Beruflich Dokumente
Kultur Dokumente
Syntax 1 Final
12/6/2010
Collaborated with: Lila Hunt, Zoe Lu, Dima Zadorozhny, Andrew Pedelty
Since we cannot have idiom chunks or dummy subjects in the upstairs subject position
without a passive with "intend," (1)-(2), intend must be a subject-subject raising verb. If it is not a
subject-subject raising verb, and still has a missing subject in the downstairs clause, it must be a
subject-subject control verb. Since you cannot passivise "intend," (3)-(4), it must not be a subject-
Since we can have dummy subjects or idiom chunks in the upstairs subject position
without a passive with "be going," (5)-(6), we can tell that "be going" is a subject-subject raising verb.
Since it is subject-subject raising, it cannot be subject-subject control. Since you cannot have a
Since we can have dummy subjects or idiom chunks in the upstairs subject position
without a passive with "be sure," (8)-(9), we can tell that "be sure" is a subject-subject raising verb.
Since it is a subject-subject raising verb, we can know it is not a subject-subject control verb. Since we
Sam Sanders
Syntax 1 Final
12/6/2010
Collaborated with: Lila Hunt, Zoe Lu, Dima Zadorozhny, Andrew Pedelty
cannot have a grammatical sentence with a passive, (10), "be sure" cannot be subject-object raising
verb.
Since we cannot have an idiom chunk in upstairs specifier position without a passive,
(11)-(12), we can tell that "refuse" is not a subject-subject raising verb. Since it is not a subject-subject
raising verb, we can tell that it must be a subject-control verb. Since we can see the idiom chunk in (13)
being raised to upstairs specifier position, we can tell that "refuse" is a subject-to-object raising verb.
Since we can have dummy subjects or idiom chunks in the upstairs subject position without
a passive with "need," (15)-(16), we can tell that "need" is a subject-subject raising verb. Since it is a
subject-subject raising verb, we can know it is not a subject-subject control verb. Since we cannot have
a grammatical sentence with a passive, (17), "need" cannot be subject-object raising verb.
Since we cannot have a an empty downstairs specifier position without a passive, (18),
we can tell that "assume" is neither subject-subject control or subject-subject raising. Since we can
have an idiom chunk from the downstairs clause after a passive, (19), we can tell that "assume" is a
Sam Sanders
Syntax 1 Final
12/6/2010
Collaborated with: Lila Hunt, Zoe Lu, Dima Zadorozhny, Andrew Pedelty
Since we cannot have a an empty downstairs specifier position without a passive, (20),
we can tell that "find" is neither subject-subject control or subject-subject raising. Since we can have an
idiom chunk from the downstairs clause after a passive, (21), we can tell that "find" is a subject-object
raising verb.
We can tell that idiom chunks can't be in the upstairs specifier position, even with a
passive, (26)-(27), so we can tell that "advise," is not a subject-subject raising verb. Because it's not
even ok to have certain other things as the upstairs specifier after a passive or as the upstairs
object/downstairs specifier before a passive, (22)-(25), we can tell that there are restrictions about
which things can be the specifier of the downstairs clause. When this is the case we can tell that the
upstairs verb is an object-subject control verb. If we say that "advise" is an object-subject control verb,
we don't currently have the tools to determine if it is a subject-subject control verb, as we have no way
of knowing whether the downstairs specifier of (23) was "the students" or "John" in the deep structure,
so I will simply assume something cannot be both subject-subject and object-subject control for now.
Sam Sanders
Syntax 1 Final
12/6/2010
Collaborated with: Lila Hunt, Zoe Lu, Dima Zadorozhny, Andrew Pedelty
Since we can see that have7 can either have or not have the downstairs specifier,
(29)-(30), we can tell that have7 works as a subject-subject control verb. We can also tell that it is an
object-object control verb, because you can be missing a downstairs direct object, but only if it's not an
idiom chunk.
Since we can have an idiom chunk or dummy specifier in the upstairs specifier
position without the proper transformation or a passive transformation, we can tell that have8 is a
Since we can have an idiom chunk or dummy specifier in the upstairs specifier
position without the proper transformation or a passive transformation, we can tell that use5 is a
II. In order to check the transformations occurring in sentences with tough-type adjectives, I will use
simpler sentences than those given in the examples to try and eliminate anything that is confounding.
The set of sentences I will be using in order for us to get a better idea about these types of adjectives, I
Sam Sanders
Syntax 1 Final
12/6/2010
Collaborated with: Lila Hunt, Zoe Lu, Dima Zadorozhny, Andrew Pedelty
What pattern do we see in these sentences? We can tell from (38) that "hard" is neither a raising or a
subject-subject control predicate. We can tell it's not an object-object control predicate because there is
no direct object in (37). It can't be object-object control for the same reason, that there is no direct
In (39) we see an idiom chunk as the upstairs specifier without a passive transformation, so there
has to be movement. Why does this tell us there has to be movement? Since the idiom chunk could not
have been in the upstairs specifier position of the deep structure, then it must have been moved there
somehow. However, there is the possibility that one may not consider (39) to be grammatical. If this is
the case then consider the following sentence, which is less dubious.
