CP

Back to C
C¢
TP
CAS LX 522 business… [–Q]
T¢
DS

Syntax I n
n
Mary is likely to leave.
Mary starts in SpecVP,
T
[pres]
VP
V¢
gets a q-role from leave. AdjP
V
be Adj¢

Week 8. Control and PRO Adj q TP

likely
T¢

T VP
to
DP i V¢
V
q Mary leave

CP CP
C¢ C¢
Recall… C
[–Q]
TP Recall… C
[–Q]
TP
SS
T¢ DP i T¢
Mary (Note how we
n Mary is likely to leave. Vj+T VP n Mary is likely to leave. Vj+T VP write multiple
n Mary starts in SpecVP, be+[pres] V¢ n Mary starts in SpecVP, be+[pres] V¢ traces)
gets a q-role from leave. tj AdjP gets a q-role from leave. tj AdjP
n Mary moves up to the Adj¢ n Mary moves up to the Adj¢
embedded SpecTP to embedded SpecTP to
Adj q TP Adj q TP
satisfy the EPP. satisfy the EPP.
likely likely
n Mary still doesn’t have DP i T¢ n Mary still doesn’t have t i¢ T¢
Case. Mary Case.
T VP T VP
to n Mary moves up to main to
ti V¢ clause SpecTP, satisfying ti V¢
V the EPP and getting Case. V
q leave q leave

CP
C¢
Recall… C
[–Q]
TP
SS Reluctance to leave
DP i T¢
Mary
n This happens because Vj+T VP
n Now, consider:
likely assigns only one be+[pres] V¢ n Mary is reluctant to leave.
q-role, an internal q-
tj AdjP
role. Adj¢
n This looks very similar to Mary is likely to leave.
Adj q TP n Can we draw the same kind of tree for it?
n Likely does not assign
likely
Case, and so Mary t i¢ T¢
must keep moving, n How many q-roles does reluctant assign?
both to satisfy the EPP T VP
and to get Case. to
ti V¢
V
q leave

1
 Reluctance to leave Reluctance to leave

n Reluctant has two q-roles to assign. n In Mary is reluctant to leave,

n One to the one feeling the reluctance (Experiencer) n Mary is doing the leaving, gets Agent
n One to the proposition about which the reluctance holds from leave.
(Proposition)
n Mary is showing the reluctance, gets
Experiencer from reluctant.
n Leave has one q-role to assign.
n To the one doing the leaving (Agent).
n And we have a problem:
n In Mary is reluctant to leave, what q-role does Mary n Mary appears to be getting two q-roles,

get? in violation of the q-criterion.

TP TP

Reluctance… Mary
DP i T¢ SS Reluctance… Mary
DP i T¢ SS
Vj+T VP Vj+T VP
is V¢ is V¢
n Mary is reluctant to leave. n Mary is reluctant to leave.
tj AdjP tj AdjP
n Reluctant assigns its q- n There must be something
roles within AdjP as ti Adj¢ there, getting the q-role ti Adj¢
required, Mary moves and satisfying the EPP.
q q
up to SpecTP in the Adj q TP Adj q TP
main clause by SS. reluctant reluctant
T¢
n But we can’t see it. T¢
? ?
n But what gets the q-role T VP n It’s a phonologically T VP
from leave, and what to
empty (Ø) DP. We will
to
satisfies the EPP for the ? V¢
call it PRO. ? V¢
embedded clause? q V q V
leave leave

TP TP

Reluctance… Mary
DP i T¢ SS Reluctance… Mary
DP i T¢ SS
Vj+T VP n Mary is reluctant Vj+T VP
is V¢ [PRO to leave]. is V¢
n Mary is reluctant to leave.
tj AdjP tj AdjP
n There must be something
there, getting the q-role ti ti
Adj¢ n PRO does not get Case. Adj¢
and satisfying the EPP.
q n *Mary is reluctant Bill to leave. q
Adj q TP Adj q TP
reluctant n In fact, PRO cannot get Case. reluctant
n But we can’t see it. DP k T¢ n *Mary is reluctant for to leave DP k T¢
PRO n Mary is reluctant for Bill to PRO
n It’s a phonologically T VP leave T VP
to to
empty (Ø) DP. We will tk n PRO refers (like a pronoun tk
V¢ V¢
call it PRO. V or an anaphor) to Mary. V
q leave q leave

