高普考/三四等/高員級◆英文題庫下載題庫

    Scientists in the United States are planning to create lie-detection software for emails and text messages. The new software is being developed by a team led by Jeff Hancock at Cornell niversity in New York State. "We asked people to come into the labs and write lies and we had other people come in and write the truth," says Hancock. "Using this method, we've developed a database of tens of thousands of emails.” 

    The scientists have analyzed the data and identified a number of methods to predict whether or not a person is lying. One of the most important methods is the length of a message. Lying emails have, on average, 28% more words than truthful messages. Another method is the overuse of sense words, such as "see," "feel," and "touch." "When you're lying, you want the other person to believe you, so you give more details," says Hancock. 

    Liars also tend to use pronouns like "we" and "they" rather than "I" as a way of distancing themselves from a lie because they feel guilty. "People often use negative emotional words too, because they feel uncomfortable when they're lying," says Hancock. "So they will use terms like "sad," "angry," "unhappy" and "stressed." 

   Previous research has shown that up to one third of communications at work involve lying of some sort. The most common workplace lie is pretending to be sick to avoid going to work. 

    But Peter Collett, a psychologist who has written a book on lying, is skeptical about Hancock's work. "You can see if somebody is lying by looking at their body language, how fast they talk, and how they speak," he says. "How can a computer program see all this? They've clearly done a lot of research, but with emails all you have is words."