FOUR

Detecting Deceit from Words, Voice, or Body

“And how can you possibly know that I have told a lie?’

“Lies, my dear boy, are found out immediately, because they are of two sorts. There are lies that have short legs, and lies that have long noses. Your lie, as it happens, is one of those that have a long nose.”—Pinocchio, 1892

PEOPLE WOULD LIE less if they thought there was any such certain sign of lying, but there isn’t. There is no sign of deceit itself—no gesture, facial expression, or muscle twitch that in and of itself means that a person is lying. There are only clues that the person is poorly prepared and clues of emotions that don’t fit the person’s line. These are what provide leakage or deception clues. The lie catcher must learn how emotion is registered in speech, voice, body, and face, what traces may be left despite a liar’s attempts to conceal feelings, and what gives away false emotional portrayals. Spotting deceit also requires understanding how these behaviors may reveal that a liar is making up his line as he goes along.

It is not a simple matter to catch lies. One problem is the barrage of information. There is too much to consider at once. Too many sources—words, pauses, sound of the voice, expressions, head movements, gestures, posture, respiration, flushing or blanching, sweating, and so on. And all of these sources may transmit information simultaneously or in overlapping time, competing for the lie catcher’s attention. Fortunately, the lie catcher does not need to scrutinize with equal care everything that can be heard and seen. Not every source of information during a conversation is reliable. Some leak much more than others. Strangely enough, most people pay most attention to the least trustworthy sources—words and facial expressions—and so are easily misled.

Liars usually do not monitor, control, and disguise all of their behavior. They probably couldn’t even if they wanted to. It is not likely that anyone could successfully control everything he did that could give him away, from the tip of his toes to the top of his forehead. Instead liars conceal and falsify what they expect others are going to watch most. Liars tend to be most careful about their choice of words. Everyone learns in the process of growing up that most people listen closely to what is said. Words receive such great attention because they are, obviously, the richest, most differentiated way to communicate. Many more messages can be transmitted, far more quickly, by words than by the face, voice, or body. Liars censor what they say, carefully concealing messages they do not want to deliver, not only because they have learned that everyone pay attention to this source but also because they know that they will be held more accountable for their words than for the sound of their voice, facial expressions, or most body movements. An angry expression or a harsh tone of voice can always be denied. The accuser can be put on the defensive: “You heard it that way. There was no anger in my voice.” It is much harder to deny having said an angry word. It stands there, easily repeated back, hard to disavow totally.

Another reason why words are carefully monitored and so often the chief target for disguise is that it is easy to falsify—to state things that are not true—in words. Exactly what is to be said can be written down and reworded ahead of time. Only a highly trained actor could so precisely plan each facial expression, gesture, and voice inflection. Words are easy to rehearse, again and again. The speaker has continual feedback, hearing what he says, and thus is able to fine-tune his message. The feedback from the face, body, and voice channel is much less accurate.

After words, the face receives the greatest amount of attention from others. People receive commentary about the appearance of their face: “Wipe that look off your face!” “Smile when you say that!” “Don’t look sassy at me.” The face receives attention partly because it is the mark and symbol of the self. It is the chief way we distinguish one person from another. Faces are icons, celebrated in photographs hung on walls, placed on desks, and carried in wallets and purses.¹ Recent research has found that one part of the brain is specialized for recognizing faces.²

There are a number of other reasons why people pay such attention to faces. The face is the primary site for the display of emotions. Together with the voice, it may tell the listener how the speaker feels about what is being said—but not always accurately, since faces can lie about feelings. If there is difficulty hearing, watching the speaker’s lips can help the listener figure out the words being spoken. Attending to the face can also provide an important signal necessary for conversations to proceed. Speakers want to know whether their listeners are listening. Looking at the speaker’s face implies that, but it isn’t the most trustworthy signal. Bored but polite listeners can watch a speaker’s face while their minds are elsewhere. Listeners also encourage the speaker with head nods and “mm-hmms,” but these too can be faked.^*

Compared to the attention lavished on the words and face, the body and voice don’t receive much. Not much is lost, since usually the body provides much less information than the face, the voice much less than the words. Hand gestures could provide many messages, as they do in the sign language of the deaf, but hand gestures are not common in conversations among northern Europeans and Americans of that background, unless speech is prohibited.^† The voice, like the face, can show whether someone is emotional or not, but it is not known yet whether the voice can provide as much information as the face about precisely which emotions are felt.

Liars usually monitor and try to control their words and face—what they know others focus upon—more than their voice and body. They will have more success with their words than with their face. Falsifying is easier with words than with facial expression because, as mentioned earlier, words can be rehearsed more readily than facial actions. Concealing also is easier. People can more readily monitor their words than their face, censoring anything that could betray them. It is easy to know what one is saying; much harder to know what one’s face is showing. The only parallel to the clarity of feedback given by hearing words as they are spoken would be a mirror always in place showing each expression. While there are sensations in the face that could provide information about when muscles are tensing and moving, my research has shown that most people don’t make much use of this information. Few are aware of the expressions emerging on their face until the expressions are extreme.^*

There is still another, more important reason why there are more clues to deceit in the face than in words. The face is directly connected to those areas of the brain involved in emotion, and words are not. When emotion is aroused, muscles on the face begin to fire involuntarily. It is only by choice or habit that people can learn to interfere with these expressions, trying, with varying degrees of success, to conceal them. The initial facial expressions that begin when emotion is aroused are not deliberately chosen, unless they are false. Facial expressions are a dual system—voluntary and involuntary, lying and telling the truth, often at the same time. That is why facial expressions can be so complex, confusing, and fascinating. In the next chapter I will explain more about the neural basis for the distinction between voluntary and involuntary expressions.

Suspicious people should pay more attention to the voice and body than they do. The voice, like the face, is tied to the areas of the brain involved in emotion. It is very difficult to conceal some of the changes in voice that occur when emotion is aroused. And the feedback about what the voice sounds like, necessary for a liar to monitor how he sounds, is probably not as good for hearing the voice as it is for the words. People are surprised the first time they hear themselves on a tape recorder, because self-monitoring of the voice comes partly through bone conduction, and it sounds different.

The body is also a good source of leakage and deception clues. Unlike the face or voice, most body movements are not directly tied to the areas of the brain involved in emotion. Monitoring of body movements need not be difficult. A person can feel and often see what his body is doing. Concealment of body movement could be much easier than concealing facial expressions or voice changes in emotion. But most people don’t bother. They have grown up having learned it was not necessary to do so. Rarely are people held accountable for what they reveal in their bodily actions. The body leaks because it is ignored. Everyone is too busy watching the face and evaluating the words.

While we all know that words can lie, my research has found that people take others at their word and are often misled. I am not suggesting that the words be totally ignored. People do make verbal mistakes that can provide both leakage and deception clues. And even if there are no mistakes in the words, it is the discrepancy between the verbal line and what is revealed by the voice, body, and face that often betrays a lie. But most of the clues to deceit in the face, body, and voice are ignored or misinterpreted. I found this out in a number of studies in which I asked people to judge others shown to them on videotape.

Some saw just the face, others just the body, still others heard the speech run through a filter that made the words unintelligible but left the sound intact, and the rest heard or read the words. Everyone saw the same people—the nursing students, described in the last chapter, who told the truth or lied about their feelings while they watched films. Remember that in the honest interviews these students had seen a pleasant film showing the ocean and had been instructed to describe their feelings frankly. In the dishonest interviews they had seen a film showing gory medical scenes and had been instructed to convince the interviewer they were looking at another pleasant film, this time showing flowers. The interviewer couldn’t see which film the nursing student was watching. These students tried very hard to mislead the interviewer, for the stakes were very high. They believed our experiment tested how well they would be able to control their emotional reactions in the emergency or operating room.

