How well can we recognise lies?
Whenever we hear someone talking, we consciously or unconsciously form an opinion as to whether we can believe them. In doing so, we rely on our eyes and ears. And rightly so?
On television, everything looks very simple. The suspect raises the corners of his mouth for a fraction of a second during the interrogation: he is happy because the investigators suspect his bomb is in the wrong place. When the right place name is mentioned, anger flashes across his face. And as he protests his innocence, he shrugs his shoulders. For the lying expert sitting opposite him, the matter is now clear. The suspect's body language contradicts his words - he is lying.
The lie expert is the alter ego of Paul Ekman, 86, the world's most famous lie and emotion researcher. He has not only advised the makers of the crime series "Lie to me", but also numerous US authorities such as the FBI and CIA. His credo: The truth is written all over our faces.
The idea has a long tradition. An Indian scripture from 900 BC describes a poisoner as follows: "He does not answer questions or evades them, he speaks nonsense, rubs his toes on the ground and trembles, his face is pale, he rubs the skin with his fingers ..."
At the beginning of the 20th century, Sigmund Freud is said to have said during a case analysis: "Whose lips are silent, he chats with his fingertips: Treachery seeps out of his every pore." Since the middle of the 20th century, security agencies in the USA in particular have been trying to get to the bottom of the truth with a lie detector. The so-called polygraph registers, among other things, how sweat production, pulse and breath change when critical questions are asked. In Germany, lie detector tests have no probative value in court; here, statements are tested for credibility using a content analysis (this also applies to Switzerland, note from the Galaxus editorial team). But in everyday life, whenever we hear another person speak, we continue to search for the truth in the same way our ancestors did: with our eyes and ears.
Ekman was the first to conduct large-scale research into directly observable evidence for lies and truth. The psychologist, who was a professor at the University of California from 1971 to 2004, formulated the theory of universal human facial expressions for the basic emotions of anger, disgust, joy, fear, sadness and surprise in the 1960s. He catalogued the facial muscles involved in a kind of identification book of facial expressions, the "Facial Action Coding System". He laid the foundation for his popular lie theory in 1969 in an article with co-author Wallace Friesen on non-verbal signals from patients. The core idea: emotions that are supposed to remain hidden are sometimes revealed in facial expressions and movements of the arms, hands, legs and feet. A prime example is a brief facial expression that lasts a quarter to half a second at most and is barely perceptible, if at all, to untrained observers.
Such micro-expressions, in which hidden feelings seep through, do not occur all that often, according to Ekman. More often, he says, we observe aborted or incomplete emotions. In the case of feigned fear or sadness, for example, the characteristic wrinkles on the forehead are missing, and the eye muscles are not involved in a feigned, fake smile. Ekman does not see such inconsistencies as proof, but merely as an indication of deception. And it requires repeated or different clues - just one is not enough. In his book "Telling lies", he claims that in laboratory experiments, lies and truth can be distinguished 80 per cent by facial expressions alone and 90 per cent when body movements, voice and speech are taken into account.
Such figures seem "simply implausible", comments Maria Hartwig from John Jay College of Criminal Justice, a criminology school in New York. According to research literature, success rates are generally only just above chance hits. Even if Ekman assumes intensive training: as far as is known, he has not published a single corresponding study that would substantiate his figures.
"Ekman's idea of using microexpressions to detect deception is not taken particularly seriously by many researchers," says legal psychologist Kristina Suchotzki from the University of Mainz. Not only because of the lack of empirical evidence: there is also a lack of theory. "If someone is afraid during an interrogation, that doesn't mean they are lying. You can't infer deception from an emotion."
Kristina Suchotzki is currently the most active German lie researcher. Above all, she looks for signs of the mental effort involved in making false statements. Because lying is not that easy: you have to conceal the truth, invent a coherent alternative story, put yourself in the other person's shoes, have any traitorous feelings under control and appear authentic at the same time. "Until now, emotions and cognitions have usually been analysed separately. I would like to bring the two together and clarify what exactly goes on in the mind when lying," she says. She does not believe that using microexpressions to recognise deception is very promising. "There is simply a lack of studies that support Ekman's claims."
One of the few independent tests was carried out in 2008 by Stephen Porter and Leanne ten Brinke, who asked their test subjects to hide their true feelings when confronted with sad, scary or happy images. When they feigned a different emotional state, their facial expressions often appeared incongruous. Microexpressions were seen in two per cent of all snapshots. They appeared at least once in 22 per cent of the test subjects - but not only when they were supposed to be masking their feelings.
Ekman and his critics agree on one thing: People are generally very poor lie detectors. The most frequently cited hit rate comes from a meta-analysis and is based on data from around 25,000 test subjects: They were correct in 54 per cent of cases - which, given that half of the statements were true and half were false, is hardly better than letting chance decide. With audio-only recordings, the hit rate was 63 per cent higher than with video recordings with or without sound - apparently the image distracted from the relevant characteristics. Whether someone had to deal with lies more frequently in their profession, be it as a police officer, judge or psychiatrist, did not play a role. The supposed experts were no better than laypeople.
Parents did not judge their children better than strangers
But what if you know someone as well as you know your own child? A Canadian experiment investigated whether parents are better at recognising their offspring's lies than other parents and students. All three groups were shown videos of 8 to 16-year-old children giving information on whether they had looked for the answers to a test in a forbidden way. The parents were no better at distinguishing between lies and truth than students and other parents. The following was true for all three groups: they would have been just as likely to flip a coin, but tended to trust their own judgement and were more likely to believe the children's statements to be true - the parents most of all.
One of the authors of the study, psychologist Kang Lee from the University of Toronto, was apparently not at ease with this. In a lecture, he presented a picture of his son lying. Lee used a device that measures blood flow in the skin to analyse what was hidden behind his neutral, meaningless facial expression. He calls what he discovered the "Pinocchio effect": Blood flow decreases in the cheeks and increases in the nose.