Since we can see that tough-type predicates move the downstairs object into upstairs specifier
I won't name this yet, because it seems to be deeply intertwined with the other possible
transformations we see occurring with tough-type predicates. One more final point before moving on -
Sam Sanders
Syntax 1 Final
12/6/2010
Collaborated with: Lila Hunt, Zoe Lu, Dima Zadorozhny, Andrew Pedelty
since (39) seems dubious, while (42) does not, there seems to me to be something about this
transformation not being possible for some when there is a downstairs specifier which is not indefinite.
Now the question becomes - what other transformations do we see? Considering (40), it also
seems possible to move the CP complement of a tough-type predicate into upstairs specifier position.
Would prove this idea wrong. However, consider the fact that we can see the same type of
This has nothing at all to do with tough-type adjectives, and everything to do with having an
indefinite or definite deep structure specifier in embedded clauses. So, if we ignore these special cases,
it's fairly clear to see that tough-type predicates raise their CP complements to upstairs specifier
position. Since this is still a tough-type predicate raising something, and it goes without saying that you
cannot raise both an entire embedded clause as well as the object of the embedded clause, we can
C: Raise either the downstairs object or the CP complement to the predicate to upstairs
specifier position.
This explains (39) and (40) readily. However, we still need to account for (41). Looking at
(41) we can tell that there is a missing DP complement to the P "on." We could question whether this is
would also be saying that there are two different subcategorizations for tough-type predicates. Why is
Sam Sanders
Syntax 1 Final
12/6/2010
Collaborated with: Lila Hunt, Zoe Lu, Dima Zadorozhny, Andrew Pedelty
this? Well, for transformations to raise something into upstairs specifier position, they must be 1 theta
role predicates that subcat against a deep structure specifier, and control predicates are the exact
opposite. Working off the concept that we are seeing the complement of a P being raised, we now need
to consider when it can be raised. The "on Murphy" portion of the deep structure of (41) may or may
These sentences show pretty strongly, if they are grammatical, that the PP does not have to
be complementing the VP, but can also be modifying it, for tough-type predicates to raise the P's
complement. So, there seems to be one more possibility we need to account for in the possible
transformations associated with tough-type predicates. To account for these I will propose one more
There is one final question to answer about these - how bounded are these transformations?
Considering these sentences, we can see that the CP movement is bounded so that it is
only the CP that is complementing the tough-type predicate that can be raised. However, we can also
tell that the raising of the DP complementing the P or the VP can be raised no matter how many clauses
below the tough-type predicate it is. This means that we should actually split this into two
CP Raising
DP Raising
C: Raise any DP that is an object of any clause more downstairs than the one containing
the tough-type predicate or is in a PP in a clause that is more downstairs than the predicate
There is still one more point to be made about these transformations, which I stated at
the start, but that I need to restate now. Even though I think it is important to have these as two seperate
transformations because of how they are bounded differently, you cannot do both of them. That is, you
cannot move one of the DPs and then move the CP that that DP was in.
III. The difference between "ready" and tough-type predicates is simple when you understand that
tough-type predicates are raising while "ready" is control. The easiest way to see this is in testing the
grammaticallity of the two in a similar surface structure with an idiom chunk. Consider the following.
Sam Sanders
Syntax 1 Final
12/6/2010
Collaborated with: Lila Hunt, Zoe Lu, Dima Zadorozhny, Andrew Pedelty
The patterning is immediately obvious with (54) and (55) - idiom chunks do not work as the
upstairs specifier with "ready." This is dichotomy is usually due to one predicate being control and one
predicate being raising, and this pattern holds in this instance. The other reason it becomes clear that
"ready" is a new type of control verb is because of the ambiguity. (57) gives us further reason why we
should consider "ready" to be a control predicate and not a raising predicate - we can have grammatical
surface structures with the downstairs clause having and not having a passive transformation. Why
does this show this? Because if it was an object-subject raising predicate like tough-type predicates,
then when we use the passive on the downstairs clause, which removes the downstairs object, we
would no longer be able to get grammatical surface structures, because we wouldn't be able to do the
required raising. I will now propose the new transformation that we see in "ready" that will also be a
stepping stone in explaining the ambiguity in (59) that does not exist in (60).
Subject-object control
D: Take a sentence with a subject-object control predicate that has an upstairs specifier co-
Sam Sanders
Syntax 1 Final
12/6/2010
Collaborated with: Lila Hunt, Zoe Lu, Dima Zadorozhny, Andrew Pedelty
Now that I have that proposed, how does this account for the ambiguity of (59)? This accounts
for it because we can tell that "ready" is also a subject-subject control predicate because of (61).
