On 10/24/06, Roger Mills <[log in to unmask]> wrote:
> I find that particular case entirely reasonable-- though on quick reading I
> assumed that the /i/ was also a conditioning factor, since t >
> fricative/affricate/sibilant before i or front V in general is very common.
> But it's also possible before other vowels too, though stress is not usually
> a factor, I think. Jap. /t,d/ > [ts,dz] before /u/ IIRC
That was a bad example, because I'm planning for /i/ to have affects
similar to what you mention. I haven't decided whether or not this
form of the language will have it though; I know that later it will be
> An Indonesian language of my acquaintance pre-aspirates stops in post-stress
> position-- /'paka/ > ['pa(h)ka]; that could easily lead to a fricative
> pronunciation like ['paxa].
That's interesting. I like it.
> I toyed with a similar idea for the development of Gwr, but couldn't decide
> if stop > fricative before or after the stress, so gave up on it. (I did
> keep both as irregular/dialect features.) In most cases, I think such
> changes happen mostly in intervocalic position; what did you plan to do
> about, say, initial unstressed position-- that is, would /ta'ti/ > [sa'ti]?
> That's probably not found in any natlang, but in a conlang, it could be an
> interesting departure from realism, and nihil obstat, IMHO :-))
I don't know yet. I'm thinking that it might change, if the previous
word ends in a vowel, or perhaps if the previous word is a clitic.
I'm thinking about making stress penultimate and then having clitics
and affixes affect which syllable stress falls on, making noun
declension and maybe verb conjugation difficult if you don't know the
rules. However, I still haven't worked out how to make it irregular
but not *too* irregular, so anything I come up with now might be
sacrificed to reach that goal. =)
I'm creating two different versions of this language, and in the later
form I think that lenition is no longer so systematic or common. But I
want remnants of it scattered all over the place.