Re: YAPT

--- Tim May <butsuri@...> wrote:

> Gary Shannon wrote at 2005-01-04 08:08:15 (-0800)
>  > More evidence for my contention that the exact
> nature
>  > of mouth noises is probably the least important
> aspect
>  > of language. Just listen to a four way
> conversation
>  > between people from Syndey, Bombay, Boston and
> South
>  > Carolina and you'll notice that they all
> understand
>  > each other just fine even though they have
> practically
>  > no mouth noises in common.
>  >
>
> Gary, you're not really saying anything substantive
> here.
I think what I'm trying to say is that for a word to
be mutually understood the mouth noises have to lie
somewhere within a rather broad region on the map of
possible sounds, rather than at some precise point
within that map.  The size of those regions might be
interesting to investigate. But I believe that the
size of the region is a significant percent of the
size of all possible regions, and that this percentage
scould be used to measure the relative magnitude of
the changes.

> You're just
> indulging your personal prejudice against phonology
> again.
Quite true. Once a year or so I like to take it out
and wave it around just for the fun of it. On the
other hand, a conclusion presented without the
suporting arguments does sound like a prejudice,
doesn't it. I've just never taken the time to
articulate my arguments, even to myself.

> Without
> specifying any standard of "exact" one could as
> accurately claim that
> the exact nature of speech sounds is the _most_
> important aspect of
> language, as it is the only thing which enables us
> to distinguish
> between two utterances.
I think if we were to compare all intelligible
pronunciations of a particular vowel (for example) in
a particular word and then map that region onto the
space of all possible pronunciations we would find
that that the region occupied something in the order
of magnitude of 20% to 50% of the all possible
pronunciations of each vowel. This is just a wild
guess, of course, but I think that in an appropriate
context the pronunciation of a particular vowel could
wander at least 25% of the available sound space and
still be understood within the context of that word.

Orthography, on the other hand, has nowhere near that
flexibility.  The space of possible 5-letter words,
for example, is huge while the space of understandable
alternate spellings of a given word is miniscule by
comparison.  Given, for example, 2 dozen alternate
spellings for a single word and 11,881,376 possible
5-letter combinations, that word is free to wander
only 0.002 percent of that space.  Far more
restrictive than the 25% freedom for each vowel sound
in the word as pronounced.

> Defining the kind of
> variation in
> pronunciation which will not prevent comprehension
> is not trivial.
Clearly. It sounds like a formidable problem.

> Even if you could define it, you'd have no standard
> by which to
> compare changes in pronunciation to changes in
> syntax or any other
> domain.
I think valid measures could be devised based on space
occupied divided by space available as in the example
above.

> Since you cannot judge whether a difference
> in one domain is
> greater or smaller than a difference in another, you
> have no grounds
> on which to say that comprehension is more sensitive
> to changes of one
> type than another.
>
I believe such grounds could be established.  I would
suspect, off the top of my head, that phonolgy has the
greatest freedom, changes most readily, and is,
therefore, the least signigifcant. After all,
Bostonian English and Bombay English are both called
English.

Grammar is next in rank, has somewhat less freedom,
and changes somewhat more slowly. "Them folks is nice"
is perfectly understandable. But I don't think a
purely isolating variant on English would be called
"English" in practise. It might, however, be called
Iso-English, linking it closely with English.

Vocabulary is the most rigid and changes most slowly.
A language which shared the grammar of English but had
a different vocabulary would certainly not be called
"English".

The standard disclaimer is that I have no formal
education in linguistics and am applying nothing more
than common sense and some skimpy mathematical
reasoning. There's a good chance I'm wrong, but I
kinda like my line of reasoning and plan to stick with
it any way. :)

--gary