Gramley, Stephan; P?tzold, Kurt-Michael. A Survey of Modern English

Подождите немного. Документ загружается.

It can also refer to the state or condition of being injured or insulted. The sentence John

got hurt, in contrast, has only the dynamic meaning.

Despite the advantage of its clearly dynamic meaning, get also has a disadvantage: it

may only be used with certain verbs (get married, get involved, get done; not *?get seen,

*?get aided, *?get instructed). Get is only suitable when a change of state is involved.

Furthermore, style is also a factor: in more formal usage get is less acceptable than in

casual conversation.

Other ‘auxiliaries’: the semi-passive In some cases get is less a passive than a

copula followed by a participial adjective. This can be conﬁrmed by the fact that an inten-

siﬁer such as very, awfully or extremely may precede the participle (Towards midnight

I got (terribly) tired). Much the same thing is true of become, grow, feel, seem and a

number of others. They resemble the passive, but are not usually regarded as true passives.

Passive-like structures Both get and have are used in a structure in which the

subject is the experiencer (but not the agent, see 5.3.2) of a passive act: John got/had his

arm broken in the ﬁght. Here John experienced something, but did not actually do it. In

a different interpretation of this structure, the subject may be understood as the person

who caused or instigated what was done (John got/had his tonsils removed, meaning

‘asked the surgeon to do this’).

Active constructions with passive meaning In conclusion it should be pointed

out that there are a number of constructions in which the subject of an active sentence is

the ‘logical’ object of the predicator. This gives these sentences a passive-like interpreta-

tion, and is possible with such verbs as blow, burn, ring, break, open etc. (The leaves

blew in the wind, ‘were blown by the wind’; The house burned down, ‘was burned down’).

A similar pattern applies to verbs which usually occur with an adverbial. Here the

agent can be left out and the object can then become the subject; hence We are selling

that book awfully fast becomes the passive-like That book sells awfully fast, ‘is being sold

fast’. This is not truly a passive, for note that the same relationship holds with intransi-

tive You can write well with that pen, which becomes That pen writes well. Here no

passive paraphrase is possible, for the subject has the semantic role of instrument. There

do not seem to be any obvious restrictions on what predicators permit this so long as

they: (a) (usually) contain an adverbial of manner and quantity; and (b) what is promoted

to subject is an objective or instrumental semantic role.

The communicative structure of sentences One ﬁnal, extremely important

question remains: what function does the passive fulﬁl? In English what comes early in

a sentence is its theme or topic and what follows is the rheme (from the Greek word

meaning ‘what is said’) or comment on this topic. Since subjects usually come at the

beginning of sentences, they also normally designate the theme. Sometimes, however,

the object is the theme. By using the passive the object can become the subject and

take the thematic position at the beginning of the sentence. For example, in John’s in the

hospital. Someone hurt him in a ﬁght, the theme of the second sentence is John, as estab-

lished by the ﬁrst sentence. The two sentences would have more cohesion if the expected

theme came ﬁrst in the second sentence and the new material, the comment, towards the

end. This can easily be accomplished with the passive: John’s in the hospital. He got

hurt in a ﬁght.

122 ENGLISH AS A LINGUISTIC SYSTEM

Other devices are also available to change the placement of elements. One of these

means of highlighting information is by fronting (Roses, I like; violets, I don’t or He

didn’t go gracefully, but quickly he did go). Notice that fronting is used especially for

contrast. Left-dislocation is a variation on this in which the usual, or canonical, position

of the element which is moved to the front/left is marked by a pronoun, e.g. Roses, I like

them; John, he didn’t go gracefully.

Another way of contrasting elements is by means of contrastive stress, which can be

indicated only imperfectly in writing by using italics or underlining for what, in speech,

would be realized by loudness, pitch change or similar (see 4.5.2).