Because it can control either the downstairs object or subject, the deep structure of (59) could have
IV. (64) The kids stopped being hard to get ready to send to school when they learned to tie their own
shoes
This sentence is hiding many, many problems and secrets. To understand fully what is going on,
we first have to step by step determine what the main verbs in the sentence are. First, we have to
determine the relation of the three words "stopped" and "being." We see "be" in the present participle,
while we see "stop" in the past tense, which is a hint at their relation. Perhaps "stop" is a new auxiliary
So, we can see from these sentences that stop is in fact a new type of auxiliary verb that subcats
against having a be1 compliment. This solves part of the problem, as we can now tell the relation
between "stopped" and "being" in (64). Since we have been discussing tough-type adjective-predicates
earlier in this paper, it is not necessary to discuss the relation of being and hard for now. Since we saw
earlier that you can pull the DP direct object of any clause more downstairs than the tough-type
predicate, that means that "the kids" must have been the direct object of one of the clauses after
"stopped being hard." I will get back to determining which verb "the kids" was the direct object of later.
Next, I will discuss the T' "to get ready." We can see that there is the adjective-predicate
"ready," as well as the predicate "get." We have discussed ready prior in this paper already, so the
bigger question becomes what type of verb is "get." Consider the following sentences.
We can see two unique things going on with "get" in these sentences. First, when you consider
(73)-(75), we can see something that we haven't seen before, but that is not expected - "get" is
subcatting against specific types of adjectives. I would assume that this would have to do with some
kind of strange theta-role, but that is far too difficult to determine for this paper. However, (75)-(77)
show us something very interesting - if you consider that (76) is equivalent to (77) and not (75), then it
Sam Sanders
Syntax 1 Final
12/6/2010
Collaborated with: Lila Hunt, Zoe Lu, Dima Zadorozhny, Andrew Pedelty
seems that "get" has another new transformation type that follows patterning similar to other control
transformations - a subject-upstairs object control verb. What does this have to do with (64)? This
means that "get" has a direct object, and if I formulated my transformations for tough-type predicates
correctly, then we have our direct object that has been raised. Consider just the clause "to get ready to
send to school." Who is the indefinite specifier getting ready? I believe that intuitively, it is "the kids."
This agrees perfectly with our observations about "get" and "hard," so it is a good choice. I say that the
specifier of the clause is indefinite because "hard" does not have an outside theta role, so it cannot be
subject-subject control, so this is easiest explained by way of indefinite embedded specifier deletion.
The last int he series of verbs before "when" is "send." What sorts of patterns do we see with
What is important in these sentences is that "send" requires a direct object. While I can still
ignore the empty specifier because of indefinite specifier deletion. However, the CP containing "send"
is complementing "ready." This means that, if my analysis earlier was correct, then we can be safe in
saying that if the specifier in the clause containing "ready" is the same as the object in the clause
complementing "ready," we can delete the downstairs object. Lets now reexamine the clause containing
"ready." This clause is "to get ready to send to school." If my earlier analysis that the thing that is being
gotten ready is "the kids," then that would mean that "the kids" is also the thing that is being sent to
school. However, this is not quite right. Consider the following sentences.
What do these two show us? If we recognize the ambiguity in (81), and the ungrammaticality of
(82), then we can see that "ready" also makes other predicates, or perhaps get ready is a single
predicate like "be going," act as object-object control. This accounts for (64) in an appropriate way. I
(83) Stopped being hard for [indefinite] to get the kids ready for [indefinite] to send the kids to
school
The most downstairs kids would be deleted because of "get ready" acting as an object-object
control predicate, then we would see the more upstairs "the kids" being moved into the most upstairs
specifier position because of the tough-type raising. This accounts for everything before the clause
starting with "when." Why do I say that that is all accounted for? Consider the fact that the sentence is
grammatical with or without "when they learned to tie their own shoes."
"Learn" is rather uninteresting, and thus the clause following "when," as it is simply a subject-
Thus, the final question is - what is "when?" First, as I showed, it seems to be modifying,
because the sentence is grammatical with or without its phrase. Admittedly, I am assuming that "when,"
is the head of the clause/phrase. Now, let's consider when you can see a when phrase. First, I will check
(86) The kids were orange when they learned to tie their shoes
(88) The man brought cookies slowly to the class on the day I slept in when I was sick
What do these sentences show us? They show us that "when" headed phrases are modifying, not
complementing, the VP, because there are not Vs that subcat against them. This suggest that "when" is
a P, and that we are now seeing PPs that have TPs as their complements. If this is true, that "when" is a
PP, then we should see it in other places that PPs normally occur. Consider the following sentence
This seems to suggest this. However, CPs can also complement NPs. What is one place that a
This shows fairly strongly that "when" headed phrases are CPs. So, then what is the final
change we need to make to our grammar to account for this? We need to change it so that CPs can
modify VPs when the have certain heads like "when," and include "when" into the category of C.