Confessions of a Lie Detector: years of theft, sex, and murder

by Jim Wygant

  –––

  In any culture, the number of words available to describe something is an indication of how important that circumstance is and how pervasive. A culture that dwells in Arctic cold for ten months of the year may have a dozen different words for snow. Each word will represent a different critical characteristic of snow in the air or on the ground, qualities like how wet or how dry it is, whether it is drifting or has crusted. In our culture there are many kinds of lies, and many degrees of damage done by them. The list of words we use to identify a lie is long: deception, falsehood, untruth, deceit, prevarication, invention, whopper, distortion, bullshit, inaccuracy, misrepresentation, white lie. That’s only a partial list. We might conclude that our extensive vocabulary for lying is an indication of how important truth is for us – and how often it is abused.

  5. The Machine

  Most people regard the polygraph instrument with a mixed sense of awe, fear, and distrust. There is a mystique that attaches to the instrument, as though it could discover secrets. In fact, the instrument is often confused with the process, so that a popular misconception ignores the role of the examiner and imagines that attaching a polygraph instrument to a human being is all that is necessary to reveal lies.

  The polygraph instrument is only an aid to the process of lie detection. As such, it is the most refined and most accurate method of many that have been tried for centuries. Long ago someone figured out that a dry mouth often accompanied lying, so a simple test was devised. The suspect was asked to hold uncooked rice in his mouth for a designated period of time. When he expelled it, the moisture was checked. Rice that remained dry condemned the suspect as guilty. Moist rice meant a reprieve.

  One of the most astute methods of primitive lie detection may be apocryphal but is too brilliant to ignore. A suspect was told to go alone into a darkened tent in which a donkey was tied. He was instructed to yank on the donkey’s tail. He understood that if the donkey brayed it would be assumed by the judges outside the tent that the suspect was guilty; if the donkey remained silent the suspect was presumed innocent.

  What was not revealed to the suspect was that the donkey’s tail had been coated with black soot. The judges had no interest at all in any braying by the donkey. However, they did want to inspect the suspect’s hands when he emerged from the tent. Clean hands meant that the suspect had not pulled the donkey’s tail and was therefore guilty. Sooty hands belonged to the innocent. Pity the poor innocent man who was too clever to follow the judge’s instructions. If he decided not to risk his innocence by pulling the donkey’s tail, he only succeeded in falsely condemning himself.

  It’s a good story, but as a method of lie detection it would be useful only one time in any community. Once word got around about the soot, the judges would encounter people who tried to “beat the test.”

  Most of the components of a contemporary polygraph instrument began to be developed and refined in a series of experiments conducted near the beginning of the twentieth century. Although initial experiments were crude, they led to the ultimate development of the respiration, skin resistance, and cardio components used in current polygraph instruments.

  The galvanometer had been around since Luigi Galvani published his work in 1791, but 100 years passed before anyone concluded that a galvanometer could reveal something about human regard for immediate circumstances. Experimenters in the 1890s began reporting the results of their work with humans. They had attached galvanometer electrodes, asked questions, and observed changes in the electrical resistance of the skin.

  In 1895 in Italy, Cesare Lombroso published his theories about blood volume changes. His apparatus consisted of a large glass beaker with a hole in one side big enough to be fitted with a rubber glove. The glove extended into the beaker and created a seal around the hole, so the beaker could be filled with water that completely surrounded the rubber glove. He instructed his suspect to insert his hand into the rubber glove. Lombroso marked the level of the water while the suspect remained in a “normal” or resting state. He then asked critical questions. If the level of the water rose, Lombroso theorized that this was because of an increase in blood volume in the gloved hand. He asserted that such an increase corresponded to lying or guilt.

  We know now that increases in peripheral blood supply at skin-level and the deeper flow to the skeletal muscles do often correspond to lying, but Lombroso’s apparatus was probably not sensitive enough to establish that.

  Sir James Mackenzie, a famous British heart specialist, developed his own “ink polygraph” around 1906 and published a description in 1908 in a British medical journal. In function, his instrument was remarkably similar to modern polygraphs. It wrote on paper with ink, rather than recording by a more common method of that time, lines scratched onto smoked paper. A spring driven clockwork mechanism moved the chart paper. The instrument monitored blood pressure, using a cuff, and breathing. Oddly, it hardly figures in the history of lie detection. Mackenzie intended it as a medical instrument to assist his diagnoses of heart disease. His instrument is not known to have ever been used as an aid to discovering truth.

  In 1908 the U.S. Congress took its own look at physiological instruments. The Committee on the Judiciary published a report entitled “Juvenile Crime and Reformation.” The report included an extensive list, with illustrations, of bizarre devices thought to be helpful in diagnosing and remedying juvenile delinquency. One instrument was identified as “the polygraph”, said to be a “French name” (the device was manufactured in Paris). The instrument was nothing more than a kymograph, a clockwork mechanism with a revolving drum for recording whatever was attached. The kinds of attachments were left up to the determination of the operator.

