It’s the wrong time and the wrong place. Though your face is charming, it’s the wrong face.
It’s not his face, but such a charming face That it’s all right with me.
—Cole Porter
There had been an automobile accident. Before the court one of the witnesses, who had sworn to tell “the whole truth, and nothing but the truth,” declared that the entire road was dry and dusty; the other swore that it had rained and the road was muddy. The one said that the automobile was running very slowly; the other, that he had never seen an automobile rushing more rapidly. The first swore that there were only two or three people on the village road; the other, that a large number of men, women, and children were passing by. Both witnesses were highly respectable gentlemen, neither of whom had the slightest interest in changing the facts as he remembered them.
—Hugo Münsterberg,
On the Witness Stand: Essays on Psychology and Crime, 1906
WHEN A VICTIM or a witness to a crime points a finger at a suspect and cries, “He’s the one! I’m sure!”, what are the odds that the accuser is right?
Eyewitness testimony is notoriously unreliable. Prosecutors know it, and so do defense attorneys. Then why is there so much of it? Because juries do not know, and the sight of the victim sitting in the witness box and pointing that quavering finger at the defendant is a powerful psychological boost to the prosecution’s case. Even if you know that eyewitness testimony is often unreliable, it’s hard to believe that this slender girl who has been horribly molested, or this elderly, frail man on crutches with his head still swathed in bandages, could be lying.
And, of course, they’re not lying. Not consciously, anyway. Consider the case of “Lord Willoughby” in Chapter 3. The ladies who appeared in court against Adolf Beck honestly believed that they had identified the right man.
In his 1908 book On the Witness Stand, Dr. Hugo Münsterberg discusses perception and memory as they pertain to eyewitness testimony. In one of his chapters, “The Memory of the Witness,” he recounts his own experience in giving testimony in a burglary case. He appeared in court not as an expert witness but as a victim. While he and his family were away at the seashore, his Boston house had been burglarized.
Münsterberg testified that the burglars had entered through a cellar window and moved about upstairs from room to room. He told the court that he had found drops of candle wax on the second floor, indicating that they had been there at night. The burglars had left a large mantel clock wrapped in paper on the dining room table, showing that they had intended to return. They had also taken some articles of clothing, and Münsterberg gave the police a list of the missing items.
Within a few days after he appeared on the stand, Münsterberg realized that his testimony had been wrong in every detail. The burglars had not come in through a window, they had broken the lock on the cellar door. The clock was on the dining room table, but not wrapped in paper—it was covered with a tablecloth. The candle had dripped not on the second floor but in the attic. And the list of missing clothing was incomplete by seven garments.
And this from a psychologist, a criminalist, and a trained observer with an excellent memory.
In his book, Münsterberg goes on to analyze why he made these mistakes: he was in a high state of excitement (he had rushed home when the police told him that his house had been broken into); he had focused too much on a specific observation (the clock, the wax) without really noticing the secondary details (what the clock was actually wrapped in, and on which floor the wax was actually spilled). When it came time to tell the story, his mind had then supplied the supporting details. As for the cellar window, the police had mentioned that burglars usually enter through the cellar window; so Münsterberg assumed that this is what he had seen. It was not until a few days later that he noted that the lock on the cellar door had been forced.
The Survival, Evasion, Resistance, and Escape programs of the United States Army conduct experiments to find out how best to prepare soldiers in the event they become prisoners of war. One of these experiments, devised by Yale psychiatrist C. Andrew Morgan, has uncovered a fascinating statistic. In the experiment, a test subject undergoes a tough interrogation and is put into an isolation cell. Later, another captor bursts into the cell. He waves a photograph and asks, “Did your interrogator give you anything to eat?”
That’s it. That’s the extent of the misdirection.
As the experiment comes to an end, the subject is asked to pick out his interrogator’s face from among nine photographs he is shown. An amazing 85 percent of the subjects choose the wrong face. Instead of the face of their actual interrogator, they pick the one they had seen briefly in the waved photograph.
Eighty-five percent.
