Born Liars (7 page)

These days Ekman teaches his face-reading techniques to police investigators, embassy officials and military intelligence officers. In these sessions, he begins by showing photos of faces in neutral poses on a computer screen. A microexpression appears for forty milliseconds and the student has to press a button to indicate which emotion it displayed: fear, anger, surprise happiness, sadness, contempt or disgust. Without training, they are invisible; with it, says Ekman, people greatly improve their ability to spot them.

Of course, we have a voluntary muscular system and thus a degree of control over our facial responses. Most of the time we can fake a smile well enough. But the more we are emotionally involved in a falsehood, and the higher the stakes – that's to say, the worse the consequences of being found out – the more likely it is that our face will give us away to the trained observer. It seems to have a mind of its own.

Silvan Tomkins used to open his lectures by declaring, ‘The face is like a penis!'

We've already seen that most of us aren't as good as we think we are at distinguishing truth from lies. You might expect that those whose jobs rely on making such distinctions would be better at it. According to Bond and DePaulo's meta-analysis, however, psychiatrists, judges, custom officials and policemen score no more highly than the general public on tests of lie-detecting ability.

Aldert Vrij, a professor at Portsmouth University and the author of
Detecting Lies and Deceit
, a seminal book in the field of deception research, believes that professional lie catchers, like most of us, focus too much on the stereotypical physical signs of deception and fail to pay enough attention to verbal ones. Some people, says Vrij, have a naturally ‘dishonest' demeanour even when they're telling the truth; conversely, those with a naturally ‘honest' demeanour can get away with a lot more. He cites the case of a Florida man, who became a prime suspect in a murder case after he appeared to be flushed in the face and embarrassed during his police interview. He was later found to be innocent.

If Ekman's research focuses on the signs of emotional stress experienced by a liar, Vrij is more interested in the effects of the mental strain that a lie imposes on the person telling it – their ‘cognitive load' – and in particular the manifestation of this strain in the liar's speech.

He believes that the tried and trusted methods used by police interviewers are seriously flawed. In their day-to-day work, says Vrij, the police often assume that if somebody is surly and uncooperative then they're probably lying. But his research indicates that because liars are concerned about not being believed, they are likely to come across as
more
helpful than truthful people during an interview. Another problem is that some police interrogators, following in the tradition of Gene Hunt from
Life On Mars
, charge aggressively into interviews, accusing the suspect of guilt from the off. Vrij's work suggests that such attempts to break down the defences of the suspect only strengthen them, because they shut down the conversation; the suspect, feeling threatened, makes short answers or clams up altogether. The more talking they have to do, the more mental effort they have to make, and thus the more likely they are to incriminate themselves. Vrij argues that the way to catch a liar is to make them talk more, not less.

The police interrogation guidelines aren't much help. The official manuals recommend several strategies to help interviewers decide whether they are being told the truth. The principal one focuses on visual signs, such as whether the suspect is making eye contact, or fidgeting a lot. But, as we've seen, there's little evidence that these cues are reliable. Another recommended approach, the Baseline Method, involves comparing a suspect's language and demeanour during small talk at the beginning of the interview with those they use in the interview proper. But, says Vrij, people naturally adopt different modes of talking at different points – whether or not they're telling the truth. A third approach, the Behavioural Analysis Interview (BAI) strategy, comprises a list of questions to which, it is suggested, liars and truth-tellers will give different responses. Again, there's sparse evidence that this works, says Vrij, and neither do the guidelines address the fact that the police, like the rest of us, carry around a host of unconscious prejudices. It's been shown that suspects are less likely to be believed if they have a foreign accent, and more likely to be believed if they're attractive, baby-faced, socially adept and articulate – even though these two last traits have been found to be specifically associated with good liars.

So what should the police be looking for? As we've established, lying is effortful. Liars have to think of plausible answers under questioning; they need to avoid contradicting themselves; they need to tell a story that is consistent with what their interrogator already knows; they must try to avoid revealing slips of the tongue; they have to remember what they've said in case they're asked to repeat it. Whilst they're doing this they also have to be monitoring their own speech and body language to ensure they're not giving themselves away, even as they know that if they show any signs of all the effort they're putting in then they will arouse suspicion. Vrij's preferred strategy is to increase the liar's cognitive load to the point where they simply can't manage to perform the mental juggling act.

One of the interviewing techniques he recommends is to ask the suspect to tell their story backwards. By putting this extra mental pressure on suspects, those already struggling with the effort to tell a consistent lie will make mistakes that give them away. In 2007, Vrij and his colleagues published the results of a study that tested the police's conventional techniques against their own. The research involved more than two hundred and fifty student interviewees and two hundred and ninety police officers. The interviewees either lied or told the truth about staged events. Police officers were then asked to spot the liars using the traditional methods. Those who focused on visual cues proved significantly worse at identifying liars than those looking for speech-related cues. The liars seemed
less
nervous and more helpful than those telling the truth. As Vrij predicted, the most reliable technique turned out to be the backwards-story test.

Another technique designed by Vrij is the sketch test: asking people to draw a scene they claim to have witnessed. While it's non-verbal, this technique also puts pressure on the cognitive facilities of the liar. In Vrij's study, thirty-one participants – all of them members of the police or armed forces – were sent on a mock mission to pick up a laptop from a ‘secret agent'. Afterwards they were asked to make a detailed drawing of the location at which they'd received the laptop. Half the participants were instructed to tell the truth, the other half to lie. Vrij hypothesised that the liars, in order to make their lie convincing, would sketch a location they'd actually been to in the past, furnishing it with the kind of detail often thought to be the hallmark of a good lie. He also predicted that in doing so they'd forget a key part of the scene: the agent. The truth-tellers would be much more likely to draw the man with the laptop, as he was such a central part of the scene they had in their minds. So it proved. On the basis of this factor alone it was possible to detect who was lying nearly ninety per cent of the time.

