The Death Penalty's Strange Career

On June 29, 1972, the U.S. Supreme Court handed down one of the most surprising decisions in its history. By a vote of 5 to 4, it ruled in Furman v. Georgia that every existing death penalty law in the United States was unconstitutional.

The ruling touched off the biggest flurry of capital punishment legislation the nation had ever seen. The day after Furman, legislators in five states declared their intention to introduce bills to resurrect the death penalty. President Richard Nixon asked the Federal Bureau of Investigation to supply him with incidents in which convicted killers had committed a second murder after being released from prison. In California, where the state supreme court had ruled that the state constitution barred capital punishment, support for the death penalty was strong enough to propel the issue to the ballot in November 1972. The voters reinstated the death penalty by a 2 to 1 margin. By 1976, four years after Furman, 35 states and the federal government had enacted new capital punishment statutes.

Public opinion on capital punishment shifted dramatically within months of the U.S. Supreme Court’s decision. In March 1972, a few months before Furman, supporters of the death penalty outnumbered opponents just 50 percent to 42 percent, according to a Gallup poll. By November 1972, the margin was 57 percent to 32 percent. An eight-point margin had grown into a 25-point margin in seven months. By 1976 supporters outnumbered opponents 65 percent to 28 percent, the widest gap since the early 1950s. The shift was uniform across all regions of the country. The belief that Americans had repudiated the death penalty—the linchpin of abolitionists’ constitutional argument in the Furman case—had been decisively disproven.

Neither the poll results nor the number of states with statutes authorizing capital punishment would change much in the ensuing decades. This suggests that the swing back to the death penalty would have taken place eventually, with or without Furman. In the long history of the death penalty, periods of strong abolitionist sentiment—some states eliminated the death penalty as early as the antebellum period—have always been followed by sharp drops in support for abolition. In the last three decades of the 20th century, growing public demand for law and order meant that an era of restoration was likely anyway.

But if Furman did not influence the direction of change, it almost certainly influenced its speed. Suddenly, capital punishment was a more salient issue than it had been in decades, perhaps ever. People who previously had had little occasion to think about the death penalty now saw it on the front page. Furman, like other landmark Court cases such as Roe v. Wade (1973), had the effect of calling its opponents to action.

The new death penalty statutes were drafted to conform to the opinions of the two justices who had held the balance of power in Furman, Potter Stewart and Byron White. What had troubled them about the death penalty was its randomness. When juries were given complete discretion to choose between life and death, the two justices concluded, the resulting pattern of verdicts had no rhyme or reason. On identical facts, one jury might sentence one defendant to death, while another jury might sentence another defendant to prison.

There were two ways to correct the problem, and some states tried each. One solution was to take discretion away from the jury by returning to the old practice of defining a class of crimes for which the penalty would always be death. In North Carolina, for instance, death became the mandatory sentence for first-degree murder and aggravated rape. The other solution was to legislate standards that would narrow the jury’s discretion in determining who would live and who would die. For guidance the states looked to the Model Penal Code, drafted a decade earlier by a group of eminent lawyers, judges, and law professors. It listed aggravating circumstances (such as a previous conviction for a violent felony) and mitigating circumstances (such as the defendant’s youth). In order to sentence the defendant to death, the jury would have to find at least one aggravating circumstance present. The jury was then to weigh the aggravating and mitigating circumstances in deciding on the sentence.

The new sentencing schemes were put to immediate use. In 1974, a total of 149 people were sentenced to death, probably more than in any year since 1942. (The U.S. Justice Department did not collect such data from 1951 to 1959.) The next year, 298 people were sentenced to death, far more than in any previous year for which data exist. The lawyers who had battled for years to persuade the Supreme Court to abolish the death penalty had inadvertently created a monster.

No executions could be carried out, however, until the Supreme Court had ruled on the constitutionality of the new sentencing schemes. The Court announced in January 1976 that it would hear appeals of five murder cases from different states, cases that would become collectively known as Gregg v. Georgia.