In formal writing especially, but not only there, use is made of cleft and of pseudo-

cleft sentences. In the ﬁrst type the rheme is introduced by It is/was and the theme follows

in a relative clause (It was my car keys which I lost). This reversal of the usual theme-

rheme order has, once again, a contrastive function. Any sentence element or part of a

sentence element except the VP may be made the rheme in this way:

it is the second question which concerns us [subject];

it is statements of this sort which the author deals with [object];

it was in a ﬂash that they left [adjunct].

In the pseudo-cleft construction, even a VP may be highlighted. Here the element which

is to be emphasized is preceded by the appropriate sort of relative clause and the verb

be. The element highlighted may be

a verb: What we did was (to) leave as fast as possible;

a noun: The person who got hurt was John;

a manner adverbial: The way (in which) he did it was by dishonesty;

a place adverbial: (The place) Where we met was (in) Ohio;

a time adverbial: The day (when) we left was Tuesday.

A further means of changing the relative order of elements in a sentence is extraposi-

tion (see 5.8.7).

5.4.5 Modality

Modality in English has to do with the world, not so much the way it is as the way it

might potentially be. This may revolve around people’s beliefs about it or around their

potential actions in it. There are various linguistic means of expressing this in English,

e.g. with AdvPs and PPs (Probably he’s coming; It’s in their power to decide the issue),

with AdjPs (It’s likely that he’s coming; They’re able to decide the issue), with NPs (the

probability of his coming; their capability to decide the issue) and with VPs (He might

come; They can decide the issue). Here we will concentrate on the modal verbs.

Morphology and syntax The modal verbs form a closed class whose central

members are will, would, may, might, can, could, shall, should and must. Somewhat less

central are need, dare, ought to, used to and had better. All the modals share the

NICE

syntactic features of the auxiliaries/operators (see 5.1.3). The central modals retain these

1111

1011

4111

20111

5111

30111

40111

47111

GRAMMAR 123

features under all circumstances, however the less central ones do not always exhibit

them. Dare, need and used to may appear with do-periphrasis. For example, I needn’t

ask is virtually the same as I don’t need to ask. In addition, had better seems to have no

widely acceptable interrogative form (

Had I better buy a new one? sounds awkward).

In addition, none of the modals has the third person singular {-S} (she may, never *she

mays), and all take the bare or unmarked inﬁnitive except ought (to) and used (to).

For these reasons needs/dares to are often not regarded as modals; no further notice

will be taken here of any of the more peripheral modals except for needn’t as a negation

of must.

The defective nature of the modal verbs vis-à-vis lexical verbs can be seen in the fact

that none of them (as modals) have either a present or a past participle. As a result they

cannot appear in the perfect, progressive or passive forms. Furthermore, while four of the

modals do form present-past pairs (will-would, shall-should, may-might, can-could), the

past tense forms function as markers of past time only in the most restricted of circum-

stances (see below ‘Tense’).

Proposition and event orientation There are two major types of modality in

English. The one, propositional, has to do with beliefs and knowledge about logically

possible or logically necessary situations, and it commonly referred to as epistemic. The

other, event-oriented, has to do with potential actions and is frequently called deontic.

Epistemic modality indicates the degree of probability of a fact or proposition. Take

a proposition such as Sally is 170 cm tall. This can be viewed as relatively unlikely or

convincingly plausible or, of course, somewhere between the two. Expressed with modals

this produces Sally might be 170 cm tall at the unlikely end and Sally must be 170 cm

tall at the certainty end. Each of these could be paraphrased as follows: It is just possible

that Sally is 170 cm tall and It is necessarily true that Sally is 170 cm tall. Since epistemic

modality is also often based on evidence, it is sometimes termed evidential.

The epistemic modals can be arranged on a scale from low to high probability: might-

may-could-can-should-ought to-will-must. This scale is a somewhat subjective estimate

and is not always accurate. The modals are too vague to allow a more precise calibra-

tion. Will, for example, makes a prediction, sometimes on the basis of evidence (The

phone’s ringing; that’ll be my sister) and sometimes ‘out of the blue’ (It’ll be nice at the

party, I hope).

