* Thomas R. Wier said on 2004-10-06 07:09:46 +0200
> From: Chris Bates <chris.maths_student@...>
> > So... if something merges morphological with another word it can't be an
> > adposition even if it satisfies your criteria? Then there are two
> > criteria at least:
> >
> > (i) Not repeated for each NP in coordinate phrases
> > (ii) Does not merge morphologically with another word
>
> The first of these is the criterion that actually distinguishes cases
> from postpositions; the second is only a practical concern not actually
> related to case-marking as such, but to boundedness in morphology.
So how about a language where all suffixes common to two NPs can move
onto the conjunction that is joining them together? There is some degree
of morphological merger as well.
t.