You are on page 1of 54

CHAPTER IV

METHODS OF SEMANTIC ANALYSIS

4.1 Componential Analysis.

A much more explicit, clear and economical way

of characterizing the lexical meaning relations is to

use the method of componential analysis which assumes

that words do not have unitary meanings but are

complexes of components. The majority of structural

semantists subscribe to some version or other of

componential analysis. Componential analysis aims at

discovering and organizing the semantic components of

the words. On this view, woman would be analysed as

having for its meaning a complex of components of


[FEMALE], [ADULT], [HUMAN],

153
spinster as having for its meaning a complex of

components of

[FEMALE], [ADULT], [HUMAN], [NEVER MARRIED].

Using the method of compontial analysis, Kempson

(1977:86) formally defines hyponymy and incompatibility.

A lexical item P can be defined as a hyponym of Q if all

the features of Q are contained in the feature

specification of P; and incompatibility can be defined

as a relation between a set of items P, Q, R, ____ if

they share a set of features but differ from each other

by one or more contrasting features.


e.g. Spinster is a hyponym of woman

and Spinster is incompatible with bachlor

Componential analysis describes the meaning of

words and phrases, as Lyons (1977:317) observes that the

sense of every lexeme can be analysed in terms of a set

of more general sense components (or semantic features),

some or all of which will be common to several different

lexemes in the vocabulary. Each lexical item is defined

in terms of the components. According to Lehrer

(1974:46) componential analysis tries to be systematic

in contrast to traditional lexicography.


Anthropologists turned to componential

analysis for kinship terms. (cf. Goodenough 1956,

Loundsbury 1956, Wallace and Atkins 1960). Kempson (18)


argues that such componential analysis is not of course

restricted to kinship terms : it can be applied in many

154
areas of the vocabulary. For example, the distinction

between murder and kill can be stated explicitly and

economically if murder is analysed as having a meaning

which is a complex of components representing intention,

causation ,death and kill as having a complex of only

the components representing causation and death. In

addition to the kinship terms, certain sets of animals

are used to illustrate and argue for componential

analysis. Hdelmlev (1961) and Lamb (1964), for example,

present the following table :

sheep ram ewe lamb

horse stallion mare colt

chicken rooster hen chick

The items in each column (except first) and each row

share a feature : [MALE] for column two, [FEMALE] for

column three, and [YOUNG] for column four. The

component common to each row can be called by the same

name as the first word. Lehrer (1974:46) points out

that componential analysis presupposes aspects of the

field theory, for instance, the necessity of looking at

a set of words in a carefully delineated area which have

basic semantic features in common but whose meanings

contrast with each other by virtue of one or more

differences in respect to several other features

(Lounsbury, 1956:193).

155
The two Important semantic' theories in which

syntax and semantics is combined are based on

componential analysis are those of Katz and Weinreich.

Katz and Fordor's semantic theory has two

components : a dictionary and projection rules. In


dictionary entries each sense of a lexical item has

sense characterization, that is, a definition expressed

as a set of components. Katz's and Fordor's early work

(1963) postulated two kinds of components - semantic

markers and distinguishers.

Weinreich explains how the meaning is derived

from fully specified meanings of its parts (1966:417).

In his theory sets of semantic features for a lexical

item may be unordered (clusters) or ordered

(configuration).

Cluster : (a,b) = (b,a)

Configuration : (a b) # (b a)

Some lexical items may have clusters and configurations

in their semantic structures, e.g. (a,b) —> (c,d).

Linking constructions are those which form

clusters. Black + dog is a linking construction and

forms a cluster (Black, Dog). Non-linking constructions

are those that yield configurations. Transfer features

constitute another aspect of Weinreich's theory. A

transfer feature transfers or supplies what is necessary

for a possible interpretation. In place of projection

rules, Weinreich proposes two seperate components for

156
the semantic process : the semantic calculator and the

semantic evaluator. The semantic calculator consists of

semantic rules that compute the meaning of a sentence

from the meanings of the idividual lexical items and the

syntactic structure. The function of semantic evaluator

is to quantitatively measure and compute the deviance of

sentences according to the number and kinds of rule

violations. A sophisticated evaluator would be

sensitive to stylistic variations, and the measurement

of device would vary according to the discourse and the


text.

Thus, Katz and Fordor (1963), Weinreich

(1966), Lamb (1964) and Nida (1964) have intergrated

semantics and syntax within the framework of

transformational grammar. They applied componential

analysis in many areas of vocabulary. Being a

systematic means of examining sense relations,

componential analysis has been widely used to study and

interpret vocabulary.

Nida (1975:29-31) observes that the meanings

of lexical units within a single domain may be related

to one another in four basic ways : 1) contiguous 2)

overlapping 3) included or 4) complementary. In the


analysis of meanings, the contiguous sets are certainly

the most important, since the meanings in sets are

maximally similar, while at the same time they differ


significantly in at least one crucial component. The

157
series of related meanings of chaiibench, stool and

hassock constitute precisely such a set. They belong to


the domain of pieces of furniture on which people sit,

but there are certain definitive distinctions between

them. The meaning of chair contrasts with the meanings

of bench in that a chair is for one person, while a

bench is for several persons. Chair also contrasts with

stool in that a chair has a back and a stool normally

does not, while a hassock, which is also for one person,


has no back and also lacks legs.