Deontic modality has to do with ability, permission, volition and obligation with regard

to an action. If you can or may do something, i.e. are allowed to do it, the possibility

exists that you will carry out this action. If you must do it, there is an obligation or a

necessity, rather than a possibility. The key terms are permission and obligation. Ability

and volition are variants of permission (they are sometimes referred to as a further type

of modality, dynamic). While permission comes from outside the agent, ability and

volition are internal to the agent (see below ‘Subjective and objective’).

Time and tense Deontic and epistemic modality have different structures for

expressing the past. For epistemic modality reference to the past is realized by means

of the perfect inﬁnitive, e.g.:

it might be true now : it might have been true back then

it may be true now : it may have been true back then

124 ENGLISH AS A LINGUISTIC SYSTEM

it could be true now : it could have been true back then

it must be true now : it must have been true back then

Note that all of the epistemic modals, even the past tense forms, have a present (some-

times future) meaning when used with the simple inﬁnitive; furthermore, all of them

may be used with the perfect inﬁnitive to express conjecture about the past. In addition,

the past form of the modal in connection with the perfect inﬁnitive has a contrafactual

or condition-contrary-to-fact effect, which may make them ambiguous, e.g. they could/

might have done it, (but didn’t); or, but I don’t know if they did (see below ‘Conditional

Sentences’).

In reported speech there is no backshift of those modals which are already past in form

or which have no past form, i.e. might remains might and must remains must; epistemic

will, can and may can be shifted to would, could and might, though this need not be the

case.

In their present tense forms the deontic modals have future or potential reference. For

past time reference there are substitute forms available; they include be permitted/allowed

to for may and can, be able to for can, be willing to or be going to for will, be supposed

to for should and have (got) to for must (also possible for epistemic must).

The two (dynamic modality) past tense forms, could and would, are the only ones

which fulﬁl all of the remoteness functions of the past tense: past time reference, distance

in reported speech, unreality in conditional sentences, and politeness and tentativeness in

statements (see 5.4.2). However, past time reference is not completely unrestricted.

Note that both auxiliaries have past time reference chieﬂy in non-assertive contexts such

as negation and questions and only in ability and volitional meanings. That is, it is the

ability meaning which is carried by the past negative couldn’t. For will, it is the volition

(rather than prediction) meaning which wouldn’t conveys. The unnegated past of can

ability is usually not expressed by could, but by be able to and the volition sense of will

not by would, but by want to or be willing to (negative He couldn’t solve the problem

as compared to positive He was able to ...; and negative She wouldn’t sell her stock as

opposed to positive She wanted to ...). The past permission meaning of can is never

expressed by a modal as in He was permitted/allowed to stay up late, while He could

stay up late has an ability meaning. Likewise the future-in-the-past of will must be the

non-modal She was (not) going to sell and not the volitional She would(n’t) sell.

Should, though past tense in form, never has past time reference. Past deontic time

reference is possible either with the perfect inﬁnitive, which has a contrafactual effect

(should have done = ‘was/were supposed to, but did not’), or with the substitute forms

was/were supposed to do/have done, which is not restricted to a contrafactual interpreta-

tion. The same applies to ought to. Should is not used as the tentative or unreal form of

shall, either; however, it is employed as a backshifted form in reported speech, e.g. A:

Shall I leave? B: She asked if she should (*shall) leave. Much the same sort of thing

applies to might, which is almost never found for past time reference and seldom as the

unreal or tentative form of may. It can, however, be used for backshifted may in reported

speech, e.g. A: You may have some cake if you wish. B: She said we might (more likely:

may) have some cake if we wished.

Negation Besides the effects of negation just mentioned for the past tense of can

and will, there are some important complications that have to do with the question of

1111

1011

4111

20111

5111

30111

40111

47111

GRAMMAR 125

whether not negates the modal or the following inﬁnitive. Sometimes this is unimportant

inasmuch as, for example, not having permission to smoke (You can’t smoke in here) is

what would usually be said. Having permission not to smoke (You can not smoke in here),

which sounds strange, would very likely be interpreted in the same way as the former

utterance. Much the same applies to the should and ought of obligation. The most com-

plicated case of negation is with deontic must. The negative of the modal is needn’t (You

needn’t go, ‘it is not necessary that you go’). The negative of the lexical verb, in contrast,

is mustn’t (You mustn’t go, ‘it is necessary that you not go’).

