amnfn at well.com
Fri Jun 13 13:29:09 UTC 2008
Language is a kind of phenomenon, I agree. It has concrete manifestations,
but it itself is abstract. It can be instantiated in many different ways:
auditory, visual and tactile. While its origin is biological, in the sense
that living entities are where it first arose, today there are non-animate
entities that are pretty good at simulating it. The other day I called
directory assistance to get the number of Staples in Rolla, Missouri. A
computer generated voice tried to ask me what listing I wanted, and then
it had trouble understanding what I said. "Do you want the listing of
Staples in Sioux Falls, South Dakota?" I tried to enunciate as clearly as
possible: "No." At this point, the AI construct switched me over to a real
Despite the difficulties with speech recognition, it works amazingly well.
We were on the same general topic, it understood I was looking for a
Staples outlet, but it couldn't understand "Rolla, Missouri" as pronounced
by me. More significantly, I understood every word it said. To me, it was
like any other conversation -- some understanding, some misunderstanding,
and a cooperative effort to get the information across.
Soon AI constructs will do even better, and we won't be able to tell them
apart from real people. The Turing test is a wonderful way to determine
what language really is.
By the same token, Alex the parrot, with his avian brain and quite
distinct articulatory system, was able to have meaningful conversations
about colors, numbers and feelings, in English.
I have had the experience of a common chimpanzee letting me know that the
bonobo sounds he was listening to were in fact English. It didn't sound
like English to me, but Bow could recognize the bonobo sounds that Kanzi
used to stand for the contrasting phonemes of English.
There is fMRI data to show that autistics have different activation
patterns from non-autistics in lexical retrieval tasks, even though their
performance is comparable.
When I say that the language code will not be cracked through close
attention to brain patterns, I am not motivated by a disdain for hard
science or biology. All the hard evidence indicates that language is a
portable code which is very flexible as to its particular instantiations.
There is more than one way to produce language correctly. It's not how
you produce it that's important.
The meaning is in the contrasts.
More information about the Funknet