Analysis of statements has played an important part in criminal investigations. However, new insights have shown that deceivers can now be identified by how they say things and not just by what they say.
The art of deception is an integral part of survival and has formed part of the natural selection process for a few million years. It has proved so successful that deception, either to advance our own interest or to avoid punishment, now forms part of everyday communication.
Nature, however, contains its own checks and balances. Just as natural selection produces deceivers, it also produces deception detectors.
There exists a popular notion that deception can be identified through the ways an author (or speaker) adds or subtracts from the perceived truth. But this notion is dangerous, because deceivers do not always say, or avoid saying, specific things.1 However, cultural and educational background, social class, sex, age, occupation, personality, and geographical location all affect how we speak and behave. There is no such thing as typical deceptive behavior.
What may be deceptive in a certain context may not be so in a different context. The presence or absence of strong emotion, or the failure to deny one’s guilt, are only indicators of deception in the context of specific persons and specific situations.2 Additionally, there is no evidence that a lack of memory about an incident is generally indicative of deception.
Many deception/detection strategies operate on the assumption that psychological guilt and anxiety, which are characteristic of deception, manifest themselves in verbal and non-verbal behavior. However, many have misconceptions about what actually constitutes deceptive behavior. As a result, most people are bad at identifying it.
While individuals are able to control certain aspects of their behavior quite well, they are generally unable to control all their behavior.3 As a result, clues to anxiety and deception “leak” out through various communication channels that deceivers find themselves either unable to control or ignore.
Generally, people are very good at controlling the content of a story. Deceivers know they need to construct a convincing tale if they are to appear truthful, so they concentrate on content and ignore the way they deliver it.
However, addressees also tend to ignore the way a story is delivered. Both groups are so focused on story content they fail to notice “leakages.” For that reason, how someone says (or writes) something, rather than what someone says, is one of the most informative and reliable indicators of deception.
Current techniques are based on the analysis of words in the context of the information in which they appear. Analysts look for words that suggest knowledge inconsistent with the information being provided.
This is best exemplified by the analysis of a statement by Susan Smith who intentionally rolled her car into a lake in 1994, drowning her two sons. She told police that her children were abducted at gunpoint. “My children wanted me,” Smith said. “They needed me. And now I can’t help them.”
Officers investigating the disappearance of the children concluded that Smith must have known her children were dead because she referred to them in the past tense whereas parents would normally continue to believe their children were still alive.4 Smith was subsequently arrested, tried, and convicted for the murder of her children.
Within context, verbal incongruities by themselves are not indicative of deception. A statement can be truthful yet contain inaccuracies from distorted perception or memory of events. Identifying incongruities merely highlights issues, which analysts should revisit and explore further to determine whether deception is indeed taking place.
While there can be no doubt that contextual analysis is important in identifying deception, studies show that linguistic style is a more reliable indicator of deception across context.
Linguistic style (how something is said) is the way that people speak and write that distinguishes them from others.5 It is constant across time, situations, and topics. It is a reflection of the author’s perception of himself or herself, and the world in which he or she lives in.
Language is composed of words belonging to different classes that work to define grammatical “rules.” These rules become so internalised as we learn to speak, that we form coherent, and often complicated sentences, without consciously having to consider structure or word choice.
As a result of this internalisation, many of the words we use are loaded with information reflecting personal experiences, understanding of events, and individual degrees of mental (cognitive) processing.
This personal perspective influences all “information-heavy” words (such as nouns, verbs, adverbs, and adjectives) in a person’s mental dictionary. For instance, someone who comes from a stable family background has a different perspective of the word “Dad” than a person who has never known a loving father figure; someone who has always lived in a big city has a different idea of what constitutes a “road,” than one who lives in the Australian Outback.
For this reason, these word classes are heavily influenced by culture, education, and geography, and therefore are not reliable for analytical purposes across groups of people.
On the other hand, certain word classes (such as pronouns, conjunctions, prepositions, auxiliary verbs, and articles) have little information content, and are not subject to these external influences.
These word classes are called “particles,” and are responsible for creating coherence and managing information flow. Although we commonly use less than 200 such words, they account for more than 50 percent of our vocabulary.6 We use them largely subconsciously and they are the words that characterise linguistic style.
As the deceiver concentrates on information-heavy words to convince the “audience” that he or she is sincere, the deceiver will pay less attention to particles that apparently do not contribute to the story. Yet it is precisely these insignificant words that are most revealing about deception.
