The End of the Lie

I recently met in London with Dr. Richard Wiseman, a British scientist who studies the behavior of lying. He said humans constantly scan each other for signs of deception, looking for subconscious “tells” that betray a fib: a twitchy lip, a hesitant answer, a tapping toe. “Lying is a way of seizing control — that’s why we want so badly to detect it,” he said. “But we’re dreadful at it. We think that if someone won’t make eye contact, or they’re fidgeting, that they’re lying. Actually, those things don’t tell us much.” So people started looking for better ways to pry the truth from each other.

In ancient India, so the story goes, a priest invented the first recorded lie detector, and maybe the best so far. After a crime in the village, he announced that his magic donkey could determine a man’s truthfulness. He led the donkey into a dark room, and sent in the suspects, with instructions to pull the donkey’s tail. The donkey would speak, the priest said, when the dishonest man gave a pull. Each man took a turn, and of course the donkey never spoke. But in preparation, the priest had secretly rubbed lamp soot on the donkey’s tail. Afterward, he asked for a show of the suspects’ hands: The man with clean hands was pronounced a liar.

The modern magic donkey — the polygraph — arrived in 1935, when a Wisconsin detective contrived a device that measured blood pressure, breathing and sweat. The world embraced it: The military, the FBI, the CIA, and law enforcement agencies everywhere started wiring up subjects and peppering them with questions. The Department of Defense set up a whole Polygraph Institute. The polygraph crept into films and television shows, and into the public consciousness: the lie detector.

A few years ago, following allegations of espionage and stolen nuclear secrets, the prestigious National Academy of Sciences started an investigation to answer a simple question: How accurate is the beloved polygraph, really?

Not very, it turns out. Not very at all.

The polygraph isn’t a lie detector, for starters. It’s an anxiety detector. It measures physiological changes — such as increased heart rate — that we expect from a liar. But if you’re an Al Qaeda informer, telling the CIA the truth about your associates might induce more anxiety than lying. Also, NAS pointed out, the polygraph can be duped. Aldrich Ames, a CIA spy and one of the most lowdown and ruthless traitors in American history, breezed through routine polygraphs for a decade while he sold out his colleagues to the Soviet Union.

In the end, the NAS determined that the polygraph isn’t worthless — it’s worse. It hinders the truth. For every spy caught by a determined polygraph tester, according to the study, a whopping two hundred innocent subjects would be swept into the net.

The polygraph’s only decent use, the NAS concluded, is as a tool of intimidation. A criminal would strive to avoid a polygraph if he thought it worked, and might even confess outright in fear. In other words, the polygraph works in the exact same way as the sooty magic donkey.

What the world needed was a machine that reached beyond the body — beyond the sweat and the quivering voice — and directly into the brain itself. What we needed, in other words, was better magic.