2
 If there’s a PRO, If there’s a PRO,
how do we know? how do we know?
n Mary is reluctant [PROm to leave] n Best method for finding PRO: Count the q-
n Maryi is likely [ ti to leave]. roles. If there appear to be fewer
arguments than q-roles (in a grammatical
sentence), there must be a PRO.
n These two sentences look very much
alike—when faced with a sentence that
looks like this, how do we know which n Another way is to try with idioms like The
kind it is? cat is out of the bag or The cat’s got your
tongue or The jig is up.

Idioms Idioms
n For something to have an idiomatic
interpretation (an interpretation not n It is ok if the pieces of the idiom move
literally derivable from its component away after DS, we can still get the
idiomatic interpretation:
words), the pieces need to be very close
n [The cat]i is likely ti to have your tongue.
together at DS.
n [The cat]i is likely ti to be out of the bag.
n It is likely that the jig is up. n [The jig]i is likely ti to be up.
n It is likely that the cat is out of the bag. n The important thing is that they are
n It is likely that the cat has your tongue. together at DS (the q-role needs to be
assigned by the predicate to the noun)

Idioms Idioms
n If we break up the pieces, then we lose the
idiomatic interpretation and can only get the n The reason for this is that the idiomatic
literal meaning. subject and the idiomatic predicate were
n The cat thinks that it is out of the bag. never together…
n The cat thinks that it has your tongue. n The cat is reluctant [PRO to be out of the bag]
n With PRO sentences (“control sentences”), we n The cat attempted [PRO to have your tongue]
also lose the idiomatic reading. n The jig tried [PRO to be up]
n #The cat is reluctant to be out of the bag.
n #The cat attempted to have your tongue.
n #The jig tried to be up. n Unlike with raising verbs:
n [The jig]i is likely [ ti to be up]

3
 Control Subject and object control

n PRO is similar to a silent pronoun; it n There are actually two different kinds of
gets its referent from somewhere “control verbs”, those whose subject
outside its sentence. In many situations, controls an embedded PRO and those
however, PRO is forced to co-refer to a whose object does.
preceding DP, unlike a pronoun.
n Billi thinks that he i/j is a genius.
n Billi is reluctant [PROi to leave]
n Billi is reluctant PRO i/*j to leave.
n reluctant is a subject control predicate
n We say that PRO is controlled (here n Johni persuaded Billj [PROj to leave]
by the matrix subject). n persuade is an object control predicate

PROarb “Control theory”

n For now, what control theory consists of is just
n Finally, there is a third use of PRO, in which marking the theta grids of specific predicates
it gets arbitrary reference and means something (persuade, reluctant) with an extra notation that
like “someone/anyone”. indicates when an argument is a controller.
n [PROarb to leave] would be a mistake. reluctant Experiencer Proposition
controller

n The conditions on which interpretation PRO i j

can/must get are referred to as Control
persuade Agent Theme Proposition
Theory, although to this day the underlying
controller
explanation for Control remains elusive.
i j k

“Control theory” The PRO conundrum

n Predicates that have a controller marked are n Back when we talked about Binding Theory, we
control predicates. When the controller is the said that DPs come in one of three types, pronouns,
external argument, it is a subject control predicate, anaphors, and R-expressions.
otherwise it is an object control predicate. n PRO is a DP, so which kind is it?
reluctant Experiencer Proposition n It gets its reference from elsewhere, so it can’t be an R-
controller expression.
n It is sometimes forced to get its referent from an
i j antecedent, like an anaphor and unlike a pronoun.
n But that referent is outside its clause, meaning it can’t be
persuade Agent Theme Proposition an anaphor (the antecedent would be too far away for
controller Principle A). Plus, it’s not always forced (PROarb), like a
pronoun.
i j k