Certain characteristics are typical of deceptive communication, and are not dependent on context: uncertainty, vagueness, nervousness, reticence, dependence, and unpleasantness (i.e., negative effect)7. Most of these characteristics dictate how and when we use particles.
Self-references (I, me, my) are an important reflection of personal association with what is being said. The use of the first person singular pronoun “I” is by far the best indicator of honesty. By using “I,” people express “ownership” of the topics of which they are writing.
Absence of ownership goes beyond just dropping the “I.” It also includes recounting events in which the author plays no active role even though he or she is at the centre of events.
Consider this extract from a taxi-driver’s account of an attempted robbery:
“The guy outside of my cab said something to the guy in the cab similar to ‘get out of the cab and help me kick this guy’s a__.’ It was similar to this but not exact. The other guy got out of the cab very quickly as the guy outside the cab started to swing at me. This guy was landing punches in my face and back. The guy who just got out of the cab didn’t look like he was going to join in but after the other guy had hit me several times he joined in and started to swing wildly at me. This guy was very drunk and was making contact but his punches were not effective. The first guy was landing his punches to my face, back, back of my head and neck area. This guy was very effective and literally beat me down.”
This extract differs from the rest of the account in that the author’s presence has changed from “I” to “me” (representing the author as recipient of the action). The focus is on a third party (the guy, this guy, the other guy – “other” reference) and not on the author. This absence of “I,” the use of “me,” as well as increased “other” references, allows the author to write an impersonal account.
Pronoun choice reflects the author’s perspective. There is a distinct difference in perspective between “he and I” and “we” or “you and them and I” and “us and them.” This identifies with whom the author is willing to associate himself.
Later in the same account, the cab driver reports,
“When the guy that was the drunkest came back at me I shovedhim away from me and fired two or more rounds at him. We were about eight feet or less apart and he was facing me.”
This is the only time the joint reference “we” occurs in the account. The change in pronoun “leaks” the author’s true perspective of events. It highlights his willingness to be associated with a man who allegedly assaulted and tried to rob him. These changes in linguistic style, irrespective of the content of the story, bring into question the author’s version of events.
Particle use is also significant in combination with other particles. Conjunctions and prepositions introduce explanation and rationale. Such function words are referred to as “explicans.” A high incidence of “I” associated with explicans normally indicates anxiety.
Another particle class, auxiliary verbs (be, do and have in combination with other verbs e.g., did run, has jumped, had been shouting), introduce shades of time and mood. Heavy use of it indicates uncertainty and distancing.
Consider this extract in which the author explains his presence in a particular location without anyone actually seeking the information. The author unexpectedly appeared at the same holiday location as the addressee, an ex-girlfriend, whom he had been harassing. He wrote to the addressee explaining his presence there, believing that she had seen him there.
“I would certainly not come over if I thought you were still here. I only arranged this visit last Friday after I had seen your car outside your home. I thought wrongly that you had returned home. There was no one I could ask to verify it, so I thought it would be an ideal time to kill two birds with one stone. Firstly to get this place off my mind once and for all, thereby completing my therapy to recover from the unjust way I have treated you [for] which I am totally ashamed. Secondly to see what it was like to go on a ‘singles’ holiday.”
The author’s linguistic style is very much defined by the combination of the first-person personal pronoun “I” and explicans (if, after, that, so; and adverbs, “firstly” and “secondly”), which indicate anxiety. In addition, he also uses a high incidence of auxiliary verbs, which indicate uncertainty. The author’s linguistic style, the need to explain his position, together with qualified actions and needless emphasis indicate the statement is deceptive.
Deceptive statements are also characterised by an increased use of negative emotion words (e.g., no, never, angry, wrong, sad, etc.)8. Some deceivers may react angrily, either offensively or defensively, particularly when information is challenged. Others may suffer anxiety because they are fearful of their deception being discovered. Leakage of these underlying emotions may result in a use of negative emotion words.
In the extract below, consider how the author’s feelings of guilt and anxiety translate into angry words (in bold).
“I told her that Betty was the culprit, I did this because my so-called friend Betty took the first opportunity to blame me for this misadventure. I was extremely disappointed and stunned to say the least. The following day, we were called to the office and confronted again by Mr. Weatherbee. Betty was so angry with me for telling the truth, that she was beet red in the face and crying. She appeared to be ashamed of herself. ... She called twice on three-way to entice me to confess which I did not! She passed on information that she is going to have me beaten up and has a friend of hers lie and say that I confessed to this crime”9
The author’s anger is not only directed at Betty (whom she accuses of lying) but also at Weatherbee, (who “confronted” both Betty and her), and Betty’s friend (“has a friend of hers lie”). The author’s anger is directed at everyone associated with the event, brought on by her anxiety at the possibility and repercussions of discovery. The author’s anxiety is reflected in her emotional use of negative words.
