Hypothesis: I believe when people are lying, they will mostly show it on their eyes and facial expressions.

May 7, 2009 11:13 PM

Lie to Me is a fairly new US drama series based on the work of Paul Ekman (previously) who studied facial movements and what they reveal about the emotional state of a person (see Malcolm Gladwell's "The Naked Face", also previously, for more background). No doubt inspired by the show, Gisela, a sixth grader at Mesa Grande Elementary School, decided for her science project to see whether she could tell if a person was lying based on their facial expressions. Here is the result: part 1, part 2 (youtube videos).

As part of the experiment, she asked people questions and observed their expressions to see if they were lying. The six questions she asked:

1. Please describe the American flag.
2. Please describe an elephant.
3. Please describe a giraffe.
4. Please describe what you're wearing right now.
5. Please describe what type of weather we're having right now.
6. Please describe what your hands look like.

As part of the experiment, Gisela and her brother Victor are filmed lying and telling the truth. Maybe the best video of all is when her father puts her and her brother through making faces drills.

posted by Deathalicious (56 comments total)

Best of the Web.

O_o
posted by felix betachat at 11:48 PM on May 7, 2009 [2 favorites]

I knew this sounded familiar.

Paul Ekman studied under Silvan Tomkins, whose work on the physical display of emotions lead to the Affect Theory and Script Theory. Both of which are very interesting approaches to framing human emotions and behavior.

Also, the videos are pretty amazing.
posted by mrzarquon at 11:52 PM on May 7, 2009 [1 favorite]

I use some of this stuff when students tell me about why they came to class an hour or so late. One of the easiest tells out there is to watch the eyes. Usually, if someone looks up and to the left, it's because they're accessing the memory centers in the brain. If they look up and to the right, they're using the creative centers of the brain, in other words, most likely they're not telling the truth.

Sicilians are great liars. The best in the world. I'm Sicilian. My father was the world heavy-weight champion of Sicilian liars... Now, what we got here is a little game of show and tell. You don't wanna show me nothin', but you're tellin me everything.
posted by Ghidorah at 12:21 AM on May 8, 2009

Experiment ends Meh.
posted by Mblue at 12:57 AM on May 8, 2009

The six questions she asked:
1. Please describe the American flag.
2. Please describe an elephant.
3. Please describe a giraffe.
...

What is this, a pop quiz? I'm as smart as the average American, but I can't answer complicated questions like this. No wonder my body is freaking out.

Friggin 6th graders with their "knowledge" and their "truth". Get off my lawn!
posted by twoleftfeet at 1:13 AM on May 8, 2009 [1 favorite]

Also, the guy at 3:14 in the Part 2 video has microexpressions that can only be described as macro.
posted by twoleftfeet at 1:31 AM on May 8, 2009 [1 favorite]

The Lie To Me site says the show is "based on the real-life scientific discoveries of Paul Ekman", but it's pretty clear that the main character ("DR. CAL LIGHTMAN, the world's leading deception expert who studies facial expressions and involuntary body language") is based on Ekman himself. It is really bizarre to see one of these campy Fox dramas be about a famous living psychologist. Usually the science in shows like Bones or Fringe or CSI is laughably bad, and the TV scientist is a supermodel who is nothing like a real scientist. It's kinda funny that Ekman is letting himself be dolled up and dumbed down in this way. I'd like to think that he'll keep the science on the show honest, but given the ridiculousness of the premise, there's no hope of that happening. But who can blame him? Having a fictional version of your life glamorized in a TV show sounds pretty fun.
posted by painquale at 1:35 AM on May 8, 2009

Analayze ? really ?
posted by Pendragon at 2:23 AM on May 8, 2009

1. Please describe the American flag.
2. Please describe an elephant.
3. Please describe a giraffe.
4. Please describe what you're wearing right now.
5. Please describe what type of weather we're having right now.
6. Please describe what your hands look like.

