I watch a lot of police procedurals and a standard scene is where an experienced detective interviews someone and then later tells a colleague that the person is lying. They say it confidently, and the viewer is led to believe that their wide experience with people who lie makes them capable of detecting when someone is telling falsehoods, that subtle clues reveal it. I don’t play poker but I am told that good players can tell when someone is bluffing by picking up on subtle indicators. There are also apparently TV shows whose central characters are people who are professional detectors of when people are lying.

But this article says that it is very hard to know when someone is lying.

Indeed, the study of lying is at least a century old, and thousands of scientific papers have been published. Researchers have mainly focused on the two key questions raised by that quote from Freud and the anecdote from India, namely: how good is the average person at detecting lies; and what, if any, are the behavioural signs of lying?

The answer? Not so much.

But how about people whose jobs involves trying to detect lying? Surely they’d be better?

[T]he research participants in the study, who had to judge which students were lying and which were truthful, were professional fact-finders – court judges, psychiatrists, criminal investigators. Despite their professional expertise, the participants were unable to distinguish which nursing students were telling the truth about their state of mind, and which were faking it – scoring either no better, or barely any better, than if they’d guessed at random.

If people whose job involves the detection of deceit weren’t competent on such a task, is there any reason to expect ordinary citizens to be so? As it turns out, we are not. The general findings of this classic study have been confirmed by a large number of laboratory studies with laypersons and professional fact-finders as participants, all pointing to the same conclusion: on average, people are not able to tell lies from truths based on how others talk or behave.
…

So, despite the fact that cultures throughout history have had quite firm ideas about how an untruthful person behaves, the science suggests people are generally poor at detecting lies.

While verbal and visual clues are not reliable indicators of lying, there are other more mundane things that can reveal lies.

So, how best to detect a lie? Well, there is one reliable procedure based on common sense, and that is to simply find out what the supposed liar says that does not fit with other stuff that you know. This is a recommended strategy for police interviews and, after a century of research, the only one to disclose lying in everyday contexts. A suspect should initially be asked to give as complete an account as they can of the alleged incident, together with their alibi or reason for their denial. In such a set-up, because they don’t know what evidence the authorities have, an interviewee can inadvertently produce a narrative that is inconsistent with other relevant and reliable information, thereby in many cases trapping themselves in a lie.

But the article says that despite this research, people are likely to continue to think that they can intuit when someone is lying. But as the case of pathological liar George Santos reveals, even the most outrageous liars can get away with it , at least for some time. What trips them up is not any signs they give while lying but discrepancies between what they say and the actual record.

Paradoxically, the idea that we can detect liars is probably what enables liars to get away with it, because we think that we can figure them out without doing the work of checking out what they say.