Philosophers on prison abolitionism: Theory versus practice

Abstract

Philosophers have long been debating the moral justifiability of punishment. However, they have seemingly ignored the adjacent question concerning the moral justifiability of incarceration, as demonstrated by the dearth of philosophical work on prison abolitionism. This silence is puzzling, given that, on closer examination, many philosophers implicitly (or even explicitly) endorse the core assumptions of prison abolitionism. By discussing some examples in the recent literature in philosophy of punishment, I argue that philosophers support prison abolitionism in practice, but not in theory. They endorse the practical strategies in the abolitionist agenda, yet refuse to engage with the normative assumptions that ground them. This ambiguity stems from a deliberate methodological choice that relegates issues concerning the reality of prisons to the margins of the philosophical inquiry. In doing so, philosophers hinder a fair assessment of prison abolitionism as a legitimate moral theory in criminal justice.

It is no exaggeration to say that, throughout the world, the prison system is in a state of disarray. ¹ Overcrowding, violence, lack of resources and staff burnout are well documented within correctional facilities (Allison et al., 2017; Garland, 2023; UNODC, 2013; Ward and Smith, 2023). ² Incarceration rates continue to grow, even in countries where crime rates are decreasing (Gramlich, 2024; Productivity Commission, 2021). Over 60% of previously incarcerated individuals return to prison within 5 years (Antenangeli and Durose, 2021). In a recent report, the UK House of Commons Justice Committee has noted the ‘enduring crisis in prison safety and decency’ and called for ‘a serious open public debate about (…) the role of prison and its affordability’. ³

There is certainly an urgent need for such debate, as few people deny that the prison system must undergo significant reforms to address the above issues. Some scholars, however, claim that the ongoing failure of the carceral system cannot be addressed through ‘reform’: rather, the goal should be to abolish imprisonment as a form of punishment altogether and replace it with non-carceral alternatives. Advocates of ‘prison abolitionism’ see incarceration as a morally fraught practice that harms individuals and communities, fails to rehabilitate, and reinforces social hierarchies. They demand that resources be diverted away from the prison system and towards social programmes that target the root causes of crime, such as poverty and mental health. For these theorists, it is time we acknowledge that incarceration constitutes an ‘obsolete’ form of punishment that promotes neither justice nor community safety (Davis, 2003).

Prison abolitionism has elicited significant interest among legal and social scholars (Ben-Moshe, 2020; Coyle and Nagel, 2022; Coyle and Scott, 2021; Mathiesen, 2014). ⁴ However, it has received very little attention from philosophers. Despite their focus on debates about the moral justifiability of punishing offenders, philosophers appear indifferent to the adjacent question concerning the moral justifiability of incarcerating offenders. ⁵ It may be tempting to attribute this indifference to an alleged theoretical weakness in prison abolitionism: hence, some may claim that philosophers do not engage with prison abolitionism because the latter is fundamentally a ‘naïve’ (Bagaric et al., 2021), ‘counterintuitive’ or even ‘absurd’ (Shelby, 2022: 2) theory. Yet, in this article I seek to provide a different explanation for the lack of philosophical work on prison abolitionism: the latter, I claim, stems from a methodological approach that relegates questions about the reality of the criminal justice system to the margins of the philosophical inquiry. Concerned solely with ideal theorising about punishment, philosophers ignore the non-ideal circumstances of existing prisons. They dismiss questions about the prison for they deem them irrelevant to their enterprise: but in doing so, they defuse prison abolitionism's potential contribution to debates in penology, by hindering the assessment of its moral claim about the wrongness of incarceration.

This treatment of prison abolitionism is all the more puzzling since, as I show in this article, many philosophers who refrain from engaging with its theoretical underpinnings nonetheless endorse its practical recommendations. Thus, I will argue, many philosophers support prison abolitionism in practice, yet not in theory. On the one hand, they endorse two of its core assumptions, that is, (1) that most criminal behaviour is caused by social factors (such as poverty and marginalisation) and (2) that prisons can do little, if anything at all, to address these social factors. One the other hand, they support the key strategies of the abolitionist agenda, that is, ‘moratorium’, ‘decarceration’ and ‘excarceration’ (Knopp et al., 1976). Nonetheless, they seemingly evade the normative question ‘Are prisons morally justifiable?’.

