The term "search for truth" was repeatedly invoked by both sides of the Simpson case. A review of the trial transcript reveals that this phrase was used more than seventy times. The prosecutors claimed that they were searching for truth and that the defense was deliberately obscuring it. Where it was in their interest to have the jury hear evidence that would hurt Simpson -- such as the details of arguments between him and his former wife -- the prosecutors argued that the search for truth required the inclusion of such evidence, despite its marginal relevance. On other occasions, they argued that the search for truth required the exclusion of evidence that demonstrated that one of their key witnesses, Los Angeles Police Detective Mark Fuhrman, had not told the truth at the trial. The defense also claimed the mantle of truth and accused the prosecution of placing barriers in its path. And throughout the trial, the pundits observed that neither side was really interested in truth, only in winning. They were right -- and wrong.

In observing this controversy, I was reminded of the story of the old rabbi who, after listening to a husband complaining bitterly about his wife, replied, "You are right, my son." Then, after listening to a litany of similar complaints from the wife, he responded, "You are right, my daughter." The rabbi's young student then remarked, "But they can't both be right" -- to which the rabbi replied, "You are right, my son." So too, in the context of a criminal case, the prosecution is right when it says it is searching for truth -- a certain kind of truth. The defense is also searching for a certain kind of truth. Yet both are often seeking to obscure the truth for which their opponent is searching. In arguing to exclude evidence that Fuhrman had perjured himself when he denied using the "N" word, Marcia Clark said just that:

This is a search for the truth, but it's a search for the truth of who committed these murders, your Honor. Not who Mark Fuhrman is. That truth will be sought out in another forum. We have to search for this truth now, and I beg the court to keep us on track and to allow the jury to pursue that search for the truth based on evidence that is properly admissible in this case and relevant to that determination. ¹

The truth is that most criminal defendants are, in fact, guilty. Prosecutors, therefore, generally have the ultimate truth on their side. But since prosecution witnesses often lie about some facts, defense attorneys frequently have intermediate truth on their side. Not surprisingly, both sides emphasize the kind of truth that they have more of. To understand this multilayered process, and the complex role "truth" plays in it, it is important to know the difference between a criminal trial and other more single-minded searches for truth.

What is a criminal trial? And how does it differ from a historical or scientific inquiry? These are among the questions posed in a university-wide course I teach at Harvard, along with Professors Robert Nozick, a philosopher, and Stephen J. Gould, a paleontologist. The course, entitled "Thinking About Thinking," explores how differently scientists, philosophers, historians, lawyers, and theologians think about and search for truth. The goal of the historian and scientist, at least in theory, is the uncovering or discovery of truth. The historian seeks to determine what actually happened in the recent or distant past by interviewing witnesses, examining documents, and piecing together fragmentary records. The paleontologist searches for even more distant truths by analyzing fossils, geological shifts, dust and DNA. Since what's past is prologue, for both the historian and the scientist, efforts are often made to extrapolate from what did occur to what will occur, and generalizations -- historical or scientific rules -- are proposed and tested.

Although there are ethical limits on historical and scientific inquiry, the ultimate test of a given result in these disciplines is its truth or falsity. Consider the following hypothetical situation. An evil scientist (or historian) beats or bribes some important truth out of a vulnerable source. That truth is then independently tested and confirmed. The evil scientist might be denied his Nobel Prize for ethical reasons, but the truth he discovered is no less the truth because of the improper means he employed to arrive at it. Scientists condemn "scientific fraud" precisely because it risks producing falsity rather than truth. But if a fraudulent experiment happened to produce a truth that could be replicated in a non-fraudulent experiment, that truth would ultimately become accepted.

Put another way, there are no "exclusionary rules" in history or science, as there are in law. Historical and scientific inquiry is supposed to be neutral as to truth that is uncovered. Historians should not favor a truth that is "politically," "patriotically," "sexually," or "religiously" correct. In practice, of course, some historians and scientists may very well skew their research to avoid certain truths -- as Trofim Lysenko did in the interests of Stalinism, or as certain racial theorists did in the interests of Hitlerism. But in doing so, they would be acting as policy-makers rather than as historians or scientists.

The discovery of historical and scientific truths is not entrusted to a jury of laypeople selected randomly from the population on the basis of their ignorance of the underlying facts. The task of discovering such truths is entrusted largely to trained experts who have studied the subject for years and are intimately familiar with the relevant facts and theories.

Historical and scientific inquiries do not require that fact-finders necessarily be representative of the general population, in race, gender, religion, or anything else -- as jurors must be. To be sure, a discipline that discriminates runs the risk of producing falsehood, since truth is not the domain of any particular group. But again, historical and scientific truths may be just as valid if arrived at by segregationists as if by integrationists. In history and science, truth achieved by unfair means is preferred to falsity achieved by fair means.

Nor are historical and scientific truths determined on the basis of adversarial contests in which advocates -- with varying skills, resources, and styles -- argue for different results. Although the quest for peer approval -- tenure, prizes, book contracts, and so on -- may become competitive, the historical or scientific method is not premised on the view that the search for truth is best conducted through adversarial conflict. ²

Finally, all "truths" discovered by science or history are always subject to reconsideration based on new evidence. There are no prohibitions against "double jeopardy." Nor is there any deference to considerations of "finality"; nor are there statutes of limitations. In sum, the historical and scientific inquiry is basically a search for objective truth. Perhaps it is not always an untrammeled search for truth. Perhaps the ends of truth do not justify all ignoble means. But the goal is clear: objective truths as validated by accepted, verifiable, and, if possible, replicable historical and scientific tests.

The criminal trial is quite different in several important respects. Truth, although one important goal of the criminal trial, is not its only goal. If it were, judges would not instruct jurors to acquit a defendant whom they believe "probably" did it, as they are supposed to do in criminal cases. The requirement is that guilt must be proved "beyond a reasonable doubt." But that is inconsistent with the quest for objective truth, because it explicitly prefers one kind of truth to another. The preferred truth is that the defendant did not do it, and we demand that the jurors err on the side of that truth, even in cases where it is probable that he did do it. Justice John Harlan said in the 1970 Supreme Court Winship decision that, "I view the requirement of proof beyond a reasonable doubt in a criminal case as bottomed on a fundamental value that it is far worse to convict an innocent man than to let a guilty man go free." ³ As one early-nineteenth-century scholar explained, "The maxim of the law … is that it is better that ninety-nine … offenders shall escape than that one innocent man be condemned."⁴ More typically, the ratio is put at ten to one.

In a criminal trial, we are generally dealing with a decision that must be made under conditions of uncertainty. We will never know with absolute certainty whether Sacco and Vanzetti killed the paymaster and guard at the shoe factory, whether Bruno Hauptmann kidnapped and murdered the Lindbergh baby, or whether Jeffrey MacDonald bludgeoned his wife and children to death. In each of these controversial cases, the legal system was certain enough to convict -- and in two of them, to execute. But doubts persist, even decades later.

Those who believe that O.J. Simpson did murder Nicole Brown and Ronald Goldman must acknowledge that they cannot know that "truth" with absolute certainty. They were not there when the crimes occurred or when the evidence was collected and tested. They must rely on the work and word of people they do not know. The jurors in the Simpson case were not asked to vote on whether they believed "he did it." They were asked whether the prosecution's evidence proved beyond a reasonable doubt that he did it. Juror number three, a sixty-one-year-old white woman named Anise Aschenbach, indicated that she believed that Simpson was probably guilty "but the law wouldn't allow a guilty verdict."⁵ Had the Simpson trial been purely a search for truth, this juror would have been instructed to vote for conviction, since in her view that was more likely the "truth" than that he didn't do it. But she was instructed to arrive at a "false" verdict, namely that although in her view he probably committed the crimes, yet as a matter of law he did not.

This anomaly has led some reformers to propose the adoption of the old Scottish verdict "not proven" instead of the Anglo-American verdict of "not guilty." Even the words "not guilty" do not quite convey the sense of "innocent," although acquitted defendants are always quick to claim that they have been found "innocent." Some commentators have suggested that alternative verdicts -- "guilty," "innocent," and "not proven" -- be available so that when jurors believe that the defendant did not do it, they can reward him with an affirmative declaration of innocence rather than merely a negative conclusion that his guilt has not been satisfactorily proved.