In our studies of how well people can detect when these students were lying, we were interested not only in which source was the best—face, body, voice, or words—but also whether suspicious people did better than those not expecting to be misled. We divided the people who were to see or hear the videotape into two groups. Some we made suspicious by what we told them about the people they were to judge, and some we tried to keep unsuspecting. The unsuspecting group were not told anything about the experiment; no mention was made of possible deceit or lying. We just told them they would be seeing or hearing people talking about a film they were watching. In order not to arouse their suspicion, we buried the judgment they were to make about honesty in a long list of judgments they had to make about friendliness, extroversion, dominance, awkwardness, calmness, and so on.

Although a few nursing students were terrible liars and were easily detected, most of the students misled the unsuspecting judges. Those who saw just the face or heard just the words did the worst: they rated the nursing students as more honest when they were, in fact, lying. Suspicious people did not do much better. These judges were told all about the instructions given to the nursing students, and they were told that the people they were to judge would be either lying or telling the truth. They were asked to make only one judgment—honesty or deceit. Very few did better than chance in spotting which was which. Those who saw just the body did the best, but even they were right on only about 65 percent of their judgments, when chance would be 50 percent.³ A few people did very well, correctly identifying 85 percent of the liars. Some of these accurate judges were highly experienced psychotherapists with reputations for being expert clinicians. Some were just extraordinarily sensitive people in other professions.^*

It is not necessary to be so misled. People who have been told some of what is in this and the next chapter did very well in judging when the nursing students were lying, as well as the most experienced psychotherapists were able to do. Clues to some deceits can be learned. The lie catcher has a better chance if the deceit involves emotion, and the liar is not a psychopath, highly practiced, or a natural liar. There are three goals: to spot a liar more often; to misjudge the truthful less often; and, most importantly, to realize when it may not be possible to do either.

The Words

Surprisingly, many liars are betrayed by their words because of carelessness. It is not that they couldn’t disguise what they said, or that they tried to and failed, but simply that they neglected to fabricate carefully. The head of an executive search firm described a fellow who applied to his agency under two different names within the same year. When asked the fellow which name should he be called, “The man, who first called himself Leslie D’Ainter, but later switched to Lester Dainter, continued his prevaricating ways without skipping a beat. He explained that he changed his first name because Leslie sounded too feminine, and he altered his last name to make it easier to pronounce. But his references were the real giveaway. He presented three glowing letters of recommendation. Yet all three ‘employers’ misspelled the same word.”⁴

Even a careful liar may be betrayed by what Sigmund Freud first identified as a slip of the tongue. In The Psycho-pathology of Everyday Life Freud showed how the faulty actions of everyday life, such as slips of the tongue, the forgetting of familiar names, and mistakes in reading and writing were not accidents but meaningful events revealing internal psychological conflicts. Slips express, he said, “. . . something one did not wish to say: it becomes a mode of self-betrayal.”⁵ Freud was not specifically concerned with deceit, but one of his examples was of a slip that betrayed a lie. The example describes the experience of Dr. Brill, one of Freud’s early and well-known followers:

I went for a walk one evening with Dr. Frink, and we discussed some of the business of the New York Psychoanalytic Society. We met a colleague, Dr. R., who I had not seen for years and of whose private life I knew nothing. We were very pleased to meet again, and on my invitation he accompanied us to a café, where we sat two hours in lively conversation. He seemed to know some details about me, for after the usual greetings he asked after my small child and told me that he heard about me from time to time from a mutual friend and had been interested in my work every since he had read about in the medical press. To my question as to whether he was married he gave a negative answer, and added: “Why should a man like me marry?”

On leaving the café, he suddenly turned to me and said: “I should like to know what you would do in a case like this: I know a nurse who was named as co-respondent in a divorce case. The wife sued the husband and named her as co-respondent, and he got the divorce.” I interrupted him, saying: “You mean she got the divorce.” He immediately corrected himself, saying: “Yes, of course, she got the divorce,’ and continued to tell how the nurse had been so affected by the divorce proceedings and the scandal that she had taken to drink, had become very nervous, and so on; and he wanted me to advise him how to treat her.

As soon as I had corrected his mistake I asked him to explain it, but I received the usual surprised answers: had not everyone a right to make a slip of the tongue? It was only an accident, there was nothing behind it, and so on. I replied that there must be a reason for every mistake in speaking, and that, had he not told me earlier that he was unmarried, I would be tempted to suppose he himself was the hero of the story; for in that case the slip could be explained by his wish that he had obtained the divorce rather than his wife, so that he should not have (by our matrimonial laws) to pay alimony, and so that he could marry again in New York State. He stoutly denied my conjecture, but the exaggerated emotional reaction which accompanied it, in which he showed marked signs of agitation followed by laughter, only strengthened my suspicions. To my appeal that he should tell the truth in the interests of science, he answered: “Unless you wish me to lie you must believe that I was never married, and hence your psycho-analytic interpretation is wrong.” He added that someone who paid attention to every triviality was positively dangerous. Then he suddenly remembered that he had another appointment and left us.

Both Dr. Frink and I were still convinced that my interpretation of his slip of the tongue was correct, and I decided to corroborate or disprove it by further investigation. Some days later I visited a neighbour, an old friend of Dr. R., who was able to confirm my explanation in every particular. The divorce proceedings had taken place some weeks before, and the nurse was cited as co-respondent.⁶

Freud said that “the suppression of the speaker’s intention to say something is the indispensable condition for the occurrence of a slip of the tongue [italics in original].”⁷ The suppression could be deliberate if the speaker was lying, but Freud was more interested in instances in which the speaker is not aware of the suppression. Once the slip occurs, the speaker may recognize what has been suppressed; or, even then, the speaker may not become aware of it.

The lie catcher must be cautious, not assuming that any slip of the tongue is evidence of lying. Usually the context in which a slip occurs should help in figuring out whether or not the slip is betraying a lie. The lie catcher must also avoid the error of considering someone truthful just because there are no slips of the tongue. Many lies do not contain any. Freud did not explain why some lies are betrayed by slips while most are not. It is tempting to think that slips occur when the liar wants to be caught, when there is guilt about lying. Certainly Dr. R. should have felt deception guilt about lying to his esteemed colleague. But there has been no study—or even much speculating—that would explain why only certain lies are betrayed by slips.

Tirades are a third way liars may betray themselves in words. A tirade is different from a slip of the tongue. The speech blunder is more than a word or two. The information doesn’t slip out, it pours out. The liar is carried away by emotion, not realizing until afterward the consequences of what he is revealing. Often, if the liar had remained cool, he would not have revealed the damaging information. It is the pressure of overwhelming emotion—fury, horror, terror, or distress—that causes the liar to give away information.

Tom Brokaw, when he was the interviewer on NBC-TV’s “Today Show,” described a fourth source of deception clues. “Most of the clues I get from people are verbal, not physical. I don’t look at a person’s face for signs that he is lying. What I’m after are convoluted answers or sophisticated evasions.”⁸ A few studies of deceit support Brokaw’s hunch, finding that some people when they lied were indirect in their reply, circumlocutious, and gave more information than was requested. Other research studies have shown just the opposite: most people are too smart to be evasive and indirect in their replies.^* Tom Brokaw might miss those liars. A worse hazard would be to misjudge a truthful person who happens to be convoluted or evasive in his speech. A few people always speak this way. For them it is not a sign of lying; it is just the way they talk. Any behavior that is a useful clue to deceit will for some few people be a usual part of their behavior. The possibility of misjudging such people I will call the Brokaw hazard. Lie catchers are vulnerable to the Brokaw hazard when they are unacquainted with the suspect, not familiar with idiosyncrasies in the suspect’s typical behavior. I will discuss ways to avoid the Brokaw hazard in chapter 6.