The Legal Defense Fund (LDF) and its principal litigator, law professor Anthony Amsterdam, led the argument against the death penalty, as they had in Furman and many earlier cases. Each of the states had its own lawyer, but they were overshadowed by Solicitor General Robert Bork, who filed a brief for the Ford administration seeking to overrule Furman. The case quickly became a contest between two of the foremost lawyers of the era: Amsterdam, on the faculty at StanfordUniversity, who had devoted his career to abolishing the death penalty, and Bork, on leave from YaleUniversity to serve as solicitor general, who had become the nation’s leading advocate of the constitutionality of capital punishment.

Amsterdam and the LDF faced a strategic puzzle. They had advanced two kinds of arguments in Furman: a procedural argument, that the means by which capital punishment was imposed rendered it cruel and unusual punishment; and a substantive argument, that the death penalty was unconstitutional regardless of how it was administered. The substantive argument had commanded only two votes on the Court in Furman, and it was not likely to do any better in Gregg. The procedural argument had been the winner, but now the states had corrected the procedural flaws the LDF had identified. To have any hope of success, the LDF would have to find procedural problems in the new statutes. But making that argument would open the LDF lawyers to the charge that by their interpretation no death penalty procedure could ever satisfy the Constitution. And if that charge were justified, the procedural argument would turn into the very substantive argument the LDF needed to avoid.

The LDF’s briefs all made the same point. The sentencing schemes of all five states purported to do away with discretion in the choice between life and death, but all they really did was shift that discretion to other parts of the criminal process. "Prosecutorial charging and plea-bargaining discretion, jury discretion to convict of one or another amorphously distinguished capital or non-capital crime, and gubernatorial discretion to grant or withhold clemency are all equally uncontrolled and uncontrollable," the LDF contended. "In its parts and as a whole, the process is inveterately capricious."

There was nothing else the lawyers could say, but the argument inevitably led Amsterdam into trouble at oral argument. Chief Justice Warren Burger was the first to pounce: "Since there is always an initial discretion on the part of the prosecutor, and...at the far end a power of clemency by an executive," he pointed out, "then no statutes can meet [your] standards."

Amsterdam was in a bind. If he agreed, he would be conceding that he was in fact arguing that capital punishment was unconstitutional under all circumstances, and he would lose. If he disagreed, he would be asked to identify the kind of statute that would meet constitutional requirements—that is, asked to identify the circumstances under which he would concede defeat. Amsterdam did the best anyone could do in the situation: He responded that he would "eventually take the position" Burger accused him of taking, but that it was "not a position that needs to be taken in this case" in order for the Court to rule in his favor.

But the issue could not be avoided. "Suppose just one crime, say, air piracy, and nothing else," Justice John Paul Stevens posited. "Would your argument about total discretion render such a statute unconstitutional?" The question put Amsterdam back in the same bind. If he said no, he would be telling his adversaries how to bring back capital punishment. If he said yes, he would be confirming Stevens’s suspicion that the LDF’s argument would have the effect of invalidating every conceivable sentencing scheme. Amsterdam struggled to answer, but the dilemma was irresolvable: Either the states could draft constitutional statutes or they could not.

At the justices’ conference two days later, most of the votes were unsurprising. William Brennan and Thurgood Marshall stuck with the positions they had taken in Furman: Capital punishment was unconstitutional, period. Burger, Harry Blackmun, and William Rehnquist stuck with their positions too. If, as they believed, the statutes at issue in Furman were constitutional, the new ones were easily so. Byron White, who had joined the majority in Furman, found that all five states had satisfied his original concern with arbitrariness, so he joined the three Nixon appointees in voting to uphold the statutes.

That left the decision in the hands of Stewart, Stevens, and Lewis Powell. "In light of what 35 states have done since 1972," Stewart explained, one "can no longer argue that capital punishment is incompatible with evolving standards of decency." Stewart and his two fellow swing voters joined Brennan and Marshall in declaring the mandatory death penalty unconstitutional. There was still too much discretion in the process. But the trio joined the four other justices in approving the sentencing schemes that guided the jury with instructions about the circumstances surrounding a crime. The opinions were published on July 2, 1976, almost exactly four years after the Court had declared the death penalty unconstitutional in Furman.

