--- In email@example.com
, "elmeras2000" <jer@...> wrote:
> I have read this several times and thought a lot about it, and
> I have to give in. That indeed seems to be the only consistent
> synchronic analysis. It means that clitics can not only arise
> from full words that are reduced in status, but also the other
> way around, from desinences that grow into more independent
The so called declensional inflections of Ossetic work
just like the English -'s in this respect, and so are
in fact enclitic case particles as well. Some of these
endings, such as the genitive -y, go back to true Proto-
Iranian inflectional endings, but at least one, the
addessive -yl (Dugoron -bæl), is supposed to have come
from the postposition *upari. So here we see both ways
in one language.
If the interrogative particle 'ti' found in some dialects
of Canadian French is considered a enclitic, then there
is still another way that one may arise. This particle
arose as a misanalysis of the differing results of liaison
in such phrases as 'il dort' [il.doR] "he sleeps", and
'dort il?' [doR.ti] "is he sleeping?", which led to 'ti'
being considered separate from 'dort' [doR], and the marker
of interrogation. My little linguistics primer calls it
an "example of restructuring from the domain of syntax".
> I wonder if that possibility has not been overlooked in cases
> where it would have provided solutions.
Yes, and surely such misanalysis and restructuring can
occur in numerous and varied circumstances.