I provide evidence of this ambiguous stance on prison abolitionism by considering some examples from the extant literature in philosophy of punishment. I first consider Tommie Shelby's recent book (Shelby, 2022), which represents the only extensive analysis, by a philosopher, of the ‘idea’ of prison abolition. Shelby rejects that idea, claiming that prisons are indeed necessary for community safety. I first highlight some reasons why his argument fails to convince, including its omission or, at times, distortion of key aspects of the abolitionist view. I then show that, for all he says against the ‘idea’ of prison abolitionism, Shelby enthusiastically endorses its ‘practice’, as he openly supports the key abolitionist strategies mentioned above. Thus, although he rejects prison abolitionism in theory, Shelby supports it in practice: while he rejects the normative claim about the need to abolish prisons, he endorses the practical recommendations that rest on that very claim. I then consider four other examples of philosophers who adopt the same ambiguous stance towards prison abolitionism. As mentioned, my goal is to show that philosophers’ failure to engage with prison abolitionism does not warrant dismissing the latter as implausible; rather, that failure bears witness to the methodological flaws in the way philosophers approach questions about the prison.

The ensuing discussion comprises three sections and a conclusion. In section ‘Prisons and crime control’, I briefly highlight problems with Shelby's rejection of the idea of prison abolition. In section ‘On the idea of prison abolition’, I articulate the prison abolitionist perspective, by describing its three key strategies mentioned above. I then return to Shelby, to show that he supports all these strategies. In section ‘Prison abolition in philosophy of punishment’, I consider other examples of philosophers who endorse prison abolitionism's practical agenda yet refuse to engage with its theoretical underpinnings. In the conclusion, I highlight the methodological roots of philosophers’ ambiguous approach to prison abolitionism, drawing on the claim that the prison represents ‘a black box in punishment theory’ (Kerr, 2019). I close by noting that what needs to be reformed is not only the way societies punish lawbreakers, but also the way philosophers talk about incarceration.

Prisons and crime control

In his recent book, Tommie Shelby (2022) has argued that the current prison crisis does not show that prisons are inherently unjust. ⁶ Rather, prisons tend to mirror the ethical milieu of the societies in which they operate. ‘The fact that prisons are functional to unjust societies does not show that prisons exist because (…) they are functional to unjust societies’ (Shelby, 2022: 106). The injustice of (most) prisons is a contingent matter. The prison crisis is but a symptom of deeper structural problems, involving racism, poverty, lack of opportunities in our societies. It would be a mistake to focus on the symptom rather than on the causes: rather than abolishing the prison, he argues, we should aim to ‘abolish the ghetto’ (Shelby, 2022: 201).

It is worth pointing out from the outset that Shelby endorses the two core assumptions of prison abolitionism mentioned earlier. On the one hand, he notes that crime and incarceration are mostly driven by social circumstances (upbringing, status, race); on the other, he recognises that prisons can do little, if anything at all, to address such circumstances. Hence, he claims that if we want to address the social drivers of crime and incarceration, we need to focus on society, not on prisons. ⁷ Nonetheless, Shelby (2022) is firmly committed to prison reform rather than its abolition, for he believes that prisons, with all their flaws, play a ‘legitimate and socially necessary’ (p. 15) role in our societies, that is, crime reduction. This claim rests on the standard consequentialist justification of punishment, according to which ‘the harm that prisons cause can only be justified by the good that they do’ (Shelby, 2022: 149). Consequentialist justifications of punishment are usually framed in terms of deterrence, incapacitation or rehabilitation, and Shelby (2022) openly embraces the first of the three, as he believes that ‘fear of prison’ is what drives people away from crime (p. 154). Rehabilitation cannot succeed without deterrence, he writes, as the threat of punishment is necessary to pressure offenders into completing their programmes; incapacitation, on the other hand, becomes necessary only after deterrence has failed, that is, with individuals who will not be deterred and who must be incarcerated ‘to protect others from grave harm’ (Shelby, 2022: 169).

On the face of it, Shelby's argument for prisons is thus extremely simple. It is reminiscent of Beccaria's ‘classical’ theory of crime, which portrays offenders as rational, self-interested, utilitarian agents, inclined to follow the law only on a cost-benefit analysis (Beccaria, 1764). This can be noted in Shelby's concern that removing the threat of incarceration would not only embolden would-be criminals, but also ‘encourage or enable harmful wrongdoing among the otherwise law-abiding’ (Shelby, 2022: 161). However, this classical account faces well-known problems, which Shelby either underplays or ignores altogether.

From a normative standpoint, deterrence-based justifications of punishment attract familiar objections concerning the treatment of offenders as means to further others’ ends. Deterrence-based arguments have been traditionally criticised for disrespecting the status of offenders as autonomous agents, and for allowing (or even requiring) disproportionate punishments and/or unfair convictions (Boonin, 2008: 41–47; Brownlee, 2012: 243–248; Duff, 2001: 3–14). ⁸ Shelby sidesteps these philosophical objections altogether, which sits oddly with his intention ‘to make use of the tools of analytic philosophy’ to assess the idea of prison abolition (Shelby, 2022: 12). Equally disappointing is Shelby's failure to tackle the empirical issues faced by a deterrence-based defence of prisons, as the literature not only questions the extent to which incarceration can deter crime, but also highlights prisons’ criminogenic effects (Cullen et al., 2011; Kirk and Wakefield, 2018; Loeffler and Nagin, 2022; Paternoster, 2010; Rodriguez and Turanovic, 2018). While noticing this is ‘worrisome and a possible problem for reformers’, Shelby insists it is a problem that can be addressed, well, through reforms (Shelby, 2022: 158). As I show below, however, the reforms he proposes are the ones prison abolitionists defend.

Shelby's rebuttal of prison abolitionism hinges, furthermore, on the risk posed by violent criminals. This is notable in his worry that abolishing prisons would leave citizens without protection against ‘people who try to kill them [or] murder those they love’ (Shelby, 2022: 179). In doing so, Shelby reiterates an all-too-common objection to prison abolitionism, the one concerning the dangerous few, that is, ‘the axe murderers and serial rapists’ who would allegedly roam the streets in a world without prison (Frampton, 2022). However, there are two problems with dismissing abolitionism on this ground.

First, it is worth noting that many incarcerated people have not committed violent crimes. In the United States, for example, almost 40% of the population in State prison, and almost 50% of that in Federal ones, have not committed a violent offence. In the United Kingdom, non-violent crimes account for 61% of prison population (Carson and Kluckow, 2023; Prison Reform Trust, 2023). Among incarcerated women, that percentage grows to over 80% (Woods, 2023). Furthermore, many of those convicted for violent offences have not committed what are commonly considered violent acts (McLeod, 2015: 1168). With reference to this point, Shelby argues that prisons should be used only for those who commit ‘the most serious and egregious crimes’, while other offences should either receive a non-custodial sentence or be decriminalised altogether (Shelby, 2022: 116). But he fails to notice this is a key claim of prison abolitionism, as I show in the next section.

This paves the way to the second problem with the ‘dangerous few’ objection to prison abolitionism, that is, that it partly distorts the abolitionist view. Contrary to what Shelby claims, prison abolitionists do not claim that ‘the state should admit no new persons to prison and should release those now incarcerated’ (Shelby, 2022: 115), as I show in the next section.

On the idea of prison abolition

My goal is now to clarify the scope and the agenda of prison abolitionism, aiming to address some misconceptions about it. After that, I return briefly to Shelby's argument to show that, on closer analysis, it endorses rather than reject prison abolitionism.

In terms of scope, prison abolitionism must not be confused with a broader ‘penal’ abolitionism, which targets the institution of punishment altogether and seeks to dismantle the entire criminal justice system (Ruggiero, 2010; Saleh-Hanna, 2008). Prison abolitionists need not harbour such radical view: they object to the way offenders are punished (i.e., the use of prisons), not to the fact that they should be punished. In its most basic form, prison abolitionism pursues the elimination of one form of state punishment, that is, incarceration. The focus on alternative (i.e., non-carceral) forms of punishment, however, remains central to prison abolitionism.

Within this narrower horizon, and as already mentioned, prison abolitionism does not demand that prisons be closed at once. Rather, abolitionists pursue ‘a transformative goal of gradual decarceration (…) in which radically different legal and institutional regulatory forms supplant criminal law enforcement’ (McLeod, 2015: 1161). They thus acknowledge that, as part of this gradual process, ‘the dangerous few’ might be kept behind bars for public safety. It is worth stressing that this point is well established: decades ago, prison abolitionists had already noted that the ‘dangerous few’ objection involves ‘a confusion about time sequences’, for ‘prisons are a present reality [while] abolition is a long-range goal’ (Knopp et al., 1976). It is unfortunate that this objection continues to be employed nowadays to defuse anti-carceral narratives. ⁹

Having clarified its scope, I now provide a sketch of prison abolitionism's agenda, which comprises three key strategies. ¹⁰ The first concerns the establishment of a moratorium on the building of new prisons, as the initial step towards ‘a world without prisons’ should be to prevent the further expansion of the prison complex. ¹¹ A fixed number of prison cells would increase the pressure towards non-carceral responses to crime, forcing societies to think seriously about alternatives to the use of imprisonment.

The second strategy in the prison abolitionist agenda is decarceration, which focuses on limiting incarceration only for dangerous (i.e., violent) offenders. ¹² This is pursued through legislative changes, such as repealing ‘three-strikes-you’re-out’ legislation, decriminalising drug offences and sex work, abolishing bail, minimising the use of short-term sentences etc. Decarceration embodies the ‘gradual’ nature of prison abolitionism vis-à-vis the ‘dangerous few’ objection, as it grants that those individuals should remain behind bars if strictly necessary for public safety. As I show in the next section, this is a proposal that most philosophers openly endorse.

Thirdly, prison abolitionism pursues a strategy of excarceration. While decarceration emphasises legal reforms to reduce the population behind bars, excarceration focuses on social reforms to prevent people from getting involved with the prison system. ¹³ From this standpoint, excarceration embodies the essence of ‘a world without prison’, as it shifts the focus from ‘prison’ to ‘community’: it calls for diverting public funds away from the prison system and towards social and welfare support, to address criminogenic factors such as homelessness, unemployment, lack of education and mental health.

I now return briefly to Tommie Shelby, to clarify my claim that, in his book, he appears to reject prison abolitionism in theory, yet he supports it in practice. I already pointed out that Shelby endorses two core assumptions of prison abolitionism, that is, that crime and incarceration are driven by social factors (such as poverty and marginalisation), and that prisons can do little, if anything at all, to address those social factors. I now show that Shelby further supports the three abolitionist strategies I have just sketched.

First, he openly endorses a moratorium on the building of new prisons: in fact, he goes further than that, as he argues for a moratorium on the use of prisons. Until we ‘dramatically alter the social environment within which prison operates’, he writes, ‘we should regard the use of prison in places like the United States as illegitimate and intolerable’ (Shelby, 2022: 113). ¹⁴ Second, Shelby (2022) expresses his support for ‘decarceration’, claiming that it would ‘make sense’ (p. 116) to release prisoners who have not committed serious crimes, for prisons should only be for ‘those who commit the most serious and egregious crimes [such as] murder, rape, kidnapping, child abuse, sex trafficking and aggravated assaults’ (116).

It is also evident that Shelby further endorses ‘excarceration’. His call for ‘abolishing the ghetto’ stems from his belief that ‘[t]he root of the problem is not the existence of prisons, but rather the pervasive and deeply unjust socioeconomic disadvantage’ of our societies (Shelby, 2022: 82). For this reason, he thinks the ‘the appropriate response’ to crime must focus on addressing that disadvantage, through policies securing equal education opportunities, employment, housing, public healthcare, wealth redistribution and so on. As already mentioned, Shelby thus embraces the driving assumption of prison abolitionism, that is, that prisons cannot solve the problem of crime. Yet, faced with the choice between ‘abolishing’ or ‘reforming’ the prison, Shelby sides with reformism.

This dichotomy between prison ‘abolition’ and ‘reform’ is, furthermore, unwarranted. While Shelby takes the two camps as mutually exclusive, in reality prison abolition hinges on the idea of prison reform. We can appreciate this point by considering the concept of ‘non-reformist reforms’ (Gilmore, 2007: 242; Gorz, 1968): the latter introduce changes which, though potentially benefiting the system in the short term, aim to undermine it in the long term. In doing so, non-reformist carceral reforms work against the very system they seek to ‘improve’, by targeting the power relations underlying the custodial environment. An example is the abolition of forced labour in Norwegian prisons, discussed by Thomas Mathiesen (2014: 223). This reform had positive effects for the prison system, as it defused ongoing public criticisms; however, it ‘did not provide the system with a new, positive addition of legitimacy’ (Mathiesen, 2014: 224). Rather, abolishing forced labour was a way to weaken prison's coercive power over its inmates, the very power on which the system relied. ¹⁵ Reforms aiming at decarceration, for example, may be deemed ‘non-reformist’ as they seek to reduce the number of individuals over whom the prison has control (Giertsen, 2014).

