I’m tired of the stories that you always tell
Shakespeare couldn’t tell a story that well
You’re the largest liar that was ever created
You and Pinnochio are probably related
– Too Much Joy, “That’s a Lie”
There’s an amusing video clip circulating this weekend of late-night talk show host Jimmy Kimmel subjecting some poor kid named Blake to a fake lie detection machine — ably aided in his task by the pinkified “Truth Fairy.” Blake gamely answers whatever question Kimmel puts to him, but the “feedback” — in the form of a blaring noise signaling the detection of a lie — is administered pretty arbitrarily. Kimmel gets the poor kid to admit he doesn’t like school, hates science, but likes math (!), sometimes he uses bad words, and that he has peed in the pool at least once — although there’s a bit of pressure being placed on the kid, as the clip below shows, to answer the way Kimmel wants him to.
At one point, Kimmel even asks, “Do you think I could get you to say anything with this machine?” Initially Blacke says “No,” setting off the “lie detector.” Then he answers yes. And that, my friends, is only a bit of an exaggeration of the problems with standard polygraph machines as a lie detection technique.
Lying is probably as old as humankind, and the elusive ability to tell when someone is lying has consumed a great deal of brainpower over the ages. Or, as David Thoreson Lykken phrases it in his classic book, A Tremor in the Blood: “If man learned to lie not long after he acquired language, we may assume that the first attempts at lie detection soon made their appearance….We are all human lie detectors; we must be to survive in our mendacious society.”
So it’s not surprising, then, that lie detection has a long and colorful history. It has its roots in instruments of torture, most notably during the European Middle Ages, when it was believed that subjecting the body to extreme physical agony would force the victim to blurt out the truth. (We now know that this is far from the case. An Italian Enlightenment thinker, Cesare Beccaria, wrote in 1764, “By this method, the robust will escape, and the feeble be condemned. These are the inconveniences of this pretended test of truth.”) In 1730, Daniel DeFoe suggested it might be possible to measure someone’s heart rate to detect deception.
The evolution of the modern lie detector, or polygraph machine, began with the first tests to determine the physical responses of the body during the act of deception. In 1895, the so-called “ ‘Father of Modern Criminology,’” Cesare Lombroso used a device called a plethysmograph to monitor changes in the blood flow of a subject during interrogation; two years later, in 1897, B. Sticker developed a method of measuring the galvanic responses of an individual under interrogation: i.e, the amount of sweat they produced as determined by the electric conductibility of their skin.
Finally, in 1914, Vittorio Benussi began to study the breathing rates of individuals, using pneumatic tubing wrapped around the subject’s chest to measure depth and rate of breath. He found that the “ratio of inspiration and expiration was generally greater before truth telling than that before lying.” So not only could blood pressure, pulse rate and sweat production be linked to the act of lying, but breathing rates as well.
All these components are combined in the modern polygraph machine, which measures physical responses such as respiration, heart rate, pulse, and electrical skin conductance to determine if a subject is lying, are notoriously unreliable.
Its invention is largely credited to William Moulton Marston, an American psychologist who in 1915 began to demonstrate a lie detection test, which determined whether the subject was being deceptive using a blood pressure cuff, or sphygmomanometer, to take measurements of systolic blood pressure during interrogation.
(He also created the comic book character Wonder Woman under the alias Charles Moulton. Wonder Woman was known for her Lasso of Truth that compelled people to tell the truth when wrapped in its coiled — clearly, Moulton had serious trust issues. Maybe it had something to do with his polyamorous lifestyle: he and his wife lived with a third woman, Olive Byrne, for many years.)
An American medical student and an employee of the Berkley police department,named John Larson, is credited with the first “polygraph” to be used in forensic science: he adapted the scientific procedure created by Marston in the Harvard Psychological Laboratory and adapted it to police procedure beginning in 1921. Like Marston, Larson recognized the importance of asking questions in the correct order, and wording them in specific ways, as being critical to lie detection — the apparatus was just the supporting device. Larson called his invention a “cardio-pneumo-psychogram,” because it documented blood pressure, pulse rate and respiratory rates, all on a drum of paper.