None of these are good questions as the answers to all of them are known in advance. She obviously knows what a giraffe looks like, so studying facial expressions is pointless if you already know they're telling a lie.

She should have asked questions to which she didn't know the answer but could verify later.
posted by splice at 2:41 AM on May 8, 2009 [1 favorite]

Been thinking of making a post about this. I wonder if there's a body of work where all the collected wisdom of the series is bundled (whether on DVD or book), that is not called "Mike Caro's Book of Poker Tells"...
posted by NekulturnY at 2:52 AM on May 8, 2009

I found question six sort of redundant, as my hands are shaped like windy giraffes made out of American flags.
posted by Uppity Pigeon #2 at 3:22 AM on May 8, 2009 [2 favorites]

I wonder if there's a body of work where all the collected wisdom of the series is bundled (whether on DVD or book), that is not called "Mike Caro's Book of Poker Tells"

Sure there is.

"Reap 'em and Reap", by Joe Navarro.
posted by splice at 3:58 AM on May 8, 2009 [1 favorite]

None of these are good questions as the answers to all of them are known in advance. She obviously knows what a giraffe looks like, so studying facial expressions is pointless if you already know they're telling a lie.

I thought the same thing, but then I watched her video. Her hypothesis isn't quite as it is presented in this post. It isn't "can you tell if someone is lying" (i.e. testing the listener). It is "do these body language variables correlate to lying" (i.e. testing the liar). She has a checklist for eyes, hands, body, etc either being still or moving. Perhaps not an ideal procedure, but I was impressed at the dedication and thoroughness from a 6th grader.
posted by DU at 4:29 AM on May 8, 2009

For those interested in Ekmans' work, you might also be interested in the work of Adam Kendon on gesture, and in Affect Control Theory, which works on predicting behavior and emotions by understanding roles and behaviors with the semantic differential.
posted by dragonsi55 at 4:52 AM on May 8, 2009 [6 favorites]

None of these are good questions as the answers to all of them are known in advance. She obviously knows what a giraffe looks like, so studying facial expressions is pointless if you already know they're telling a lie.

Alright now, it's not like this is the peer review process for her dissertation, it's a sixth grade science experiment.

Also, is Gisela available for babysitting, because she seems like the kind of kid I want my daughter to be hanging around (plus I can tell when she's lying).
posted by Pollomacho at 5:12 AM on May 8, 2009 [4 favorites]

Where's the Ira Glass signal?
posted by Potomac Avenue at 5:25 AM on May 8, 2009 [2 favorites]

Whoa, I had this on a background tab, and at 3:23ish in the first clip, I thought my cell phone was ringing. Yeah, that's my ringtone. Got a problem with it?
posted by uncleozzy at 5:49 AM on May 8, 2009

Usually the science in shows like Bones or Fringe or CSI is laughably bad, and the TV scientist is a supermodel who is nothing like a real scientist.... It's kinda funny that Ekman is letting himself be dolled up and dumbed down in this way.

Isn't it almost exactly the same thing that Kathy Reichs -- who was a serious forensic anthropologist once, and whose crime novels reputedly get the science correct -- let Fox do with her life in Bones?
posted by aught at 6:21 AM on May 8, 2009 [1 favorite]

Watching a person's eye movement to tell if they're lying is considered by some to be a myth.
posted by drezdn at 6:24 AM on May 8, 2009

I imagine it could be interesting for a sixth-grader to know exactly when her teachers are being economical with the truth.
posted by permafrost at 6:36 AM on May 8, 2009

I use some of this stuff when students tell me about why they came to class an hour or so late. One of the easiest tells out there is to watch the eyes. Usually, if someone looks up and to the left, it's because they're accessing the memory centers in the brain. If they look up and to the right, they're using the creative centers of the brain, in other words, most likely they're not telling the truth.

