[YG Conlang Archives] > [jboske group] > messages [Date Index] [Thread Index] >
On Sat, Dec 28, 2002 at 07:21:12PM -0500, John Cowan wrote:
> la .and. cusku di'e
> > Which rule does the parser apply?
> >
> > A. Where KU is found, insert {ku}.
> > B. Where KU is found, insert any member of KU.
> >
> > If A, then there is no problem in having other cmavo in KU.
> > If B, then there is a problem, but why would B be preferable to A?
>
> B. But that's not the point. The point is that "le ninmu cu cusku"
> is considered grammatical because it can be unambiguously transformed
> into "le ninmu ku cu cusku". If KU had both ku and ku'u'u as members,
> then we wouldn't know whether "le ninmu cu cusku" meant what it formerly
> meant, or meant "le ninmu ku'u'u cu cusku", and we would have to stop
> saying "le ninmu cu cusku" -- which is not going to happen.
This is a bit much. Of course we could have "ku'u'u" and mandate
that elided cmavo from KU are semantically "ku". It does *not*
mean we'd need to stop saying "le ninmu cu cusku".
However, the fact that we can doesn't mean we should. It's a lot
less clean to put And's shortening crap into terminator selma'o
than to use substitution rules before parsing each line (e.g. have
a list of s/goi'a/goi ko'a/g s/goi'e/goi ko'e/g type foo which are
just applied to each line).
Of course, I would prefer not having these at all without some
actual data about which things need to be shortened first (e.g.
a goi'a type thing for goi ko'a might be needed, but it's highly
unlikely that "goi fo'u" is said frequently enough).
--
Jordan DeLong - fracture@hidden.email
lu zo'o loi censa bakni cu terzba le zaltapla poi xagrai li'u
sei la mark. tuen. cusku
Attachment:
binmIJw4Adsqi.bin
Description: application/ygp-stripped