On 26. jan. 2016 12:56, J S Jones wrote:
> Jan17 has a few applicatives, usually increasing the valence.

Heh, I just finished the chapter on applicatives in Basic Linguistic 
Theory yesterday evening.

> I label the arguments of the verb X, Y, and Z. When the verb is
> intransitive, the applicative redefines the X argument to have the
> semantic role specified by the applicative (such as instrument,
> beneficiary, destination, etc.).

According to RMWD@BLT, *canonical* applicatives redefine an existing 
peripheral argument (location, instrument, comitative, goal, recipient) 
as a patient/object, with language-dependent effects on the original 
patient/object if any. If the valency changes and that affects the 
subject, that is due to the valency change, not the applicative. A 
nominative stays a nominative and an absolutive changes to an ergative 
on going from intransitive to transitive for instance.

What he calls "quasi-applicatives" is when there is no possibility for 
an existing peripheral argument, and the applicative is the only way to 
connect the argument to a verb.

> Here's the odd thing: when the instrumental applicative is used on a
> transitive verb and the X argument is a person,

Which type of instrumental applicative? (I'll summarize the types of 
applicatives in the book when I get home.)

According to BLT, applicatives don't mess with the agent/subject, so if 
your stuff does that, you don't have applicatives but something new, 
nice and shiny.