DomainKey-Signature: a=rsa-sha1; c=nofws;        d=gmail.com; s=beta;        h=received:message-id:date:from:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references;        b=PWQUiRuwTDxGU3GXwL0SKyT26QPJueN+s/PZYu8HuhtcnNVKemDizn2aGnUAfmqh3Sni6ZZ3VD35GI7o5W0XXFokw4Ec23/wukrAc51lqsgddYoD71aUZQYl5f2hw+1qlWWH0c/zAIVQ0JchOJUjaQp+YIy513BpXYBmAmH9dCQ=
Message-ID: <925d17560707160809j7cd27c87ve3506d54831c66f7@mail.gmail.com>
Date: Mon, 16 Jul 2007 12:09:34 -0300
From: "=?ISO-8859-1?Q?Jorge_Llamb=EDas?=" <jjllambias@gmail.com>
To: lojban-beginners@lojban.org
Subject: [lojban-beginners] Re: double letters
In-Reply-To: <127184.51719.qm@web56406.mail.re3.yahoo.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
References: <925d17560707140740j4b9c3422s247213f558703a38@mail.gmail.com>	 <127184.51719.qm@web56406.mail.re3.yahoo.com>
Sender: lojban-beginners-bounce@lojban.org
Errors-to: lojban-beginners-bounce@lojban.org
Precedence: bulk
Reply-to: lojban-beginners@lojban.org
Content-Length: 5200

On 7/15/07, Nathaniel Krause <nathanielkrause@yahoo.com> wrote:
>
> Naturally, "ea" and "oe" must be distinguished not only from "eia" and "oue"
> but also from "e.a" and "o.e". I've no doubt that someone, somewhere can
> make this three-way distinction, and I'll take your word for it that you can
> do it (although, like Matt, I'm interested in hearing how it sounds), but I
> strongly suspected it would seem problematic to a lot of people in the
> world.

I don't have a working microphone at the moment, but I will try to get
one and make some recordings.

In Spanish, the equivalent of ea and eia are contrastive, for example in the
minimal pair "sea" and sella". I don't think there is any word with oue,
but we can still have the contrast with oe for example in "o eso" vs.
"o hueso". In Spanish no glottal stop is inserted between the vowels, and
inserting one makes it sound quite different.

In any case, there are many things in Lojban phonology that are problematic
for a lot of people in the world, so having this distinction (which I don't
advocate) would not seem to be specially noteworthy.

> In jbovla phonotactics (by which I mean, the phonotactics of gismu,
> lujvo, and cmavo, excluding cmevla) two vowels can appear in a sequence
> precisely because one of them becomes a semi-vowel -- there's no precedent
> for having a sequence of two or more vowels where none of them is a
> semi-vowel.

That's right. And in fact there is never any sequence of more than two vowels
in those core word forms, and also the iV and uV diphthongs only occur without
a preceding consonant.

> (for instance, judging by Wikipedia, it seems like the Albanian
> word for Albania would be closer to {ctcipyrias} than {ctiipyrias}; I'm not
> sure how many people would argue that the former is valid)

Both would be valid under my proposal, but only because they
are cmevla. Neither {ctciperia} nor {ctiiperia} would be valid fu'ivla
(I give a sketch of the proposal below in case anyone is interested).


> I definitely agree that English is particularly permissive with final
> consonant clusters (I remember particular incident in which I found myself
> struggling to pronounce the cluster at the end of "sixths" correctly).  But
> this seems like all the more reason to place some restrictions on what we
> consider valid, since the English-speaking majority of Lojbanists will
> naturally have a tendency to produce combinations that other people find
> difficult to pronounce.

Yes, but it's hard to say where to put the limit. Just a single consonant?
But then what about -rt., -rk., -nd., -st.? And then what about -jd., -mf.?
-rks.?, -rst.?

Here is a sketch of Lojban phonotactics as I see it.

A (vocalic) syllable consists of three parts: onset, nucleus and coda.

There are exactly 10 valid nuclei: a, e, i, o, u, ai, au, ei, oi, y.

There are exactly 18 valid codas: b, c, d, f, g, j, k, l, m, n, p, r,
s, t, v, x, z and null.

There are exactly 137 valid onsets:

., ', i, u,
b, c, d, f, g, j, k, l, m, n, p, r, s, t, v, x, z,
tc, ts, dj, dz,
br, cr, dr, fr, gr, jr, kr, mr, pr, sr, tr, vr, xr, zr,
bl, cl, fl, gl, kl, ml, pl, sl, vl, xl,
zb, zd, sf, zg, sk, sm, zm, sn, sp, st, zv,
jb, jd, cf, jg, ck, cm, jm, cn, cp, ct, jv,
zbr, zdr, sfr, zgr, skr, smr, zmr, spr, str, zvr,
jbr, jdr, cfr, jgr, ckr, cmr, jmr, cpr, ctr, jvr,
zbl, sfl, zgl, skl, sml, zml, spl, zvl,
jbl, cfl, jgl, ckl, cml, jml, cpl, jvl,
bi, ci, di, fi, gi, ji, ki, li, mi, ni, pi, ri, si, ti, vi, xi, zi,
bu, cu, du, fu, gu, ju, ku, lu, mu, nu, pu, ru, su, tu, vu, xu, zu.

So there are exactly 137x10x18 = 24660 valid vocalic syllables.

In addition there are 64 consonantal syllables, that consist of any of
the 17 consonants followed by one of l, m, n, r (except ll, mm, nn, rr).

The phonotactic constraints for syllable combinations are as follows:

1- A syllable with onset ' must be preceded by an open syllable.
2- A syllable with voiced coda cannot be followed by one with
unvoiced onset, and viceversa.
3- A syllable with sibilant coda (s, c, z, j) cannot be followed
by one that starts with a sibilant.
4. A syllable with coda "x" cannot be followed by one that starts
with "k" or "c", and viceversa.
5. A syllable with coda "m" cannot be followed by one that starts with "z".
6. A syllable with coda "n" cannot be followed by one with onset "tc", "ts",
"dj" or "dz".

(Rules 4, 5 and 6 I find really arbitrary, but what can we do.)

I think that would be a complete list of the phonotactic constraints, if
I'm not forgetting anything, but an additional provision needs to be made
for any final consonant clusters in cmevla (which the above would not allow
at all), and also for some initial consonant clusters in cmevla. The
official rule
is not very clear about that. My take on the latter is that a cmevla can begin
with a coda plus permitted syllables, so {c,tci,py,rias} and
{c,tii,py,rias} would
both fall under this exception. To make these pronounceable within the
constraints, a "y" is allowed to be added at the start: {yc,tci,py,rias}
and {yc,tii,py,rias} without changing the name.

I still haven't made up my mind about what's my preference for final clusters
in cmevla.

mu'o mi'e xorxes