The problem with polygraph tests is that they are notoriously inaccurate and people can train themselves to beat the machine. Most notably, they only measure physiological responses; determining whether those responses indicate a lie is the job of the person administering the test — which makes the results highly subject to interpretation. Or as the American Civil Liberties Union puts it: “The lie detector does not measure truth-telling; it measures changes in blood pressure, breath rate, and perspiration, but those physiological changes can be triggered by a wide range of emotions.”
The most common countermeasures to beat the polygraph include sedatives, putting antiperspiranton the fingerprints, biting the tongue, lips or cheek, or placing tacks in one’s shoe. In Ocean’s 13, for instance, a character beats a polygraph test by stepping on a tack whenever he answers a question truthfully, skewing the machine’s readings and making it harder to determine the difference between lies and truth. But it’s not a simple matter either; it requires a bit of skill to beat the polygraph. The Mythbusters notoriously attempted to fool a polygraph in one of their episodes, and failed miserably.
Anyone still supporting the accuracy of the polygraph is going to have to come up with some seriously convincing solid evidence to win over the scientific community at this point. In 2003, the National Academy of Sciences released a report called The Polygraph and Lie Detection, concluding that most such research was “unreliable, unscientific and biased,” based on the group’s analysis of 57 research studies on which the APA bases its reliability.
Okay, those studies didn’t completely bomb: the report concluded that polygraphs can detect a lie “at a level slightly greater than chance, yet short of perfection”; however, correct results were habitually over-stated, “almost certainly higher than actual polygraph accuracy of specific-incident testing in the field.”
Lots of devices work wonderfully in the carefully controlled conditions of the laboratory, but “slightly greater than average” accurate readings doesn’t inspire a great deal of confidence in the technique. A common misconception is that, when properly conducted, a polygraph is accurate 80-99% of the time. The NAS report contradicts that widespread belief.
As recently as April 2009 researchers at the University of Florida conducted a study in the Journal of Forensic Sciences demonstrating the inaccuracy of standard lie detection techniques. The researchers hooked up 78 test subjects (men and women of all ages) to voice stress analyzes and used those devices to analyze vocal frequency of the speakers to determine when they were lying. The volunteers were instructed to lie while undergoing small electric shocks to simulate stress.
The result? “[T]he ‘true positive’ (or hit) rates for all examiners averaged near chance (42-56%) for all conditions,” the researchers concluded. “Most importantly, the false positive rate was very high, ranging from 40% to 65%.” That was true even when representatives from the device manufacturers conducted the tests, as opposed to the scientists.
The shortcomings of traditional polygraph techniques were succinctly demonstrated in a scene from Lie To Me‘s first season, when the fictional Cal Lightman debunks a new handheld lie detector device under demonstration. There actually is such a device being used by US Department of Defense called the Preliminary Credibility Assessment Screening System (PCASS); apparently it relies less on the judgment of a polygraph examiner and more on a special algorithm to determine whether a subject is lying, based on the measured physiological responses. But we’ve just seen that those responses can be misleading, and are not always indicators of deception.
In the episode, the (male) test subject performs quite well when being interrogated by a bland male examiner, but then Lightman sends in a sexy young woman to ask the same questions in a more flirtatious, suggestive manner — and the subject exhibits a physiological response similar to “lying” when he is really feeling self-conscious about his sexual arousal. Lightman likens the handheld device to a West African tribal custom in which a bird’s egg is passed to someone suspected of a crime. If the suspect broke the egg, s/he was found guilty, because obviously they broke it out of nervousness, and if they were nervous — well, then they must be guilty!
Lightman’s character is based on real-life scientist Paul Ekman, who pioneered the use of so-called “microexpressions” to determine whether or not someone is lying; he calls his approach the Facial Action Coding System, and it classifies every human expression, including the unconscious body mechanics of deception. For instance, there are telltale arm and movements, eye contact, and verbal contexts, all of which combined can reveal whether or not someone is being truthful (in theory, anyway).
A liar won’t make eye contact, and may compulsively touch his/her face, throat or mouth, or touch or scratch the nose or behind the ear. S/he will not be likely to touch the chest or heart area with an open hand. If someone says “I love you” while frowning, s/he is likely lying — the gestures or expressions don’t match the verbal statement. The timing and duration of gestures and expressions are also useful determining factors. And we’ve all met that person whose smile never quite reaches their eyes, making us feel like their cheerfulness is insincere.