A componential analysis of the meaning of this

series indicated clearly that they are closely related,

in that they share a number of common components; e.g.

they are all objects, countable, artifacts, furniture


and made for sitting. But they also differ from one

another in at least one crucial component. This type of

contiguous semantic structuring of a domain may be


diagrammed as in the following figure :

158
Nida (19T5 b : 64) states that the actual

linguistic procedures employed in componential analysis

consist of four types : Dnaming 2) paraphrasing

3)defining and 4) classifying.

The process of naming is in certain respect

similar to reference, though the perspective is somewhat

different. Reference is usually described as the

relation established between a linguistic unit and a

referent, while naming is the specific act of

designating such as referent. The distinctive features

of any semantic unit can be defined by employing certain

types of paraphrase. Uncle may be paraphrased as My

father's brother or my mother's brother .The

distinctive feature of repentance may be given as he felt

sorry for what he had done and determined to change his

way of life. Defining is a highly specialized form of

paraphrase and is rarely used in actual language

situation. It consists essentially in combining all the


various specific paraphrases into a single statement

based on the diagnostic components of the particular

meaning in question. Classification involves (1)

lumping together those units which have certain

features in common, (2) separating out thse units which

are distinct from one another and (3) determining the

basis for such grouping.

159
According to Goodenough (1965), componential

analysis begins with extentional and intensional

definitions of significational aspects of meaning. For

him, the extensional definition of the word uncle may

list the following relationship : mother's brother,

father's brother, mother's half brother, mother's

sister's husband, father's sister's husband and father's

half-sister's husband.

The intensional definition defines the word

uncle as any kinsman by blood or marriage which is

simultaneously (a)male, (b)two degrees in geneological

distance from ego, (c)not lineal, (d)in a senior

generation and (e)not connected by a marital tie in

other than senior generation of relationship. The

meanings of the term is thus specified conjunctively as

a combination of all these definitive attributes. If we

vary any one of these, the resultant meaning cannot be

attributed to the word uncle. For example, if we vary

the attribute (a) keeping the other attributes common,

the resultant meaning would be that of the word aunt; if

we vary only (b), we get the meaning of great uncle; and

so on.

Morris (1955), one of the proponents of

componential analysis distinguishes between denotative

or significational meaning on the one hand and


connotative or associational meaning on the other. The

distinction is Ddesignation 2) denotation

160
3) signification 4) connotation.

For example, cloudy sky —

1) designates a class of images or concepts,

2) denotes a specific image on any one occasion of its

use (e.g. the cloudy sky of last evening),

3) signifies definitive attributes of the sky, and

4) connotes images or concepts associated with cloudy

sky (e.g. rain, chill, etc.)

Leech (1981:89) defines componential analysis

as the analysis of word meanings often seen as a process

of breaking down the sense of a word into its minimal

components. He cites, as an example, the words-man,

woman, boy, girl and other related words which belong to

the semantic field the human race. The relations


%

between them ' may be represented by a two dimentional

field diagram'.

Male Female

1 1 l
1 1 i

adult ! Man ! Woman j


I
«
. \1 \I
t 1 1
1 i l

Young ! • boy ! girl !


1
1
1
_1
1
l

human

The diagram shows two dimensions : that of sex and that

of adulthood; a third dimension is presupposed by the

isolation of the field as a whole : that between human

and non-human species.

161
Another way of representing realtions is by

using feature symbols like HUMAN and ADULT :

man : + HUMAN + ADULT + MALE

woman : + HUMAN + ADULT - MALE

boy : + HUMAN - ADULT + MALE

girl : + HUMAN - ADULT - MALE

These formulae are <sailed the componential definitions

of the items concerned; they can be regarded as

formalized dictionary definitions. The dimensions of

meaning themselves are semantic oppositions.

Using formulae like these, we can show the

synonymy of two items by giving them both the same

componential definition. For example, both adult (in

its human sense) and grown up can be given the same

definition + HUMAN + ADULT, even though they clearly

differ in stylistic meaning, the one being rather

formal, the other colloquial. Thus, if two expressions

have the same semantic representation and thus are

componentially synonymous, we get a case of componential

synonymy (Leech 1969:21).


Leech (1981 : 99-107) describes the semantic

oppositions :

1) binary taxonomy : e.g. alive - dead.

2) Multiple taxonomy : e.g. types of metal- gold,


copper, iron, silver; types of noses - bang, crash,

clatter.

162
3) polar oppositions : e.g. small-large, rich-poor,

deep-sha11ow;

4) realtive opposition : e.g. parent-child, own-

belong to, up-down, above-below, before-after, left-

right, west-east;

5) hierarchic oppositions : e.g. one-two-three,


January-February__ ..., inch-foot-yard, and

6) inverse oppositions : e.g. all-some, possible

necessary, allow-compel, be willing-insist, still-

already.

