In response to another question over email, about the questions on the practice midterms that ask about how may [uP*] features are in the trees.

The answers to these basically hinge on whether the PP in the tree is getting a theta role or not. If a PP in the tree is getting a theta role, it must have been Merged into that position, which would have to happen in order to check a [uP*] feature.

The way to tell is basically to look at where the PP is in the tree, and what its mother and sister nodes are. First of all, the only kind of theta role that we can assign to a PP is Goal (PP sister of V). So, if any PP gets a theta role, it’ll be in that position in the tree.

Another way to determine this is to look to see if the sister node and mother node are the same (somethingP). That would indicate that the PP was attached to the structure using Adjoin. So, if the sister is vP and the mother is also vP, then the PP does not get a theta role, and so it was not introduced into the tree to check a [uP*] feature. Same if the sister is NP and the mother is also NP, no [uP*] feature was checked when the PP was added to the tree.

Then, you basically just count them up. In a simple sentence with a single verb, the answer really only could have been “zero” or “one”, since introducing a Goal is the only place where we would find a [uP*] feature.

Side note: Above I said Goal is PP sister of V. For ditransitive verbs like give and put, this is the same as saying PP daughter of V′ (which is what I said in class and what Adger says in the book). However, I think it might be more accurate to say that it is the PP sister of V, based on the structure we’d probably want to assign to something like Pat glanced at the handout. I won’t ask any questions on the midterm that rely on this difference, but I just wanted to note it in case anyone looked back at how we handled glance (back before we introduced the “little v“) and wondered how that would work with the UTAH.