Received-SPF: pass (google.com: domain of daeldir@gmail.com designates 2a00:1450:400c:c03::22f as permitted sender) client-ip=2a00:1450:400c:c03::22f;
Date: Sun, 9 Jun 2013 08:53:47 +0200
From: Daeldir <daeldir@gmail.com>
To: lojban@googlegroups.com
Subject: Re: [lojban] Number of unique symbols needed to write lojban
Message-Id: <20130609085347.30689d8aded469d57d5a7967@gmail.com>
In-Reply-To: <C87E7070-9168-4F82-9546-D08CEF7A114B@aol.com>
References: <C87E7070-9168-4F82-9546-D08CEF7A114B@aol.com>
Mime-Version: 1.0
Reply-To: lojban@googlegroups.com
Precedence: list
Mailing-list: list lojban@googlegroups.com; contact lojban+owners@googlegroups.com
Sender: lojban@googlegroups.com
Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: quoted-printable
X-Spam_score: -0.1
X-Spam_score_int: 0
X-Spam_bar: /

On Sat, 8 Jun 2013 13:36:41 -0500
Terry <tsofian@aol.com> wrote:

> How many unique symbols would be needed for all sounds and punctuation in=
 written lojban?

Do you plan to do some stuff with lojban alphabet?

I did some work on lojban symbols a while ago to pass the time (when I had =
some=85). An HTML page to compare all proposed alphabets for lojban. Curren=
tly, there is latin alphabet, cyrillic alphabet, tengwar mode, and I added =
braille, just to see what it gives.

I wanted to add hiragana, but the conversion algorithm I found was=85 A lit=
tle too complex/undefined for me to work on it. I also thought about a hang=
ul system, as hangul has some interesting linguistic properties (so says wi=
kipedia), but=85 I don't know hangul, so I didn't spend time on it neither.

I also tried to design an alphabet from scratch, with sound features in min=
d but=85 Well, Tolkien did a better job, mine was useless compared to tengw=
ars (even if tengwars are more complicated that needed for lojban, using a =
subset works fine).

For now, stressed characters are in uppercase (it doesn't recognize accents=
). Unknown characters are replaced with =93..=94 and numbers are replaced w=
ith their name (1 is =93pa=94, 45 is =93vomu=94=85).

While that doesn't answer your question, I can say that for that work, I us=
ed 27 lowercase characters (including punctuation and *space*).  Uppercase =
was checked only for latin letters, not the apostrophe (so, I don't recogni=
ze the "h" as a letter in my program), which gives 23 more symbols. So, 50 =
symbols are used (counting space, no =93h=94).

However, at least in the tengwar mode, there are more (or less) letters: 12=
 vowels (=93full=94 vowels, which are a symbol, and =93small=94 vowels, whi=
ch are like accents), 18 consonants, one punctuation (=93'=94 is counted as=
 a letter, and with tengwar, the comma is useless). Because each full lette=
r (six vowels, 18 consonnants) can have an =93accent=94, that is, can be fo=
llowed by a small vowel, we have 24=D77=3D168 letters, (we count 7 =93vowel=
s=94, one being =93no vowel=94), that, time two, for stressed syllables: 33=
6 symbols are used in the tengwar mode. However, it is really 32 symbols, s=
ince each of the 336 final symbols are from one to three superposed symbols=
 (24 full letters + 6 small vowels + stress + space). So, that is less than=
 the 50 symbols used in the latin alphabet. Also, we could think about the =
way tengwar are designed, and decompose them in even less symbols (each one=
 being a feature of the represented sound).

Here is the code I used to =93classify=94 each latin letters:

> maps.latin =3D {
>   letters:      "abcdefgijklmnoprstuvxyz.,' ",
>   vowels:       "aeiouy",
>   consonants:   "bcdfgjklmnprstvxz",
>   punctuations: ".,' " // (=93'=94 should be a consonant and not a punctu=
ation, but, he!=85)=20
> };

What matters for =93counting=94 the symbol is only the =93letters=94 field.=
 And adding stress (uppercase characters).

You can give it a try here:

http://daeldir.ninm.net/lojban-alphabets/alphabets.html (needs javascript t=
o work, and a recent navigator, at least for displaying tengwar)

I wanted to make the code more readable before releasing it, but as I don't=
 have time, and will spend the small amount I have on translations rather t=
han on this code, I guess I won't wait more=85

I have another page with my attempt at a new alphabet, but it is not on the=
 net, and not very interesting.

That may not answer totally your question, but it shows that the number of =
symbols needed to write lojban can vary from 32 to 336, depending on how yo=
u analyze the language and the alphabet (are =93a=94, =93e=94, =93i=94, =93=
=E4=94, =93=EB=94 and =93=EF=94, six symbols, or four? Is the dot really ne=
eded, as it is said (at least in the wave lessons, I don't remember for the=
 CLL) that it is always optionnal?).

Also, keep in mind that I'm a beginner=85 But you already have better answe=
rs from more experienced people. Mine is a=85 Complement.

mu'o mi'e la .daeldir.lurni'acadz.

--=20
You received this message because you are subscribed to the Google Groups "=
lojban" group.
To unsubscribe from this group and stop receiving emails from it, send an e=
mail to lojban+unsubscribe@googlegroups.com.
To post to this group, send email to lojban@googlegroups.com.
Visit this group at http://groups.google.com/group/lojban?hl=3Den.
For more options, visit https://groups.google.com/groups/opt_out.