Front Page > Entertainment Page

CSI: Language analysis unit

By Ann Geracimos
THE WASHINGTON TIMES
January 12, 2006

    "You have to know the context," notes Mr. Leonard. "There is no meaning without context."
    "When you face a law case, you don't know which tools you will need; it could be syntax in one, phonology in another; sometimes you just puzzle through," says Mr. Shuy, who has worked with the FBI on the Unabomber case, among other high-profile investigations. His latest book, from Oxford University Press, is "Creating Language Crimes: How Law Enforcement Uses (and Misuses) Language."
    While claiming not to be an expert in acoustic phonetics — the breaking down of recorded speech to the smallest unit of sound — he recalls a time before a jury in Reno, Nev., when he had to count syllables — "beats of sound" — to make clear a disputed transcript in a suspected bribery case. The difference hinged on whether the words were "I'll take the bribe, wouldn't you?" or " I wouldn't take a bribe, would you?"
    James R. Fitzgerald, acting unit chief in the FBI's Behavioral Analysis Unit-1 who has been with the agency for almost 20 years, recalls how a transposition of verbs in the manifesto written by the Unabomber helped lead to a closer identification of Ted Kaczynski in April 1996.
    The latter used the phrase "You can't eat your cake and have it, too," instead of the usual form, which is "You can't have your cake and eat it, too." Like most people, Mr. Fitzgerald thought Kaczynski had made a mistake. But examination of other letters by him contained a similar feature, which, Mr. Fitzgerald says, "is actually a traditionally middle English way of using the term. He technically had it right and the rest of us had it wrong. It was one of the big clues that allowed us to make the rest of the comparison and submit a report to the judge who signed off on a search warrant."
    Textual material helps in often quite subtle ways in narrowing a field of suspects. A later case he worked on involved a heart surgeon who murdered his wife but had tried pinning the deed on another person through some anonymous letters. Among other textual similarities, Mr. Fitzgerald found both the anonymous letters and the doctor's own writing samples contained similar and unusual spacing between words.
    This time Mr. Fitzgerald had the use of the bureau's new communicated threat assessment database, or CTAD, that, by breaking down the current repository of 1,500 communications into 23 different categories, provides what he says is the only comprehensive repository of its kind in the United States.
    The only person in the bureau with an advanced degree in linguistics — he received his master's degree in science from Georgetown last year — he says special agents once received only two hours' training in statement analysis. That since has evolved into a weeklong workshop open to members of the law enforcement profession at all levels. His own unit, he says, reviews between 500 and 600 cases a year, 90 percent of which are anonymous threats in the form of letters, tapes, e-mails and voice mails.
    "In the old days it took 14 separate steps to write a threat," he says. "Today you sit at a computer and, with the power on, write 'I'm going to kill you,' and in three steps you have committed a felony. Today it's much easier to send a threat, and it can be harder to trace one, depending on the method. Even so, measured lexical features are still important and tell a lot about the offender."
    

Want to use this article? Click here for options! Copyright 2006 The Washington Times