[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [lojban] Named multiples
On Wed, May 19, 2010 at 5:12 PM, Oleksii Melnyk <lamelnyk@gmail.com> wrote:
> 2010/5/19 Jorge Llambías <jjllambias@gmail.com>
>
>> Can you give an example? ... Are you talking about human parsing?
>
> Yes. Let's take an example from NORALUJV.txt. I've spotted "backemselrErkru"
> (: it is long enough to either got out of air in the middle of it, or just
> need to look into the dictionary to recall the next rafsi, whatever :).
A human parser will either recognize the word, in which case any minor
errors in pronunciation will probably be ignored, or (most likely)
they won't recognize it, in which case they will say "ki'a" whether
there were any errors in pronouncing that monster-word or not.
> Any
> pause after the vowel and the stressed "lrE" leaves us with "rkru" as the
> last rafsi, so here was some error. Any pause after the consonant will alert
> us about "no LA before the cmene"(not morphology level, but easy/close
> enough for humans).
There are many contexts in which cmevla can occur, not just after LA.
You could be talking about the word "zo backemselrerkru" for example.
> Pause after the "CVV" rafsi (not in an example), will be
> noticed as "no stress in previous word". Only in "la backemselrErkru" the
> pause between the rafsi's will go unnoticed. In the cmevla, after the
> required LA, we do expect the sequence of names, so breaking one into
> several will cause almost no harm, we'll just glue them together up to the
> last consonant ending word.
I'm not sure I follow what you are saying. LA can be followed by more
than one cmevla, so if you say for example: "la pip backemselrerkru",
and you pause somewhere after a consonant of the lujvo, it will be
taken (at least by a mechanical parser) as a cmevla.
> So, we are almost immune to the extra pauses inside the long words. Now,
> without the required LA, we'll get:
>
> ba.ckemselrErkru - can fail
> bac.kemselrErkru - OK
> back.emselrErkru - can fail
That's a weird place for an unplanned pause.
> backe.mselrErkru - can fail
Another weird place to stop.
> backem.selrErkru - OK
> backems.elrErkru - can fail
Another weird place to stop.
> backemse.lrErkru - can fail
Another weird place to stop.
> backemsel.rErkru - OK
> backemselr.Erkru - can fail
Another weird place to stop.
> backemselrE.rkru - can fail
Another weird place to stop.
> backemselrEr.kru - OK, can eat the next word
> backemselrErk.ru - OK
Another weird place to stop.
> backemselrErkr.u - OK
Another weird place to stop.
> Note, that all the "fails" only "_can_ be"; if there are the otherwise
> _allowed_ pause after the entire word, it will parse, giving us 2 wrong
> meaningful chunks.
I still don't get what the point of this is.
The only realistic unintended pauses are:
ba.ckemselrerkru
bac.kemselrerkru
backem.selrerkru
backemsel.rerkru
backemselrer.kru
The last one the least realistic, because of the stress.
All the others are not reasonable places for unintended stops.
The human hearer will have to decide whether or not to take any such
pauses seriously based on the resulting meaning.
>> The exact same problem exists with or without the change. The
>> change has no relevance to word parsing.
>
> So, the change affects an error detection. If the humans all were the
> reliable electronic devices with the error detection codes in the
> communication channel, that would be irrelevant. However they are not, and
> the language was meant to be "error detecting communication channel" for
> them.
Are you saying that with the current grammar the human will say: "lo
backem.selrerkru" is ungrammatical, therefore I will attempt a
possible correction to "lo backemselrerkru, which fixes the problem",
whereas with the change the human will say "lo backem.selrerkru" is
grammatical, therefore I will not attempt a correction, even though
what I'm hearing makes very little sense". Is that the point?
>> Why would a text be full of cmevla? As you say, cmevla are a
>> cumbersome type of word, so they don't blend well with normal Lojban
>> words. Simplifying their syntax would not make them morphologically
>> prettier, it would only make the syntax simpler.
>
> I do like the idea. I just looking for the bad consequences. The result of
> the change would be the wider usage of cmevla. Otherwise, there are no
> reason to make it simpler.
That's not the motivation though. If we don't want cmevla, we should
remove them from the language, not complicate their grammar
needlessly. (Not that their grammar is too complicated, just more
complicated than what it needs to be.) The motivation is not to use
more cmevla, but to allow things like "la cmalu djan" for a name like
"Little John". There is no reason why the change should encourage more
cmevla, since anything that can be said with the change can already
be said in some other way without the change.
> Trying to read them aloud would push "the live
> language" towards toki pona (as the best case).
I don't understand what you mean by that. Trying to read what aloud
would push the language towards toki pona?
> "." is OK in writing. Not so in speech. We do need to think/breath/rest/etc.
> sometimes. That sounds just as ".......". So, using "." as the syntax marker
> looks bad for a human-spoken audio-video isomorphic language.
Long words are bad for Lojban, I agree, whether they are lujvo,
fu'ivla or cmevla. But the proposed change doesn't really affect that.
mu'o mi'e xorxes
--
You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to lojban@googlegroups.com.
To unsubscribe from this group, send email to lojban+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban?hl=en.