Professor Cervigni is correct to call into question the suggestions of an
external linguistic source for the enclitic possessives. There appears to
be is no evidence for origins outside of Italy, but a rather lengthy list
of circumstantial evidence to place hypotheses of external influence in
very serious doubt.
I won't bore everyone with such a list. But I will note that it is
implausible that a borrowing so syntactically constrained (enclitic to the
noun, unstressed) could occur, even if the linguistic and socio-historical
facts could provide a scenario of sufficiently intense contact with a
language containing enclitic possessives to be able to support the
hypothesis of borrowing. The major glitch in the borrowing hypothesis,
however, is that it's unnecessary.
A good source for a quick overview of the historical development of
enclitic possessives is chapter 4 of Pavao Tekavcic's 1972 Grammatica
storica dell'italiano. Vol 2: Morfosintassi, in particular pp. 186-188.
There we find, as Professor Cervigni suggested, that the construction goes
back to Latin, thus there is no need to posit an external source.
Latin syntax permitted many freedoms, of course, but normal position for
the full, bisyllabic forms in Latin seems to have been postposition. It's a
short step phonologically for these to develop unstressed alternative
forms. It's a much longer -- or better, linguistically more complex -- step
for these to grammaticalize as highly constrained syntactic clitics, but
it's fairly unremarkable in historical perspective once the monosyllabic
unstressed variant exists. Tekavcic points out that (presumably unstressed)
monosyllabic possessives are attested along with the expected full forms in
popular Latin as early as at least 2nd Century (MATER MA, PATER TUS), and
that both the old Tuscan and the modern Central-Southern dialect enclitic
forms are the direct continuation of these.
Although once common at least as far north as Lucca, they may appear
striking or mysterious today because their use is highly restricted in
geographical terms, i.e. they are typologically or statistically somewhat odd,
especially if viewed in pan-Romance perspective. Their historical loss in
northern areas of Central Italy, though, and the modern restriction in at
least some dialects to a few terms of family relationship and only first
and second person is just the sort of narrowing of scope to be expected of
an archaism.
As an historical linguist by trade, I find that if there's a interesting
mystery to be solved here, it's not so much their origin, but why and/or
how they have receded.
Tom Cravens
University of Wisconsin-Madison
[log in to unmask]
|