Print

Print


Despite the stated non-IAL-ness of at least the original Loglan,
I am somewhat surprised to hear it spoken of now as a credible IAL ...
 
Robin Turner wrote:
 
> If by species, you mean plant and animal species, if it's common, make =
up a lujvo
> (compound word), if not use a fu'ivla (loan word) - the same goes for a=
ny large and
> specialised area.  A fu'ivla contains a prefix indicating the broad sco=
pe of the word,
> so people who haven't come across it before can still have a stab at th=
e meaning.
 
OK, that's a mechanism, and it is defined:
http://www.animal.helsinki.fi/lojftp/reference-grammar/chap4.html
But I can't stop and go through a 5-step algorithm to say "spaghetti".
This is the difference between theory and practice, I think.
 
> > About the 3-letter abbreviations of gismu, the combining forms;
> > why is there not *one* standard form for each? Are they needed at all=
?
 
> I don't think there can be one form (e.g. "take the first three letters=
 of the gismu")
> as we would end up with a lot of gismu having the same combining form. =
 Yes, they are
> necessary for forming lujvo - otherwose the words would just get too lo=
ng.
 
That seems backwards. Gismu are defined to have only 3 significant letter=
s,
which should be the official form; then generate the longer form for
noisy-room-talk. The 3-consonant-root method of Arabic could work nicely,
giving at least 8000 basic roots, much more realistic for an auxlang.
Then fill in vowels using "vowel harmony" or something ... Just an idea.
Otherwise as it is, recognizing/creating compounds looks too difficult,
more like a puzzle than a conversation.
 
> The problem with English is that it is so metaphorical.  I think this i=
s not an
> inevitable consequence of an isolating language (though I could be wron=
g) but of having
> an isolating language which presses too many lexical words into grammat=
ical use.
 
I definitely disagree here. Metaphor isn't the problem, irregularity is.
And the spelling, specifically the vowels. All natlangs are doomed to
succumb to these evils, over time. An auxlang, if it remained true
to its purpose, would not; e.g., the Lingua Franca did not, and I only
think it is odd that LF did *not* quite survive the 19th century.
 
> Synthetic languages tend not to have this problem, because even if affi=
xes derive from
> lexical words, their morphology has changed - very few Turks realise th=
at the
> progressive aspect suffix -yor- comes from the old Turkish root yoru- ,=
 meaning "walk"
> (modern Turkish y=FCr=FC-).
 
If even a relatively regular natlang such as Turkish, with such transpare=
nt
derivations, is still opaque to its speakers, how can one hope to avoid
use of metaphor rather than grammaticalized or lexicalized logic?
 
(Do they also use "sit" or "stand" in some kind of stative sense?)
 
> Lojban atempts to get round the problem by having a large number of cma=
vo (structural
> words) which are derived from gismu (root words) but have a different f=
orm, so you know
> you're talking grammar, not content.  For example, the cmavo for the pr=
esent tense,
> {ca} comes from {cabna}, meaning "now".
 
Hmm. I think metaphor defeats logic, in at least two senses.
 
> I think this is less of a problem in an isolating language, especially =
when a large
> proportion of grammatical words are derived from lexical words.  At lea=
st with Lojban,
> only knowing a few grammatical words won't make your sentences wrong, j=
ust less
> detailed (I certainly don't know all the cmavo, and I don't intend to l=
earn them all in
> the near future).
 
Well, I heard this is a 5-year testing period for Lojban.
May only the fittest features and words survive.
 
> Lojban has no punctuation as such - some Latin punctuation is used for =
phonological
> purposes e.g. a full stop is a pause.
 
I meant the whole issue of suprasegmentals. People use pauses and
stress, indicated by written punctuation, to mark a word or phrase.
This kind of thing, "na'e zei .a zei na'e zei by. livgyterbilma";
it does look like a liver disease, but not in a nice way.
 
> > The arbitrary argument structures are inferior to natlang case.
 
> Argument structures are not all that arbitrary - they usually follow a =
nominative -
> accusative - dative - ablative pattern where relevant.  I must admit I =
was very
> sceptical about place structure when I first encountered Lojban, but no=
w I find it
> fairly natural. Natlang case can also be arbitrary - there's no real re=
ason why in
> Turkish "hit" takes the dative but "beat" takes the accusative, or why =
"be angry" takes
> the dative but "be afraid" takes the ablative.
 
Even assuming a strict NOM ACC DAT ABL order, which would be an improveme=
nt,
one still needs particles to sort them out. This is more difficult than
natlangs, so it can't be right for an auxlang. Maybe the error is
in thinking that one language can do everything optimally.
 
> > Attitudinals (disjuncts) as a separate morphological class seem wrong.
 
> Actually, I find this one of the most attractive and practical features=
 of the language
> - I even throw attitudinals into my English postings.  But there's no r=
ule that says
> you have to use attitudinals for affect or judgement - it's just more c=
oncise, and
> somtimes more precise.
 
OK.
 
> I find very little to be grammaticalised.  by this count, Chinese would=
n't be isolating
> either.
 
I don't understand this statement. Chinese is problematic (from my point
of view) in its writing and homonymy, not in its isolating (non-)morpholo=
gy
and very (relatively) logical syntax. Full word, empty word; nice.
 
> > Precision is advertised, but every compound is imprecise
> > and yet required by having so few roots (gismu).
 
> As I said, this is why we have lujvo.  If I say {kurji mikce} it is amb=
iguous, since it
> just means "taking-care-of-kind-of medic".  But the lujvo {kurmikce} is=
 precise - it
> means "nurse" (or to be more accurate, it will do if it gets into the L=
ojban
> disctionary - this is one I made up myself).
 
If it were precise it would not need to be approved or listed.
The only language I have seen that delivers on the promise of precision,
is the translation inter-language designed by Rick Morneau.
I don't think it is really normal-human-learnable, and it does not
aspire to be an auxlang. If not for that lone example, I would have
said that the goal of precision is chimerical, unattainable, and
quite silly; of course, my saying so means nothing anyway.
 
> English doesn't distinguish between IF and IFF, or between OR and XOR -=
 Lojban does.
 
Computer geeks say "xor" and "ior"; "iff" can't happen in a procedural la=
ng.
 
> Lojban is also more precise about causation, distinguishing between phy=
sical causation,
> motivation, justification and logical implication.
 
Which never seems to happen in real life. Sorry, but when one is
haggling over the price of a fish, it just doesn't matter!
 
> I'm happy with Lojban orthography because (i.e. (mu'i}
> - with motivation!) I don't like capitals much, but I understand that s=
ome people may
> not like it.   There is also an Elvish alphabet for Lojban, but I don't=
 think it's been
> used much.  Personally, I'd like to see a specific Lojban alphabet deve=
lop, but that's
> way in the future, and probably won't happen while most people are stil=
l using ASCII.
 
{i'd prefer to mark sentences this way (and parentheticals like so) but
it doesn't matter much} {a prettier alphabet might help, though}
 
> gleki je prami =3D be-happy-and-love
>=20
> as opposed to
>=20
> gleki prami =3D happy-type-of-love
 
But all natlangs can do this without difficulty. The more complex
expressions look ugly, no matter how said. People naturally avoid them.
"That girl has a school, such a pretty and little one ... let's burn it."
 
> I think once you've learnt the basic bridi (predicate) structure, place
> conversion and forking the bridi with {gi}, you have all you need synta=
xwise.  I don't
> bother with the really flashy stuff like termsets and bridi-tails.
 
With more work, it could become the world's best auxlang.
Whether it could remain a loglang, I doubt. For now, I think
Ido is a far better auxlang, and Glosa could become so.