Professor Elizabeth Loftus holds appointments at the University of California, Irvine, in the departments of Psychology and Social Behavior and Criminology, and in Law and Society and Cognitive Sciences as well as at the Center for the Neurobiology of Learning and Memory. Loftus has been called upon to consult or testify on questions of memory in hundreds of cases, including the McMartin Preschool molestation case, the Hillside Strangler case, and the Oklahoma City bombing case. She is an expert—the expert, in fact—on memory and what can be done to distort it. Her main finding, after three decades of research, is that it does not take much. In a 2002 article in Skeptical Inquirer, she wrote, “A memory is, of course, not proof of the event it purports to recall. We all ‘remember’ things that never actually happened, as ample scientific evidence has demonstrated.”
Eyewitness testimony, Loftus finds, not only is unreliable but can be influenced by the way questions regarding the incident are phrased. In one study she showed her subjects films of traffic accidents and then asked questions. If she asked how fast the cars were going when they smashed into each other, the speeds reported by her subjects were higher than if she merely asked how fast they were going when they hit each other. Smashed into also resulted in the subjects’ recalling broken glass when in fact there had been none.
The ways of inducing misinformation into memory are myriad. Subjects have been convinced that a stop sign at an intersection was a yield sign, that a blue car was white, and even that Mickey Mouse was Minnie Mouse.
In addition to its innate unreliability, one of the seldom-stated problems with eyewitness testimony is the habit of the police and prosecuting attorneys of cherry-picking their eyewitnesses. If there are three eyewitnesses to a crime and one says the perpetrator had brown hair, one says blond, and the third says he was bald, and the defendant on trial has brown hair, which of the three witnesses do you suppose will testify for the prosecution? In theory the prosecution is supposed to turn over all evidence, both inculpatory and exculpatory, to the defense. But all too often in practice a great deal of information the defense might find useful is discarded along the way. In 1924 J. Collyer Adam, in his introduction to Hans Gross’s book Criminal Investigation, addressed the difficulty of eyewitness testimony.
It must be admitted that at the present day the value of the deposition of even a truthful witness is much overrated. The numberless errors in perception derived from the senses, the faults of memory, the far-reaching differences in human beings as regards age, sex, nature, culture, mood of the moment, health, passionate excitement, environment—all these things have so great an effect that we scarcely ever receive two quite similar accounts of one thing; and between what people really experience and what they confidently assert, we find only error heaped upon error.
Out of the mouths of two witnesses we may arrive at the real truth, we may form for ourselves an idea of the circumstances of an occurrence and satisfy ourselves concerning it, but the evidence will seldom be true and material; and whoever goes more closely into the matter will not silence his conscience, even after listening to ten witnesses. Evil design and artful deception, mistakes and errors, most of all the closing of the eyes and the belief that what is stated in evidence has really been seen, are characteristics of so very many witnesses that absolutely unbiased testimony can hardly be imagined.
At the end of the nineteenth century, Hans Gross wrote that “A large part of a criminalist’s work is nothing more than a battle against lies. He has to discover the truth and must fight the lie. He meets the lie at every step. Utterly to vanquish the lie, particularly in our work, is of course, impossible, and to describe its nature exhaustively is to write the natural history of mankind.”
The trick, of course, is knowing how to recognize a lie when you hear it.
Involuntary changes in blood pressure are a measure of, among other things, psychological stress. In 1854 the German physiologist Karl von Vierordt (1818–1884) developed a device he called a sphygmograph, which measured instantaneous blood pressure. In 1863 Etienne-Jules Marey (1830–1904) devised a portable version of the sphygmograph, which could trace its readings onto a sheet of paper.
These were further improved upon and then used for the detection of psychological stress by the Italian criminalist Cesare Lombroso. His apparatus consisted of a glove that would monitor the pulse and changes in blood pressure of the person wearing it. The glove was in turn connected to a recording pen that made marks on a revolving drum. In the 1890s Lombroso conducted experiments in lie detection. He came to his task with a heavy load of preconceptions, however—he believed that criminals possessed inherited criminal traits and could be recognized by their appearance. He documented his results, claiming success in using his lie-detection device against real suspects in real criminal cases, in the 1895 edition of his book L’Homme Criminel (The Criminal Man).
The next step in lie detection was taken in 1914 by Vittorio Benussi, who published the results of research on the differences in breathing rates between the “honest and dishonest states of being.” He divided the time it took to take a breath by the time it took to exhale. He discovered that the ratio was greater before telling the truth and decreased afterward, but was smaller before telling a lie and increased afterward.
