The notion of restriction in

English grammar
K. R. Narayanaswamy

Downloaded from https://academic.oup.com/eltj/article/36/1/48/430946 by guest on 22 February 2022

'Restriction' and 'non-restriction' in accounts of English grammar are
notions usually applied only to relative clauses, which, we are often
told, may be restrictive (defining) or non-restrictive (non-defining).
In this article a case is made out for seeing the dichotomy between
restriction and non-restriction as a more pervasive feature of English
grammar that affects many other grammatical structures, and even
lexical Hems.

The notion of restriction/non-restriction comes into sharp focus in the two

types of relative clause, the restrictive (or 'defining') and the non-restrictive
(or 'non-defining'), which pedagogic grammars usually have a great deal to
say about. A restrictive relative clause is one which restricts or limits the
noun or pronoun it modifies to a particular selected type:
1 The man who lives next door is a millionaire.
The man is a millionaire.
What man ?
The man who lives next door (and not the man who lives three doors
away, etc.).
The embedded clause 'who lives next door' helps to identify the man who is
being talked about. It supplies, in other words, essential information about
him. Blot out the clause and the message is incomplete, hardly a meaning-
ful sentence. The non-restrictive relative clause, on the odier hand, offers
incidental, non-essential information about the referent, as in the follow-
ing example:
2 My next-door neighbour, of whom I have had occasion to speak to you in the
past, is a millionaire.
The clause 'of whom . . . in the past', far from being an integral part of the
sentence, in fact breaks the continuity of the idea expressed by the main
clause: 'My next-door neighbour is a millionaire'. Blot out the intrusive
clause and the essential message is still clear.
However, the notion of restriction/non-restriction is not confined to
relative clauses alone, as most pedagogic grammars seem to suggest. It
appears to run through the entire spectrum of English grammar, a kind of
pervasive 'deep structure' feature realized in dependent clauses odier dian
the relative, in certain types of co-ordinate constructions, in phrases, in
individual words even. Compare, for instance, the dependent clauses in
each of die following pairs of sentences:
3a. The police fired on the crowd when they turned riotous.
b. The police fired on the riotous crowd when all attempts at persuading them
to disperse/ailed.
or
The crowd must have turned riotous because the police had tofireon them.

48 ELTJournal Volume 36/1 October 1981

4a. The boy was punished because he was insolent.
b. The insolent boy was punished because the teacher wanted to make an
example of him to the class.
5a. I never seem able to please my master no matter how hard I try.
b. My master is such a perfectionist that I never seem able to please him
no matter how hard I try.
While the dependent clauses in 3a, 4a, and 5a are restrictive, those in 3b,
4b, and 5b are non-restrictive and therefore more or less redundant. What
makes the clauses in die (b) examples non-restrictive is not any formal
syntactic feature, but the extra lexical information supplied by the main
clause.
Sometimes a dependent clause may be construed as non-restrictive even

Downloaded from https://academic.oup.com/eltj/article/36/1/48/430946 by guest on 22 February 2022

when there is no such explicit lexical information available in the main
clause:
6 He will come, whether you invite htm or not. (Compare: He will come ifyou
invite him.)
The clause 'whedier you invite him or not' is non-restrictive—unlike the
clause in parendieses 'if you invite him'. Here die determining factor, it
would appear, is the implicit knowledge die speaker has of die 'he' of the
main clause, who presumably is a known gate-crasher.
This is true of die non-restrictive relative clause as well. There is nothing
in die syntax diat would account for die relative clause in 2 above—the
restricdve as well as die non-restrictive clauses share die same syntactic
features. The diing, however, diat is peculiar to 2 and differentiates it from
1, is die extra lexical information contained in die noun phrase of die main
clause ('My next-door neighbour'), which renders die information
supplied by the dependent clause redundant or nearly so.
At times die question of whedier a relative clause is restrictive or non-
restrictive can only be decided widi reference to die implicit socio-cultural
assumptions and values of die language user. Thus, an utterance like:
7 My wife who lives in the village is arriving today.
will mean one thing in Hausaland in Northern Nigeria and usually a
different diing in Soudi India. The same embedded clause widi die same
antecedent noun phrase in die main clause allows two different interpreta-
tions according to whedier die clause is construed as restrictive or non-
restricdve, which again would depend on what die language user and his
community consider die accepted matrimonial norm. This makes it
obvious diat, in dealing widi the notion of restriction/non-restriction (and,
for diat matter, generally widi language in use), we are dealing widi
questions way beyond syntax. True, diere are phonological/graphological
markers to distinguish one meaning of 7 from die odier, but these are
merely a way of making explicit to die listener/reader die choice of
meaning that die speaker/writer has already made from die same un-
differentiated syntacdc structure. And in any case no such help is available
when attempting to resolve die ambiguity of an utterance like:
8 My country wife is arriving today.
The way such an utterance is understood would depend entirely on die
cognitive and experiential reality of die listener/reader.
To return to the point made early on, all types of dependent clauses are
capable of being restrictive or non-restrictive, die only exception being
noun clauses, which are invariably restrictive. Even co-ordinate construc-
tions, notably parendietic clauses, can at times be non-restrictive:

Restriction in English grammar 49

9 My late uncle—well do I remember him—enjoyed in his time a reputation
as a philanthropist, wholly undeserved in my view, for he was a mean old
man.
10 He told me the story—or perhaps it was his son.
I l l come from Lancashire, as you have probably guessed by now.
. As with clauses, so with phrases. All types of phrases—infinitive, par-
ticipial, appositive, absolute—can function as restrictive or non-restrictive
in their given environment; die phrases in parendieses below are all non-
restrictive :
12 She was too deeply shocked to say anything. (Compare: To judge by her
silence, she must have been deeply shocked.)