Both abolitionists and reformers want prison ‘reform’: however, they want it for different reasons. ¹⁶ The disagreement between prison reformists and abolitionists concerns therefore the ends (i.e., a world with ‘better’ vs a world with ‘no’ prisons), not the means. ¹⁷ To be sure, prison abolitionists fear that an exclusive focus on carceral reform would encourage ‘the stultifying idea that nothing lies beyond the prison’ (Davis, 2003: 20), and that mass incarceration, racism, and inhuman treatment are mere flaws in an otherwise benign institution. However, this should not be mistaken for an uncompromising aversion to reformism. Shelby eventually acknowledges this point, as he writes that the gap between the two camps represents ‘a red herring’, ‘a distraction’ from their common target, which is the fact that ‘far too many people [are] in prison due to structural injustice’ (Shelby, 2022: 198). Nonetheless, he refuses to address prison abolitionism as a legitimate theoretical view, as his quick and, as I argued above, unconvincing dismissal indicates.

In the next section, I will show that this approach to prison abolitionism is emblematic of a broader stance in philosophy of punishment. Shelby is not the only philosopher who dismisses prison abolitionism in theory, while endorsing it in practice.

Prison abolition in philosophy of punishment

The ambiguity permeating Shelby's discussion of prison abolitionism stems from a common methodological approach to the problem of incarceration in philosophy of punishment. Simply put, that approach glosses over the reality of the prison system. Confined to defending an ideal theory of punishment, philosophers are untroubled by the fact that existing prisons do not resemble that ideal. ¹⁸ And yet, on the rare occasions when they turn their attention to imprisonment as a form of punishment, many philosophers endorse (like Shelby does) the abolitionist agenda, particularly its call for excarceration and decarceration. Nonetheless, virtually no philosopher has devoted any attention to the moral justifiability of imprisonment as a form of punishment. ¹⁹ In this section, I consider four examples of philosophers who endorse prison abolitionism in practice.

In presenting a normative defence of punishment, philosophers usually posit some necessary preconditions that must obtain for the practice to be morally justifiable. ²⁰ Among these is the familiar assumption that the societies imposing the punishment should themselves be reasonably just, that is, founded on a commitment to treating their members as equals. Real-world societies, however, fall short of this ideal, as large sections of the population live under chronic conditions of poverty and marginalisation, whose criminogenic role, in turn, is well documented (Battams et al., 2021; Camacho, 2023; McCausland and Baldry, 2023). The philosophers I briefly mention in this section acknowledge that much crime stems from social disadvantage; furthermore, they express support for not building new prisons, for decarceration and/or for excarceration. And yet, much alike Shelby, they fail to note that their normative arguments overlap, to a serious extent, with prison abolitionism. While they endorse the abolitionist agenda in practice, philosophers do not engage with its theoretical underpinnings concerning the wrongness of incarceration.

Let me first consider Gregg Caruso's rejection of retributive justifications of punishment (Caruso, 2021). Caruso claims that criminal behaviour is akin to ‘poor health outcomes’, as it is ‘more a characteristic of places and circumstances than of people’ (Caruso, 2021: 229). Unfortunate upbringing, lack of education, and history of abuse, for example, are well-known predictors of criminal behaviour, and fall largely outside the individual's sphere of control. To properly respond to crime, it is thus vital to address its social causes (which, for Caruso, also include poor diet and ‘environmental health’). An effective crime reduction strategy must focus on delivering tailored social justice policies, rather than on ‘punishing criminals and building more prisons’ (Caruso, 2021: 188). ²¹ To this end, Caruso outlines eleven proposals for criminal justice reform (Caruso, 2021: 257–268): notably, only one of these proposals concerns the prison system, while the rest focus on social reforms. The scarcity of prison-focused recommendations is no accident, as it shows that Caruso, like Shelby, endorses the driving assumption of prison abolitionism, that is, that prisons can do very little, if anything at all, to address the causes of crime. This is evidenced by his claim that ‘the only legitimate role’ prisons can play is to temporarily incapacitate seriously dangerous individuals, for rehabilitative (not punitive) purposes, and only if less restrictive alternatives (e.g., monitoring) are not available (Caruso, 2021: 265). All the prison can do, Caruso argues, is to act as a ‘quarantine’ detention facility (Caruso, 2021: 260). ²²

