You can't tell if someone is lying by reading their facial expressions

So you've watched every single episode of Lie to Me and read every article from here to Mars on how to spot a fake smile and gotten pretty damn fantastic at spotting these fleeting, involuntary facial movements called microexpressions. You're pretty much ready to bet the farm that you could spot a liar just by reading the clues written all over his stupid little liar face. Except no. No you can't.

Microexpressions, as defined by psychologist Paul Ekman (who coined the term "microexpression," basically wrote the book on the little bastards, and has been studying their use in detecting deception for going on half a century, now), are:

...very brief facial expressions, lasting between 1/25th and 1/15th of a second. They occur when a person either deliberately or unsconsciously conceals an emotion being felt. Any one of the seven emotions found to have a universal signal may appear in a micro expression: anger, fear, sadness, disgust, contempt, surprise and happiness.

Microexpressions therefore fall under the umbrella of "body language" ("nonverbals," if you're one for parlance), and are distinguishable in that they refer explicitly to the face and specific situations in which they're likely to appear, viz. a situation where the emotion being felt is being either intentionally or unintentionally suppressed.

The big upshot is that microexpressions have the potential to reveal hidden sentiments (NB: They were first detected by psychologists Ernest Haggard and Kenneth Isaacs when the two of them were reviewing footage of people undergoing couple's therapy. Go ahead and mull that one over for a second), and so could serve some use to people (law enforcement, parents, poker players, a suspicious spouse) who stand to gain from knowing where someone really stands on an issue – or, more specifically, whether that someone is lying. The whole premise of the show Lie to Me, starring Tim Roth as an uncannily capable human lie-detector, is based almost entirely on research surrounding microexpressions.

It of course bears mentioning that while Lie to Me is a fictional TV show, microexpressions are 100% real. They're difficult to detect, and even more difficult to analyze, but they're real. But we repeat: you probably can't use them like Tim Roth.

There are several reasons for this, but arguably the most important one is that a microexpression, on its own, can really only tell you that a person is experiencing some emotion. Even if you can accurately determine which emotion it is – let's say it's fear – all you can really say from your assessment is that the person you're talking to is experiencing fear. Let's say you're interrogating someone suspected of murder and you pick up on their fearful microexpression. Are they afraid that you're close to figuring out they committed the crime, or afraid of being interrogated by some creep who is goggling at their face the way a dog might stare at a rack of ribs?

The point being that context – situational context, emotional context, the context of your relationship with the person you're scrutinizing – is incredibly important when reading someone's microexpressions. The importance of context applies to the analyses of other forms of body language, too, by the way. According to Joe Navarro, a former FBI counterintelligence agent and an expert on nonverbal cues, a person's feet tend to be more accurate in revealing sentiments and intentions than her face. And loads more important than someone's face or feet, on their own, is the concurrent reading of her entire body, which Navarro says "is constantly transmitting vital information."

But even if you have been trained to pick up on every last one of a person's corporeal clues, it remains incredibly unlikely that you are equipped to divine when they are or are not being deceitful. To quote Ekman, the pinoeering microagression researcher, again, this time from his book Telling Lies: "Most liars can fool most people most of the time." That includes people who have been taught to spot body language.

"Our research," he goes on to explain, "and the research of most others, has found that few people do better than chance in judging whether someone is lying or truthful." His research has also shown, and I'm going to bold this, just so that I can refer to it easily when I see someone bring this up in the comments "that most people think they are making accurate judgments even though they are not." Allow us to reiterate: YOU ARE NOT TIM ROTH.

Okay but so wait... like... what if you really CAN tell when people are lying? What if you really are Tim Roth? Well, then perhaps you are one of the "few exceptional people who," Ekman claims, actually "can quite accurately spot deceit."

And this is where things get a little weird.

Some years ago, Ekman teamed up with fellow UC San Francisco psychologist Maureen O'Sullivan to establish a research project that would not only investigate people's ability to detect lies, but actively seek out "Truth Wizards" – that is, people able to identify deception in other people in at least 80% of standardized trials. "Although most groups, including police officers, CIA and FBI agents, lawyers, college students and therapists, do little better than chance," said O'Sullivan in a 2004 statement, wizards can usually detect whether a person is lying, whether that lie is about an opinion, and "how someone is feeling about a theft."

By 2004, "The Wizard Project," had tested more than 13,000 people and O'Sullivan and Ekman had identified 31 wizards. By those figures, fewer than 25 people in 10,000 are good enough at detecting deception to merit wizard status.

But then, the fact that there are any wizards to speak of whatsoever is certainly surprising, is it not? This is the real question – are there, in fact, people who possess the uncanny ability to read fleeting facial expressions, hand gestures, or shifts in body weight and divine your true feelings, opinions, or knowledge on a subject? Could you really be Tim Roth?

A statistical critique of The Wizards Project by psychologists Charles F. Bond Jr. and Ahmet Uysal found O'Sullivan and Ekman's findings to be statistically and methodologically suspect. "Analyses reveal that chance can explain results that the authors attribute to wizardry," the researchers write in a 2007 issue of Law and Human Behavior. "Thus, by the usual statistical logic of psychological research, O'Sullivan and Ekman's claims about wizardry are gratuitous."

In the same issue of Law and Human Behavior, O'Sullivan cam to Ekman's defense as well as her own with a perspective piece titled "Unicorns or Tiger Woods: are lie detection experts myths or rarities?" The piece takes Bond and Uysal to task, claiming misrepresentation or misinterpretation by the researchers' of The Wizards Project and its findings.

The "lively debate" over Ekman and O'Sullivan's work on deceit detection, is ongoing – and on the subject of microexpressions, in particular, Ekman remains a contentious figure. His efforts to teach microexpression-analysis to TSA Agents has produced uninspiring results. Boise State University psychologist Charles Honts, a former DoD polygrapher who was trained by Ekman and now specializes in the study of deception, claims that every one of his attempts to replicate Ekman's experiments have failed. "There's not a lot of science to back up Ekman's claims," said Jay Nunamaker, head computer engineer of the digital lie-detecting Embodied Avatar project, in an interview with WIRED published earlier this year. "Applying them to deception detection is a reach."

In light of this contention, it is perhaps sufficient, in the debate over whether you are or are not capable of telling whether another person is lying, to refer to the title written by O'Sullivan in defense of her and Ekman's Wizard Project findings: "Unicorns or Tiger Woods: are lie detection experts myths or rarities?" When an ability becomes so rare as to be confused with myth, it's probably safe to assume you don't possess it.

In brief: No. You are not Tim Roth.