intervocalic consonant clusters in Lojban & Vorlin

["Mark E. Shoulson" <cbmvax!uunet!pucc.princeton.edu!shoulson>]

"homorganic" is a word which I came across studying Skt, which has all
sorts of rules enforcing what was considered euphony, and many had a lot to
do with homorganity.  Your definition is right, Rick, leastways right
enough for our purposes (it's easier in Skt, where the alphabet is ordered
by points of articulation and within those by type: voiced, unvoiced, etc.
Very cool.  Ask me about it one of these days).

And's complaint is very cogent and, in my opinion, bordering on damning to
Lojban's phonology.  Lojban's rules on consonant clusters are as follows:
Define "voiced consonants" as b,d,g,j,v,z, and unvoiced consonants as
c,f,k,p,s,t,x, and nasal/liquids are l,m,n,r.  For a consonant cluster
C1C2, (1) C1 cannot equal C2.  (2) If C1 is voiced, then C2 must either be
voiced or nasal/liquid, and if C1 is unvoiced, then C2 must either be
unvoiced or nasal/liquid.  (3) Both C1 and C2 cannot be among c,j,s,z.  (4)
cx, kx, xc, xk, mz are forbidden also.

As you can see, these rules say nothing about homorganity, and would be
quite a bit more complex if they did.  Personally, with the possible
exception of "kx", I find all the clusters of rule 4 easy to pronounce.
Note that "xk" is a homorganic fricative/stop conjunct.  Lojban "n" may be
pronounced as [ng], but "m" may not.  It doesn't include that phone.

Note also that the rules permit some clusters which (sorry, Lojbab) are not
so easy to work with.  One which I found just recently and marveled at was
"pm".  This is a real toughy.  Homorganic stop/continuant clusters are
usually fairly touch, as opposed to homorganic continuant/stop clusters,
which tend to be easy.  Something like "pm" would likely get swallowed up
into the nose; not what we'd like.  Ditto "tn" or "dn" or "bm".  "dl" and
"tl" have their own problems.

I should point out that these rules are not sufficient to determine
permissible *initial* clusters; those form a much smaller set.

I've included the lojban list on the loop here, because this is something
that ought to be discussed there as well (though I'd hope that pure "should
we change?"/"No, we shouldn't"/"Yes, we should" discussions would stay
there and not cross over too much).

I don't know about Arabic, but I do know that Hebrew does, in at least a
few cases (nowhere near the overwhelming complexity and ubiquity of Skt),
change letters in grammatical constructions for the sake of euphony, to
allow easier pronunciation (e.g., the root z-q-n, meaning "old".  when used
with the reflexive, meaning "to grow old":  the reflexive adds "hit-" to
the beginning, among other things.  There's a transposition: the "t" of
"hit-" becomes a "d", and moves *after* the "z" of the root, giving
"hizdaqen")  More commonly, the vowels (which are a trifle more complex in
Hebrew than Arabic) often change according to rules of euphony (certain
letters force the grammar to use slightly different vowels than otherwise),
or even change the pronunciation of a vowel without changing its
orthography (only one vowel is like this: the schwa (the English word comes
from the Hebrew).  It has two flavors: a "moving" schwa, which is
pronounced more or less like a schwa, only shorter, and I usually hear it
more fronted, almost like a very very short [I], and a "resting" schwa,
which has no sound at all and goes between letters because you have to put
something there (cf. Arabic sukun, tho someone who knows better will
correct me)).

English does seem to be more forgiving of non-homorganic clusters
straddling morpheme boundaries; I remember hearing somewhere that the only
Eng word with "-tk-" in it was "catkin" (yeah, a commonly used word).
English speakers sometimes do tend to go for the Skt sort of homorganizing
(?) nasals, even in words like "manpower", some will nasalize the "a" and
make the "n" sort of a short "m".  A more striking example is the
pronunciation of "sandwich" as "samwich" or even "samwidge".

I think most assimilation is likely to be regressive (i.e. the earlier
consonant changes to be in concord with the later, so [mt] wouldn't go to
[mp]).  This is how the Skt rules go, and also how I find more natural.
There are sure to be exceptions, though.

~mark