Re: CHAT: Brainstorming! Relative clauses

On Tue, 5 Oct 1999, taliesin the storyteller wrote:

> I'm currently struggling with understanding the rel. cl. of my lang
> correctly, and thought I'd look at how it works in other langs.
> Unfortunately, it seems that relative clauses is part of the grammar
> that is 'not yet online' :) Does it change too often?
>=20
It doesn't change much, really... I try never to change anything
(although I think having gender in plural pronouns, but not in
singular pronouns is having to go.), but relative clauses are not
much used in Denden, where nominalising a phrase and then
subordinating it is preferred. The first is a clear example:

>=20
> rel. subject:
> "The dog that saw the cat was large."
>=20
  (In most of these cases the nominalisation particle _ga_ NOM after
  _zi_ can be omitted; it is not necessary, except for the rythm of the
=09sentence. One syllable words like _zi_ somehow give the impression
=09the phrase isn't finished.)

The most idiomatic construction is with the relative clause preposed,
finished with _ga_.

  quna qir.e   ga  kal zi  ga
=09cat  see.PRT NOM dog big NOM
  The dog that saw the cat was big. 'the cat having seen dog was big'

This is possible, too, a simple juxtaposition of two clauses, with
a topic marker that focusses on the dog:

  kal ka  quna qir.e,  zi  ga
  dog TOP cat  see.PRT big NOM
  'As for the dog, it saw the cat, it was big'

Without the topic marker it is not clear whether the remark _zi ga_
applies to the dog or the cat, and the sentence is ambiguous:

=09kal quna qir.e,  zi ga
=09dog cat  see.PRT big NOM
  'The dog saw the cat and was big.'
=09'The dog saw the cat, it was big.'

Putting the topic marker after _qir.e_ shifts the focus:
=09
  kal quna qir.e   ka  zi  ga
=09dog cat  see.PRT TOP big NOM
  'As for the dog that saw the cat, it was big'

Putting the topic marker after _quna_ is interesting, too:

=09kal guna ka, qir.e,  zi  ga
=09dog cat  TOP see.PRT big NOM
=09'As for the dog and the cat, it saw/ it was seen, it was big
=09As for the dog and the cat, they say/they were seen; they were big

Omitting the pause after qir.e would give the meaning: 'it was seen to
be big'

Participially (rather a foreign influence, southern mostly):

  quna qir.an  kal zi  ga
=09cat  see.AGP dog big NOM
=09'The cat-seeing dog was big.'

There exist special deictics for use in relative clauses, and these are
often a mark of influence from Classical Charyan or from easteren non-Chary=
an
languages:

  kal yohox quna qir.e   zi  ga
=09dog RSUB  cat  see.PRT big NOM
=09The dog who saw the cat was big

The subject relative pronoun _yohox_ is most often used with
human referents, and for a dog the proximal deictic _yo_ 'near
the hearer', which is the most unmarked deictic in Denden would
be preferrred:

  kal yo   quna qir.e   zi  ga
=09dog that cat  see.PRT big NOM

Although in this sentence _yo_ could also refer to the cat.

> rel. object:
> "The dog saw the cat that killed the mouse."
>=20
The situation with relative objects is as complicated as with=20
relative subjects. A common construction is with the nominalisation
particle _ga_ NOM - it appears that this construction is incompatible
with the topic marker _ka_, which would occupy the same place in the
sentence. The relation between _ka_ TOP and _ga_ NOM is worth a
separate study...

  kal ga  quna kela  mozhaz qir.e
=09dog NOM cat  mouse kill   see.PRT
=09'the dog's seeing the cat killing the mouse.'

While this is a perfectly idiomatic expression, but the following
construction is far more common:

  quna kela  mozhaz ga  kal qir.e
=09cat  mouse kill   NOM dog see.PRT
=09The dog saw the cat that killed the mouse.

There exists also the object relative pronoun _danran_, although the
shortened proximal deictic _d_ 'near the speaker', can also be used.
I don't exactly know why 'relative to the subject' is equal to 'near
the hearer', and 'relative to the object' is equal to 'near the speaker'.

  kal qir.e   quna danran kela  mozhaz
=09dog see.PRT cat  OREL   mouse kill
=09The dog saw the cat kill the mouse/the cat that killed the mouse.

