Quoting JS Bangs <[log in to unmask]>:

> Andreas Johansson sikyal:
> > The possessive, finally, is formed by infixing an _-i-_ , turning the stem
> > vowel into an diphthong. Pronunciation; _ii_ [ej], _ei_ [ej], _ai_ [aj],
> _oi_
> > [oj], _ui_ [uj]. Yes, _oi_ is ambigious between [wi] and [oj]. The
> possessive
> > goes after the thing possessed; _chreach chain_ "the castle of the lord".
> To
> > top it off, it, out of misplaced sympathy, echoes any accusative ending on
> the
> > thing possessed, giving us things like _chreanco chainon_ "the castles
> (acc)
> > of the lords".
>
> This is quite similar to the Yivrian possessive, though Yivrian lacks the
> twisted  complications :-). It seems clear from this and the behavior of
> plural -n- that the proto-language must have had several infixes. The
> plural -n- was probably imperfectly generalized to a suffix, while the -i-
> remained as it was. Does this fit with your plan?

The idea was that final fricatives represent primitive short stops, which got
a -nV adding strapped upon them, with subsequent metathesis and loss of final
vowel in the pl, and fricativization of a short stop preceeded by a vowel in
the sg, whereas the present final stops represent primitive long ones, which
demanded a epenthetic -a- before the -nV.

I'm considering a /-w-/ infix for some verbal inflections, tho'. Would create
another source of Vw diphthongs if I take Christophe's suggestion of having
accusatives of stems ending in vowels simply add -o /-w/.

                                                        Andreas