Re: Optimum number of symbols

Christophe Grandsire <christophe.grandsire@...> wrote:

>En réponse à "Mike S." <mcslason@...>:
>
>>
>> It's only problematic I think if you consider it very important
>> to encode morphemic information in addition to, and on top of,
>> the phonemic information.  If morphemic coding starts getting
>> too complex, I feel it's always a viable option to back off
>> and go totally phonemic.  If different morphemes are not
>> distinguished in the speech, or if the same morpheme
>> alternates, why should we fret so much when the same thing
>> occurs in writing?
>>
>
>Because speech offers more clues for disambiguating than writing. During a
>communication, more than 50% of the message is actually transmitted non-
>verbally. Tone, prosody, facial features and body language account for a
>disambiguating context that writing cannot offer. Hence the need to
>disambiguate in writing what is not disambiguated in speech.
Non-verbal communication is mostly register, not semantics.  I say
something like "oh, I had great time"  and you know from my tone
whether I am serious or sarcastic.  No matter how I change my tone
though, you will not hear it as "I had a grate time" because that
interpretation makes no sense.  Nearly all the time, context alone
establishes literal meaning;  tone expresses the speaker's attitude
towards his words.  (Note the phrase "literal meaning": it's the
meaning that is most obvious given the letters alone.)

Thus, although a great deal of information is lost to writing,
it is typically not the sort of information that could be used
to successfully disambiguate a homonym, say, "paws" and "pause".

A special remark on prosody--that indeed is important.  The best we
can do here pretty much though is use punctuation marks.  Spelling
has no effect on prosody that punctuation also couldn't accomplish.


>The only
>disambiguating context writing can offer is the previous sentences written
>before, and it's far from equating the context present in speech.
I would argue that context is more salient and more sufficient in written
speech than in spoken.  More salient, because the reader has no other
clues to go by, and more sufficient because writing is normally composed
with more care than spontaneous speech.


>In short,
>actual speech can afford more ambiguities than writing, since it has more
ways
>to disambiguate (it's well-known that a single word can be a fully
unambiguous
>declaration in speech, while the same word, even accompanied by
punctuation,
>will never give a fully unambiguous message in writing, since this word is
not
>accompanied by the prosody, tone and posture of the speaker at the moment,
and
>often also lacks the spatial and temporal context when it was uttered).
I think I know what you mean, but examples would be helpful.


>In short, writing needs to disambiguate more than speech, which is why
purely
>phonemic systems are not only less advantageous but also not advisable.
Not advisable?  I guess someone forgot to tell the Greeks and Romans.
They didn't even use punctuation marks or spacing, the fools.


>And I
>think it's something people know nearly instinctively. After all, why
didn't
>the Japanese throw away all the kanji while their kana are far enough to
write
>everything they want to write?
That's like asking why English orthography isn't thrown away. It's
the historical momentum, nothing more.  It doesn't mean that English
orthography isn't a disaster.


>Simply because the burden of ideograms is
>compensated by the possibility of disambiguating homophonous words which
are
>easy to disambiguate in speech thanks to things like prosody, body
language and
>context, but not in phonemic writing where all those things are absent and
the
>context of the previous sentences may be far from enough.
I find this a little far fetched when applied to morphemic ambiguity.
In order for me to accept this, you would have to explain in a little
more detail how body language could disambiguate a pair like "aisle"
and "isle" that context had already failed to disambiguate.


>Add to that a much
>faster reading and writing, and suddenly the use of ideograms because more
>optimal than a purely phonemic system.
I too have a sense there could be enhanced readability in an ideogrammatic
script, but dispute that it is worth the other costs.  Ideograms are
far too complex and too difficult to learn.

You may have some sort of case buried in this post, but as it stands,
I can't imagine what it is.  Perhaps some concrete illustrations
would illuminate the matter.


Regards

Replies