That is Eye Accessing Cues from NLP by Bandler and Grinder. Eckman studied Facial Expressions and subsequently Micro-Expressions.
posted by P.o.B. at 6:41 AM on May 8, 2009

A friend of mine is what some call a natural. It's a pain to be around her but either way. She says looking at eyes has nothing to do with spotting a lie. She has noticed that when people lie they usually look you straight in the eye perhaps to counter the myth. She looks for incongruity, for example if someone says that they had a flat tire with a slight look of elation on their faces. Then she would ask which tire and the liar would usually take longer to answer that, because s/he wasn't prepared for that. She says she doesn't call out on people's lies more than half the time.
posted by Lucubrator at 6:41 AM on May 8, 2009

I use some of this stuff when students tell me about why they came to class an hour or so late. One of the easiest tells out there is to watch the eyes. Usually, if someone looks up and to the left, it's because they're accessing the memory centers in the brain. If they look up and to the right, they're using the creative centers of the brain, in other words, most likely they're not telling the truth.

Yeah, that's total bullshit. Eye movements do not correlate to activation of "memory centers" (and I'd be further interested to hear where you think these "memory centers" are). Bandler and Grinder were pretty much total hacks. The study of micro-expressions is a hell of a lot more subtle, and to do serious analysis, video that can be broken into discrete frames is required.
posted by solipsophistocracy at 6:54 AM on May 8, 2009 [3 favorites]

I'm not going to vouch for Bandler and Grindler, but I wouldn't call it total bullshit. He didn't come up with that eye stuff willy nilly. They came up with the idea after going through quite a few people. You could actually go out and test the stuff your self and see there is some validity to it. Also, Looking straight ahead is one of the signs according to Grinder.
posted by P.o.B. at 7:19 AM on May 8, 2009

I just started watching Lie to Me, and I love it. I'd like to read more about Eckman's work. Though I've noticed, if I watch more than one Lie to Me ep in one night, I'm painfully self conscious about my own "micro-expressions" for awhile the next day.

I did anthropology undergrad. In a forensic anthro class, we were given novels by Aaron Elkins to read, and then use our class notes to figure out whether his forensic anthropologist protagonist could have figured out the clues he did, from the evidence presented.
Best assignment EVER!
posted by SaharaRose at 7:21 AM on May 8, 2009

Here is Eye Accessing Ques broken down.
posted by P.o.B. at 7:22 AM on May 8, 2009 [1 favorite]

She looks for incongruity, for example if someone says that they had a flat tire with a slight look of elation on their faces. Then she would ask which tire and the liar would usually take longer to answer that, because s/he wasn't prepared for that. She says she doesn't call out on people's lies more than half the time.

That technique is a little law enforcement secret, particularly in law enforcement dealing with lots of different kinds of people because different cultures may have different conventions about eye contact, body language, authority figures, etc.
posted by Pollomacho at 7:24 AM on May 8, 2009

I would have loved to be at the pitch meeting for Lie To Me.

It's a police drama, OK? And the star is a guy who can always tell when someone is lying! So picture the scene at headquarters... They're questioning a suspect. And when he says he wasn't involved in the crime, the star says, "You're lying!" And that solves the case! He can do it every week!
posted by Joe Beese at 8:05 AM on May 8, 2009

No scientific studies have corroborated the eye-accessing cues theory of NLP.

(NLP is one of my pet peeves.)
posted by wittgenstein at 8:08 AM on May 8, 2009 [3 favorites]

Oh, I love this show, but I'd hate to always know when people were lying. It'd be a little like having the power to read minds, and I don't want to know what most people are thinking.
posted by desjardins at 8:11 AM on May 8, 2009

Part of my act (previously) involves a lie detection routine that totally like fer real uses all these sophisticated NLP techniques and microexpression cues to determine when someone is lying. I must have keen insight into the human mind...my psychological powers can't possibly be explained by carefully-hidden sleight of hand and misdirection dressed up in psuedo-scientific buzzwords! In fact, my only real question is why the FBI hasn't already hired me as a consultant...I have the VESTS!