Leech (1981:110) tries to combine the features

of different semantic oppositions and formulates the

relationship through a redundancy rule. For example the

features + MALE and - MALE presuppose the presence of

the feature + ANIMATE; this may be expressed as follows:

± MALE reqires + ANIMATE

(i.e if + MALE or - MALE is present in a componential

formula, then + ANIMATE is also present). The effect of


this rule is to add the feature + ANIMATE automatically

to any formula in which the feature + MALE or - MALE

occurs. Leech observes that through redundancy rules

great economies can be made in the definition of words,

and yet one can still show their interrelation with many

other words.

Further, it is also observed that componential

analysis is based on the principle of lexical

decomposition. Analysing the- senses of words into

163
components that represent the sense of other words

assume that the sense of complex words are represented

in the mind in terms of the sense of simpler words. The

sense of the complex word is said to be decomposed into

components. For example, the sense of bachelor is

decomposed into components like : MALE, HUMAN, ADULT &

NOT MARRIED. But there remains the same complexity, as

the words having implicit negatives (doubt, deny).

Leech (1967) proposes some important

principles in componential analysis :

1) That the technique of componential analysis can,

with certain additions, be extended to the semantic

description of vocabulary as a whole.

2) That semantic statements should be referable to

the empirical evidence of important testing procedures.

3) That semantic comparisons between languages can

be made on the basis of common features of meaning.

Bendix (1966) has used componential analysis

to study and interprete a set of verbs in English, Hindi

and Japanese. He has explored a way of operating with

only those semantic components which can be extracted

from the mutual oppositions of the particular verbs. He

says, " we have gone beyond conventional componential

analysis in selecting a set of words that is only part


of a larger system of mutual oppositions which, unlike a

kinship terminology for example, is not clearly

delimitable (at least not yet) “ (Bendix : 3). He

164
I

modifies the apparatus of componential analysis by

defining a set of components as statements.

Nida (1975:132) has used componential

analysis to analyse the various meanings from the

standpoint of the glosses and the components which ,are


!

to be found in the relations which exist between

meanings. He tries to study the same meanings of the

Biblical Hebrew kbd charted with five culturally

relevant components, and concludes that the traditional

method of defining meaning would not give the clear

meaning of the words, but componential analysis makes it

possible to reject such an either'-or proposition. It

gives the clear meaning of the words.

Limitations: ,

Componential analysis appears, at first sight,

to be an attractive way of handling semantic relations.

But it raises far too many difficulties to be at 1 all

workable.

Kempson (1977:20) argues that ‘so stating jthe

inter-relationships between words in terms of more

primitive semantic components, one is transferring the

burden of semantic explaination from word meaning onto


the components which together, in different

combinations, constitute word meanings. Indeed what


remains to be explained in such componential analysis is

the relationship between the words of a given language

165
and the apparently independent components'.She observes

the central problem of the relationship between the word

human and the component [HUMAN]. She says that the

account of the semantic components themselves given by

linguists using the methodology of componential analysis

is often no more substantial that de Saussure's or

Sapir's characterization of word meaning.

Lee (1968:192) criticies componential analysis

for treating language as if it is a collection of

separable and self-sufficient parts. Lyons (1977:334)

observes that componential analysis leaves unexplained

at least as much as it succeeds in explaining. For

example, if the meaning of the lexemes man, woman,

adult, girl, boy and child are analysed in terms of the

sense components HUMAN, ADULT & FEMALE, male child

should be synonymous with boyi but it is not. An

eighteen-year old boy is certainly not a child. Lyons

says that componential analysis is always liable to fall

victim to rather facile overgeneralization. Whenever

we appeal to such allegedly common sense-components as

HUMAN, ADULT & FEMALE, we must ask ourselves what their

cross -lexemic status is , how they are to be identified

and what their explainatory power is.

There are certain problems in componential


analysis, for which componential analysis is criticised.

It is often said that compoential analysis

accounts for only some parts of a language's vocabulary.

166
It suffers from a various circle in that it meerly

explains one set of symbols by another set of symbols.

It postulates abstract semantic entities (semantic

features) unnecessarily. It postulates universal

features of meaning, and therefore relies upon the

strong assumption that same semantic features are found

in all languages.

Leech (1981: 117) tries to defend these

objections. He says that componential analysis cannot

have wider goal : It is meant to explain word sense, not

the encyclopedic knowledge which must enter into theory


*

of reference. It is not plausible that componential

analysis could deal with the whole vocabulary of a

language but at the same time componential analysis can

be fitted into a more powerful model of meaning, with

additional levels of analysis apart from componential

analysis.

Leech (1981:117) agrees with the fact that


componential analysis as a theory of word meaning is

controversial. Its limitations, which have been widely

commented on, should not lead one to abandon it. Leech

tries to look for ways in which it may be fitted into a

larger and more adequate model of conceptual semantics.

Thus, componential analysis has made

considerable contribution to the development of

semantics. It is treated as a useful and revealing

technique for demonstrating relations of meaning between

167
words. Componential analysis has developed a technique

for analysing the meaning of words. Bendix (1966) and

Leech (1981), have tried to integrate componential

analysis into the system of predication analysis.

4.2 Predication analysis :

Some scholars believe that the meaning of a

sentence is merely the sum of the meanings of the words

and other constitutes which compose it. Leech

(1981:123) proves it to be false. He argues : firstly,

we could not, if this were the case, tell the difference

of meaning between -

My wife has a new dog,

My new wife has a dog,

My new dog has a wife,

My dog has a new wife, etc.-

all these would have to be regarded as synonymous.

Further, he notes that - if componential analysis

applied not only to words but to sentences as a whole,

then a sentence like :

The woman likes the puppy

would have to be regarded as meaningless : the features

+ ADULT in woman and - ADULT in puppy would contradict

one another.

Projection rules which amalgamate the meanings

of sentence constituents in particular ways in order to

produce the meanings of whole sentences, is worked out

by Katz and Fordor. But this approach is misleading.

168
For example if we merely add components together as we

use the projection rules then it will follow that

cats chase mice

and

mice chase cats

have exactly the same meaning. Chase is essentially

relational. Secondly, a problem arises in that the

same component may at times merely provide the

environment for amalgamation, at others be part of the

derived path. For example - pretty child, buxom

neighbour where the noun phrase is presumably (-male)

but the nouns child and neighbour are not.


Weinreich (1966) talks about a special

transference rule which transfers the relevant component.

But such examples show that componentlal analysis does

not provide a simple way of proceeding from the meaning

of lexical items to the meaning of sentences by a

process of the adding together of the components through

amalgamation.

Palmer (1976:106) observes that an analysis in

relational terms seems to offer a far more satisfactory

solution to the problem of sentence meaning than

componential analysis. Such analysis will have much in

common with the predicate calculus used by logicians.


Leech (1969:19-25) uses the terms systemic analysis and

structural analysis for componential analysis and

predication analysis. He tries to integrate

169
componential analysis into the system of predication

analysis. Predication analysis is interrelated with and

complementary to componential analysis.

The meaning of a sentence is a PROPOSITION.

Propositions consist of TERMS, which are of two types,

PREDICATES & ARGUMENTS. The predicates are the

relational terms and usually correspond to verbs in

sentences; the arguements are the terms that are

related, and usually correspond to nouns. For example


in John loves Mary we have a relation (a predicate)

expressed in loves and two items (two arguments)

expressed in John and Mary. We get a formula :

[Love] John, Mary.

The agrumenis are ordered in that

[Love] Mary, John.

will be the formula for something different

- Mary loves John.

Thus using X and Y as variables we can say that

[Love] X, Y is distinct from

[Love] Y, X.

It is also true that some relations are symmetrical.


For example

John is married to Mary entails

Mary is married to John.

Thus.

[Married to] X, Y entials [Married to] Y, X

but this is a property of this particular predicate, and

170
not one of predicates in general.

Palmer (1976:108) uses predicate calculus to

account for modifiers such as brave in brave men. He

analyses the noun phrase : the brave man -

(Def X) ([Man] X [Brave] X ) -

i.e. The entity X such that X is a man and X is brave.

The use of the round brackets indicated the whole of X

is a man and X is a brave is part of the such that

qualification (or in more technical terminology is within

the scope of (Def X)). The formula for the more

complex items like The brave man ran away will be;

[Ran away] (Def x) ([man] x [Brave] x);

i.e The entity X such that X is man and X is brave ran

away.

For The young boy loves the beautiful girl

Palmer (109) proposes :

[Love] (Def x) ([Boy] x.[Young] x), (Def y) ([Girl] y

[Beautiful] y)

Leech (1981:128) proposes the following three-

tire structure for the analysis of sentence meaning. He

prefers predication to proposition.

ARGUMENT !
I ______ _____

[PREDICATION ! consist of ! consist of ! FEATURE!


I I I I I
I.................... <
PREDICATES! ' '

He gives a fairly complete analysis of a plf<?dication,

171
that is ,except for the omission of redundant features :

A man is/was in front of a woman

PREDICATION

Predicate

i
f —> DIRECTIONAL
I
I LATERAL
I

in front of
Agrumentl Agrument2

+ HUMAN + HUMAN
+ ADULT + ADULT
+ MALE - MALE
+ SINGULAR + SINGULAR

a man a woman

All these contrastive features play a role in the

meaning of that sentence. Predicate and Argument have

different roles in the whole predication. The

predicate is the major or pivotal element, and may be

said to GOVERN the arguments, which are in relation of

DEPENDENCY to it. According to the number of argument,

we have one place predicate, two place predicate, or

even three place predicate. There may also be four or


five - place predicates. We can also have no-place

predicates, e.g. It is raining. It is semantically

empty. These can be diagrammed as follows :

172
—> p <— p P
/ \ / \ /
/ \ / \
a b a b

two place predicates one place pardicates

One place predicates govern only one argument

For example : Alsatians are large

PREDICATION

Predicate

i " BIG !
1 I I
f — f - - - t

/
/ be large
Agrument /
--------------- /
* BREED |
- SINGULAR !
I
...... I

Alsations

Another example is : Cats are small animals.

