engage with competing viewpoints is therefore essential for individuals to evaluate the merits of different ideas and ultimately arrive at a better understanding of these “political truth[s].”³⁹

Importantly, the marketplace only functions when there is open debate and a free flow of information.⁴⁰ Holmes argued that the government should not have the power to suppress speech on the grounds that the speech is controversial or unpopular. In his view, the government should only step in when the speech poses a clear and imminent danger to the public or national security.⁴¹ Nearly a decade after his Abrams dissent, Justice Holmes joined Justice Brandeis in the latter’s concurring opinion for the famous 1927 case Whitney v. California, highlighting why the Founders believed strongly in protecting free speech:

[The Founders] believed that freedom to think as you will and to speak as you think are means indispensable to the discovery and spread of political truth; that without free speech and assembly discussion would be futile; that with them, discussion affords ordinarily adequate protection against the dissemination of noxious doctrine; that the greatest menace to freedom is an inert people; that public discussion is a political duty; and that this should be a fundamental principle of the American government.⁴²

Book bans are fundamentally incompatible with Holmes’s marketplace of ideas theory and the First Amendment protections that safeguard the right to create, distribute, and access written works without government interference. Public libraries play an important role in maintaining an open marketplace of ideas. Because libraries are institutions that provide knowledge and information to “the people,” they must contain a wide range of ideas available for exploration, free from government interference based on ideological or political preferences. Unchecked government interference in the marketplace of ideas will manipulate the flow of knowledge and ensure that only certain ideas are permitted to reach “the people.”

A. Viewpoint Discrimination

Part of the First Amendment’s speech-protecting safeguards is the protection against efforts to restrict speech based on the speech’s content or viewpoint.⁴³ The government is, in nearly all circumstances, prohibited from engaging in viewpoint discrimination.⁴⁴ In 1989, the Court held in Texas v. Johnson that “[i]f there is a bedrock principle underlying the First Amendment, it is that the government may not prohibit the expression of an idea simply because society finds the idea itself offensive or disagreeable.”⁴⁵

A law is considered viewpoint-based if it restricts speech due to a specific ideology, opinion, or perspective expressed by its speaker, rather than regulating speech based on content-neutral criteria.⁴⁶ Viewpoint discrimination is particularly dangerous because if permitted, the government is able to silence certain viewpoints based solely on subjective disapproval, which undermines the foundational prohibition on government censorship of speech with which the government disagrees.

In the context of book bans, viewpoint discrimination occurs when government actors remove or restrict access to books because they disagree with the ideas those books express, rather than because of neutral concerns such as age-appropriateness or educational relevance. This often happens when books are targeted for addressing topics that some find controversial, such as race, politics, or sexuality. For example, the removal of books that discuss LGBTQ+ identities or systemic racism reflects viewpoint discrimination when those books are excluded specifically to suppress those perspectives, while books expressing more politically favored viewpoints remain available.

B. Government Speech

Despite the Court interpreting the First Amendment to prohibit viewpoint discrimination, the government-speech doctrine exempts the government from the usual First Amendment restrictions on content- and viewpoint-based speech limitations, allowing it to engage in such discrimination.⁴⁷ The concept of government speech was articulated in the 1991 case Rust v. Sullivan, even though the term “government-speech” was not utilized in the opinion.⁴⁸

In Rust, the Court upheld a government regulation that any healthcare professional receiving Title X funding was prohibited from providing information about abortion, consequently limiting what healthcare professionals were permitted to say to their patients about abortion.⁴⁹ The Court reasoned that because the government is not obligated to fund all viewpoints, and because it has the ability to select which programs to fund (here, family-planning that excluded abortion), the government’s actions would inevitably conflict with the perspectives of other programs it chooses to promote.⁵⁰ Essentially, the government should be able to communicate its own messages or express its own policies and particular viewpoints without violating the First Amendment.⁵¹ Two decades later, in the 2009 case Pleasant Grove City v. Summum, the Court identified government speech when actions are “meant to convey and have the effect of conveying a government message.”⁵² In 2022’s Shurtleff v. City of Boston, the Court applied various factors to determine if governmental action is in fact government speech: (1) “the history of the expression at issue”; (2) “the public’s likely perception as to who (the government or a private person) is speaking”; and (3) “the extent to which the government has actively shaped or controlled the expression.”⁵³

While there may be some truth to the idea that effective governance would be impossible if the government were unable to support or oppose certain viewpoints when implementing programs,⁵⁴ scholars have noted an inherent contradiction between the government speech doctrine and the First Amendment:

To satisfy traditional First Amendment tests, the government must show that it is not discriminating against a viewpoint. And yet if the government shows that it is condemning or supporting a viewpoint, it may be able to invoke the government speech defense and thereby avoid constitutional scrutiny altogether. Government speech doctrine therefore rewards what the rest of the First Amendment forbids: viewpoint discrimination against private speech.⁵⁵

Seemingly for this reason, the Court has warned that courts should “exercise great caution before extending . . . government-speech precedents.”⁵⁶

C. Minor’s Right to Free Speech

An individual’s status as a minor cannot serve as a blanket justification for censorship. The Supreme Court has recognized that minors enjoy a degree of expressive liberty under the First Amendment, although it is not as expansive as that of adults.⁵⁷ In Tinker v. Des Moines Independent Community School District, the Court famously held that students do not “shed their constitutional rights to freedom of speech or expression at the schoolhouse gate.”⁵⁸ Also, in Erznoznik v. City of Jacksonville, the Court held that lawmakers cannot suppress speech for the sole purpose of shielding minors from ideas they do not find appropriate.⁵⁹

III. HISTORY AND PURPOSE OF PUBLIC LIBRARIES

Libraries have long stood as pillars of American civic life, serving the public as “forums for information and ideas.”⁶⁰ Founding Father Benjamin Franklin championed the first iteration of the American library system in 1731 with the founding of the Library Company of Philadelphia, which permitted individuals to borrow books for a fee.⁶¹ By 1800, this concept grew in popularity, and there were over forty fee-based libraries established across the United States.⁶² In his autobiography, Franklin highlighted the important role libraries play in the furtherance of democracy, noting how the Library Company “improved the general conversation of the Americans” and “made the common tradesmen and farmers as intelligent as most gentlemen from other countries.”⁶³

Today, there are more than 9,000 libraries in the United States.⁶⁴ Although libraries had long served a public civic role, the role was formally embedded in public life through tax-funded libraries, beginning in 1833, when the Peterborough Town Library in New Hampshire became the first such institution in the United States.⁶⁵ While the libraries of Franklin’s era were far from the inclusive institutions we strive for today, the heart and soul of the library’s purpose has remained and grown. Libraries serve our society as equalizers, ensuring access to knowledge for all regardless of background or circumstances while strengthening the democratic foundation of society.⁶⁶

For instance, consider libraries in the mid-19th century. As industrialization and urbanization took shape during this period, there was a significant expansion of subscription libraries and public library systems, fueled in part by demand from the working class and their employers for self-education.⁶⁷ Employers funded libraries as investments in an “efficient . . . working class,” and trade-specific libraries such as the Mechanics’ Institute opened to “stretch the mind and teach new skills.”⁶⁸ One such proponent of expanding libraries to the working class was Andrew Carnegie, who provided an employee library at one of his steel plants in Pennsylvania.⁶⁹ Carnegie became one of the richest men in the world for his role in oil, steel, and railroads, and used much of his wealth for philanthropic endeavors⁷⁰—one of which was the creation of a free public library system. He referred to libraries as “palaces for the people”⁷¹ and spent more than $56 million to create more than 2,500 libraries.⁷² He worked with local governments to establish government-funded libraries, for which he donated buildings through a grant. Localities then instituted taxes to fund the libraries’ operations and acquisitions of books.⁷³

Emily Knox, a professor at the University of Illinois Urbana-Champaign who researches censorship and library science, has stated that a guiding principle embraced by librarians is the “right of every individual to both seek and receive information from all points of view without restriction.”⁷⁴ This commitment to receiving information has been evident for more than a century. In a brief New York Times article from 1918 titled Not a Public Library Ban, the director of the New York Public Library (“NYPL”) sought to dispel any suggestion of censorship by responding to claims cited in an earlier article that a patron had been denied access to The Unpardonable Sin. He clarified that the book was available and requested identifying information about the individual who had suggested otherwise.⁷⁵ This early defense of open access underscores the long-standing dedication of libraries to resisting censorship and ensuring the public’s right to access information. Later, in response to “growing intolerance, suppression of free speech, and censorship affecting the rights of minorities and individuals,” the American Library Association adopted its first Library Bill of Rights in 1939, emphasizing that libraries are neutral spaces for the free exchange of ideas.⁷⁶ David Leonard, President of the Boston Public Library (the third largest library behind the Library of Congress and the NYPL) states that “Free to All”—as is inscribed on many of the library buildings—“encapsulates what we [are] all about.”⁷⁷

The evolution of libraries from small lending collections to vast public institutions reflects a commitment to ensuring that information remains accessible to all, without government interference or ideological gatekeeping. At their core, libraries represent a belief in the power of knowledge to uplift individuals and strengthen society. However, this very openness and inclusivity have made them a target for those who seek to control the flow of information. The ongoing battle over book removals and censorship in libraries threatens to undermine these fundamental principles that have endured since this country’s founding. The fight to keep libraries truly “free to all” is not just about books—it is about preserving the foundational freedoms on which this country was built.

IV. HISTORY OF BOOK BANS AND CENSORSHIP

The banning of books and the censorship of ideas are not new phenomena. Both have persisted throughout the course of human history. In 213 B.C., Chinese emperor Qin Shi Huang ordered various books of poetry and history to be burned;⁷⁸ Ovid, a Roman poet, was banished from Rome in A.D. 8 for his The Art of Love; in 1526 England, thousands of copies of the New Testament were burned; for more than 400 years from 1564 to 1966, the Catholic Church maintained the Index Librorum Prohibitorum—Index of Prohibited Books—that forbade Catholics from reading books by authors such as Victor Hugo, John Locke, and Jean-Paul Sartre; and from 1788 to 1820, King George banned Shakespeare’s King Lear from stage performance.⁷⁹ In modern times, book banning has steadily continued. In 1939, John Steinbeck’s novel The Grapes of Wrath was removed from a Kansas City library for being “indecent”;⁸⁰ during the Nazi regime, thousands of books including Jack London’s The Call of the Wild and numerous Ernest Hemmingway titles were banned and burned;⁸¹ in the 1980s, Beatrix Potter’s The Tale of Peter Rabbit was banned from schools in London County, England; in 1987, at a North Carolina high school, Maya Angelou’s I Know Why the Caged Bird Sings was removed from the required reading list;⁸² lastly, Toni Morrison’s Beloved has been included in the American Library Association’s top ten most challenged books of the year list seven times since 2006.⁸³ While justifications for banning books may evolve to reflect the politics and ideologies of a particular time, those who seek to control what others can read share similar claims of acting in the interest of morality, public order, or protecting children as their reasons for book banning.

Although bans are not new, “[they are] definitely getting worse.”⁸⁴ The American Library Association’s Office for Intellectual Freedom documented 4,240 unique titles challenged in 2023, making it a 65% increase from 2,571 titles in 2022.⁸⁵ This was the highest level since the organization began tracking twenty years ago.⁸⁶ Comparatively, the number of unique titles challenged in 2014, just over 10 years ago, was 235.⁸⁷ In 2020, the number of unique titles was still less than 300, but that number shot up to 1,858 in 2021 and has been growing since.⁸⁸ Between July 2021 and June 2023, PEN America reported that 5,894 books have been banned across 41 states.⁸⁹ The trend is most prominent in “Republican-leaning” states—Florida and Iowa had the most book bans from 2023 to 2024⁹⁰—reflecting an increasingly organized and politically motivated effort to control access to literature.

The motivations behind banning books often stem from deeply rooted psychological and sociological factors—“an act of control[] driven by . . . fear.”⁹¹ Emily Knox, who studies book banning and censorship, reported that communities try to ban books because words have power, and words can change who an individual is.⁹² Throughout history, censorship has been a tool for controlling narratives.⁹³ Economic and social turmoil often lead to widespread efforts to restrict the rights of marginalized individuals,⁹⁴ and silencing their voices by restricting their access to books is a powerful tool for accomplishing this goal. There is a desire to preserve societal norms, and restricting books keeps them away from readers whom censors fear might be “changed” by their content.⁹⁵ Encountering ideas that may conflict with an individual’s deeply held beliefs can create a mental discomfort that many avoid by challenging books that challenge their worldview.⁹⁶ “Activist” groups pushing for parents to challenge books often use the term “indoctrination” to describe book content they object to—instead of the word “education”—to emphasize that their worldview does not comport with the worldview expressed by the challenged book.⁹⁷ Because books introduce unfamiliar and challenging ideas that may be uncommon in certain communities or regions, the differing views and lives represented in the pages of the challenged book are often seen as a threat.⁹⁸ Thus, individuals justify restricting these ideas as a desire to protect readers from radical ideas.⁹⁹

V. PROLIFERATION OF BANS AND LEGISLATION

The rise in book bans has become a significant national issue, sparking debate regarding parental rights and minors’ First Amendment rights. As of 2022, more than one-hundred bills have been proposed for various forms of book banning at the state level.¹⁰⁰ Florida’s HB 1557, also known as the “Parental Rights in Education” law and referred to by opponents as the “Don’t Say Gay” law, restricts discussions of sexual orientation and gender identity in certain grade levels.¹⁰¹ With respect to book banning, the Florida law increases the ease with which parents, or any resident, can object to books used in classrooms and school libraries.¹⁰² It requires school districts to house an “objection form” on the school district’s homepage website, and directs districts to remove books objected to as “pornographic” within five days of receiving an objection.¹⁰³ Charlotte County school district, acting on the law, prohibited books with LGBTQ+ characters in all schools with limited exceptions for high schools; Charlotte’s superintendent and school attorney stated, “These characters and themes cannot exist.”¹⁰⁴ Two years after the enactment of this law, Florida was required through a settlement to provide direction to schools regarding the scope of the legislation; subsequently, it has been clarified that the law does not apply to libraries in which no educational instruction is taking place.¹⁰⁵

In addition, Texas’ HB 900, the READER Act, attempted to enact stringent guidelines on educational materials by requiring school book vendors to rate books that contain “sexually explicit” or “sexually relevant” content in order to restrict access to these books.¹⁰⁶ Specifically, the Act required book publishers, bookstores, and internet-based book retailers to rate books in active use that they sold to public schools.¹⁰⁷ These independent, private entities received minimal guidance on how to implement the ratings despite the fact that books labeled by the sellers as “sexually explicit” would consequently be removed from public school libraries.¹⁰⁸ In May of 2024, the Fifth Circuit in Book People v. Wong upheld a district court injunction that held the READER Act to be unconstitutionally vague and compelled speech.¹⁰⁹ The legislation in Florida and Texas have served as models for similar measures in other states such as Arizona, Missouri, Tennessee, and Idaho, signaling a broad trend.¹¹⁰

