Re: tonal languages

>Douglas Koller, Latin & French wrote:
>
>>  Natch, there are exceptions. "Ki3", "to go", should morph into "ki2",
>>  but this is usually realized as "ki1" ("ki1si2" = "drop dead"). "Lai5",
>>  "to come", and "beh4", "to want", operate similarly. Too, these sandhi
>>  rules may vary in different varieties of Hokkien (like I think in Tainan
>>  tone 5 (24)  becomes tone 3 (21), not 7 (33), or something like that).
>>  H.S.'s Hokkien may look considerably different.
>[snip]
>
>Indeed it's considerably different. For one thing, tone 3 usually sandhi's
>into tone 1, not tone 2. Actually, now that I think of it, there appears
>to be no fixed pattern: hi3 (show/movie) + tai5 (stage) --> hi1tai5
>(cinema);
As I mentioned above, "ki3" usually sandhis into "ki1" rather than
the expected "ki2". Perhaps what you're describing is a broader
application of that phenomenon.

>  however, gong3 (foolish) + kia~2 (child) --> gong3kia~2 (no
>sandhi!).
You mention below that for you, 7 and 3 have merged, which would make
sense in that for me, "gong7" is originally tone 7, so "gong3kia~2"
is actually sandhied. But lo, we get the same result (at least
here)....:)

>Perhaps it's just dependent on context. Also, tone 2 in my
>idiolect is (35) not (52), so this may be a result of euphonic change
>based on this different tone shape.
As Malaysia also has a Cantonese speaking population, could this be
Cantonese influence? The Cantonese shang tones are both rising. It
screwed up my head totally for months on end to have to pronounce yin
shang as a falling (52) tone in Taiwanese. Having it (35) would have
felt a lot more natural.

>This may be why I'm
>getting strange sandhi rules here.
Not *so* strange.

>So there you have it. The Hokkien I speak in my hometown consists of these
>tones, plus a few sound changes (such as mutation of some initial g's:
>gua2 ("I") --> wa2; jit4ge1 ("this one") --> ji4le1; li2 ("you") --> lu2).
These are common variants in Taiwan as well.

Kou

Reply