Several high-profile criminal cases, such as the first trial of Erik and Lyle Menendez, which resulted in a hung jury, have persuaded many people that courts in the United States have become excessively tolerant of excuses. In his 1994 book The Abuse Excuse: Cop-Outs, Sob Stories, and Other Evasions of Responsibility, the Harvard law professor Alan Dershowitz decries the growing list of excuses that defendants have offered to explain their behavior, from drug and alcohol addiction to possession of XYY chromosomes.
Most criminologists and law-school professors, however, would take strong exception to that criticism. To them, American criminal justice is, if anything, excessively punitive. They point out that most people accused of crimes are convicted, that prison populations have risen dramatically in recent years, and that attempts to employ such excuses as the insanity defense rarely succeed.
In my view, both sides are right -- up to a point. Most criminal defendants wind up being convicted, often as a result of plea bargains in which the charges are reduced but clever excuses are rarely accepted. However, a few criminal defendants -- notably, those who are wealthy or otherwise attractive to skilled defense attorneys -- enjoy advantages denied to less-celebrated suspects: expert witnesses (who are often academics), lengthy trials, and elaborate appeals. In these cases, the criminal trial has changed fundamentally, from a process that seeks to judge the accused into one that tries to explain his or her behavior.
As some critics have long claimed, we do have two kinds of criminal trials, one for the rich and another for the poor. But the difference lies not so much, as these critics claim, in the penalties that defendants receive as in the strategies that they employ. The celebrated defendant takes a chance by seeking a jury trial -- jurors can render a harsh verdict as well as acquit. The average defendant negotiates a deal -- almost certain conviction, but a modest penalty.
I believe that some"expert witnesses” have helped widen the gap by agreeing to provide paid testimony in which they inflate their own scientific credentials or claim scientific certainty where none exists. Some of these witnesses are scholars; others are full-time, self-appointed experts. Their credentials range from weak to strong, from advanced degrees to a variety of other kinds of experience. The celebrated defendant -- one who is either wealthy or notorious enough to attract high-powered attorneys -- can take on a lengthy trial in which such expert witnesses play a large role. The law does not give equal access to such witnesses to all criminal defendants (with the possible exception of those facing the death penalty). Generally, only defendants who can afford them enjoy them.
But what do they enjoy? Not, I think, real science, but often merely"pretend” science. Real science involves testing theories by repeated independent experiments that, if they never manage to refute the theory, lead one to accept (provisionally) the theory as stated. Real science has been used to establish, so far as we can tell, that every person has a unique set of fingerprints and a distinctive DNA profile. And real science has been used to show that polygraphs -- lie detectors -- do not always detect lies, and that voice prints do not always identify the speaker. Real science is most likely to be produced (although not always) when scholars study inanimate objects, as do biologists and chemists.
The science that seeks to explain human behavior -- that practiced by psychologists, psychiatrists, and the like -- is much less empirically exact. At best, it consists of little more than an assertion that, in a certain percentage of cases, a person with certain traits (such as a particular score on the Minnesota Multiphasic Personality Inventory) is likely to be part of a group that will be more likely than others to behave in a certain way. A person who receives a particular score on the inventory, for example, might be statistically more likely than the average person to be paranoid. This statistical probability has little legal meaning, however, because it does not say whether the person is, in fact, paranoid.
At worst, behavioral science is often much less than this. Academics and other advocates who care about social problems such as battered women, abused children, and intoxicated men will interview people who have been battered or abused or drunk and then describe the characteristics that those people share. Because these advocates care deeply about mobilizing public opinion to combat battering or abuse or drunkenness, they frequently pay more attention to telling a compelling story than to making a careful effort to show in what ways (and with what frequency) women who have and haven’t been battered, or children who have and haven’t been abused, differ. Earnest scientists who study these problems acknowledge that they can rarely predict how a particular battered woman, abused child, or drunk man will behave in a particular circumstance.