Much of the book banning debate has focused on public school libraries and academic curriculum choices, but there has been a growing trend in regulation and litigation regarding public community libraries. For instance, Llano County in Texas and Pasco County in Florida have begun expanding book banning efforts from school libraries to community libraries. For example, in September of 2024, Pasco County removed 130 children’s books with LGBTQ+ content from its community library.¹¹¹ In January of 2025, South Carolina introduced a bill in its state senate requiring libraries to certify each quarter that they “do not offer books or materials in children’s, youth, or teen sections that might be considered sexually inappropriate” in order to receive state funding they need.¹¹² In April of 2024, Idaho passed House Bill 710 to require that books deemed harmful to children be relocated to “adult only” areas in public community libraries,¹¹³ causing some rural libraries to respond by prohibiting children’s access to libraries altogether because their libraries are too small for a separate section.¹¹⁴

Critics of the laws that restrict access to books contend that they infringe upon First Amendment rights, including the right to receive information. They argue that these regulations constitute censorship and limit readers’ exposure to diverse perspectives. The debate centers on whether a state government’s interest in shielding children from content it deems inappropriate outweighs a broader societal commitment to the free exchange of ideas.¹¹⁵ States that are traditionally more liberal are responding to this increase in bans by enacting “anti-book ban” laws for both school and community libraries.¹¹⁶ California, Illinois, Minnesota, Maryland, and New Jersey have all enacted protective laws,¹¹⁷ and in June of 2023, Illinois became the first state to enact legislation that outlaws book bans.¹¹⁸ The Illinois law requires libraries to create formal policies or adopt the American Library Association’s Bill of Rights that asserts books could not be removed because of partisan disapproval.¹¹⁹ More states are beginning to introduce protective measures. In February of 2025, for example, a bill was introduced in the New York State Assembly to prohibit libraries from banning books based on “partisan or doctrinal disapproval.”¹²⁰

This proliferation of legislation and regulation has received federal attention, with hearings in both the House Committee on Oversight and Reform in 2022 (entitled “Free Speech Under Attack”) and in the Senate Committee on the Judiciary in 2023 (called “Book Bans: Examining How Censorship Limits Liberty and Literature”).¹²¹ Supporters of states’ legislation and regulation limiting access to books argue that governments have the authority to regulate access to content deemed inappropriate and to protect children from material they consider harmful.¹²² These arguments often frame the issue as one of parental rights.¹²³ In the 2023 Senate Judiciary hearing, the committee heard from five “witnesses” with varying points of view on the issue. Nicole Neily, one of the hearing’s witnesses and President of “Parents Defending Education,” stated before the committee that “families’ concerns about books in schools is not ‘book banning.’ ”¹²⁴ She indicated that their concern with “age appropriate[ness]” is not “radical,”¹²⁵ and stated that the debate around book banning “is a manufactured crisis that distracts from families’ valid concerns about the quality of their children’s education and whether students are safe from drugs, assaults, and bullying in schools today.”¹²⁶ Neily, betraying her own argument that this is about children’s safety and not her own disapproval of LGBTQ+ content, concluded, “Please listen to the concerns of families who want their children to learn basic grammar, rather than be policed on pronouns.”¹²⁷

On the other hand, those against the book ban legislation and regulation argue that it is a violation of First Amendment rights and a threat to democracy. Illinois Secretary of State Alexi Giannoulias spoke about his initiation of House Bill 2789, which proposed that Illinois libraries will not receive funding if they ban books:

This legislation is important because both the concept and the practice of “banning books” contradicts the very essence of what our country stands for and what our democracy was founded on. It also defies what education is all about: [t]eaching our children to think for themselves. If the book banners care to, they can go to our libraries and check out the Federalist Papers, the U.S. Constitution and even Supreme Court cases on the First Amendment. What they will learn is that our democracy depends on the “marketplace of ideas.” That “marketplace of ideas” will not function if we ban books—because we will be banning ideas and preventing our children from thinking for themselves and having the ability to debate [and] learn [and] understand different perspectives.

[ . . . ]

I could never imagine a world where I would tell another family what books their kids should or should not be allowed to read.¹²⁸

This surge in book bans is an extension of broader culture wars currently gripping the nation, particularly around issues of “wokeism,” so-called “traditional” family values, and the targeting of transgender individuals by conservative politicians.¹²⁹ These bans reflect a systematic effort to reshape public discourse by restricting access to materials that address topics such as LGBTQ+ identities, systemic racism, or gender diversity. On January 24, 2025, the U.S. Department of Education’s Office for Civil Rights announced the elimination of its book ban coordinator position, which had been established in June of 2023 to investigate complaints regarding school removals of books claimed to be “racially divisive.” Along with eliminating the position, the department dismissed eleven related complaints, stating that the concerns over book bans were a “hoax” and a “false narrative” perpetuated by then-President Biden.¹³⁰ Additionally, on February 7, 2025, the Department of Defense (“DoD”) announced that it would be reviewing books “potentially related to gender ideology or discriminatory equity ideology topics” in DoD-funded schools located in seven states and eleven countries.¹³¹ The DoD removed books from these school classrooms and libraries, including titles such as Freckleface Strawberry written by actress Julianne Moore about a young girl learning to accept her freckles and No Truth Without Ruth, a biography about Justice Ruth Bader Ginsburg.¹³²

One of the activist groups leading the charge on book restrictions is Moms for Liberty.¹³³ Moms for Liberty was founded in 2021 and began by speaking out against mask mandates and COVID vaccination requirements in schools. It later shifted gears to focus on the ways in which topics related to LGBTQ+ individuals, racism, and religion are presented in schools.¹³⁴ The group is similar to that of PONYU discussed earlier in regard to the Pico case, and the Southern Poverty Law Center has compared Moms for Liberty to “pro-segregationist parent groups that flourished in the wake of . . . Brown v. Board of Education.”¹³⁵ The group’s influence, for example, led to a parent in South Carolina sending a list of ninety-seven books to her child’s school district for removal in early 2024. The list of books was obtained from a site called “Book Looks,” which was created by a former member of Moms for Liberty to enable parents to “find out what objectionable content may be in [their] child’s book before they do.”¹³⁶ The school librarian was subjected to numerous threats from parents who wanted to turn the librarian in to the police and the FBI for distributing pornography, despite the school having a program in which parents could fill out an opt-out form so their children would not be able to access the book.¹³⁷ Similarly, in the year preceding the enactment of the Idaho House Bill 710 discussed earlier, the Idaho Falls Library received twenty-one complaints from a group called “Parents Against Bad Books.”¹³⁸

Although framed by supporters and activist groups as measures to protect children and uphold parental rights, these laws reflect fears that exposure to certain ideas undermine “traditional” family values. Critics argue that these actions amount to censorship, threaten and violate First Amendment rights by limiting access to diverse perspectives, and politicize public spaces like libraries that are supposed to function as neutral fora for freedom of thought and expression. Art Spiegelman, a Pulitzer Prize winner and author of widely banned book Maus, a graphic novel depicting his father’s survival of the Holocaust, says of the current situation, “It [i]s a culture war that [i]s totally out of control.”¹³⁹

VI. PROLIFERATION OF LAWSUITS

There have been numerous legal challenges to legislation and book removals in schools and libraries. However, this Note focuses on two cases that have specifically raised government speech arguments and have gone before the appellate courts: the Eighth Circuit case of GLBT Youth in Iowa Schools Task Force v. Reynolds and the Fifth Circuit case of Little v. Llano County. Notably, a newer case, PEN America Center, Inc. v. Escambia County School District, has been ongoing in front of the Northern District of Florida, in which the school district is also advancing a government speech argument.¹⁴⁰ This represents a possibility for this argument to now make its way through the Eleventh Circuit.

A. Eighth Circuit: GLBT Youth in Iowa Schools Task Force v. Reynolds

In 2023, Iowa enacted Senate File 496 (“SF496”), a contentious law that reshaped policies governing public school libraries and classrooms.¹⁴¹ The legislation introduced significant restrictions and prohibitions such as requirements to remove books deemed age-inappropriate from school libraries.¹⁴² Two lawsuits emerged in response to SF496. The first was brought by GLBT Youth in Iowa Schools Task Force, a group of students who asserted that SF496 violated the First Amendment. The second lawsuit was brought by Penguin Random House, authors, and educators, who raised similar First Amendment violations. Because “both cases related to the same legislation . . . and had considerable overlap such that the cases would benefit from joint administration,” the District Court for the Southern District of Iowa consolidated the cases.¹⁴³

In response to the plaintiffs, the State argued that the removal of books from school libraries constituted government speech, which is not subject to First Amendment viewpoint-neutrality requirements when the government speaks on its own behalf.¹⁴⁴ The District Court then issued a preliminary injunction on SF496 due to the likelihood that the law violated the First Amendment.¹⁴⁵ On appeal, the Eighth Circuit Court of Appeals overturned the district court’s injunction but, importantly, rejected the application of the government speech doctrine while noting that the Supreme Court has not held that library curation is government speech.¹⁴⁶ Also, applying the Shurtleff factors, discussed earlier in Section II.B, the Eighth Circuit stated that the public does not typically associate this activity with government action, as the library usually hosts a broad and diverse set of materials instead of a curated set of government-endorsed messages.¹⁴⁷ Additionally, Iowa has not historically exercised significant control over the selection and removal of books from libraries, and the Supreme Court has directed the use of caution when extending the doctrine of government speech.¹⁴⁸

B. Fifth Circuit: Little v. Llano County

In August of 2021, three Llano County, Texas, residents, Rochelle Wells, Eva Carter, and Jo Ares, filed complaints with Judge Ron Cunningham, head of Llano County Commissioners Court, which supervises the Llano County library system, about children’s library books that they found to be “pornographic and overly sexual.”¹⁴⁹ Wells, in order to prevent others in the community from checking out books she deemed unacceptable, checked them out continuously for months. Cunningham then directed the librarian to remove these books from Llano County library shelves.¹⁵⁰

Several months later, another community member sent Cunningham a list of books that she found objectionable, calling them “pornographic filth.”¹⁵¹ Cunningham further instructed the librarian to remove these additional books. In total, the library removed seventeen books by the end of 2021, and these books became accessible only if an individual asked the librarian for them. However, the books were not listed in the catalog system, making it impossible to know if the books were available for checkout.¹⁵²

In 2022, the library board dissolved, and Cunningham appointed Wallace and Wells to the new board. The new board prohibited the Llano County librarian from attending the board’s meetings and required her to obtain approval for any new books she sought to include in the library.¹⁵³ Subsequently, plaintiffs brought suit in the United States District Court for the Western District of Texas and alleged that their First Amendment rights had been violated because the book removals occurred based on the defendants’ disagreement with the content of the books.

In March of 2023, Judge Robert Pittman of the United States District Court for the Western District of Texas issued a preliminary injunction, concluding that the plaintiffs were likely to succeed on the merits of their First Amendment claim based on viewpoint discrimination. The Court mandated the return of the book titles in question and prohibited further removal of books while the case proceeded.¹⁵⁴ The decision was initially affirmed by a 2-1 Fifth Circuit panel but was quickly vacated and ordered to be heard en banc.¹⁵⁵ Notably, the panel dissent endorsed the government speech argument.¹⁵⁶

Oral arguments before the en banc court took place in September of 2024 and opened with the lawyer for Llano County, Jonathan Mitchell, arguing that the Court should overturn Campbell v. St. Tammany Parish School Board, previously discussed in this Note’s Section I. Campbell held that the constitutionality of removing books from a library hinges on the substantial motivation of officials involved. Specifically, Campbell ruled that decisions driven by a desire to suppress certain viewpoints or ideas violate the First Amendment. This holding closely aligns with the plurality’s reasoning in Pico.

Mitchell and the County argued that library decisions to acquire, retain, or remove books are choices that should be categorized as government speech—and thus should be exempt from restrictions on viewpoint discrimination.¹⁵⁷ Drawing on Moody v. NetChoice LLC,¹⁵⁸ the County compared the library’s book collection to “speech that belongs to the curator,” asserting that the inclusion or exclusion of particular materials is expression.¹⁵⁹ They compared the content selected in libraries to a curated broadcast or a social media page, which is deemed expressive activity.¹⁶⁰ The defendants argued that by exercising discretion over library content, the government is communicating its own message about what is appropriate for the community it serves. Under this framework, the decision regarding the removal of books would be insulated from a First Amendment challenge.

Plaintiffs countered that holding library curation to be government speech would “recast government censorship as protected affirmative speech” and “expand the government’s power to extinguish controversial ideas.”¹⁶¹ They argued that the act of curating library books is not government speech because it does not meet the Supreme Court’s test from Shurtleff v. Boston. Specifically: (1) the county has not historically censored books based on viewpoint; (2) the public does not perceive the government to be speaking because the books provide for conflicting views; and (3) the county has not engaged in viewpoint curation for every book in the library.¹⁶²

The en banc panel overruled Campbell and a plurality held that library book curation is government speech in their final ruling on May 23, 2025.¹⁶³ On September 9, 2025, the attorneys for plaintiffs, now notably including Elizabeth Prelogar, the former Solicitor General of the United States, filed a petition for a writ of certiorari.¹⁶⁴

VII. BOOK BANNING CASES ARE HEADED TO THE SUPREME COURT

The Supreme Court should grant certiorari in Little because of the unresolved legal questions stemming from Pico, the steady rise in legislation targeting book removals, and the newly divided circuits on government speech. Pico established that school boards cannot remove books solely because they disagree with their ideas, but the fractured nature of the ruling left room for interpretation—especially regarding the distinction between school and public libraries. Now, with a circuit split on whether library book curation constitutes government speech, the need for Supreme Court clarification is even greater. The Fifth Circuit’s decision in Little directly contradicts the Eighth Circuit’s ruling in GLBT, which held that library curation is not government speech. If courts allow the government to claim that book removals are a form of government speech, it will fundamentally reshape public libraries by transforming them from spaces of free inquiry into spaces of government-endorsed messages. Given the implications for First Amendment rights and the increasing legal battles over library censorship, the Supreme Court will likely need to resolve these critical constitutional issues.

A. Decades-Old Nonbinding Supreme Court Plurality Opinion In Pico, Revisited

As previously discussed in Section I, Pico involved the removal of several books from a public school library, in which the school board argued that the books were “anti-American” and “vulgar.”¹⁶⁵ In its plurality opinion, the Court held that there was a genuine issue of material fact as to whether the school board’s removal of the books violated the First Amendment.¹⁶⁶ Justice Brennan, joined by Justice Marshall and Justice Stevens, held that the books could not be removed solely because of the government’s disagreement with the ideas contained in the books. A plurality of the Justices also reaffirmed that students have a right to receive information protected by the First Amendment, even in a school setting.¹⁶⁷ However, the Pico decision left room for discretion on the part of the schools, in that they could remove books deemed educationally unsuitable. As in, removal would be permissible if it was based on educational suitability rather than ideological disapproval.

The Justices differed on the legal standard for determining when a school could remove books. Justice Blackmun concurred in the judgment but did not agree that the right to receive information was a necessary consideration.¹⁶⁸ Justice White concurred in the judgment only and would have held that there was a genuine issue of material fact without identifying the viewpoint limitation on school boards.¹⁶⁹ As mentioned throughout this Note, the fractured nature of the Pico ruling—in addition to the rise in legislation on book banning—indicates that lower courts will now need clarification from the Supreme Court on the standards of book removal in school and public libraries.

