Homicide may have a political dimension.Illustration by Anders Wenngren

Steven Hayes and Joshua Komisarjevsky, who met three years ago in a Hartford drug-treatment center and shared a room in a halfway house in between stints in prison, were both seasoned burglars, though Hayes, a forty-four-year-old crack addict, was quite a bit older than Komisarjevsky, who was twenty-six, and the great-grandson of a Russian princess. In the spring of 2007, both men were paroled. Hayes, whose arrest record stretches back to 1980, had served about three years of a five-year sentence for third-degree burglary, and Komisarjevsky had finished half of a nine-year sentence for burglary in the second degree. Hayes moved in with his mother, in Winsted, in Litchfield County; Komisarjevsky went back to his home town, Cheshire, a suburb about fifteen miles north of New Haven. They kept in touch. On July 23, 2007, authorities say, Hayes and Komisarjevsky broke into the Cheshire home of William Petit, Jr., an endocrinologist, and tortured the family through the night, raping Petit’s wife, Jennifer Hawke-Petit, and at least one of the couple’s two daughters. In the morning, Hayes and Komisarjevsky are said to have forced Hawke-Petit, a school nurse who suffered from multiple sclerosis, into the family car and taken her to a local bank, where she withdrew fifteen thousand dollars, after which a suspicious teller alerted the police. The two men allegedly then took Hawke-Petit back to the house, killed her, set the house on fire, and fled in the Petits’ S.U.V., though not far: they crashed into a police barricade, just past the driveway.

Inside the house, a four-bedroom Colonial, police found three bodies. Hawke-Petit, forty-eight, had been strangled. Seventeen-year-old Hayley Petit, who, that September, was to start college at Dartmouth, died of smoke inhalation. Her eleven-year-old sister, Michaela, was found tied to a bed, her body badly burned after having been doused with gasoline. Only William Petit, who had been bound with rope, beaten in the head with a baseball bat, and left for dead in the cellar, survived.

Hayes and Komisarjevsky have been charged with kidnapping, sexual assault, arson, and murder. Jury selection for Hayes’s trial is scheduled to begin in January, in New Haven. William Petit, who is expected to testify about what happened that night, had asked not to be put through that ordeal twice, but his request for a single trial was denied. A trial date for Komisarjevsky has not yet been set. The state is seeking the death penalty.

Every murder raises terrible questions that no trial, no law, no punishment can answer. What forces make it possible for one human being to take the life of another? Murders can be solved and even explained—at least, that’s the operating assumption of criminal investigation and the narrative logic behind every whodunit—but to think about a specific murder with any clarity, or for very long, can be difficult, and viscerally painful. Maybe the brisk trade in lurid violence as spectacle has something to do with it: one either watches or averts one’s eyes; dispassionate reflection rarely enters into it. Scholars ranging from theologians and psychologists to evolutionary biologists have offered theories about murder—theories of evil, theories of disease, theories of disposition—but the analytical burden placed on any general discussion of murder, freighted, as it is, with atrocity, is nearly unbearable. Nothing suffices, or can.

Between the convulsive emotional response to a single murder and an elusive general theory of murder lies another kind of contemplation: the study of the murderousness of nations. The United States has the highest homicide rate of any affluent democracy, nearly four times that of France and the United Kingdom, and six times that of Germany. Why? Historians haven’t often asked this question. Even historians who like to try to solve cold cases usually cede to sociologists and other social scientists the study of what makes murder rates rise and fall, or what might account for why one country is more murderous than another. Only in the nineteen-seventies did historians begin studying homicide in any systematic way. In the United States, that effort was led by Eric Monkkonen, who died in 2005, his promising work unfinished. Monkkonen’s research has been taken up by Randolph Roth, whose book “American Homicide” (Harvard; $45) offers a vast investigation of murder, in the aggregate, and over time. Roth’s argument is profoundly unsettling. There is and always has been, he claims, an American way of murder. It is the price of our politics.

In the archives, murders are easier to count than other crimes. Rapes go unreported, thefts can be hidden, adultery isn’t necessarily actionable, but murder will nearly always out. Murders enter the historical record through coroners’ inquests, court transcripts, parish ledgers, and even tombstones. “Fell by the hands of William Beadle / an infatuated Man who closed the / horrid sacrifice of his Wife / & Children with his own destruction,” reads the headstone of Lydia Beadle, of Wethersfield, Connecticut, who was murdered, along with her two children, in 1782. The number of uncounted murders, known as the “dark figure,” is thought to be quite small. Given enough archival research, historians can conceivably count, with fair accuracy, the frequency with which people of earlier eras killed one another, with this caveat: the farther back you go in time—and the documentary trail doesn’t go back much farther than 1300—the more fragmentary the record and the bigger the dark figure.

Pieter Spierenburg, a professor of historical criminology at Erasmus University, in Rotterdam, sifts through the evidence in “A History of Murder: Personal Violence in Europe from the Middle Ages to the Present” (Polity; $24.95). In Europe, homicide rates, conventionally represented as the number of murder victims per hundred thousand people in the population per year, have been falling for centuries. Spierenburg attributes this long decline to what the German sociologist Norbert Elias called the “civilizing process” (shorthand for a whole class of behaviors requiring physical restraint and self-control, right down to using a fork instead of eating with your hands or stabbing at your food with a knife), and to the growing power of the centralizing state to disarm civilians, control violence, enforce law and order, and, broadly, to hold a monopoly on the use of force. (Anthropologists sometimes talk about a related process, the replacement of a culture of honor with a culture of dignity.) In feuding medieval Europe, the murder rate hovered around thirty-five. Duels replaced feuds. Duels are more mannered; they also have a lower body count. By 1500, the murder rate in Western Europe had fallen to about twenty. Courts had replaced duels. By 1700, the murder rate had dropped to five. Today, that rate is generally well below two, where it has held steady, with minor fluctuations, for the past century.

In the United States, the picture could hardly be more different. The American homicide rate has been higher than Europe’s from the start, and higher at just about every stage since. It has also fluctuated, sometimes wildly. During the Colonial period, the homicide rate fell, but in the nineteenth century, while Europe’s kept sinking, the U.S. rate went up and up. In the twentieth century, the rate in the United States dropped to about five during the years following the Second World War, but then rose, reaching about eleven in 1991. It has since fallen once again, to just above five, a rate that is, nevertheless, twice that of any other affluent democracy.

What accounts for this remarkable difference? Guns leap to mind: in 2008, firearms were involved in two-thirds of all murders in the United States. Yet Roth, who supports gun control, insists that the prevalence of guns in America, and our lax gun laws, can’t account for the whole spread, and a few scholars have argued that laws allowing concealed weapons actually lower the murder rate, by deterring assaults. Some Europeans suspect that Americans haven’t undergone the same “civilizing process,” as if, unmoored from Europe, Colonial Americans went murderously adrift. Spierenburg speculates that democracy came too soon to the United States. By the time European states became democracies, the populace had accepted the authority of the state. But the American Revolution happened before Americans had got used to the idea of a state monopoly on force. Americans therefore preserved for themselves not only the right to bear arms—rather than yielding that right to a strong central government—but also medieval manners: impulsiveness, crudeness, and fidelity to a culture of honor. We’re backward, in other words, because we became free before we learned how to control ourselves.

Perhaps unsurprisingly, not everyone buys these arguments, and Monkkonen himself took a different, though equally conjectural, approach. At the time of his death, he had been working on an article called “Homicide: Explaining America’s Exceptionalism,” which hypothesized that four factors accounted for the centuries-long differences between American and European homicide rates: mobility, federalism, slavery, and tolerance. Mobility breaks social ties; federalism is a weak form of government; slavery not only rationalized a culture of violence among white Southerners (where the murder rate has been disproportionately high, as it has, and remains, in many of the so-called law-and-order states) but also infected American culture; and American judges and juries have historically proved less willing than their European counterparts to convict murderers, tolerating, among other crimes, racial murders and killings by jealous spouses.