Capital punishment was back. Six and a half months later, Gary Gilmore of Utah became the first person to be executed in the United States in a decade.

The death penalty’s popularity held steady for the rest of the century.

Between 1977 and 1998, the percentage of those polled who favored capital punishment for murder fluctuated between 66 and 76 percent. The percentage who opposed it fluctuated between 19 and 28 percent. (Some people report no opinion, so the percentages do not add to 100.) The level of public support was higher than at any time since the first polls on the issue were taken, in the 1930s. It was remarkably consistent across regions and demographic groups. The only significant disparity in attitudes turned on race, unsurprisingly, but people of all races tended to favor the death penalty. White people just liked it more. Whites annually favored capital punishment by approximately a 4 to 1 margin, while the margin was much smaller among nonwhites. There were other demographic differences, but none were very large. Men favored the death penalty a bit more than women, Republicans a bit more than Democrats, the rich a bit more than the poor.

If only a small minority of Americans considered themselves opponents of the death penalty in principle, a majority harbored reservations about it when presented with alternatives. In the late 1980s and early 1990s, when polls were rephrased to ask whether murderers should be sentenced to death or to life in prison without parole, slightly less than half of respondents expressed a preference for the death penalty. When the alternative to death was life in prison without parole plus restitution to the victim’s family, support for the death penalty dropped to around 30 percent.

It was nevertheless true that in the 1980s and 1990s the great majority of Americans, in all parts of the country, favored the death penalty at least as an option. For an elected official to disagree with that sentiment in public was often tantamount to giving up hope of continuing a career in public office. In 1988, many observers concluded that Michael Dukakis lost any chance of winning the presidency after he emphasized his opposition to capital punishment during one of his televised debates with George H. W. Bush. Four years later, in the midst of the 1992 campaign, Governor Bill Clinton made it a point to return to Arkansas to sign the death warrant for Ricky Rector, a brain-damaged inmate so oblivious to his fate that he planned to save the dessert from his last meal to eat after his execution.

While there was not much regional variation in public opinion, there were striking differences in practice. By the 1990s there were 38 states with death penalty statutes, only three more than in 1976. Of the 12 without such laws, nine were in New England or the northern Midwest. (Seven had abolished capital punishment long before: Michigan, Rhode Island, and Wisconsin before the Civil War; Iowa and Maine in the late 19th century; and Minnesota and North Dakota in the years before World War I.) New England and the northern Midwest were the only parts of the country where homicide rates were considerably below the national average. It may be that in those regions capital punishment was popular but not particularly salient—most supporters simply did not consider the issue important.

There were also pronounced regional differences in the pattern of executions and death sentences. Of the 598 executions conducted between 1977 and 1999, all but a handful took place in the South. Texas was the leader, with 199, followed at some distance by Virginia (73), Florida (44), and Missouri (41). The leader among the northern states was Illinois, with only 12.

If capital punishment as a general policy was no more popular in the South than in the North, why did the southern states have so many more death sentences? And why was the distribution of executions so much more uneven than the distribution of death sentences? Race was not the answer. By the 1980s and 1990s, black defendants were no more likely than white defendants to be executed in most states. So why were executions so much more frequent in the South?

One cause was the fact that the murder rate was much higher in the South than in the North. In most years between 1976 and 1998, the homicide rate in the four-state area of Texas, Louisiana, Arkansas, and Oklahoma was three to four times greater than in New England. The number of death sentences in a state in the decades after Furman was closely correlated with the number of homicides in that state. Southerners had more opportunities to impose the death sentence than northerners did, and the prevalence of murder may have made them more willing to impose it in any given case.