Subjective and objective As mentioned above, the sentences We said they were

able to help out and We said they could help out do not mean the same thing. In the ﬁrst

case the ability to help out is introduced as an objective possibility. In the second case

the subject of the main clause (we) attributes this ability to the subject of the reported

clause they; it is an assessment and as such subjective.

There are several such pairs, for example: subjective must and objective have to;

subjective needn’t and objective not have to/not need to; subjective can and objective

be allowed to. The distinction is, however, not always clearly maintained, and this can

lead to ambiguity, as when the Australian Prime Minister said of a politician who was

accused of deceiving the Parliament: Senator Withers may have misled Parliament.

As a case of subjective modality the Prime Minister would be committing himself to

this possibility; as a case of objective modality [the actual intention] the Prime Minister

would merely be admitting that others had made this accusation (example from

Huddleston 1984: 167).

The deontic modal must is subjective, which means that it imposes an obligation stem-

ming from the speaker; hence it is relatively forceful. Objective have to invokes an outside

obligation, which takes the onus off the speaker. This may well be the reason why some

dialects of English (for example, some varieties of Scottish English) do not have must,

but only have to. Certainly, the difference between the impolite sounding You must leave

immediately and the more neutral You have to leave immediately is evidence for this.

If, however, something pleasant is expressed as an obligation, must is unproblematic

(You must try our new sauna).

Conditional sentences Four types of conditional sentences may be distinguished

in English, and the modals play an important role in all of them. The types are given in

Table 5.6. Modals do not normally appear in the if-clause. However, will may when it

expresses volition. Generally, the modal auxiliaries are restricted to main clauses. Real

conditional clauses refer to the future; since the modals are commonly used for future

reference, they appear frequently in such sentences. Will is the most common, but other

central modals occur in real conditional clauses as well. Unreal conditional sentences

differ from real ones only inasmuch as they express less likelihood. The use of the past

tense therefore ﬁts very well as a marker of remoteness.

Contrafactual conditionals, the third type, are used to say something about the past.

Here a conclusion is drawn about a hypothetical state of affairs. The main clause always

contains a modal (chieﬂy would, often could, sometimes might) plus a perfect inﬁnitive.

This combines the remoteness of the past with epistemic modality, which expresses

the likelihood of something given the right conditions in the past. Since the past is not

repeatable and cannot be changed, this is ‘contrary to fact’ or contrafactual.

126 ENGLISH AS A LINGUISTIC SYSTEM

The ﬁnal type, the implicational conditional, states a relationship which the speaker

logically supposes to be true. It is basically a variant of the real condition. The if-clause

represents not a possibility but a circumstance whose truth is not deﬁnitively known.

Hence if it’s raining implies that the speaker does not know whether or not it is raining.

However, if the condition is true, then by logical implication it is also true that they must

be getting wet. Note that the predicator in the main clause will either be the epistemic

must of logical necessity or it will be the straightforward indicative (i.e. they are getting

wet).

Conjunctions Conditional clauses have, for convenience and clarity, been referred to

as if-clauses. While it is true that if is the most common conjunction employed, others

are also to be found, for example: supposing, suppose, in case, allowing that, in the event

that, on (the) condition that (all stylistically formal); unless, negative (= ‘if not’, e.g. This

can’t be true unless I’m dreaming!); lest (‘in order that not’; extremely formal, chieﬂy

AmE, e.g. Plan in advance lest ill fortune bring you to a fall, ‘if you do not want ill

fortune to . . .’). In addition, conditional clauses may be introduced by subject-operator

inversion, (e.g. Were he to agree/Should he agree), I would be very astonished; also

1111

1011

4111

20111

5111

30111

40111

47111

GRAMMAR 127

Table 5.6 Conditional clauses

Type of conditional if-clause main clause

I real condition present will (or other modal)

Ex: If you study hard, you will/can get a good grade. [= future]

IIa unreal condition past would (or other modal)

Ex: If I studied hard, I would/might get a good grade.