There is no difference between the two translations, and I don't know
whether Denden can make clear the dog saw the cat in question, but not
the killing of the mouse, or the cat in action, without use of topic
markers or circumlocution.  Interestingly, this is one area where the
SVO origins of Denden show up without fail - Denden is clearly moving
towards SOV, but=20

  *kal quna qire danran kela mozhaz

and
=20
  *kal quna danran kela mozhaz qire

are impossible.

Another idiomatic rendering using a participle and a topic marker
would be:

  kela  mozhaz.dan quna ka, kal qir.e
  mouse kill.AGP   cat  TOP dog see.PRT
  As for the mouse killing cat, the dog saw it.

_kela mozhaz quna ka, kal qir.e_ would mean 'As for the mouse
killing the cat, the dog saw it.'

Compare:

  quna kela  mozhaz ka, kal qir.e
  cat  mouse kill   TOP dog see.PRT
  As for the cat killing the mouse, the dog saw it.

However, a construction with _tan_ RTV also exists, and can only be used
for object relative sentences, and also demands a SVO sentence pattern,
this time in both parts:

  kal qir.e   quna tan mozhaz kela
=09dog see.PRT cat  RTV kill   mouse
=09The dog saw the cat's killing the mouse,
=09The dog saw the cat kill the mouse

This construction is not used in careful writing, but appears to be
quite popular in spoken Denden, especially in Broi, which is curious,
since it is in Broi that the move towards SOV has progressed furthest.

Comparable is:

  quna kela  mozhaz ga  ka, kal qir.e
=09cat  mouse kill   NOM TOP dog see.PRT
=09As for the cat that killed the mouse, the dog saw it.

The juxtaposition of _ga_ and _ka_ is considered inelegant, and the
same construction without _ka_, (see above) is preferred.

> nested:
> "The dog that saw the cat that killed the mouse that was large drank
> from the river that John put the poison in."
>=20
Nobody'd write a sentence like that in Denden ;-). (Easy escape!) An
idiomatic rendering would be:

  Hamal laulau yuhan tan lus.ca     ka, kal sh=FCmogh.=20
  Hamal poison put   RTV river.ILLA TOP dog drink

  quna qir.e   ga  kal ka, zi  ga.  Quna kela  mozhaz ga  yo   quna ga
  cat  see.PRT NOM dog TOP big NOM  cat  mouse kill   NOM that cat  NOM

As for the river Hamal put poison in, the dog drank from it.  It was the
dog that saw the cat, the big dog. It was the cat that killed the mouse,
that cat.

> rel. indirect object:
> "The dog that John gave the ball to."
>=20
This can best be rendered using a topic-comment construction, since the
English sentence isn't complete:

kal ka, Hamal tado jerat.
dog TOP Hamal ball give
As for the dog, Hamal gave it a ball.

As part of a larger construction, the following would be preferred:

Hamal tado jerat ga  kal
Hamal ball give  NOM dog
The by Hamal a ball given dog

> rel. oblique:
> "The mouse that the cat chewed on." or,
Quna kela  verat ga...
cat  mouse gnaw  NOM
The mouse the cat gnawed=20

> "The cat that the dog was bigger than."
This is strange, since the sentence isn't finished. It looks like an
ordinary comparison, but _tan_ comes after _kal, instead of before it:

Quna Kal tan zi  ga
cat  dog RTV big NOM
The cat the dog was bigger [than]

> rel. posessor:
> "The dog saw the cat whose teeth were huge."
>=20
Kal qir.e   quna, tan quna rash  zi.zi   ga
dog see.PRT cat   RTV cat  tooth big.big NOM
The dog saw the cat, the teeth of the cat were real big.

(Hmm... I wonder how this shapes up with the essay on _ga_ is sent to
the list in July. Still haven't written that chapter...  Let me amend
that bit on changes - I never change anything on purpose! And I've been
to busy with Kura to write much Denden, might have made a few mistakes in
this.)

Boudewijn Rempt  | http://denden.conlang.org/~bsarempt