A couple of things I think might be worth saying as I’m reviewing the homework I gave out.

First, as I pointed out in class, the leadup to the last problem in the homework (having to with the “disturbing alternative derivation”) refers to a paper by Fox & Pesetsky that we haven’t read, although I don’t think that should harm your ability to do the problem, since it explains what is meant at the same time.

Second: This homework is fairly solidly cast in terms of the more recent Minimalist syntax in a couple of ways. So, for people who didn’t “grow up with” Minimalist syntax, let me just highlight a couple of things.

At the bottom of p. 2: “Lastly, we’ll assume that Q is introduced as soon as possible as the tree is built up.” This is to be thought about in the context of trees being built from the bottom up, by taking things and linking them together to form constituents (which can then be linked to other things, etc.). E.g., you take the verb see and the pronoun him and put them together to form the VP see him, and then continue on. So, generally, “sooner” is the same as “closer to the bottom of the tree.”

On the problem concerning the “Russian doll” questions, part of this hinges on what a trace actually is. As a reminder (I think this has come up briefly in class a couple of times), the normal view in the 80s was that a trace was really something like a t, a little silent thing that had no particular internal structure, just possibly an index to link it to the thing it is a trace of. One of the things that was introduced (or, perhaps, re-introduced from a long, long time ago) in the original Minimalist Program paper was the idea that the trace of “movement” is not really a trace, but rather just another copy of the thing that moved. This way, we don’t need to have any kinds of rules about how to create traces; moving something is just about attaching (usually “Merging”) a new copy into a higher position in the tree. Then, a “trace” is just something that is c-commanded by a copy of itself.

Thus, (8) on the homework handout. Kolki studenti po kakvo ot Bulgaria is moved, by which we mean it is put into SpecCP. It remains where it was before moving as well, but it isn’t pronounced because it is c-commanded by a copy of itself (this is indicated by the strikethrough).

One of the things I’m trying to encourage you to ponder is the question of what happens to the lower copy/trace. The point of (9) is to highlight the fact that if you have to move two wh-words, the second one will presumably have to move past the trace of the first one, meaning that the trace of the first movement must be somehow “invisible” to the movement operation. It’s laid out perhaps a bit more perspicuously in (10).

I throw around some “checking” and “feature” terminology at this point. Remember this: C needs wh-words, and it is the drive to satisfy this need that forces wh-words to move. One way to say “C needs a wh-word” is to say that “C has a feature that needs to be checked” (where it is understood that moving a wh-word is what is necessary to “check the feature,” i.e. to satisfy the need C had). The “Agree process” referred to under (10) is basically the process of C “looking for a wh-word” to solve its problem. It looks down the tree until it finds something with the property of being a wh-word (That is, having a [wh] feature) and when it finds that thing it “Agrees” and forces movement of the wh-word into SpecCP. C basically “attracts” wh-words.

Under (11) I say “We need to assume something like what Fox & Pesetsky assumed,” though you don’t have any way to know what that is, since we haven’t read that paper yet (though I hope to get to it). The point there is just this, though: We can think about this process of movement as being one where you find something in the tree, make a copy of it, and attach the copy higher in the tree. That’s how I’ve been talking about it so far. Another possibility is something a bit harder to draw, though, which is that you don’t make any copies, but rather just take the thing you found and attach it to another point in the tree. If you do this, the constituent that is undergoing “movement” now basically has two different parent nodes (the one it started under, and the one it ended under). The difference is subtle.

I think this problem might end up being a bit hard, whether you were familiar with feature checking beforehand or not. Give it a try though, we can talk about it in class.