Category Archives: syntax

Syntactic intrigue…

An intriguing session today*, in several ways. I was doing straightup elicitation and translation, which in itself is surprising since that hasn’t been very successful in the past, but it was working extremely well today.

Second, I got confirmation of the obviation (or better to call it quasi-obviation). Bardi has this thing where there are two forms of the direct object and oblique agreement markers. One has a jarr- in front of it, basically. So, you can say either ana=ngay or ana=jarrngay for ‘give it to me’. It was my impression from texts and other stuff that there was a strong preference for the jarr-forms when the subject is third person. (They’re also used for a bunch of other things, which I’ve talked about in an article in grammatical and discourse reference – it’s on my web site and is in press, I believe.) Today I was able to get confirm that, indeed, there’s a person hierarchy component to this marking.

HOWEVER, and this is a change in Bardi from even 10 years ago, these forms can now be used with 3rd person oblique marking too. They weren’t possible when Gedda Aklif tested for them, and when I tested them (with another speaker) in 2001. But today I got this sentence:













“John told Mary about himself.” [Yes, I’m doing reference again…]

This looks a bit like the third person plural possessive and the lenited third person singular oblique, but there’s no plural in the phrase. It’s the form that you’d expect for a third person oblique –jarr-form (as I’ve been calling them).

Further testing shows that you can get forms like this with 3rd subjects (co-reference is not necessary) but not if there’s a 1st person subject, which is cool. They don’t seem to have the same behaviour as the speech act participant forms (as in they don’t do topic tracking) and I still can’t find any 3rd person direct object forms, or 3pl oblique forms.

So, what’s unusual here? On the one hand, not much, since it was always odd that these forms didn’t exist with overt third person forms, the fact that they do exist isn’t weird. If they really are obviation-type items, it’s not surprising they are concentrated on speech act forms, but since the third person forms so exist, it’s *very* surprising there isn’t an example in any of the other materials. If it’s a ‘last speaker’ reanalysis, it’s a little surprising, since the corpus has been pretty constant syntactically for the last 30 years or so.


*I’m a bit ahead with blogging, so ‘today’ was actually Friday.


Cancellable predicates

An off-chance English comment about ‘drowning’ led to an opportunitistic question or two. It turns out that Bardi has the same sort of cancellable semantics in certain predicates that have been reported for some Asian languages.

For example, in English, the sentence “He killed the dog, but it didn’t die” cannot be true in any real world sense, because ‘kill’ implies that the person does actually die. But in a number of languages, the semantics of ‘kill’ are ‘cancellable’, in that the outcome of verbs like ‘kill’ can be ‘cancelled’ by subsequent sentences. This is true in Bardi too, for verbs like ‘die’ and ‘drown’. In Bardi there are sentences like ‘The kid drowned, but we pulled him out and he’s fine now.’