It is difficult to ignore the strong overlaps between Caruso's argument and prison abolitionism. The quick summary I just sketched already shows an endorsement of the three abolitionist strategies discussed earlier. Yet, Caruso does not even mention the word ‘abolitionism’ once, throughout his book: when it comes to prisons, all he says is that they should be turned from places of ‘retribution’ to places of ‘rehabilitation’ (Caruso, 2021: 260), adopting the ‘normality principle’ of the Norwegian prison system (van de Rijt et al., 2023). In doing so, he glosses over the momentous social changes the latter would require: the Norwegian prison model might work in Norway, a country with a strong welfare culture (and a long history of prison abolitionism activism), but not in Western, neoliberal countries, characterised by an enduring punitive culture (Papendorf, 2006). ²³ For all he has to say about crime and the appropriate way to respond to it, Caruso ends up fully supporting the prison abolitionist agenda, while sidestepping its theoretical underpinnings altogether. ²⁴

This ambiguous stance about prisons also characterises the work of Nicola Lacey and Hanna Pickard. In a series of influential articles, they have also claimed that crime should be treated as a health issue (Lacey and Pickard, 2013, 2015b, 2019). By presenting a conception of ‘responsibility without blame’, they have argued that offenders should be held accountable for their law-breaking conduct, yet without the ‘affective blame’ that characterises retributive discourses on punishment. Affective blame targets the offenders’ character rather than their conduct, stigmatising them as ‘evil’ persons: in turn, this undermines the main goals of punishment, which for Lacey and Pickard are reparation and rehabilitation (Lacey and Pickard, 2015b: 668). Hence, our penal policies must account for the fact that most offenders have ‘suffered terrible harm [for example, through childhood adversity, past trauma, physical and psychological abuse] (Lacey and Pickard, 2013: 24), and that such harm plays an important role in determining their criminal behaviour. We must therefore acknowledge that ‘many offenders are also patients’, and that patients need help rather than punishment. ²⁵

We can see parallels between this account of punishment and the one by Caruso. Both approaches draw attention to offenders’ specific circumstances, though Lacey and Pickard place most of the emphasis on past experiences, while Caruso also discusses ongoing disadvantages. Both approaches also call into question the extent to which retributive ideals can drive an appropriate and legitimate response to crime. And both approaches seemingly underplay the radical scope of their proposals. This is even more surprising in the case of Lacey and Pickard (2015b: 668), who contend that punishment should be conceptualised as ‘an institutionalised form of forgiving rather than blaming’, which does not entail ‘hard treatment’, that is, incarceration. In their view, punishment calls for ‘the imposition of consequences’, which will be ‘no doubt typically negative, but occasionally not, so long as [these consequences] are serious and appropriate to the crime and the context’ (Lacey and Pickard, 2015b: 672; 2013: 3). Lacey and Pickard do not articulate what this means in practice, though they are eager to clarify that their argument does not support the ‘abolition’ of punishment, only its ‘reconception’ (667). ²⁶ In the one instance they mention incarceration, they simply note that prisons should be redesigned to ‘allow offenders to maintain relationships with significant others’ (Lacey and Pickard, 2015b: 694). Yet, they sidestep the question of how prisons can be the site of the ‘detached blame’ they deem central to the practice of punishment. Forgiveness, rehabilitation and restoration are inherently relational concepts, and it is unlikely, to say the least, that an institution designed to remove offenders from society can properly pursue those goals. If hard treatment is not constitutive of punishment, then why incarcerate offenders? If punishment aims for reparation and rehabilitation, how can prison support that? It is unfortunate that Lacey and Pickard, like Caruso, choose to ignore these questions as well as their argument's practical implications for the use of prisons.

The same emphasis on the ‘social’ causes of crime characterises Vincent Chiao's analysis of the criminal law (Chiao, 2018). Chiao defends a principle of ‘inclusive aggregation’, according to which penal policies ought to equally consider the interests of victims and those of criminals (Chiao, 2018: 97). This claim rests, in turn, on the recognition that the state is often complicit in the crimes committed by socially disadvantaged citizens (Duus-Otterström and Kelly, 2019; Ewing, 2018; Tadros, 2009). Public institutions not only create the laws, Chiao argues, but also, through their policies, ‘contribute to defining the social circumstances under which individuals perceive the options that are realistically open to them, choose their actions, and plan their lives’ (Chiao, 2018: 98). In this way, Chiao questions the legitimacy of placing the responsibility for crime entirely on the offenders: to the degree that state policies affect the social causes of crime, criminal responsibility should be viewed ‘as broadly shared rather than as a matter of individual fault’ (Chiao, 2018: 98). ²⁷ This leads to Chiao's claim that, in designing its criminal justice policies, the state should not discard offenders’ interests on the ground they have committed a crime ‘no matter how serious’ (Chiao, 2018: 96). If anything, those policies should give priority to the interests of those whose access to basic freedoms (or, to use Chiao's expression, to ‘central capability’) is least secure, that is, the worst off. In this regard, Chiao's principle of inclusive aggregation forbids neglecting the interests of offenders, for example, via punitive policies centred on incarceration. Rather than focusing on ex post punishment, the state should prioritise ex ante strategies to mitigate the risk of people committing crime in the first place (Chiao, 2018: 99).

