Re: An arabo-romance conlang?

En réponse à Vasiliy Chernov <bc_@...>:

> Sorry, I dropped out quite unexpectedly! I see lots of interesting
> messages in this thread, and I'll give an all-in-one reply to some.
> (Thus, again, beating the records for mesage length).
>
Well, I'll beat them again with my first post on Arabo-Romance phonology, be
sure of it :) .

>
> Consonants (use fixed-width fonts, pls.):
>
> f (or p?)  t   t_l (or s_l?)  s               t_S   k   X         H   h
> b          d   (d_l ??)       z               d_Z   g   R         3   '
> (p' ?)     t'  t'_l (or s'_l) t'_s (or s' ?)  t'_S  k'
> w          r   l                              j
>
> Modern correspondences:
>
> f          t    S             s                T    k    X         H  h
> b          d                  z                D    d_Z  R (ghain) 3  '
> (?)        t.   d.            s.               z.   q
> w          r    l
>
> Vowels: a(:), i(:), u(:), aw, aj - total 8.
>
Thanks! With this I get a broad view of Arabic sound changes. I think it won't
be that difficult to match it with early Romance phonology (even though some
sound changes will have to work hard to produce glottalized consonnants as well
as post-velar ones :) ).

> Here's another funny problem for Arabo-Romance: if there had been no
> contractions yet, and further development in Arabic is to be imitated,
> you'll need loads of combinations with intervocalic [j] and [w] to
> proceed from. In Latin they are *much* less frequent, and there are many
> restrictions. Perhaps you'll need some preliminary changes yielding more
> glides. (I guess you'll agree that half of the fun with Arabic, and
> other Semitic langs, is about the 'weak' roots which are very common).
>

Indeed :) . Well, Proto-Romance is so full of glides, I should be able to find
places to put them right :) .

>
> A sample:
>
> http://www.worldscriptures.org/pages/syriacancient.html
>
> Not too nice calligraphically, but may give you some idea of it. The
> title line is the older variety, the rest - the newer one. I always
> forget which of them is called Serto and which Estrangelo :( .
>
Indeed, it looks quite nice, especially the diacritics.

> >What about Greek /pt/ and /ps/? Did they stay like that or become
> >emphatic too?
>
> Dunno about /p/ in general. Ethiopic and late Aramaic created a
> "borrowed
> phoneme" for it: glottalized [p']. Arabic could do the same at first,
> then merge it with either [f] or [b]. Hard to say, since most early
> borrowings seem to be mediated by Aramaic which wasn't very consistent
> here.
>
Well, then Arabic could be unconsistent too, so if I decide to drop /p/ (not
sure, after all, it's a Romance language with Arabic substratum, not an Arabic
language with Romance superstratum), I may also be inconsistent :) .

> But note that Greeks themselves identified the first components in _ps_
> and _ks_ with *aspirated* [p_h], [k_h].
>
So ps could yield easily to /fs/, and ks to /xs/?

> >For initial consonnant clusters, I will use an epenthetic vowel [i]
> <...>
>
> My point was primarily that perhaps you'll *lack* some clusters :)
> (medially).
>
Indeed. Well, some vowel deletion will have to do (I've already thought of the
ominous -min- sequence - like in homo, hominis, femina, dominus, etc... -
leading to so many different things in different Romance langs - -mbr- in
Spanish and Portuguese I believe, -mm- in French, etc... -. I've thought of a
development: -min- -> -mn- -> -bn- through some kind of dissimilation. I don't
know how plausible it is, but it could be a source of -bn- clusters).

> >If they are three-consonnant clusters (I know only
> >two: scr and str) <...>
>
> s + p/t/c + r/l  ([stl] rare: _stlatarius_ and a couple of other weird
> words).
>
I think I can do a lot of things with these ones :) .

> >What is the stress pattern of Arabic by the way?
>
> Same rule as for Latin (the word being taken in its 'context' form).
> With some variation between local traditions concerning the impact of
> sentence-final modifications of the word etc. No reservation for 'muta
> cum liquida' - all clusters lengthen the syllable.
>
At least I won't have to much trouble to match the stress patterns then. Good
(but also bad, since the mismatching due to different stress pattern between
the substratum and the superstratum could have yielded to nice sound
changes :) ).

> >> (Besides, I'd love [Ng] as an additional source for [3], very
> frequent
> >> in Arabic).
> >
> >Well, that's a strange sound change! How did it happen? (interested
> look)
>
> No, I don't think it's been ever attested in the Semitic langs ;)
>
Oh, I just misunderstood your sentence. Sorry!

> But I do believe [N] > [3] would be probable. I think the closest thing
> to identify with a (post-)velar resonant is an unclearly pronounced [3],
> for an early Semitic lang speaker. And did you notice that [3] is one
> of the most frequent consonants in all positions in Arabic? I just
> thought
> you may like a (Eurocentrically) weird change like that. Just imagine
> some _*ego plango_ > 'a: Sa:3(u) %D (the last two chars being a
> smiley...)
Well, I'm planning to achieve such effect yes, so your idea is definitely
nice :) .

> >I chose to put the maximal inventory here, so that if there's a big
> mistake
> I
> >would have to substract phonemes rather than add them.
>
> Then it's OK; indeed Arabic has no trace of some distinctions (s : ts,
> sl : tl, l : dl, sl' : tl'), but you can always merge these sounds
> *later* ;)
>
Indeed I did :) . I won't use the Proto-Semitic inventory anyway, just the 1st
century Arabic one :) .

> >
> >This would give this inventory:
> >
> >   t  ts  tS  tl  k     ?
> >b  d  dz  dZ      g
> >   t' ts' tS' tl' k'
> >
> >f  s              x  H  h
> >                  G  3
> >
> >w  r      j   l
> >m  n
> >
> >a  i  u
> >a: i: u:
> >aw aj
> >
> >Well, this is probably a mess, but it was worth the try.
>
> Not at all. Some *phonetic* details are questionable (dz, tl, etc.), but
> the inventory is OK with me.
I agree with you. I would rather think they were /z/ and /sl/, but then it
would be more difficult to merge with Early Romance phonology, so full
of /dz/, /dZ/ clusters (a little less of /tl/ ones :) ), so since I have to
stretch a little bit with history to find a reasonable way to explain a Latin
settlement in the middle of the desert, I think I can stretch a little bit with
the phonology I will use (let's just say that the Arabic speakers those Latin
people came into contact spoke a strange dialect of Arabic that didn't survive
in our timeline :) ).

 And indeed, the exact quality of x, G
> (uvular/velar) isn't so important. (I guess G is fricative?)
>
It was really a mistake of me (indeed G is the voiced velar fricative). I had
forgotten that Arabic had /x/ but no /G/, rather the voiced *uvular* fricative.
Still it seems very plausible to me to have /G/ backing to uvular while /x/
stays the same (maybe only because I have trouble producing voiced velar
fricatives, but voiced uvular fricatives are fine with me :) ). It also drops
out the need of a voiceless uvular fricative to balance the system and helps
with the merging with a Proto-Romance phonology (at least I think so).

Well, Vasiliy, you're of great help. Thanks a lot for your information.

Christophe.