A (final draft) render of my Phil. of Law final paper.


In this essay, I will present a criminal sentencing framework that goes beyond our current system where the judge only decides the duration of imprisonment. I will argue that sentences should specify access to counseling services, family visits, and reentry programs as well as the duration of imprisonment. I will first explain the way sentencing works right now from a historical and systematic perspective, then explain the important aspects that it leaves ambiguous. I will argue why these conditions are important according to some relevant theories of punishment, and also consider some alternative solutions.

Background on Sentencing

According to Cornell Legal Information Institute’s Wex database, ‘a criminal sentence refers to the formal legal consequences associated with a conviction.’ This can be a fine, restitution, community service, or any number of other things, but in this paper I will only be addressing prison sentences. The United States Sentencing Commission reports that 91.7% of federally convicted offenders were sentenced with imprisonment. Assuming that state rates are even close to this, prison sentencing concerns the majority of convicted criminals. And we have many prisoners: the Bureau of Justice Statistics (U.S. Dept. of Justice) reports that there were 1.4 million in 2017 (the latest year available).

Sentence guidelines are not a new concept. In 1987 the Comprehensive Crime Control Act was passed (Cornell Wex). It was intended to be a specific and uniform system for assigning sentences (Cornell Wex). However, the Act was demoted to a set of guidelines by United States v. Booker, 543 U.S. 20 (2005) (Cornell Wex). However, binding sentencing guidelines are in place in many states (Washington: RCW 9.94A, California PEN, Virginia 18.2). Insofar as they relate to imprisonment at all, these guidelines almost exclusively relate to the duration.

Constitutive Conditions

I claim here that there are morally significant aspects of a prison sentence besides duration. I will call these constitutive conditions. A good starting point for this is Emma Kaufman’s Harvard Law Review article ‘The Prisoner Trade’.

Kaufman describes the system of interstate prisoner transfer, describing the wide variation in prison conditions between departments and facilities. Facilities like Granite in Vermont are relatively humane, and facilities like Hancock in Virginia overly populated and lacking in programs (Kaufman 1817). A prisoner can be moved from the former to the latter with a minimal rationale and without their (or the sentencing Judge’s) consent (Kaufman 1817).

Kaufman lists harms caused by interstate prisoner transfer:

  • Distance from home (1857)— many incarcerated people have families and friends that can visit them, and if they are several states away, they can become harder or even impossible to visit.
  • Reentry programs (1858)— job training and certification programs are frequently state-specific, and prisoners are deprived of access to programs relevant for ‘back home,’ as well as potentially disadvantaged by differences between qualification standards.
  • More everyday considerations can be different between prisons as well, especially between prisons governed by different organizations. Prisons can have different levels of crowding, violence, qualities of medical care, and so on (1859).

All of these problems are especially notable in the case of interstate prisoner transfer, but they can also come into play when laws, corrections funding, staffing, or other factors change (I think Kaufman mentions this, but I can’t find where.). The normative upshot of Kaufman’s paper is this: prisoner transfer must be limited in scale and mechanism (p. 1874).

Philosophical Implications

It is obviously undesirable from a judicial perspective for the consequences of every sentence to potentially vary wildly based on factors totally out of a prisoner or judge’s control, but it is also incredibly unethical. This is because the current system substantially risks disproportionality between punishment and desert, whether you are a retributivist or a utilitarian.


I will make my argument serially on the basis of different theories, and Jeremy Bentham is a good place to begin. Bentham is known as a utilitarian because he considers the justice of a punishment to be determined by how it affects the individuals involved (Bentham 674). Bentham says that ‘the end of the law is, to augment happiness’, and he says that we should refrain from enacting punishments that are ‘unprofitable’ (Bentham 674-675).

What Bentham means by ‘unprofitable’ relates, of course, to happiness (Bentham 678). A punishment will be unprofitable when the unhappiness experienced by the punished outweighs the benefit gained overall (Bentham 677). Furthermore, Bentham also specifies (similarly to Michael Davis) that the punishment should outweigh the benefit an offender would gain by breaking the rule (678). These rules might seem to be contradictory, but all they are saying is that the punishment must be an effective deterrent so that someone could not commit a crime, be caught and punished, and still profit thereby, but that the collective benefit by that deterrence should still outweigh the unhappiness it causes in the person suffering the punishment.

Now, this implicates the variability of prison conditions in two main ways:

  • First, that the benefit of a punishment to society is a product not only of the deterrence it provides, but also its rehabilitative effectiveness. If, as Kaufman argues, reentry programs and access to social support structures are associated with lower recidivism, then the benefit to society is as much a product of those conditions as is the duration of a sentence.
  • Second, that time imprisoned can be vastly different experientially based on access to family, employment, risk of violence, and other such day-to-day concerns. A sentence in a ‘nice’ facility might cause less suffering for a prisoner than it would create benefits through deterrence, but an awful and dangerous sentence far from home could well cause more collective (consider the friends and family of the incarcerated as well) suffering than happiness.