IIb stronger unreal condition were + inﬁnitive would (or other modal)

Ex: If I were to study hard, I would . . ..

[were is a subjunctive form indicating unreality]

III contrafactual condition past perfect would (or other modal)

(contrary-to-fact; irrealis) + perfect inﬁnitive

Ex: If I had studied hard, I would/could have had a good grade.

IVa implicational condition

(no tense restrictions) present present or must + inﬁnitive

past past or must + perfect inf.

Ex: If that is the evening star, it is (must be) Venus. [modal of logical necessity]

If you studied hard, you got (must have gotten)

good grades.

IVb implicational condition present present

(iterative) past past

Ex: If (= Whenever) I study hard, I get a good grade.

Also possible with habitual would: If (= Whenever) I studied hard, I got/would get

good grades.

Had he agreed I would have been . . . . Finally, conjoined clauses (chieﬂy second person

and frequently imperative) may have a conditional effect, e.g. [threat]: (You) come here

again and you’ll get to know me.

The semi-modals This label is used to refer to a number of (non-defective) verbs

which have meanings parallel to those of the modal auxiliaries. They include, above all,

the substitute forms mentioned above: have (got) to, be willing to, want to, be allowed/

permitted to, be supposed to, be able to and be going to. Evidence for the auxiliary nature

of many of these may be seen in the irregularity caused by the assimilation of the inﬁni-

tive marker to with the preceding element to form a single phonetic unit, i.e. have to is

really hafta /h

ftə/ (has to is hasta and had to is hadda), have got to is gotta, supposed

to is supposta, going to is often gonna, want to may be wanna; even the less central

modals with to have usta for used to, a possible needa for need to and oughta for ought

to. This kind of reduction does not apply to other verbs followed by an inﬁnitive; there

is no *beginna for begin to or *lofta for love to.

5.5 THE NOUN PHRASE (NP)

The NP can consist of up to four parts: the obligatory noun head, pre-head determiners

with several different positional possibilities, pre-modiﬁers and, ﬁnally, post-modiﬁers,

also consisting of several possible elements.

5.5.1 The noun head

Nouns can be divided into common nouns, themselves subdivided into count and non-

count, and proper nouns.

Count and non-count nouns Count nouns are prototypical nouns because they

have plurals and, in many cases, possessives (see 5.1.1). They also take all forms of the

determiner (the, a/an and zero). Non-count nouns, in contrast, have no plural and an indi-

vidualized singular. As a result they cannot appear with the indeﬁnite article (e.g. there

is no *a snow). They may, instead, have either the deﬁnite article with speciﬁc reference

or zero article with generic reference (see below). Non-count nouns may be concrete and

include such mass nouns as coffee, sand, wheat or mud. More often, however, they will

be abstract nouns such as loudness, strength or entrepreneurship. Furthermore, a large

group of nouns includes both concrete and abstract nouns which are sometimes count and

sometimes non-count, e.g. cabbage, denial or sound, as in I bought a cabbage (= a head

of cabbage), where cabbage is a count noun, and I don’t like cooked cabbage, where it

is a non-count mass noun. In addition, mass nouns may be used as count nouns, when,

for example, you order two coffees or a writer talks about the snows of yesteryear.

Normally, however, this is accomplished by preﬁxing some measure or quality expres-

sion as in two cups of coffee, a kind of snow, or a grain of sand. Article usage with count

and non-count nouns is summarized in Table 5.7.