In 1915 the psychologist William Moulton Marston (1893–1947) began researching the effects of lying on systolic blood pressure (the higher number on the blood pressure fraction). In 1917 he wrote a report of his experiments for the Journal of Experimental Psychology. By the early 1920s, with the aid of his wife and fellow psychologist Elizabeth Holloway Marston, he believed he had devised a useful mechanism for detecting lies. A Harvard graduate with both a law degree and a Ph.D. in psychology, Marston was a man of many interests. He wrote essays on popular psychology, lived with two women—his wife and his mistress Olive Byrne—ardently supported women’s rights, and created the comic book character Wonder Woman, believed to be an amalgam of the two women in his life. Wonder Woman’s “lasso of truth” was certainly inspired by the Marston lie detector.
In 1938 Marston wrote a book called The Lie Detector Test to support his place in the development of what eventually became the polygraph machine. His publisher sent a copy to J. Edgar Hoover at the FBI, where it was reviewed by Quinn Tamm of the Technical Laboratory. It was not regarded with favor. The reviewer found Marston egotistical and self-aggrandizing. Tamm was particularly displeased with the chapter called “Love and the Lie Detector,” in which Marston claims to have settled marital difficulties through the use of the instrument. Tamm commented:
It is noted that throughout the book the author points out that the blood pressure test for the detection of deception in the hands of a trained operator is infallible and that once the deception has been detected it has been his experience that if this is pointed out to the subject he will admit his guilt and it will have the psychological effect on him of making him always in the future tell the truth. This . . . exemplifies the same egotistical ridiculous strain in which this book is written.
Marston’s systolic pressure recorder became only the first step toward the device that we think of today as the lie detector.
In 1921 Dr. John A. Larson, working under the direction of Berkeley, California, police chief and famed criminalist August Vollmer, cobbled together a device he called a “cardiopneumo psychograph.” It could measure blood pressure, pulse, and breathing rates, but it was large, bulky, and took half an hour to set up. Pens on the device scratched lines on paper that was “smoked” with carbon black and then passed over large recording drums. If a permanent record of the session was required, the paper was shellacked to prevent it from smudging, and it was then stored in cans.
In 1923 a young college student named Leonard Keeler showed Chief Vollmer his plans for an improved lie-detecting device. It was more compact, easier to use, and added galvanic skin response to the measurements it took. “If you build it,” Vollmer told him, “we’ll try it.” Keeler and a couple of his friends built it and dubbed it the “emotograph.” It looked, said Vollmer, “like a crazy conglomeration of wires, tubes, and old tomato cans.”
According to Ken Alder in his book The Lie Detectors, the first success of the emotograph in a murder case involved a bit of highly unorthodox role-playing.
In the early hours of the morning of December 16, 1923, in Los Olivos, California, a small town some forty miles north of Santa Barbara, the shack of itinerant blacksmith John J. McGuire was blown to smithereens. McGuire was asleep inside at the time. The logical suspect, the town’s barber and postmaster William H. Downs, who had feuded with McGuire for years, had an unimpeachable alibi—he had spent the night with his family in a hotel in Hollywood.
Williard Kemp, an assistant district attorney of Santa Barbara County, had heard of Keeler’s emotograph and decided to try it. The question was how to get Downs to agree to be tested.
At the time the Ku Klux Klan was still very big in rural California, and Kemp was king kleagle of the local chapter. He suggested to Downs that a good way to keep the police off his back would be to join the Klan. And indeed, at that time and in that place, this was probably true. So a fake initiation ceremony was arranged wherein Downs was strapped into Keeler’s emotograph and told that he had to tell the truth: “In the solemn secrecy of this room you may answer the questions put to you without fear or favor.” Thus Downs took the test but did not tell the truth. Still, when Keeler analyzed the results, he noticed an interesting pattern. When asked if anyone on a list of twenty-five men had been involved in the bombing, Downs had reacted strongly to seven of the names.