Here, I'll teach you guys a simple one that will work for you about 75% of the time:

IMPROMPTU LIE DETECTION
1. Ask someone about a trip they recently took. (Or whatever.) Have them list four facts about the trip, with one catch: one of the facts should be a lie. Rush them along a little bit...explain the concept then make them start as soon as you can.
2. The third one they list is the lie.
3. TA-DA!

Top Tip: If they begin any of the facts with a really long "um" or "uh"...that one's the lie, Sherlock. Also, watch their eyes: if they look away just before they state a fact, they're about to lie. This isn't because of a hardcore physiological fact about "accessing different parts of the brain" or whatever, though; it's because they're trying to think of a lie and they're not very good at it.

Confidential To You: Yes, I know you personally would never do something so obvious as put the lie in the third spot. I wasn't talking about you, I was talking about all the dull-eyed easily-duped sheeple out there. You're not one of those, are you? Of course you're not. You're a creative and unique individual with refreshingly original thoughts and experiences. You'd NEVER pick the third slot.
posted by Ian A.T. at 8:49 AM on May 8, 2009 [6 favorites]

I'm not going to vouch for Bandler and Grindler, but I wouldn't call it total bullshit. He didn't come up with that eye stuff willy nilly. They came up with the idea after going through quite a few people. You could actually go out and test the stuff your self and see there is some validity to it. Also, Looking straight ahead is one of the signs according to Grinder.

I wasn't calling out NLP on the whole as total bullshit. I think it's fascinating, though definitely a pseudo-science. While I've never seen any empirical evidence to back it up, I owe it a debt of gratitude in that, when I was in high school, hearing about NLP is what turned me on to the study of mind and brain. However, I can say, without reservation, that correlating eye movement to "memory centers" or "creativity centers" is total bullshit. Where is the "memory center" in the brain? Are you talking about the hippocampus? It could be considered the memory center in some respects, though memories aren't really stored there. The lateral temporal lobes? Temporal poles? Medial parietal cortex? Medial prefrontal cortex? I guess you could refer to the "frontoparietal-temporo-occipital" network (i.e. the brain) as the memory center if you wanted to. There are dozens of anatomical regions involved in different kinds of memory retrieval, and autobiographical memory in particular relies on a diffuse network. NLP might be nifty, but it sure as hell isn't informed by neuroscience.
posted by solipsophistocracy at 9:07 AM on May 8, 2009 [7 favorites]

"Usually, if someone looks up and to the left, it's because they're accessing the memory centers in the brain. If they look up and to the right, they're using the creative centers of the brain, in other words, most likely they're not telling the truth."

If they look back and to the left...back and to the left...back...and to the left, they're being shot in the head by someone other than Lee Harvey Oswald.
Best way to catch a lie is to know the answer beforehand and lead someone to predicate everything on that false bit of information. Once you reveal that, and everything falls like a house of cards, they'll usually assume you know the whole story already. And you 'prove' to them you know the whole story by revealing pieces selectively. Stuff you have questions about, but have a slight chunk of. People tend to want to make sense of things, so you just let their mind create the order in the events, faces, whatnot, for you, so they'll pad it out. All you have to do is keep quiet and listen. (S'how they got Zubaydah to give up KSM...as opposed to the b.s. about 'enhanced interrogation).
posted by Smedleyman at 9:12 AM on May 8, 2009 [1 favorite]

If you watch the face drills video with the sound off, it looks like she's taking a dump.

Just sayin'...
posted by orme at 9:30 AM on May 8, 2009

Neurolinguistic programming? Bah!

An enormous 2003 review of the research literature (not anecdotal stuff) on cues to deception (available here in PDF form) demonstrates pretty clearly that eye-gaze aversion is pretty useless as a cue to lying, and that there are many, many other cues that are far better. On p. 20 of the PDF, the authors of the review specifically point out the uselessness of eye-gaze aversion, eye-contact, and eye-shifts as pretty poor indicators of deception.
posted by anaphoric at 9:43 AM on May 8, 2009 [2 favorites]

Obviously I'm an idiot for having brought this up in the first place. The way I heard it (and no, I'm not a scientist having anything to do with studying the brain) is that it had to do with the different hemispheres of the brain controlling different aspects of thinking. The right hemisphere, as I was taught in school (however briefly) is the creative, language processing part, and the left contains memory and functions like math.