  The report also described a “myograph” built by a Frenchman named Marey. It functioned much like our simpler blood pressure cuff. Another apparatus, attributed to Hugo Münsterberg, a Harvard professor of psychology, was described as a tiny car that was pulled up and down a measured track, supposedly in response to changes in “muscular sense.” Mosso’s plethysmograph and sphygmomanometer were updated versions of Lombroso’s glove. And Marey’s pneumograph was an elaborate mechanical device to monitor respiration, as a much simpler flexible rubber tube does today. Marey also devised a series of wrist-mounted contraptions to monitor pulse rate and blood pressure.

  All of the devices in the Congressional report reflect a growing fascination with the possibility of relying upon physiological changes as an indication of mental or emotional states. Although we can see precursors of some polygraph components in these devices, there was not yet any reliable means of determining truth or deception with any of these isolated gadgets.

  Of greater interest than the Congressional report was the book “On the Witness Stand,” issued in the same year as the report, 1908, by Münsterberg, the Harvard psychology professor. Münsterberg speculated in his book about the possible benefits of incorporating several different physiological measurements into a single lie detector device. He described in remarkable detail something that did not exist with all of his suggested components until 1931. He would have used a smoked drum to record variations in respiration, obtained with a flexible rubber tube around the chest. He would have included measurements of cardio responses, and he even suggested use of galvanic skin response to record changes in electrical resistance in the skin. Cardio and respiration elements were incorporated in the first real recording polygraph built by Dr. John Larson in the early 1920s, but it was Leonarde Keeler who finally added to the polygraph a measure of skin resistance in 1931.

  Münsterberg was more interested in false confessions and the unreliability of eyewitness accounts, and he is not known to have attempted to build the lie detector instrument he described. His vision had no known impact on the development of the polygraph, which might otherwise have been accelerated.

  In 1914 in Italy, Vittorio Benussi fastened a rubber tube around the chest of a suspect and concluded that lying was revealed when breathing
became slower. His findings could not be replicated. Today a decrease in rate of respiration is not by itself regarded as a clear indication of lying, although it is likely to be given weight based upon the context in which it appears. Benussi’s focus on respiration, even if his theory was a little incomplete, gave direction to research that later established clear and valid indicators of lying.

  In the United States in 1917, William Marston published his theory that increases in blood pressure corresponded to lying. Marston’s technique was too crude to be reliable. He used a standard medical blood pressure cuff and determined blood pressure with the aid of a stethoscope, just as doctors do today. He had no means of making a continuous recording of changes, so he repeatedly inflated and deflated the cuff, noting any changes in blood pressure while asking questions.

  Criminal defense attorneys have always been alert to developments that might help their clients. Such an attorney represented a man named James A. Frye, who in 1921 was accused of murder. The attorney asked Marston to use his lie detection procedure with Frye.

  Frye was a young black man in Washington, D.C., charged with killing a wealthy black physician. He confessed to the murder during police questioning. With his admission, all that appeared to remain was a guilty plea and a sentence. Before that could happen, Frye recanted. He claimed he had only confessed because he was misled by the police. He said they promised to drop a separate robbery charge, and minimized the significance of the murder charge by telling him there was insufficient evidence to convict of that.

  Marston met Frye in jail and tested him there with a blood pressure cuff that he repeatedly inflated and deflated. He concluded that Frye was telling the truth when he denied the murder.

  The defense attorney attempted to introduce Marston’s testimony at the trial. The judge refused it. The jury convicted Frye of second degree murder and he went off to jail, where he served nearly 18 years.

  The case is notable because of the “Frye v. United States” decision rendered in 1923 by the U.S. Court of Appeals for the District of Columbia. The defense had appealed Frye’s conviction on several grounds, including the refusal of the judge to admit the “lie detector” testimony. In deciding against the defense, the appeals court established what has become know as the “Frye test” for admissibility of scientific evidence. One of the most critical elements of that test is that a procedure must have “gained general acceptance in the particular field in which it belongs.” The court concluded that Marston’s opinion of truthfulness, based only upon his own research with a blood pressure cuff, did not meet that standard.

  The “Frye test” continued for years to be the primary gauge for admissibility of any testimony or evidence that purported to be scientifically based. Only in the 1980s did other decisions finally begin to replace it. The 1992 U.S. Supreme Court decision in “Daubert et al. v. Merrell Dow Pharmaceuticals” was generally regarded as the last nail in the coffin of the Frye test. The court concluded, “That austere [Frye] standard, absent from and incompatible with the Federal Rules of Evidence, should not be applied in federal trials.” This was an abrupt change, in language that was almost a rebuke to the lower courts, both federal and state, for relying on Frye for years.

  Instead of Frye, the court said that the usual considerations of relevancy and reliability should be the standard. The court directed trial judges to decide admissibility in consideration “of whether the reasoning or methodology underlying the testimony is scientifically valid and of whether that reasoning or methodology properly can be applied to the facts in issue.”

