genitive affixes

[stassen]

Dear colleagues,

to tide you over the holiday season, I have a question to ask you, which
may very well be a stupid one. Somehow I have the feeling I should already
know about this, but if I should, I can't remember the relevant literature,
so I hope you can give me some pointers.
  The question is this. Does anybody have evidence for the hypothesis that
markers of adnominal NP's ('genitive' markers, whether they be adpositions
or affixes) derive from locational case markers, such as "to", "from" or
"at,on"? Of course, this is a reasonably assumption for Dutch and German
("van"/"von"), which clearly derives from the ablative marker "from", and
the same goes for English "of" and Romance "de/di". In fact, I think the
common opinion is that there was just one case suffix for the genitive and
the ablative in Proto-Indo-European (at least in the singular). But can a
same sort of analysis be defended for other language families, for example
for Uralic "-n" , Turkish
"-un/-in", and similar items in languages that have overt genitive marking,
like for example Dagestanian and Tibetic?
  The reason I ask is the following. I am setting up a typology of
predicative possession. One of the types I have to distinguish anyway is
something I call the Locational Possessive: it's the Latin/Russian type of
"At/to John, there is/exists a horse". Now, in the same areas where I find
this locational type, I also find instances of what might be called the
Genitive Possessive, that might be rendered as "John's horse is/exists".
However, if I can reduce genitive marking to locational
(ablative/dative/locative) marking, I can collapse the two types into one,
and I would like that.

Thanks in advance. And, after all this, can I still wish you all Happy
Holidays? Like some poet once said (and I assume this is particularly
appropriate to linguistic typology): "Let's party like it's 1999!"

Leon Stassen.