For the jury, it was a close call whether to find Rene Rodriguez guilty of selling any marijuana at all.

On the third day of deliberations, the jurors told the judge that they were unable to reach a unanimous verdict. But finally, after being given an Allen charge, they brought in a conspiracy conviction. It was apparently based on prosecution evidence implicating Rodriguez in a 10-ounce marijuana sale, together with the judge’s instruction that all the jury had to find to convict was that Rodriguez had conspired to sell a "measurable" amount.

The sentence: life without parole.

How’s that? Rodriguez would have faced only 18 months in prison, increased to 4 1/2 years by his two prior drug convictions, if his sentence had been based on the 10-ounce sale that was proven beyond a reasonable doubt at trial.

What jacked his sentence up to life without parole, under as-early "three strikes and-you’re out" statute enacted by Congress, was a finding by the judge, at the post-verdict sentencing hearing, that Rodriguez had probably conspired to sell more than 1,000 kilograms of marijuana, not just 10 ounces.

He probably had. But the judge, Thomas Curran of the U.S. District Court in Milwaukee, did not suggest that this had been proven beyond a reasonable doubt, or even by clear and convincing evidence. Nor did the prosecution claim to have that kind of proof.

Rather, under the current federal sentencing regime, the prosecution had to prove only that it was more probable than not that Rodriguez had sold 1,000 kilograms, in order to condemn him to die in prison.

That seemed unjust to Chief Judge Richard Posner of the U.S. Court of Appeals for the 7th Circuit, the Reagan appointee whose vast body of scholarly work ranks him among the most brilliant intellects on the federal tench. In a lucid dissent from the 7th Circuit’s Jan. 4 denial of Rodriguez’s petition for rehearing en banc, Posner wrote this:

[T]here is a serious question whether it is permissible to sentence a person to life in prison, without possibility of parole, at the end of a brief and casual sentencing hearing in which there is no jury, in which the rules of evidence are not enforced, in which the standard of proof is no higher than in an ordinary civil case, and in which the judge’s decision will make the difference between a light punishment and a punishment that is the maximum that our system allows short of death.

Posner stressed that "I have no desire to attempt an end run" around what he called "[t]he extraordinary severity of the punishments prescribed by Congress for sellers of marijuana." But he suggested that in cases like this one, in which "the real trial occurs at the sentencing stage," and the resulting increase in the defendant’s sentence is so large, the court should consider requiring prosecutors to prove the alleged conduct by clear and convincing evidence, rather than the usual 51-49 odds.

Posner suggested that such a rule might be required as a matter of due process, especially in light of the fact that "evidence presented at a sentencing hearing is often unreliable, both because the rules of evidence are not enforced in such hearings and because evidence of quantity of drugs sold is ordinarily given by criminals, the defendants’ former associates."

Alternatively, he suggested (as have some other federal appellate judges) that courts could adopt a clear-and-convincing evidence standard for cases like Rodriguez’s "simply as an intelligent rule of the federal common law of criminal procedure." He noted that neither the sentencing statutes nor the Federal Sentencing Guidelines specify the quantum of proof at a sentencing hearing, and that while the preponderance standard is the norm, the Supreme Court has "left open the possibility of imposing a higher standard of proof in cases in which the sentencing tail is wagging the guilt dog."

Indeed, in principle, it’s hard to see why any defendant should ever be sentenced on the basis of conduct allegedly related to the offense, but not proved at trial. And Posner acknowledged "the difficulty of defining the scope" of his suggested rule, which would raise the standard of proof only for especially dramatic increases in a defendant’s sentence. But it was probably politic of Posner to suggest only a modest, incremental change in prevailing practice.

Posner’s dissent adds a powerful voice to a growing chorus of proposals, by some of the best federal judges, of creative ways to temper the mechanistic harshness of the Federal Sentencing Guidelines and statutes as they have widely been interpreted so far.

And the Rodriguez case illustrates the extreme consequences-probably unforeseen by Congress- that sometimes flow from one aspect of the regime ushered in by the Sentencing Reform Act of 1984 and subsequent developments: the requirement that judges generally must give defendants the sentences prescribed not merely for the "offense of conviction" proven beyond a reasonable doubt to the jury, but also for any and all evidence of other "relevant conduct" that the prosecutor chooses to throw in at the sentencing stage.

As Marc Miller and Ronald Weich noted in a 1989 article in the Sentencing Law Reporter, "Laymen and law students are often surprised to learn that defendants may receive sentences based, in part, on uncharged conduct, conduct contained in dismissed counts of an indictment, and even conduct that a jury has found not proved beyond a reasonable doubt."

There is nothing new about judicial latitude to consider a broad range of evidence of untested reliability at sentencing hearings. The seminal case is Williams v. New York in 1949; the Supreme Court upheld as consistent with due process the traditional freedom of sentencing judges to consider all evidence bearing on the conduct that led to the conviction, along with all other "information about the convicted person’s past life, health, habits, conduct, and mental and moral propensities … even though obtained outside the courtroom from persons whom a defendant has not been permitted to confront or cross-examine."

Underlying Williams is the sensible proposition that the process of making the punishment fit the crime and the criminal should involve consideration of a wide range of factors regarding the defendant’s character that cannot feasibly be proven with the procedural rigor required to find a defendant guilty of a specified criminal charge at trial.

But the Federal Sentencing Guidelines, and the mandatory minimum sentencing statutes adopted over the past decade, have ushered in an important change: While Williams gave trial judges discretion to enhance sentences based on evidence aired for the first time at sentencing hearings, federal law now requires trial judges to crank all evidence of "relevant conduct" into their computations, with no explicit exceptions for cases in which the resulting increase in a sentence is disproportionate to the offense of conviction or otherwise unjust.

The purpose of this requirement is to serve the larger goal of reducing the unwarranted sentencing disparities associated with the old regime of untrammeled judicial discretion to consider or to ignore such evidence.

But the new regime has had an unanticipated consequence: It gives prosecutors incentives in some cases to seek enormous prison terms based on evidence too flimsy to be proven beyond a reasonable doubt to a jury.

Indeed, a substantial number of defendants around the country have been acquitted by juries of some counts of their indictments, only to find themselves being sentenced as though they had been convicted on those counts. That follows from the law of most circuits that if a judge finds at a sentencing hearing, by a mere preponderance of the evidence, that, in fact, you committed all the crimes charged in your indictment, the judge should sentence you on that basis even if the jury acquitted you of some of those charges.

While harsh results are sometimes required by the new sentencing regime, however, there is also considerable room for judges to show some courage and creativity in interpretation-including judges who (in Posner’s words) "accept absolutely the power of Congress to adopt these policies,…. if not their wisdom."

For example, the 9th Circuit has held, and judges including Chief Judge Jon Newman of the 2nd Circuit have argued in dissent, that the guidelines cannot constitutionally be read as mandating increases in a defendant’s sentence on the basis of "relevant conduct" of which he was acquitted. And now comes Judge Posner with a cogent legal basis for ensuring greater reliability in sentencing when the stakes are extremely high.

What remains to be seen is whether the Supreme Court will encourage those lower court judges who are seeking to bring some sense to the sentencing regime, or will acquiesce in the Justice Department’s view that judges have little choice but to rubber-stamp the sentences proposed by prosecutors. Some answers may come in two important sentencing cases that the Court is set to hear on Feb. 20 and 27: Koon v. United States and Melendez u United States.