Re: Evolution of Romance (was: **Answer to Pete**)

2008/2/1, John Vertical <johnvertical@...>:
> On Fri, 1 Feb 2008 17:20:24 +0100, Benct Philip
> Jonsson wrote:
>
> >The only thing I see speaking against a development g' >
> >j > g' > dZ is Occam's razor! Clearly [g;] or [J\] and
> >[j] can both develop out of and into each other, but to
> >posit a to-and-fro development seems a bit suspicious.
>
> But you need to set up j >> dZ anyway, so then you have a
> palatal > palatovelar > palatal to-&-fro development
> there. And for the same, a continuant > stop > continuant
> development is not only possible, but necessary!
No, that would be a straight palatal fricative >
palatal stop > alveopalatal affricate > palatoalveolar
affricate shift.

> However, it just occurs to me that starting from j > gj)
> rather than j > J\ directly would be symmetrical with w >
> gw)... or "/gj)/" could have been phonetically a simplex
> [J\] since the beginning anyway...
The possible distinction between [g;] and [J\] is academic.
I doubt kids learning a language could distinguish them, so
you may have /g;/ realized as [J\] and/or [J\] phonemicized
as /g;/. Also the thing with the Germanic /w/ > VL /gw/ is
that after w > B /_V /gw/ was the closest thing VL had to
[w], not that it was in any way parallel to /j/ which
existed in both languages. FWIW /gw/ may have been realized
[Gw] / V_V, which made it even more like [w].

> Is there any evidence on which of the kj) tj) and gj) j
> mergers took place first?
I've got the impression that spelliongs like NACIONES appear
earlier than G/I confusion. Z for affricates unfortunately
only occurs in words that were or were thought to be Greek,
though often significantly erroneously from an etymolgical
POV, like ZABULUS (diabolos) and ZODORUS (Theodoros).
Conversely you find spellings like Zoulia = Iulia in Greek.

> >The relative infrequency of dj compared to tj is probably
> >a better and sufficient explanation why the voiceless
> >palatals develop differently in Western Romance.
>
> I'm afraid I don't quite see the logic behind this
> argument. It's a merger, not a chain shift, so there can
> be no pull effect due to either palatalized coronal.
No, but kids learning to speak would hear many more tokens
of [t;] to confuse
[c] with than they'd hear tokens of [d;] -- the frfequency
    of [d;] tokens was simply to low to make a mental
    imprint, or alternatively the few that appeared stood
    out enough to preserve their identity. Perhaps the
    preservation of /dz/ in Italian was due to kids there
    being called to PRANDIU at least once a day, while kids
    in Gaul were callled to DESIEIUNUM! :-)

Also it seems that [c] is more perceptually similar to [t;]
than [J\] is to [d;] for whatever reason. The Hungarian
spellings _ty_ for /c/ but _gy_ for /J\/ are probably
significant for whatever reason. Also /ts/ is more
widespread crosslinguistically than /dz/. I'm not sure
what's at work here, though something clearly is. Perhaps
palatalization is more audible as such in voiced sounds,
while in [t;] the frequent secondary affrication ([t;] >
[ts;] > [ts]) is more salient? Dirk, are you reading this?

> >For all that I want an alveolar diacritic _a\ seems a
> >poor choice unless a\ stands for some alveolar sound. I
> >dislike _t for breathy voice for the same reason. To me
> >_t for alveolar and _h\ for breathy voice would make much
> >better sense. I guess I'll have to add that one to my
> >list of 'BXS' modifications. Of course the only goal of
> >BXS is to be inherently consistent, unlike Z-SAMPA and
> >even less than CXS caring about backwards compatibility.
>
> We all have our SAMPA idiolects, don't we?
Sure. There have always been IPA dialects
(diagraphies? :-) too.

>  Like me with my preference to use ) not _ for an actual
>  tie bar; tS_w_h would become thSw))) or maybe even
>  tWS)) ;)
Actually I think [tS_h_w], since labialization would persist
through the aspiration phase! ;-)

> John Vertical
>

--
/ BP