A good judge, on Bentham’s account, would have to reckon with these factors, yet as things are now they would be unable to count on their prisoners getting anything more than the minimum possible severity of punishment or less than the maximum possible severity (Kaufman 1878). They would either have to vastly under-punish, or run the risk of unjustly over-punishing anyone who happened to be transfered or who experienced any of the other situational changes that can happen in a department of corrections. This makes for an egregiously inefficient and unjust justice system.


John Rawls did not argue for a utilitarian position in his magnum opus A Theory of Justice but he did temporarily defend one in the context of punishment (Rawls, A Theory 20). In ‘Two Concepts of Rules’, Rawls asserts that utilitarianism can be profitably applied to the establishment of the institutions of punishment, though not used to justify the decisions of individual judges (Rawls 685).

We can take after Rawls' example and assess our present system, where durations are fixed by judges while the other constitutive conditions of imprisonment are determined by departments of correction (or even private prisons) with neither party consulting the other. We can’t just compare this principle with the absence of a justice system. We should follow Rawls' methodology in A Theory of Justice and present the strongest possible alternative principles. My alternative principle, which I will argue for more holistically later on, is that judges decide a prisoner’s right to these things when they impose a sentence. From a utilitarian perspective, I think it is clear that if we trust judges to make these decisions in more just fashion than wardens, we should expect a better outcome.


Immanuel Kant was a retributivist. In ‘On the Right to Punish’, he argues that we determine the just punishment for an offense based solely on the offense, rather than on the public benefit from the punishment (Kant 702). This means that rather than justifying punishment on the basis of its overall outcome, a punishment is justified regardless of whether society will benefit from imposing it.

Kant talks in terms of specific punishments. There is a specific punishment appropriate for murder, and a specific kind of punishment appropriate for theft (Kant 702-703). He not only argues that a punishment could be inappropriate to a crime, but that it could be inappropriate because it is not qualitatively related enough to the offense (Kant 702-703).

Kant would be disgusted with our justice system because it often punishes murderers with imprisonment instead of death. However, he would also prefer a system that allowed a judge to finely control more features of imprisonment. A corrupt teacher could be barred from education programs, a thief could be disallowed work programs, a child-killer could be disallowed family visits, et cetera. These restrictions seem draconian, but they are clearly well-suited to Kant’s stringent and unique sense of justice, so it is clear that Kant would prefer my alternative system to the one presently in place.


In ‘Harm and Retribution’, Michael Davis argued for a theory that is retributivist but is otherwise quite different from Kant’s. Davis argues that we can and should conceptualize criminal punishments as retribution for the ill-gotten gains of law-breaking (Davis 713). Davis says that we should imagine a license that would grant pardon for a crime, to be bought before committing it. We can ‘price’ an act in comparison to other acts based on how likely potential ‘consumers’ would value various licenses (Davis 713).

So, how is this relevant to variable prison conditions? Well, if we put the judge in the position of the merchant who sets the price of a license, then a corollary situation would be a merchant who can’t tell you how much something costs. The merchant might be able to set the price for something to between ten and twenty dollars, but he can’t give a more accurate price than that. Instead, you have to buy it and see how much money leaves your bank account over the course of a few months or years. In other words, in this situation judges are put in a position where they have to allocate a punishment that could end up being very bad or only somewhat bad, merely based on chance factors outside of their control.

Now this might amount to gambling in a merchant scenario and thus a good tactic, but in a legal scenario we actually want to sell no licenses. If a criminal commits a crime and considers the price worth it in the same way that a satisfied customer does, then they will have profited from their misdeed. On Davis' theory, this means the law has failed. The punishment of the law should leave all rule-breakers not better off for their having committed a crime, so keeping open the possibility for it to be worthwhile risks at best injustice, and at worst will encourage risk-takers to try their luck, creating more crime.

Sentencing Process

Considering that the present variability of these constitutive conditions is so easily construable a serious detriment, I suggest that judges be empowered to stipulate other constitutive conditions of imprisonment alongside the duration. Their decision could require that correctional departments keep them close to their family, allow them job paying more than a certain amount, allow or even required to participate in counseling and education programs, or deny them all of these things. Different situations would call for different rules, but key aspects such as access to parole programs, job training, health services, counseling, proximity to communities, in-person and phone visit access, and so on all should at least be either minimally specified or explicitly omitted from the sentence. Furthermore, judges should also be required to justify their deviations from these conventions.