But differences in murder rates were most likely too small to account for the North-South disparities in death sentences and executions. By the end of the century the southern states were conducting as many executions as they had in the 1940s, but executions were still rare in the North. The remainder of these regional differences was probably attributable chiefly to disparities in the way states provided defense lawyers. Defendants charged with capital murder were almost always too poor to pay a lawyer. In most of the northern states, capital trials were handled by experienced public defenders, many of whom specialized in capital cases. In most of the South, by contrast, capital defendants were represented by lawyers in private practice who were appointed by trial judges to handle individual cases. Compensation was so low that it often attracted the least-skilled segment of the bar—and many of these lawyers had no experience in criminal matters. Many made no effort to gather evidence that might help their client avoid a death sentence. Horror stories abounded of defense lawyers who slept through parts of the trial or used racial epithets to refer to their own client before the jury. Similarly inept appellate counsel ensured that death sentences were upheld on appeal. In these states, someone accused of a capital crime might obtain a competent lawyer only after his execution date had been set—too late to make a difference in most cases.

Of all the aspects of capital punishment’s popularity, perhaps the most curious was the increasing irrelevance of what had once been a crucial question: whether capital punishment deters murder any more than prison does.

That issue, a staple of the debate since the early 19th century, was taken over in the 1970s by economists. They created equations expressing the murder rate as the product of a host of different variables, one of which was the likelihood of being executed. They then used the statistical technique called multiple regression to measure the effect on the murder rate of changes in that one variable, while holding all the others constant. The first economist to use this technique was Isaac Ehrlich, who in 1975 calculated that each execution prevented approximately eight murders. The finding received enormous public attention, because it appeared just as the Supreme Court was preparing to consider the post-Furman death penalty cases.

Ehrlich did not lack for critics. They pointed out, among other things, that his list of factors was short; including more factors that contributed to the murder rate might reduce or even eliminate the measured impact of the death penalty. Other critics demonstrated that Ehrlich’s results were sensitive to tiny changes in the data used. He had studied the period from 1933 to 1969, for example, but if the five most recent years were removed, the deterrent effect disappeared.

Neither Ehrlich’s work nor that of his critics had much effect on the Court’s decision in Gregg. The debate did attract a swarm of social scientists to the attempt to measure deterrence, and by the end of the century there was an abundant literature in academic journals of law and economics. A few studies found a deterrent effect, but most did not. There was a raging methodological disagreement over how best to pick the variables, and a nagging suspicion that researchers’ own attitudes toward capital punishment were subconsciously influencing the forms of equations. The lack of academic consensus allowed advocates in the public-policy arena to choose the studies that supported their own views. Much of the public, meanwhile, stuck to the pervasive folk wisdom that the death penalty had to have a deterrent effect.

It soon became apparent, however, that the popularity of capital punishment had little to do with deterring crime. Surveys conducted between 1983 and 1991 uniformly indicated that a large majority of supporters would still favor the death penalty even if it had no effect whatsoever on the murder rate. They valued capital punishment for two other purposes. Both were very old, and both had been in abeyance for some time.

One was retribution. Long rejected as a legitimate goal of punishment in academic and policymaking circles, retribution made an astonishingly fast comeback. Part of its rise was a reaction to the widespread loss of faith in the power of prisons and other institutions to rehabilitate criminals. Part grew out of the resurgence of causal models of crime that rested on the free will of the criminal rather than on social or biological forces beyond the criminal’s control.

Speaking in favor of the death penalty before a committee of the New Jersey Senate in 1982, a representative of the state attorney general’s office made it clear that his opinion had little to do with deterrence: "The idea that the punishment must fit the crime is something more than . . . the idea that somehow we ought to try to discourage others from committing crimes by imposing prison sentences and other forms of punishment.... Somehow society needs to feel that when a criminal act has been committed, its interests have been vindicated."

The point was made again and again: Capital punishment was a moral imperative, regardless of whether it reduced the murder rate or cut murderers off from the possibility of rehabilitation. Sometimes retribution was cited as an instrumental value, as in previous centuries. The anger people felt toward criminals, disparagingly labeled "revenge" by the previous generation of criminologists, was in fact the glue that held society together, argued political scientist Walter Berns. The criminal law "must remind us of the moral order by which alone we can live as human beings," Berns concluded, "and in our day the only punishment that can do this is capital punishment." But it was probably more common to think of retribution as an end in itself, as an emotional need that only an execution could fulfill.