Hey, maybe I'm wrong, and if so, apologies. I was relaying an anecdote that rarely, if ever, has failed for me. On the other hand, I've always been pretty good at reading body language (unless that's all a fraud as well) and being able to figure out if someone is lying.
posted by Ghidorah at 10:09 AM on May 8, 2009

Quackwatch.com on NLP:

"Neurolinguistic programming (NLP) is a variable system of procedures purported to enable people to communicate more effectively and influence others. It is said to involve modifying the patterns or "programming" created by interactions among the brain (neuro), language (linguistic), and the body that produce both effective and ineffective behavior. Proponents claim that NLP has cured phobias, allergies, and other problems in one or a few brief sessions. Its core postulates are: (a) people are most influenced by messages that reflect how they internally represent whatever they are doing; and (b) this representation is reflected by eye-gaze patterns, posture, tone of voice, and language patterns. The internal representation can be visual (picturing what they are involved with), auditory (hearing it sounded out), or can involve other senses. Proponents claim, for example, that a someone experiencing a mental image might use the words "I see," whereas someone in an auditory mode might say "that sounds right to me. Scientific studies have demonstrated no correlation between eye movements and visual imagery, reported thoughts, or language choices. A National Research Council committee has found no significant evidence that NLP's theories are sound or that its practices are effective."

Unfortunately because of it's "catchiness", I'm afraid this little meme is going to be with us for a long, long time.
posted by storybored at 10:19 AM on May 8, 2009

I was relaying an anecdote that rarely, if ever, has failed for me.

How do you know? Maybe you've been calling people out wrongly all these years?
posted by nasreddin at 10:46 AM on May 8, 2009

Sicilians are great liars. The best in the world. I'm Sicilian. My father was the world heavy-weight champion of Sicilian liars... Now, what we got here is a little game of show and tell. You don't wanna show me nothin', but you're tellin me everything.

You fell victim to one of the classic blunders. The most famous is never get involved in a land war in Asia. But only slightly less well-known is this: Never go in against a Sicilian when death is on the line!
posted by rokusan at 10:56 AM on May 8, 2009 [1 favorite]

Somebody call me when Penn Jillette does a special about this.
posted by rokusan at 10:58 AM on May 8, 2009

I don't care what you all say, Tim Roth is awesome.
posted by Justinian at 11:00 AM on May 8, 2009

I quite like Lie to Me, but just because Tim Roth is in it.
posted by ob at 11:11 AM on May 8, 2009 [1 favorite]

Victor is adorable.
posted by spec80 at 11:41 AM on May 8, 2009

Quick question to those who watch Lie to Me: are they overselling the tells in order to make the TV audience feel good about themselves? I've pretty consistently been able to spot the tells right away (knew who the bad guy was in the most recent episode pretty much right away, for example). I assume this is because if they made it too difficult then people wouldn't enjoy it as much.
posted by Deathalicious at 11:57 AM on May 8, 2009

Quick question to those who watch Lie to Me: are they overselling the tells in order to make the TV audience feel good about themselves?

It's probably because any purposefully acted "microexpression" is going to be a lot less micro and a lot more expression. Or, yeah, it could be done on purpose to make the audience feel good.
posted by Justinian at 12:11 PM on May 8, 2009 [1 favorite]

are they overselling the tells in order to make the TV audience feel good about themselves?

Honestly, I don't even watch for the microexpressions because I know it's going to be blatantly pointed out to me. In the most recent episode, I only knew what was going on because [hover for spoiler], not because I was watching for any tells. This show has somewhat of a unique premise, but it's still subject to the formulaic plots and predictability of any crime TV show out there.
posted by desjardins at 12:14 PM on May 8, 2009

Pretty sure you can guess easily because the plots are generic TV drama plots complete with telecast "twists" in all the expected places, not because of any overacted tells.

