Re: "He opened the door and he (same referent) left the room"

En réponse à Trebor Jung :


>How do natlangs handle sentences like in the subject?
Many people have given plenty of answers to that already, so I just want to 
add a little bit here.

As some have said, many langs simply omit the second subject, like in 
English. However, even then, interesting things can happen, because you 
have to take into account the syntactic orientation of the language. 
Indeed, just as a language can be grammatically accusative or ergative (and 
all the other possibilities around them, but let's for now concentrate only 
on those two), i.e. in the way it handles the relationship between the 
grammatical slots available for the main actants and the corresponding 
semantic roles, a language can be syntactically accusative or ergative, in 
the way referents are considered when something is underspecified. 
Basically, in a syntactically accusative language, the sentence: "he opened 
the door and left the room" would mean unambiguously: "he opened the door 
and he (same referent) left the room". However, in a syntactically ergative 
language, the same sentence could only mean "he opened the door and the 
door left the room" (the unspecified subject of an intransitive verb always 
refers back to the *object* of the previous transitive verb. In other 
words, absolutives keep refering to each other).

The interesting thing is that grammatical and syntactical orientation 
needn't match. For instance, Basque is grammatically as ergative as one can 
imagine. However, syntactically it's strongly accusative (so that the 
sentence "he opened the door and left the room" means exactly the same 
thing in English and in Basque, despite the fact that English is 
grammatically accusative and Basque grammatically ergative).

Now add to that all the possibilities between full accusativity and full 
ergativity (Split-S, Fluid-S, Active, or some different things like trigger 
systems) and think that you can have each of those, relatively 
independently (I said "relatively" because there seem to be limitations. I 
for instance know no grammatically accusative yet syntactically ergative 
language) on both the grammatical and the syntactical dimension, and you 
can already have quite a few exotic things going on ;)) .

Christophe Grandsire.

http://rainbow.conlang.free.fr

You need a straight mind to invent a twisted conlang. 

Replies