At one level, we understand -- and most agree with -- the requirement of proving criminal guilt by a more demanding standard than that required for other decisions in which the risk of error is equivalent on both sides. Yet at another level, we rebel at the notion that a different "truth" may be found in different kinds of proceedings. Imagine the public reaction, for example, if Simpson were to be found liable by a jury in the civil case now pending against him by the heirs of the murder victims for the very same acts of which he was acquitted by the jury at his criminal trial. FRONTLINE editor's note: Simpson was, in fact, found liable for the deaths of Nicole Brown Simpson and Ronald Goldman in the civil trial. Would that mean "he did it" for purposes of the civil suit, but "he didn't do it" for purposes of the criminal prosecution? Most Americans would surely believe that it only went to prove that "the law is a ass," as Mr. Bumble put it in Dickens's Oliver Twist. But such a result, were it to occur, would rather show that the law is a relatively subtle instrument capable of making refined distinctions between the standards of proof required to deprive a person of his liberty, on the one hand, and to deprive him of money, on the other. As Justice Harlan further commented in his Winship opinion:

If, for example, the standard of proof for a criminal trial were a preponderance of the evidence rather than proof beyond a reasonable doubt, there would be a smaller risk of factual errors that result in freeing guilty persons, but a far greater risk of factual errors that result in convicting the innocent. ⁶

The burden of proof in a criminal case is "beyond a reasonable doubt," while the burden of proof in a civil case is "by a mere preponderance of the evidence." Simply put, this means that it takes more and better proof to convict a criminal defendant of a crime than to hold a civil defendant liable for monetary damages. How much more and how much better are not subject to precise quantification. We know what proof by a preponderance is supposed to mean: Even in a close case, the side that is more persuasive wins. In civil cases, truth is supposed to prevail, without the law's thumb on either side of the scales of justice.

We don't, however, have a very good idea of what "proof beyond a reasonable doubt" means in criminal cases. We know that the law's thumb is on the side of the defendant in a criminal case, but the courts are reluctant to tell us how heavy that thumb is supposed to be. My law students debate the following hypothetical case: A fatal accident is caused by a blue bus in a town where 90 percent of the blue buses are owned by the A Company and 10 percent by the B Company. That is all the evidence presented at the criminal trial of the A Company. Is this 90 percent "likelihood" that an A bus caused the accident enough to prove that fact "beyond a reasonable doubt"? Are we willing to convict a company -- or an individual -- in the face of a 10 percent likelihood of innocence?

Students often respond in the negative to this question, arguing that a clear statistical likelihood of innocence in the range of 10 percent is too high for a criminal conviction. I then ask these students if it would be enough for conviction without the statistics -- if an eyewitness were to testify for the prosecution that he was "sure" it was an A Company bus because he saw the A Company's logo. Most of these same students then say yes, it would be enough, because there was an eyewitness and eyewitnesses can be certain, whereas statistics are always probabilistic. I then ask them if their minds would change again if the defense introduced an acknowledged expert who testified that the kind of eyewitness testimony introduced by the prosecution is accurate in only 85 percent of cases. They still say yes, thus apparently preferring to convict with an 85 percent likelihood of truth rather than convict with a 90 percent likelihood. The debate goes on with numerous variations on these themes. And these are the very sort of difficult questions that the courts rarely address, because they do not have entirely satisfactory answers.

We do understand the reasons for permitting a lower standard of proof in civil than in criminal cases. In civil cases, the risk of error on each side is equal and we do not prefer one type of error over another. But in criminal cases, we prefer the type of error under which a possibly guilty defendant would go free to the type of error under which an innocent defendant would go to prison or be executed.

Nor are different standards of proof limited to legal cases. We apply varying standards in our daily lives. Consider, for example, a woman applying for a baby-sitting job who shows you a certificate proving that she was unanimously acquitted of child molesting. You would never hire her, because you are unwilling to take any chances with your child's baby-sitter.

