The Weekly Abolitionist: Prisons Without Punishment?

For libertarian prison abolitionists, Randy Barnett’s The Structure of Liberty: Justice and the Rule of Law is an indispensable book. Not only does Barnett offer a persuasive series of arguments for a stateless legal order, he further argues against the legitimacy of punishment altogether.

However, even as crucial as Barnett’s work is for libertarian prison abolitionists, it is not a prison abolitionist book. Despite opposing punishment, Barnett does still defend hypothetical prisons on two grounds: incapacitation and restitution. In this post, I’ll try to explain why even Barnett’s very limited defense of prisons still fails.

What is a prison?

Before doing so, it’s important to remember what exactly we mean by “prison” when we say that we are “prison abolitionists.” By prisons, I refer to large compounds where people are involuntarily confined (typically with many other criminals) in response to their having committed a crime, without the right to voluntarily transfer to another location if that other location would confine them just as well, and where the administration has almost total control over those confined. A given prison might not clearly and precisely fit all of these conditions exactly, but this at least gives us a general idea of what we’re talking about. As will be important in what follows, it also shows that one can hypothetically be forcibly confined without the location of their confinement being a prison.


Barnett holds[1] that one non-punitive purpose for prison and prison sentences can be incapacitation. In other words, he believes that a prison sentence can be legitimate as a means of extended self-defense. If it is legitimate to use force in direct self-defense, then it is clearly legitimate to prevent someone from (for example) leaving a room if you know they would immediately aggress against someone if they did. If this is this the case, and if we accept that certain kinds of actions can communicate intentions toward future actions, then forcibly confining someone based on crimes they’ve committed in the past can also be potentially legitimate.

So far, Barnett’s argument is sound as far as it goes. However, the next step that Barnett takes is where he goes too far. Namely, he contends that this leads to the re-emergence of the crime-tort distinction[2], and consequently, prisons.

While it may be legitimate to forcibly confine people who pose ongoing threats, this is only going to be a very small subset of criminals. Even for violent crimes like murder, the fact that someone has aggressed against another person is not good enough of a reason to believe they will do so again. The average crime is a product of things like passion or circumstance, not of the psychotic nature of someone who genuinely poses an ongoing threat. Furthermore, most people who are ongoing threats are typically ongoing threats to particular people, not the general public, and can often be handled with a restraining order. Those who are ongoing threats to the general public are also not necessarily especially serious ongoing threats, and could potentially be handled in some way other than simply confining them.

When dealing with those remaining serious ongoing threats to the general public, forced confinement can be legitimate. However, it is not legitimate to confine them to a particular location when another location might do just as well. Nor is it legitimate to place them there automatically as a result of the particular crime they committed – Barnett is right to suggest that actions can communicate future actions, but this must be as part of a larger, highly contextual interpretation of those actions.

It is also not legitimate to place them there in such a way that puts them under the near total control of the administration. In fairness to Barnett, he seems to agree there, saying that prisoners “would be entitled to take legal action to ensure that their rights are respected”. It seems difficult, though, to see how that entitlement could be actualized without also allowing even a limited right of exit to another location (provided that the alternative location would work just as well.)

Finally, why this issue gives us the re-emergence of the crime-tort distinction is left unexplained. Surely, tort law is able to handle self-defense just as well as criminal law, and it’s unclear what necessary thing that bringing back the crime-tort distinction actually does in this instance.

Working off Restitution Debts

Barnett’s second, much more far-reaching and problematic justification for prisons involves cases where an aggressor is not immediately able to pay restitution, or might be expected to try avoiding payment[3]. In such cases, Barnett proposes that criminals could be sentenced to prison, where those debts could be worked off. Importantly, Barnett believes[4] that “they would be released only when full restitution had been made or when it was adjudged that reparations could more quickly be made by unconfined employment.”

The main problem with this approach, and especially with Barnett’s use of it, is that it clashes with inalienability. Without a sturdier justification for bringing back the crime-tort justification than the one he gives, restitution debts become legally indistinct from any other debt. Thus, unless Barnett is willing to sanction debt slavery for a debt generated by hitting someone’s car, or even failure to repay a loan, he cannot justify debt slavery for a debt generated by actions we currently think of as crimes. Wages can be garnished, and all sorts of other, more normal methods of debt collection can be used, but you cannot imprison someone and force them to work.

Notice, by the way, that even if you accept Barnett’s limited re-emergence of the crime-tort distinction, this does not justify imprisoning someone simply to collect their restitution debt. This is because Barnett’s justification is only argued for on the basis of incapacitation, and at no time does he provide a reason for also allowing restitution debt as an independent justification for incarceration. If the only reason a particular criminal in Barnett’s hypothetical is still in prison is because they have not yet paid off their restitution debt, then they are there for a reason that Barnett has not argued for.

Inalienable & Nonforfeitable Rights

Barnett might also resist the claim that his defense of inalienability is inconsistent with his defense of restitution-based imprisonment is by drawing a distinction between alienating and forfeiting one’s rights. As he states earlier[5] in the book:

“A claim that a right is inalienable must be distinguished from a claim that it is nonforfeitable. ‘A person who has forfeited a right has lost the right because of some offence or wrongdroing.’ One who wishes to extinguish or convey an inalienable right may do so by committing the appropriate wrongful act and thereby forfeiting it. But notwithstanding the consensual nature of such an action, it is the wrongfulness or injustice of the right-holder’s act, and not the right-holder’s consent, that justifies the conclusion that an inalienable right has been forfeited.”

All this may be enough to justify a distinction between the way we define the concepts of inalienable and nonforfeitable rights, but it does not justify a difference in the way we judge them. Virtually every argument for inalienability will also be an argument for nonforfeitability. Barnett’s own argument[6], that you are indeed physically unable to give up control of yourself, whereas you can easily physically give up control of alienable goods, is one such example. Committing a crime no more makes you able to give up control of your body than saying “I hereby give up control of my body.”

A defender of the claim that criminals can forfeit their rights by committing crimes might say that we need forfeitability in order to account for the justice of self-defense. If I’m coming at you with a clenched fist, clearly with the intent to swing it at your face, I forfeit my right to run in your general direction, and to swing my fists. Without forfeitability, someone might say, we cannot make that judgment, and are reduced to total pacifism.

Where this defense goes wrong is that it mistakes what’s morally going on in this situation. Someone who physically prevents me from attacking you is not acting against a right that I’ve forfeited, but one that I never had in the first place. If I never raised my fist or even had the thought of hitting you, I would still not have the right to use my fists in that way. If I did hit you, I would still retain all the rights that I had previously had to my fists – for instance, you couldn’t retributively cut them off – that I had before doing so. At no point does the concept of forfeitability need to be appealed to in order to explain self-defense.

For these reasons, Barnett’s defense of prisons without punishment does not succeed. Even still, because these reasons do not have to appeal to the concept of punishment, I take them to be the strongest reasons given for prisons. Thus, it is important to take the time to specifically address them independently of more standard justifications.

[1] pp. 187-191.

[2] pp. 190-191

[3] 177-181.

[4] 178.

[5] 78.

[6] 78-79.

Anarchy and Democracy
Fighting Fascism
Markets Not Capitalism
The Anatomy of Escape
Organization Theory