Re: Lojban Orthography (has been several other things)

["Mark E. Shoulson" <mark@xxx.xxxx]

>Date: Wed, 22 Sep 1999 11:49:56 -0700
>From: Ivan A Derzhanski <iad@math.bas.bg>
>Organization: Institute for Mathematics and Computer Science
>CC: lojban@onelist.com
>
>From: Ivan A Derzhanski <iad@math.bas.bg>
>
>mark@kli.org wrote:
>> [OBTW, take everything I [...] will say with the following proviso:
>> my exposure to Devanagari is (almost) strictly via Sanskrit [...].]
>
>Whereas mine is predominantly via Hindi (that being the only
>DN-written language wherewith I have any practical familiarity).

Nifty; so we complement each other (hey, nice shoes, Ivan).

>> Thing is, the ligatures aren't so neat and tidy as we'd hope. There isn't
>> really every possible conjunct ligature; sometimes you fall back on virama
>> to make them. Especially when dealing with a foreign language (like
>> Lojban) which may construct clusters alien to Indic languages.
>
>Indeed. The ugly thing is that DN has several different strategies
>for handling consonant clusters (both letters written in full with a
>virama under C1, C1 written with a diacritic for C2=<r>, C2 written
>with a diacritic for C1=<r> or a homorganic nasal, a more or less
>transparent ligature, an unanalysable letter for the whole cluster),
>and while that may be linguistically motivated in Sanskrit, it would
>make no sense in Lojban.
>
>Btw, would there still be an unwritten vowel, and if so, which one?
>{a} (Sanskrit-like) or {y} (Hindi-like)? The latter makes more sense
>to me, since in Lojban {y} is a better candidate for special treatment,
>being structurally different from the other vowels, though it's rare.

Mmm. NEED there be an unwritten vowel? Maybe we can just decree that all
vowels shall be written. Makes the virama unnecessary, but then the period
in lojban is also not strictly *needed*, and obviates the assymmetry in the
existence of conjunct ligatures for SOME but not all consonant-clusters.
We'd use the long-A for {a}, of course. Possibly even use long-form vowels
for all of them, and using the visually salient short-i for {y}. Just a
thought.

I note that in the Sanskrit I learned, although the vowel was spelled {a}
in transliteration, it was pronounced something very close to Lojban {y}
(even in tapes from learned and practiced speakers).

>> It'll look like really strange text no matter what, since you
>> pretty much HAVE to break after each word in Lojban, unless you mark
>> stress, and that looks *weird* in Devanagari, with all those tiny words.
>
>It doesn't look weird at all; it looks like Hindi, where words of
>one, two or three letters (plus diacritics) are very frequent.

(*flips through some Hindi texts*)... Yes, good point. I got too used to
those looooong strings of letters Sanskrit produces.

>> The vowel-hiatus gets weird with diphthongs too... [...]
>> And the rising diphthongs even worse. Falling diphthongs,
>> SOME are written as vowels (ai, au), some are not (ei, oi).
>
>The falling diphthongs are not a problem; scores of Indic, Dravidian
>etc. languages have augmented the system with their own conventions
>for writing vowels that Sanskrit doesn't have. How about a combination
>of the diacritics for <e>/<o> and <i>? Not at all sure how one would
>handle the rising diphthongs, though.

OK, my limited exposure taking its toll again. Hmm, since DN vowels
are pronounced *after* the consonant they're written on, it would almost
make more sense to use the <i> diacritic on full-mode <e>/<o>. And it
would look no worse (better, even, since <e> and <o> diacritics are more
prone to confusion, differing as thet do only in a vertical stroke). 

>> Erk, we can't even use the existing semivowels for rising diphthongs,
>> since the [w] is also what you'd have to use for [v].
>
>Not necessarily; you can recycle <bh> for {v} (that's only natural,
>since <ph> would probably become {f}) and then use <v> for prevocalic
>asyllabic {u}.

Eh, I suppose. Would take some getting used to, but then the whole thing
would.

~mark