B. The Supreme Court Has Not Yet Differentiated Between School Libraries And Public Libraries

School libraries and public libraries both play essential roles in supporting free speech, but they differ in scope when it comes to book removal challenges. Public libraries are community institutions that serve people of all ages and provide access to a vast array of viewpoints and information. In contrast, school libraries function within an educational setting, where materials are evaluated based on curricular goals and suitability for students. Because of this distinction, parents have a greater influence over school libraries as they have a recognized right to direct their children’s upbringing.¹⁷⁰ However, this parental right does not extend to controlling what is available in a public library, where individuals—children and adults alike—retain personal autonomy over what they choose to read.

Public libraries operate on the principle of voluntary access. Unlike schools, in which students may be required to engage with certain materials, public library users have complete discretion over what they may choose to check out. If a parent does not want their child reading a particular book, they can simply prevent their child from borrowing it without imposing that decision on the rest of the community. This opt-in nature of public libraries makes government-imposed book removals particularly troubling, as these removals do not merely reflect an effort to structure a child’s education, but also limit access to information for everyone in the community.

Pico narrowly addressed book removals in school libraries, but the differences between school and public libraries demand a more nuanced legal approach. The dissent in Pico expressed concerns about a school board’s ability to maintain control over the educational environment, but that is not relevant here.¹⁷¹ How might the dissent have ruled had this been a community library? The fact that Pico left these unresolved questions highlights the need for clearer guidance from the Supreme Court on how book removal standards apply to public libraries. Furthermore, the Supreme Court should uphold an even stronger determination against viewpoint discrimination than it did in Pico, given the dangers posed by an expanded understanding of the government speech doctrine, as discussed in Section D below.

C. Llano County’s Book Removals Are Viewpoint Discrimination

Although Pico held that books cannot be removed from school libraries solely because of disagreement with the ideas they contain, the case was remanded for further proceedings, and subsequent proceedings never determined whether the school board’s actions constituted viewpoint discrimination. After the remand, the school board faced pressure from 1,200 parents who petitioned to end the case and return the books to the library shelves.¹⁷² In response, the board reinstated the books and implemented a parental notification requirement for checking out these books. However, the Civil Liberties Union challenged this notification policy, arguing that it violated laws protecting confidentiality of library records.¹⁷³ Facing continued public pressure, with some parents even leaving the district over the issue, the school board eventually abandoned its efforts and restored all books to the shelves.¹⁷⁴

Book bans are unconstitutional when they are based on viewpoint discrimination.¹⁷⁵ The Supreme Court, however, has yet to articulate a clear standard for identifying when book removals cross the constitutional line. Supporters of book bans often claim that they are motivated by concerns about age appropriateness, but modern challenges overwhelmingly focus on specific themes and perspectives—particularly those related to race, gender, and sexuality¹⁷⁶—suggesting they are driven by ideological opposition rather than neutral educational concerns. For instance, in Georgia, a challenger criticized a book featuring a same-sex couple, stating, “Books like this [are] where teens get the idea that [same-sex relationships are] okay”¹⁷⁷—a clear expression of disagreement with the viewpoint that same-sex relationships are acceptable. Similarly, in Tennessee, a branch of Moms for Liberty opposed a book detailing the story of civil rights activist Ruby Bridges, written by Bridges herself, as teaching “anti-American values.”¹⁷⁸ In Idaho Falls, most of the twenty-one complaints that its public library received in 2023 were for books that dealt with LGBTQ+ characters or race.¹⁷⁹ LGBTQ+ books are frequently mischaracterized as being about sex, even when the content is entirely age-appropriate—such as the children’s picture book And Tango Makes Three, which tells the true story of two male penguins raising a chick together.¹⁸⁰

The rapid rise of book bans over the past several years is directly connected to the broader political efforts used to suppress discussions of race and gender. This trend is underscored by Donald Trump’s executive orders targeting diversity, equity, and inclusion (“DEI”) initiatives, which frequently incorporate books on these subjects. DEI programs are being vilified as promoting “antiwhite racism” and undermining “national unity.”¹⁸¹ Schools that receive federal funding are being threatened with withholding of those funds if they do not abandon their DEI programs.¹⁸² The 2025 DoD directive to remove books from DoD schools based on “gender” and “discriminatory equity” ideologies makes explicit the political-and viewpoint-based motivations behind the legislation targeting books in schools and libraries. The directive, echoing the Trump administration’s hostility toward DEI, is obviously rooted in content- and viewpoint-based discrimination. It is difficult to justify restricting access to children’s picture books about a girl with freckles or the life of Ruth Bader Ginsburg as anything but obvious partisan viewpoint censorship.

The same viewpoint-based targeting is evident in Little v. Llano County, making it an ideal case for the Supreme Court to clarify when book removals constitute unconstitutional viewpoint discrimination. The books removed from Llano County libraries overwhelmingly focus on LGBTQ+ identities and racial equity. County officials referred to these books as “pornographic filth” and “disgusting.” Similar to the school board defendants in Pico who used a list of books from an outside source to recommend books for removal because they disagreed with the contents, here, the defendants in Little created a list of books they “personally opposed,” including books that had been identified by an outside party as about “politics, race, sexuality, and gender.”¹⁸³ These books included Caste: The Origins of Our Discontents, They Called Themselves the KKK: The Birth of an American Terrorist Group, Being Jazz: My Life as a (Transgender) Teen, and Freakboy, among others.¹⁸⁴

Several factors in Little make clear that the book removals were based on viewpoint discrimination. First, the targeted books overwhelmingly dealt with themes that have been politically contentious, indicating the removals were not neutral but ideological. Second, officials ignored standard library weeding procedures and instead relied on personal opposition and lists of objectionable books obtained from outside sources to justify removals.¹⁸⁵ Third, internal emails show that librarians did not agree with several of the removals and thought they might be “illegal.”¹⁸⁶

The Supreme Court has long recognized that the government cannot suppress speech simply because it disagrees with its message. Yet, the rise of book bans driven by ideological opposition demonstrates the need for a more explicit and enforceable standard. Little may present the Court with an opportunity to provide that clarity and reaffirm that viewpoint-based censorship in public libraries is unconstitutional. The Court must make clear that the specific targeting of LGBTQ+ and racial equity themes is a constitutional violation, and that not all books containing these themes are inherently age-inappropriate.

D. The Dangerous Rise of the Government Speech Argument

An issue that was not raised in the Pico case, but remains a serious and dangerous argument gaining prominence, is that of government speech. Should the Supreme Court grant certiorari to Little, affirm the Fifth Circuit’s decision to overrule Campbell, and hold that library book selection is government speech, the outcome of Little could have profound consequences. Holding that library book curation is government speech would fundamentally distort the marketplace of ideas, which is a fundamental building block of our country’s democracy.

As mentioned, the Eighth Circuit held in GLBT that library book curation is not government speech. Therefore, the Fifth Circuit’s plurality holding in Little that library book curation is government speech and its overturning of Campbell, which closely aligns with the Pico holding, make it very likely that the Supreme Court will grant certiorari. In the original Fifth Circuit opinion that was vacated, dissenting Judge Stuart Kyle Duncan wrote, “There is a simple answer to the question posed by this case: [a] public library’s choice of some books for its collection, and its rejection of others, is government speech.”¹⁸⁷ This split with the Eighth Circuit and divergence from Pico, the latter of which has been adhered to across circuits, requires clarification from the Court.

Seventeen states signed onto an amicus brief on behalf of Llano County endorsing defendants’ argument that library book curation is government speech. The states wrote, “The county’s decisions over which books to offer its patrons in its public libraries, at its own expense, are its own speech.”¹⁸⁸ In April of 2024, Florida made the government speech argument in PEN America v. Escambia County School District, in which a district judge held that whether library curation decisions are government speech is an unsettled matter.¹⁸⁹ Florida, one of the states to sign the amicus in Little, traveled to the Fifth Circuit to make the argument for government speech after Llano County yielded some of its speaking time to Florida.¹⁹⁰ In a social media post, The Florida Freedom to Read Project alerted followers to “pay attention” to the Little case; it stated that Florida would keep presenting this argument “until they find a friendly court.”¹⁹¹

Legal scholars have raised concerns about the ambiguous scope of the government speech doctrine and its potential to undermine First Amendment rights.¹⁹² Cases such as Little and GLBT are prime examples illustrating these scholars’ justified fears. The consequences of states characterizing library book curation as government speech would turn spaces dedicated to the free exploration of ideas into “silos of partisanship.”¹⁹³ Government speech is particularly serious in this context because, in the case of libraries, there are few other places where individuals can access such a breadth of ideas freely. Libraries are physical manifestations of the marketplace of ideas, where people can go to find information on nearly any topic they might think of with little to no cost. Allowing the government to pick and choose what ideas line the shelves based on viewpoint is directly removing those ideas from the marketplace. For instance, while Rust v. Sullivan (previously discussed in Section B of this Note) allowed the government to restrict healthcare professionals from putting information about abortion into the marketplace, families could still turn to other sources—like libraries—to find information on abortion. For many members of the public, the library is the last stop to gather information on these politically restricted ideas.

Libraries are unique in that they offer a wide array of materials, often without any cost or significant barriers to access. Bookstores and online research can be expensive, and in many areas, libraries are the only option for communities to obtain information. For seniors or those who might not be tech-savvy, the library is a place where they can easily access ideas that may not be readily available elsewhere. Calling library book curation government speech is therefore particularly risky because it grants the state power to limit the flow of information in spaces dedicated to public knowledge, thereby undermining the core purpose of that space.¹⁹⁴

Further, library book curation is not government speech as analyzed by the test in Shurtleff. The Little plaintiff pointed out that Llano County has not historically engaged in viewpoint-based censorship and echoed GLBT in that the government “has not historically spoken by censoring books at public libraries.”¹⁹⁵ Libraries have historically stood in contrast to censorship, providing the public with free access to ideas.¹⁹⁶ Defendants, in response, focused on “curation,” which they argued has “always been the prerogative of government employees.”¹⁹⁷

The Little defendants’ argument is wrong. Curation does not equal government speech in the context of libraries. One aspect of a librarian’s job duties is to review the library collection. Many libraries, including those in Llano County, follow a standardized process of “Continuous Review, Evaluation and Weeding” (“CREW”).¹⁹⁸ Librarians are tasked with removing—weeding—books deemed outdated or duplicated. This removal is subject to neutral criteria, most typically by following “MUSTIE” factors.¹⁹⁹ MUSTIE stands for: (1) “misleading” (factually inaccurate); (2) “ugly” (worn out and unfixable); (3) “superseded” by a new edition (or better source); (4) “trivial” (of no literary or scientific merit); (5) “irrelevant” to the needs and interest of the community; and (6) “elsewhere” (material can be easily borrowed elsewhere).²⁰⁰

A book satisfying one factor alone would likely not be weeded; instead, it would require meeting a combination of these factors for removal. In fact, Llano county required two MUSTIE factors to be met before removal.²⁰¹ Libraries engage in MUSTIE practices to weed out certain library books, and of course, such weeding involves making some selection choices because libraries do not have unlimited physical space. However, MUSTIE relies on relevance, demand, and educational value to make removal decisions and include various (and likely opposing) viewpoints. These choices are not made to represent a government-endorsed message. As Little plaintiffs pointed out in their brief, “any regulation of speech could be reframed as the government” if choices about quality are considered government messages.²⁰²

Second, the public does not perceive the government to be presenting a message or speaking through the combination of books they are presented with in a library. Libraries are meant—and understood by the public—to facilitate access to a diverse range of perspectives, not to endorse or suppress ideas. As plaintiffs in Little argued, the public does not perceive the government to be endorsing the messages of each book in the library.²⁰³ They point to the Eighth Circuit decision in GLBT to highlight that libraries contain books with a “wide range of conflicting views”²⁰⁴ and that it is not possible for the public to believe the government is endorsing them all. Otherwise, the government would be “babbling prodigiously and incoherently.”²⁰⁵ Defendants in Little countered this argument by insisting that the issue in GLBT is not about the endorsement of the speech within the library books, but rather about the status of the person who selected them (i.e., a government employee or a private citizen).²⁰⁶ They argue that it is “inconceivable” that anyone would attribute this speech to the actions of a private citizen.²⁰⁷ While it is likely true that the public would view a librarian as acting in their scope of employment, rather than acting as a private citizen, defendants’ argument is dangerous and greatly expands the reach of the government speech doctrine, contrary to the U.S. Supreme Court’s warning in Matal v. Tam that government speech should be applied with caution.²⁰⁸

The mere fact that government employees are involved in a process does not automatically transform the process into one of government speech. For example, contrary to Little’s majority opinion, just because museum curators choose exhibits does not mean their actions are viewed and understood by the public as government messaging.²⁰⁹ Shifting the debate from the public perception of government endorsement of a message to that of the identity of the selector would give the government near immunity from First Amendment violations. When the government issues a press release or policy, it is widely recognized as an official message because it comes from a government employee and explicitly communicates the government’s position. Unlike an official government-issued report or press release, a library’s collection does not serve as a direct communication of any particular message.²¹⁰ Further, when a person checks out a book from the library, they do not view the book as a message from the government. Library patrons do not understand the selection of books in a library to be a message from the government saying, “we think this is a good book” or “we approve of this book.” While it is understood that the librarians have selected the book for inclusion in the library’s collection, the message is not the librarian’s personal endorsement of the viewpoints contained in the book, but rather of the ideas presented by the author. The library’s role is not to act as a representative for government messages, but to facilitate access to diverse ideas. Thus, the public sees the speech within the book as belonging to the author, not as a government-endorsed message.

Third, libraries do not extensively engage in viewpoint discrimination when selecting books for their library shelves. Little plaintiffs correctly argue that the weeding process that libraries engage in is based on a neutral set of criteria and is not an effort to remove books based on viewpoint.²¹¹ The plaintiffs analogized this work to that of maintenance, not intentional control of specific content.²¹² Again, here, the defendants argue that this is based on the public status of the librarian’s employment, stating that the selection and removal decisions “will always be shaped and controlled by . . . government employees.”²¹³ Defendants missed the point here. Per aforementioned reasons, there is no viewpoint discrimination involved in the weeding process and the point of weeding is not to shape a particular expression.

The Court should make it unequivocally clear that the curation of public library collections by professional librarians is not government speech. Librarians are trained professionals guided by ethical standards. Their role in selection and curation of books is fundamentally about serving the public’s needs by providing them with unrestricted access to diverse perspectives. Libraries follow the principles of the Library Bill of Rights, which was created in response to censorship. Librarians are not curating a government message; they are curating a space where individuals can “seek and receive information from all points of view without restriction.”²¹⁴

CONCLUSION

The First Amendment serves as a foundation of democracy, protecting access to information and the free exchange of ideas. Throughout the nation’s history, it has been vigorously defended as a means of fostering an informed citizenry. Democracy thrives when individuals can explore, challenge, and expand their understanding of the world around them through exposure to wide-ranging perspectives. This right is not reserved for adults alone—children, as future stewards of our society, are also entitled to these fundamental protections.

Community libraries embody the First Amendment’s principles in their purest form. Libraries are not merely collections of books, but institutions dedicated to providing neutral spaces where people of all backgrounds can engage with a variety of perspectives, free from political bias or government interference. Politicizing libraries by framing the curation of their collections as government speech would fundamentally undermine the heart of the marketplace of ideas exemplified and provided by libraries.

The Supreme Court’s plurality decision in Pico rightly recognized that students have a constitutional right to access information. However, given the increasing frequency of book bans and the evolving legal landscape, the Court must reaffirm and strengthen the Pico decision. Librarians are trained professionals who adhere to ethical standards, and their role in curating books is centered on ensuring unrestricted access to knowledge, not advancing political agendas. Treating library curation as government speech would open the door to political ideological control over library collections and erode public trust in government institutions.