In addition to the requirement of proof beyond a reasonable doubt in criminal cases, there are numerous other barriers to absolute truth that have been deliberately built in to the criminal process to serve other functions. Some of these barriers can be justified as perhaps contributing to the search for truth in the long run, while probably sacrificing truth in a particular case. The exclusion from evidence of a coerced confession may produce falsity in a case where the confession, although coerced, is nonetheless true and can be independently corroborated. Consider, for example, a case in which a defendant is coerced into admitting not only that he killed the missing victim but also where he buried the body. By excluding this true, coerced confession -- and thus possibly freeing this guilty murderer -- the law may be seeking to increase the long-term, truth-finding goals of the system, since many other coerced confessions may turn out to be false.

We could, of course, satisfy both long- and short-term truth goals by adopting a rule -- which we have not adopted in this country -- under which coerced confessions that can be independently corroborated will be admitted into evidence. Under such a rule, a coerced confession which produces merely a statement that "I did it" would not be admitted, but a coerced confession that leads to the victim's body covered with the defendant's fingerprints and DNA would be admitted. Or, perhaps as a fail-safe, only the body and the physical evidence would be admitted, but not the coerced confession itself, even though it has been proved to be true.

Instead, we have opted for an exclusionary rule under which the coerced confession and all its fruits are excluded, even if the fruits prove the truthfulness of the confession.⁷ Such a broad exclusionary rule is not designed to serve only the goals of truth, either long- or short-term. It is also intended to serve an important set of values entirely unrelated to truth. Those include privacy (or, in the eighteenth-century language of the Fourth Amendment, "security"), freedom from unreasonable governmental intrusion, and the integrity of the mind and body. These values were regarded by those who introduced the "exclusionary rule" as being more important, at least on occasion, than truth. The exclusionary rule explicitly recognizes that the guilty will sometimes have to be freed in order to send a message to police and prosecutors that the noble end of seeking the truth does not justify ignoble means such as unreasonable searches or coerced confessions.

If the only goal of the adversary system were to find "the truth" in every case, then it would be relatively simple to achieve. Suspects could be tortured, their families threatened, homes randomly searched, and lie detector tests routinely administered. Indeed, in order to facilitate this search for truth, we could all be subjected to a regimen of random blood and urine tests, and every public building and workplace could be outfitted with surveillance cameras. If these methods -- common in totalitarian countries -- are objected to on the ground that torture and threats sometimes produce false accusations, that objection could be overcome by requiring that all confessions induced by torture or threats must be independently corroborated. We would still never tolerate such a single-minded search for truth, nor would our constitution, because we believe that the ends -- even an end as noble as truth -- do not justify every possible means. Our system of justice thus reflects a balance among often inconsistent goals, which include truth, privacy, fairness, finality, and equality.

Even "truth" is a far more complex goal than may appear at first blush. There are different kinds of truth at work in our adversary system. At the most basic level, there is the ultimate truth involved in the particular case: "Did he do it?" Then there is the truth produced by cases over time, which may be in sharp conflict. For example, the lawyer-client privilege -- which shields certain confidential communications from being disclosed -- may generate more truth over the long run by encouraging clients to be candid with their lawyers. But in any given case, this same privilege may thwart the ultimate truth -- as in the rare case where a defendant confides in his lawyer that he did it. The same is true of other privileges, ranging from the privilege against self-incrimination to rape shield laws, which prevent an accused rapist from introducing the prior sexual history of his accuser.

Even in an individual case, there are different types -- or layers -- of truth. The defendant may have done it -- ultimate truth! -- but the police may have lied in securing the search warrant. Or the police may even have planted evidence against guilty defendants, as New York state troopers were recently convicted of doing, and as some jurors believed the police did in the Simpson case.

The Anglo-American criminal trial employs the adversary system to resolve disputes. This system, under which each side tries to win by all legal and ethical means, may be conducive to truth in the long run, but it does not always produce truth in a given case. Nor is it widely understood or accepted by the public.

One night, during the middle of the Simpson trial, my wife and I were attending a concert at Boston Symphony Hall. When it was over a woman ran down the center aisle. We thought she was headed toward the stage to get a close look at Midori, who was taking bows. But the woman stopped at our row and started shouting at me: "You don't deserve to listen to music. You don't care about justice. All you care about is winning." I responded, "You're half right. When I am representing a criminal defendant, I do care about winning -- by all fair, lawful, and ethical means. That's how we try to achieve justice in this country -- by each side seeking to win. It's called the adversary system."