The second purpose that seemed to be served by the death penalty was harder to defend intellectually but may have been more important. Back in the days of public hangings, an execution had been a vehicle for a collective condemnation of crime. Going to a hanging was a way of siding with the community against the criminal, a means of broadcasting the seriousness with which one took crime and its consequences. When the ceremony was moved indoors, the execution lost much of its purpose as a vehicle of denunciation. Beginning in the 1970s that symbolic function returned, this time attached not to the ceremony of execution but to support of capital punishment as an abstract policy. To declare in favor of capital punishment was often implicitly to announce that one wanted to "get tough on crime" in order to reduce its frequency, that criminals ought to be held morally responsible for their actions, that crime was chosen by criminals rather than forced upon them by genes or their environment, and that the worst criminals were unlikely candidates for reintegration into society. These were the same symbolic statements that had once been made by spectators at public executions.

What was unfortunate about this shift was that it greatly muddied the debate by permitting support for capital punishment to be invoked in situations in which the death penalty could not conceivably be applied. When the New Jersey legislature was considering a bill to reinstate the death penalty, one senator said that he had been deluged by pro-capital punishment letters and telephone calls from people who recounted their own experiences as assault victims. "Almost all these letters ask the same questions: Why don’t our laws protect us? . . . What has happened to justice in our country?" The frequency of assault and other low-level crimes could hardly be affected by capital punishment for murder, but that was not the point. Support for the death penalty was a shorthand way of expressing concern about crime generally.

Elected officials were quick to capitalize on this symbolism. By 1998, the federal criminal law included no fewer than 46 capital crimes, virtually all of them variations of murder defined so narrowly and yet with so much overlap among them that one suspects Congress was motivated chiefly by a desire to claim credit for putting a large number of death penalty laws on the books.

Capital punishment after Gregg was not only a political issue. The Supreme Court’s involvement turned it into a constitutional issue as well. Within a very short time the Court constructed an intricate Eighth Amendment jurisprudence on the foundation of Furman and Gregg, a body of cases distinguishing the practices that would or would not amount to cruel and unusual punishment. The result was a significant shift in decision-making authority among the three branches of government. Issues that had once been decided by legislatures, or by governors during the clemency process, now became constitutional questions to be decided by courts.

Was capital punishment too severe for crimes less grave than murder? The question had been the subject of fierce political debate within legislatures since the late 18th century. Governors had always considered the gravity of the crime in deciding whether to grant clemency. But after Furman and Gregg, the issue was recast as a constitutional question: Would it violate the Eighth Amendment to execute a criminal for committing a crime short of murder? In Coker v. Georgia, only a year after Gregg, the Court held that the death penalty was a cruel and unusual punishment for rape. What about a defendant technically guilty of murder who was not the actual killer? The criminal law had always held accomplices guilty of the crime they helped another commit, but a defendant’s minimal participation had always been a factor tending toward clemency. Now it became a constitutional question: Was it cruel and unusual to execute the accomplice? In 1982 the Court held that it was, by a 5 to 4 vote; in 1987, after Justice White switched sides, the Court held that it was not, also by a 5 to 4 vote.

What if the defendant were very young? The Court held that the Eighth Amendment permitted the execution of a defendant who was 16 years old at the time he committed the crime. What if the defendant were mentally retarded? The Court held that the Eighth Amendment did not prohibit executing people with mental retardation. These had been classic legislative or clemency issues for hundreds of years, but now they were novel constitutional questions.

New questions arose. There was endless controversy over whether the states’ definitions of aggravating and mitigating circumstances were too broad or too narrow to guide juries. In the end, the Court held that the states could not restrict the jury’s consideration of mitigating evidence—that the jury must be allowed to consider any kind of evidence that might point against a death sentence, not just the evidence relevant to one of the statutory mitigating circumstances. That conclusion went halfway toward undermining the constitutional regime created by Furman and Gregg, under which state statutes were supposed to channel the jury’s consideration of evidence at sentencing to prevent the random imposition of death sentences.