Roth, who teaches at Ohio State, wants to bring into this debate hard facts and rigorous methods. He rejects arguments about the “civilizing process” by pointing out that people didn’t necessarily intend to murder one another more often in the premodern world; they merely succeeded more often. Given modern medicine—emergency response, trauma surgery, antibiotics, and wound care—three out of every four people murdered before 1850 would probably survive today. Roth heads a collaborative project, dedicated to Monkkonen, called the Historical Violence Database, which has assembled reports of murders in several of the original thirteen colonies; nineteenth-century records from five states, seven cities, and thirty-four counties; and a wealth of twentieth-century statistics, chiefly from the Uniform Crime Reports kept by the F.B.I. beginning in 1930. As a discussion of the available data, “American Homicide” is rich, fascinating, and unrivalled. As an explanation, though, it gets dubious. Roth’s work involves three steps: first, he uses his database to count murders (he’s primarily interested in homicides among unrelated adults); then, using surviving censuses to count people, he calculates the homicide rate; finally, he attempts to explain what factors correlate with that rate, across four centuries. It’s the last step that’s the most wobbly.

Historians haven’t studied murder much, but criminologists have. Although most criminologists trace the homicide rate back only a few decades, Roth takes his lead from their work. The fluctuations in the homicide rate since the nineteen-forties have at least something to do with demography. A vastly disproportionate number of murderers and murder victims are young adult men. When baby boomers reached that age bracket, the homicide rate soared. Now that they’ve aged out of their most lethal years, the rate has fallen. To Roth, the demographic explanation of the postwar crime boom and bust falls short, but, where other social scientists have investigated economic conditions like joblessness or government policies like gun control to fill the explanatory gap, Roth favors the argument made by a criminologist named Gary LaFree, in a book called “Losing Legitimacy: Street Crime and the Decline of Social Institutions in America” (1998). LaFree observed that the crime rate correlates, inversely, with public faith in government and trust in elected officials. So, for instance, the Vietnam era, marked by declining confidence in elected officials, experienced a rising crime rate. He measured that faith and trust by consulting national opinion surveys taken beginning in 1958, which asked questions like “How much of the time can you trust the government to do what is right?”

Roth attempts to graft LaFree’s argument onto all of American history. He has determined that four factors correlate with the homicide rate: faith that government is stable and capable of enforcing just laws; trust in the integrity of legitimately elected officials; solidarity among social groups based on race, religion, or political affiliation; and confidence that the social hierarchy allows for respect to be earned without recourse to violence. When and where people hold these sentiments, the homicide rate is low; when and where they don’t, it’s high.

Whatever you think about the value of public-opinion polls, LaFree at least had them. Roth doesn’t. How do you measure the belief that government is stable in 1695 or 1786 or 1814 or 1902? You can’t. You can only look at what was happening in those years and tell a story about what you think people believed about their government, and, if you know what the homicide rate is, it’s easy to find a story that fits your data. The homicide rate in New England fell from a high, in 1637, of a hundred and twenty to under one, in 1800, chiefly by dropping, rather dramatically, after the Pequot War and King Philip’s War. Roth argues that the rate fell, over all, as judicial institutions were established and people developed faith in them, and that the rate fell, sharply, after these wars because conflicts with hostile neighbors brought the colonists together. But it seems equally plausible to argue that the homicide rate in Colonial New England tracks the European decline quite nicely, over all, and drops, in a stepwise fashion, after wars because they diminish the population of young men, leaving fewer potential murderers and murder victims around. Both interpretations make sense; neither has been demonstrated.

The implications of Roth’s argument are, as he realizes, distressing. Democracy requires dissent. If a high American murder rate is a function of not placing our trust in government, are we doomed to endure a high murder rate? Roth takes his case all the way to the White House: “The statistics make it clear that in the twentieth century, homicide rates have fallen during the terms of presidents who have inspired the poor or have governed from the center with a popular mandate, and they have risen during the terms of presidents who presided over political and economic crises, abused their power, or engaged in unpopular wars.” The homicide rate appears to correlate with Presidential approval ratings. If Roth is right, electing a bad President is dangerous and inciting people to hate any President, good or bad, could be deadly. But which is the cart, and which the horse? The Presidential approval rate might be a proxy for all sorts of measures of a well or poorly adjusted society. Or maybe there’s another horse, somewhere, some third factor, that determines both the Presidential approval rate and the homicide rate. It’s hard to say, partly because, in using quantitative methods to make an argument about the human condition, Roth has wandered into a no man’s land between the social sciences and the humanities. After a while, arguments made in that no man’s land tend to devolve into meaninglessness: good government is good, bad government is bad, and everything’s better when everything’s better. Correlating murder with a lack of faith and trust may contain its horror, but only because, in a bar graph, atrocity yields to banality.