4
 The PRO conundrum The PRO conundrum
n Back when we talked about Binding Theory, we
n These weird properties of PRO are sometimes
said that DPs come in one of three types, pronouns,
anaphors, and R-expressions. taken to be the cause of another generalization
about PRO (the “PRO theorem”)
n PRO is a DP, so which kind is it?
n PRO cannot get Case.
n Conclusion: It doesn’t seem to be any one of the
three. It doesn’t seem to fall neatly under Binding n That is, PRO is forbidden from any position
Theory where Case would be assigned to it (hence, it
cannot appear in SpecTP of a finite clause—only a
n …hence, we need “Control Theory” to deal with nonfinite clause)
the distribution and interpretation of PRO.

PRO: One possible

Control Theory
piece of support
n Despite the fact that PRO does not submit n Let’s think back to Binding Theory.
to Binding Theory, there are some binding- n Principle A says that anaphors must be
theory-like requirements on control of PRO. bound within their binding domain, and we
take binding domain to be the clause.
n PRO is only obligatorily controlled by a c- n *Bill wants [Mary to meet himself]
commanding controller. n However, now consider:
n Bill is reluctant to buy himself a gift.
n Bill promised Mary to buy himself a gift.
n [Billj’s mother]i is reluctant [PROi/*j to leave]
n Why are these allowed?

PRO: One possible

PRO: recap
piece of support
n Billi is reluctant [PRO i to buy himselfi a gift]
n Although we can’t see that PRO is there, all
n Billi promised Mary [PROi to buy himselfi a gift]
of our theoretical mechanisms point to its
n *Billi promised Mary j [PROi to buy herselfj a gift]
being there.
n *Billi promised Mary j [PROi to buy him i a gift]
n EPP says that clauses need a subject.
n Billi promised Mary j [PROi to buy herj a gift]
n *Billi is reluctant [PROi to buy him i a gift]
n The q-criterion says that there must be exactly as
many arguments as q-roles.
n While it’s true that Bill is outside of the binding
n Binding Theory indicates something is present
domain of himself, and hence Bill cannot be the inside embedded clauses.
antecedent for himself, PRO is in the binding domain
and its reference is controlled. n If the rest of our theory is right, it seems that
PRO must be there.

5
 Back to raising Back to raising
n In fact, nothing keeps us from piling raising
n So far, we’ve only talked about is likely, but there
are a couple of other raising verbs as well.
verbs one atop the other:
n [The cat] i seems [TP ti to be out of the bag]. n [The cat]i seems [ ti¢ likely [ ti to get his tongue]].
n [The cat] i appears [TP ti to have his tongue]. n [The jig]i began [ ti¢¢ to seem [ ti¢ likely [ ti to be
n [The jig]i proved [TP ti to be up]. up]]]
n [The cat] i began [ TP ti to get his tongue].
n What these verbs have in common is that they n In these cases, the subject moves from
have no external q-role and an internal
Proposition q-role.
SpecTP to SpecTP, only receiving Case at the
last stop, satisfying the EPP at each TP.

Back to raising Side note: Chains

n Raising verbs will cause anything in a
n [The jig]i began [ ti¢¢ to seem [ ti¢ likely [ ti to be up]]]
complement TP that isn’t getting Case to move
up to their SpecTP. n Some time ago we saw the term chain applied to
n Passive arguments: the concept of positions occupied by a (moving)
n [The sandwich]i seems [ ti¢ to have been [ eaten ti]] constituent in a structure.
n Even expletive it: n Here, the chain for The jig is:
n It i began [ t i to rain] n ( [The jig]i, ti¢¢ , t i¢ , ti )
n It i began [ ti¢¢ to seem [ ti¢ likely [ ti to rain]]] n …referring to all the places its been in the tree.
n Here, it was inserted to satisfy the EPP in the
most embedded TP, but then raised from SpecTP
to SpecTP to satisfy the rest of their EPP
conditions.

