diphthong. Significantly, however, cliticisation is not possible in sentences such as (344a) below, as we see from the fact that the sequence say have cannot contract to say’ve in (344b) (and so isn’t pronounced in the same way as save): (344) a.
Which students would you say have got most out of the course?
b.
*Which students would you say’ve got most out of the course?
Movement
299
What prevents have from cliticising onto say here? Let’s assume that prior to being moved to the front of the sentence by operator movement, the operator phrase which students is the subject of have, as in the echo question counterpart to (344a)
in (345):
(345)
You would say which students have got most out of the course?
If we also assume that when the phrase which students is fronted, it leaves behind a silent trace copy of itself (= t) in the position out of which it moves, then the structure of (344a) will be (346):
(346)
Which students would you say t have got most out of the course?
This being so, we can account for why have cannot cliticise onto say: it is not immediately adjacent to it, the two words being separated by the intervening trace – hence the ungrammaticality of (344b).
An interesting extension to our analysis of the syntax of operators is suggested by complement clause questions such as that bracketed in (347): (347)
I’m not sure [which senators the president has spoken to]
The bracketed interrogative (i.e. question-asking) clause in (347) is a complement clause since it serves as the complement of sure. In (347), the wh-operator expression which senators clearly originates as the complement of the preposition to (as we see from echo questions such as The president has spoken to which
senators?). But where does it move to? So far we have assumed that wh-operator expressions move into the specifier position within CP, to the left of C. But how can this be the case in (347), since the bracketed complement clause contains no overt C
constituent? A natural answer to give to this question within a theory which posits that specific positions in a structure can be occupied by empty categories is to suppose that the head C position in the bracketed CP in (347) is filled by a covert complementiser φ, so that the bracketed clause in (347) is a CP derived as in (348):
(348)
CP
DP
C'
which senators
C
TP
ϕ
DP
T'
the president
T
VP
has
V
PP
spoken
P DP
to
t
300
senten ces
There are a number of reasons for suggesting that the bracketed complement
clause in (347) contains a covert complementiser. One is that this enables us to maintain a unitary characterisation of operator movement as involving the movement of an operator expression into a specifier position to the left of an (overt or covert) C constituent. Another is that such an analysis provides a straightforward account of why auxiliary inversion is not permitted in complement clause questions in (standard varieties of) English, as we see from the ungrammaticality of (349):
(349)
*I’m not sure [which senators has the president spoken to]
Recall that in relation to the ungrammaticality of speaker B’s utterance in (333) above, we suggested that the presence of an overt complementiser like if
blocks auxiliary inversion: it seems a natural extension of this idea to suppose that the presence of the covert complementiser φ also prevents an auxiliary
from moving from T to C. A third reason is that, as observed by Alison Henry for a variety of English spoken in Belfast, we can find complement clause
questions which contain an overt complementiser, as in (350) (where the %
sign indicates that this type of structure is found only in some varieties of English):
(350)
%I’m not sure [which senators that the president has spoken to]
Since it is clear that in structures such as (350) the operator expression which senators is positioned to the left of the italicised complementiser that, it is reasonable to suppose that in structures like (347) which senators is positioned to the left of a covert complementiser φ.
An interesting question to ask at this stage is why wh-operators should be
moved to the front of the relevant interrogative clause in wh-questions. We can put this question rather differently by asking ‘What is it that makes us interpret the bracketed clause in (350) as a question?’ The answer clearly isn’t the choice of complementiser heading the clause, since that isn’t interrogative (hence the that-clause in I didn’t know that he was cheating can’t be interpreted as a question in Belfast English). So, it would seem that it is the presence of the interrogative phrase which senators in the specifier position of CP which ensures that the bracketed clause is interpreted as interrogative. Generalising, we can hypothesise that a clause is interpreted as a question in English if it has an interrogative specifier. We can then say that the wh-operator expression which senators in
(348) moves into spec-CP in order to ensure that the clause containing it has an interrogative specifier and so is interpreted as a question.
But why should it be that in questions containing more than one wh-operator, such as (351) below, only one wh-operator can be preposed, not more than one?
(351) a.
Who do you think will say what?
b.
*What who do you think will say?
The sentence in (351a) is derived as in (352):
Movement
301
(352)
CP
D
C'
Who
C
TP
do
D
T'
you
T
VP
t
V
TP
think
D
T'
t
T VP
will
V
D
say
what
Following assumptions made in recent work in syntax, let us suppose that our theory of grammar incorporates an Economy Principle along the lines of (353): (353)
Economy Principle
Minimise grammatical structure and movement operations (i.e. posit as little structure as possible, and move as few constituents as possible the shortest distance possible)
Obviously, (353) is consistent with general scientific guidelines which require us always to seek the simplest and most elegant theory which is consistent with the data we need to explain. Now, if a clause is to be interpreted as a question, it requires an interrogative specifier in spec-CP. It does not require more than one such interrogative specifier, and it follows from (353) that we therefore need to prepose only one of the two interrogative operators (who or what) in (352) in order to satisfy the requirement for CP to have an interrogative specifier: preposing both would be superfluous (in that it would involve two applications of wh-operator movement rather than one) and hence is ruled out by the Economy Principle.
Furthermore, (353) requires that it is the nearest wh-operator expression which moves to spec-CP in a multiple wh-question (because 353 favours shorter movements over longer ones). Thus, we can account for why it is who and not what that moves to spec-CP in (352). It is clear from the schematic structures in (354) that what must move further than who to get to the spec-CP position:
(354)
a.
Who do you think t will say what?
b. *What do you think who will say t?
302
senten ces
Yes–no questions
The assumption that questions are CPs which contain an interroga-
tive specifier runs into apparent problems in relation to yes–no questions such as (355):
(355)
Are you having any problems?
Even though (355) is obviously a question, it doesn’t seem to contain an interrogative specifier of any kind. So, it would appear that our existing analysis wrongly predicts that sentences such as (355) can’t be interpreted as questions.
Читать дальше