American Scientist

To the Editors:

Over 60 years ago, when I was an engineering student at the University of Connecticut, I was part of an experiment to determine which part of the human voice spectrum communicated information. It was obvious, even then, that while the voice spectrum is many kilocycles wide, the information transfer is at teletype speeds, typically under 100 words per minute. The brain processes a simpler signal generated in the structure of the ear. My job, along with other volunteers (paid 50 cents per hour), was to listen to artificial speech written by hand.

Your article brought memories of this experiment back to me. One of the principal experimenters could almost miraculously speak to us with a calligraphy pen on a strip of paper processed by the frequency synthesizer. Not only was he easily understood, he could speak to us in accents from New England to the deep South. The spectrograms in the article on bird speech (“The Complex Call of the Carolina Chickadee” by Freeberg, Lucas, and Krams in the September–October 2012 issue) look much like what was processed by the machine we listened to.

If they keep at it, someday these experimenters may be able to talk to the birds in their own language.

Roland Boucher
Irvine, CA