Re: Help in Determining Asha'ille Typology

Andreas Johansson sikyal:

> Quoting JS Bangs <jaspax@...>:
>
> > These are the wrong kind of examples to use for deciding whether a
> > language is accusative, ergative, or active. These terms have to do with
> > the marking of arguments to a verb and transitivity, so we'd need to see
> > some examples of that. Can you post translations of these three sentences:
> >
> > 1) I eat food.
> > 2) I run.
> > 3) I fall.
> >
> > In most general terms:
> >
> > An accusative language is one in which "I" in all three sentences is
> > marked the same (nominative), while "food" is marked differently
> > (accusative).
> >
> > An ergative language is one in which "food" from (1) and "I" from (2) and
> > (3) are marked the same (absolutive), while "I" from (1) is marked
> > differently (ergative).
> >
> > An active language is one in which "I" from (1) and (2) is marked the same
> > (agentive), while "food" and "I" from (3) are marked the same
> > (patientive). This is subject to a lot of language-specific variation,
> > though, so beware.
>
> What would we call a language that marks "I" from (1) the same as "I" in (3),
> and "I" in (2) the same as "food" in (1)? Beyond weird, that is.
Maggel? I doubt that such a situation is common enough to have its own
name.

I put a little more thought into this today, and realized that you could
deductively arrive at the conclusion that there are three broad types of
languages, categorized by how agency and transitivity interact in the
syntax.

The four roles I indicated above are equivalent to the four quadrants in
the following diagram. The top row represents the agency of the argument,
and the side row represents the transitivity of the verb:

       +agt    -agt

+trans  (1)    (2)
-trans  (3)    (4)


If we allow only two morphosyntactic categories in which to divide these,
what are the possibilities? If we allow anything, there are eight ways to
group them. However, (1) and (2) are the only two quadrants that can
co-occur in the same utterance, so we will probably want these to be
separate for intelligibility. That brings the possibilities down to four.

(Of course, there is at least one counterexample to this, the "monster
raving loony language" in Iran. But it proves the rule: there's only *one*
counterexample.)

The four remaining possible groupings are these:

A: 1,  B: 2,3,4 = ergative
(A is ergative, B is absolutive)

A: 1,3,4, B: 2 = accusative
(A is nominative, B is accusative)

A: 1,3, B: 2,4 = active
(A is agentive, B is patientive)

A: 1,4, B: 2,3 = ????

This last one is what Andreas talked about. However, we should notice that
this groups the points that have nothing in common together, which is
completely counter-intuitive and unlikely. Andreas' Law of Freaks demands
that some language does it, but I don't know what it is.

Additional thoughts: One really should add a third dimension "topicality"
to this, and see what possibilities occur there.

Also, what happens if you use three categories? I know that this occurs:

A: 1, B: 2, C: 3,4 = Overspecified, but possible

But what about

A: 1, B: 2,4, C: 3
A: 1,3, B: 2, C: 4

Interesting ideas for a conlang . . .

--
Jesse S. Bangs jaspax@u.washington.edu
http://students.washington.edu/jaspax/
http://students.washington.edu/jaspax/blog

Jesus asked them, "Who do you say that I am?"

And they answered, "You are the eschatological manifestation of the ground
of our being, the kerygma in which we find the ultimate meaning of our
interpersonal relationship."

And Jesus said, "What?"

Reply