Tel Aviv University - Buchmann Faculty of Law; Harvard Law School; Harvard University - Edmond J. Safra Center for Ethics
October 13, 2009
The article discusses comparative sentencing in criminal trials - namely, correlating punishment severity with the probability of guilt. The criminal proceeding is conducted in two phases: the guilt-innocence phase, in which the issue of the defendant's culpability is addressed, and the sentencing phase, in which the punishment is determined. There is an underlying assumption of separation between the decision-making processes in each of these phases, in that all questions of guilt are supposed to be resolved at the conviction phase of trial, and are not to be re-considered during sentencing. Once the defendant is found guilty beyond reasonable doubt, the extent to which the court was persuaded of her guilt is not meant to be included amongst the sentencing considerations: the severity of the punishment is considered to be detached from the probative strength of the evidence underlying the conviction. Likewise, when the incriminating evidence fails to cross the beyond reasonable doubt threshold, the defendant is not supposed to be subjected to any sanction. The court is prohibited from convicting her but expressing the epistemic doubt in a partial lenient sentence. This article will challenge the separation ideal both descriptively and normatively, and will contest the derivative all-or-nothing punishment regime, which dictates uniformity of punishment in the epistemic space above the reasonable doubt threshold and non-punishment in the space below this threshold. On the descriptive front, the article will demonstrate that prevalent criminal law doctrines and practices - such as the residual doubt doctrine, the recidivist sentencing premium, and the jury trial penalty - effectively tear down the boundaries between the decision-making processes in the two phases of the trial creating a de-facto correlation between certainty of guilt and severity of punishment. On the normative plane, the article will point to the sub-optimality of the bipolar all-or-nothing regime. The article will show that correlating magnitude of punishment with certainty of guilt is preferable to uniform punishment in the epistemic space above the reasonable doubt threshold. It will also demonstrate that when the level of certainty as to the defendant's guilt does not meet the reasonable doubt standard of proof, the imposition of partial punishment, reflecting the epistemic doubt, can also lead - in certain circumstances - to better outcomes than the presently existing alternative of full acquittal and no punishment. The article will conclude by contending with the expressive, retributivist, and institutional objections that can be raised against a comparative sentencing regime.
Number of Pages in PDF File: 51
Keywords: criminal procedure, punishment, residual doubt, criminal trial
Date posted: October 13, 2009 ; Last revised: October 30, 2013