In a conversation with a fellow language teacher today the following question came up. When one hears one of the increasingly common TV commercials with a voice-over by a famous actor (as we are writing one for a telecommunications company with the voice of Al Gore’s old roomie, Tommy Lee Jones came on, for example), within the first few words, without seeing a face or receiving any additional cues, and going exclusively on memory of recorded, broadcast and poorly reproduced sound, one can identify the person who is talking. This is what the advertisers are counting on. Now, think of how many voices you can do this with. Probably thousands. People you know, actors and singers, people you watch on TV,
The question was, what qualities of a person’s speech make this incredible feat of aural recognition possible? Words like “pitch” and “timber” and “accent” contain some of the answer, although they are hardly scientific. Stress, timing and phrasing are important but inadequate because often the identification can be made on the basis of a single word or syllable. What makes each voice unique?
In part, this phenomenon is a testimony to the incredible audio instrument which is the human ear. In a crowded ballroom with a hundred people involved in fifty conversations, human hearing can pick out and follow any one of those conversations, filtering out the rest in a brilliant feat of semantic cognitive processing.
But it is not simply an exquisitely sensitive and wide-spectrum audio input system. The input device is also connected to a software system (or system software) of deep semantic structures, viral language memes, which allow human beings to exchange data in a multimedia stream so dense and nuanced that the most advanced computers today cannot decode, decipher or independently produce more than a fraction of the content carrying capacity of human speech.
Talking about talking. What a bunch of language geeks we are….