Upholding the First Amendment’s protection over access to library books is essential to maintaining a society that values open inquiry and free expression. Libraries are not arenas for partisan agendas, but sanctuaries for knowledge and intellectual growth. As the Foundation for Individual Rights and Expression put it, “Regardless of whether book banning campaigns target the Bible or Judy Blume, politicized efforts to restrict access to information cannot be reconciled with the Founders’ faith in the free exchange of ideas and our national commitment to freedom of expression.”²¹⁵

Editor’s Note:

After publication of this issue, the Supreme Court denied certiorari in Little v. Llano County, making it easier for government officials in Texas, Louisiana, and Mississippi to engage in viewpoint discrimination with respect to the removal of books from public libraries.

99 S. Cal. L. Rev. 203

Download

Kelcey Sholl*

*Editor-in-Chief, Southern California Law Review, Volume 99; J.D. Candidate 2026, University of Southern California Gould School of Law; M.A. 2016, California State University, Northridge; B.A. 2013, California State University, Long Beach. I thank my advisor, Professor Rebecca Brown, for her support and guidance; Professor Rebecca Brown and Professor Lee Epstein for the opportunity to serve as a research assistant, which sparked my interest in the First Amendment; and the editors of the Southern California Law Review for their hard work and thoughtful suggestions on this Note.

Artificial Incompetence? Unpacking AI’s Shortcomings in Contract Drafting and Negotiation

January 2026January 2026Carly Snell*

INTRODUCTION

This Note was inspired by my time as a data center procurement contracts intern during the summer after my first year of law school. In this role, I assisted contract analysts and attorneys with their procurement of space in data center facilities by contracting with data center suppliers. I regularly reviewed contract redlines from suppliers, identified non-market or disadvantageous terms in those contracts, and suggested changes for the next “turn of the redlines,” or when the company would return the contract to the supplier with new edits to the document. An impactful conversation with my manager about artificial intelligence’s potential as a useful tool in a transactional lawyer’s toolbelt inspired a deeper dive into the benefits and drawbacks of applying artificial intelligence (“AI”) to the contract drafting, redlining, and negotiation space—ultimately leading to the development of this Note.

After the internship concluded, I began my second year of law school. While the most noticeable change upon my return was that I was no longer a first-year student, I also immediately observed a greater emphasis on AI in legal education than before. My law school offered a course on AI’s legal applications, peers used AI to supplement their studies, and professors emphasized the importance of mastering AI during law school, as it would be an essential tool in future legal practice. Similarly, students at other law schools honed their negotiation skills against AI chatbots¹ and even developed their own AI-driven case briefing technology.²

As with the implementation of any new technology, however, there are some points of contention that arise when applying AI to the law—especially in the context of contract drafting, formation, and negotiation. This Note covers four main challenges to applying AI to contract drafting: (1) contract law principles, (2) equity concerns, (3) accuracy issues, and (4) legal profession challenges. Additionally, this Note presents the results of a novel empirical study designed to test AI technology’s tendency to discriminate when tasked with negotiating a contract on behalf of different types of clients. Interestingly, ChatGPT, a popular AI chatbot,³ appears to favor corporations and nonprofit organizations over individuals when acting as a negotiation assistant.⁴ This finding suggests that the excitement surrounding AI’s potential uses in the legal field⁵ is premature, and professionals should hesitate to implement this technology in contract drafting and negotiation until algorithmic discrimination is adequately addressed.

Part I of this Note introduces the historical development of AI technology and its rise to stardom that began with the public release of ChatGPT in 2022.⁶ Part I then describes early applications of AI technology to the contracting space, such as Spellbook, Harvey, and LegalSifter.⁷ After that, Part I discusses fundamental contract law principles, such as mutual and constructive assent, that AI contract drafting may not readily align with.⁸ Finally, Part I concludes by orienting the reader with basic legal profession concepts, such as the lawyer’s duties of confidentiality, communication, competence, and diligence.⁹

Part II introduces several illustrative examples of AI in contract drafting and negotiation that pose unique questions about the key differences between human and AI-driven contracting. These differences make it difficult to apply existing contract law to AI and raise important concerns about AI’s potential to discriminate when contracting and negotiating on behalf of different clients.¹⁰ Part III of this Note expands upon AI’s usurpation of traditional contract law principles. Fundamental contract law concepts, such as the “meeting of the minds” required to form a valid contract, do not readily apply to wholly AI-driven contracting.¹¹ Principally, AI’s application in contract drafting and negotiation can present novel complications when determining whether or not the parties to a contract mutually agree on its terms. These issues persist regardless of whether a party performs some of its obligations under an AI-driven contract and despite the controversial doctrine of constructive assent.

Part IV covers the equity concerns that arise when applying AI technology to contracting. In general, applications of AI technology in the contracting space raise concerns about “algorithmic discrimination”—AI’s tendency to produce discriminatory outputs as a consequence of being trained on tainted data.¹² AI in contracting also raises ethical issues regarding enforcement of fully automated contracts. A pervasive issue in the AI space is ensuring proper alignment between an AI model’s goals and those of its operator.¹³ Given that AI technology regularly suffers from misalignment problems, would it be ethical and equitable to enforce contracts drafted by these models? Another ethical dilemma that arises in the AI contracting context concerns legal liability and accountability if a party is injured by an AI-formulated contract. If harm results from an AI-drafted contract, who should be held accountable for these harms? Between the AI model itself, its designer, its user, and other parties, there is no readily apparent answer. Finally, the implementation of AI in contracting—a setting that involves a plethora of sensitive information—presents serious data privacy and security concerns.¹⁴

In Part V, this Note reviews the accuracy issues apparent in current and potential applications of AI technology. Simply put, AI technology can behave unpredictably and output inaccurate results known as “hallucinations.”¹⁵ In the litigation context, several lawyers, including Michael Cohen’s attorney, have recently been sanctioned or publicly admonished for citing fabricated cases generated by ChatGPT in their filings.¹⁶ In the contracting space, in which exact language and minor details can govern the legal meaning of an agreement, AI’s tendency to hallucinate can cause major problems.

Part VI presents the challenges to the legal profession that arise when using AI technology in contract drafting and negotiation. For example, overreliance on AI technology to draft and negotiate contracts may violate an attorney’s professional duties of competence and diligence—much like the actions of the lawyers who cited fabricated cases in their court filings. Overreliance may also violate an attorney’s professional duty of communication if they cannot explain their reasoning for a recommended course of action to a client due to reliance on ChatGPT in their decision-making. Additionally, since AI models operate as “black boxes,” their use may raise concerns about duty of confidentiality violations if client information is input into these systems without proper safeguards.¹⁷

Part VII discusses the empirical findings that resulted when the author “hired” ChatGPT to assist various types of fictitious clients with negotiating a standard commercial real estate lease. These research findings suggest that ChatGPT discriminates against individual clients by tending to recommend renegotiation less often and to a smaller degree when advising individual clients than when assisting corporate or nonprofit clients. These findings have immense equity implications for contract drafting and negotiation in an AI-driven world, as AI models that disfavor individual clients may exacerbate existing market power or resource inequalities between individuals and more sophisticated corporate or nonprofit clients.¹⁸ Finally, Part VIII discusses some strengths and potentially useful applications of AI technology in legal work in light of this Note’s theoretical discussion and empirical findings. Part VIII posits that, although AI technology excels at summarization,¹⁹ concerns about its ability to exercise discretion and judgment suggest that it may be best suited for administrative tasks.

I. A CRASH COURSE IN AI AND RELEVANT LEGAL THOUGHT

A. What Is Artificial Intelligence and How Can It Contract?

There is no widely accepted definition of what constitutes artificial intelligence, which is partially a byproduct of how technological capabilities have rapidly improved in recent years.²⁰ To oversimplify, computer programs were historically classified as artificial intelligence if they successfully mimicked human rational thought.²¹ An early example of this concept is the Turing test for artificial intelligence, which was developed by the “father of modern computer science,” mathematician Alan Turing.²² The Turing test assesses how well a machine can imitate human thought and behavior via a competition that Turing called the “Imitation Game.”²³ In the game, a machine and human compete by answering questions asked by a human interrogator; at the end of the game, the interrogator must identify which competitor is a human and which is a machine.²⁴ If the interrogator gets it wrong—i.e., says that the machine is the human—then the machine is thought to demonstrate human-level thought and intelligence.²⁵

This Note utilizes a relatively expansive definition of artificial intelligence that is reminiscent of the Turing test. For the purposes of this Note, artificial intelligence is any computer software program that demonstrates human-like behavior or intelligence. As discussed below, the focal point of artificial intelligence in this Note is large language models, which are some of the best modern examples of AI that would likely pass Turing’s test for artificial intelligence, given their language-based design and applications.²⁶

Artificial Intelligence’s Rise to Prominence: The “AI Boom”²⁷

Artificial intelligence has taken the public consciousness by storm since the release of ChatGPT, OpenAI’s text-generating chatbot, in November 2022.²⁸ ChatGPT is an AI model trained to engage in natural language conversations, which means that when users interact with ChatGPT, it converses with them by generating textual responses comparable to that of a human.²⁹ The model’s successful imitation of human-sounding speech captured the public’s imagination,³⁰ prompting increased interest in potential applications of AI technologies from the general public³¹ and software developers³² alike.

ChatGPT can complete a variety of academic tasks in a matter of seconds, such as writing essays, generating ideas, and answering mathematical problems.³³ It is no surprise, then, that students from primary school to collegiate grade levels were some of the model’s most prevalent initial users, asking ChatGPT to write papers and complete homework assignments on their behalf.³⁴ Students’ widespread use of ChatGPT to complete assignments led many schools and universities to initially ban the AI model altogether,³⁵ although it was difficult, if not impossible, to enforce AI bans—especially outside of the classroom.³⁶ A new industry of tools meant to detect the use of AI in students’ writing emerged to combat this issue, but their accuracy remains widely disputed.³⁷

Although initial widespread applications of ChatGPT were somewhat rudimentary in nature, such as students’ use of the tool to complete assignments,³⁸ OpenAI’s introduction of the model to the public sphere was instrumental in prompting other AI developers to invest in the creation and public release of their own large language models (“LLMs”).³⁹ After witnessing OpenAI’s successful launch of ChatGPT, prominent tech industry leaders such as Google and Meta immediately sought to turn AI technologies into tangible, profitable products that they could sell to individuals and companies.⁴⁰ Although these major technology companies had already been developing (and, in some cases, even released, to little success⁴¹) their own AI technologies before November 2022, ChatGPT’s successful public launch prompted an expansion of the AI industry like never before.⁴² By the following spring, a flurry of new LLMs had emerged on the market: Meta’s LLaMA model, Google’s PaLM-E, and even OpenAI’s newest iteration of its LLM: GPT-4.⁴³

In essence, large language models are AI models designed to interact with and produce language.⁴⁴ “Large” refers to the increasing trend to train these models on large quantities of data stored in massive data sets that are usually housed in collocated data centers.⁴⁵ While ChatGPT, LLaMA, PaLM-E, and GPT-4 are all generally considered LLMs, much like AI more broadly, a concrete definition of what constitutes a large language model remains an open question.⁴⁶ There are no exact parameters for how large an AI model must be or how it must interact with language in order to be categorized as an LLM.⁴⁷

On the other hand, LLMs are generally considered to be a subset of generative AI.⁴⁸ Generative AI is defined as artificial intelligence capable of producing new creations, such as graphic images, text, and audio, based on training data inputted into the model.⁴⁹ Therefore, generative AI enables a user to generate substantial quantities of work product with minimal effort by prompting a generative AI model and letting it create content for them based on the query. This is partly why ChatGPT became wildly popular in a short period of time⁵⁰—and why the generative model caused concerns about students using it to complete homework and other assignments on their behalf.

Beyond their avocational applications as homework helpers⁵¹ and joke writers,⁵² LLMs are being increasingly used by industry professionals to improve and expand the potential of their products and services.⁵³ For instance, Amazon Web Services implemented an externally facing AI chatbot on its Amazon.com retail site designed to handle returns, provide shipment tracking information, and generally improve the site’s customer service capabilities⁵⁴ (albeit the chatbot has garnered mixed reviews⁵⁵). Similarly, in 2024, Target Corporation launched an internally facing generative AI model, called Store Companion, to assist with employee training, store operations management, and general problem-solving tasks.⁵⁶ Meanwhile, social media platforms such as Instagram use AI models to filter content and craft feeds that are better personalized to users’ individual preferences.⁵⁷

Early Applications of Artificial Intelligence to Legal Contracting

Naturally, the ever-increasing implementation of LLMs in a variety of businesses, industries, and settings includes applications in the legal field as well.⁵⁸ For example, AI has already been used to create legal workflow companions with suites of legal skills,⁵⁹ contract lifecycle management software programs,⁶⁰ and contract redlining and drafting assistants.⁶¹ A simple Google search for AI contracting services yields a plethora of (interestingly named) AI-powered software programs that purport to assist an attorney with redlining (e.g., Harvey,⁶² Lawgeex,⁶³ Superlegal,⁶⁴ Ivo,⁶⁵ Screens,⁶⁶ and Spellbook⁶⁷) or managing (e.g., Evisort,⁶⁸ Ironclad,⁶⁹ Sirion,⁷⁰ and LegalSifter⁷¹) their legal contracts. Even companies that operate widely used legal research databases, such as LexisNexis and Thomson Reuters, have created and marketed their own generative AI-powered legal assistants.⁷²

Legal professionals are generally excited about new and potential future applications of AI to the legal world.⁷³ Many believe the technology will increase efficiency in a time-intensive industry by synthesizing documents and reducing the time a human attorney needs in order to perform certain legal tasks.⁷⁴ Some hopefuls even view AI as infallible—capable of outperforming humans, whose work is prone to errors, because AI can craft perfectly completed and accurate work product.⁷⁵ Finally, AI is thought by some to make legal services more affordable and accessible to the general public⁷⁶ by reducing the number of billable hours an attorney must dedicate to any given task, enabling individuals to access legal services without hiring a human attorney, or both. In fact, Utah and Arizona have already implemented pilot programs that allow non-lawyer entities, such as AI chatbots, to provide legal services, and Washington may be the next state to institute such a program.⁷⁷

Despite this enthusiasm about AI, the immediate application of LLMs to the legal space has not been without its challenges. Some attorneys have wrongfully used LLMs to shirk their responsibilities by asking AI models to conduct legal research or write briefs on their behalf.⁷⁸ This practice has resulted in massive sanctions and fines for attorneys who cited “bogus” cases that were fabricated by ChatGPT in documents that they later submitted to a judge.⁷⁹ Furthermore, as discussed later in this Note, issues regarding lawyers’ ethical and professional duties, algorithmic discrimination, AI’s inaccuracies, and the subversion of traditional contract law principles also arise when large language models are applied to the legal field.

B. A “Meeting of the Minds” Regarding Contract Law Theory

An orientation into the foundational principles underlying contract law theory is needed before one can take a proper deep dive into the applications of AI in contracting. A great place to start is the traditional contractual theory of mutual assent, colloquially known as the “meeting of the minds.”⁸⁰ Mutual assent is one of many requirements that must be demonstrated for a court to hold that a given contract is legally valid and enforceable.⁸¹ “Meeting of the minds” refers to the idea that both parties must mutually agree to the terms of a contract in order for the agreement to be legally binding.⁸² That is, the parties’ minds must, in a sense, “meet in the middle” at the moment when the contract is formed. For that reason, mutual assent may not be found when one or both of the parties to a contract entered into the agreement based on a misunderstanding or a mistake of law or fact.⁸³ Intuitively, this makes sense; it would not be good public policy to bind people to a contractual agreement if they did not fully understand the obligations and consequences they allegedly agreed to when the agreement was executed. Beyond equity justifications, it may also be inefficient to hold a party accountable for obligations that they did not intend to undertake and may not be equipped to fulfill. Relatedly, to create a binding agreement, the parties to the contract must specifically mutually assent to the material terms of the contract.⁸⁴ Without a “meeting of the minds” between the parties to any given contract regarding the essential provisions of the agreement, the contract is invalid and not legally binding on the parties.

In some instances, courts have imputed assent to a party based on their conduct even if they did not explicitly agree to or approve of the terms of an agreement.⁸⁵ This doctrine is known as “constructive assent,”⁸⁶ and it is common among online transactions.⁸⁷ For example, if a user of an online webpage affirmatively acknowledges the page’s terms of use by clicking an “I accept” or “I agree” button without actually reading the agreement, the user is usually found to have constructively assented to the terms of the agreement despite not actually being aware of its contents.⁸⁸

Although many people make light of the fact that nobody ever reads various websites’ terms of use or, more notably, Apple’s Terms and Conditions,⁸⁹ constructive assent is no laughing matter. In these types of situations, constructive assent can be used to essentially waive the traditional contract theory requirement of a “meeting of the minds,” instead holding individuals accountable for the contracts that they sign even if they do not fully understand or have knowledge of the terms that they allegedly agreed to.⁹⁰ Unsurprisingly, the doctrine of constructive assent is controversial—especially its application to consumer contracts⁹¹ and form contracts more broadly.⁹² Further, the ethics of constructive assent are hotly debated among scholars, with some arguing that applying constructive assent to a contested contract unfairly disadvantages the weaker party (e.g., the consumer) to the benefit of the dominant party (e.g., the retailer) whose greater market power enables them to force the weaker party to consent to the dominant party’s preferred terms.⁹³

C. Attorneys as Ethical and Professional Fiduciaries

Another important factor to consider when analyzing the potential applications of AI to the contracting space is the ethical and professional complications that arise due to attorneys’ special fiduciary duties to their clients. In general, attorneys are held to a higher standard than those who work in many other professions.⁹⁴ Specifically, attorney conduct is governed by each state’s bar association, many of which have adopted the Model Rules of Professional Conduct—the generic rules promulgated by the American Bar Association.⁹⁵ The Model Rules serve as a fundamental guideline for attorney conduct by prescribing various professional and fiduciary duties to attorneys, such as client confidentiality, competence, diligence, and communication.⁹⁶ The Model Rules also address various topics relating to an attorney’s practice—like conflicts of interest, the formation of an attorney-client relationship, the scope of one’s representation, and how to interact with unrepresented persons⁹⁷—and explain how model attorneys should approach these issues. Importantly, the Model Rules detail practices that constitute misconduct, like engaging in dishonesty or fraud, violating the Model Rules of Professional Conduct, or committing a criminal act.⁹⁸ For the purposes of this Note, it is important for one to keep the Model Rules of Professional Conduct in mind when considering how an attorney may use AI technology in drafting or negotiating contracts, as certain applications of AI may subvert the underlying goals that the Model Rules were designed to support in more traditional applications.

II. ILLUSTRATIVE EXAMPLES

Several ethical, practical, and theoretical questions arise when one considers various applications of AI to contract drafting, formation, and negotiation. To better illustrate the issues that arise from applying AI to the contracting space, consider the following numbered examples and the questions they raise regarding their implications for the contract law principles and legal profession concepts that we have discussed:

Example #1: Laypeople Using AI to Draft a Contract⁹⁹

Two laypeople (i.e., not attorneys) are doing business together. Interested in summarizing their deal in a written form, they “draft” a contract by asking ChatGPT to do so for them. Once ChatGPT has drafted the contract, the two parties both read and sign the contract, despite not understanding the agreement’s legalese or terms. Later, something goes wrong, and the contract’s validity and enforceability are disputed.

Was there a “meeting of the minds,” or mutual assent, here?

Is this a case of AI-assisted human contracting, or was this effectively an entirely AI-created contract?

Is the contract enforceable?

Should society want the contract to be enforceable?

Example #2: AI as a Contract Drafting Tool for Attorneys¹⁰⁰

As is industry practice, a lawyer in a corporate law firm normally uses a standard form contract from prior deals as a starting point when drafting new contracts. However, for a particular deal, she decides to use ChatGPT to draft the initial form contract instead.

Is this an example of AI as a tool that assists humans in contract drafting, or is this a wholly AI-drafted agreement?

Does this distinction have important implications for the contract’s validity and enforceability?

Is there any significant difference between this attorney using AI to create a form contract or pulling a precedent contract out of her firm’s database?

Would this amount to a breach of the attorney’s professional duties of competence, diligence, or anything else?

Example #3: Human Error Versus AI-Drafted Terms

Overwhelmed with his busy workload, a lawyer mistakenly inserts a clause in a contract he is drafting for his client. Both his client and the other party to the contract sign the agreement; neither party nor the attorney knows at the time the agreement is executed that the accidental provision is included in the contract.

Is the extra provision in the agreement enforceable (i.e., did the parties mutually assent to the term)?

Is this scenario any different from if AI completely drafts and executes a contract without humans involved in the contracting process?

How are these two examples reconciled in terms of mutual assent? Are they the same, or fundamentally different in any way?

Example #4: AI Automatically “Agreeing” to Online Terms

Annoyed with websites’ many Terms of Service and Cookies pop-ups, an inventor creates an AI-driven “ad blocker” software that automatically clicks through and “agrees” to these pop-ups on the software user’s behalf so that they never have to see them again.

Would this constitute the user’s assent to various websites’ Terms of Service?

Does the answer to this question depend on how long the user has had the software, or whether they knew or reasonably should have known that specific websites had Terms of Service or Cookies pop-ups?

* * *

There are two possibilities when applying AI technology to contract drafting and negotiation: (1) AI effectively functions as an assistant, aiding humans with their contracting, and (2) fully automated decision-making, in which AI completely takes over contracting, from start to finish, with no humans involved in the process. Under either scenario, four categories of problems arise when implementing AI in contract drafting and negotiation: the subversion of contract law principles, equity concerns, accuracy issues, and legal profession challenges.

III. AI’S SUBVERSION OF CONTRACT LAW PRINCIPLES

If AI functions as a mere contract drafting and negotiation assistant, mutual assent concepts would apply in the same manner that they do for purely human-conducted contracting. An underlying principle of the mutual assent requirement for a valid contract is the notion that the parties to a given contract must understand the terms of the agreement and have a “meeting of the minds,” or mutual agreement, that they find the terms acceptable.¹⁰¹ If AI technology merely assists an attorney with drafting or negotiating a contract, this does not affect the portion of the dealmaking process that mutual assent concerns. The only point in time that is relevant for mutual assent is when the parties come to a consensus that the contract’s terms are agreeable and subsequently execute the agreement.¹⁰² By that point in time, the drafting and negotiating phases of the process are complete (and, truthfully, long gone)—the agreement is in its final drafted form and will not undergo further redlines or revisions. Thus, the implementation of AI as a mere assistant in the contracting and negotiation process is not within the timeline or contextual scope that mutual assent concerns. AI’s use as a contracting assistant is therefore akin to any personal opinions the drafting attorney may have (outside of their thoughts and duties as a fiduciary of their client) regarding the deal at hand—i.e., irrelevant to questions about mutual assent.

While some may argue that the cyclical drafting, redlining, and negotiation process drives the parties to a contract toward the ultimate goal of mutual assent at the end of the contracting cycle, it is not a necessary component of mutual assent that agreements are modified and negotiated by the parties. If one party presents a complete agreement to another party, who signs it without criticizing its contents or insisting on revisions, it is still a valid contract. Furthermore, in many instances, an attorney drafts and negotiates on behalf of their client, who signs the final contract without a comprehensive legal understanding of the negotiations and redlines that were made during the dealmaking process. This is arguably like Example #1 in Part II, in which the two laypeople used AI to draft a contract that they then signed. Although the individuals did not negotiate between themselves, mutual assent was arguably satisfied because the humans—not ChatGPT—assented to the agreement at the end of the contracting process.

On the other hand, if contracting is entirely managed by AI—without humans involved in the process—then the contract law requirement of mutual assent is not satisfied. Arguably, if the laypeople in Example #1 did not understand the contract because ChatGPT performed a substantial portion of the legal lift for them (which is possible, considering that they did not understand the AI-drafted agreement’s legalese or terms), then the mutual assent requirement may not be satisfied because the contracting process was effectively completed without human involvement. Example #4 details a more abstract example of this concept. In Example #4, the inventor’s software “agrees” to websites’ terms of use on its users’ behalf. In this situation, the human user never sees, let alone reads, the terms of service that they allegedly agreed to through the AI-driven software. Although some might argue that there is mutual assent because a person who installs the software knows that it will “agree to” the terms on any site that the person visits, this argument does not hold up to pragmatic scrutiny. Given how often and extensively people surf the Internet, it is highly likely that, over time, the person would not know which websites had pop-up advertisements or terms of use that the AI bot “agreed” to on their behalf, let alone the content of those agreements.

Therefore, the contract law requirement of mutual assent goes unsatisfied when AI fully takes over the contracting process. This flaw in solely AI-executed contracting becomes even more apparent when considering contracts that involve multimillion- or multibillion-dollar transactions, fundamental changes in a company’s structure or dealings, or changing the client’s financial or business practices in any substantial way. Without providing notice of these changes to the client and securing their informed assent to new and material contractual terms, solely AI-driven contracting is unlikely to satisfy traditional contract law principles.

Some might argue that a party’s performance of its obligations under a fully AI-driven contract would justify its validity and waive the mutual assent requirement, much like the traditional contract law enforcement principles surrounding the Statute of Frauds.¹⁰³ However, a fully automated contracting process differs from classic applications of the Statute of Frauds—such as when a party denies a prior verbal agreement, claiming that they never agreed to the deal because no written proof of it exists.¹⁰⁴ Rather, if AI completely drives the contracting process, then the parties to a contract would likely never be aware of, let alone read, the AI-drafted and executed agreement. Due to this disconnection, it is highly unlikely that the parties would completely perform their obligations under the agreement—simply because they would not know what their obligations are. Even if the parties were generally aware of their performance obligations (e.g., because the AI model contracted an extension of an existing purchase agreement between a purchaser and supplier), they would still not know the specifications of the agreement to a high enough degree for public policy to justify holding them to the transaction.

Furthermore, although some may argue that the doctrine of constructive assent can waive the mutual assent requirement in the purely AI-driven contracting setting, this argument is specious. Constructive assent is a highly controversial doctrine in its current limited uses, such as form contracts.¹⁰⁵ Scholars have raised particular concerns about constructive assent eliminating the need for mutual assent in online transactions, such as clickwrap agreements,¹⁰⁶ because the doctrine can infer an Internet user’s assent from their decision to click “I agree”—regardless of how “ill-informed and not well considered” that decision might have been.¹⁰⁷ Therefore, because constructive assent is thought by many to subvert traditional contract law theory, especially in online transactions, it provides a weak justification for waiving the mutual assent requirement in a purely AI-driven contracting setting.

Therefore, the distinction between AI as a contracting assistant and wholly AI-driven contracting carries significant contract law implications. In Example #2 in Part II, the legal difference between an attorney using a precedent contract from prior deals and relying on an AI-generated form contract is crucial, even though practicing attorneys may see little to no practical difference between the two. As AI technology continues to advance, the line between human-driven and AI-driven contracting will increasingly blur, raising questions about contract validity, enforceability, and an attorney’s professional obligations. Whether AI serves merely as a drafting tool or takes on a more autonomous role could have far-reaching legal consequences.

IV. EQUITY CONCERNS

A. Algorithmic Discrimination

Algorithmic discrimination occurs when ostensibly impartial AI technology produces discriminatory results because it was trained on tainted inputs.¹⁰⁸ Put more simply, algorithmic discrimination is a perfect example of “Garbage In, Garbage Out.”¹⁰⁹ Proponents of AI argue that even if algorithmic discrimination occurs, automated decision-making is preferable to human decision-making because humans are biased.¹¹⁰ However, algorithmic discrimination can perpetuate and amplify existing biases or stereotypes in an AI model’s training data, with the dangerous added implication that the tainted model appears facially objective and neutral.¹¹¹ Furthermore, because of their reliance on human inputs, algorithms will arguably never be fully bias-free and nondiscriminatory, but perpetually flawed as “partially human.”¹¹² Additionally, due to its highly advanced pattern-detection abilities, AI technology has the potential to develop new forms of discrimination by extracting patterns from its inputted data that humans alone would not have been able to detect.¹¹³

Algorithmic discrimination is also concerning because current legal theories do not supply satisfactory remedies for discrimination by AI systems.¹¹⁴ For example, imagine that an online job search site, such as LinkedIn, uses an AI-driven algorithm to “match” employers with potential interview candidates on the site by recommending certain user profiles to employers.¹¹⁵ If a user believed that the algorithm discriminated against them in choosing not to suggest their profile to employers, they would have limited options to seek legal redress. In the employment space, discrimination claims are separated into two categories: (1) disparate treatment and (2) disparate impact.¹¹⁶ Disparate treatment is focused on combating explicit discrimination, which requires a finding of intent.¹¹⁷ In a traditional, non-AI setting, explicit discrimination may be demonstrated by a qualified job candidate who was denied employment by a firm that refused to hire her by proving that the refusal was based on one of the candidate’s protected characteristics, such as race or gender.¹¹⁸ Conversely, to claim disparate treatment in the case of an AI algorithm, the disgruntled LinkedIn user would have to demonstrate that the algorithm had the intent to discriminate, which may be incredibly difficult, if not impossible, to prove in the case of a nonhuman entity. Thus, algorithmic discrimination is likely thought to be a product of unintentional or incidental discrimination.

Alternatively, disparate impact claims do not require the plaintiff to prove discriminatory intent;¹¹⁹ rather, the doctrine considers whether there is a disparate impact on members of a protected class, any business necessity for the impact, and a less discriminatory alternative means of achieving the same result.¹²⁰ Therefore, given the aforementioned difficulty of ascribing any particular cognitive motivations to an AI model, disparate impact discrimination is the only potential mode of existing discrimination law that

might provide legal redress for members of protected classes who experience algorithmic discrimination in the employment context.

In the contracting space, algorithmic discrimination has the potential to create disastrous consequences. If an AI model is trained on discriminatory data or its algorithm is improperly weighted by its human developers, it may tend to favor one type of party over another, such as men over women.¹²¹ This bias may then prompt the AI model to negotiate more favorable deals for certain parties than it would for others. This potential for AI to act as a discriminatory advocate may exacerbate existing inequalities, especially if the model’s reliance on tainted training data causes it to reinforce biases that disproportionately harm certain groups. Particularly sensitive communities include women, racial or ethnic minorities, and people who are socioeconomically disadvantaged. In the contracting setting, where every word in a contract has an important implication for the meaning of the agreement, a tainted AI model could selectively include unfavorable terms—or simply choose terms that are not the most favorable—in an agreement when “hired” by a party that the model’s data disfavors. The individual who experiences discrimination by receiving the “short end of the stick,” or undesirable contract terms, would likely never know that they were discriminated against by the model they used to contract. Even if the disadvantaged individual later became aware of the discriminatory term selection, it is likely that they would not have the ability or resources to advocate for themselves.

