The Language of Deception

Separating lies from truths is key to successful negotiation, and research designed to uncover markers of deception can aid in this goal. Below, Michael T. Braun, Lyn M. Van Swol, & Deepak Malhotra consider how the analysis of linguistic patterns between liars and truth-tellers can betray our true intentions, and suggest the effect of this research on business.

Separating lies from truths and bluffs from honest offers is key to successful negotiation. Negotiators must know when to push for more and when to settle and accept an offer. Research designed to uncover markers of deception can aid in this goal. Research in this area has often focused on behavioral patterns (e.g., fidgeting, avoiding eye contact) and has achieved mixed results. New research, started in the late 1990s and accelerating today, tries a different tack: analysing linguistic patterns between liars and truth-tellers to see if the words we speak can betray our true intentions. This research uses computer software like Linguistic Inquiry Word Count (LIWC) that can quickly count the percentage of words used in specific categories (such as negative emotion words, swear words, pronouns, etc.) from transcripts.

How do these patterns manifest themselves? To understand that, we must first look at how researchers explain the language differences between truth tellers and liars. Two explanations, more complimentary than contradictory in nature, have emerged. First, some researchers suggest that liars use different linguistic patterns strategically to reduce the likelihood of the receiver detecting their lie. For example, one of the most supported findings in the literature is that liars use more words than truth-tellers, called the Pinocchio Effect.¹ Researchers explain this result by suggesting liars must speak more as a way to control the interaction. As the liar speaks more, his partner speaks less, thus enabling the liar to head off questions and otherwise control the flow of conversation. The greater verbosity of the liar may also help create a believable context for the lie, as the liar adds realistic, concrete details.  The amount of words may not be a reliable indicator, however, as some research has found the opposite pattern. Why might liars sometimes use more words and other times less? Speculation centres on whether the lie is verifiable or not.² For example, imagine a negotiation over the purchase of a piece of property in which the seller previously lied about increasing property values. A seller might wish to avoid discussing a recent news report about property values declining in the area. Because the potential buyer can uncover this information on his own, the seller may wish to speak less to avoid contradicting this knowable reality. In this way, the seller avoids revealing herself to be a liar.

In another circumstance, a seller might wish to encourage the potential buyer to make an offer right away. She might lie and state there are other showings later that day. The potential buyer has no way to verify if this is the case. Under these circumstances the seller can speak with impunity, use words to control the interaction, and add believable details about these other interested parties.

This difference might also occur in a written property listing. For a seller wishing to unload a piece of property that needs substantial upgrades, she may write a listing that is very short. This way, when a potential buyer sees the property, he cannot claim he was misled by the listing. On the other hand, if a flaw is undetectable (for example, very obnoxious neighbors), the seller can expand widely to control the interaction, hook the buyer, and make the sale with no worries that the truth will be revealed.

This idea has been used to explain contradictory findings between studies on deceptive linguistic patterns. In one study, deceptive online dating profiles were found to contain fewer words.² In another study, lying negotiators were found to use more words than their truthful counterparts.¹ How can the two findings be reconciled? For online daters, the goal of the lie is to attract more attention and get more dates. More words mean more potential lies that can be uncovered by romantic suitors when the couple actually meets face to face. Saying less may mean less chance that the suitor will feel misled and the relationship will fail. But in a negotiation where certain facts cannot be uncovered, the liar is free to speak at will without risk of jeopardising the interaction. Thus both speaking more and speaking less can represent a strategic use of language.^3,4

The strategic use of language is one way a liar may be detected using linguistic cues. Non-strategic language use may also expose a liar. This ‘leakage’ occurs because a liar is unable to control (or does not know to control) patterns like pronouns, conjunctions, and negative emotion words. Non-strategic linguistic patterns may reflect underlying emotional or cognitive differences between liars and truth-tellers.

What is the cognitive or emotional state of a liar? Interpersonal deception theory suggests that liars have more variables to pay attention to during the interaction and, thus, have a higher cognitive load. Lying is a difficult business. First, the liar must construct the lie, complete with details to make the fabrication believable. Second, the liar must remember any past lies or details he has fabricated to avoid contradiction. Third, the liar must monitor his partner for any signs of disbelief, so that he can preemptively counter them. And fourth, the liar has to monitor himself for any nervousness or nonverbal behavior that may be seen as a ‘tell’ by his partner. With cognitive resources tied up in this effort, the liar may speak using simpler sentences and overall reduced linguistic complexity.³ For example, a liar may speak in shorter sentences, use sentences that tend to repeat the patterns of previous sentences, and use a less diverse, more concrete vocabulary.