Not all deceivers will exhibit verbal characteristics associated with deception. Individuals have different emotional responses to lying.
Influences on linguistic style as a result of deception depend very much on the complexity and consequences of the lie and the discovery of the deception, and the character of the individual.
Planned lies are easier to tell than spontaneous lies. People with certain personality disorders do not experience guilt or anxiety associated with wrongdoing, nor do people who lie about what they believe are inconsequential matters or who believe their lies are justified.
However, the problem all deceivers have in common is that of “disclosure” – how much of the truth can be withheld, or a lie fabricated, that will achieve the deception yet contains the lowest risk of discovery.
Deception occurs by manipulating the amount, truthfulness, relevance, and clarity of the information conveyed10. To achieve this, deceivers use a number of linguistic styles to construct their deceptions.
There are three main linguistic styles,11 which are generally predictive of deception:
This style is characterised by semantically rich and complex sentences, in which the information contained is not clear or relevant. Deceivers try to hide and dilute their lie with many words, qualifying and offering other information in exchange for the information they do not want to provide. The deceiver may use:
This style is characterised by a reticent attitude and indirect responses. Deceivers avoid discovery by saying the bare minimum. The deceiver frequently may use short sentences, which lack in fluency but appear assertive or substitute a different focus to avoid a direct response. Also, the subject may be missing and the action implied rather than explicit.
This style is characterised by the author’s disassociation from the messages and shifting the focus to an external context (in which the situation does not involve the author or any personal input). Overall, deceivers prefer to remove themselves from personal involvement in their deceptive messages. The deceiver:
The statement may consist of just one deceptive linguistic style, or a combination of them, depending on the situation, and what the deceiver finds to be most effective.
Consider the response to a request to clarify information regarding ownership of a property, and whether the author derived any income from it.
“As I said in our meeting in January my mother was the beneficial owner of the flat and it was transferred to her last year. I have asked her to confirm if she rented out this property and she confirms that she has for a short period before she decided to sell it. As to the amount she believes that it has not exceeded £400 a month, however once payments had been made by her the actual rent was much less than that. None of this rent has been paid to me. I also asked her if she permitted anyone else to live there during the period before I purchased the flat and she said she believes she has.”
In this extract, the linguistic style is characterised by increased use of other references (“she”) and reduced self-references ( “I”). Auxiliary verbs and qualifiers “soften” actions, and there is high variation between active and passive voices. Sentences are long and wordy, but information is vague and mostly irrelevant. The author focuses on his mother’s actions and income instead of his own.
One sentence (“None of this rent has been paid to me.”) is much shorter than the others. The subject (performer of the action) is missing; the focus (income derived from the flat generally) is substituted with a narrower definition (“this rent” paid to his mother), thereby avoiding a direct response.
High-consequence deception requires a great deal of mental planning. To lie successfully, the deceiver has to control all his or her channels of communication to avoid leakage of underlying emotions. This is simply not possible.
Information-heavy words (such as nouns, regular verbs and modifiers) have not shown consistent social or psychological results, probably due to their susceptibility to external influences.
On the other hand, particles (pronouns, conjunctions, prepositions, articles, and auxiliary verbs) carry tremendous social and psychological meaning, and have shown themselves to be constant over time, topic, and situations. Consequently, they are true markers of linguistic style.
Statement analysis has come of age. Deceivers can now be identified by how they say things and not just by what they say. In the right hands, statement analysis provides a powerful new tool in the fight against fraud.
Isabel Picornell, CFE, is a principal with QED Limited, in the British Channel Islands. Her e-mail address is:
isabel.picornell@qedfraud.com.
Unlock full access to Fraud Magazine and explore in-depth articles on the latest trends in fraud prevention and detection.
Read Time: 12 mins
Written By:
Annette Simmons-Brown, CFE
Read Time: 16 mins
Written By:
Dick Carozza, CFE
Read Time: 6 mins
Written By:
Patricia A. Johnson, MBA, CFE, CPA
Read Time: 12 mins
Written By:
Annette Simmons-Brown, CFE
Read Time: 16 mins
Written By:
Dick Carozza, CFE
Read Time: 6 mins
Written By:
Patricia A. Johnson, MBA, CFE, CPA