I mean, you can guess every Law and Order and CSI twist, too, right?

But, yes, Tim Roth makes all fine with me too.
posted by rokusan at 12:27 PM on May 8, 2009

Yes, I know you personally would never do something so obvious as put the lie in the third spot. I wasn't talking about you, I was talking about all the dull-eyed easily-duped sheeple out there. You're not one of those, are you? Of course you're not. You're a creative and unique individual with refreshingly original thoughts and experiences. You'd NEVER pick the third slot.

I'm actually such a nerd about these kinds of things that I almost never trust myself to randomly generate numbers or permute lists mentally. One trick that I use when I don't have a computer around or I can't write the options down and shuffle them is to use my watch's seconds hand as a pseudo-random number generator. So for a single 1-4 random choice I could just map each quadrant of the watch face to the four possible values. The watch method has the advantage that it's not obvious that I'm even generating a random number (unlike, say, flipping a coin) which can come in handy in situations like making a pseudo-random choice in the middle of a poker game.
posted by burnmp3s at 12:34 PM on May 8, 2009 [1 favorite]

[hover for spoiler]
posted by Potomac Avenue at 12:51 PM on May 8, 2009

It's part of my plot to encourage use of the abbr tag, Potomac Avenue.
posted by desjardins at 1:02 PM on May 8, 2009

In 9th grade, I was assigned a project at school to research a topic by reading a how-to book, and present an instructional presentation to the class on anything we wanted. I chose reading body language, and proceeded to not read the book and base my entire presentation off things I had read online about a year ago. So my presentation consisted of an awkward 7 minutes in which I listed all the random things about body language I could think of, and then it was time for the practical part of the presentation.

I had three volunteers come up and, after I left the room, agree on one person to tell a lie. Then when I came back in, they'd each tell a story about some event in their childhood. I was supposed to read their body language and determine who was lying.

I failed. Then we did it three more times, and I got each of those wrong. I did worse than I should have, speaking from probability. I was sabotaging my own success by looking for these certain things, instead of actually using the methods of common sense and deep evaluation we were all born with. I'm not sure there's a book that can really teach you to read body language. You really just have to hang around a lot of liars. Or watch C-Span.
posted by Bleusman at 1:41 PM on May 8, 2009

Alright, I was waffling as to whether or not to watch this show. Now I'm sold. Paul Eckman co-wrote a book with THE DALAI LAMA. Therefore, this is automatically awesome.

I'm also now a fervent Patriots fans since The Dalai Lama's appearance at Gillette Stadium last week. (self-link)
posted by grapefruitmoon at 2:54 PM on May 8, 2009

The idea of this show bothers me. I picture people who watch it, and think they know something scientific because they watched a 10-second fictionalized explanation of something that would take years of study to genuinely understand, making wild accusations:

"You're lying! You touched your face! Your hand turned a certain way!"
posted by sageleaf at 3:15 PM on May 8, 2009

Don't get me wrong, here. I'm not espousing NLP by any means. I also don't think the eye accessing stuff is the greatest. Everything I've heard it is only good about 50% of the time. By a far margin there are other clues that are more useful. The thing is, none of this stuff is sound enough to be admissible in court, and when you run into people who have mental conditions a lot of it goes out the window. A good sense of intuition is almost just as useful.
posted by P.o.B. at 4:46 PM on May 8, 2009

When googling around for links about Ekman and NLP, I came across this blog post, which is one of the more sober and interesting things I've read about NLP.
posted by painquale at 6:41 PM on May 8, 2009 [2 favorites]

I like Lie to me because it reminds me of House before House got bad, not because I think I can learn things from it. It's just a cognitive heat sink.
posted by edbles at 7:36 PM on May 8, 2009