Patrick Littell <[log in to unmask]> writes:
> On 8/13/05, Henrik Theiling <[log in to unmask]> wrote:
> >...
> > Why would this forbid two relative clauses in a sentence?
> Not the subject of the relative clause; the subject of the main clause.
> ...

Ah! *bang*  Now I understand.

> So, if only one participant per clause can be realized as a subject, and
> only main clause subjects can be heads of IHRCs, each main clause may have
> only one IHRC.

Yes, indeed!

> Actually, S11 (Tesäfköm) will work like this: the topic (=fronted
> > noun) in the relative clause is the reference.
> This sounds like the Malagasy pattern. ...

Yes, I think so.

>... This is my "favorite" relativization construction, insofar as a
>non-creative linguist like myself has "favorites". ...

I also like the Chinese style relative clause very much.  I abstracted
it a bit and used it in Qthyn|gai, where every word and clause((!) by
means of its head) is marked with a case marker.  So you just mark the
modifying relative clause with genitive case and use it as an adjunct.
I liked that.

> I find it very elegant, for some value
> of elegant; it doesn't require "extra" machinery within the language to
> work. ...

Hmm, depends, I think.  The Bambara example is very nice.  Although it
seems that you could also have a reference marker (Bambara), or a
resumptive pronoun (Korean).  This is extra machinery just like 'de'
in Chinese.  And actually, Japanese and Korean EHRCs come with *less*
extra machinery: they lack any modification marker between the
relative clause and its modified noun IIRC.

I also love minimalistic machinery in my conlangs.  Tyl Sjok, for
example (with IHRCs) has an optional reference marker, but can do
without.  I also found this very elegant for an isolating language.

Tesäfköm, OTOH, will have a resumptive pronoun for IHRCs and use
construct state on the modified noun for EHRCs, so there is some extra
machinese for both.  Actually, with Chinese in mind, I was quite
surprised to see that in Japanese a 'no' is missing between relative
clause and modified noun.  And because I personally find it strange,
Tesäfköm will mark the modified with construct state.

> So long as there's a way to topicalize/front/realize-as-subject/etc. a
> given argument, it can be relativized upon.

The optional reference marker in Tyl Sjok is indeed the focus marker
when used at top level (does Tyl Sjok have a topic marker?  I don't
recall...  Marking as topic seems more appropriate for IHRCs). :-)

Anyway, it's optional -- so you *can* do without marking which
argument you relativize, although that probably becomes a bit
ambiguous.  But then, that was a design feature for Tyl Sjok. :-)

> Since we're on the topic of relative clauses, do your languages indicate the
> difference between restrictive and non-restrictive relative clauses (and
> adjectives)? ...

A week ago, I would not have been able to tell you for Tesäfköm (it's
still embryonic), because at that time I found that it can easily have
an EHRC and I wondered which relative clause to use (it already had
its IHRC from the very beginning).

After reading about Japanese (again -- I forgot almost everything), I
decided that having *both* is also a good option: the IHRC for
descriptive relative clauses and the EHRC for restrictive ones.  Both
embed nicely into Tesäfköm's structure.  I'm very happy with that
decision. :-)

> (This distinction is also possible with adjectives; ...

Adjectives are verbs are relative clauses in Tesäfköm. :-) All the
same, yes.

> I believe in most languages this can only be distinguished by intonation
> (or, of course, by words like "incidentally"), but in a few it's
> grammaticalized. ...

German is usually said to not distinguish, I think, but it can
optionally mark a clause as descriptive by the word 'ja'.

   Der Mann, der nicht arbeitet, ist krank.
   the man   who not   works     is  sick

      - ambiguous, but usually understood restrictive, I think.  At
        least without context.

   Der Mann, der ja nicht arbeitet, ist krank.

      - purely descriptive

And I thought English preferred 'which' for descriptive relative
clauses and 'that' for restricted ones (we also learned different
comman rules at school, but those you can probably not hear when

   The book that burns was funny.

      - restrictive?

   The book, which burns, was funny.

      - descriptive?

Native speakers?

> Since the fashion these days seems to be for avoidance of any
> ambiguity, ...

Yippie!  Tyl Sjok is not fashion!

> ... do any of your languages distinguish this? (You meaning
> anyone, not Henrik in particular.) ...

For me, only Tesäfköm so far.  (Tyl Sjok has pragmatic means of
distinguishing, like using two sentences instead of a relative clause,
but that probably does not count...)

OTOH, Qthyn|gai could do very easily, since it has an appositional
case, so I might think about using that instead of genitive for
descriptive clauses.  Very, very interesting idea!