Re: Relative clauses

On 8/6/05, Patrick Littell <puchitao@...> wrote:
> 
> 
>  Yes, there are natlangs that do this; as Jeffrey said these are sometimes 
> called "internally headed" relative clauses. My linguistic literature and my 
> internet connection are far away from each other right now, but give me a 
> week and I'll give you languages and relevant examples. 
> 
Here are some examples of internally-headed relative clauses, if anyone 
still wants some. The following is from Bambara, a Mande language of West 
Africa, adapted from Comrie (1989).

N ye so ye.
I PAST house see.

Tye be [ n ye so min ye ] dyo.
man PRES I PAST house MARK see build.

Marking the internal head is useful for reducing ambiguity. Consider this 
example from Imbabura Quechua:

[ Kan kwitsa-man kwintu-ta villashka ]-ka sumaj-mi.
you girl to story ACC telling -TOP pretty-VALIDATOR.

This could mean either "The girl to whom you told the story is pretty" or 
"The story that you told the girl is pretty".

As for disambiguation of what role the internal head plays in the main 
clause, the same strategies are available as for languages with external 
heads. (The goal is opposite, however; the need in an external-headed 
language is to indicate the role the external head plays in the subordinate 
clause.) If you have strict word order, the solution is pretty simple: just 
plunk the relative clause in the appropriate place for, say, a direct 
object, as in the Bambara example above. Otherwise, you may wish, for 
example, to mark the entire clause with a case marker. Here's an example 
from Diegueño, which iirc is a Yuman language.

Tenay ?wa :?wa :w.
Yesterday house I-saw.
"I saw the house yesterday."

?wa :-pu -Ly ?ciyawx.
house -DEF -LOC I-will-sing.
"I will sing in the house."

[Tenay ?wa :?wa :w] -pu -Ly ?ciyawx.
"I will sing in the house I saw yesterday."

In essence, the entire relative clause is just plunked down where the root 
would have been. Also pretty simple. So it's actually pretty simple for 
word-order-y and case-y languages. What about for agreement-y languages? 
Well, I can't think of any language with interally-headed relative clauses 
that also relies primarily on agreement; every example I can find is (as 
Thomas noted) adjunct-head languages, and adjunct-head languages tend to be 
case-y. My guess would be that IHRCs wouldn't require any special treatment; 
the verb would agree with the head in the same manner whether it was 
internal or external to the relative clause. 

For any of these strategies, you could also try...

1) Do the inverse of Malagasy: require that only the subject can be 
expressed by an IHRC and use the inventory of voices to realize the argument 
as subject. I like this one, although it forbids two IHRCs in a sentence. 
But plenty of languages have similar restrictions, like forbidding the 
question "Who killed whom?"

2) Require "equi-type" relative clauses, in which the head plays the same 
role (semantic or grammatical) in both the main clause and the relative 
clause.

3) Do nothing. Examples in which real ambiguity occurs are pretty rare, 
except when taken out of context and used as examples of ambiguity.

Anyway, there're some ideas for ya.
-- 
Patrick Littell
PHIL205: MWF 2:00-3:00, M 6:00-9:00
Voice Mail: ext 744
Spring 05 Office Hours: M 3:00-6:00

Reply