Downloaded from https://academic.oup.com/eltj/article/36/1/48/430946 by guest on 22 February 2022

13 The guests, bored by the host's interminable monologue, left as soon as diey
decendy could. (Compare: The bored guests, in a hurry to get away, left
widiout saying good-bye to the host.)
14 Finding himself alone and in a strange country, die man was at a loss to know
what to do. (Compare: Judging by his record to date, he was a most un-
fortunate choice as party secretary.)
15 The only child of his parents, he was pampered into a spoilt brat.
(Compare: Even Shakespeare, the well-known English playwright, could at
times be tiresome.)
16 The weather being favourable throughout the morning, the climbers had little
difficulty in reaching die summit. (Compare: The old men, some of them
in their late seventies, were left to fend for themselves.)
While in most cases clauses and phrases are easily interpreted as
restrictive or non-restrictive, it may not be die case widi single words,
especially prenominal nouns and adjectives. Prenominal nouns and
adjectives are usually restrictive but diey can have a non-restrictive
function, as in Keats' 'leaden-eyed despairs':
17 Where but to think is to be full of sorrow And leaden-eyed despairs.
(Ode to a Nightingale)
Again, in die following sentence:
18 Biped man has been die enemy of every odier species on eardi
die prenominal 'biped' is non-restrictive, since die sentence does not
provoke the question 'What of triped or quadruped man?' The same word,
however, occurring in a sentence like diis one:
19 Biped man, from an evolutionary point of view, has been somediing of
a mistake
could be construed as restrictive to mean diat man or his immediate
forebear made a mistake when in die course of his evoludon he decided to
stand up and turn biped. In
20 Mad men are dangerous
'mad' is clearly restrictive, while in
21 Mortal men have but a short time to live
'mortal' is, just as dearly, non-restrictive. The fact diat all men are under-
stood to be mortal while all men are not understood to be mad would
account for one adjective being assigned a restrictive, and die odier a non-
restrictive, significadon.
However, in some cases, die ambiguity persists and cannot be resolved
widi reference to die immediate linguisdc context. Consider, for example,
die prenominals in die following sentences:
22 Rash young men are not to be trusted.
Does it mean: 'Young men are rash and are not to be trusted' or 'Young
men who are rash are not to be trusted' ?

50 K. Narayanaswamy
23 Fatalistic orientals see no point in kicking against the pricks.
An observation of this kind occurring, say, in a travelogue by a Westerner,
would almost certainly mean diat all orientals are fatalistic, but by itself it
could carry a restrictive meaning.
Determiners too, like prenominal nouns and adjectives, can be restric-
tive or non-restrictive (i.e. specific or universal), as in the following
examples:
24a. Some books were found missing from die shelves.
b. Some men are difficult to work with.
25a. A boy has been reported lost.
b. A boy is usually more mischievous than a girl.
26a. The films shown were decidedly boring.

Downloaded from https://academic.oup.com/eltj/article/36/1/48/430946 by guest on 22 February 2022

b. Films are a popular pastime.
To sum up, the notion of restriction/non-restriction is a pervasive deep
structure feature that is realized in a variety of ways in the language, cer-
tainly in more ways than pedagogic grammars seem to allow for, and the
categorization of a given clause, phrase, or word as restrictive or non-
restrictive depends not so much on syntax as on lexis and, in die absence of
lexical help, on die language user's understanding and experience of
reality. Secondly, a restrictive string (clause or phrase) modifies the
meaning of the NP or VP to which it relates. It has a defining, specifying or
identifying function and is dierefore integral to die overall idea diat is
being expressed. The non-restrictive string, on the odier hand, is essen-
dally linear, additive, and more or less supererogatory; it seeks to expand
die field of reference by trying to include more and more of peripheral
informadon. At die level of single words or morphemes (prenominal
nouns or adjectives and determiners of die obligatory class), die restricdve
specify, while die non-restrictive tend to universalize. Finally, as users of a
human language, we particularize as well as generalize, 'restrict' as well as
'let go'. The restricdve/non-restrictive strings, it would appear, are the
devices diat die language makes available to us to do just diat—a point well
worth remembering when teaching adult or advanced learners of
English. D
Received April 1981.

The author
K. R. Narayanaswamy has taught English language research in this area. He has an MA in English
and literature for over thirty years, and has also Language and Literature from the University of
trained college and university teachers of English in Madras, and a Postgraduate Research Diploma in ESL
India. Since 1972 he has been teaching at Bayero from the Central Institute of English, Hyderabad. He
University, Kano, Nigeria, and is currently working on is the author of Reading Comprehension at College Level
a three-stage reading course based on his own (OUP, 1972).

Restriction m English grammar 51