Handle po

[lynn]

This is much like #4 but also there's this "po opacity" thing to worry about.

[solpahi]

po being opaque means that quantified expressions in the complement of po are bound below po, rather than above it as they are in the case of ordinary verbs.

Naively:

This analysis is problematic for two reasons: 1) semantically, there is a type mismatch between the QP and the DP 2) it lacks an explanation for why QP can't take higher scope

We know that CPs are scope boundaries: In hâo sa rảı, sa rảı must be bound by a QP within the same CP; it can't be bound by a QP in a higher CP.

If we assume that CPs are the only scope boundaries, then the fact that in po sa rảı, sa can't be bound by a QP higher than po points to the presence of a C somewhere between po and sa.

For example:

The CP can be thought of as a free relative or a property. The CopV head probably just means ıq, and its spelling tells us what the null V in the CP is.

So it comes out as "to be such_i that ∃raı. hao(PRO_i, raı)"

Likewise for jeı:

"to be such_i that ∃raı. jeıga(PRO_i, raı)"

And finally mea, this time without QP:

[lynn]

I implemented this: