Andrew Radford - Linguistics An Introduction [Second Edition]

(307) below:

(307)

We didn’t know [φ he had resigned] or [that he had been accused of

corruption]

We can then say that (306) is grammatical because it involves the co-ordination of two declarative CPs (more precisely, two CPs headed by a declarative

complementiser).

The null C analysis can be extended from finite embedded clauses to main

(= root = principal = independent) clauses like (308) below: (308)

I am feeling thirsty

This sentence is declarative in force (by virtue of being a statement). If the force of a clause is marked by a complementiser introducing it, this suggests that such declarative main clauses are CPs headed by a null complementiser marking declarative force. And indeed, theoretical considerations require us to assume this, if Luigi Rizzi’s suggestion that the set of UG principles wired into the Language Faculty include a Uniformity Principle is correct. This principle

requires that all expressions of the same type belong to a uniform category (so that all clauses with the same force belong to the same category): since a

declarative that-clause like that bracketed in (304b) is clearly a CP, it follows from the Uniformity Principle that all other declarative clauses (including

declarative main clauses) must be CPs. This leads to the conclusion that a

declarative main clause like that in (308) is a CP headed by a null declarative complementiser. This means that the relevant clause has the structure shown

below:

(309)

CP

C TP

ϕ

PRN

T'

I

T VP

am

V A

feeling thirsty

Empty categories

281

Under the CP analysis of main clauses in (309), the declarative force of the overall sentence is attributed to the fact that the sentence is a CP headed by a null declarative complementiser φ.

It might at first sight seem strange to posit that declarative sentences in English contain an ‘empty’ or ‘silent’ CP layer of structure with no overt head or specifier.

However, this CP layer is not always empty – as can be illustrated in relation to the following sentence:

(310)

This kind of behaviour, nobody can tolerate it

As we saw in section 18, the italicised expression in this type of sentence represents the topic of the sentence, and the clause following the comma is the comment clause. While the comment clause is a simple TP (comprising the pronoun nobody, the present tense T-auxiliary can and the verb phrase tolerate it), the topic this kind of behaviour appears to be positioned somewhere outside the comment clause. But where? If we suppose that clauses are CPs, we can say that the topic occupies the specifier position within CP, and hence that (310) has the structure

(311) below:

(311)

CP

DP C'

This kind of

behaviour

C TP

ϕ

PRN

T'

nobody

T VP

can

V PRN

tolerate it

It follows that the CP ‘layer’ of clause structure in declaratives is not always empty.

The more general conclusion to which our discussion thus far leads us is that all finite clauses (whether main clauses or complement clauses) are CPs headed by an (overt or null) complementiser which marks the force of the clause. But what about non-finite clauses? It seems clear that for–to infinitive clauses such as that bracketed in (312a) are CPs, since they are introduced by the infinitival complementiser for – but what about the type of (bracketed) infinitive complement clause found after verbs like want in sentences such as (312b)?

(312) a.

I will arrange [for them to see a specialist]

b.

She wanted [him to apologise]

At first sight, it might seem as if the bracketed complement clause in sentences like

(312b) can’t be a CP, since it isn’t introduced by the infinitival complementiser for.

However, it is interesting to note that the complement of want is indeed introduced

282

senten ces

by for when the infinitive complement is separated from the verb want in some way – e.g. when there is an intervening adverbial expression like more than

anything as in (313a) below, or when the complement of want is in focus position in a pseudo-cleft sentence as in (313b):

(313) a.

She wanted more than anything for him to apologise

b.

What she really wanted was for him to apologise

(Pseudo-cleft sentences are sentences such as ‘What John bought was a car’,

where the italicised expression is said to be focused and to occupy focus position within the sentence.) This makes it plausible to suggest that the complement of want in structures like (312b) is a CP headed by a null counterpart of the complementiser for (below symbolised as φ), so that the complement clause in

(312b) has the structure shown in simplified form in (314) below: (314)

She wanted [

φ

CP [C ] [TP him [T to] apologise]]

In standard varieties of English, the null variant of the complementiser for found in (314) is generally used only when the complementiser immediately follows the verb want (with for being used where the verb and complementiser are separated by intervening material). The more general conclusion which our discussion here leads us to is that infinitive clauses are CPs, headed either by the overt infinitival complementiser for or by a null infinitival complementiser.

Our conclusion that infinitive clauses are CPs has important implications for the syntax of control infinitive clauses with a null PRO subject like that bracketed in

(315) below:

(315)

I will arrange [PRO to see a specialist]

It means that control clauses like that bracketed above must be CPs headed by a null infinitival complementiser. This would mean that there is parallelism between the structure of a for infinitive clause like that bracketed in (312a) above, and that of a control infinitive clause like that bracketed in (315), in that they are both CPs and have a parallel internal structure, as shown in (316a, b) below (simplified by not showing the internal structure of the verb phrase see a specialist):

(316)

CP

TP

T'

C PRN T

VP

a.

for

them

to

see a specialist

b.

ϕ

PRO

to

see a specialist

The two types of clause thus have essentially the same CP+TP+VP structure

and differ only in that a for infinitive clause like (316a) contains an overt for complementiser and has an overt accusative subject like them, whereas a

Empty categories

283

control infinitive clause like (316b) has a null φ complementiser and a null PRO

subject.

Some evidence in support of claiming that a control clause with a null PRO

subject is introduced by a null complementiser comes from co-ordination facts in relation to sentences such as the following:

(317)

I will arrange [to see a specialist] and [for my wife to see one at the same time]

The fact that the italicised control infinitive can be conjoined with the bold-face CP headed by for suggests that control infinitives must be CPs (given the traditional assumption that only the same types of constituent can be conjoined).

Overall, the conclusion which our analysis leads us to is that all ordinary clauses (whether finite or infinitival) are CPs headed by an overt or null complementiser (C), with C serving to mark the force of a sentence in finite clauses, and serving to mark the clause as non-finite in infinitives.

Empty determiners

The kind of reasoning we have used here to argue that all clauses are

CPs can be extended to the analysis of nominal (i.e. noun-containing) expressions.

In this connection, consider the italicised nominals in the two replies produced by speaker B in the dialogue below: