THE APPLICATION OF NON-VERBAL COMMUNICATION METHODS FOR LIE DETECTION
According to Trovillo (Reference Trovillo1939), the interest in lie detection methods based on non-verbal communication is very antique. According to the author, the first written document about this topic is the Vedas, an ancient Indian papyrus dated 900 BC. This document includes specific behavioral instructions for detecting poisoners, an important and transversal problem at that time:
“A person who gives poison may be recognized. He does not answer questions, or they give evasive answers; he speaks nonsense, rubs the great toe along the ground, and shivers; his face is discolored; he rubs the roots of the hair with his fingers, and he tries by every means to leave the house…”
This first description of deceptive behavior already included areas of investigation for future scientific research, including non-verbal indicators (i.e. “he rubs the roots of the hair,” “he rubs the great toe along the ground”), verbal indicators (i.e. “he speaks nonsense,” “he does not answer questions”) and neuro-vegetative indicators (i.e. “his face is discolored”). We can easily hypothesize the early interest of man towards lie detection through behavioral observation. According to Sommer (Reference Sommer1999:11), lie detection skills played an evolutionary role for the human species: “lies and deception are part of the fight for the survival of species.”
Despite practitioners’ transversal interest in lie detection methods through observational methods, recent scientific research has shown weak correlations between lies and non-verbal cues (DePaulo et al. Reference DePaulo, Lindsay, Malone, Muhlenbruck, Charlton and Cooper2003; Vrij, Granhag, and Porter Reference Vrij, Granhag and Porter2010; Vrij, Hartwig, and Granhag Reference Vrij, Hartwig and Granhag2019). Moreover, research has shown that investigators, police officers, and experts believe that non-verbal communication can contribute to lie detection and because police manuals often recommend the analysis of non-verbal behavior (Boogard et al. Reference Boogard, Meijer, Vrij and Merckelbach2016; Vrij and Semin Reference Vrij and Semin1996).
A NEW FRAMEWORK FOR THE APPLICATION OF NON-VERBAL COMMUNICATIONS IN CRIMINAL INVESTIGATIONS. WHY SCIENTIFIC LITERATURE FAILS IN IDENTIFYING A CORRELATION BETWEEN NON-VERBAL COMMUNICATION AND LIES
This paragraph will summarize the basic assumptions of a new framework for applying non-verbal communication in lie detection and criminal investigations. The goal is not only to promote non-verbal communication as support for acquiring information during a forensic interview but also to encourage further scientific research in this field in order to help practitioners in the correct understanding of the strengths and weaknesses of this method.
The first important step for promoting the new framework for applying non-verbal communication in criminal investigations is identifying possible weaknesses in scientific research claiming the absence of correlation between lies and non-verbal communication signs.
We can find contradictions between research and practice comparing these aspects:
(1) Most research calculates Pearson’s correlations (r) between non-verbal communication indicators and lies or compares frequencies of non-verbal signs in liars and truth-tellers through the t test.
(2) Manuals of non-verbal communication agree on the impossibility to have a single indicator directly associated with deception (Ekman Reference Ekman1985).
(3) Manuals of non-verbal communication invite practitioners to compare non-verbal communication meaning and verbal statements.
Although many manuals of non-verbal communication include checklists applicable for lie detection purposes, none states that single indicators can be predictive of lies because deception is indicated by the differences in the meaning expressed by words and body language. These contradictions suggest that, probably, research designed for the application of non-verbal communication in this field should consider additional factors, summarized in our framework:
(1) Non-verbal communication indicators have different meanings based on the exact timing and context of facial muscle contractions and body movements.
(2) Lie detection behavioral techniques should be based on a constant qualitative comparison between verbal statements and non-verbal communication.
(3) A sequence of non-verbal communication signs reveals additional meaning that a single movement’s analysis cannot explain.
These three factors cannot be investigated through a general Pearson’s correlation between non-verbal communication signs and lies or by t test comparisons between liars and truth-tellers. The crucial role of these three factors is explained in the following paragraphs.
Non-verbal Communication Indicators Have Different Meanings Based on the Exact Timing, Context of Facial Muscle Contractions and Body Movements
One of the most underestimated factors in the analysis of non-verbal behavior is timing. The general area investigating the relationship between time and non-verbal communication is called chronemics (Bruneau Reference Bruneau and Ritchie Key1980). The research focused on comparing frequencies of non-verbal behavior in liars and truth-tellers does not consider chronemics, despite research showing that timing is predictive of deliberate facial expressions. Specifically, this research’s object includes several chronemics factors: duration, onset, apex, and offset of facial expression. For example, research by Schmidt et al. (Reference Schmidt, Zara, Cohn and Ian Reed2006) identified that the timing of contractions could discriminate deliberate and spontaneous happy facial expressions. A study by Prkachin (Reference Prkachin1992) identified temporal differences in spontaneous and deliberate expressions of pain. A similar study by Hess and Kleck (Reference Hess and Kleck1990) investigated the expressions of happiness and disgust in the same direction, finding that facial expressions’ temporal features can contribute to the identification of deliberate facial expressions. Identifying deliberate facial expressions is a valid support in understanding the intention of concealing real emotions since deliberate actions are performed to address a communication process’s goal.
Diana et al. (Reference Diana, Zurloni, Elia, Cavalera, Realdon, Johnsson and Anguera2018) completed an exploratory study on the application of temporal pattern (T-pattern) and cognitive load to understand low-stake lies in a sample of 40 students. T-patterns are hierarchical sequences of behaviors in time, identified through an algorithm applied by Theme Software. This algorithm determines if arbitrary events sequentially repeat, within a specific timeframe, at a rate greater than expected by chance. The same method has been used successfully to analyze deception in doping-related interviews (Zurloni et al. Reference Zurloni, Diana, Cavalera, Argenton, Elia and Mantovani2015).
Lie Detection Behavioral Techniques Should be Based on a Constant Qualitative Comparison Between Verbal Statements and Non-verbal Communication
The main role of non-verbal communication analysis during forensic interviews is to give additional information about the interviewee’s intentions and emotions. This additional information can be coherent to the verbal statement, suggesting to practitioners a consistency between words and internal feelings, or can be the opposite of the statement, revealing additional meaning that the person decided not to state in that particular situation.
Theoretically, if the contribution of non-verbal communication indicators is compared with verbal communication, the measurement of general frequencies of non-verbal signs in a generic timeframe, without any information about the verbal content, is not sufficient to understand the contribution of non-verbal behavior for lie detection purposes.
A meta-analysis conducted by DePaulo et al. (Reference DePaulo, Lindsay, Malone, Muhlenbruck, Charlton and Cooper2003) showed a significant weak correlation between deception and facial expression called ‘chin raiser’. This movement is called action unit (AU) 17 according to the Facial Action Coding System (Ekman, Friesen, and Hager Reference Ekman, Friesen and Hager2002a), a standardized method for describing facial actions, and it can be interpreted as a sign of perplexity. Of course, at a general level, perplexity should be correlated with deceptive statements. If the subject is uncertain about his statement, it is reasonable to think that the statement’s accuracy is scarce, and this would happen more frequently during deceptive statements. However, this facial expression’s contribution cannot be understood totally without comparing the exact verbal content. For example, a facial expression of perplexity is congruent to ambiguous verbal content. If a person says: “I am not sure where I left the keys” and contracts AU 17, there is no contradiction with the verbal statement. However, if a person gives a solid verbal statement (i.e. “I left the keys in that room at 5 p.m.”) with a facial expression of perplexity, the same action would suggest a deceptive statement, since there is a contradiction between perplexity expressed through facial expression and the solidity of the statement. Adopting statistics that uniquely consider the frequency of a facial sign and ignoring the verbal content is not useful to clarify if non-verbal communication can contribute to forensic interviews.
A Sequence of Non-verbal Communication Signs Reveals Additional Meaning That a Single Movement’s Analysis Cannot Explain
The analysis of single facial movements concerning lies ignores another important factor: the sequence of facial signs and other actions has an additional meaning that the analysis of single movements cannot explain.
