Many criminal defendants are held in detention while they await trial. Though conditions in pretrial detention are much like those in prison, detention is technically not punishment. Since detainees are merely accused of crimes, they are presumed innocent.1 Their detention is not intended to punish them, and so, the Supreme Court has said, it is not punishment at all.2 Rather, detention is a means of promoting public safety, reducing witness intimidation, and preventing people accused of crimes from fleeing before trial.
Nevertheless, defendants who are convicted generally receive credit at sentencing for time served in pretrial detention.3 An offender who deserves two years of incarceration as punishment will have his sentence cut in half if he has already spent one year in detention. Such offsets appear to conflict with principles of proportional punishment. Taken literally, giving credit for time served leads us to systematically underpunish detainees by reducing their punishment by time spent unpunished. To unlock the mystery of credit for time served, defenders of proportionality need to square the widely held view that offenders should receive credit for time served with the widely held view that punishment should be proportional to blameworthiness.
One seemingly plausible solution that I suggest in Part I is to modify our understanding of proportionality. Courts and legal scholars sometimes draw too sharp of a distinction between formal punishment and other harms inflicted by the state. On more careful reflection, however, what they really seem to care about is not proportional punishment but proportional "harsh treatment," which includes not only the suffering and deprivation of formal punishment but also the suffering and deprivation of "punishment look-alikes," like pretrial detention. Even though pretrial detention is technically not punishment, it is harsh treatment, and most people are inclined to give offenders credit for it.
Shifting to proportional harsh treatment, however, solves one problem at the expense of several others. As I argue in Part II, once we understand punishment severity in terms of harsh treatment rather than a more neatly bordered but inaccurate construct like days in prison, we must consider the actual amount of harsh treatment we inflict.4 The amount will vary based on (1) the particular facilities to which inmates are assigned, (2) how inmates experience those facilities, (3) how confinement harms them relative to their unpunished baselines, and (4) how they are affected by the collateral consequences of incarceration for decades to come. Even if these harms do not constitute punishment, they contribute to sentence severity as surely as pretrial detention does.
While we could try to salvage proportional harsh treatment by taking all of this variation into account, I argue in Part III that when we look closely at proportional harsh treatment, it becomes much less appealing and consequentialist punishment theories that do not depend on proportionality look comparatively more appealing. Even though retributivist notions of proportionality are central to sentencing systems around the world and are widely thought to undergird core notions of criminal justice, proportionality has profoundly counterintuitive implications.
I. THE MYSTERY OF CREDIT FOR TIME SERVED
In this Part, I describe what I mean by proportional punishment and how it conflicts with giving offenders credit for time served. I then propose five responses proportionalists might offer to reconcile the two policies. I argue that the best reconciliation requires us to replace proportional punishment with the more general concept of proportional harsh treatment.
A. Background on the Meaning of Punishment
In United States v. Salerno,5 the Supreme Court upheld the federal statute that provides for pretrial detention of dangerous offenders.6 Crucial to the Court's determination was its view that Congress intended pretrial detention to be "regulatory" rather than "punitive. …