I did not try to persuade my critic, since I have had little success persuading even my closest friends of the morality of the Vince Lombardi dictum as it applies to the role of defense counsel in criminal cases: "Winning isn't everything. It's the only thing."

There are several reasons why it is so difficult to explain this attitude to the public. First, hardly anybody ever admits publicly that winning is their goal. Even the most zealous defense lawyers proclaim they are involved in a search for truth. Such posturing is part of the quest for victory, since lawyers who candidly admit they are interested in the truth are more likely to win than lawyers who say they are out to win. Second, although defense attorneys are supposed to want to win -- regardless of what they say in public -- prosecutors are, at least in theory, supposed to want justice. Indeed, the motto of the U.S. Justice Department is "The Government wins when Justice is done."⁸ That is the theory. In practice, however, each side wants to win as badly as the other. Does anyone really doubt that Marcia Clark wanted to win as much as Johnnie Cochran did? She told the jury during her closing argument that she had stopped being a defense attorney and became a prosecutor so that she could have the luxury of looking at herself in the mirror every morning and knowing that she always told juries the truth, and that she would only ask for a conviction where she could prove that the defendant was, in fact, guilty.⁹ But notwithstanding these assertions, Clark and other prosecutors put Mark Fuhrman on the stand after having been informed that he was a racist, a liar, and a person capable of planting evidence even before they called him as a trial witness. An assistant district attorney, among others, warned the Simpson prosecutors about Fuhrman. The prosecutors also saw his psychological reports, in which he admitted his racist attitudes and actions. The only thing they didn't know was that Fuhrman -- and they -- would be caught by the tape-recorded interviews that Fuhrman gave an aspiring screenwriter, Laura Hart McKinny. If the tapes had not surfaced, the prosecutors would have attempted to destroy the credibility of the truthful good Samaritan witnesses who came forward to testify about Fuhrman's racism. Only the tapes stopped them from doing that.

Clark behaved similarly with regard to Detective Philip Vannatter. Any reasonable prosecutor should have been suspicious of Vannatter's testimony that when he went to the Simpson estate in the hours following the discovery of the double murder, he no more suspected Simpson of the killings than he did Robert Shapiro. That testimony had all the indicia of a cover story, and yet Clark allowed it to stand uncorrected.

In practice, the adversary system leads both sides to do everything in their power -- as long as it is lawful and ethical -- to win. Since most defendants are guilty, it follows that the defense will more often be in the position of advocating ultimate falsity than will the prosecution. But since the prosecution always puts on a case -- often relying on police testimony -- whereas the defense rarely puts on any affirmative case, it follows that the prosecution will more often be in the position of using false testimony in an effort to produce its ultimately true result.

Outrage at Simpson's acquittal is understandable in those who firmly believe that he did it. No one wants to see a guilty murderer go free, or an innocent defendant go to prison. But our system is judged not only by the accuracy of its results, but also by the fairness of the process. Indeed, the Supreme Court has said that our system must tolerate the occasional conviction, imprisonment, and even execution of a possibly innocent defendant because of considerations of finality, federalism, and deference to the jury. The United States Supreme Court recently recognized that "our judicial system, like the human beings who administer it, is fallible" and that innocent defendants have at times been wrongfully convicted. The Court concluded that some wrongful convictions and even executions of innocent defendants must be tolerated "because of the very disruptive effect that entertaining claims of actual innocence would have on the need for finality in capital cases, and the enormous burden that having to retry cases based on often stale evidence would place on the States." ¹⁰

While reasonable people may, and do, disagree with that conclusion, it surely must follow from our willingness to tolerate some innocents being wrongly executed by our less than perfect system that we must be prepared to tolerate the occasional freeing of defendants who are perceived to be guilty. This is a Rubicon we, as a society, crossed long before the Simpson verdict -- although one might not know it from the ferocity of the reaction to that verdict. As I mentioned earlier, the exclusionary rule is based on our willingness to free some guilty defendants in order to serve values often unrelated to truth. It is interesting to contrast the public reaction to the jury's acquittal with what would have happened if Simpson had gone free as a result of the judge's application of the exclusionary rule.