Furthermore, the contracting setting presents a multitude of consequential and important situations in which a person’s livelihood depends on the degree of favorability they are able to negotiate for themselves in a given contract. For example, in an employment contract, the starting salary, amount of paid family leave, and inclusion of any noncompete provisions may have huge implications for a prospective employee’s financial stability and future wellbeing. If an AI model poorly negotiates on a potential employee’s behalf, that potential employee may experience a lower quality of life than they would have otherwise—and if the reason for AI’s poor performance is discriminatory conduct, these disadvantaged outcomes will only exacerbate existing inequalities in our society.

B. Ethics of Enforcing Automated Deals

Another serious concern that arises when using AI in contracting is the ethical dilemma of deciding when to enforce completely automated deals. If we get to the point in which contracting is an entirely AI-driven task, do we feel comfortable holding humans accountable for the deals that an AI model entered into on their behalf?

A critical consideration when determining accountability in this circumstance is AI (mis)alignment. Broadly speaking, direct alignment refers to the ability to program an AI system so that it pursues goals consistent with the goals of its operator.¹²² There are a plethora of difficulties in ensuring proper direct alignment, including (1) determining the operator’s goals, (2) conveying those goals to the AI software, and (3) getting the AI model to correctly translate those goals into actions.¹²³ It is often incredibly difficult for an AI user to overcome these challenges, and efforts to do so sometimes cause AI programs to take unexpected actions that result in adverse consequences.¹²⁴

In the contracting context, holding the user of an AI contracting software to an agreement that the AI model drafted on their behalf can have especially inequitable consequences. Much like Example #3 in Part II, in which the human attorney mistakenly added language to the contract he was drafting, if an AI program is misaligned with its user’s goals, then it may draft contracts that do not reflect those goals. Both general intuition and contract law theory suggest that in a scenario like Example #3, the parties to the contract should not be bound by terms to which they did not assent. Similarly, in the case of misaligned AI contracting software, intuition suggests that it would be unethical to bind a party to an agreement if the AI model that contracted on their behalf did so in a manner that did not align with the user’s intentions.

C. Who Is Liable or Accountable?

If and when AI-assisted or wholly automated contracting goes wrong, who should we hold liable for breached contracts? Would we want to differentiate between the AI developer, the human who “hired” the AI to contract on their behalf or otherwise used the model to contract, and the AI model itself?

These questions are especially difficult to answer because traditional liability frameworks are designed with an inherent assumption that a human decisionmaker caused the alleged harm.¹²⁵ In the contracting setting, we would hold this human decisionmaker accountable for their breach of a contractual promise. If AI functions as a contracting agent, however, a human may not have made decisions that directly caused the complaining party’s harm. If an AI contracting program enters into agreements on a human’s behalf, that may not be enough under traditional liability frameworks to justifiably say that the human caused the alleged harm and hold them liable for it.

For similar reasons, it also appears unreasonable to hold an AI developer liable for breaches of contracts that its AI contracting software simply aided in drafting. To oversimplify, in order to prove causation of harm due to a breached contract, a plaintiff must demonstrate that the defendant’s breach was more than just an actual cause of the plaintiff’s harm.¹²⁶ Rather, the plaintiff has a higher burden: they must prove that the defendant’s act was the proximate cause of their harm.¹²⁷ To demonstrate proximate cause, the plaintiff must show that the harm was a foreseeable consequence of the defendant’s breach of contract.¹²⁸ In the AI context, a developer and its AI software may be actual, or but-for, causes of the harm suffered by a party who contracts with the software. However, the broad applicability of AI contracting software and its limitless potential uses suggest that, in many cases, the developer’s creation of the software would not be the legal, or proximate, cause of the injury because the alleged harm was not foreseeable.

Given these uncertainties about holding either the user or developer of AI-driven contracting software accountable, a plaintiff’s final potential avenue in a breach of contract claim might involve asserting that the AI program itself is liable for the harm. However, while holding the contracting algorithm liable may initially appear to be a plausible approach, it poses two serious concerns.

First, there is no legal precedent for holding a completely nonhuman entity liable for a person’s harm. Although corporations have been found liable for various harms, they are not analogous to AI-powered software programs. As “legal fictions,” corporations achieve legal personhood by “acting” through the actions of their human agents (that is, their officers, directors, promoters, and employees).¹²⁹ AI contractors differ significantly from corporations and operate in an almost entirely opposite manner. Instead of operating through human agents, AI software operates on behalf of humans. As a result, efforts to attribute liability to AI software by drawing analogies to corporate liability may be both inaccurate and misguided.

Second, if an AI model is held liable for contract breaches and required to pay damages to compensate for the resulting harms, this could expose AI software developers to above average or substantial levels of risk.¹³⁰ This increased risk may discourage AI developers from investing in further innovation, fearing that their investments could be lost to breach of contract, product liability, or other lawsuits. Additionally, if AI companies or algorithms were exposed to liability in this way, potential entrants to the AI contracting industry might hesitate, hindering further technological advancements. This suppression of innovation could cause greater harm to society than that posed by the inability of those alleging harm from breached contracts to obtain damages.

Thus, preserving innovation and investment into AI technology and its legal applications may involve specially protecting AI software, its users, and its developers from liability for harm-causing AI contracts—or, at the very minimum, maintaining existing standards of proof that prevent plaintiff-victims with lower socioeconomic statuses from securing damages in these types of cases.¹³¹ Under the current legal framework, only those individuals with higher socioeconomic statuses would be able to secure the costly expert testimony needed to demonstrate that an AI’s contract drafting did not satisfy the standard cost-benefit analysis used in determining liability in product warning, instruction, or design liability cases.¹³² Lowering the burden of proof would combat this issue, but such a change is unlikely to occur as it would expose AI software, its developers, and its users to substantial liability due to the highly unpredictable nature of AI-created risks.¹³³ Although there are numerous instances in recent history when the American public has accepted negative consequences for a minority group to achieve broader benefits for society as a whole,¹³⁴ the benefits of AI contracting do not outweigh its disproportionate harms.

Another issue in the context of assigning liability for AI contracting-related harms is allocating fault between the multiple parties that were involved in the contract’s creation and implementation. Parsing out which party should be held liable—whether it be the AI software itself, its designer, seller, or user, or another party altogether—inherently includes a significant policy decision as to how society chooses to (dis)incentivize AI technology’s development, usage, and applications.¹³⁵

D. Data Privacy and Security Concerns

When you log into ChatGPT to ask it a question, the prompt that you send the model does not stay on your laptop. It does not even stay on ChatGPT’s webpage.¹³⁶ By the time your query has been answered by the LLM (which is within seconds), your information is long gone—out into the ether of wherever OpenAI stores the many gigabytes of data it uses to train its AI models.¹³⁷ In reality, the information likely ends up in a remotely located and highly classified data center, where it sits on a server until OpenAI uses it to train its next LLM.¹³⁸

The average person may not care that their question asking ChatGPT to craft a new diet for them may get stored somewhere.¹³⁹ However, sophisticated legal clients commonly include their proprietary information—such as property addresses, purchase prices, and highly technical engineering or software information—in high-level contracts. Thus, legal clients are typically very protective of the private information in their contracts and subsequently include confidentiality clauses in their agreements to safeguard against disclosure to third parties.¹⁴⁰

For cases in which legal clients have highly sensitive information, AI’s “black box” can become a major issue. The “black box” problem refers to the fact that we are unable to see how LLMs make their decisions.¹⁴¹ Although the inputs and outputs of LLMs are observable, given the algorithms’ ever-evolving nature, their internal workings are a mystery—including what input data they retain.¹⁴² AI models’ mysterious inner workings may interfere with the efficacy and implementation of AI in the contract redlining and negotiation space because legal clients who are protective of their proprietary information may object to an AI model’s use in the contracting process. Even if a law firm used an “internal” AI software program, clients with sensitive information may not be comfortable with such a program because their information would be stored within the firm’s model for perpetuity.

There is an inherent tension between training an LLM and protecting clients’ confidential information. LLM models are trained on inputted data—and they improve if provided with greater quantities of training data.¹⁴³ Therefore, without clients who are willing for their information to be input into an LLM, the model’s efficacy will not improve. This may create problematic incentives for law firms to encourage their clients to commingle their sensitive information with that of other clients in the firm’s AI model in order to produce a better-quality software program for the firm.

Finally, LLMs’ greatest skill is their ability to recognize patterns in data. With more and more sensitive client information inputted into and stored by an LLM, the potential for an AI model to identify connections between data increases. In the case of an outsourced AI model not owned by a law firm, these recognized patterns may be disclosed to third parties for nefarious purposes. For instance, an LLM may analyze contracting patterns to determine which companies are economically successful, leading a third party to misappropriate this information and engage in fraudulent or deceptive dealings. In a more alarming scenario, third parties who gain access to confidential company addresses or security details that an LLM extracted from contracts—such as the location of a technology company’s classified data center—could use this information to break into the facility and steal servers.

V. AI: ARTIFICIAL INTELLIGENCE OR ACCURACY ISSUES?

Artificial intelligence is widely known to “hallucinate,” or misinterpret patterns in its data and create inaccurate or nonsensical outputs.¹⁴⁴ When an LLM hallucinates, it can fabricate legal cases, contradict itself, or provide outright wrong answers to questions.¹⁴⁵ In the contracting space, minute missteps when negotiating or redlining an agreement can have enormous consequences.¹⁴⁶ Therefore, AI’s tendency to hallucinate presents a major barrier to its successful implementation as a contractor. Given its pattern recognition functionality, AI is also known to provide different answers to the same question if it is asked multiple times, with slightly different wording, or by different people. These inaccuracies and inconsistencies are unacceptable in a detail-oriented field such as contract law, where “the devil is in the details.”

Furthermore, there are currently no regulatory compliance standards that would require AI models to be regularly updated with new case law, statutes, and other sources of law. On the other hand, state bar associations require attorneys to remain knowledgeable about updates in the law and complete continuing legal education (“CLE”) courses.¹⁴⁷ The nonexistence of regulation that would mandate AI models to remain up to date on new laws presents major challenges in the contracting space. Just like an attorney who refuses to complete their CLEs, an AI model that is not fully updated on what the current law is cannot adequately contract or negotiate for a client. Even if regulations were eventually implemented that required regular updates to AI models so that they included new case law, statutes, and other laws, this would be difficult to administer. Since it would be incredibly difficult, if not impossible, for an AI model to be instantaneously updated as new laws came into effect, this time lapse means that these models will always be somewhat out of date and not fully updated on the newest laws. Additionally, such regulations, if they came into effect, would place immense compliance costs on AI developers to continually update their models and may even discourage certain developers from entering the legal contracting space altogether.

Finally, LLMs are not sufficiently accurate to be used in contracting because of their technical limitations. AI technology lacks the ability to exercise judgment and is known to struggle with customization, context, and complexity (“CCC”)¹⁴⁸—all of which are highly relevant aspects of contracting. In fact, CCC is a major reason in-house counsel as a general concept exists; businesses that are highly technical or complex in nature often prefer to have their own attorneys who are better suited than outside counsel to understand the company’s unique situation and needs. Thus, AI would not serve well as a legal assistant because it would not understand the context or complexity of a prospective client’s specific contracting needs.

VI. LEGAL PROFESSION CHALLENGES

As fiduciaries for their clients, lawyers are held to a high professional standard. Subsequently, lawyers’ use of AI technology poses unique challenges to the legal profession, particularly in the context of contract drafting and negotiation.

A compelling argument can be made that an attorney who relies on AI technology to draft contracts violates their professional duties of competence and diligence.¹⁴⁹ Although the AI-toting attorney may argue that an LLM is a tool that they use to aid their work, much like Microsoft Word or Excel, such an analogy is misplaced. Generative AI differs from these types of technologies because it allows lawyers to create substantive work product with minimal effort.¹⁵⁰ Therefore, relying on ChatGPT for contract drafting may undermine an attorney’s obligation to provide competent and diligent representation for their client.

Furthermore, an attorney’s reliance on AI technology to draft and negotiate contracts may create communication gaps between the attorney and their clients. If an attorney blindly accepts an LLM’s output as the best possible redline or negotiation strategy in a given situation, the attorney may be incapable of explaining to their client why they undertook the AI-suggested action.¹⁵¹ This blind acceptance of an AI model’s output is very likely if an attorney uses an AI model to contract because we often cannot look into an LLM’s inner workings or see why they generate the outputs that they do.¹⁵² The black box problem exacerbates this duty of communication issue if an AI model executes contracts without humans involved in the contract drafting and negotiation process, as the model would provide little to no legal reasoning to its client to explain its outputted action.

As mentioned in Section IV.D, serious duty of confidentiality concerns arise when clients’ data is input into an LLM.¹⁵³ Even if placeholder information is used in an effort to protect confidential client data, an AI model may be able to use its ability to detect patterns to extract confidential information from the provisions and context that are inputted into it. This is especially possible if an attorney or law firm inputs substantial amounts of client data into an AI model, as in the case of AI-driven contract lifecycle management programs or internal AI programs more broadly.

Finally, AI is not suited for the ethical and emotional dilemmas that are inherent in legal contracting and negotiation. Attorneys regularly encounter ethically and emotionally intense situations when negotiating and contracting for their clients. If an AI model is tasked with contracting in an ethically ambiguous situation, it would lack the human touch necessary to appropriately respond. Even if the model was trained to provide canned outputs in specific scenarios, it would be impossible for the model’s programmers to predict all potential ethical dilemmas that the AI model may encounter in practice. Additionally, in emotionally intense contracting settings, such as mergers and acquisitions, partnership agreements, or certain real estate transactions, clients are likely to value the human touch of an attorney over the detached and indifferent nature of an AI model.

VII. EMPIRICAL RESEARCH: “HIRING” CHATGPT IN A CONTRACT NEGOTIATION

To test AI’s current capabilities in the contract drafting and negotiation space, the author conducted novel empirical research using OpenAI’s Application Programming Interface (“API”). The experiment was designed to imitate “hiring” ChatGPT¹⁵⁴ as a legal assistant by tasking it to assist with a client’s negotiation of a commercial real estate lease. To investigate whether ChatGPT suggests different negotiation recommendations depending on its type of client, the author selected four general client types for this experiment: (1) an individual; (2) a small, privately held corporation; (3) a large, publicly held corporation; and (4) a nonprofit organization. ChatGPT was not provided with additional information about each client, and the rest of the experiment—including the exact prompt language, base contract structure, and output scale—was held constant across all client types in order to control for differences in the AI model’s responses.

