Does a greater control over the vocal chords lead to greater mastery of talesmanship?
Over the past few weeks, I have been training (and been trained by) Dragon NaturallySpeaking, a voice recognition program. In the course of this process, I have probably talk out loud more than I have at any other comparable period in my life!
I wonder – where will this lead?
Possibly it will lead nowhere. And yet I can’t help but suspect that there is some greater connection between the act of speaking and a fluency with words themselves. I may of course be entirely wrong about this – it might be equally true that writing the words down, whether on a keyboard or longhand, is just as strong and instructive a pathway to the deeper art of finding words, summoning them up from memory, and using them in sequences that are both communicative and artistic.
And yet it cannot be denied that the language that we create when we speak will perforce differ from the language that we create when we silently sat down words on paper or screen. The language as it is spoken is the oldest form of language, saving only for the gesture – if we can really say a selection of gestures without any verbal language is a language at all.
I find, for example, that even as I dictate old essays from 30 years ago, I will sometimes change the phrasing because I’m sure that the voice recognition program will misinterpret what I say.
This is something that I realized quite some time ago, when I was studying foreign languages. We never really hear all the words that someone else is saying. Instead, we hear certain key words and then we assume that these keywords are being used in their standard, phrases – the phrases in which the words almost always occur – and we also tie in these few words that we hear to the context of the subject which we are discussing. This is what I concluded was the most difficult thing for anyone to get past when listening to the conversational speech of native speakers in a foreign language. They just speak too fast! It seems impossible to hear every discrete word – that’s why I came to the conclusion that normally we don’t hear every single word, but maybe one out of four or one out of three words. But since we know that usual phrases those words are found, and since the conversation follows normal guidelines, we can function quite adequately – if we are fully-conversant or fluent in the language.
When I was writing longhand, I always took pride in tweaking my grammar and my phrases, so that they would be a little bit unusual, a little bit distinctive. This is all well and good, when the words are there to be read, and they can’t be misunderstood, and every single word can be read, and the reader, if thrown off track for a moment, can always back up and reread a phrase or two.
But in speaking these same thoughts, it works to a disadvantage of communication to alter the typical patterns of speech and phrases and clichés. The voice-recognition software never fails to understand my speech so often as when I am trying to speak an unusual phrase or grammatical construction. Whenever I come upon such a phrase in these old writings of mine, I have to choose between two paths to achieve a near-perfect recognition on the part of the program.
The first path is to slow down. I have to speak each word separately and distinctly, and not as part of any phrase. This works very well when the words are polysyllabic and therefore have far fewer other words that sound even remotely similar. The second path is to rewrite as I dictate, and rephrase these unusual and distinctive phrases into more common and typical phrases, such as any other native speaker of my language would recognize immediately as the first choice of expression of the same idea. Dragon NaturallySpeaking is not, of course, as intelligence or flexible as another human being, therefore I have to be even more careful – it’s rather as if I were speaking to someone slightly hard of hearing, or someone who knows the language very well, but is not completely fluent in it.
This means that if I go on to compose directly into the voice-recognition software, I will probably (by a process of reinforcement) tend to compose phrases that are simpler and more common and expected that I would if I composed on a keyboard or by hand.
It also means that my relationship to language is now being shaped primarily (at least actively – i.e., in the process of composition) through my tongue and mouth and vocal chords.
(For example – if I just said the word, ‘cord’ the program would not recognize the word as, ‘chord’ – but if I say the phrase, ‘vocal chords’ the second spelling is what comes up first. And interestingly enough, if I say instead the phrase, ‘vocal cord’ the first spelling comes up first! Knowing this, I am being guided and ‘trained’ by the program to speak in terms of phrases more than individual words. And the phrases had damn well better be common phrases such as most people use – because those of the phrases that have been built into Dragon NaturallySpeaking by its creators.)
(Composed by dictation Tuesday, October 7, 2008)