Most of the other half of the decision, the identification of aggravating circumstances, was cut loose from statutory guidance not long after, when the Court allowed sentencing juries to consider nonstatutory aggravating evidence as well. By this point, all that was left of the constitutional framework was the require-ment that the jury find a single statutory aggravating circumstance before proceeding to what had become a virtually unguided exercise of discretion.

For a time the Court did exclude one kind of evidence from sentencing, evidence of the effect of the murder on the victim’s family and friends, but that was by a 5 to 4 vote. In 1991, the Court overruled its prior cases and let in such "victim impact" evidence as well. After 1991, well-conducted capital sentencing hearings normally included emotional presentations by both sides, matching the defendant’s weeping relatives against the victim’s weeping relatives, in an effort to gain the sympathy of the jury. Any pretense that this was a rational process of distinguishing degrees of culpability was long gone.

In the 20 years after Gregg, capital punishment occupied a significant percentage of the Court’s time; accruing, as a result, were scores of cases making up a complex and ever-shifting body of law. Justice Antonin Scalia, among other critics, complained of "the fog of confusion that is our annually improvised Eighth Amendment, ‘death is different’ jurisprudence."

Much of the fog was produced by the Court’s ceaseless effort to reconcile two irreconcilable goals—consistency across cases, which is best achieved by formal rules restricting jury discretion; and attention to the unique characteristics of each case, which is best achieved by allowing the jury unrestricted discretion. In 1994, a few months before he retired, Justice Blackmun finally gave up. "Over the past two decades, efforts to balance these competing constitutional commands have been to no avail," he despaired. "From this day forward, I no longer shall tinker with the machinery of death." But the rest of the Court tinkered on.

Many areas of the law are complex, but the tragedy of the Court’s Eighth Amendment jurisprudence was that all the complexity served scarcely any purpose. Trials were long and expensive, lawyers had to master bodies of arcane doctrine, every case raised several issues that could be plausibly litigated on appeal, and yet, for all that, the process of distinguishing the murderers who would be executed from those who would be sent to prison seemed no less haphazard than it had been before.

There was one piece of good news. Before Furman, it was common knowledge that black defendants were sentenced to death at higher rates than white defendants. Econometric studies conducted after Gregg revealed a less consistent pattern. In some states, the race of a defendant was no longer a factor influencing the likelihood of a death sentence. In some states black defendants were still disadvantaged, but in others white defendants were now disadvantaged.

This change almost certainly had little to do with the new sentencing schemes. It was most likely a product of two other developments. First was the Court’s holding in Coker v. Georgia that the Eighth Amendment barred capital punishment for rape. Rape had always been the crime for which the race of the defendant made the biggest difference in the sentence. Second was the fact that blacks gained better representation on juries after the 1960s, especially in the South, where most of the death sentences were imposed.

Capital sentencing was not free from racial disparities, however. Econometric studies revealed a pronounced bias based on the race not of the defendant but of the victim. The first and most extensive of the studies, conducted in Georgia, showed that when all other variables were held equal, a death sentence was 4.3 times more likely when the victim was white. Similar results were obtained in other states. Abolitionists quickly adopted these findings as an argument against the death penalty. Capital punishment, they contended, undervalued the lives of black victims. But the implications were not entirely clear. Would things be better if more killers of black victims were sentenced to death? Because most murders involved criminals and victims of the same race, that would cause more black defendants to be sentenced to death.

The race-of-victim disparity was the vehicle for the LDF’s last serious effort to persuade the Supreme Court to declare capital punishment unconstitutional, in McCleskey v. Kemp (1987). Race discrimination had been the original reason for the LDF’s involvement in death penalty litigation in the 1960s and it had been that silent specter that had prompted the Court to require statutory standards to guide the jury’s discretion.

The persistence of racial differences, the LDF argued, demonstrated that the "post-Furman experiment has failed."