All of these factors play a substantial role in imprisonment on all of the theories of punishment that I have mentioned, so leaving these factors up to chance is a harmful and immoral omission. However, this is not to say that a sentence should have to specify all of these constitutive conditions in detail. Instead, constitutive conditions not specified in the sentence should be stipulated by a de jure standard of imprisonment. This standard could be authored by judges or legislators, and any specification in a sentence besides the duration could be assumed to follow the de jure standard unless specified otherwise.

This way, judges would still have the power to tailor the punishments to either fit the crime (for a retributivist) or fit a rehabilitation scheme for the prisoner (for a utilitarian), without creating unnecessarily complicated individual guidelines for correctional facilities to obey.


This is a radical change from the present state of affairs, so there are many objections worth considering. Below I have described three compelling ones and my rejoinders.


In ‘The Prisoner Trade’, Kaufman proposes ‘consent as a cure’ for the ills of interstate prisoner trade, suggesting that prisoners should either have a right not to be transferred out of state, or even a right to apply to be transfered (1873-1874). This solution would substantially curb transfer rates, especially of those transfers likely to be detrimental to the inmate.

The objector might also say that this is the most common-sense and implementable way of addressing these concerns, certainly to be preferred over the more complex solution I have proposed. This case could be made for prison transfers, but it’s not feasible to get consent from prisoners every time you change the availability of a reentry program, adjust staffing, or lose health service funding. If we want to address these more minute, widely enacted and more necessary changes that still make a difference to prisoners, we need a solution that does not rely on their consent.

Universal Standard

It might also be objected that it is better to establish a more robust baseline for prisoner care and apply that universally. The objector might say that it doesn’t make sense for us to specify things other than duration for individual prisoners because it would impose too great a cost on correctional facilities.

I respond that it also costs correctional facilities money to feed their prisoners and allow them sunlight, but they must do so regardless. Correctional facilities are not intended to make prisoners miserable for as little of the federal budget as possible, we are only justified in punishing them within certain standards. While ideally all of these standards besides duration would be universal, this is not the case. In the same way that different cases justify different durations, different cases also can justify other variant circumstances. All this objection really suggests is that somehow this requirement for tailored circumstances is superseded by an imperative so save the state’s money in the process. Let’s review the literature.

For someone with views like Kant’s, it is clear from his writings on punishment that he would not accept a modification of a sentence in favor of expediency, as evidenced by his stance on the hypothetical proposal where a prisoner sentenced to death could be instead subjected to dangerous medical experiments. Davis likewise does not mention allowing any modification of a sentence in favor of cost.

However, Bentham and Rawls do potentially allow for such a change. Their theories both being utilitarian, it is easy to see how they could permit the risk of one sentence being less just than another in every way but in its financial cost, so long as the good consequences of that financial cost outweighed the ways in which it made the punishment worse. Yet we have to weigh the potential financial detriment against the possibility of improved rehabilitative measures. It is easy to imagine that judges will be in a better position than correctional officials to devise a rehabilitative program, especially given the budgetary incentives correctional facilities have to avoid doing so. There may be situations where the financial costs of a more specific sentence outweigh the rehabilitative benefits, but it is also hard to imagine that such situations are many, given the immense cost of recidivism.

Ultimately it seems that within a retributive framework, the more specific a judge’s sentence can be, the more just we can expect it to be, so a single baseline standard would not work well. However, within a utilitarian framework the value of a single standard might be more arguable, depending on an analysis of the financial costs measured against rehabilitative benefits I would like to imagine that spending money banking on rehabilitation is unlikely to be a bad idea at any reasonable scale of expenditure. Furthermore this objection simply will not satisfy retributive concerns as well as my proposal.

The Potential for Injustice

It has also been objected that giving judges an increased ability to specify the conditions of imprisonment will result in increased variability in constitutive conditions even between those with similar offenses, because there will be more opportunities for the individual differences between judges to come out: corruption, bias, and irrationality all will potentially have an increased role to play in sentencing in my proposed system.

First, it has to be said that the objector must still advocate for a baseline system such as I have outlined. A baseline system might be said to be preferable to the status quo on the grounds mentioned above, but our present system of unspecified conditions is simply indefensible, because the constitutive conditions people are imprisoned under currently can vary wildly according to the biases of correctional facilities and officers, to the extent that programs can be freely adjusted and prisoners freely transferred.

I have two counterpoints. The first is that my system which offers judges the opportunity to deviate more from convention, but also documents their reasoning will potentially give a convicted person a greater body of material that can be searched for bias writing an appeal. This material will allow society to hold corrupt and biased judges accountable.