‘What does this mean concretely?’, Chiao then asks. It means an open endorsement of the idea of excarceration, as Chiao argues that societies should invest on education rather than on punishment ('schools, now’ rather than ‘prisons, later’) (Chiao, 2018: 101). The children who, without decent education, would go on to commit crime and be incarcerated for it, have a claim of justice that the state prioritise building schools rather than prisons. ²⁸ Such policy, Chiao (2018) concludes, better respects the status of those children as equals, compared to one that ‘waits for them to commit crimes and then punishes them for doing so’ (101).

Chiao's argument also supports the notion of decarceration. Echoing Caruso, he claims that prisons should be used only if strictly necessary to promote the community's overall wellbeing. The resort to incarceration is warranted only where it would ‘optimally [promote] effective access to central capability for all, when compared to other feasible alternatives (including doing nothing)’ (Chiao, 2018: 90). ²⁹ On that point, Chiao (2018) believes our societies have the means to respond to crime ‘in increasingly less destructive ways’ than by incarcerating offenders, through early intervention ('schools now’ rather than ‘prisons later’) and other ‘more sophisticated’ methods, involving mental health support, drug courts and restorative justice (90). The use of prisons, he concludes, should be limited to narrowly defined contexts and only ‘as a backup’ once non-custodial alternatives have failed. Importantly, Chiao is aware of the abolitionist overtones of his argument, as he briefly considers what he sees as ‘a kernel of truth to abolitionism: much, even most, criminal punishment might not be justifiable here and now’ (Chiao, 2018: 95). ³⁰ If we could develop non-custodial alternatives to prison, able to achieve similar results in ‘less destructive ways’, then, Chiao concludes, ‘the best thing to do about the criminal law might be to get rid of it’ (Chiao, 2018: 96).

Finally, let me consider some aspects of Antony Duff's work. Duff famously defends state punishment as the process whereby the polity communicates to offenders the censure they deserve qua members of the community (Duff, 2001, 2018a). However, Duff has long been mindful of the ‘considerable force’ of the abolitionist challenge, admitting that ‘our present practices of criminal punishment are, if not wholly unjustified, very largely unjustified’. ³¹ Recently, he appears to have partially yielded to the view that ‘our existing systems of criminal law are so deeply oppressive, damaging, and unjust that they cannot be saved – not even by radical reform; they must be abandoned, or abolished, as lost causes’ (Duff, 2020: 497). Like the authors mentioned above, Duff is aware of the criminogenic role of social disadvantage, and wonders whether we can hope to do penal justice in our unjust societies (Duff, 2018b: 785). Achieving the latter would require ‘the (re)creation of genuine communities whose members are united by shared values and mutual respect and concern’, something that would be necessary also if we were to follow the abolitionist proposal (Duff, 2001: 198).

Once again, what we have here is an (albeit implicit) nod at the idea of excarceration, as Duff emphasises that to achieve penal justice we must change our societies. He supports more explicitly the idea of decarceration, as he claims that imprisonment should have ‘a (very limited) role in a justifiable system of criminal punishment’ (Duff, 2001: 31). Prisons should only hold dangerous offenders, that is, those ‘who [have] defined [themselves] as a serious danger to others by [their] persistent commission of crimes of serious violence against the person’ (Duff, 2001: 170). But this raises two issues. The first is a shift from a communicative justification of punishment to one now based on incapacitation: granted dangerous offenders should be placed behind bars (much alike dangerous animals), it is not clear how that contributes to the ‘communicative enterprise’ central to Duff's account of punishment. This leads to the second issue, that is, that nothing in Duff's theory of punishment requires or presupposes the use of prisons (Lacey and Pickard, 2019: 238; Matravers, 2011; Tasioulas, 2011). What offenders deserve, Duff (2003) writes, is ‘the pain of remorse, and the burden of making moral reparations’ (190), but this does not imply imprisonment. Unfortunately, the question ‘Are prisons necessary?’ never appears in Duff's ‘communicative’ theory of punishment, as he never engages with the idea of prison abolition. In this regard, Duff represents another philosopher whose normative account of punishment does not imply the use of prisons, who endorses prison abolitionism in practice (by supporting strategies of decarceration and excarceration) and who, nonetheless, pays no attention to prison abolitionism as a theory (i.e., does not assess the moral justifiability of incarceration).

