Chris: Does one-place eat necessarily imply an unexpressed second argument?

Joshua: What's wrong with leaving a predicate argument unexpressed? If you have a two place relation and one is left unexpressed.

Dan: Could certainly do that for passive. That doesn't seem wrong at all. I was imagining more interesting places like causative -> inchoative, where it seems maybe wrong to have the inchoative as two place; maybe you should have had a two-place predication. But if you have reason to say the two-place predicate is the underlying one.

Joshua: So not formally wrong, just feels wrong?

Dan: There may be regularities in the language that cause you to regret that.

Ron: A parallel semantic operation like existential closure?

Emily: That's why Dan picked causative/inchoative.

Dan: Could tell people to go back through the lexicon and

Ron: How did the door open if nothing caused it too? ... The door opened itself.

Emily: What about a language where the 'drop'/'fall' pair has this property. Doyou really want 'fall' to have an agent?

Martin: The door opened/The door was opened. One implies an agent and the other one doesn't.

Francis: Can't we put in something that says, "I'm the thing that doesn't exist."

Ron: That's exactly what I wasn't thinking about.

Chris: The eaxmple from the VLAD session was benefactive in Lakota combined with a repreated de-objective. I cooked food. I cooked food for my family. I cooked-deobj-deobj (things for someone).

Guy: In that case it seems quite valid to have them in the semantics as unexpressed arguments. Wouldn't it be sensible to then ask "What did you cook?" But The leaf fell./#What dropped it?

Joshua: Or #/Who did it fall for?

Dan: Maybe bracket this as a genuine worry regarding maintaining integrity in the arity of these predicates, but if the issue is only with causative, we'll waive our hands.

Alex: I want to pour cold water on using follow up questions as test for what the MRS shoudl look like. That's really testing whether the presuppositions that come with the question can be accommodated:

The door opened. Who opened it? The flower opened. #Who opened it.

=> difference turns on world knowledge.

Guy: Is there a better test?

Alex: Can't think of one, but don't use that one.

Berthold: Farrell Ackerman on Moro where indirect and direct objects can passivize & there's a generalized applicative, turning an adjunct into yet another complement that can passive. Did you look at which additional argument will/will not be available for the next layer of valence changing rule?

Chris: Starting from the supposition that they're all available, because Lakota. There are also limits that I need to look

Mike: Can you drop more than one argument? / You bet you can.

Emily: That's argument optionality, not Chris's problem (but see Safiyyah's library).

Emily: [Scribe's post-script --- see VLAD session on this. ]