LECture 3 - That Complements (II)
LECture 3 - That Complements (II)
C.3. Extraposition
The first syntactic property we are going to discuss is EXTRAPOSITION.
The extraposition structure is a characteristic syntactic structure of the English complementation
system. In this structure, the complement clause appears at the right periphery of the
sentence, while the pronoun it appears in the position which should have been occupied by
the clause, thus indicating its syntactic function.
- the clause which has been moved to the right periphery is said to have been ‘extraposed’
- as a typical DP, it occupies the SpecIP position (as a DP, it receives case there)
!Mind you! In cases of Extraposition from P.O., the CP (the complement clause) moves to the
right periphery and ‘it’ is inserted in its place; since 1) the selecting predicate is a
prepositional one (i.e., insist ON, be worried ABOUT, be happy ABOUT/FOR, etc.) and 2) ‘it’
is a DP (and therefore needs case), the preposition must re-appear/be overt in cases of
extraposition from P.O. (27b); when there is no extraposition (27a), the P.O. complement
clause does not move, hence no preposition is needed, since CPs do not need case.
► Syntactically, the extraposed (subject) clause in (28) is adjoined to the VP. It is base-
generated in SpecVP but moves out from there via extraposition and it adjoins to the right
periphery of the VP. Because English sentences require an overt subject, the pronoun it is
inserted in SpecIP, the Nom case position, and the transitive verb checks Acc on the object (me).
Thus, the CP occupies a non-case-marked position, in accordance with the CRP
1
IP
2
DP I’
→it 2
I° VP
-s 2
VP CP
2 [that they like syntax]
tCP V’
2
V DP
surprise me
N.B.: When the matrix predicate is an unaccusative verb, there is no movement of the
complement clause (i.e., no VP-adjunction): being unaccusative, the verb cannot assign
accusative case to its complement, so the complement clause does not need to move to avoid
receiving case because of the Case Resistance Principle. It thus remains in the complement
position of V and ‘it’ is inserted because sentences in English need an overt subject (the EPP
property):
IP
2
DP I’
→it 2
I° VP
-s 2
V’
2
V0 CP
seem [that they don’t hate syntax that much]
► Extraposed direct object clauses are also assumed to be moved to the periphery of the VP,
i.e., VP-adjoined & ‘it’ is introduced in the place initially occupied by the complement clause:
(30) a. John expected [that Mary would accept his marriage proposal]CP → the CP moves to the right
> b. John expected tCP [that Mary would accept his marriage proposal]CP → the CP is VP-adjoined
(by moving, it leaves behind a trace)
> c. John expected it [that Mary would accept his marriage proposal]CP → it insertion
2
Step 1 – movement to the right of the that-complement
IP
2
NP I’
John 2
I° VP
-ed 2
VP CP
2 [that Mary would accept his marriage proposal]
tNP V’
2
V DP
expect tCP
Step 2 - insertion of it
IP
2
NP I’
John 2
I° VP
-ed 2
VP CP
2 [that Mary would accept his marriage proposal]
tNP V’
2
V DP
expect it
OBS: sentences with (direct) object extraposition have a different interpretation from those
without extraposition: whenever it+CP appears (instead of just CP), the event in the complement
clause is seen as a fact already known to the speaker.
3
C.4. The difference between subject it and object it in extraposition structures
1. Subject it
- in subject position it is an expletive pronoun, whose only role is to provide the sentence with
a subject, be it a ‘dummy’ subject (the EPP, see below). So, subject it bears case (Nom.) but it
does not receive a theta-role. As a result, it appears in contexts where lexical DPs (which are
thematic) are banned:
(32) a. It was widely believed [that the world was flat]
b. *The hypothesis was widely believed [that the world was flat.]
- because expletives are not theta-marked, they cannot be questioned, only theta-marked
constituents can be questioned:
(33) a. [That he came] was a blessing. / What was a blessing? > the CP can be questioned
because it bears the theta-role proposition
b. It was a blessing [that he came.] / *What was a blessing that he came? > expletive it
cannot be questioned because it is not thematic (it does not bear a theta role)
2. Object it
- when it is an object, it is not an expletive, but an ordinary pronoun which receives a theta-
role: a Theme of Event.
(34) a. I regretted it that he got fired. (theme – specific event)
b. I resent it every time you say that. (event) – for every event of you saying that, I regret
that event
4
This is related to the information structure of sentences: The Topic > Focus articulation (37):
constituents which represent old information (known to the speaker) appear to the left
(topicalized), whereas constituents which represent new information appear to the right (in the
focus position)
(37) TOPIC > > FOCUS
old info new info
This dictionary, I will not buy. (the dictionary has been mentioned before)
Oh, John, everybody likes him (both speaker & hearer know John)
Everybody likes….John (John is the new info, not known to the hearer)
Compared to DPs (e.g., his illness), CPs (e.g., that he is ill) are longer, hence heavier. As a result,
they are subject to the principles of End-Focus & End-Weight >> being heavy, they have to
move to the right periphery of the sentence:
TOPIC > > > FOCUS
(38) a. His illness worries me. (the DP is light > not constrained by
End-Focus & End-Weight)
b. [That he is extremely ill] worries me. (the CP is heavier > has to abide by the
principles of End-Focus &End-Weight,
especially since CPs usually represent new
information >>
►Extraposition from object position is not motivated by the same considerations, since object
clauses already satisfy the principles of END FOCUS and END WEIGHT: a DO complement
clause is ALREADY to the right, so it already occupies the Focus position >> DO/PO
extraposition is functionally superfluous and hence infrequent.