Conclusion: The ‘black box’ in philosophy of punishment

The philosophers I briefly considered in the previous section do not discuss what role, if any, the prison plays in their normative arguments. Despite endorsing key elements of the prison abolitionist agenda (the pursuit of excarceration, the call for decarceration and for a reduced reliance on prison), they do not engage with the normative claim grounding prison abolitionism, concerning the alleged immorality of incarceration. It goes without saying that those authors may have good reasons to believe that prisons should not be abolished: the problem is that those reasons are never articulated. ³² The prison appears to be taken for granted in philosophy of punishment.

Tommie Shelby might claim this is evidence that prison abolitionism belongs to the ‘absurd’ proposals (Shelby, 2022: 2). But I would argue that this is, rather, the expected by-product of philosophers’ exclusive focus on ideal theorising. On this point, Lisa Kerr has noted how philosophical accounts of punishment lack ‘any minimal fit with the realities of prison administration’: the prison, she argues, is ‘a black box in punishment theory’ (Kerr, 2019: 101). What Kerr means by this is that philosophers offer an ‘input’ (i.e., an account of when prisons should be used, on what grounds and for how long) which, they assume, leads to a certain ‘output’ (i.e., the realisation of a certain ideal of justice). Yet, philosophers bypass the process leading from the ‘input’ to the ‘output’, that is, the very process of incarceration: it is in this way that they treat the prison as ‘a black box’, whose ‘inner workings and methods’ remain obscure (Kerr, 2019: 87). Philosophers, Kerr continues, focus on punishment's normative principles (e.g., ‘the guilty deserve to suffer’ and ‘the punishment must fit the crime’) but do not engage with the adjacent issue of whether and how these principles may be pursued in our societies and, more to the point, whether incarceration is the appropriate means to pursue them. This is not an accident, she claims, but rather the deliberate result of philosophy's ‘abstract mode’, one that reduces punishment to ‘an unwanted experience, divorced from time and space’, vaguely conceptualised as ‘hard treatment’ (Kerr, 2019: 97). Oblivious to prisoners’ subjective experiences, the abstract mode takes punishment as a mere matter of duration: the worse the offence, the longer the deserved sentence (Kerr, 2019: 101–103). But this overlooks, for example, the sentence's impact on different prisoners: one's personal history (race, gender and past trauma), the specific institution (e.g., minimum vs maximum security facilities), the political context (e.g., a Scandinavian vs an American prison) will significantly affect the quantum of ‘pain’ each offender will experience despite facing the same ‘punishment’ (Bell, 2024; Garland, 2011; Kolber, 2009; Lacey and Pickard, 2015a).

Philosophy's ‘abstract mode’, Kerr (2019: 106) insists, stems from a ‘separability presumption’, which assumes that questions about penal allocation (e.g., who should be punished and why) should be kept separate from those about delivery (e.g., how the punishment should be administered in practice). This presumption is so pervasive in philosophy of punishment that, she writes, ‘to point it out is somewhat like showing water to fish’ (Kerr, 2019: 106). This suggests that philosophers’ reluctance to engage with prison abolitionism might have little to do with the latter's plausibility, and rather stem from a deliberate methodological choice, whereby debates concerning the reality of prisons are excluded altogether from the purview of the theoretical inquiry. ³³

These considerations reveal that the problem with an approach to prison abolitionism à la Tommie Shelby is that abolition is not a mere ‘idea’, to be assessed in abstraction from the reality of the ‘here and now’. We cannot debate the morality of incarceration from the ‘abstract mode’, as philosophers do, for prison abolitionism rests on, and develops from, a non-ideal theory of criminal justice. Calls for ‘reform’, therefore, should concern not only the way societies punish offenders: what needs to be reformed is also the way philosophers approach the problem of incarceration altogether.