Nevertheless, expert witnesses -- academics and independent practitioners such as psychologists and psychiatrists -- are generally allowed to give courts their opinions on behavioral issues after undergoing only the most cursory effort to establish their professional standing and scientific objectivity. Federal and state rules of evidence differ in this regard, but in many cases the rules say a person can become an expert witness (and thus be entitled to offer opinions, not simply recount observations) because he or she is"anyone” who is"qualified” by"knowledge, skill, experience, training, or education” to be an expert. In short, people with a Ph.D. or some training that approximates that degree are considered to be experts whether or not their views are based on sound and compelling empirical evidence.
Some states, such as California, have used a tougher test, known as the Frye rule, which holds that the substance of expert testimony must meet a test of"general acceptance in the particular field in which it belongs.” That is a useful restriction, but even states that claim to employ it frequently have made exceptions. In 1989, the Supreme Court of California allowed a psychologist to testify that a woman who had been charged with sexual perversion against children had a “normal personality” and had been"falsely charged” -- solely on the basis of two written tests that he had given her.
In 1993, in Daubert v. Merrill Dow Pharmaceuticals Inc., the U.S. Supreme Court tried to set tougher standards by urging other courts to use more-scientific measures in judging would-be expert testimony: For example, has an expert’s scientific assertion withstood serious efforts to falsify it by experiments? Was it published in a reputable, peer-reviewed journal? But just how this new test should be applied was left to lawyers to argue over, as they subject expert testimony to vigorous cross-examination.
Criminal trials that produce legal appeals have increased sharply in recent decades. Anyone with a grasp of how organizations behave should be able to guess what this means. Many judges, eager to avoid successful appeals, allow defendants to admit as much evidence as possible, including that presented by expert witnesses. Today, courts in many states have been instructed either by appeals courts or by legislators to allow women who kill their husbands or lovers to use"battered women’s syndrome” as a mitigation or excuse.
A large and tragic number of women have been battered, and helping them is a matter of great importance. But converting the fact of being battered into a"syndrome” that has known explanatory power is suspect. If an abused woman kills a man who is attacking her, she can properly claim self-defense. But the situation is much less clear if she kills her sleeping husband and then claims that, according to experts, she suffers from a syndrome that has dictated her action. This syndrome has, in fact, no explanatory power, in part because no one has explained who is and who is not affected by it, or why it adds anything to our conventional theories of self-defense.
Matters are much the same with child abuse and intoxication. There are, alas, many abused children; there are even more people who drink too much. But why should these facts, taken alone, provide a way of mitigating a penalty when the many other causes of crime are not considered mitigating circumstances? The likelihood of people killing or raping is affected by countless factors, yet courts have accepted some excuses and not others, making little attempt to justify their inconsistency.
The effect of excuses on juries can be quite profound and worrisome. Many experiments have tested how jurors in mock trials react to different versions of a defense trial. The versions differ in how much information jurors are given about the legally irrelevant -- but socially important -- features of a defendant. Suppose a woman kills her husband. What is legally relevant is whether she acted in self-defense, under duress, or was in the grip of insanity so great that she was hallucinating. What is not relevant is whether she was a good mother or a bad one. But when mock-trial jurors were told about her status as a mother, the penalties they supported changed dramatically, with lesser punishment given to a good mother.
This helps explain an otherwise puzzling fact. Americans, by and large, are deeply upset about crime and are quick to criticize judges and jurors who acquit apparently guilty defendants or give them a light sentence. But take 12 people more or less randomly from our society, put them on a jury, and expose them to lengthy accounts of why the defendant acted as he or she did, and many of them will vote for the lightest penalty.
American criminal courts have not accepted a long list of excuses based on abuse. What they have accepted is a long list of claims about the plight of defendants who are rich or famous enough to enlist supposedly expert witnesses. The task for our legislatures is partly to purge false excuses from the criminal code and partly to place meaningful restrictions on who may be summoned as an expert and what they may testify about. The key is to reduce the amount of social-science speculation that may be introduced as expert testimony in a trial. I would like to keep almost all social scientists -- myself included -- out of the courtroom.
James Q. Wilson is a professor of management and public policy at the University of California at Los Angeles and the author, most recently, of Moral Judgment, published last month by BasicBooks.