TOPIC > > > FOCUS
(39) a. I expected [that they would grow to like syntax]
b. I expected [it [that they would grow to like syntax]]
However, when DO/PO extraposition does occur (39b), the resulting structure has characteristic
semantic and pragmatic properties (i.e., a mark of factivity, information in the complement
clause was previously known to the speaker – see above)
C.4.2. Structural considerations relate to the Extended Projection Principle (EPP, Chomsky,
1981), which requires clauses to have subjects.
English is an SVO language: it requires an overt subject in preverbal position.
in the case of extraposed subject clauses (where the complement, being heavy, is moved to the
right according to End-Focus and End-Weight (40b)), the EPP makes sure that the sentence does
5
not remain subjectless >> it is inserted to satisfy EPP and extraposed subject clauses cannot
function without this expletive subject (40c).
(40) a. [That he is ill] worries me.
b. __ worries me [that he is ill] → the heavy complement clause moves to the right BUT the sentence is
incorrect without an overt subject (!! because English is SVO >
sentences in English require an overt subject (the EPP) >>
c. IT worries me [that he is ill]
While subjects must always be overt in English (41a), objects may be missing from a
sentence (41b) (even though they are recovered via the subcategorization frame of the verb):
(41) a. John came yesterday. / * __ came yesterday
b. John ate lunch an hour ago. / John ate ___ an hour ago.
! Similar constraints apply in case the subject or the object is a CP (a complement clause) instead
of a DP: while the subject must always be overt, the direct object need not be overt:
(42) a. [That John came] is wonderful. / * ___ is wonderful [that …] / IT is wonderful [that….]
b. I know [that John came]. / I know ___ / I know it. / I know (it) that John came
↓
need not be overt for syntactic reasons
.
► Hence, object it is not obligatory in extraposed object clauses. If/when it does appear, the
sentence acquires a different interpretation (see above on the contribution of it + CP (i.e.,
factivity, old info known to speaker)
C.5. Topicalization
The reverse of extraposition is the so-called TOPICALIZATION structure.
If Extraposition = movement of the complement clause to the right, Topicalization = movement
of a complement clause (from its base position) to the left periphery of the clause (i.e., in the
complementizer domain, in a TopP)
6
Remember! A subject that-complement in clause-initial position (43a) is not actually in subject
position but in topic position (see section C.2. above & examples (21) & (22))
Syntactically, in the case of subject clauses as in (43) above, the CP, which is base-generated
in the specifier position of the VP has two options: a) either it moves leftwards (passing
through SpecIP (a case position) and moving forward above it (in a TopP) (= Topicalization) or
b) it moves rightwards and adjoins to the periphery of VP (and ‘it’ is inserted in SpecIP
because of EPP) (= Extraposition, see (28) above)
Topicalization of Subject clauses is the most frequent case, but there is also topicalization of D.O
& P.O clauses. When this occurs, it is usually in the literary register (and also signalled by
specific intonation and punctuation (see the comma in (44a)):
(44) a. [That Mary would betray him], John would never have imagined tCP (topicalization of
D.O. complement clause)
b. [That Mary would betray him], John had never worried (about) tCP (topicalization of
P.O. complement clause)
C.6. HEAVY NP SHIFT and CLAUSE SHIFT
The phenomenon we are discussing represents the occurrence of an object (DP or CP) at the right
periphery of the sentence, in a position different from its initial theta-position.
C.6.1. Heavy NP Shift = nouns (NPs) which are long (hence heavy) need to appear to the end of
the sentence (in the Focus position)
(45) a. He threw [the letter]light NP into the basket
b. *He threw [the letter which he had just decoded]heavy NP into the basket. > the heavy NP has to
shift (= be moved) to the right periphery to avoid ungrammaticality/ambiguity
c. He threw theavyNP into the basket [the letter which he had just decoded] = HNPS
7
Not only DPs, but also clauses, which are by definition complex, may appear at the right
periphery = CLAUSE SHIFT
Clause Shift is very similar to HNPS; they differ in that the former involves complement clauses
while the latter involves heavy/complex NPs (i.e., long, containing modifiers)
(47) a. (?) Mary saidV [that she wouldn’t deny it]d.o.CP flatly (awkward; ambiguous: the adverb
can, in principle, relate to both verbs (i.e., say flatly or (?) deny flatly)
b. Mary saidV tCP flatly [that she wouldn’t deny it.]→ Clause Shift
Clause shift, therefore, involves the movement of a complement clause to the right periphery
of the sentence. BUT: so does Extraposition!!
Q: Is Clause Shift the same as Extraposition?
A: Clause Shift differs from Extraposition in 2 relevant respects:
a) it applies to objects only (never to subjects);
b) no (expletive) pronoun marks the initial position of the clause.