Five of the seven were rounded up for questioning. The questioning went on for several days, and the men were hooked up to the emotograph from time to time in order to add to the pressure on them. Finally Harvey Stonebarger, the owner of a local machine shop, broke down and confessed. It had been Downs and he along with Downs’s father and a local rancher named William Crawford. (Crawford’s was one of the names that Downs had reacted to.) Downs and his father had purchased the explosive, Stonebarger had helped place it, and, after Downs left town, Crawford had lit the fuse with his cigar.
But at the trial Stonebarger retracted his confession, saying it had been obtained by the police using coercive methods. The jury believed him. They discounted the newfangled emotograph and found the men not guilty.
Fred Inbau (1909–1998), a criminologist and law professor at Northwestern University, was a firm believer in the use of lie detectors, but only as an awe-inspiring prop. Inbau joined Northwestern Law School’s Scientific Crime Detection Laboratory in 1933. By the time the lab was taken over by the Chicago Police Department in 1938, he had become its director.
Inbau developed deceit, deception, trickery, and outright lies as interrogation techniques that would be used to trick suspects into confessing. He saw this as an acceptable substitute for the third-degree methods applied in the back room of the police station. These police tactics succeeded in eliciting confessions but, as with other methods of torture, compelled the innocent as well as the guilty to say what the questioner wanted to hear. Inbau’s book, Criminal Interrogation and Confessions, written with John E. Reid, has been the textbook of choice on the subject since it was first published in 1962. Inbau held, to put it mildly, conservative opinions. He believed that, short of physical persuasion, any method of obtaining a confession was acceptable.
When Miranda v. Arizona was decided in favor of Miranda and the Supreme Court mandated that all suspects had to be read their rights—now known as “Miranda rights”—Inbau was in violent opposition and wrote many scholarly articles arguing against it. As Professor Yale Kamisar put it, “One of Professor Inbau’s pervasive themes was that, no matter how the officer misbehaved, so long as his or her misconduct did not affect the reliability of the evidence, the evidence should be admissible. It should suffice that the officer was disciplined in a separate proceeding.”
Inbau saw the polygraph as a wonderful psychological tool for inducing suspects to confess. And with any luck the investigators would not have to use it at all. As he says:
There have been innumerable instances of confessions made as a result of a mere proposal to have a suspect submit to a lie-detector test. On many occasions suspects have confessed their guilt while waiting in the laboratory to be tested.
There also are many instances when suspects have confessed immediately after the operator had adjusted the instrument preparatory to making the test.
He then goes on to reassure the reader that none of these have been false confessions. But he does not say how he knows this.
The sound spectrograph was developed by Bell Telephone Laboratories in 1941 to analyze the frequency of the sound of the human voice as words are spoken. It seemed possible that the characteristics of each person’s voice were individual enough that someone could be identified by the graph of his voice. This proved to be so. During World War II voice analysis was used to identify the speakers on German military radio networks. Knowing which radio operator was using which radio helped in keeping track of troop movements and in sorting out the relationships between various units.
In the 1960s, at the urging of the FBI, Bell Labs did further work on voice spectrograms. Bell engineer Laurence Kersta did much of the research and became convinced that “voiceprints,” as he called them, could be used for identification with a high degree of accuracy. The word “voiceprint” was dropped, but the technique gained forensic acceptance after a 1972 Michigan State University study that quantified its effectiveness. The study found a false identification rate (in which the analyst picks the wrong person) of only 2 percent, and a false elimination rate (in which the analyst says there is no match when in fact there is) of only 5 percent.
The sound spectrograph was used in 1971 at the behest of a group of reporters who wished to verify that the voice they were listening to over the phone was in fact that of the billionaire recluse Howard Hughes. He was trying to debunk a spurious biography of him that had just been published, and reporters wanted to be sure that it was indeed Hughes doing the debunking. (It was.) The first time most of the public became aware of the technique was in 1972 when Richard Nixon’s White House tapes were authenticated.
Since 2002, analysts at the National Security Agency and the Central Intelligence Agency have been busy verifying the voices of Osama bin Laden and other al Qaeda leaders. This is especially true whenever a new tape of their threats is released.
The sound spectrograph is not to be confused with the Voice Stress Evaluator, or Psychological Stress Evaluator (PSE). Purportedly, this device can determine if a person is lying by analyzing the stress patterns in his voice. The efficacy of the PSE is currently in great dispute.