Every September, the F.B.I. issues a report on crime, a compilation of statistics for the previous year. It does not offer an interpretation of this immense quantity of data. “We leave that up to the academics and the criminologists and the sociologists,” an F.B.I. spokesman said, upon the release of this year’s report. For all the number crunching, it’s clear that there is no such thing as an average murder. Even if there were, what happened at the Petits’ house in Cheshire, Connecticut, on July 23, 2007, wouldn’t be it, and not just because of that crime’s particular depravity. Much about the case is out of the ordinary. The victims were white and wealthy; murder victims are disproportionately black and poor. Exceptional, high-profile crimes often lead to legislative action driven by citizen initiative. California’s controversial three-strikes law, a ballot measure, was proposed by a Fresno photographer whose daughter was murdered. Last year, after the Petit murders, the Connecticut legislature doubled and tripled mandatory penalties for second- and third-time offenders. “Big cases make bad laws” is a criminological axiom, and one with which Mark A. R. Kleiman agrees, in “When Brute Force Fails: How to Have Less Crime and Less Punishment” (Princeton; $29.95).* Kleiman blames big cases and bad laws for another distinctive feature of American life: 2.3 million people are currently behind bars in the United States. That works out to nearly one in every hundred adults, the highest rate anywhere in the world, and four times the world average. Prison crowding may have been one reason that Steven Hayes and Joshua Komisarjevsky were paroled. Although the crime rate today is fifteen per cent lower than it was twenty-five years ago, the incarceration rate is four times as high. At what point, Kleiman wonders, will incarceration be a greater social ill than crime? He proposes, for lesser offenders, punishments that are swift and certain but not necessarily severe: a night in jail, instead of a warning, for missing a meeting with a parole officer, say, and ten nights the next time. Whether or not Kleiman’s recommendations are practical, Connecticut, reeling from the Petit murders, is heading in the opposite direction.

The F.B.I. may leave the analysis of crime to academics, but, in the past few decades, the government has, increasingly, left the punishment of criminals up to public opinion. William Petit and his sister* Johanna Petit-Chapman serve as the honorary co-chairs of Three Strikes Now, a grass-roots organization lobbying the state legislature to adopt California-style mandatory sentencing of life without the possibility of parole for third-time violent offenders. The Cheshire case has also dominated the state’s death-penalty debate, a debate that, nationwide, has long centered on race. In Connecticut, whose population is eighty-four per cent white, six of the ten men on death row are black. (Both Hayes and Komisarjevsky are white.) Earlier this year, the Connecticut legislature voted to abolish the death penalty. William Petit publicly denounced the bill, and Jodi Rell, the state’s governor, a Republican, vetoed it.

Capital punishment has been on the books in Connecticut since 1642. Three strikes has been tried before, too. In Colonial America, many crimes, including murder, were punishable by death and, for lesser crimes, Connecticut, like many colonies, mandated the death penalty for third-time offenders. That began to change on September 7, 1768, when a burglar named Isaac Frasier was hanged in Fairfield. Frasier had shown early evidence of a “thievish Disposition.” “Men go from one degree of wickedness to another,” the town’s minister said in a sermon at the gallows titled “Excessive Wickedness, the Way to an untimely Death.” Convicted of burglary in New Haven, Frasier was whipped and branded and had his ears cropped. Caught again in Fairfield in 1766, he received the same punishment “and was solemnly warned . . . that death would be his punishment on a third Conviction.” When Frasier robbed another house, he was sentenced to death. “The Government of Connecticut have always been remarkably tender of putting persons to Death,” one observer noted. But when Frasier applied to the legislature for clemency, he was denied. Said the pastor at the gallows, “Justice requires that you should suffer.”

An outcry followed. Two weeks after Frasier’s death, a Hartford newspaper published an essay called “An Answer to a very important Question, viz. Whether any community has a right to punish any species of theft with death?” The writer’s answer—an emphatic no—borrowed extensively from Cesare Beccaria’s treatise “On Crimes and Punishments,” published in 1764. Beccaria, an Italian nobleman, argued against capital punishment—which was, at the time, widespread in Europe, too—on two grounds: first, in a republic men do not forfeit their lives to the government; and, second, capital punishment does not deter crime. Beccaria argued (and Kleiman has merely revisited that argument) that punishments, to be effective, must be swift and certain but not necessarily severe. Punishments, he insisted, should be proportionate to crimes, whose dangerousness could be measured, in “degrees,” by their injury to society. For the crime of murder, Beccaria considered life in prison to be both more just and a more effective deterrent than execution.

