Tricky to judge body language

The apex court recently issued a warning about something judges do routinely but is seldom commented upon.

The Court of Appeal warned judges against relying too much on witness demeanour to detect lying. The warning came in its written grounds for overturning a High Court decision granting an investment firm's claim to $1 million in dividends.

How much store should a judge set by witness demeanour in making a decision? This has serious ramifications in a criminal case where a guilty verdict can see the accused jailed or caned.

While "appearance" means the outward aspect, "demeanour" refers to how one acts towards others, consciously or not. The demeanour includes one's facial expressions, eye contact, gestures, tics and body language.

Little has been said on how judges should deal with witness demeanour. Perhaps we trust judges to use cold logic in following the rules of evidence and courtroom procedures, by which they draw rational conclusions from the evidence to reach a correct and just decision. Even judges may believe this about themselves as impartial assessors of fact who look only at the evidence presented on the witness stand.

But even if the rules of evidence and courtroom procedures might be adhered to assiduously, the courtroom is not a sterile operating theatre where a case is dissected with surgical precision.

Rather, it is reality theatre where participants play their assigned roles for the judge to determine guilt (in criminal cases) or fault (in civil cases). That is why demeanour matters, for the judge cannot help but notice it. Being human, judges are prone to being influenced by perceptions, no matter how hard they try to be objective.

Clearly, what lawyers say in court and whether they deliver it convincingly matter in the outcome of a trial; similarly, a witness' verbal and non-verbal behaviour, mannerisms, dressing and speech. Even spontaneous reactions in the gallery or the presence of the victim's family can add colouration to one's perception of truth.

So, what may not be formal evidence could still affect a trial's outcome, although judges are expected to assess such non-testimonial "evidence" judiciously.

But law professors tend to argue that demeanour should not be considered as it gives rise to the potential for prejudice, unless witness conduct in court is so brazen that no one can miss it.

In the early 1990s, based on research in psychology dating from the 1960s, a consensus emerged in the legal academy that observing demeanour is, at best, inaccurate and, at worst, useless for assessing credibility and culpability.

That older research found that people who used demeanour cues to detect lies were accurate only half the time. However, more nuanced studies recently show different results.

First, they affirm that demeanour cues by themselves indeed do not help detect deception.

For example, cues believed by many to show lying such as gaze aversion, speech pauses, blinking, smiling and fidgeting, in fact, are not linked to deception at all.

This was shown in two meta-analyses of newer studies on deception cues: one published in 2003 by University of Virginia psychologist Bella DePaulo and co-workers in Psychological Bulletin; the other by psychologists Siegfried Sporer and Barbara Schwandt of the Justus-Liebig University of Giessen, Germany, in Psychology, Public Policy, And Law in 2007.

Meta-analyses aggregate many studies on a particular issue and analyse their findings collectively to pick out patterns, if any. These meta-analyses of newer studies show that when the assessor has additional information about an event the witness is unaware of, then the assessor is more accurate in picking out witness lying.

This suggests that appropriate context improves lie detection a great deal. Such contextual information may come from third parties, physical evidence and confessions. With such information in hand, the judge can more accurately determine whether a witness is likely to be lying or not.

Conversely, if the witness is also aware of that contextual information which the judge has, then the judge's accuracy in detecting lying drops way under 50 per cent. The reason for this might be that liars tend to tell stories that are less consistent and also have fewer details. So, if you have an edge in contextual information, you can formulate more detailed questions to detect deception.

Where the judge does not have this edge, he won't be able to detect deception accurately. In that case, it is his initial bias about the witness that will incline him to believe whether the witness is lying.

The newer studies also show that if contextual information is not available, people are more accurate at identifying truths than lies. The 2003 DePaulo meta-analysis showed that people correctly figured out true messages to be true in 61 per cent of cases. In contrast, they actually detected false messages to be false in just 48 per cent of cases. Later studies show similar results.

What this means is that when we think someone is being truthful, we are more likely to be correct than when we think that he may be lying.

This is probably because, in life, people say things that are true most of the time, perhaps because of general etiquette. In life, there are probably more truthful statements made than lies. Picked at random, the chances for a truthful statement are higher than a lie.

For this reason, when we use deception cues to judge a witness, it is still a coin flip but one biased for truthful statements - simply because there are just more truthful ones around in general.

What does all this mean?

First, demeanour cues without contextual information don't make lie detection more accurate.

Second, demeanour cues on their own are more likely to help one pick out truths than lies.

Third, if there is contextual information, which is likely if a crime has been well investigated, lying is easier to detect. Even so, one may be wrong at least a third of the time.

All in, demeanour cues may sometimes be useful as a guide to deception detection. But there is much room for error, so the court's warning is prudent. Judges would be wise to rely on demeanour contextually.

This article was first published on June 26, 2014.
Get a copy of The Straits Times or go to for more stories.