"The idea that blood flow in the nose could be an indication of lying is absurd," comments Kristina Suchotzki. "Such claims are dangerous because they suggest that the characteristics are also useful in practice, for example at the airport." In a controlled situation in the laboratory, such an effect might be detectable. But no technology can solve the problem that supposed lie characteristics also appear in suspects who are telling the truth. "There is no clear sign of lying, only evidence that may indicate lying."
In a meta-analysis by a team led by Bella DePaulo, 14 out of 50 non-verbal traits recorded were associated with lying more often than others, most notably dilated pupils and tension. Most telling, however, was the impression of the statements themselves: False statements tended to come across as hesitant, ambivalent and uncertain. A German meta-analysis of 41 studies produced somewhat different results. The psychologists from the University of Giessen found evidence of increased self-control when lying: fewer movements of the hands, legs and feet as well as reduced head nodding.
"The effects are so small and unstable that they do not help to recognise lies in practice," says Kristina Suchotzki. Linguistic characteristics have proven to be somewhat more meaningful. "But even these effects are not large and the findings may be too optimistic."
Psychologist Aldert Vrij from the University of Portsmouth, one of the most active international researchers into lying, doesn't think much of non-verbal lying characteristics either. "Weak and unreliable" is his conclusion in a review paper together with Maria Hartwig and Pär Anders Granhag from the University of Gothenburg. They are hoping for more from linguistic cues - although these are hardly more closely related to lies than non-verbal characteristics. However, they can be elicited and reinforced with questioning techniques, as several of their own experiments have shown. There is no such extensive research for non-verbal characteristics.
Not surprising, as speech is easier to record. In order to reliably record facial expressions and gestures, trained observers or complex wiring of the face and body are required. It is only in recent years that researchers have been increasingly experimenting with computer-aided methods such as automatic facial recognition. These promise new insights because they can process large amounts of data and identify complex patterns.
Vrij, Hartwig and Granhag concede that more nuanced non-verbal characteristics may have been overlooked or ignored until now, such as subcategories of gestures as once defined by Ekman. If you look more closely, for example, you will find more pointing gestures for true statements and more metaphorical gestures for lies, such as the fist as a symbol of strength. Of all the non-verbal characteristics, Vrij believes that signs of physical tension are the most meaningful, as he writes 2020 with a colleague. But perhaps more or a combination of signs will be discovered if they are recorded using other methods,
When Hartwig and Bond combined various behavioural characteristics in a meta-analysis with over 9,000 test subjects, they were able to correctly classify around two thirds of the lies. However, most studies only test selected characteristics. A typical laboratory experiment does not usually create a realistic situation. There is no real interaction between the interviewer and the interviewee, and more importantly, the deception is carried out under instruction. And no one can say with certainty to what extent and under what conditions laboratory findings can be transferred to real offences.
In order to ensure that the test subjects have something at stake, they are usually promised money if they appear convincing. Kristina Suchotzki has already resorted to harsher means in the service of research and threatened her test subjects with mild electric shocks if the computer found their statements to be unbelievable. With the expected effect, as she and Matthias Gamer from the University of Würzburg discovered in an experiment. In the group that had committed a bogus theft, they observed a slower pulse and increased sweating of the hands with false answers, and the fear of the consequences intensified these differences.
Of course, if the subjects did not appear credible, they had nothing bad to fear in the lab. The situation was much more serious for the unwilling subjects in a field study at the University of British Columbia. Leanne ten Brinke and Stephen Porter analysed video recordings of 78 people who had contacted the public in search of missing relatives. Just under half were later found guilty of killing the missing person.
In terms of body language, the guilty relatives did not differ from the innocent ones, as the comparison of 75,000 still images revealed. But more signs of hidden emotions such as joy and feigned sadness appeared on the faces of the guilty, the two authors report. "Innocent people showed real, full-face grief and distress."
In addition, the guilty used more than twice as many vague phrases, for example: "Somebody's got to know something, somewhere. I think so. I think there's somebody who's got to be running scared, who knows what they're doing." Real appeals sounded clearer, more direct: "You can't imagine what Sarah means to us. We are a strong family, and we don't survive well apart. We need her home now, today, quickly as we possibly can."
A confident lie can be more credible than a stuttered truth
As impressive as such studies are: They do not solve the problems of lie research. The differences are small and the evidence is ambiguous. They only represent the average and at best offer rough indications for individual cases. A confident lie can appear more credible than a stuttered truth. This is because most people base their judgement on how competent, clear and unambiguous a statement appears, according to Maria Hartwig and Charles Bond in a meta-analysis. If they overlook lies, it is not because they are looking for the wrong signals. They fail more often when a person who inspires confidence lies or when an untrustworthy person speaks the truth.
Not knowing what is going on in others can cost us dearly. A good sense for the truth should therefore have prevailed in evolution. And yet we are easily fooled. Perhaps this is the downside of civilised coexistence: the harmless little everyday lies have taught us to be gullible.
Why do many people still think they can see lies? If you turn the question around, the answer is obvious: what if lies and truth were like two peas in a pod? If the guilty got away with it and the innocent paid in their place? The thought is hard to bear, writes Maria Hartwig. "We want to believe that liars betray themselves."
Tip from the Galaxus editorial team: Ricky Gervais showed what a world in which no one lies looks like in his comedy "The Invention of Lying".
Spectrum of Science
We are partners of Spektrum der Wissenschaft and want to make well-founded information more accessible to you. Follow Spektrum der Wissenschaft if you like the articles.
[[small:]]
Experts from science and research report on the latest findings in their fields – competent, authentic and comprehensible.