Although Ekman and Vrij place different emphases on what to look for, both agree on the importance of taking a holistic approach. When assessing truthfulness, a person's voice, hand movements, posture and speech patterns should all be taken into account, and it's vitally important to put all of this in context: do these behaviours contrast in revealing ways with how the person usually acts, and how do they square with everything else that's known about the situation? Such judgements require many fine and fallible calculations; there is no single, all-purpose tell we can use as a short-cut. Pinocchio's nose remains a fairy tale.

In 2008 a group of Norwegian researchers ran an experiment to better understand how police investigators come to a judgement about the credibility of rape claims. Sixty-nine investigators were played video-recorded versions of a rape victim's statement, with the role of victim played by a professional actress. The wording of the statement in each version was exactly the same, but the actress delivered it with varying degrees of emotion. The investigators, who prided themselves on their objectivity, turned out to be heavily influenced in their judgements by assumptions about the victim's demeanour: she was judged most credible when crying or showing despair. In reality, rape victims react in the immediate aftermath of the event in a variety of different ways: some are visibly upset; others are subdued and undemonstrative. It turns out there is no universally ‘appropriate' reaction to being raped. The detectives were relying on their instincts, and their instincts turned out to be constructed from inherited and unreliable notions about women in distress.

Shakespeare's warning about how hard it is to read ‘the mind's construction in the face' is supported by a host of empirical evidence, yet interrogators remain stubbornly convinced of their ability to tell if a person is truthful by observing them, and relying only on their gut instincts. The lawyer and fraud expert Robert Hunter calls this misapprehension the ‘demeanour assumption'. He cites the case of the American student Amanda Knox, arrested in 2007 for the murder of Meredith Kercher. The Italian police concluded she was guilty based almost entirely on their assessment of Knox's demeanour under intense questioning: ‘We were able to establish guilt,' declared Edgardo Giobbi, the lead investigator, ‘by closely observing the suspects' psychological and behavioural reactions during the interrogations. We don't need to rely on other kinds of investigation as this method has allowed us to get to the guilty parties in a very quick time.' Giobbi's logic is dangerous, because people do not behave in police custody or in court in the same way as they might outside it and, guilty or innocent, some people will always behave suspiciously.

Of course, it's not just police investigators who suffer from this bias. We all have a tendency to make instant judgements about a person's integrity based on received ideas about appropriate demeanours. Italian prosecutors were quick to leak stories about Knox doing cartwheels while in custody, and when the press published pictures of her with a smile on her face readers around the world reacted the same way: no innocent person accused of a crime would behave like this. But people react to intense pressure in unpredictable ways, and a single photo yields no reliable information about a person's inner thoughts. But as unscientific as it is, the demeanour assumption plays a part in some of society's most important mechanisms. Hunter points out that it underpins the notions of oral evidence and jury trials; those who
watch
witnesses give evidence are assumed to be best placed to judge whether they are telling the truth.

What does it stem from, this over-confidence in our intuitions about lying? It probably has something to do with our innate tendency towards self-absorption, and our difficulty in recognising that other people are as fully rounded and complex as we are. Emily Pronin, a psychologist at Princeton University, reminds us that when two people meet there is a fundamental asymmetry about the way they relate to one another. When you are talking to someone, there are at least two things more prominent in your mind than in theirs – your thoughts, and their face. As a result you tend to judge others on what you see, and ourselves by what you feel.
You
know when you're hiding your true thoughts and feelings – pretending to be fascinated by your boss's endless anecdote, or grinning your way through a terrible first date – but you nonetheless tend to assume the other's appearance tells the full story of how
they
feel – if she's smiling, it's because she's really
enjoying herself. It's been found that people over-estimate how much they can learn from others in job interviews, while at the same time maintaining that others can only get a glimpse of them from such brief encounters. The model we tend to work with is something like this:
I
am infinitely subtle, complex and never quite what I seem,
you
are predictable and easy to read. ‘I suppose no one truly admits the existence of another person,' sighs the narrator of Fernando Pessoa's
Book of Disquiet.

Paradoxically, this asymmetry makes us less confident than we should be in our own ability to lie, because we assume other people can read our faces as easily as we suppose we can theirs. In Edgar Allan Poe's short story
The Tell-Tale Heart,
a man who has committed a murder is being interviewed by clueless detectives. He becomes convinced that they can just
tell
he's guilty, and breaks into a needless confession. It's a dramatic example of what the psychologist Thomas Gilovich terms the ‘illusion of transparency' – the irrational but often irresistible conviction that others can read our minds. A dinner guest suspects that her hostess can read her distaste for the food she's being served; a secret admirer guesses the object of his crush must have an inkling of how he feels; a business executive gets the overwhelming feeling that everyone in the room can tell how nervous she feels about making a presentation. We have a powerful tendency to exaggerate such fears and intimations, because although we mentally compensate for the fact that we have better access to our inner states than others do, we find it hard to compensate
enough
.

Gilovich carried out a series of experiments to demonstrate that we're much harder to read than we imagine. In one of them, groups of participants played a round-robin lie detection game – a version of Call My Bluff. Each person told either a lie or a truth and the rest of the group had to guess which was which. The ‘liars' in each group consistently over-estimated the extent to which the other would guess they were lying. The effect was particularly pronounced amongst those who scored higher on a separate test for self-absorption.