Andrew Radford - Linguistics An Introduction [Second Edition]

predicts that when we combine a verb like reduce with a noun like taxes, the resulting phrase reduce taxes has verb-like properties (as opposed to noun-like properties or properties which are neither verb-like nor noun-like). This can be seen from the fact that the phrase reduce taxes can occupy the same range of positions as a verb like resign, and hence, for example, occur after the infinitive particle to:

(250) a.

The government ought to resign

b.

The government ought to reduce taxes

257

258

senten ces

By contrast, reduce taxes cannot occupy the kind of position occupied by a

plural noun such as taxes, as we see from (251):

(251) a.

Taxes are at the heart of the debate about policy

b.

*Reduce taxes are at the heart of the debate about policy

So, it seems clear that the grammatical properties of a phrase like reduce

taxes are determined by the verb reduce, and not by the noun taxes. We can say that the verb reduce is the head of the phrase reduce taxes, and conversely

that the phrase reduce taxes is a projection of the verb reduce (i.e. a larger expression whose head word is the verb reduce). Since the head of the resulting phrase is the verb reduce, the phrase reduce taxes is a verb phrase: and in the same way that we abbreviate category labels like verb to V, we can abbreviate the category label verb phrase to VP. If we use the labelled bracketing technique

(section 10) to represent the category of the overall phrase reduce taxes and of its component words reduce and taxes, we can represent the structure of the

resulting phrase as in (252):

(252)

[VP [V reduce] [N taxes]]

What (252) tells us is that the overall phrase reduce taxes is a verb phrase (VP), and that it comprises the verb (V) reduce and the noun (N) taxes. The verb reduce is the head of the overall phrase, and the noun taxes is the complement of the verb reduce. The operation by which the two words are combined to form a phrase is called merger.

Although we have used the labelled bracketing technique to represent the

structure of the verb phrase reduce taxes in (252), we have seen in section 10

that an alternative way of representing this sort of structure is in terms of a labelled tree diagram such as (253):

(253)

VP

V

N

reduce

taxes

The tree diagram in (253) is entirely equivalent to the labelled bracketing in

(252), in the sense that the two provide us with precisely the same information about the structure of the phrase reduce taxes: so (253) – like (252) – tells us that reduce is a verb (V), taxes is a noun (N) and reduce taxes is a verb phrase (VP).

The differences between a labelled bracketing like (252) and a tree diagram like

(253) are purely notational: each category is represented by a single node (i.e.

point) in a tree diagram, but by a pair of brackets in a labelled bracketing.

We can generalise our discussion at this point and hypothesise that all phrases are formed in essentially the same way as the phrase in (253), namely by merging two categories together to form a larger category. In the case of (253), the resulting phrase is formed by merging two words. However, not all phrases contain just two words, as we see if we look at the structure of B’s reply in (254):

Sentence structure

259

(254)

speaker a: What’s the government’s principal objective?

speaker b: To reduce taxes.

The italicised phrase in (254) appears to be formed by merging the infinitive particle to with the verb phrase reduce taxes. What’s the head of the resulting phrase to reduce taxes? There is evidence which indicates that this head is the infinitive particle to, so that the resulting string (i.e. continuous sequence of words) to reduce taxes is an infinitive phrase. The evidence is that strings such as to reduce taxes have a different distribution from verb phrases, as is indicated by sentences such as those in (255) and (256):

(255) a.

They ought [to reduce taxes]

b.

*They ought [reduce taxes]

(256) a.

They should [reduce taxes]

b.

*They should [to reduce taxes]

If we assume that reduce taxes is a verb phrase whereas to reduce taxes is an infinitive phrase, we can then account for the data in (255) and (256) by saying that ought is the kind of word which requires an infinitive phrase after it as its complement, whereas should is the kind of word which requires a verb phrase as its complement.

But what kind of word is infinitival to? It is traditionally termed an infinitive particle, and this terminology implies that it is a unique kind of word unrelated to any other kind of word in English. But is this so? In some respects, infinitival to seems to resemble an auxiliary like will, in that both are typically used in a clause with future time reference (as you can see from the fact that the bracketed

complement clauses in (257) both refer to a future event): (257) a.

Everyone is expecting [the government will reduce taxes before the election]

b.

Everyone is expecting [the government to reduce taxes before the election]

Moreover, the fact that the auxiliary wil and the infinitive particle to occupy the same position in the two clauses (between the subject the government and the verb phrase reduce taxes before the election) makes it plausible to suggest that the two are different exponents of the same category. The core function of an auxiliary is to mark tense – as we see from the fact that the present/past tense distinction in sentences such as He is/was lying is marked by use of the present tense auxiliary is or the past tense auxiliary was. Let us therefore assume that finite auxiliaries and infinitival to both belong to the category T of Tense-marker and differ only in that auxiliaries are finite (and so are overtly inflected for the past/non-past distinction), but infinitival to is non-finite (and so is not inflected for the past/non-past distinction).

After this brief digression about the status of infinitival to, let’s return to consider the structure of speaker B’s utterance to reduce taxes in (254). This is an infinitive phrase formed by merging the infinitival tense particle to with the verb phrase to reduce taxes. Using T as a convenient abbreviation for infinitival tense particle and TP as an abbreviation for infinitival tense phrase, we can say

260

senten ces

that the phrase to reduce taxes is a TP formed by merging the infinitival

tense particle (T) to with the verb phrase (VP) reduce taxes and so has the structure in (258):

(258)

TP

T

VP

to

V

N

reduce

taxes

The resulting TP is headed by the T to (indicating that the action of reducing taxes is intended to take place at some unspecified time in the future), and the VP reduce taxes is the complement of to.

What is implicit in our discussion up to this point is the idea that we can build up complex structures by successively merging pairs of categories to form ever larger phrases. For example, by merging the infinitive phrase to reduce taxes with the verb try, we can form the phrase produced by speaker B in (259): (259)

speaker a: What will the government do?

speaker b: Try to reduce taxes.

The resulting phrase try to reduce taxes is headed by the verb try, as we see from the fact that (like a typical verb phrase) it can be used after the infinitive particle to in sentences like those in (250) above (The government ought to try to reduce taxes). This being so, the italicised phrase in (259) is a VP which has the structure in (260):

(260)

VP

V

TP

try

T

VP

to

V

N

reduce

taxes

The head of the overall VP is the verb try, and its complement is the TP to reduce taxes. Now, (260) illustrates the important property of recursion, which we introduced in section 10, when discussing English compounds. Our analysis is claiming that try to reduce taxes is a VP which itself contains another VP, reduce taxes, and it is easy to see that further applications of merger will yield a larger VP–expect to try to reduce taxes including the VP in (260). We thus see that this simple operation of merger, as a core operation in the theory of grammar,