A commercial real estate lease was selected for this experiment because all four of the selected client types could plausibly negotiate and enter into a commercial real estate lease as a tenant. To simulate a real-world commercial real estate contract, the author provided ChatGPT with thirty generic boilerplate provisions typically found in a commercial real estate lease, such as assignment, security deposit, renewal option, and maintenance provisions.¹⁵⁵ For each provision, the AI software was asked whether it would recommend renegotiation to its client. To facilitate objective comparisons between ChatGPT’s responses for different client types, the query solicited numerical responses by specifically asking ChatGPT to output its response on a scale from 0 to 100. On this scale, 0 indicated that ChatGPT would recommend to the client that the language was acceptable and should not be renegotiated, while 100 signified that ChatGPT would recommend that the language was unacceptable and the client should renegotiate the provision.¹⁵⁶ The 0 to 100 scale was chosen to prevent ChatGPT from outputting renegotiation advice in plain English. With numeric outputs, the author did not need to make subjective judgments about the quality of ChatGPT’s negotiation recommendations—which would have been necessary if they were in plain English—in order to compare the outputs across client types.

ChatGPT was selected as the AI chatbot for this experiment due to its popularity.¹⁵⁷ Because ChatGPT is pervasive, the results of an experiment utilizing it are more easily generalized to real-world applications and settings than the results of an experiment conducted with a less popular AI program. Put simply, the author chose to use ChatGPT for this research because this experiment seeks to replicate laypeople’s use of AI to negotiate contracts and laypeople are more likely to use ChatGPT than other AI programs.

The author also selected OpenAI’s API to conduct this experiment rather than prompting ChatGPT manually because the API provided an efficient and cost-effective method of testing the author’s algorithmic discrimination hypothesis.¹⁵⁸ In general, an API is a set of protocols that connects software programs, devices such as computers, and applications by enabling them to more easily communicate with each other.¹⁵⁹ APIs are useful because they enable a researcher to automate repetitive tasks such as scraping information from webpages or, in this case, prompting ChatGPT repetitively.¹⁶⁰

To conduct this experiment, the author drafted Python code that prompted ChatGPT for each client-provision pairing through its API and saved the AI model’s outputted numbers in an Excel file. Notably, iterating prompts through OpenAI’s API enabled the use of its log probabilities (“logprobs”) feature to construct more accurate data as compared with the data that would result from manual prompting.¹⁶¹ Logprobs is a feature in OpenAI’s API that responds to a particular prompt with both ChatGPT’s most likely outputs and the corresponding log probabilities for those responses.¹⁶² In essence, the logprobs feature enables a researcher to determine the estimated probability that ChatGPT would respond to any given prompt with particular responses.¹⁶³ For instance, in the context of this experiment, when ChatGPT is tasked with advising an individual client about whether to renegotiate the “Premises” provision of the provided lease agreement, the AI program is 78.629% likely to output “25,” 11.181% likely to output “50,” and 6.966% likely to output “75” on the 0 to 100 scale.¹⁶⁴

The logprobs feature allowed the author to construct a weighted response output for each inputted client-provision pairing that represents ChatGPT’s landscape of potential responses in a single number. The author created each client-provision prompt’s corresponding weighted response by utilizing the five most common responses for each prompt. For example, the mathematics behind the average weighted response when ChatGPT advises an individual client about the “Premises” provision of the lease is shown in Figure 1 and described below.

Figure 1. Weighted Response Calculation for Individual Client “Premises” Provision

First, each of the top five response values were multiplied by their corresponding probabilities, which were extracted from the log probabilities provided by OpenAI’s API. Then, these individually weighted values (shown in Figure 1 under the “Response × Probability” column) were summed. For the “Premises” provision and individual client prompt in Figure 1, this sum totaled approximately 31.095. Then, the individual probabilities of the five most likely outputs were summed; in Figure 1’s example, that total equaled approximately 0.9798, or 97.98%. This total conveys that approximately 97.98% of ChatGPT’s responses to this particular client-provision prompt were either 25, 50, 75, 20, or 85. Finally, the “Response × Probability” sum (approximately 31.095) was divided by the probability sum (approximately 0.9798) to calculate the weighted average response for this particular client-provision combination, or 31.73. Therefore, when ChatGPT is tasked with assisting an individual client and the provided provision of the lease agreement is the “Premises” provision, the AI program’s weighted average response is 31.73. Qualitatively, a result of 31.73 on the 0 to 100 scale facially suggests that ChatGPT may not be highly likely or enthusiastic to recommend to the individual that they should renegotiate this provision. However, the nature of this experiment was to derive comparisons between client types, so although the 31.73 value might suggest that ChatGPT is unlikely to be a zealous advocate,¹⁶⁵ this value must be compared with the AI program’s average weighted responses for other client types with the same “Premises” provision to be able to draw substantive conclusions about ChatGPT’s propensity to discriminate against certain types of legal clients.

As demonstrated above, this math derived a single numerical response for each client-provision pairing, facilitating objective comparisons between ChatGPT’s outputs when it is “hired” by different clients. The individual client’s average weighted response was used as a baseline measure by taking each non-individual client response and subtracting the corresponding individual response for the same lease provision to calculate a difference between the two values for each provision. Then, these difference calculations (one value for each provision of the lease agreement) were plotted. The visual representations of the differences between the average weighted responses for an individual client and a small corporation, large corporation, and nonprofit organization were constructed by plotting these differences on the following histogram plots.¹⁶⁶

Small Corporation Versus an Individual as a Client

Figure 2. Histogram of Differences in Average Weighted Responses Between a Small Corporation and an Individual Client

The histogram of differences between ChatGPT’s average weighted responses for a small corporation and those of an individual client demonstrates a few takeaways. First, the differences are clustered around zero, where zero indicates no numerical difference between ChatGPT’s responses when hired by either an individual or a small corporation. This finding suggests that, for the most part, ChatGPT treats individual and small

corporate clients similarly when tasked with advising them in a contract negotiation.

However, the histogram includes some instances of large differences between individual and small corporate responses, such as one provision where ChatGPT output a renegotiation suggestion for a small corporation that was over thirty points larger than the recommendation it provided the individual client. Notably, there were no instances of ChatGPT outputting a weighted response for the individual client that was greater than or equal to ten points higher than its corresponding small corporate output. On the other hand, there were multiple provisions where ChatGPT output renegotiation suggestions for small corporate clients that were ten or twenty points higher than the provision’s corresponding individual-client responses. These provisions, in addition to the rightward-skewed shape of the histogram in Figure 2, suggest that ChatGPT tends to recommend renegotiation for small corporate clients more often and to a greater extent than it does for individual clients.

Large Corporation Versus an Individual as a Client

Figure 3. Histogram of Differences in Average Weighted Responses Between a Large Corporation and an Individual

Figure 3, which shows the differences between ChatGPT’s responses for large corporate clients and individual clients, demonstrates similar patterns. Much like the small corporate client example in Figure 2, Figure 3 includes clustering around zero. This suggests that for a variety of provisions, ChatGPT will provide similar renegotiation recommendations for both individual and large corporate clients.

However, Figure 3 also includes the most dispersed results of the three client comparisons conducted in this experiment. The histogram includes a wide variety of difference values, most of which are relatively numerically different from one another—so different, in fact, that they fall into individual difference bins in Figure 3’s histogram. The dispersed nature of these results suggests that, while there is some clustering around zero, ChatGPT provides a wider range of negotiation recommendations when advising large corporate clients compared with other client types. This variability may indicate that ChatGPT’s training data assumes that large public corporations are more varied and complex than smaller, privately held corporations¹⁶⁷ and subsequently require a broader variety of negotiation advice or have greater market power to exert its will in a contract negotiation.¹⁶⁸ Additionally, the broader spread of the differences in responses for large corporate clients as compared with individual clients might also suggest that ChatGPT views large corporate clients as having more nuanced or varied negotiation capabilities and needs compared with individual clients.

Nonprofit Organization Versus an Individual as a Client

Figure 4. Histogram of Differences in Average Weighted Responses Between a Nonprofit Organization and Individual Client

Figure 4 visualizes the difference in weighted responses for a nonprofit organization as ChatGPT’s client as compared with an individual as its client. Here, we see the strongest clustering of results around zero of the three client comparisons studied in this experiment.¹⁶⁹ This suggests that, between corporations and nonprofit organizations, ChatGPT considers a nonprofit to be most analogous to an individual in the contracting space. This makes some intuitive sense if ChatGPT assumes that both individuals and nonprofit organizations tend to have less financial and political resources, market power, and influence over negotiations than large public or small private corporations.¹⁷⁰

However, despite this stronger clustering of differences around zero for nonprofit organizations, the histogram in Figure 4 continues to demonstrate the same trend seen for both corporation types: a rightward shift. This again suggests that ChatGPT favors nonprofit organizations over individuals in the negotiation space by more strongly or commonly recommending renegotiation to them, potentially because the model perceives individuals as having less power than nonprofit organizations to effectively negotiate for favorable provisions.

D. Overall Trends and Conclusions

Figure 5. Histogram of Differences in Average Weighted Responses Across All Four Client Types

Figure 5 is an overlay of the results from Figures 2, 3, and 4. Taken as a whole, while there is some clustering around zero, the rightward shift in the data demonstrates that ChatGPT tends to recommend renegotiation to (1) large, public corporations; (2) small, privately held corporations; and (3) nonprofit organizations more often and to a greater extent than it does when its client is an individual. Additionally, there are few occurrences of negative values on the combined histogram, which represent when ChatGPT outputted an individual client renegotiation value that was higher than the value outputted for any of the other client types for a given provision. Collectively, these trends suggest that ChatGPT may discriminate against individuals when “hired” to consult a contract negotiation by recommending

less favorable terms or negotiation strategies to an individual than it would to other types of clients.¹⁷¹

Interestingly, the minimum differences for the small corporation, large corporation, and nonprofit organization clients were -5.82, -8.42, and -5.36, respectively. These values represent the provisions for which ChatGPT most strongly recommended negotiation to an individual client as compared with other client types. Conversely, the maximum differences, which represent the instances when ChatGPT most strongly recommended the small corporation, large corporation, and nonprofit organization to negotiate as compared with an individual client, are significantly larger than the minimum differences. The maximum differences for the small corporation, large corporation, and nonprofit organization were 39.28, 22.68, and 29.43, respectively. Taken together with each client type’s mean differences (3.98, 2.99, and 3.71, respectively), this data demonstrates the systematic disadvantage in negotiation advising that individual clients experience compared with their corporate or nonprofit counterparts when using ChatGPT to assist in a contract negotiation.

E. Shortcomings

Although the findings of this empirical study are intriguing, there are some important caveats to note as well. First, the author chose to specifically use OpenAI’s GPT-4 Turbo model for this experiment, meaning that its results may not be readily generalizable to other OpenAI or AI models. Additionally, to best balance creativity with coherence, the author set the API’s temperature to 0.7. Temperature is a parameter value that controls how often ChatGPT outputs a less likely response; in essence, it is a measure of how random or creative the model’s responses are.¹⁷² The author initially tested the experiment with GPT-4 Turbo’s default temperature of 1 but ultimately tamped the parameter down to 0.7 in an effort to replicate the deterministic nature of legal advising.¹⁷³ The author also decided to use only the top five logprobs, rather than more, in conducting this analysis.¹⁷⁴ Both the temperature and top logprob decisions were made in an effort to replicate an individual user’s experience on ChatGPT while maintaining consistency across various API code executions.¹⁷⁵

Unfortunately, while these decisions were necessary to conduct the research, they also inherently shaped its results. Any modification of the temperature or number of requested logprobs alters ChatGPT’s renegotiation recommendations. Furthermore, this style of research does not easily facilitate demonstrating statistically significant findings—such as with a p-value used in traditional statistical analyses—because the model generates different outputs each time the code is run. As a result, these findings are not readily replicable, which is an unfortunate nature of conducting social science experimentation with the black boxes that are AI models.¹⁷⁶

Beyond technical limitations, other factors may impact the generalizability of this study’s findings. Only one type of contract, a lease agreement with thirty boilerplate provisions, was used in this research. Future scholars can expand upon this work by incorporating new and additional types of contracts and more detailed or varied provisions into this study’s framework to investigate if AI models discriminate against individuals when contracting in different contexts or with multiple types of contracts. Additionally, given that ChatGPT is a large language model, it is likely that the exact phrasing of the prompts used in this research impacted the model’s recommendations. Therefore, future scholarship can include a greater diversity of prompt language to determine if these findings hold across different prompting styles and approaches.

Similarly, additional research can incorporate more specific details about the AI model’s client when soliciting negotiation advice, whether in the contract itself or by expanding on the details included when contextualizing the prompt for the AI model. Inclusion of greater detail in a future study may determine if the use of specific company or individual names or other information results in similar algorithmic discrimination patterns. Greater contextualization is also more likely to align with real-world uses of AI modeling in contract negotiation, as the user would probably provide information about themself, the other party, and the deal at hand while soliciting assistance from an AI model.

Additionally, another version of this research might request AI’s assistance in renegotiating a contract that initially includes blatantly favorable (or unfavorable) provisions for the client. This arrangement may demonstrate different findings than an experiment conducted with relatively neutral starter provisions would, like those used here. The author intentionally used neutral lease provisions in this case to facilitate easier comparisons between client types and force ChatGPT to rely on its training data in making renegotiation recommendations rather than following an implicit suggestion to renegotiate provisions that are blatantly unfavorable (or vice versa).

Another alternative experiment design might use iterative follow-up prompts, rather than a single prompt, to solicit advice from the AI model because the language and structure of the prompt used to solicit advice may influence the AI model’s recommendations. For example, uploading a contract to ChatGPT and asking it a leading question such as “Should I negotiate Provision A?” may result in the AI model suggesting renegotiation more often or to a stronger degree than a broadly phrased prompt that asks ChatGPT what it thinks about the provision. Furthermore, this experiment used a numeric scale to gather ChatGPT’s outputs in a form that was easily and objectively comparable across client types. The 0 to 100 scale used in this Note’s empirical framework inherently assumes that this continuum is representative of the quality and strength of the renegotiation advice that ChatGPT would output in plain English to a real-world client. In real life, an AI model’s output would be substantive—it would tell the user in plain English what it thinks of the provision, whether or not to renegotiate it, and why. Therefore, it may be worthwhile for future research to solicit and examine substantive outputs and assess whether those outputs are equally clear, definite, and confident across different client types.

Although this study’s findings have limitations that are common to empirical research, this Note offers novel insights into algorithmic discrimination in the contracting space. Plausibly, ChatGPT discriminates against individuals when tasked with advising them in a contract negotiation—as evidenced by the AI model suggesting renegotiation to individual clients less often and to a smaller degree than it does when advising other types of clients.

As noted above, additional scholarship can expand upon the research implemented in this Note to strengthen this conclusion. If future research confirms algorithmic discrimination in the contracting space, then AI models must be retrained to prevent further exacerbation of existing inequalities. If AI models discriminate against individuals as their contracting client, this behavior may worsen inequities between those who have the resources to renegotiate favorable contract terms (such as corporate firms) and those who do not (individuals, for example) and are therefore more likely to rely on AI as an accessible contract negotiation tool.¹⁷⁷

VIII. ENOUGH NEGATIVITY—WHAT IS AI GOOD AT?

While AI has a plethora of disadvantages that hinder its applicability to contract drafting and negotiation, it does have advantages in limited legal applications. For instance, given its ability to summarize information quickly and accurately, AI is a prime candidate for administrative, clerical, or other summary tasks. A number of these types of AI applications already exist, such as Evisort,¹⁷⁸ a contract workflow management program. AI can also streamline a law firm’s tracking of its billable hours (e.g., Clio AI¹⁷⁹). Furthermore, AI technology can prove useful in speeding up legal research by summarizing documents, as seen with LexisNexis’s Protégé.¹⁸⁰ As a rule of thumb, AI is best suited for tasks that do not require judgment. Unlike billing or other administrative tasks, contract drafting and negotiation requires immense judgment, which is why AI technology is better suited for legal uses other than contracting.