Side note: Chains Italian subjects

n [The jig]i began [ ti¢¢ to seem [ ti¢ likely [ ti to be up]]] n Many languages have the property that
n Chain: ( [The jig]i, t i¢¢ , ti¢ , ti ) when the subject is understood (often in
n If we consider the chain as a coherent entity, we can the cases where in English we would use a
state conditions in a slightly nicer way:
pronoun subject), it can be just left out
n Every (argument) chain gets exactly one q-role.
n Every (argument) chain receives Case. (except PRO’s)
entirely. For example, Italian:
n Doing this allows us to avoid saying every
argument gets case at some point, and a q-role at n Parlo. Parli.
some different point. speak-1s speak-2s
‘I speak’ ‘You speak’

6
 Italian subjects Little pro
n So what about the EPP and the q-criterion? n There is one important difference between
Clearly ‘speak’ assigns a q-role, and the Italian null subject and PRO, namely
presumably the Italian SpecTP needs to be the null subject in Italian appears in a
filled as well. position that gets Case.
n Io parlo.
I speak-1s ‘I speak’
n This sounds like a familiar question… n Since PRO cannot appear in a Case-marked
should we hypothesize that the subject in position, we have to take this to be
these sentences is PRO? something similar but different: Little pro.

Little pro Features and checking

n Little pro is really just a regular pronoun, only null. n An elaboration…
It doesn’t have the fancy control properties
exhibited by PRO, it appears in Case-marked
positions. n We assume that we have a lexicon full of
items (“words”) that get inserted into
n Languages seem to be divided into those which terminal nodes of the tree. These items can
have little pro and those which don’t, often be considered to be little collections of
correlating with the amount of agreement on the
verb (rich agreement makes it more likely that a properties, or “features.”
language will have pro). Languages with pro are
often called “pro-drop languages” or “null subject
languages”.

Features and checking Features of T

n Now, let’s think about T.
n What do we know about she in English? n English T has features like [past] or [pres],
n It’s a D
and sometimes we’ve written [past] as -ed
n It’s pronounced “she”
to indicate its pronunciation.
n It has nominative Case
n It is 3rd person
n But what determines the (regular)
n It is singular
pronunciation of the affix in T?
n It is feminine n I walk. You walk. He walks. They walk.
n These things we know are all properties, or n I walked. You walked. He walked. They
walked.
features, of the lexical item she. (These are the
grammatically relevant properties anyway…)

7
 Features of T Spec-head agreement
n It seems that both the tense feature and the
person specification of the subject affects how T n The reason it is important for the features to be
is pronounced. close to each other is that the syntax needs to be
n Why? able to check to make sure the features match.
n The modern approach to this phenomenon Spec-head counts as “close”.
(which often goes by the name of Spec-Head n *I walks. *He walk.
Agreement) is to suppose that there are features n If the subject has different person features from
both on T and on the subject (for person, number)
and that when they are in a Spec-Head the tense/agreement suffix in T, then the
relationship, the features are close to each other. sentence is ungrammatical.

SpecTP SpecTP
n Another thing SpecTP is famous for its ability to
host nominative case-marked subjects. n This is really just another way to state the Case
n This is implemented in the same way, by analogy Filter (“DPs need (to check their) Case”) but it’s
to agreement. now in terms of a more specific understanding
n To say that finite T is a nominative case assigner is of what it means to “assign Case”.
to say that it has a feature [(Assign) Nom], and
DPs like I and he have a feature [Nom]. n This also means that the “government radius” is
n A subject “getting Case” in SpecTP is then not a way to characterize the positions which are
exactly getting Case so much as it is checking to be close enough for feature checking to occur.
sure that the Case it has is the right one.
n Case has to be checked (guilty until proven
innocent).