PREDICATION
Prdicate

! BIG
+ ANIMAL
- SINGULAR

/
/ be small animals
/
/
/
Agrument /
/
---------------- /
! * SPECIES !
! - SINGULAR !
1 I
( -.... ......... ... -................. . ............. 1

cats

173
A one place predicate is typically realised by a nominal

or adjectival complement. It does not contain

relational features such as —> DIRECTION, since a

relation implies that there are two arguments to be

related.

Leech (1981:130) defines predication analysis

as a kind of semantic analysis, which is Interrelated

with, and complementary to, componential analysis, which

consists in breaking down predication into their

constituents. Componential and predication analysis

together enable us to represent the greater part of the

meaning of sentences. For example, the sentence A man

is/was in front of a woman could be placed in a simple

linear arragement :

A man = a

in front of = <— P

a woman = b

Two place predication is : (a. —> P.b)

i.e. Predications with their componential content :

( a -- > P b )

HUMAN DIRECTION HUMAN


ADULT - LATERAL ADULT

MALE - MALE

SINGULAR SINGULAR

174
and to place the components In a simple linear

arragement :

(HUMAN ADULT MALE SINGULAR.DIRECTION LATERAL. HUMAN

ADULT - MALE SINGULAR) A man in front of a woman.

In the same manner the sentence : Alsations

are large could be placed in a simple linear

arrangement.

One place predication : (a : P )

i.e predication with their componential content :

(a : P )

! * BREED j BIG !
i - SING ! |
1 I I
I................... < ......... »

Alsatians be large

and to place the components in a simple linear

arrangement :

(BREED - SINGULAR : BIG ) Aleations be large.

These are linear predication-componential

formula, which gives the whole meaning in a linear form.

Leech (1981:138) observes that meaning seenjs

to flow sideways from one part of a sentence to another;

that certain features of meaning are predictable from

semantic environment and that any contradication of such

features will result in an unacceptable utterance. This

phenomena is discussed under the heading of selection

restrictions (collocation). According to him,

predication analysis provides a connection between

meaning relations.

175
4.3 Collocation
4.3.0. Origin and Background

In the present section an attempt has been

made to develop the Flrthian concept of collocation and

to define it somewhat more closely.

Firth's entry into the field of linguistics in

the early 1930's was certainly not a conspicuous one and

most important for an understanding of his later work is

the fact that he participated in the seminars conducted

at that time by the anthropologist Bronislaw Malinowski

at the University of London (Langendoen 1968 : 1).

Malinowski

In order to understand most of the important

aspects of Firth's work, it is necessary to be

acquainted with Malinowski's linguistic studies.

Malinowski was one of the few important anthropologists

of his time, and the only one in Great Britain, to have

had an abiding interest in language itself.

In his paper Classificatory particles in the

language of Kiriwina Malinowsky expressed the hope that

he would someday be able to write a grammer of

Kiriwinian (1920, P.67). He asserted that there is a


need for the development of a theory of semantics

connected closely with enthnographic theory. He tried

to connect language with culture. According to

Malinowski, "In reality the problem of defining the

176
meaning of a single word and of processing correctly in

the translating of terms is as difficult as any which

will face us... It will be obvious to anyone who has so

far followed my argument that isolated words are in fact

only linguistic figments, the products of an advances

linguistic analysis. The sentence is at all times a

self contained linguistic unit, but not even a sentence

can be regarded as a full linguistic datum. To us, the

real linguistic fact is the full utterance within its

context of situation " (1935 : 11). For Malinowiski the

meaning of any linguistic unit consisted in its


interrelations with its environment : exactly as a

single word Is - save In exceptional circumstances -

meaningless, and receives Its significance only through

the context of other words, so a sentence usually

appears in the context of other sentences and has

meaning only as a part of a larger linguistic whole. I

think that it Is very profitable in linguistics to widen

the concept of context so that it embraces not only

spoken words but facial expression, gesture, bodily


activities, the whole group of people present during an

exchange of utterances and the part of the environment

on which these people are engaged. (1935:22).


Malinowski put forward a scheme for a two­

fold contextual definition of each utterance-

l)an utterance belongs to a special context of


culture, i.e. it refers to a definite subject matter.

177
2) the another context is that the situation in

which the words have been uttered. A phrase, a saying

or a few sentences concerning famine may be found in a

narrative, or in a magical formula, or in a proverbial

saying. But they may also occur during a famine,

forming an integral part of some of those essential

transactions wherein human-beings co-operate in order to

help one another (Malinowski 1935 : 51-2).


Thus, Malinowski's theoretical position was

based on the tenets of behavioristic psychology. This

is most clearly seen in his Coral Gardens, published in

1935 after the seminars in which he and Firth

participated. According to Langendoen, "it is

reasonable to suppose that Firth, always effective in

arguing his position, actually had a considerable

influence on Malinowski during the time that they worked

together. This suggests that Firth had a well defined

theoritical framework of his own and that in formulating

his linguistic position he simply fitted key concepts

from Malinowski into this framework." (Langandoen


1968:2-3). The single most important notion in Firth's

early writings is that of context. One might say that

Firth took the current notions of phonological,


morphological and lexical contexts added the

Malinowskian notion of context of situation.

There are restrictions of various sorts on

what words can co-occur, i.e. words seldom occur alone.

178
There are following hypotheses on the occurrences of

words.

4.3.1. The lexical Position :

The London school of linguistics was the

creation of J.R.Firth. Its date of origin can be given

as 1944. In this year Firth acceded to the chair of

General Linguistics at the University of London.

Because of Firth's long association with the university

of London the school has come to be known as the London

school. It was the dominant school of descriptive

linguistics.

It was in Modes of meaning, Essays and

studies (The English Association) 1951 (reprinted in

papers in Linguistics 1934-1951 Oxford 1957) that J. R.

Firth proposed to bring forward as a technical term,

meaning by collocation, and to apply the test of


collocabill ty.

The term collocation was not originally

Firth's and he may well have been influencced in its

selection by H.E.Palmer, who from Tokyo wrote a

monograph on the subdect in the nineteen thirties

the term appears in Palmer and Blandfords Grammar of

spoken English and Plamer doubtless did much to shape

the second Interim Report on English collocations

(Department of English, Tokyo), mentioned in J.E.

Mansion's edition of Harrap's Standard French and

English dictionary and R.P.L. Ledesert's recent

179
updating of Mansion in the introductory bibliography of

works consulted. But the more or less technical

employment of the term antedates Palmer, too, by .upwords

of 325 years according to the N.E.D. although the

application of the concept and term to features of

linguistic structure seems to be comparatively recent.

Palmer appeared quite properly to see collocations as a

highly abstract order of compatibility between

linguistic elements but did not define the term with any

degree of precision. Firth, for his part, appropriately

thought of it as primarily lexical, as means of

restricting the vagrancy of words and of providing

stylistic delineation of his restricted languages. The

lexical emphasis has been taken further by the neo-

Firthians, and notably by M.A.K. Halliday and J.McH.

Sinclair, to the point of regarding collocational study

as independent of grammar (Mitchell 1975:134).

Firth concerned himself mainly with the

development of two new ideas, one concerning phonology

and the other semantics. These ideas came to be

considered the hallmarks of London school of

Linguistics. These ideas are the notions of prosodic

analysis in phonology and of meaning by collocation in


semantics. Throughout his later period Firth maintained

intact his understanding of the notion of meaning. At

first he expressed this notion in his paper Technique of

semantics.

180
In his paper Modes of meaning, however, one of

the five dimensions of meaning, the lexical dimension

(or mode), received much greater attention. The meaning

that was supposedly contributed by this mode was given a

new name, "meaning by collocation." In this essay he

argues that one level or aspect of the meaning of a word

is the set of other words that it collocates with.

Since he believed that a linguist's task is to analyse

text, rather than to state generative rules for novel

sentences, many of his examples are concerned with

specific texts, especially literary works. He was also

aware of the potentiality of collocations and proposed

the test of collocability, in which various lexical

items are substituted and the results judged. He gives

the following sentences with the word ass :

i) As ass like Bagson might easily do that.

ii) He is an ass.

iii) You silly ass!

iv) Don't be an ass!

" One of the meanings of an ass is its

habitual collocation with an immediately preceding you


silly, and with other phrases of address or of personal

reference.... There are only limited possibilities of


— ....... . ■■ —— - .

collocation with preceding adjectives, among which the

commonest are silly, obstinate, stupid, awful,

181
occasionally egregious. Young is much more frequently

found than old. The plural form is not very common."

(Firth 1969:195).

According to Firth, it is seen that

collocation is not to be interpreted as context by which

the whole conceptual meaning is implied. It should not

be confused with citation. The habitual collocations in

which words under study appear are quite simply the mere

word accompaniment, the other word material in which

they are most commonly or most characteristically


embedded. It can safely be stated that part of the

meaning of cows can be indicated by such collocations as

They are milking the cows, cows give milk. The words

tigresses or lionesses are not so collocated and are

already clearly seperated in meaning at the

collocational level (Firth 1968:180).

It must be pointed out that meaning by

collocation
« ■
is not at all the same thing as
......................... ■■ “ 1 1 * **"
contextual
r-.i_

meaning, which is the functional relation of the

sentence to the processes of a situation in the context

of culture (Firth 1969:195). Firth gives the example

of the language of Lear's limericks where man is

generally preceded by old, never by young. Person is

collocated with old and young.

Stating collocational meanings, according to

Firth, does not involve definition or paraphase.

"Meaning by collocation is an abstraction at the

182
syntagmatic level and is not directly concerned with the

conceptual or idea approach to the meaning of words.