  So what does all of that have to do with the development of the polygraph instrument? I have wandered off into this tale of Frye only because of the false mythology that has surrounded it. The procedure used by Marston on Frye has often mistakenly been regarded as a polygraph test. In fact, the instrument used by Marston was only a simple blood pressure cuff which he had to continually re-inflate, a grossly imprecise method for measuring slight variations in blood pressure. It has been mistakenly reported that Frye’s innocence was proven by Marston’s test and that Frye was acquitted. It has also been wrongly claimed that someone else confessed to the murder and Frye was released from jail after only a few years. Marston himself contributed to this confusion with his own subsequent writings. Unfortunately, among his other sins he was not a reliable reporter.

  –––

  Around 1921 we encounter the earliest example of something that functioned like a contemporary polygraph instrument and was used for lie detection. Dr. John A. Larson, who worked with the Berkeley, California, police department, built a primitive instrument that recorded respiration and cardio function at the same time. The original recording medium was a smoked drum. As it revolved, lines were scratched onto the surface. That messy procedure was soon replaced by paper and ink. A former policeman, Leonarde Keeler, became interested in Larson’s work and formed an informal partnership with him. Together they continued to develop both the equipment and the methods under which it was employed.

  The two moved to Illinois, where Larson became assistant state criminologist and Keeler joined the staff of the Scientific Crime Detection Laboratory at Northwestern University in Evanston, a few miles north of Chicago.

  In 1931, Keeler added to his instrument the last of the three primary elements of a modern polygraph, the galvanometer. It is a mysterious tool. While the blood pressure and respiration portions of a polygraph instrument rely upon mechanical linkages to physiological functions we can feel within ourselves, the galvanometer has no mechanical sensors and records something we can not feel. It uses fingertip electrodes to record changes in electrical resistance or conductance in the skin, a measure once known as galvanic skin response or G.S.R.

  Although other gadgets have been added to polygraph instruments since Keeler’s days, the primary components are still those devised by Keeler and Larson: monitors for respiration, blood pressure, and skin resistance or conductance. The primary difference between the earlier instruments and those in use now is sensitivity and precision. A contemporary computerized instrument gathers the analog data using electronic sensors and then converts it to the digital form necessary for a computer to accumulate and display. This has also permitted creation of computer software that evaluates charts, offering an examiner a second opinion that is neutral and disinterested and based entirely on physiological data.

  The polygraph instrument does not detect lies. An examiner can only attempt to do that within the context of a validated test procedure. To hook someone up to a polygraph instrument and then engage in an unstructured conversation with them would reveal very little that was meaningful. Polygraph tests have become highly structured.

  Many people who want quick and easy answers are impatient with the demands of a structured process. Several decades ago the “voice stress analysis” procedure was offered as a speedier alternative to a polygraph examination. It eventually became popular as a joke item. “Lie detector” wristwatches with flashing lights were advertised in popular magazines and “apps” were offered for cell phones. No attachments to a suspect were needed, only the sound of his or her voice. What could be handier? Unlike a polygraph, there was no structure needed for a voice stress analysis procedure. The device could be used in regular conversation, over the phone, or from any other audio source, including statements made on television. Expensive versions of these gizmos were sold to some police agencies. The manufacturers made outrageous claims about accuracy. When research repeatedly showed results no better than chance, the manufacturers claimed that polygraph examiners were trying to discredit them.

  Some police agencies still rely on voice stress analysis. In 1998 a 12-year-old girl, Stephanie Crowe, was murdered in her home in California. Police questioned a homeless man seen in the area but released him after confiscating his sweatshirt. They quickly focused their attention on Stephanie’s 14-year-old brother and two of his teenaged male friends. During long videotaped interrogations the boys init
ially denied any involvement in Stephanie’s murder. They were given voice stress analysis tests, which they were told revealed deception. Under hours of continuous questioning, they finally made incriminating partial admissions, enough to charge all three with murder. Two years later, on the eve of trial, a defense attorney had the homeless man’s sweatshirt analyzed. Blood on the shirt was matched to that of Stephanie. Charges against the boys were dropped and the homeless man was indicted. His trial was prolonged at considerable public expense because of his opportunity to use the charges made against the boys in his own defense. He was convicted and remains incarcerated.

  The families of the boys sued the manufacturer of the voice stress equipment and the police agencies involved. The insurance company for the manufacturer quickly paid off in an undisclosed amount, although the manufacturer asserted that his machine did not make mistakes and it must have been used improperly. The case against the others was tossed out but was then reinstated by an Appeals Court that concluded, that the boys had been subject to “hours of grueling, psychologically abusive interrogations.” The court further stated that murder charges had been placed against, “innocent teenagers for a crime they did not commit.” A psychologist who had consulted with the police settled for nearly a million dollars in 2010. The police agencies and the officers who worked for them appealed to the U.S. Supreme Court, which in January 2011 refused to hear the case, effectively setting the stage for either an unwinnable trial or a settlement, at taxpayer expense.