The argument fell one vote short of a majority. Lewis Powell, who wrote the majority opinion, firmly believed that the pattern of results in thousands of cases should never upset the verdict in a single case. "My understanding of statistical analysis—particularly what is called ‘regression analysis’—ranges from limited to zero," he confessed to his law clerk. But he was well aware that allowing statistical attacks on criminal convictions promised to open a Pandora’s box. What about other minority groups? What about gender disparities? Everyone knew that women were very rarely executed—did that violate the constitutional rights of men? What if there were racial or other disparities in the length of prison sentences? The LDF was "attacking the jury system," Powell noted to himself.

Suffusing the Court’s opinion in McCleskey was a weariness, a pessimism about the possible. "Apparent disparities in sentencing are an inevitable part of our criminal justice system," Justice Powell wrote. "The Constitution does not require that a State eliminate any demonstrable disparity." Fifteen years after Furman, the Court had given up hope of eliminating the arbitrariness that had once been the motor of constitutional change.

If the constitutionalization of capital punishment failed to impose any order on the task of distinguishing which criminals would live or die, it had a profound impact on the death penalty considered more broadly. Clemency rates suddenly declined, chiefly because most of the factors that governors once weighed in considering whether to commute a sentence were now handled in the courts. Abolitionists, seeing that success in the political arena was unlikely, put more energy into courtroom battles. One result was that the average period between sentencing and execution grew from 51 months during 1977–83 to 134 months by 1995.

Jury selection could take weeks. The sentencing phase of a capital trial, if conducted skillfully on both sides, was a battle of philosophies. The prosecutor told a story of free will, of a criminal with the opportunity to choose between good and evil. Defense counsel countered with a narrative of determinism, of social and biological forces that would have driven anyone to crime. This was a very old battle, dating back to the late 18th century, but it was a battle that had always been fought in the public, political arena, over whether capital punishment ought to exist at all. Now it was fought in the capital trial itself.

The constitutionalization of capital punishment created an enormously complicated, expensive, and time-consuming apparatus that had little real effect on the outcome of cases. Being executed was still, as Justice Stewart had put it in Furman, like being struck by lightning; the only difference was that it now took a decade and millions of dollars of public money for the lightning to strike.

At the beginning of the 21st century, capital punishment is once again a firmly established part of American criminal justice. Death sentences and executions have become so commonplace in some states that they are no longer news. And the execution rate, which has dropped slightly in the past three years, seems poised to skyrocket. As of January 1, there were 3,711 prisoners on death row. The annual number of death sentences regularly exceeds the annual number of executions by a factor of three. As more and more of these inmates reach the end of their appeals, if all else stays the same, the execution rate is likely to reach several hundred per year. The abolitionist movement is weak, and the Supreme Court seems unlikely to introduce any new constitutional limits on capital punishment. The death penalty looks as if it is back to stay.

If there is any hint of a possibility of change, it is in the mounting number of innocent people turning up on death row. The risk of executing the innocent has haunted capital punishment for centuries, but until the post-Furman era it was a problem handled by executive clemency. With the decline of clemency, there is no longer any routine mechanism for resolving post-trial claims of innocence. When an innocent person is sentenced to death, his best hope is that his cause will be taken up by someone with the time and resources to conduct a thorough investigation. Such people are rare, but they have nevertheless produced some startling results. Between 1987 and 1999, a total of 61 condemned inmates were released from prison because they were discovered to be innocent. A few were beneficiaries of DNA testing, a technology unavailable when they were convicted, but most were not. Most had been victims of dishonest witnesses, prosecutors, or police officers, whose lies were found out only years later.

Support for capital punishment has diminished only slightly as a result of the revelations. The execution of Timothy McVeigh this past June demonstrated that when a criminal is clearly guilty and his crime especially horrible, the death penalty is as popular as ever.

Yet if any current development has the potential to alter public opinion, the execution of innocents is it. If even more such cases were to come to light, or if a sympathetic and apparently innocent person were to be executed, support for the death penalty could conceivably dwindle quickly. In the past, when the market for news was largely local, high-profile cases could quickly tip public opinion in particular states one way or the other. Today, with national media spreading information about a single crime or a single defendant to every part of the country, the right case might have a similar effect—this time nationwide.