It’s an admittedly inexact science, something the show makes clear: Lightman and his team are not infallible. For instance, he mistakenly concludes a mother is not sufficiently grieving for her child and hiding some guilty truth because of the lack of the telltale microexpressions accompanying such emotion. Then Lightman realizes that she is hiding something: her age. The mother has received Botox treatments, which numb the tiny facial muscles that give rise to microexpressions in the first place.
Further complicating matters is the fact that people lie for various reasons and motivations — not just because they are guilty of some crime. In the case of the aforementioned episode, “He Said, She Said,” Lightman determines that a female soldier has made a false accusation of rape against a male colleague. But she is lying on behalf of another woman who is too terrified to come forward — a noble impulse, even if the ends don’t justify the means.
Nonetheless, scientists are still looking for the equivalent of Wonder Woman’s magical lasso of truth. Scientists have been looking into using functional magnetic resonance imaging (fMRI) to achieve a kind of “brain fingerprinting” as a means of lie detection. In fMRI, when certain parts of the brain are engaged during a specific cognitive activity, those areas light up in the brain scan — and if a person happens to be “dissembling,” it should be possible to tell that they are lying just by looking at the scan.
Brain fingerprinting seems to offer something closer to an objective analysis of whether or not not someone is lying. How can a brain scan lie, after all? Well, maybe the scan doesn’t lie, but how we interpret those images is prone to human error, particularly since we don’t fully understand how this complicated organ called the brain actually works.
Chief among the naysayers of this new “mind reading” technology is Melissa Littlefield of the University of Illinois, who argues that the technique is based on fundamentally wrong assumptions, most notably that “truth” is the baseline, the natural state of being, and lying is adding “a story on top of the truth.”
An fMRI scan might reveal a lie if the person knew he or she was lying — if it were a conscious decision. But “some people don’t actually know that they’re lying, or have a told a lie for so long that it becomes their subjective interpretation of reality,” Littlefield explains. And just as with the polygraph test, it’s possible to cheat and beat the machine: just clench your teeth or move your head slightly. FMRI requires the subject to hold perfectly still to get a usable image.
Ed Yong recently blogged about this issue, noting some of the potential pitfalls of relying too heavily on what is still very much a nascent technology:
That hasn’t stopped fMRI from being marketed as a tool for lie detection – two companies called Cephos and No Lie MRI currently offer such services, the latter under a tagline of “New truth verification technology”. Nor has it deterred brain scans from being presented in courtrooms, with varying success. In recent years, two US judges have dismissed fMRI-based evidence, but a murder suspect in India was sentenced to life imprisonment after brain scans supposedly revealed that she had knowledge about a crime that only the killer could have possessed.
There are defenders of the technique’s potential for lie detection as well. Recent fMRI work on truthfulness comes to us via Joshua Greene of Harvard University, who published his results in 2009 in the Proceedings of the National Academy of Sciences. He found that honest subjects showed almost no additional brain activity when telling the truth, as might be expected — you’re not inventing a lie, after all. But dishonest subjects did show extra brain activity… even when they were telling the truth. Greene’s conclusion: “Being honest is not so much a matter of exercising willpower as it is being disposed to behave honestly in a more effortless kind of way.”
Then again, compulsive honesty in all situations has its drawbacks. Would any of us really want to live in a world where nobody has any kind of filter and hurtful truths are uttered on a daily basis? Lie To Me‘s Eli Loker (Brendan Hines) has taken a vow of radical honesty in response to his work — which includes admitting to his female love interest, Ria, that he’s only slightly above average in terms of sexual performance. Ria gets in her own zinger when Lightman asks if she has any specific training in spotting deception: “Well, I’ve dated a lot of men.” Little white lies are an integral part of our social fabric. As the Mitchell & Webb sketch below illustrates, sometimes one can be too much of a stickler for the truth.
Images: (top) Marston (right) asks questions, H. C. Weaver inspects polygraph record made by Berkley Psychograph on table. C. D. King takes blood pressure with Tycos sphygmomanometer. Olive B. Richard records testee’s answers (Marston 1938, frontispiece). Source. (center) Joe Larson, one of the inventors of the modern polygraph machine, demonstrates it in the 1930s. Source: How Stuff Works (bottom) No-Lie MRI.