CONCLUSION

Artificial intelligence technology has taken the world by storm in recent years. Nearly every industry has experimented with new and innovative applications of AI technology, and the legal profession is no exception. Despite this enthusiasm, transactional attorneys should pause and seriously consider the negative implications and serious challenges involved when applying AI technology to the contracting space before they attempt to implement AI models into their practice. At the same time, it is important to remain mindful of the distinction between the “practice of the . . . [law]” and the “business of . . . [a law] firm[].”¹⁸¹ Given the contract law issues, equity concerns, legal profession challenges, and accuracy problems that abound when AI models draft and negotiate legal contracts, AI may be better suited to assist attorneys with administrative business tasks rather than the practice of law itself. This limitation on the use of AI in the contracting space is further underscored by ChatGPT’s tendency to discriminate against individuals when asked to assist them in contract negotiations, as demonstrated by the empirical research presented in this Note.

On the other hand, those determined to use AI in the contracting space may find it more useful in an in-house setting than in a traditional law firm. The typical in-house counsel functions as a “jack-of-all-trades” for their employer, managing multiple projects and legal practice areas simultaneously. Additionally, in-house counsel usually manages standard form contracts, particularly in cases when their business holds significant market power in negotiations with other parties. Maintaining a consistent client (i.e., the business) and contractual structure over multiple contract cycles would allow an AI program to detect familiar patterns and better understand the context and complexity needed to tailor contracts to the business’s needs. Furthermore, an experienced human in-house attorney may be able to manually adjust for any discriminatory patterns in an AI model’s outputted negotiation suggestions and provisions. Finally, the research presented in this Note indicates that large public and small private corporations face a lower risk of AI-driven discrimination in contract drafting and negotiation compared with other clients, such as individuals. Therefore, in an in-house attorney’s busy, consistent, and controlled setting, AI models may prove to have some utility.

However, technological innovation has its limits, and AI models are not yet suited for broad applications in legal contracting and negotiation. While this author is eager to see how AI developers and legal professionals address the current challenges of applying AI to contract drafting and negotiation—particularly, AI’s discriminatory tendencies—she is also reassured that transactional attorneys still enjoy some level of job security, at least for now.

Attachment A: Commercial Real Estate Lease Provisions

PREMISES

Landlord hereby leases to Tenant and Tenant hereby leases from Landlord those certain premises (the ‘Premises’) consisting of approximately _______ square feet located at _______________________, as more particularly described in Exhibit A attached hereto and incorporated herein by reference.

TERM.

The term of this Lease shall be for a period of ______ years, commencing on ____________, 20___ (the ‘Commencement Date’) and ending on ____________, 20___ (the ‘Expiration Date’), unless sooner terminated as provided herein.

BASE RENT.

Tenant shall pay to Landlord as Base Rent for the Premises, without any setoff or deduction, the annual sum of $_______________ payable in equal monthly installments of $_______________ in advance on the first day of each month during the Term.

SECURITY DEPOSIT.

Upon execution of this Lease, Tenant shall deposit with Landlord the sum of $_______________ as security for the faithful performance by Tenant of all terms, covenants, and conditions of this Lease. If Tenant fails to pay rent or other charges due hereunder, or otherwise defaults with respect to any provision of this Lease, Landlord may use, apply or retain all or any portion of the Security Deposit to cure such default or to compensate Landlord for any loss or damage resulting from such default.

PERMITTED USE.

Tenant shall use and occupy the Premises solely for _______________________ and for no other purpose without the prior written consent of Landlord.

OPERATING EXPENSES.

In addition to Base Rent, Tenant shall pay as Additional Rent Tenant’s proportionate share of all Operating Expenses. ‘Operating Expenses’ shall mean all costs and expenses incurred by Landlord in connection with the ownership, management, operation, maintenance, repair, and replacement of the Building and Property, including but not limited to: property taxes and assessments, insurance premiums, utilities, management fees, common area

maintenance, landscaping, and repairs and maintenance not required to be performed by Tenant.

MAINTENANCE AND REPAIRS.

Landlord shall maintain in good repair the structural portions of the Building, including the foundation, exterior walls, structural portions of the roof, and common areas. Tenant shall, at Tenant’s sole cost and expense, maintain the Premises in good condition and repair, including all interior non-structural portions of the Premises, such as doors, windows, glass, and utility systems exclusively serving the Premises.

ALTERATIONS AND IMPROVEMENTS.

Tenant shall not make any alterations, additions, or improvements to the Premises without the prior written consent of Landlord, which consent shall not be unreasonably withheld for non-structural alterations costing less than $____________. All alterations shall be made at Tenant’s sole cost and expense and shall become the property of Landlord upon the expiration or termination of this Lease.

INSURANCE REQUIREMENTS.

Tenant shall, at Tenant’s expense, obtain and keep in force during the Term of this Lease a policy of commercial general liability insurance with coverage of not less than $____________ per occurrence and $____________ general aggregate. Tenant shall also maintain property insurance covering Tenant’s personal property, fixtures, and equipment. Landlord shall be named as an additional insured on Tenant’s liability policies.

INDEMNIFICATION.

Tenant shall indemnify, defend, and hold Landlord harmless from any and all claims, damages, expenses, and liabilities arising from Tenant’s use of the Premises or from any activity permitted by Tenant in or about the Premises. Landlord shall indemnify, defend, and hold Tenant harmless from any and all claims, damages, expenses, and liabilities arising from Landlord’s negligence or willful misconduct.

ASSIGNMENT AND SUBLETTING.

Tenant shall not assign this Lease or sublet all or any part of the Premises without the prior written consent of Landlord, which consent shall not be unreasonably withheld. Any assignment or subletting without such consent shall be void and shall constitute a default under this Lease.

DEFAULT AND REMEDIES.

The occurrence of any of the following shall constitute a material default and breach of this Lease by Tenant: (a) failure to pay rent when due if the failure continues for ____ days after written notice has been given to Tenant, (b) abandonment of the Premises, or (c) failure to perform any other provision of this Lease if the failure is not cured within ____ days after written notice has been given to Tenant. Upon any default, Landlord shall have all remedies available under applicable law.

QUIET ENJOYMENT.

Landlord covenants that Tenant, upon paying the rent and performing the covenants herein, shall peacefully and quietly have, hold, and enjoy the Premises during the Term hereof.

ENTRY BY LANDLORD.

Landlord reserves the right to enter the Premises at reasonable times to inspect the same, to show the Premises to prospective purchasers, lenders, or tenants, and to make necessary repairs. Except in cases of emergency, Landlord shall give Tenant reasonable notice prior to entry.

SIGNAGE.

Tenant shall not place any sign upon the Premises without Landlord’s prior written consent. All signs shall comply with applicable laws and ordinances.

COMPLIANCE WITH LAWS.

Tenant shall comply with all laws, orders, ordinances, and other public requirements now or hereafter affecting the Premises or the use thereof. Landlord shall comply with all laws, orders, ordinances, and other public requirements relating to the Building and common areas.

ENVIRONMENTAL PROVISIONS.

Tenant shall not cause or permit any Hazardous Materials to be brought upon, kept, or used in or about the Premises by Tenant without the prior written consent of Landlord. Tenant shall indemnify, defend, and hold Landlord harmless from any and all claims, judgments, damages, penalties, fines, costs, liabilities, or losses arising from the presence of Hazardous Materials on the Premises which are brought upon, kept, or used by Tenant.

SUBORDINATION.

This Lease is and shall be subordinate to all existing and future mortgages and deeds of trust on the property. Tenant agrees to execute any subordination, non-disturbance and attornment agreements required by any lender, provided that such lender agrees not to disturb Tenant’s possession of the Premises so long as Tenant is not in default under this Lease.

FORCE MAJEURE.

Neither party shall be deemed in default hereof nor liable for damages arising from its failure to perform its duties or obligations hereunder if such failure is due to causes beyond its reasonable control, including, but not limited to, acts of God, acts of civil or military authority, fires, floods, earthquakes, strikes, lockouts, epidemics, or pandemics.

HOLDOVER.

If Tenant remains in possession of the Premises after the expiration or termination of the Term without Landlord’s written consent, Tenant shall be deemed a tenant at sufferance and shall pay rent at _____ times the rate in effect immediately prior to such expiration or termination for the entire holdover period.

SURRENDER OF PREMISES.

Upon expiration or earlier termination of this Lease, Tenant shall surrender the Premises to Landlord in good condition, ordinary wear and tear and damage by fire or other casualty excepted. All alterations, additions, and improvements made to the Premises by Tenant shall remain and become the property of Landlord, unless Landlord requires their removal.

DISPUTE RESOLUTION.

Any dispute arising under this Lease shall be first submitted to mediation, and if mediation is unsuccessful, then to binding arbitration in accordance with the rules of the American Arbitration Association. The costs of mediation and arbitration shall be shared equally by the parties.

NOTICES.

All notices required or permitted hereunder shall be in writing and may be delivered in person (by hand or by courier) or sent by registered or certified mail, postage prepaid, return receipt requested, or by overnight courier, and shall be deemed given when received at the addresses specified in this Lease, or at such other address as may be specified in writing by either party.

OPTION TO RENEW.

Provided Tenant is not in default hereunder, Tenant shall have the option to renew this Lease for ____ additional period(s) of ____ years each on the same terms and conditions as set forth herein, except that the Base Rent shall be adjusted to the then-prevailing market rate. Tenant shall exercise this option by giving Landlord written notice at least ____ days prior to the expiration of the then-current term.

OPTION TO EXPAND.

Subject to availability, Tenant shall have the right of first offer to lease additional space in the Building that becomes available during the Term. Landlord shall notify Tenant in writing of the availability of such space and the terms upon which Landlord is willing to lease such space. Tenant shall have ____ days from receipt of such notice to accept or reject such offer.

RELOCATION.

Landlord reserves the right, upon providing Tenant with not less than ____ days’ prior written notice, to relocate Tenant to other premises within the Building or Project that are comparable in size, utility, and condition to the Premises. In the event of such relocation, Landlord shall pay all reasonable costs of moving Tenant’s property and improving the new premises to substantially the same standard as the Premises.

PARKING AND TRANSPORTATION.

Tenant shall be entitled to use ____ parking spaces in the Building’s parking facility on a non-exclusive basis. Landlord reserves the right to designate parking areas for Tenant and Tenant’s agents and employees.

BUILDING RULES AND REGULATIONS.

Tenant shall comply with the rules and regulations of the Building adopted and altered by Landlord from time to time, a copy of which is attached hereto as Exhibit B. Landlord shall not be responsible to Tenant for the non-performance of any of said rules and regulations by any other tenants or occupants of the Building.

GOVERNING LAW.

This Lease shall be governed by and construed in accordance with the laws of the State of ______________. If any provision of this Lease is found to be invalid or unenforceable, the remainder of this Lease shall not be affected thereby.

ENTIRE AGREEMENT.

This Lease contains the entire agreement between the parties and supersedes all prior agreements, whether written or oral, with respect to the subject matter hereof. This Lease may not be modified except by a written instrument executed by both parties.

Attachment B: Excel Spreadsheet & Python Code

The Excel spreadsheet of OpenAI’s API outputs and the Python code used to obtain this data is on file with the author and available upon request.

99 S. Cal. L. Rev. 239

Download

Carly Snell*

*Executive Articles Editor, Southern California Law Review, Volume 99; J.D. Candidate 2026, University of Southern California Gould School of Law; Master of Public Policy Candidate 2027, University of Southern California Sol Price School of Public Policy; B.S., Mathematics, 2023, University of Arizona; B.A., Political Science, 2023, University of Arizona. I extend my sincere gratitude to Professor Jonathan H. Choi for his invaluable guidance, my friends and family for their unwavering support, and the editors of the Southern California Law Review for their hard work and dedication in preparing my Note for publication.

INTRODUCTION

I. The Harm of Ideological Employment Denial

A. Four Traditional Approaches

B. Threats to Personal Integrity

C. Two Public Harms: Affective Polarization and Censorship

D. Summing Up

II. Intending to Change Behavior through Social Pressure

A. Corrupting Aims

B. Punitive Aims

III. Inadvertent Social Pressure and Employer Integrity

A. Non-Complicity

B. Employer Non-Association and Speech

IV. Inadvertent Social Pressure and Employer Economic Interests

A. Customer and Employee Retention

B. Workplace Efficiency

C. Merit

V. Final Reflections

A. Do We Need Legal Regulation?

B. Moral Uses of Social Pressure

C. Revisiting the Examples

D. But What About the Nazis?

99 S. Cal. L. Rev. 1

Introduction

I. Conceptualizing Forgiveness in Criminal Justice

A. Individual Forgiveness in Criminal Justice

B. Should Forgiveness Have a Role in Criminal Justice?

II. Individual Forgiveness and Communal Absolution in Religious Tradition

A. Individual Forgiveness in World Religions

B. Christian Individual Absolution

C. Communal Absolution in World Religions

Communal Cleansing Rituals

Periodic Absolution Practices

Communal Absolution in American History

III. Forms of Absolution Existing in Criminal Justice

A. Sentencing and Charging Revision

Second Look Resentencing

Deferred Adjudication

Safeguarding Judicial Discretion

B. Automatic Expungement

C. Clemency

Pardon Power

Amnesty

IV. The Role of Systemic Absolution in Criminal Justice

A. Perpetrators of Crime Avoiding Adequate Accountability

B. Societal Shifts on Forgiveness

C. Punishment Levels Need Adjusting

Conclusion

99 S. Cal. L. Rev. 41

Introduction

II. The Study59For a more detailed description of the study methodology, see Appendix A.

III. Findings I: Carceral Conditions Impeding Sleep in Prison

A. Uncomfortable Beds/Bedding/Mattresses

B. Hunger

C. Extremes of Heat and Cold

D. Noise

E. Light

IV. Findings II: Meta-Conditions Impeding Sleep in Prison

A. Fear of Violence

B. Trauma

C. Poverty

D. Overly Intrusive Rules Enforcement

E. Daily Humiliations

V. Implications: Prisons, Punishment, Law, Policy

A. Prisons: Sleep Deprivation and Carceral Life

B. Punishment: Sleep Deprivation and the Carceral Penalty

C. Law: Is Systematic Interference with Sleep Unconstitutional? Is it Torture?

D. Policy: Challenges and Complications

Conclusion

99 S. Cal. L. Rev. 95

INTRODUCTION

I. Board of education v. PICO

II. THE FIRST AMENDMENT

A. Viewpoint Discrimination

B. Government Speech

C. Minor’s Right to Free Speech

III. HISTORY AND PURPOSE OF PUBLIC LIBRARIES

IV. HISTORY OF BOOK BANS AND CENSORSHIP

V. PROLIFERATION OF BANS AND LEGISLATION

VI. PROLIFERATION OF LAWSUITS

A. Eighth Circuit: GLBT Youth in Iowa Schools Task Force v. Reynolds

II. The Study⁵⁹

Artificial Intelligence’s Rise to Prominence: The “AI Boom”²⁷