> Well, that is basically the reason why I do not regard that anHowever, the later *-tor suffix could have been added later, at my
> acceptable solution.
> The difference of accent did. Assuming that the change of e to oWhat caused the "difference of accent"? That's the question I'm
> postdated the creation of zero-grade brings you nowhere as far as I
> can see, while the opposite chronology really does the trick. I see
> no motivation for your belief that the nominative lengthening was
> later than the full development of zero-grade. Things can be made
> to work if you tune the chronology in a constructive manner.
> I think it is the case. It is only unaccented where the rootSo you surmise that accent was retracted where there was underlyingly-
> vocalism is underlyingly long, i.e. in cases where the accent would
> shift to the initial even in weak-case forms.