What would the public reaction have been if the trial judge had ruled that the original search of Simpson's estate had been unconstitutional and all its fruits had to be suppressed? Such a ruling might have wounded the prosecution's case -- although perhaps not mortally. It would have excluded from evidence the bloody glove found behind Simpson's house, the socks found in his bedroom, the blood found in the driveway. It might also have tainted the warrants, which were based, at least in part, on the evidence observed during the initial search. These warrants produced a considerable amount of evidence which might also have had to be suppressed. Indeed, had the search of Simpson's estate been declared unconstitutional, virtually everything found in and around the estate might have been subject to exclusion.

That would still have left the other half of the prosecution's case-everything found at the crime scene-since no probable cause or warrant was required for searches and seizures at Nicole Brown's condominium. But the quantity of the prosecution's evidence against Simpson would have been considerably reduced if the evidence seized at the Simpson estate had been suppressed as the fruits of an unconstitutional search.

Had the trial judge suppressed all the Simpson estate evidence, there would have been a massive public outcry against the judge, the exclusionary rule, the Constitution, and the system. This outcry would have increased in intensity if this suppression had led -- either directly or indirectly -- to the acquittal of the defendant. "Guilty Murderer Is Freed Because of Legal Technicality," the headlines would have shouted. Conservatives would have demanded abolition of the exclusionary rule. But many liberals and civil libertarians who today rail against the jury verdict in the Simpson case would have defended the decision as the price we pay for preserving our constitutional rights.

This is all, of course, in the realm of the hypothetical, since it is unlikely that any judge -- certainly any elected judge with higher aspirations -- would have had the courage to find the search unconstitutional and thus endanger the prosecution's case. Recently, I had lunch with a former student who was seeking to be appointed to the California Superior Court. I asked her how she would answer the following question if it were put to her by the judicial nominating committee: "Would you have ruled the search unconstitutional if you believed the police were lying about why they went to Simpson's house, climbed the gate, and entered?" Without a moment's hesitation she responded: "No way. No judge would -- are you kidding?"

I think my former student overstated the case in saying that no judge would have had the guts to find the police were lying in the Simpson case, but I believe that most judges would do what the two trial judges almost certainly did here: assume a variation of the position of the three monkeys, hearing no lies and seeing no lies. And judges speak the lie of pretending to believe witnesses who they must know are not telling the truth. What does it say about our system of justice that so many judges would pretend to believe policemen they know are lying, rather than follow the unpopular law excluding evidence obtained in violation of the Constitution? I am not alone in believing that the judges in the Simpson case could not really have believed what they said they believed. As Scott Turow argued in a perceptive op-ed piece the day after the verdict:

The detectives' explanation as to why they were at the house is hard to believe. … Four police detectives were not needed to carry a message about Nicole Simpson's death. These officers undoubtedly knew what Justice Department statistics indicate: that half of the women murdered in the United States are killed by their husbands or boyfriends. Simple probabilities made Mr. Simpson a suspect. … Also, Mark Fuhrman had been called to the Simpson residence years earlier when Mr. Simpson was abusing his wife. …

The fact that the district attorney's office put these officers on the witness stand to tell this story and that the judge accepted it is scandalous. It is also routine. …

Turow then went on to blame the prosecutor and the judges:

To lambaste only Detectives Fuhrman and Vannatter misses the point. … It was the Los Angeles District Attorney's Office that put them on the stand. It was Judge Kennedy-Powell [the judge who presided at the preliminary hearing] who took their testimony at face value rather than stir controversy by suppressing the most damning evidence in the case of the century. And it was Judge Lance Ito who refused to reverse her decision. …¹¹

Neither the prosecutors nor the judges were searching very hard for the truth of why the detectives went to the Simpson residence. They apparently thought that the disclosure of that truth would make the proving of what they believed was a more important truth -- that the defendant was guilty -- more difficult. Thus, some people believe that the search for one truth in a criminal case can be served by tolerating other half-truths and even lies. But I believe the prosecution's decision to call Detectives Vannatter and Fuhrman to the witness stand may have been the final nail in a coffin that had been built even earlier by the police. That costly decision was thoughtlessly made by prosecutors who have become so accustomed to police perjury about searches and seizures that they did not even pause to consider its possible impact on this jury.