Thoughts on AI

Thoughts on the future of humanity, usually posted while I am drunk.

Thursday, November 25, 2010

They sound the same to me.

So I'm a voice guy, I love listening to people's voices. I love the distinctions, the subtle expressiveness of the tones, the tone color, varying accentuations and emphasis people use. Its music, really, and usually no two voices are alike.

But something bothered me today. I bought something at a nearby gas station, and the old woman who worked there had the exact same voice as a woman who worked at a different gas station nearby. The were both older Asian women who spoke English as a second language, but the patterns of tones and volume changes was so identical it was alarming, though I could see the women were different people.

It came to me, that in learning a language, we establish a feature library, an alphabet out of the communications of people around us as children. This tonal and volume library goes far beyond the mere words (for most of us) and as we communicate, we unconsciously choose elements from this library to express ourselves. Both of those Asian women learned their own libraries in their youth, libraries that use features alien to English, and are simply not recognized therefore by my English brain...So it wasn't that they were speaking identically, its that they were both speaking within their culturally established feature ranges, but because those feature ranges are not part of my tonal library, my brain couldn't distinguish them. A little like when you come upon a foreign web page with a different encoding, so it just shows up as a bunch of question marks: each such page looks a lot like the last.

0 Comments:

Post a Comment

Subscribe to Post Comments [Atom]

<< Home