On Sat, 12 Feb 2011 21:07:49 +0000, David McCann <[log in to unmask]>

>On Sat, 2011-02-12 at 08:54 -0500, 1 2 wrote:
>> - What are other possible strategies for non-subject relativization?
>There are four forms of relatives, each less restricted than the former
>in the circumstances in which it may be employed:
>1. non-finite, as in Lezgi (Caucasus)
>2. relative pronoun, as in English
>3. relativising conjunction and embeded pronoun, as in my Liburnese
>4. the correlative construction, mostly Indian, but in Bambara (Africa)

What do you make of internally (or double) headed relative clauses, or
adjoined relative clause constructions?  My own attempt to try to fit them
into your scheme would involve splitting the latter part of the list in (at
least) two, according to whether the full expression of the head comes
inside the relative clause or the main one or both.  

I wonder whether it's attested for a language of the internally-headed or
correlative types to have a distinct series of pronoun used only for
relative clause heads in their main clauses.  This seems to be about what
the analogue of relative pronouns would be.

Also, some examples I tossed off in an earlier discussion of these things
got preserved at .
It would be good to amplify that with the results of the present discussion.