No other sources of leakage and deception clues in words have been uncovered as yet by research. I suspect that not many more will be found. It is too easy, as I described earlier, for a deceiver to conceal and falsify words, although errors do occur—careless errors, slips, tirades, and circumlocutious or indirect speech.

The Voice

The voice refers to everything involved in speech other than the words themselves. The most common vocal deception clues are pauses. The pauses may be too long or too frequent. Hesitating at the start of a speaking turn, particularly if the hesitation occurs when someone is responding to a question, may arouse suspicion. So may shorter pauses during the course of speaking if they occur often enough. Speech errors may also be a deception clue. These include nonwords, such as “ah,” “aaa,” and “uhh”; repetitions, such as “I, I, I mean I really . . .”; and partial words, such as “I rea-really liked it.”

These vocal clues to deceit—speech errors and pauses—can occur for two related reasons. The liar may not have worked out her line ahead of time. If she did not expect to lie, or if she was prepared to lie but didn’t anticipate a particular question, she may hesitate or make speech errors. But these can also occur when the line is well prepared. High detection apprehension may cause the prepared liar to stumble or forget her line. Detection apprehension may also compound the errors made by the poorly prepared liar. Hearing how badly she sounds may make a liar more afraid of being caught, which only increases her pauses and speech errors.

Deceit may be revealed also by the sound of the voice. While most of us believe that the sound of the voice tells us what emotion a person feels, scientists studying the voice are still not certain. They have discovered a number of ways to distinguish unpleasant from pleasant voices but don’t yet know whether the sound of the voice differs for each of the unpleasant emotions: anger, fear, distress, disgust, or contempt. I believe such differences will, with time, be found. For now, I will describe what is known, and what looks promising.

The best-documented vocal sign of emotion is pitch. For about 70 percent of the people who have been studied, pitch becomes higher when the subject is upset. Probably this is most true when the upset is a feeling of anger or fear. There is some evidence that pitch drops with sadness or sorrow, but that is not as certain. Scientists have not yet learned whether pitch changes with excitement, distress, disgust, or contempt. Other signs of emotion, not as well established, but promising, are louder, faster speech with anger or fear and softer, slower speech with sadness. Breakthroughs are likely to occur measuring other aspects of voice quality, the timber, the energy spectrum in different frequency bands, and changes related to respiration.⁹

Changes in the voice produced by emotion are not easy to conceal. If the lie is principally about emotions felt at the very moment of the lie, then there is a good chance for leakage. If the aim of the lie was to conceal fear or anger, the voice should sound higher and louder, and the rate of talk may be faster. Just the opposite pattern of voice changes could leak feelings of sadness a deceiver is trying to conceal.

The sound of the voice can also betray lies that were not undertaken to conceal emotion if emotion has become involved. Detection apprehension will produce the voice sounds of fear. Deception guilt might be shown to produce the same changes in the sound of the voice as sadness, but that is only a guess. It is not clear whether duping delight can be isolated and measured in the voice. I believe that excitement of any kind has a particular vocal signature, but that is yet to be established.

Our experiment with the student nurses was one of the first to document a change in pitch with deceit.¹⁰ We found that pitch went up during deceit. We believe this occurred because the nurses felt afraid. There were two reasons why they felt this emotion. We had done everything possible to make the stakes very high so they would feel strong detection apprehension. And, watching the gory medical scenes generated empathic fear in some of the nurses. We might not have found this result if either source of fear was lessened. Suppose we had studied people whose career choice was not involved, for whom it was only an experiment. With little at stake, there might not have been enough fear to cause any change in pitch. Or, suppose we had shown the nursing students a film of a child dying, which would be more likely to arouse sadness than fear. While their fear of being caught would have acted to raise their pitch, this reaction could have been canceled out by sad feelings lowering their pitch.

Raised pitch is not a sign of deceit. It is a sign of fear or anger, perhaps also of excitement. In our experiment, a sign of those emotions betrayed the student’s claim that she was feeling happily contented in response to a film showing flowers. There is a danger in interpreting any of the vocal signs of emotion as evidence of deceit. A truthful person who is worried she won’t be believed may out of that fear show the same raised pitch a liar may manifest because she is afraid of being caught. The problem for the lie catcher is that innocents also are sometimes emotionally aroused, not just liars. In discussing how this problem confuses the lie catcher’s interpretation of other potential clues to deceit, I will refer to it as the Othello error. In chapter 6 I will discuss this error in detail, explaining how the lie catcher can guard against making it. It is, unfortunately, not easy to avoid. The voice changes that may betray deceit are also vulnerable to the Brokaw hazard (individual differences in emotional behavior), mentioned earlier in regard to pauses and speech errors.

Just as a vocal sign of an emotion, such as pitch, does not always mark a lie, so the absence of any vocal sign of emotion does not necessarily prove truthfulness. The credibility of John Dean’s testimony during the nationally televised Senate Watergate hearings hinged in part on how the absence of emotion in his voice—his remarkably flat tone of voice—was interpreted. It was twelve months after the break-in at the Watergate Democratic National Committee headquarters when John Dean, counsel to President Nixon, testified. Nixon had finally admitted, a month earlier, that his aides had tried to cover up the Watergate burglary, but Nixon denied that he had known about it.

In the words of federal judge John Sirica: “The small fry in the cover-up had been pretty well trapped, mostly by each other’s testimony. What remained to be determined was the real guilt or innocence of the men at the top. And it was Dean’s testimony that was to be at the heart of that question.... Dean alleged [in his Senate testimony] that he told Nixon again that it would take a million dollars to silence the [Watergate burglary] defendants, and Nixon responded that the money could be obtained. No shock, no outrage, no refusals. This was Dean’s most sensational charge. He was saying Nixon himself had approved the pay-offs to the defendants.”¹¹

The next day the White House disputed Dean’s claims. In his memoirs, published five years later, Nixon said, “I saw John Dean’s testimony on Watergate as an artful blend of truth and untruth, of possible sincere misunderstandings and clearly conscious distortions. In an effort to mitigate his own role, he transplanted his own total knowledge of the cover-up and his own anxiety onto the words and actions of others.”¹² At the time the attack on Dean was much rougher. Stories, reputedly from the White House, were leaked to the press, claiming that Dean was lying, attacking the president because he was afraid of being homosexually attacked if he went to jail.

It was Dean’s word against Nixon’s, and few knew for certain which one was telling the truth. Judge Sirica, describing his doubts, said: “I must say I was skeptical of Dean’s allegations. He was obviously a key figure himself in the cover-up. . . . He had a lot to lose. . . . It seemed to me at the time that Dean might well be more interested in protecting himself by involving the President than in telling the truth.”¹³

Sirica goes on to describe how Dean’s voice impressed him: “For days after he read his statement, the committee members peppered him with hostile questions. But he stuck to his story. He didn’t appear upset in any way. His fiat, unemotional tone of voice made him believable.”¹⁴ To other people, someone who speaks in a flat tone of voice may seem to be controlling himself, which may suggest he has something to hide. Not misinterpreting Dean’s flat voice would require knowing whether or not this tone of voice is characteristic of him.

The failure to show a sign of emotion in the voice is not necessarily evidence of truthfulness; some people never show emotion, at least not in their voice. And even people who are emotional may not be about a particular lie. Judge Sirica was vulnerable to the Brokaw hazard. Recall that newscaster Tom Brokaw said he interprets circumlocutiousness as a sign of lying, and that I explained how he could be mistaken because some individuals are always circumlocutious. Now Judge Sirica could be making the opposite mistake—judging someone to be truthful because he fails to show a clue to deceit, not recognizing that some people never do.