My second counterpoint relies on the majority opinion of the United States Supreme Court in the case of United States v. Booker 543 U.S. 20 (2005) (accessed through the Cornell Legal Information Institute), written by Justice Stevens. The Court held that no set of legislative rules (in this case a part of the Crime Control Act) could affect a sentence based on a fact a jury had not reviewed. Rules that systematically affect sentences must be factually reviewed by the jury, and it seems that any baseline system that I have already objected to ould also need be reviewed by the jury if those baseline rules were not at least subject to a judge’s discretion. In other words, so long as the baseline affects every sentence, it must be weighed in the context of the individual case by the jury, if not the judge.

In this case, it seems to me that much more bias (and also cost, though this shouldn’t be a major determinant) would likely be introduced by leaving such decisions up to juries. As for the fact that juries do not currently consider constitutive conditions: though juries do not directly decide sentences, they can decide whether someone is guilty of murder in the first or second degree, or felony murder, a decision whether that can have a drastic impact on sentence (Cornell Wex). If they decide sentences indirectly, they should consider constitutive conditions for the same reason judges should, or, more efficiently and likely less vulnerable to bias, the judge should have discretion over constitutive conditions in the same way judges have discretion over the impact of jury determination of guilt in normal criminal trial proceedings now.

In conclusion, as long as we recognize that conditions outside of duration are morally relevant features of a term of imprisonment, we must recognize that convicted people have a right to have those conditions considered, either by a judge or jury. This may be more challenging or costly than the present system, but if we want to practice justice in a way that is compatible with any philosophically deep and consistent framework for punishment, we must be willing to endure additional labor and expenditure.


In bibtex.

	address = {New York, NY},
	edition = {1. paperback ed},
	title = {What money can't buy: the moral limits of markets},
	isbn = {978-0-374-53365-6},
	shorttitle = {What money can't buy},
	abstract = {Sandel argues that we have drifted from having a market economy to being a market society and examines one of the biggest ethical questions of our time: What is the proper role of markets in a democratic society, and how can we protect the moral and civic goods that markets do not honor and money cannot buy?},
	language = {eng},
	publisher = {Farrar, Straus and Giroux},
	author = {Sandel, Michael J.},
	year = {2013},
	note = {OCLC: 935524500}

	title = {Judicial {Compensation}},
	url = {https://www.uscourts.gov/judges-judgeships/judicial-compensation},
	urldate = {2020-05-05},
	institution = {Administrative Office of the U.S. Courts}

	title = {Prisoners in 2017},
	urldate = {2020-04-20},
	institution = {Bureau of Justice Statistics, Office of Justice Programs, U.S. Department of Justice},
	author = {Bronson, Jennifer and Carson, E. Ann},
	month = apr,
	year = {2019}

	title = {Overview of {Federal} {Criminal} {Cases} : {Fiscal} {Year} 2019},
	url = {https://www.ussc.gov/research/data-reports/overview-federal-criminal-cases-fiscal-year-2019},
	urldate = {2020-04-20},
	institution = {United States Sentencing Commission},
	author = {Schmitt, Glenn R. and Russell, Amanda},
	month = apr,
	year = {2020}

	title = {Cornell {Wex}},
	url = {https://law.cornell.edu/wex},
	publisher = {Cornel Legal Information Institute},
	year = {1992}

	title = {The {Prisoner} {Trade}},
	journal = {Harvard Law Review},
	author = {Kaufman, Emma},
	year = {2020}

	title = {Harm and {Retribution}},
	volume = {15},
	issn = {00483915, 10884963},
	url = {www.jstor.org/stable/2265211},
	number = {3},
	urldate = {2020-04-14},
	journal = {Philosophy \& Public Affairs},
	author = {Davis, Michael},
	year = {1986},
	pages = {236--266}

	title = {Two {Concepts} of {Rules}},
	volume = {64},
	issn = {00318108},
	url = {https://www.jstor.org/stable/2182230?origin=crossref},
	doi = {10.2307/2182230},
	number = {1},
	urldate = {2020-04-13},
	journal = {The Philosophical Review},
	author = {Rawls, John},
	month = jan,
	year = {1955},
	pages = {3}

	address = {Cambridge, Massachusetts :},
	title = {A theory of justice},
	isbn = {978-0-674-00077-3},
	publisher = {The Belknap Press of Harvard University Press,},
	author = {Rawls, John},
	year = {1999}

	address = {New York Oxford},
	title = {The philosophy of law: classic and contemporary readings with commentary},
	isbn = {978-0-19-515512-9},
	shorttitle = {The philosophy of law},
	language = {eng},
	publisher = {Oxford University Press},
	editor = {Schauer, Frederick and Sinnott-Armstrong, Walter},
	year = {1996},
	note = {OCLC: 834299380},
	keywords = {law, philosophy}