One of the meanings of night is its collocability with

dark, and of dark, of course collocation with night"

(Firth 1969:196). Therefore, Leech (1981:17) says that

collocative meanings consist of the associations a word


acquires on account of the meanings of words which tend

to occur in its environment. Pretty and handsome share

common ground in the meaning good-looking, but may be

distinguished by the range of nouns with which they are

likely to collocate :

->girl —>boy
boy man
woman car
pretty flower handsome vessel
garden overcoat
colour airline
village typewriter
—> etc. —> etc.

The ranges may well, of course, overlap : handsome woman


and pretty woman are both acceptable, although they

suggest different kind of attractiveness because of the

collocative associations of the two adjecties. Further

examples are quasi-synonymous verbs such as wander and

stroll (cows may wander, but may not stroll) or tremble

and quiver (one trembles with fear, but quivers with


excitement) collocative meaning is simply an

idiosyncratic property of individual words.

Collocations of a given words are statements

183
of the habitual or customary r>laces of that
.................... - - T-mwi—
words
mmm
in

collocational order but not in any other contextual

order and emphatically not in any grammatical -order. The

collocation of a word or a piece is not to be regarded

as mere juxtaposition, it is an order of mutual

expectancy. The words are mutually expectant and

mutuallv_prehended. (Firth 1968:181). The collocability

of lady is most frequently with young, but person with

either old or young. Thus collocation is the

conventional associatinn-of words together in discourse.

A stock example nowadays is the use of the word maiden,


■ ilMII-JII I--"**

which rarely appears in present day English as meaning

girl, but most often appears in collocations as in

maiden name, maiden speech (in the House of commons),

maiden voyage (of a ship), maiden over (in cricket),

maiden aunt, and so on.

Firth (1969:195)makes further distinction

between usual collocations (i.e. commonly found in

various generes or modes of discourse) and personal

ones. The later would be important for many kinds of

stylistic studies of literary forms and generes and of

particular authors and works. Firth says that

the study of the distribution of common words may be

classified into general or usual collocations and more

restricted technical or personal collocations. The

commonest sentences in which the words horse, cow, pig,

swine, dog are used with adjectives in nominal phrases,

184
and also with verbs in the simple present, indicate

characteristic distributions in collocability which may

be regarded as a level of meaning in describing the

English of any particular social group or indeed of one

person. The word time can be used in collocations with

or without articles, determinatives, or pronouns. And

it can be collocated with saved, spent, wasted,

frittered away, with presses, files, and with a variety

of particles, even with no. Just as phonetic,

phonological and grammatical forms well established and

habitual in any close social group provide a basis for

the mutual- expectancies of words and sentences at those


1 ~~!ll III IL I. _mi |MI ■'■ l_ Him

levels , and also the sharing of these common features,

so also the study of usual collocations of a particular

literary form or genre or of a particular author makes

possible a clearly defined and precisely stated

contribution to what I have termed the spectrum of

descriptive linguistics, which handies and states

meaning by dispersing it in a range of techniques

working at a series of levels."

Halliday has extended the notion of a

collocational level and discusses it in terms of a


general theory of grammar. According to Halliday, in

grammar we cannot distinguish between the items table,


desk and bench, table does not equal not desk or not

desk or bench . Here, we move over to the other formal

level, that of lexis. table, desk and bench are lexical

185
items. They are members of a lexical set. Lexical sets

are not bounded in the way that grammatical systems are

in grammar we can say : at this place in structure,

these terms are possible, and all others are impossible,

in lexis we can never say : only these Items are

possible. Lexical sets in fact are bounded only by

probabilities. Given the item chair, we are more likely

to find in the same utterance the items sit or

comfortable or high than, say haddock or reap, though

no one could maintain that the later are impossible.

This tendency to co-occurrence is the basic formal

pattern into which lexical items enter. It is known as

collocation, and an item is said to collocate with

another item or items. A lexical set is simply a

grouping of items which have a similar range of

collocation. Chair^ seat and settee belong to the same

lexical set because^, they have a number of highly

probable collocations in common : they collocate

readily, for example, with comfortable and sit


(Halliday :33). ^Therefore Halliday is of the opinion

that the collocation of words is the basic formal

relation in lexis. According to him the citations are

purely formal. They describe a word in relation to its

linguistic environment. This relation between one word

and another with which it is associated is called

collocation (Halliday 1966:19). For him collocation is

outside grammar; it has no connection with the classes

186
of the wordy It is the lexical item, without reference

to grammar, that enters into^collocations. We can say

Open the window, or an open window or the opening of the

window, it is in each case the same collocation of the

item window with the item open. The relation of

collocatioin enables us to group items into lexical set.

(Halliday 1966 : 20).

Strong and powerful are examples of items that

collocate with arguement, but the two do not share the

same collocations; powerful collocates with car and

strong with tea. Phrases like strong car or powerful

tea will either be Judged less acceptable than strong

tea and powerful car or will be Judged as different and

contrasting. But the same patterns^appear in different

grammatical contexts : He argued strongly, I don't deny

the strength of his arguement, His argument was

strengthened by other factors, versus The power of her

tea. Halliday claims that the lexical


_—
items
v
need
5
not

have any formal relationship to one another for they may

be in different sentences, such as I wasn,t altogether


convinced by his argument. He had some strong points

but they coud all be met. (Halliday 1966:150-51).