Both mistakes arise from the fact that individuals differ in their emotional expressiveness. The lie catcher is vulnerable to errors unless he knows what the suspect’s usual emotional behavior is like. There would not be a Brokaw hazard if there were no reliable behavioral clues to deceit. Then lie catchers would have nothing to go on. And, there would not be a Brokaw hazard if behavioral clues were perfectly reliable for all, rather than for most, people. No clue to deceit is reliable for all human beings, but singly and in combination they can help the lie catcher in judging most people. John Dean’s spouse, friends, and co-workers would know whether he is like most people in showing emotion in his voice, or is unusually able to control his voice. Judge Sirica, having no prior acquaintance with Dean, was vulnerable to the Brokaw hazard.

Dean’s flat-voiced testimony provides another lesson. A lie catcher must always consider the possibility that a suspect might be an unusually gifted performer, so able to disguise his behavior that it is not possible to know whether or not he is lying. According to his own account, John Dean was such a gifted performer. He seemed to know in advance just how Judge Sirica and others would interpret his behavior. He reports the following thoughts as he planned how he would act when he testified: “It would be easy to overdramatize, or to seem too flip about my testimony.... I would, I decided, read evenly, unemotionally, as coldly as possibly, and answer questions the same way. . . . People tend to think that somebody telling the truth will be calm about it.”¹⁵ After he finished his testimony and cross examination began, Dean said he became quite emotional. “I knew I was choking up, feeling alone and impotent in the face of the President’s power. I took a deep breath to make it look as if I were thinking; I was fighting for control You cannot show emotion I told myself. The press will jump all over it as a sign of unmanly weakness”.¹⁶ The fact that Dean’s performance was contrived, that he was so talented in controlling his behavior, does not necessarily mean that he was a liar, only that others should have been wary of interpreting his behavior. In fact, the subsequent evidence suggests that Dean’s testimony was largely true, and that Nixon, who, unlike Dean, is not a very talented performer, was lying.

The last topic to consider before leaving the voice is the claim that there are machines that can automatically and accurately detect lies from the voice. These include the Psychological Stress Evaluator (PSE), the Mark II Voice Analyzer, the Voice Stress Analyzer, the Psychological Stress Analyzer (PSA), the Hagoth, and the Voice Stress Monitor. The manufacturers of these devices claim that they can detect a lie from the voice, even over the telephone. Of course, as their names suggest, they are detecting stress, not lying. There is no voice sign of lying per se, only of negative emotions. The manufacturers of these rather expensive gadgets have not been too forthright in cautioning the user about missing liars who feel no negative emotions and misjudging innocent people who are upset. Scientists specializing in the study of voice and those who specialize in the use of other techniques for detecting lies have found that these machines do no better than chance in detecting lies, and not even very well at the easier task of telling whether or not someone is upset.¹⁷ That does not seem to have affected sales. The possibility of a sure-fire, unobtrusive way to detect lies is too intriguing.

The Body

I learned one way body movements leak concealed feelings in an experiment done during my student days more than twenty-five years ago. There was not much scientific evidence then as to whether body movements accurately reflect emotions or personality. A few psychotherapists thought so, but their claims were dismissed as unsubstantiated anecdotes by the behaviorists, who dominated academic psychology at the time. Many studies from 1914 to 1954 had failed to find support for the claim that nonverbal behavior provides accurate information about emotion and personality. Academic psychology took some pride in how scientific experiments had exposed as a myth the layman’s belief that he could read emotion or personality from the face or body. Those few social scientists or therapists who continued to write about body movement were regarded, like those who were interested in ESP and graphology, as naive, tender-minded, or charlatans.

I could not believe this was so. Watching body movement during group therapy sessions, I was convinced I could tell who was upset about what. With all the optimism of a first-year graduate student I set out to make academic psychology change its view of nonverbal behavior. I devised an experiment to prove that body movements change when someone is under stress. The source of the stress was my senior professor, who agreed to follow a plan I devised in questioning my fellow students about matters on which I knew we all felt vulnerable. While the hidden camera recorded their behavior, the professor asked these budding psychologists what they planned to do when they finished their training. Those who mentioned research were attacked for hiding in the laboratory and shirking their responsibility to help people suffering from mental illness. Those who planned to give such help by practicing psychotherapy were criticized for wanting only to make money and shirking their responsibility to do the research needed to find a cure for mental illness. He also asked if the student had ever been a patient in psychotherapy. Those who said yes were asked how they hoped to help others if they were sick themselves. If they had not obtained psychotherapy he attacked them for trying to help others without first knowing themselves. It was a no-win situation. To make matters worse, I had instructed the professor to interrupt, never letting the student complete a reply to one of his barbs.

The students had volunteered for this miserable experience to help me, their fellow student. They knew it was a research interview, and that stress would be Involved, but that did not make it any easier for them once it began. Outside of the experiment this professor, who was now acting so unreasonably, had enormous power over them. His evaluations were crucial for their graduation, and the enthusiasm of his recommendations determined what job they might get. Within a few minutes the students floundered. Unable to leave or to defend themselves, seething with frustrated anger, they were reduced to silence or inarticulate groans. Before five minutes went by I instructed the professor to end their misery by explaining what he had been doing and why, praising the student for taking the stress so well.

I watched through a one-way mirror and operated a camera to record permanently the body movements. I could not believe what I saw in the very first interview. After the third attack, the student was giving the professor the finger! She kept her hand in that position about one full minute. And yet she didn’t look mad, and the professor was acting as if he didn’t see it. I rushed in when the interview was over. Both of them claimed I had made it up. She admitted she had been angry but denied expressing it. The professor agreed that I must have imagined it because, he said, he would not miss an obscene gesture. When the film was developed my proof was there. This gestural slip, the finger, was not expressing an unconscious feeling. She knew she was mad, but the expression of those feelings was not conscious. She did not know she was giving him the finger. The feelings she was deliberately trying to conceal had leaked.

Fifteen years later I saw the same type of nonverbal leakage, another gestural slip, in the experiment in which nursing students tried to conceal their reactions to the gory medical films. It was not the finger gesture that slipped this time, but a shrug. Nursing student after nursing student gave away her lie by a slight shrug when the interviewer asked “Do you want to see more?” or “Would you show this film to a young child?”

Figure 1

The shrug and the finger are two examples of actions that are called emblems, to distinguish them from all of the other gestures that people show. Emblems have a very precise meaning, known to everyone within a cultural group. Everyone knows that the finger means “fuck you” or “up yours” and that the shrug means “I don’t know,” “I’m helpless,” or “What does it matter?” Most other gestures don’t have such a precise definition, and their meaning is vague. Without words most gestures don’t mean much. Not so for emblems—they can be used in place of a word, or when words can’t be used. There are about sixty emblems in common usage in the United States today. (There are different emblem vocabularies for each country and, often, for regional groups within a country.) Examples of other well-known emblems are the head-nod yes, head-shake no, come-here beckon, wave hello/goodbye, finger-on-finger shame on you, hand-to-ear louder request, hitchhiker’s thumb, and so on.¹⁸

Emblems are almost always performed deliberately. The person who makes an emblem knows what she is doing. She has chosen to state a message. But there are exceptions. Just as there are slips of the tongue, there are slips in body movement—emblems that leak information the person is trying to conceal. There are two ways to tell that an emblem is a slip, revealing concealed information, and not a deliberate message. One is when only a fragment of the emblem is performed, not the entire action. The shrug can be performed by raising both shoulders, or by turning the palms up, or by a facial movement that involves raising the brows and drooping the upper eyelid and making a horseshoe-shaped mouth, or by combining all of these actions and, sometimes, throwing in a sideways head tilt. When an emblem is leakage, only one element will be shown, and even it won’t be complete. Only one shoulder may be raised, and not very high; or only the lower lip may be pushed up; or the palms may be turned up only slightly. The ringer emblem not only involves a particular arrangement of the five fingers, but the hand is thrust forward and upward, often repeatedly. When the finger emblem was not performed deliberately but leaked a student’s stifled fury, the movement component was not there, only the arrangement of the fingers.

The second tip-off that the emblem is a slip rather than a deliberate action is that it is performed out of the usual presentation position. Most emblems are performed right out in front of the person, between the waist and the neck area. An emblem can’t be missed when it is in the presentation position. A leakage emblem is never performed in the presentation position. In the stress interview when the student gave the professor the finger, it was not shoved out in space but instead was lying on the student’s knee, out of the presentation position. In the experiment with the nursing students, the shrugs that leaked their feelings of helplessness and inability to conceal their feelings were small rotations of the hands, while the hands stayed in the lap. If the emblem was not fragmented and out of the presentation position, the liar would realize what was happening and would censor the emblem. Of course, these characteristics that distinguish the leakage emblem—fragmentation and out-of-presentation position—also make it hard for others to notice. A liar can show these leakage emblems again and again, and usually neither the liar nor her victim will notice them.

There is no guarantee that every liar will make an emblematic slip. There are no such sure-fire signs of deceit. There has been too little research to yet estimate how often emblematic slips will occur when people lie. Subjected to the hostile professor, two of the five students showed an emblematic slip. A little more than half of the nursing students showed an emblematic slip when they were lying. I don’t know why some people had this form of leakage while others did not.^*

While not every liar shows an emblematic slip, when emblematic slips occur they are quite reliable. The emblematic slip can be trusted as a genuine sign of a message that the person does not want to reveal. Their interpretation is less vulnerable than most other signs of deceit to either the Brokaw hazard or the Othello error. Some people always talk in a circumlocutious fashion, but few people make emblematic slips regularly. Speech errors may signify stress of many kinds, not necessarily just the stresses involved in lying. Because the emblem has a very specific message, much like words, emblematic slips are usually not so ambiguous. If the person slips the message “fuck you” or “I’m mad” or “I don’t mean it” or “over there”—all of which can be shown by an emblem—there shouldn’t be much of a problem in interpreting what is meant.

What emblem will slip during a lie, which message will leak out, will depend upon what is being concealed. The students in my hostile professor experiment were concealing anger and outrage, so the emblematic slips were the finger and a fist. In the medical training film experiment the nursing students were not feeling angry, but many felt they were not adequately concealing their feelings. The helpless shrug was the emblematic slip. No adult needs to be taught the vocabulary of emblems. Everyone knows the emblems shown by members of their own culture. What many people do need to learn is that emblems may occur as slips. Unless lie catchers are alert to this possibility, they won’t spot the emblematic slips that will escape their notice, because they are fragmented and out of the presentation position.

Illustrators are another type of body movement that can provide deception clues. Illustrators are often confused with emblems, but it is important to distinguish between them, for these two kinds of body movements may change in opposite ways when people lie. While emblematic slips may increase, illustrators usually will decrease.

Illustrators are called by that name because they illustrate speech as it is spoken. There are many ways to do so: emphasis can be given to a word or phrase, much like an accent mark or underlining; the flow of thought can be traced in the air, as if the speaker is conducting her speech; the hands can draw a picture in space or show an action repeating or amplifying what is being said. It is the hands that usually illustrate speech, although brow and upper eyelid movements often provide emphasis illustrators, and the entire body or upper trunk can do so also.

Social attitudes toward the propriety of illustrators have gone back and forth over the last few centuries. There have been times when illustrating was the mark of the upper classes, and also times when they have been considered the mark of the uncouth. Books on oratory have usually depicted the illustrators required for successful public speaking.

The pioneering scientific study of illustrators was not undertaken to uncover clues to deceit but to challenge the claims of the Nazi social scientists. The results of that study can help the lie catcher avoid mistakes due to a failure to recognize national differences in illustrators. During the 1930s, many articles appeared that claimed illustrators were inborn and that the “inferior races,” such as the Jews or gypsies, made many large, sweeping illustrators compared to the “superior,” less gesturally expansive Aryans. No mention was made of the grand illustrators shown by Germany’s Italian ally! David Efron,¹⁹ an Argentinian Jew studying at Columbia University with the anthropologist Franz Boas, examined the illustrators of people living on the Lower East Side of New York City. He found that immigrants from Sicily used illustrators that draw a picture or show an action, while Jewish Lithuanian immigrants used illustrators that give emphasis or trace the flow of thought. Their offspring born in the United States who attended integrated schools did not differ from one another in the use of illustrators. Those of Sicilian parentage used illustrators similar to those used by children of Jewish Lithuanian parents.

The style of illustrators is acquired, Efron showed, not inborn. People from different cultures not only use different types of illustrators, but some illustrate very little while others illustrate a lot. Even within a culture, individuals differ in how many illustrators they typically show.^* It is not the sheer number of illustrators or their type, then, that can betray a lie. The clue to deceit comes from noting a decrease in the number of illustrators shown, when a person illustrates less than usual. More needs to be explained about when people do illustrate, to avoid misinterpreting why someone shows a decrease.

First consider why people illustrate at all. Illustrators are used to help explain ideas that are difficult to put into words. We found that people were more likely to illustrate when asked to define zigzag than chair, more likely to illustrate when explaining how to get to the post office than when explaining their occupational choice. Illustrators also are used when a person can’t find a word. Snapping the fingers or reaching in the air seems to help the person find the word, as if the word floats above the person captured by the illustrator movement. Such word-search illustrators at least let the other person know that a search is under way and that the first person hasn’t given up his turn to speak. Illustrators may have a self-priming function, helping people put words together into reasonably coherent speech. Illustrators increase with involvement with what is being said. People tend to illustrate more than usual when they are furious, horrified, very agitated, distressed, or excitedly enthused.

Now consider why people show less than their usual level of illustrating, for this will make clear when such decreases can be a clue to deceit. The first reason is a lack of emotional investment in what is being said. People illustrate less than usual when they are uninvolved, bored, disinterested, or deeply saddened. People who feign concern or enthusiasm can be betrayed by the failure to accompany their speech with increased illustrators.

Illustrators also decrease when a person is having trouble deciding exactly what to say. If someone weighs each word carefully, considering what is said before it is said, there is not much illustrating. When giving a talk for the first or second time, whether it be a lecture or sales pitch, there will not be as many illustrators as there later will be when not much effort has to be spent finding words. Illustrators decrease whenever there is caution about speech. It may have nothing to do with deceit. There may be caution because the stakes are high: the first impression made on a boss, the answer to a question that could bring a prize, the first words to a person passionately admired previously from a distance. Ambivalence also makes for caution about what to say. A timorous person may be terribly tempted by a much more lucrative job offer but be afraid to take the risks involved in a new work situation. Torn by whether he should or shouldn’t, he is afflicted with the ponderous problem of what to say and how to say it.

If a liar has not adequately worked out her line in advance she also will have to be cautious, carefully considering each word before it is spoken. Deceivers who are not rehearsed, who have had little practice in the particular lie, who failed to anticipate what would be asked or when, will show a decrease in illustrators. Even if the liar has worked out and practiced her line, her illustrators may decrease because of the interference of her emotions. Some emotions, especially fear, can interfere with speaking coherently. The burden of managing almost any strong emotion distracts from the processes involved in stringing words together. If the emotion has to be concealed, not just managed, and if it is a strong emotion, then it is likely that even the liar with a well-prepared line may have trouble speaking it, and illustrators will decrease.

The student nurses in our experiment illustrated less when they were trying to conceal their reactions to the amputation-burn film than when they honestly described their feelings about the flower film. This decrease in illustrators occurred for at least two reasons: the students were not practiced in making the required lie and had been given no time to prepare their line, and strong emotions were aroused, both detection apprehension and emotions in response to the gory film they were watching. Many other investigators have also found illustrators less apparent when someone is lying as compared to when someone is telling the truth. In these studies little emotion was involved, but the liars were ill prepared.

In introducing illustrators I said that it is important to distinguish them from emblems, for opposite changes may occur in each when someone lies: emblematic slips increase and illustrators decrease. The crucial differences between emblems and illustrators are in the precision of movement and message. For the emblem both are highly prescribed: not any movement will do; only a highly defined movement conveys the quite precise message. Illustrators, by contrast, can involve a wide variety of movements and may convey a vague rather than a precise message. Consider the thumb-to-first-finger A-OK emblem. There is only one way to do it. If the thumb goes to the middle finger or pinky it would not be very clear. And, the meaning is very specific—”OK,” “that’s good,” “all right.”^* Illustrators don’t have much meaning independent of the words. Watching someone illustrate without hearing the words doesn’t reveal much about the conversation; that is not so if the person makes an emblem. Another difference between emblems and illustrators is that although both are shown when people converse, emblems can be used in place of a word or when people cannot or do not speak. Illustrator movements, by definition, occur only during speech, not to replace it or when people don’t talk.

The lie catcher must be more cautious in interpreting illustrators than emblematic slips. As described earlier, both the Othello error and the Brokaw hazard influence illustrators but not emblematic slips. If the lie catcher notes a decrease in illustrating, he must rule out all the other reasons (apart from lying) for someone wanting to carefully choose each word. There is less ambiguity about the emblematic slip; the message conveyed is usually sufficiently distinct to make it easier for the lie catcher to interpret. And, the lie catcher does not need previous acquaintance with the suspect to interpret an emblematic slip. Such an action, in and of itself, has meaning. Since individuals differ enormously in their usual rate of illustrating, no judgment can be made about them unless the lie catcher has some basis for comparison. Interpreting illustrators, like most of the other clues to deceit, requires previous acquaintance. Spotting deceit is very difficult in first meetings. Emblematic slips offer one of the few possibilities.

The reason for explaining the next type of body movement, manipulators, is to warn the lie catcher about the risk of interpreting them as signs of deceit. We have found that lie catchers often mistakenly judge a truthful person to be lying because they show many manipulators. While manipulators can be a sign that someone is upset, they are not always so. An increase in manipulator activity is not a reliable sign of deceit, but people think it is.

Manipulators include all those movements in which one pitt of the body grooms, massages, rubs, holds, pinches, picks, scratches, or otherwise manipulates another body part. Manipulators may be of very short duration or they may go on for many minutes. Some of the brief ones appear to have a purpose: the hair is rearranged, matter is removed from the ear canal, a part of the body is scratched. Other manipulators, particularly those that last a long time, seem to be purposeless: hair is twisted and untwisted, fingers rubbed, a foot tapped. Typically the hand is the manipulator. The hand may also be the recipient, as can any other part of the body. Common recipients are the hair, ears, nose, or crotch. Manipulator actions also can be performed within the face—tongue against cheeks, or teeth slightly biting lips—and by leg against leg. Props may become part of a manipulator act—match, pencil, paper clip, or cigarette.

While most people were brought up not to perform these bathroom behaviors in public, they haven’t learned to stop doing them, only to stop noticing that they do them. It is not that people are completely unconscious of their manipulators. If we realize someone is looking at one of our manipulator acts, we will quickly interrupt, diminish, or disguise it. A larger gesture often will deftly cover a fleeting one. Even this elaborate strategy to conceal a manipulator is not done with much awareness. Manipulators are on the edge of consciousness. Most people cannot stop doing them for very long even when they try deliberately to do so. People are accustomed to manipulating themselves.

People are much more proper as observers than as performers. The person making a manipulator movement is given the privacy to complete this act, even when the manipulator begins right in the midst of a conversation. Others look away when a manipulator is performed, looking back only when it is over. If the manipulator is one of those seemingly pointless activities, like hair twisting, which goes on and on, then of course others don’t look away forever, but people won’t look for long directly at the manipulator act. Such polite inattention to manipulators is an overlearned habit, operating without thought. It is the manipulator watcher rather than the performer who, like a Peeping Tom, creates the offense to manners. When two cars pull up at a stop sign, it is the person who glances over at the person in the adjacent car who commits the offense, not the person who is vigorously cleaning his ear.

I and others studying manipulators have wondered why people engage in one manipulator rather than another. Does it mean anything if it is a rub rather than a squeeze, a pick rather than a scratch? And, is there some message that can be read from whether it is the hand, ear, or nose that is scratched? Part of the answer is idiosyncrasy. People have their favorites, a particular type of manipulator that is their hallmark. For one person it may be twisting a ring, for another picking cuticles, and for another twisting a mustache. No one has tried to discover why people have one versus another favorite manipulator, or why some people have no special idiosyncratic manipulator. There is a bit of evidence to suggest that certain manipulator actions reveal more than just discomfort. We found picking manipulators in psychiatric patients who were not expressing anger. Covering the eyes was common among patients who felt shame. But this evidence is tentative, compared to the more general finding that manipulators increase with discomfort.²⁰

Scientists have reasonably well substantiated the layman’s belief that people fidget, make restless movements, when they are ill at ease or nervous. Body scratching, squeezing, picking, and orifice cleaning and grooming manipulators increase with any type of discomfort. I believe that people also show many manipulators when they are quite relaxed and at ease, letting their hair down. When with their chums, people don’t worry as much about being proper. Some people will be more likely than others to burp, manipulate, and indulge in behaviors that in most situations are at least partially managed. If this is correct, then manipulators are discomfort signs only in more formal situations, with people who are not so familiar.

Manipulators are unreliable as signs of deceit because they may indicate opposite states—discomfort and relaxation. Also, liars know they should try to squelch their manipulators, and most will succeed part of the time. Liars do not have any special knowledge of this; it is part of the general folklore that manipulators are discomfort signs, nervous behavior. Everyone thinks that liars will fidget, that restlessness is a deception clue. When we asked people how they would tell if someone were lying, squirming and shifty eyes were the winners. Clues that everyone knows about, that involve behavior that can be readily inhibited, won’t be very reliable if the stakes are high and the liar does not want to be caught.

The student nurses did not show more manipulator actions when lying than when telling the truth. Other studies have found an increase in manipulators during deceits. I believe it is differences in the stakes that account for this contradiction in findings. When the stakes are high the manipulator actions may be intermittent, for contrary forces may be at work. High stakes make the liar monitor and control accessible and known clues to deceit, such as manipulators, but those high stakes may make the liar afraid of being caught, and that discomfort should increase this behavior. Manipulators may increase, be monitored, squelched, disappear for a time, reappear, and then after a time again be noticed and suppressed. Since the stakes were high, the nursing students worked hard to control their manipulator actions. There was not much at stake in studies that found manipulators increased during lying. The situation was a bit strange—being asked to lie in an experiment is unusual—and so there could have been enough discomfort to increase manipulator actions. But there were no important gains or losses for success or failure in these deceits, little reason for the liar to expend the effort to monitor and suppress manipulator actions. Even if my explanation of why contradictory results were obtained is incorrect (and such after-the-fact interpretations must be viewed as tentative until confirmed by further studies), the contradictory findings themselves are sufficient reason for the lie catcher to be cautious about interpreting manipulators.

In our study of how well people can catch lies, we found that people judged those who showed many manipulators as liars. It didn’t matter whether the person showing the manipulator was actually telling the truth or lying; those who saw them labeled them as dishonest if they showed many. It is important to recognize the likelihood of making this error. Let me review the multiple reasons why manipulators are an unreliable sign of deceit.

People vary enormously in how many manipulators and what kinds of manipulators they usually show. This individual difference problem (the Brokaw hazard) can be countered if the lie catcher has some previous acquaintance and can make behavioral comparisons.

The Othello error also interferes with the interpretation of manipulators as deception clues, since manipulators increase when people are uncomfortable about anything. This is a problem with other signs of deceit also, but it is especially acute with manipulators, since they are not just discomfort signs but sometimes, with buddies, comfort signs.

Everyone believes that showing many manipulators betrays deceit, and so a motivated liar will try to squelch them. Unlike facial expression, which people also try to control, manipulators are fairly easy to inhibit. Liars will succeed in inhibiting manipulators at least part of the time if the stakes are high.

Another aspect of the body—posture—has been examined by a number of investigators, but little evidence of leakage or deception clues has been found. People know how they are supposed to sit and stand. The posture appropriate for a formal interview is not the posture assumed when talking with a friend. Posture seems well under control and successfully managed when someone is deceiving. I and others studying deceit have found no differences in posture when people lie or tell the truth.^* Of course, we might not have measured that aspect of posture that does change. A possibility is the tendency to move forward with interest or anger and backward with fear or disgust. A motivated liar should, however, be able to inhibit all but the most subtle signs of postural clues to those emotions.

Autonomic Nervous System Clues

So far I have discussed bodily actions produced by the skeletal muscles. The autonomic nervous system (ANS) also produces some noticeable changes in the body with emotional arousal: in the pattern of breathing; in the frequency of swallowing; and in the amount of sweating. (ANS changes registered in the face, such as blushing, blanching, and pupil dilation, are discussed in the next chapter.) These changes occur involuntarily when emotion is aroused, are very hard to inhibit, and for that reason can be very reliable clues to deceit.

The polygraph lie detector measures these ANS changes, but many of them will be visible without the use of a special apparatus. If a liar feels afraid, angry, excited, distressed, guilty, or ashamed there may well be rapid breathing, heaving chest, frequent swallowing, or the smell or appearance of sweating. For decades psychologists have disagreed about whether or not each emotion has a distinctive set of these ANS changes. Most psychologists think there is not; they believe that one breathes more rapidly, sweats, and swallows when any emotion becomes aroused. ANS changes mark how strong an emotion is, not which emotion it is. This view contradicts most people’s experience. People feel different bodily sensations when they are afraid, for example, as compared to when they are angry. That, many psychologists say, is because people interpret the same set of bodily sensations differently if they are afraid than if they are angry. It is not proof that the ANS activity itself actually differs for fear versus anger.²¹

My most recent research, begun when I had almost finished writing this book, challenges this view. If I am correct, and ANS changes are not the same for, but instead are specific to, each emotion, this could be quite important in detecting lies. It would mean that the lie catcher could discover, either with a polygraph and even to some extent just by watching and listening, not just whether a suspect is emotionally aroused but which emotion is felt—is the suspect afraid or angry, disgusted or sad? While such information is available from the face as well, as the next chapter explains, people are able to inhibit many of the facial signs. ANS activity is much harder to censor.

We have published only one study as of now (see page 117), and some eminent psychologists disagree with what we have found. My findings are considered controversial, not established, but our evidence is strong and in time I believe will be accepted by the scientific community.

Two problems had stood in the way, I thought, of discovering convincing evidence that emotions have different ANS activity, and I thought I had solutions to both. One problem is how to obtain pure samples of emotion. To contrast the ANS changes in fear with those in anger the scientist must be certain when his research subject experiences each emotion. Since the measurement of ANS changes requires elaborate equipment, the subject must provide the emotion samples in a laboratory. The problem is how to elicit emotions in a sterile, unnatural setting. How do you make people afraid and angry, and not both at the same time? That last question is very important—not making them afraid and angry at the same time in what I and others call an emotion blend. Unless the emotions are kept separate—unless the samples are pure—there would be no way to determine if the ANS activity differs for each emotion. Even if it did, if the anger samples always included some fear, and the fear sample some anger, the result would be that the ANS changes would appear to be the same. It is not easy to avoid blends, in the laboratory or in real life. Blends are more common than pure emotions.

The most popular technique for sampling emotions has been to ask the subject to remember or imagine something fearful. Let’s suppose the subject imagines being attacked by a mugger. The scientist must be certain that in addition to fear the subject doesn’t become a bit angry at the mugger, or angry at himself for having been made afraid or having been stupid enough to put himself in jeopardy. The same hazards of blends occurring rather than pure emotions happens with other techniques for arousing emotions. Suppose the scientist shows a fear-inducing movie, perhaps a scene from a horror movie like Alfred Hitchcock’s Psycho, in which Tony Perkins suddenly attacks Janet Leigh with a knife while she is taking a shower. The subject might become angry at the scientist for making him afraid, angry at himself for getting afraid, angry at Tony Perkins for attacking Janet Leigh, disgusted by the blood, distressed by Janet Leigh’s suffering, surprised by the action, and so on. It is not so easy to think of a way to obtain pure emotion samples. Most scientists who have studied the ANS simply assumed, I think incorrectly, that subjects did what they wanted when they wanted them to, producing with ease the desired pure emotion samples. They failed to take any steps to guarantee or verify that their emotion samples really were pure.

The second problem is produced by the need to sample emotions in a laboratory, and results from the impact of the research technology. Most research subjects are self-conscious about what is going to happen to them when they come in the door. Then it gets amplified. To measure ANS activity, wires have to be attached to different parts of the subject’s body. Just to monitor respiration, heart rate, skin temperature, and sweating requires attaching many of these wires. Sitting there hooked up, having scientists scrutinize what is going on inside their body, and often having cameras record any visible changes, embarrasses most people. Embarrassment is an emotion, and if it produces ANS activity, those ANS changes would be smeared across every emotion sample the scientist is trying to obtain. He may think the subject is remembering a fearful event at one moment, and an angry memory at another point, but what actually may happen is embarrassment during both memories. No scientist took steps to reduce embarrassment, none checked to be certain that embarrassment did not spoil their pure emotion samples.

My colleagues and I eliminated embarrassment by selecting professional actors as our research subjects.²² Actors are accustomed to being scrutinized, and they don’t get upset when people watch their every move. Rather than being embarrassed, they liked the idea that we would attach wires to them and monitor the insides of their body. Studying actors also helped in solving the first problem—obtaining pure emotion samples. We could make use of the actors’ years of training in the Stanislavski acting technique, which makes them skilled in remembering and re-experiencing emotions. Actors practice this technique so that they can use sense-memories in portraying a particular role. In our experiment we asked the actors, while they were hooked up, with video cameras trained on their faces, to remember and re-experience as strongly as they could a time when they had felt the most anger in their lives, and then a time when they had felt the most fear, sadness, surprise, happiness, and disgust. Other scientists have used this technique before, but we thought we had a better chance of success because we were using professionals, trained in the technique, who wouldn’t be embarrassed. Furthermore, we didn’t just take for granted that our subjects did what we asked; we verified that we had obtained pure samples, not blends. After each memory retrieval, we asked the actors to rate how strongly they had felt the requested emotion, and whether they had felt any other emotion. Any attempts in which they reported re-experiencing any other emotion nearly as strongly as the one requested was not kept in our sample.

Studying actors also made it easier for us to try a second technique for sampling pure emotions, one that had never before been used. We discovered this new technique for arousing emotions by accident, years earlier, when doing another study. To learn the mechanics of facial expressions—which muscles produce which expression—my colleagues and I systematically made thousands of facial expressions, filming and then analyzing how each combination of muscle movements change appearance. To our surprise, when we did the muscle actions that relate to emotions, we would suddenly feel changes in our bodies, changes due to ANS activity. We had no reason to expect that deliberately moving facial muscles could produce involuntary ANS changes, but it happened again and again. We still did not know whether or not the ANS activity differed with each set of facial muscle movements. We told the actors exactly which facial muscles to move. There were six different instructions, one for each of six emotions. Not embarrassed by making facial expressions on demand or by being watched when they did it, and skilled in facial expression, they met most of our requests with ease. Again, we didn’t just trust them to produce pure samples of emotion. We videotaped their facial performances and only used their attempts if measurements of the videotapes showed they had produced each set of requested facial actions.

Our experiment found strong evidence that ANS activity is not the same for all emotions. The changes in heart rate, skin temperature, and sweating (which is all we had measured) are not the same for every emotion. For example, when the actors made the muscle movements on their face for anger and those for fear (and remember, they weren’t told to pose these emotions but instead just to make specific muscle actions), their heart beat faster, but different things happened to their skin temperature. Their skin became hot with anger and cold with fear. We have just repeated our experiment with different subjects and obtained the same results.

If these results hold up when other scientists try to repeat them in their laboratories, they could change what the lie detector tries to learn from the polygraph. Instead of just trying to know whether a suspect is feeling any emotion, the polygraph operator could tell by measuring a number of ANS activities which emotion. Even without a polygraph machine, just by looking a lie catcher may be able to notice changes in the pattern of breathing or in sweating that could help to spot the occurrence of specific emotions. Errors in catching lies—disbelieving-the-truth-ful and believing-the-liar—could be reduced if ANS activity, which is very hard to inhibit, could reveal which emotion a suspect feels. We don’t yet know whether emotions can be distinguished just by the visible and audible signs of ANS activity, but now there is a reason to find out. How signs of specific emotions—whether they be from face, body, voice, words, or ANS—can help to determine whether someone is lying or truthful, and the hazards of making mistakes and precautions to avoid making them, is the topic of chapter 6.

Chapter 2 explained that there are two principal ways to lie—to conceal or to falsify. So far this chapter has considered how attempts to conceal feelings may be betrayed by the words, voice, or body. A liar may falsify when no emotion is felt but one is required, or to help cover a concealed feeling. For example, a fellow may falsify a look of sadness when he learns that his brother-in-law’s business has failed. If he is totally unmoved, the false expression simply shows the proper countenance, but if he was secretly delighted by his brother-in-law’s misfortune, the false look of sadness would also mask his true feelings. Can the words, voice, or body betray such false expressions, revealing that an emotional performance is not felt? No one knows. Defects in false performances of emotion have been investigated less thoroughly than leakage of concealed emotions. I can only give my observations, theories, and hunches.

While words are made for fabricating, it is not easy for anyone, truthful or not, to describe emotions in words. Only a poet conveys the nuances revealed by an expression. It may be no more difficult to claim in words a feeling not felt than one that is. Usually neither will be very eloquent, elaborate, or convincing. It is the voice, the body, the facial expression that give meaning to the verbal account of an emotion. I suspect that most people can put on the voice of anger, fear, distress, happiness, disgust, or surprise well enough to fool others. While it is very hard to conceal the changes in the sound of the voice that occur with these emotions, it is not so hard to falsify them. Most people probably are fooled by the voice.

Some of the changes produced by the autonomic nervous system are easy to falsify. While it is hard to conceal the signs of emotion in respiration or swallowing, it takes no special skill to falsify them, breathing more quickly or swallowing often. Sweating is a different matter, hard to conceal and hard to falsify. While a liar could use respiration and swallowing to falsely give the impression of negative emotions, I expect few do.

While a deceiver could increase manipulators in order to appear uncomfortable, most people probably won’t remember to do so. The failure to include these actions, which could be easily performed, might by their absence betray an otherwise convincing claim to be feeling fear or distress.

Illustrators can be put on but probably not very successfully, to create the impression of involvement and enthusiasm for what is being said when none is felt. Newspaper accounts said that former presidents Nixon and Ford were both coached to increase their illustrators. Watching them on TV, I thought the coaching often made them look phony. It is hard to deliberately place an illustrator exactly where it should be in relation to the words; they tend to come in too early or late or stay on too long. It is much like trying to ski by thinking about each action as you do it; the coordination is rough and looks it.

I have described behavioral clues that may leak concealed information, indicate that the person has not prepared his line, or betray an emotion that does not fit the line being taken.

Slips of the tongue, emblematic slips, and tirades can leak concealed information of any kind—emotions, past deeds, plans, intentions, fantasies, ideas, etc.

Indirect speech, pauses, speech errors, and a decrease in illustrators may indicate that the speaker is being very careful about what is said, not having prepared the line being taken. They are signs of any negative emotion. A decrease in illustrators also occurs with boredom.

Raised voice pitch and louder, faster speech occur with fear, anger, and perhaps excitement. The voice changes in the opposite way with sadness and perhaps with guilt.

Changes in breathing or sweating, increased swallowing, and a very dry mouth are signs of strong emotions, and it may be possible in the future to determine which emotion from the pattern of these changes.

*Most people, when they talk, are dependent upon these listener responses and if deprived will quickly ask, “Are you listening?” There are a few people who are closed systems, talking heedless of whether their listeners provide any encouragement responses.

†Among sawmill workers, for example, who must communicate but can’t do so with words because of the noise, a very elaborate system of hand gestures is used. Pilots and landing crews for the same reason use an elaborate system of gestures.

*Neuroscientists are not certain about the circuitry that provides us with information about changes in our own expression or about whether it is changes in muscle or in the skin that are registered. Psychologists disagree about how well people can feel their own facial expressions as they emerge. My studies suggest that we don’t feel the expressions we make very well and that most of the time we don’t pay much attention to the sensations in our face.

*Many psychologists have attempted to identify what it is that makes someone a good or bad judge of people. Not much progress has been made. For a review of this research, see Maureen O’Sullivan, “Measuring the Ability to Recognize Facial Expressions of Emotion,” in Emotion in the Human Face, ed. Paul Ekman (New York: Cambridge University Press, 1982).

*It is hard to know what to make of this and other contradictions in the research literature on deceit, since the experiments are not themselves too trustworthy. Almost all have examined students, who lied about trivial matters, with little at stake. Most of the experiments on lying have shown little thought about just what type of lie they might be examining. Usually the lie studied is one selected because it is easy to arrange in a laboratory. For example, students have been asked to argue convincingly an opinion about capital punishment or abortion contrary to their own. Or, students were asked to say whether they would like or dislike a person shown to them in a photograph and then were asked to pretend that they have the opposite attitude. Typically these experiments fail to consider the liar’s relationship to the target, and how this might influence how hard the liar tries to succeed. Usually the liar and target were not acquainted and had no reason to think they would ever meet each other again. Sometimes there was no actual target, but instead the liar spoke in a misleading fashion to a machine. For a recent, but not sufficiently critical, review of these experiments, see Miron Zuckerman, Bella M. DePaulo, and Robert Rosenthal, “Verbal and Nonverbal Communication of Deception,” in Advances in Experimental Social Psychology, vol. 14 (New York: Academic Press, 1981).

*Unfortunately, none of the other investigators who have studied deceit have checked to see if they could replicate our finding on emblematic slips, I feel optimistic that they would, having twice over a twenty-five-year period found leakage through emblematic slips.

*Immigrant families from cultures that make frequent use of illustrators often train their children not to speak with their hands. Their children are cautioned that if they illustrate they will look like they are from the old country. Not illustrating will make them resemble the northern European, older American stock.

*This emblem has a quite different, obscene meaning in some southern European countries. Emblems are not universal. Their meaning varies with culture.

*One study of deceit found that people believe that those who shift their posture frequently are lying. In fact, though, posture proved to be unrelated to truthfulness. See Robert E. Kraut and Donald Poe, “Behavioral Roots of Person Perception: The Deception Judgments of Custom Inspectors and Laymen,” Journal of Personality and Social Psychology 39 (1980): 784–98.