For some people, the act of lying may also result in feelings of guilt and anxiety because the act of lying conflicts with their self-image or because of the possibility of being caught.⁴ This too may be reflected in linguistic patterns. The Newman-Pennebaker model suggests liars may use fewer first-person pronouns and more third-person pronouns because the liar wishes to psychologically distance herself from her lie.⁵ Whereas a truth-teller may be fine to identify with her statements (e.g., ‘I’m offering 5% below market value’), a liar may assuage her guilt and anxiety through pronoun use (‘That offer is 5% below market value’). This distancing may also occur using negations (e.g., ‘not’, ‘never’); for example, a liar may say something is ‘not bad’ rather than ‘good’. Guilt and anxiety may also spill over into words with a negative twinge. For example, some researchers have hypothesised that liars should use more negative emotion words (e.g., ‘sad’).

Some research has supported these findings. Analysis of half a million Enron emails found that first person pronouns and exclusion words could successfully distinguish false from true emails.⁶ And analysis of statements about the Iraq war by Bush administration officials used first person pronouns, exclusion words, and negative emotion words to correctly classify a majority of the statements as true or false.⁷ This rate was higher than the classification by humans and higher than chance.

But again, there have been inconsistent results, leading some researchers to suggest that the context of the interaction is necessary to consider when using linguistic patterns to distinguish between lies and truths.^1,3,4 How might this work? Some research investigating lies and truths in online dating profiles found that deceptive profiles contained fewer negative emotion words, opposite of what was expected. Taking context into account, researchers reasoned that negative emotion words might cast the deceptive dater in a poor light; thus deceptive daters are motivated to avoid negative emotion words and their presence is a reliable indicator of truthfulness.²

Where to go from here? One challenge to dynamic use of these findings is that all come from a substantial body of similar texts to analyse and compare, texts that also have been objectively classified as truth or lie. Therefore, researchers can establish a base rate of different linguistic patterns among a large number of truths and lies.  When engaged in an interactive negotiation, a baseline of, say, first-person pronouns may be impossible to establish. Is the seller using fewer first-person pronouns than a truth-teller would? It’s impossible to say.

Yet many negotiators may repeatedly engage in similar interactions and have access to the truth or deception in those interactions after the fact. In these cases, careful analysis of past interactions may be used to establish a baseline against which to assess negotiations in momento.

This information may also be useful to those wishing to avoid detection of deception. While researchers have argued that some linguistic cues are difficult to control, much of the patterns revealed might simply be lack of knowledge about how linguistic patterns may differ. The clever liar (who may also be known as simply a skilled negotiator), armed with knowledge of deceptive linguistic patterns, should have no trouble adjusting speech to avoid detection. Given the rudimentary measures that are used (for example, total count of first-person pronouns), simple shifts in speech can evade attempts to ferret out falsehoods.

About the Authors
Michael T. Braun is a visiting assistant professor in the Department of Communication at Millikin University. He received his Ph.D. from the Department of Communication Arts at the University of Wisconsin-Madison in 2013. His research focuses on communication technology perceptions and decision making across the lifespan, with special attention on deceptive interactions.

Lyn M. Van Swol is an associate professor in the Department of Communication Arts at the University of Wisconsin-Madison. She received her Ph.D. from the Department of Psychology at the University of Illinois-Urbana Champaign in 1999. Her research centres on dyadic and group interaction and examines what factors are likely to increase acceptance of information during an interaction.

Deepak Malhotra is a professor in the Negotiations, Organizations and Markets Unit at the Harvard Business School. His research focuses on negotiation, deal-making and conflict resolution. He is also the best-selling author of books Negotiation Genius and I Moved Your Cheese.

References
1.Van Swol, L. M., Braun, M. T., & Malhotra, D. (2012). ‘Evidence for the Pinocchio effect: Linguistic differences between lies, deception by omissions, and truths’. Discourse Processes, 49, 79-106. doi: 10.1080/0163853X.2011.633331
2.Toma, C. L., & Hancock, J. T. (2012). ‘What lies beneath: The linguistic traces of deception in online dating profiles’. Journal of Communication, 62, 78-97. doi: 10.1111/j.1460-2466.2011.01619.x
3.Braun, M. T., Van Swol, L. M., & Vang, L. (2012). ‘His lips are moving: Pinocchio effect and other linguistic indicators of political deceptions’. Paper presented at National Communication Association Conference, Orlando, FL.
4.Van Swol, L. M. & Braun, M. T. (Forthcoming). ‘Communicating deception: Differences in language use, justifications, and questions for lies, omissions, and truths’. Group Decision and Negotiation.
5.Newman, M. L., Pennebaker, J. W., Berry, D. S., & Richards, J. M. (2003). ‘Lying words: Predicting deception from linguistic styles’. Personality and Social Psychology Bulletin, 29, 665-675. doi: 10.1177/0146167203029005010
6.Gupta, S. & Skillicorn, D.B. (2006). ‘Improving a textual deception detection model’. Proceedings of the 2006 Conference of the Center for Advanced Studies on Collaborative Research, article 29. doi: 10.1145/1188966.1189005
7.Hancock, J. T. Unpublished manuscript.