Features and checking Features and checking

n There is a distinction between features n Another point worth observing about checking
that need to be checked and features that features like [Nom] on a DP is that it only
happens once. Once you’ve checked to be sure
do not. that the Case is right, you’re fine—in fact, you
n Case features like [nominative] need to be can’t check it a second time.
checked. These are the kinds of features which
often motivate movement. n For this reason, sometimes people think of the
n Category features like [D] on a determiner are features as being removed when checked (like
fine as they are, they don’t need to be checked on a checklist). Either way, you only check them
against anything else. once.

8
 Case checking seems
Moving to Case positions
symmetrical
n Recall that we said T has a feature [Assign Nom], n Consider:
and this is checked against the [Nom] feature of n It is likely that we will leave.
a subject like we in order to validate the Case on
n *Wei are likely that ti will leave.
the subject.
n There is actually reason to think that both the
[Assign Nom] feature on T and the [Nom] n What’s the problem with the second one?
feature on the DP need to be checked—and that
each can happen only once.
n Finite T needs to check Nom on a DP.
n DPs need to check Case.

Moving to Case positions Moving to Case positions

n *We i are likely that ti will leave. n It is possible to move solely for the EPP if there is
n We moved up to the finite SpecTP, and checked no Case to check (i.e. in a nonfinite TP).
off its [Nom] feature with [Assign Nom] feature of n [The sandwich]i is likely ti¢ to have been eaten ti.
T. Both are now inactivated. n So, we could have moved we to the matrix
n But then we is moved up to the matrix SpecTP. Yet SpecTP—something else went wrong.
we no longer has an active [Nom] feature (it’s been n *We i are likely that ti will leave.
checked already), so the matrix T can’t get rid of n And what went wrong is that this leaves the
its [Assign Nom] feature. matrix SpecTP without a DP to check its [Assign
Nom] feature against.

So where are we? Wh-questions

n The generalizations here are:
n The Case Filter: n This was kind of complicated, but it was
n DPs are inserted into the structure with a Case feature worth going through in order to set up
which must be checked. wh-questions for next time.
n Case assigners are inserted into the structure with a
Case-assignment feature which must be checked.
n To get us started:
n A DP cannot move from a Case-checking
position to another Case-checking position.
n Wh-questions are information-seeking
n Case checking only happens once (de-activating the
(not yes/no) questions, in English
Case feature on both the Case-assigner and the DP) involving one or more of the “wh-words”
(who, where, what, when, why, how, …)

9
 Wh-questions Wh-questions

n In a wh-question, we find that we do the same n With yes-no questions, we posited a [+Q] C
inversion that happens with yes-no questions… at the head of CP, which caused the
(moving T to C). movement of T to C.
n Willi Bill ti eat lunch? n For wh-questions, we can think of a different
n …plus, we move the wh-word into SpecCP: kind of C, a [+Q, +WH] C, which prompts
n Whatj willi Bill ti eat tj ? both the movement of T to C and the
n This movement of wh-words is similar, but movement of the wh-word into SpecCP.
different, from the DP movement we’ve seen so n (So, yes-no questions would have a [+Q, –WH]
far with passives and raising verbs. C)

Wh-questions Wh-questions
n Interestingly, looking at English, [+WH]
n What causes the movement of the wh-word to feature checking appears not as
SpecCP is considered to also be a case of
symmetrical as Case checking. In
feature checking.
particular, moving just one wh-word to
n In this case, the C has a [+WH] feature to
check, and the wh-words have [+WH] SpecCP seems to be sufficient.
features that can be checked against it. n Who gave what to whom?
n So, the wh-word is brought up into SpecCP to n That is, all of the other wh-words can
bring the features close enough for checking, remain, seemingly “unchecked”.
and then presto! everybody wins.

Wh-questions T
T T
n [+WH] C must check its [+WH] feature.
n Wh-words may check their [+WH] feature. T
T T
n In a sense, English wh-movement provides a T T
pretty good motivation for a “feature” view
of these phenomena. It appears that [+WH] C
has a “need” which a wh-feature can satisfy,
T
and once satisfied (even with other wh-words T
around), everything is fine.

10