[YG Conlang Archives] > [romconlang group] > messages [Date Index] [Thread Index] >
Okay, how about this?
Step 1
shch, zh, ch - retroflex < str, dr, tr
q, j, x - alveolopalatal < ki, gi, si
c, z, s - coronal < ti, di, sV[-i]
Step 2
(thanks to Adam for mentioning Tsakonian, where I found this rule!)
SC > C_h
shch > chh
sp > ph
st > th
sc > kh
This gives a voiceless-voiced-voiceless aspirated contrast, which seems more Tibetan than Mandarin, but it is getting closer, and I can keep q/j/x and c/z/s. The ch/zh/chh series is peculiar, but I suppose a putative scholar of this stage of the language would analyze it as mixed system, if q/j/x and c/z/s are Type I and p/b/ph, t/d/th, and k/g/kh are Type II.
--- In romconlang@yahoogroups.com, "Anthony" <mamercus88@...> wrote:
>
>
>
> --- In romconlang@yahoogroups.com, Adam Walker <carraxan@> wrote:
> >
> > On Tue, Aug 31, 2010 at 1:13 PM, Anthony <mamercus88@> wrote:
> >
> > >
> > >
> > > --- In romconlang@yahoogroups.com <romconlang%40yahoogroups.com>, Eric
> > > Christopherson <rakko@> wrote:
> > > >
> > > > On Aug 30, 2010, at 10:26 PM, Anthony wrote:
> > > >
> > > > > I've been rethinking the Lim1guam1 La2ti2nam1 series of fricatives.
> > > Here is my current arrangement, presented for your consideration:
> > > > > ch [tS] < ti, te, ki, ke
> > > > > zh [dZ] < di, de, gi, ge
> > > > > sh {S] < si, se
> > > > > q [t_s\] < [s_\t_s\] < str
> > > > > j [d_z\] < dr
> > > > > x [s\] < [t_s\] tr
> > > >
> > > > Are these meant to be the same sounds as in Mandarin? If so, _ch zh sh_
> > > should be retroflexed, and retroflexes are quite likely to arise from
> > > coronal stop+r -- rather than the alveolopalatal series you have coming from
> > > that.
> > > >
> > > > > Which do you like more for the nasal level tone? lim1 or lin1?
> > > >
> > > > Personally I kind of like the -m1 form, because it reminds me of
> > > Portuguese, but apart from Portuguese it's unusual.
> > > >
> > > I prefer the -m1 form, because it reminds me that this is still a romlang
> > > (I might simplify it to -m, since the 1 is implicit; but tone sandhi might
> > > invalidate that consideration).
> > >
> > > It seems to me that a romlang would be more likely to have an
> > > alveolopalatal series and a retroflex series, even if the romanization would
> > > be similar to Mandarin.
> > > So what about this?
> > >
> > > ch, zh, sh - retroflex < str, dr, tr
> > > q, j, x - alveolopalatal < ki, gi, si
> > > c, z, s - coronal < ti, di, sV[-i]
> > >
> >
> > I think you might concider flipping str and tr. It just seems more logical
> > to me. YMMV.
> >
> > Adam
> >
> >
> > [Non-text portions of this message have been removed]
> >
> I know the ch, zh, sh < str, dr, tr looks weird, but it comes from the following steps.
>
> tr > ch
> dr > zh
> str (since there is no Latin sr) > shch
>
> /shch/ doesn't really fit Mandarin-y Chinese phonology, so I had a choice: to combine /shch/ with /ch/, thereby leaving a gap in the /q,j,x/ vs. /ch, zh,(sh)/ contrast, or to initiate a chain shift of /shch/ > /ch/ > /sh/, creating parallel series /q, j, x/ and /ch, zh, sh/. The problem with the chain shift is whether it would drag the other series along with it (/q/ > /x/ > /h/? /c/ > /s/ > /h/?). But then the phonology would look even less like Mandarin! Is a chain shift in only one series (/ch zh shch/) possible? On the other hand, a shift /shch/ > /sh/ would also allow a shift of /st/ > /ss/ > /s/ and /sk/ > /sx/ > /x/, and /sp/ > /sf/ > /f/, thus ridding me of initial consonant clusters.
>
> Clearly, I need to rethink the GMP for Lim1guam1 La2ti2nam1. Any (more) advice on this matter would be appreciated.
>