The first American edition of Beccaria’s treatise was published in 1777, and it reached a wide audience in Connecticut beginning in 1786, when it was serialized in a New Haven newspaper. “If we glance at the pages of history, we will find that laws, which surely are, or ought to be, compacts of free men, have been, for the most part, a mere tool for the passions of some,” Beccaria wrote. This argument held particular appeal for a people who had just finished waging a war against the passions of King George; adopting Beccaria’s recommendations came to seem, in a fundamental sense, American, as if the United States had a special role to play, as a republic, in the abolition of capital punishment. In 1784, the Yale senior class debated whether the death penalty was “too severe & rigorous in the United States for the present Stage of Society.”

In the seventeen-nineties, five states abolished the death penalty for all crimes except murder. By the eighteen-twenties, all Northern states reserved capital punishment for first-degree murder. When incarceration replaced all corporal and most capital punishment, Americans built prisons, and sentenced criminals to jail time. In 1846, Michigan became the first state to abolish the death penalty. Twice, in the middle of the nineteenth century, the governor of Connecticut asked the state’s legislature to do the same, to no avail.

In the course of the twentieth century, capital punishment was abolished in much of the world, including all of Western Europe, but not in the United States. Germany, Austria, and Italy stopped executing criminals after the Second World War. Beginning in the nineteen-fifties, other European countries began limiting capital punishment. Denmark abolished it entirely in 1978; the Netherlands, Australia, and New Zealand in the nineteen-eighties; Britain, Canada, and Belgium in the nineteen-nineties. In many parts of the United States, the death penalty was, if not outlawed, abandoned. Except for a serial murderer named Michael Ross, who was killed by lethal injection in 2005, after he waived his right to appeal because he wanted to die, no one has been executed in Connecticut, or anywhere else in New England, since 1960.

Not so elsewhere. Since 1976, more than a thousand people have been executed in the United States, a third of them in Texas. If Hayes and Komisarjevsky are found guilty and sentenced to death instead of life in prison without the possibility of parole, they will be killed by lethal injection. China, Iran, and Saudi Arabia execute more criminals, but, among affluent democracies, the death penalty, like the U.S. homicide and incarceration rates, marks an American exception, or, looked at another way, an anachronism.

Long ago, Beccaria pointed out the meaningfulness of the correspondence, over time, between crime and punishment, between one kind of violence and another. If the history of murder contains a lesson, Beccaria believed, it was this: “The countries and times most notorious for severity of punishment have always been those in which the bloodiest and most inhumane of deeds were committed.”

Murder has a history, but it isn’t always edifying, and sometimes the history of crime and punishment has a chilling sameness. The prospect of death didn’t deter Barnett Davenport, a Connecticut murderer who was hanged in 1780, at the age of nineteen. “No man becomes a devil in a minute,” Davenport said, in a confession made a week before he mounted the gallows. His life of crime began when, at the age of twelve, he stole some watermelons from a neighbor’s garden. More than once, he was caught. But by the time he was eighteen he had advanced from pilfering eggs and potatoes to stealing horses. He fought in the Revolution and then deserted. He went to live in the house of a man named Caleb Mallery, near Litchfield. On February 3, 1780, “a night big with uncommon horror” (and a year with an elevated homicide rate), Davenport killed Mallery, Mallery’s wife, and their seven-year-old granddaughter, beating their heads in with a pestle and a rifle. Next, he pried open the family’s money chest and took from it a pile of bills and a handful of coins. Then he set the house on fire, leaving inside two more children, ages six and four. He was captured, and swiftly hanged. In his confession, he recalled that Caleb Mallery had cried out, in between blows, “Tell me what you do it for!” History does not record the murderer’s reply. ♦

*Correction, November 4, 2009: “When Brute Force Fails: How to Have Less Crime and Less Punishment” is published by Princeton University Press, not Current Affairs, as originally stated.

*November 10, 2009: Johanna Petit-Chapman is William Petit’s sister, not his sister-in-law, as originally stated.