y********o
发帖数: 2565 | 1 Two movements are involved in
* I wonder who pictures of were bought by Mary.
Right?
One is A-movement, one is A'-movement.
What ordering do they have? In other words, which should be moved before
which? You have an idea? |
|
d*****u
发帖数: 17243 | 2 Yes, there are both passive and WH-word, so there must be A- and A'-movements
involved.
Given the data, it is not clear which one happens first.
This example is ususally used to illustrate the "subjacency violation". i.e. A
word cannot be extracted across two bounding nodes. |
|
y********o
发帖数: 2565 | 3
A'-movements
A
Yes, you are right!
The D-Structure of this sentence is:
(1) I wonder were bought by Mary pictures of who
Right?
Then, I think we have to posit that NP movement of [NP pictures of who] must
occur before the wh-movement. i.e. A-movement occurs before A'-movement.
We get the following by moving the NP.
(2) I wonder [NP pictures of who_i ] were bought by Mary t_i
Now, up to this point, we cannot move [who], because this would violate the
CED constraint (?) since [who] is |
|
d*****u
发帖数: 17243 | 4 Well, if you want to use the subject island condition, it seems to be true.
However, even if you assume wh-movement occurs first, "who" has to be
extracted out of NP and IP at the same time, and it violates the subjacency.
You may also consider a sentence in which a complex NP is a true object and
contains a wh-word. It is still hard to move the wh-word.
i.e.
(3) |
|
y********o
发帖数: 2565 | 5
Would you please explain why "who" has to be extracted out of NP and IP at the
same time? What constraint forces this? Pied-piping of complex NPs?
I am asking questions, not challenging you. Just in case I am misunderstood.
must
the |
|
d*****u
发帖数: 17243 | 6 "Who" is a part of the complex NP "picture of who". And it has to move to
SpecCP.
However, object is easier to extract, so maybe you can use the data to
construct an argument.
true.
subjacency.
and
the
violation".
A'-movement.
get |
|