For example, if a person starts to smile by contracting the major zygomaticus (AU 12) and then presses his lips (AU 24), the sequence can be categorized as a controlled smile (Keltner Reference Keltner1995; Keltner and Buswell Reference Keltner and Buswell1997; Reed, Sayette, and Cohn Reference Reed, Sayette and Cohn2007). We use smile control when it is not convenient to show happiness or dominance. For example, during a negotiation, one party may have an advantage over the other, but it would not be convenient to show this advantage for strategical motivations. For this reason, the face contracts in sequence the major zygomaticus (related to happiness) and then muscles related to control. Considering only quantitative aspects of facial actions, a general statistic would mix actions with different meanings if considered single contractions. If considered as single actions, AU 12 is a sign of a polite smile, AU 24 is interpretable as a subtle sign of anger; if understood in sequence, the meaning would be a controlled smile, which has a completely different meaning.
THE FACIAL ACTION CODING SYSTEM’S CONTRIBUTION IN ANALYZING SUSPECTS’ STATEMENTS IN A HOMICIDE CASE IN SOUTHERN ITALY
In this section, I will introduce the Facial Action Coding System’s application in the analysis of suspects’ statements regarding a homicide case to show the possible contribution of this technique in the field of criminology.
In 2010, a 15-year-old girl was killed in a small municipality in Southern Italy. The case was reported initially as a missing persons case and had a large impact on the Italian media.
The initial investigation concentrated on a possible voluntary running away of the girl from home. Investigators focused on her personal life, the analysis of her secret diaries and her activity on social media.
After more than one month, on September 29, 2010, her uncle found the victim’s cell phone partially burned in a location near to his house. This information led the investigators’ attention toward the family of the victim’s uncle.
Subsequently, the uncle of the victim confessed to having killed the girl after a rape attempt. He admitted to having abused the dead body of the victim. Moreover, he led the law enforcement members to the location of the body. The victim’s body was found in a well. The lower part of the body was immersed in the water; the higher part of the body was outside. In this condition, it was impossible to determine if sexual abuse had been committed. The coroner determined that the girl was killed by strangulation with a belt. Moreover, the injuries on the victim’s neck were compatible with the strength of a not particularly thin woman.
The coroner’s evaluations oriented the investigations toward other family members of the victim: the aunt and the cousin. In summary, the main hypothesis of the uncle’s guiltiness has been excluded for several reasons:
The uncle of the victim was a strong man. It is unlikely to think that he lost control during the rape attempt and strangled the victim with a woman’s strength. When he led Law Enforcement to the body’s location, he removed the 90 kg stone covering the well.
In his first confession, the uncle declared to have strangled the girl with a rope. The analysis of the coroner revealed the use of a belt.
The uncle’s confession was very specific and included behaviors of necrophilia and pedophilia. Investigation about the suspect’s life history did not reveal any necrophilic or paedophilic behavior before this event. Paraphilic behavior usually has an early onset. The uncle was 55 years old on the day of the victim’s death. The hypothesis of onset around 55 years is unlikely.
The uncle changed his version many times.
º He confessed the homicide with pedophilia and necrophilia acts.
º He stated that the victim’s death was caused by accident during a game between the two cousins.
º He suggested his daughter killed the victim, and he helped only in hiding the body.
º He confessed again to have killed the girl.
During one interview, he declared: “we do not know when the cell phone fell,” revealing the presence of at least two people during the homicide.
The uncle of the victim also confessed the crime during the trial. The defense attorney of his daughter tried to demonstrate that the uncle killed the victim for sexual motivations.
In this section, I will comment on the statement in which the victim’s uncle declared again to be guilty and described the homicide.
Transcription of the Statement of the Victim’s Uncle
Suspect: “I do not know if she wanted to tell me if I can ring or maybe she was asking me something else.”
Lawyer: “Therefore, you did not see her going downstairs, but you noticed her near you in the garage?”
Suspect: “Yes, I did not understand what she wanted from me, but she bothered me. I do not know why. I remember to have said: ‘Go away,' but she insisted. I do not know what she wanted. So, I lifted her to turn her back and make her go away, and she kicked me. Everything started from this kick… Maybe… She did not do this voluntarily… I do not know why … So, there was a piece of a rope on…”
Lawyer: “Wait, so she felt your touch and, therefore, reacted with this kick.”
Suspect: “Yes, I lifted her from behind.”
Lawyer: “You even did not realized of what…”
Suspect: “No, I did not realize how many times I wrapped the rope around her neck.”
Lawyer: “But you are talking about a rope… You know that”
Suspect: “It was a rope.”
Lawyer: “Rope, belt… Do you remember that…”
Suspect: “It was a rope.”
Lawyer: “It was a rope that you found there in the garage, and you wrapped this rope around the neck of…”
Suspect: “But I do not know how much strength I impressed on the neck. I only know that the cell phone was ringing in her hand and in that moment that I… I even do not know how to say… Then she fell on a compressor.”
Lawyer: “But how long did this event last?”
Suspect: “But even I do not know how long it lasted. I know that I was shocked and did not know what to do. In that moment… Because for how people know my, my shame, the… I even do not know how to say because in these cases, the situation must occur in order to make you understand.”
According to the criteria included in the official manual (Ekman et al. Reference Ekman, Friesen and Hager2002a), we utilised the Facial Action Coding System, and the interpretation of coding is based on the Investigator’s Guide (Ekman, Friesen, and Hager Reference Ekman, Friesen and Hager2002b:136–74).
Facial Action Coding System analysis shows an interesting facial reaction during this statement. Specifically, when the suspect states, “it was a rope,” it is possible to notice AU 14 (dimpler) on the right part of the face. This action is related to contempt or moral superiority (Ekman and Heider Reference Ekman and Heider1988; Matsumoto Reference Matsumoto1992).
Contempt happens when the subject recognizes that a target is morally wrong (Ekman and Friesen Reference Ekman and Friesen2003). The authors also define this emotion as “moral superiority” because people who give negative moral judgments toward another person or behavior indirectly state their superiority on an ethical side.
Research shows that contempt is predictive of aggression (Matsumoto, Hwang, and Frank Reference Matsumoto, Hwang and Frank2014a,b). Similar to disgust and anger, contempt is activated by moral violations. However, what is not clear in this analysis is what the victim should have done to be considered morally wrong by the offender. The uncle confessed several times to have killed his niece, putting all his efforts to convince others to have killed and abused her. The victim’s uncle’s most frequent facial expression while stating guiltiness was contempt, the representation of moral violation by the victim. Nevertheless, during all confessions, the suspect had never explained what he considered morally wrong in the victim’s behavior.
In summary, the analysis of verbal and non-verbal statements contradicts the following areas:
Investigations about the suspect’s past behavior did not reveal any paraphilic behavior in the direction described (necrophilia and pedophilia). The onset of this paraphilia after the age of 55 years is unlikely.
He changed his version several times.
During the interrogation, he declared: “We did not know when the cell phone fell,” suggesting the presence of at least two people on the crime scene.
He used impersonal sentences for the description of the homicide (i.e. “it was a rope”).
The description of the homicide is incompatible with the report of the coroner (the strength applied to the neck of the victim is compatible with the strength of a woman; the weapon used was a belt).
The facial expression of contempt during the description of the homicide is incompatible with the fact that, during all confessions, the suspect never stated what he considered morally wrong of the victim’s behavior. Questions about this topic could have clarified this fact eventually.
The possible reconstruction of the crime scene is based on many incongruences. The allegation hypothesizes that the suspect based his behavior on two paraphilias, pedophilia and necrophilia, a tendency repressed for years and released with a much weaker strength than expected, according to the report of the coroner.
The coroner’s report has been determinant because, together with several contradictions based on evidence, verbal and non-verbal statements of the victim’s uncle moved the investigator’s attention toward his daughter, the cousin of the victim.
The allegation toward the victim’s cousin is based on the fact that she could have killed her for jealousy. The witnesses consistently declared that she was obsessed with a boy, and they started a relationship. After the end of their relationship, the boy started messaging through the cell phone with the victim, which would have triggered the cousin’s aggression toward the victim and determined her death.
She always claimed to be innocent. The analysis of facial expressions shows interesting additional information about the concomitant emotions during her denial of allegations. Interestingly, the cousin expressed contempt during one of her statements while talking about the victim. Specifically, she described an event in which the victim was waiting for her, messaging with the cell phone. The exact statement is: “She was waiting for me as usual with her cell phone.”
The victim was not doing anything wrong while waiting for her with the cell phone. Still, the cousin’s facial reaction reveals contempt toward her, which is consistent with the allegation’s jealousy motivation.
When the victim was theoretically considered missing by the family, the victim’s mother found some secret diaries revealing that the victim was in love with the same boy having a history with her cousin. All the family members wanted to give these documents to the Law Enforcement to help the investigation and possibly find relevant information for finding the victim, who was believed to be missing. The victim’s cousin was the only family member not wanting to give these documents to the Law Enforcement. This fact raised doubts about her version at the trial. When confronted on this topic, the suspect stated that she just turned some pages and did not read in detail the diaries. The analysis of verbal and non-verbal communication showed some contradictions. During the statement “I just turned some pages,” the suspect contracted AU 1 and AU 2, moving up the eyebrows, and AU 15, the lip depressor. The combination of these three AUs can be interpreted as perplexity. The suspect wanted to highlight doubts about the importance of that information source. However, especially in the case of missing people, when nobody has a clue about the possible real motivation of the victim’s disappearance, it is not logical to state that the diary was not relevant for the investigation and, at the same time, to declare to have just turned some pages. It would be possible to determine if the source of information is not relevant only by reading the diaries thoroughly.
Moreover, by analyzing the full statement, right after this first sentence, she claimed: “Sincerely, I have not read the diaries so much.” Nevertheless, even though the verbal declaration seems to show that the suspect believed that the source of information was just not relevant, her facial reaction associated with the word “diaries” is emotionally charged by a combination of facial actions, indicating a “mixed emotion” (Ekman and Friesen Reference Ekman and Friesen2003), including:
Lower eyelid lifted (AU 7) – related to anger.
Upper lip lifted (AU 10) – related to disgust.
Lip stretcher (AU 20) – related to fear.
It is very interesting to notice the following contradictions:
The diaries are verbally described as “not relevant” or “not useful” for the investigation about the victim.
Simultaneously, the suspect’s facial reactions include synchronous facial expressions of fear, anger, and disgust while talking about the diaries.
The diaries contained information that the victim was in love with the boy who had had a relationship with the suspect. At that time, the victim talked to him, whereas the suspect was not communicating with him anymore.
The suspect considered this source of information “irrelevant” and “useless,” but she only turned some pages of that diary without any careful reading.
Another important point of analysis is that the suspect expressed fear while talking about the victim’s feelings toward the boy. In the video analysis, it is possible to notice the contraction of AU 20, the AU described in the Facial Action Coding System as the lip stretcher. The same reaction appears when the suspect remembers that the victim’s mother decided to keep the diaries at home.
Another interesting facial reaction by the suspect is the contraction of AU 12, indicating happiness. It is very curious to notice when this action appears. This facial action is concomitant when the suspect said that the relationship between the victim and the contended boy was “absurd,” “impossible.” Smiling is related to pride (Tracy and Matsumoto Reference Tracy and Matsumoto2008). The expression of pride while stating the impossibility of a relationship between the victim and the boy is coherent with the jealousy motivation hypothesis.
The analysis of non-verbal communication is a technique commonly recommended by police manuals and books on interviewing techniques in forensic settings; however, scientific research shows low correlations between non-verbal cues and deception. This article contains a framework for the application of this technique in forensic settings and related scientific research. The standardized analysis of non-verbal communication provides additional information to the interviewer, supporting him in formulating specific questions. Future research should, however, clarify the relationship between non-verbal communications and deception, including:
The factors outlined in previous scientific research.
The introduction of timing factors.
The analysis of the combination of facial actions.
The analysis of the sequence of facial actions.
The comparison between non-verbal communication meanings and verbal content.
I acknowledge Professor Emilio Viano for the opportunity to introduce the subject of this publication during the XIX World Congress of Criminology, organized by the International Society of Criminology, held in Doha, on October 28–30, 2019. Participation as a speaker at this event allowed me to enter into contact with experts from different countries, and exchange good practices in the field of criminology. I acknowledge the Editor of the Annals for the comments, editing, and proofreading.
Igor Vaslav Vitale is a psychologist, international trainer and European Union project manager. Dr Vitale has been an external lecturer in Applied Psychology and Criminology in the Post-Degree Master Courses at LUMSA University, University of Padova, and Tor Vergata University. He has been a key expert in more than 30 Erasmus+ projects in several areas of Applied Psychology, including school, clinical, work and organizational psychology and criminology.