Nounds used in a generic sense refer to typical representatives of a class. With count

nouns there are three common ways in which generic reference is realized, namely, the

128 ENGLISH AS A LINGUISTIC SYSTEM

singular with the deﬁnite article (The unicorn is a mythological beast), the singular with

the indeﬁnite article (A unicorn has a single horn) and the plural with no article (Unicorns

do not exist). The three differ from each other in that the + singular refers to the class

represented; it has an informal alternative with your (Your unicorn is a mythological

beast). A + singular is used much like any. It may only occur in subject position. Zero

article and the plural, also restricted to subject position, is the same as saying all (or

no, if negative). A variant of this is the deﬁnite article and plural nouns of nationality

(Germans are orderly = The Germans are orderly = All Germans are orderly). Non-count

nouns are generic when they occur without an article (Snow is white).

Proper nouns can be identiﬁed as those which are capitalized. Capitalization is a sign

that each such noun refers to a unique entity (even if, in fact, there is more than one, say,

Joe or Barbara in the world). Because the people, places or things referred to are unique,

they cannot be further speciﬁed and therefore occur without an article. There are two

important exceptions to this. The ﬁrst and most general is that a name of a person can

occur more than once. If we want to distinguish between two people with the same name

we can specify further by using the deﬁnite article and a post-modifying phrase and saying,

for instance, the Joe I know or the Barbara with the red hair. In this case plurals are

also possible, e.g. the Anna’s of this world. The second exception involves places. Many

place names which are derived from common nouns retain their article, so the United

States, the Metropolitan (opera), the Mississippi (river) etc. Unfortunately, there is no

easy rule to distinguish between, for example, Buckingham Palace (no article) and the

White House (with the deﬁnite article).

5.5.2 Pre-head

The elements of the pre-head are divided ﬁrst into those which specify (tell which of a

group) or quantify (tell how many or how much). These are the pre-determiners, central

determiners and post-determiners, the last of these divided into two subgroups.

Following the determiners come the adjectives and participles, which are all designations

of quality.

Determiners Many of the determiners are uncountable (much, less (snow) – the latter

increasingly plural as well, e.g. less people), singular (a/an, every, each, this, that), dual

(= reference to two: both, either, neither) or plural (all, many, fewer, these, those).

The most common central determiner is the article as in all the fresh eggs or what

a nice car. There are, however, various other sets of central determiners, for example:

1111

1011

4111

20111

5111

30111

40111

47111

GRAMMAR 129

Table 5.7 Articles with count and non-count nouns

deﬁnite count non-count indeﬁnite non-count

singular the (song) the (music) a (song) music

plural the (songs) – (songs) –

demonstrative determiners: this, that, these, those

possessive determiners: my, your, her; Ruth’s, (the) boy’s, whose

indeﬁnite determiners: some, any, no, every, each, (n)either

relative determiners: which(ever), what(ever)

Both the demonstrative and the possessive determiners may be preceded by the pre-

determiners, which are quantiﬁers of scope (sometimes with, sometimes without of ), for

example all, half, a/one third, both, double, twice etc. (e.g. both (of) the students). There

is, in addition, the quantiﬁer many and a small set of qualifying pre-determiners (such,

quite, what) which may precede the indeﬁnite article (such a day). With so, how, too and

enough adjectives are moved in front of the indeﬁnite article along with the determiner

(so/how/too big a problem/big enough a problem). The post-determiners are exclusively

expressions of quantity. They consist of relativizing items including the ordinal numbers,

e.g. ﬁrst, second, third, but also next, last, further, additional, other etc., which precede

the cardinal numbers and a few other quantiﬁers of number (one, two, three, few, several).

No more than one representative of each of the four positional classes may appear in any

one NP, e.g. half (of) the next three (weeks) or all (of) her second three (attempts).

5.5.3 The order of adjectives

This is not strictly ﬁxed, but the dominant principle is that the more accidental, subjec-

tive and temporary qualities are named before the more essential, objective and permanent

ones. This means that evaluative adjectives (beautiful, important, stupid) tend to come

ﬁrst, and those which name the substance out of which something is made (wooden, metal)

or the subject matter something refers to (economic, religious) come last. In between

come size (tiny, tall, fat), then shape (round, ﬂat, sharp), then participles (blazing, ruined)

followed ﬁrst by age (old, new, young) and then colour (red, green, blue). After that

comes nationality or provenance (British, American, African). Adjectives which are grade-

able may be preceded by adverbial intensiﬁers (somewhat, astonishingly, pretty, very).

