(sending it again as it went to Pete only by mistake. Pete, I suggest you
check your "reply-to" settings, they're overwriting the list's)

On 12 February 2014 11:30, PETER BLEACKLEY <[log in to unmask]>wrote:

> I've been thinking about a the possibility of a split subject system where
> the choice between Nom/Acc and Erg/Abs was made by the telicity of the
> verb. The question is, which way round should it go? I can think of
> arguments for both choices.
> Argument A
> Animate agents are more likely to be the agents of telic verbs, inanimate
> agents are more likely to be the agents of atelic verbs. Since agency is
> more marked on inanimates than animates, atelic verbs should take Erg/Abs,
> and telic verbs should take Nom/Acc.
> Argument B
> The result of a telic verb is important, so that makes the patient more
> salient, meaning it should be the unmarked argument with a telic verb, and
> the agent should be marked. Therefore, telic verbs should be Erg/Abs,
> atelic verbs should be Nom/Acc.
> Does anyone have any opinions on which of these arguments holds more
> water, or have any examples of natlangs that do something similar? I know
> that Austronesian languages have (to simplify things a lot) a voice system
> that switches the alignment between Erg/Abs and Nom/Acc. Does that involve
> telicity at all?
> Pete

Mmm... I'm not sure, as I don't think telicity is in any way related to the
nature of the arguments as such. See, so far the best definition of
telicity I found is the one using quantisation and/or cumulative reference,
which makes telicity something very much like the distinction between mass
nouns and count nouns, only for verbs rather than nouns. To simplify: if
you can cut an event described by a phrase P into two pieces, and the
pieces can still be correctly described as P, then P is atelic. If you
can't, it's telic. For instance, think about the event description: "to sit
on a chair". If, for instance, you've been sitting on a chair for 20
minutes, you can describe that by saying you've been sitting on a chair for
10 minutes, and then you've been sitting on a chair for 10 more minutes:
the event "to sit on a chair" can be divided into shorter events that still
can accurately be described as "to sit on a chair". So "to sit on chair" is
atelic. On the other hand, the expression "he built a house" is telic,
because if you cut that event into two subevents, the first subevent cannot
be accurately be described as "he built a house" (since the house cannot be
finished by then, it will only be finished at the end of the second

Things get complicated very quickly with telicity, and while the exact
nature of the arguments of a verb can have an influence on its telicity
(the phrase "he built a house" is telic, while "he built houses" is
atelic), it's not in a way that can easily be captures in a split-ergative
manner in my opinion. It might not be impossible, but it doesn't feel
really natural either way.

The problem I think is that you are talking of telic or atelic *verbs*,
while telicity is more properly a property of verb *phrases*. As I showed
in the "he built a house/he built houses" example, the same verb in the
same tense (at least in English) can be telic or atelic depending on the
object's singular or plural form! So speaking of telic or atelic *verbs* is
really incorrect: you can only say something once you take the whole verb
phrase into account, including possibly number and definition of the object
(the subject is often less relevant to the telicity of an action).
Christophe Grandsire-Koevoets
President of the Language Creation Society (

Personal Website: