February 11, 2007

The perils of transcribing spoken language

Heidi Harley's recent analysis of why some listeners heard Jimi Hendrix sing "Scuse me while I kiss this guy" when what he apparently sang was "Scuse me while I kiss the sky" reminded me of the many wrong transcriptions of spoken language that pop up in government transcripts of tape recorded undercover conversations and court hearings. Sometimes a local expression is the problem, as when the government transcribed "it's deeper than a post hole toad" as "is steeper than a postal code" in a Texas sting operation some years ago. No, the bad guys weren't plotting to steal post office files. They were simply using a colorful, but not broadly recognized, Texas expression about a totally benign topic.

This week the legal affairs writer for the Associated Press called me  to talk about the problems media and government witnesses were having as they tried to decypher their own notes in the perjury trial of Lewis "Scooter" Libby. Somehow our conversation turned to the problems that court reporters have when they transcribe the proceedings of trials and hearings. The writer then told me about Prosecutor Fitzgerald's statement to the judge when he explained that there would be no deal made in this case. The court reporter's transcript had Fitzgerald saying that he couldn't do this because "it's a thicket of hope." I suppose that if you really worked at it, this version might make some sense but what the prosecutor actually said was "it's a pig in a poke."

It's possible, even likely, that the court reporter had never heard this expression before. But the Dictionary of American Regional English (DARE) shows that it's used in New England, the Midwest and throughout the South to mean "something one cannot see or evaluate before deciding to actually accept or purchase it." And it's been around for a long time. Even George Washington used it as early as 1788, according to DARE:

I am not fond of buying a Pig in a Poke (as the phraze is)

These misperceptions of "post hole toad" and "pig in a poke" weren't all that crucial but sometimes faulty transcriptions can have  potentially disastrous effects. Some years ago, in a trial of two Nevada brothel commissioners who were indicted on charges of trying to solicit a bribe from a madam who wanted to get a license to set up a trailer for "business purposes," the case boiled down to what the government's transcript reported that one of the two commissioners had said to the other one during their brief conversation with the madam:

Commissioner A: I would take a bribe, wouldn't you?

The tape, badly recorded in a Sparks restaurant and with lots of clanging dishes and blaring music, was very hard for jurors to understand. In fact, it was hard for anyone to understand. After listening to it many times, however, I concluded that what this brothel commissioner actually said was:

Commissioner A: I wouldn't take a bribe, would you?

Now my problem was how to communicate this vitally important difference to linguistically untrained jurors. I finally decided that the best way would be to call on their sense of timing and arithmetic. Most people know what a syllable is and most people can recognize a brief pause and most people can count. So I had them listen with me as I played the sentence several times and asked them to count the syllables before the pause and then count the number of syllables after it.

Together, as the tape was playing, we heard:

___    ___   ___   ___   ___   ___  /pause/ ___   ___
  1         2       3      4       5       6                  7       8

This, of course, corresponded with my version of what was said:

I     would   n't    take    a   bribe /pause/  would  you?
1       2         3       4       5      6                    7         8

The government's transcript had transported the negative contration of "would" from the third syllable to the seventh. It's version had 5 syllables, a pause, and then 3 syllables. In contrast, my transcript had 6 syllables, a pause, then 2 syllables. I played the passage a couple more times and the jury got it. I could see them counting with their fingers and they seemed to be having fun. Language can be fun, even in jury trials. They were now able to overcome the difficult tape and transfer the negative contraction to where it was actually said, which made all the difference in the case.

So what causes people to mishear spoken language? The "post hole toad" transcription problem probably happened because the transcriber had never heard of the quaint expression and did the best that could be done with it. The humorous "pig in a poke" transcription error probably came about because the transcriber was unfamiliar with the expression and tried to make some sense out of the prosecutor's words. But the "I would take a bribe" transcription most likely has a different root cause. Undercover tape transcriptions are usually made by secretarial assistants, then reviewed by law enforcement agents. When agents are predisposed that defendants are guilty, as is common in such cases, it's likely that they hear what they want to hear.

Posted by Roger Shuy at February 11, 2007 02:33 PM