Here are some illustrative examples (more than three or four adjectives in one NP would

be rare in actual use):

both my last two very worthless old British pennies;

all your shapeless old-fashioned felt carpet slippers;

the second dozen small somewhat wilted yellow roses;

his confusing modern poetic works.

5.5.4 Post-modiﬁers

The noun head can be followed by several types of modifying expression. A few adjec-

tives can be post-positive, be they ﬁxed expressions (secretary general, president elect)

or adjectives and participles with complements (a woman true to her principles, a jacket

made to order). Those adverbs of time and place which can modify nouns also follow

(the valley beyond, that car there, years before). Many of these adverbs are variants of

PPs, which are the most frequent post-modiﬁers (the valley beyond ours, a story about

130 ENGLISH AS A LINGUISTIC SYSTEM

love and war, a person of distinction). PPs themselves are sometimes, but by no means

always variations of clauses, both non-ﬁnite (the valley situated beyond ours, the woman

sitting beside you, the valley to visit) and ﬁnite (the valley which is situated beyond ours,

a student who comes from Ghana).

The relative order of post-modiﬁers is generally from short to long. The main excep-

tion to this is that PPs which provide information about the nature or provenance of the

head noun (a woman of virtue, the currency of Japan) come before participles and adverbs

(a friend of the family waiting outside in the yard).

Post-modiﬁers can be restrictive (deﬁning) or non-restrictive (non-deﬁning), whereas

pre-modiﬁers are more likely to be restrictive only (on restrictive and non-restrictive, see

5.5.5). As restrictive elements they often introduce new and distinguishing or identifying

features. On further mention these restrictive, deﬁning elements, now known or given,

may become the initial member of a compound in much the fashion of the theme-rheme,

given-new distinction (see 5.4.4). For example, a program may be introduced as a

program using a computer; further mention can then be of the computer program. More

detail may be added as in a computer program for the checking of spelling, which can

then lead to the computer spelling-check program and so on. This is parallel to the way

information is introduced via a predicative complement: Her car is black which then

becomes her black car.

5.5.5 Relative clauses

The restrictive vs non-restrictive distinction just mentioned is also made by relative

clauses. Restrictive relative clauses are used if the identity of the referent is not clear.

If someone says, ‘my mother’, it is not possible to ask ‘Which of your mothers?’; the

identity is clear, so a post-modifying relative will be non-deﬁning (My mother, who loves

ﬂowers, is president of the garden club). However, if someone says ‘my brother’ or ‘my

sister’ and has more than one, the relative clause may be necessary to identify which one

of them is meant (My sister who is at university just turned 20, but my other sister is

only 16). However, the distinction between referents which are already identiﬁed and

ones which need further deﬁnition is often far from clear. One result is that many writers

fail to indicate the distinction, namely that non-restrictive clauses are conventionally sepa-

rated off by commas while restrictive ones are not.

Non-restrictive pronominal relatives consist of only the items who(m), which and

whose. Restrictive relative pronouns include these as well as that and zero relative [= 0

]

(e.g. the book that/[0

] I read). Non-restrictive relative clauses are always ﬁnite. Restric-

tive pronominal clauses, in contrast, may be ﬁnite (We don’t like people who complain

all the time) or non-ﬁnite (the hotel at which to stay/the hotel [0

] to stay at).

Sentential relative clauses In this type of construction, the antecedent is a whole

clause and only which can be used (Grammar is interesting, which is why I study it). It

is always non-deﬁning. This is, of course, not a post-modiﬁer in an NP, but a kind of

sentence adverbial.

A further important distinction between relatives is that there are both pronominal rela-

tive clauses, that is, ones which are introduced by a relative pronoun (who(m), whose,

which, that) and adverbial ones, introduced by a relative adverb (when, where, how, why).

1111

1011

4111

20111

5111

30111

40111

47111

GRAMMAR 131