Halliday of course recognizes that frequency

of occurrence and collocability are not necessarily the

same, but he seems to favour the study of large corpora

as a source of data on collocation and set. Since

lexical collocations are comparatively rare, several

187
million words of running text would be necessary for

enough samples to draw conclusions.

Sinclair (1987) contributed to the extended

discussion of the nature of lexis, and its relation to

grammar. He discusses the relation of lexis to

semantics. Firth's concept of collocation is central in

his discussion as the concept of collocation is accepted

as an insight and celebrated in his papers. Halliday

was one of the principal celebrants, and lexis was

neatly fitted into his first full scale model, as most

delicate grammar. Nowadays, the advent of computers

made it possible in principle to study lexis and

Halliday inspired and encouraged the study.

In his paper - English lexical collocations -


A study in computational linguistics - Sinclair starts

his work by quoting Firth's meaning of collocation and


defines collocation as a technical term. [According to

Sinclair, collocation is the co-occurrence of two items

in a text within a specified environment. Significant

collocation is regular collocation between items, such

that they co-occur more often than their respective

frequencies and the length of text in which they appear


would predictT^CSinclair 1974:19).

Goran KjellmerC1987:133) defines a collocation

as illustrated in the following diagram :

188
Recording to KJellmer " a collocation is a

sequence of words that occurs more than once in

identical form (in the Brown cropus) and which is

grammatically well structured. This definition,

incidently also suggests the procedure by means of which

collocations are identified and listed.

According to Kjellmer it is true that

combinations that^ would ordinarily be accejpted as

collocations - last night, dark night, try to -occur

frequently in the Brown corpus. However combinations

like although he, but too and hall to also occur

frequently. Such combinations (1982:25) would not

normally be called collocations. Thus the frequency


alone should not be our guide to identify collocations.

There should- be another criterion viz. that of

grammatical structure. The last night type could be

said to be grammatically structured , while the although

189
he type is not . The criterion of grammatical structure

thus admits last night. However, it would also admit

yesterday evening and green ideas (which do not occur in

the corpus) as well as ebbing night and moving day

(which occur once each) combinations that would hardly

be desirable in an inventory of collocations, actual

words in habitual company. The joint use of the two

criteria, that of frequency and that of grammatical


structure, proved to be a way out of this dilemma. Only

those combinations that satisfied both criteria were to

be admitted. The operation of the criteria is

illustrated by showing the following figure.

Criteria to identify the collocations

Recurrence

+ —

a b
last night dark Yesterday evening
+ night try to green ideas ebbing
night
Gramatical
Structure c d
although he but green strayed gun
too hall to because rhions
when

Figure from (Kjllmer 1982 : 27)

It is only the type of combination found in

the upper left hand square (a) that qualifies for

inclusion. Kjellmer refers to such sequences as

collocations

Kjellmer further adds that syntanctio

190
ambiguity is a very frequent phenomenon in English.
~T-"L-— '■ - *- ..
The

superficially idential sequences will often be

structurally different. It therefore often happens that

one out of several such sequences is admissible as a

collocation while the rest are not.


For example :

When words can be used in a more fresh and

primitive way

Howard ... said he needed more fresh air, and

butter ointment was used for burns and for bruises

cereal grains have been used for centuries to

prepare.

In each pair, it is only the first of the two

marked off sequences that can be accepted as a

collocation on the strength of Itsstructural

properties, the second such sequence in each pair will

thus have to be disregarded (Kdellmer 1982:32). This

kind of identification of collocation is possible

through the context of a given sequence.

It is also taken into


1 1
consideration
' i i r .....
that
|

collocations are surface-structure phenomena. They are

not only grammatically determined but also lexically

selected. Thus the term collocation is also used in a

stricter sense to denote a special kind of lexical

relationship.
In its broadest sense, the term collocation is

more or less equivalent to recurrent word combination.

191
In the Firthian tradi. .tion of British linguistics,

however it is generally used in the stricter sense of

habitually co-occurring lexical items or mutually

selective lexical items. Both interpretations imply a

syntagmatic relationship between linguistic items, but

whereas the broad sense focuses on word sequences in

texts, the stricter sense goes beyond this notion of

textual co-occurrence and emphasizes the relationship

between lexical items in language (Greenbaum 1974:80).

It follows that the former, but not necessarily latter,

includes idioms_, compounds, and complex words, and that

the latter, but not necessarily the former, cuts across

word classes (drink heavily, heavy drinker, heavy

drinking), applies to discontinuous items (he drinks

pretty heavily) and presupposes lemmatization

(drink/drinks/drank/drinking heavily).
jEhere are also various intermediate uses of

the term collocation. Collocations are not regarded as

a probabilistic phenomenon but defined as fixed,

identifiable non-idiomatic phrases and constructions


(Benson et.al.l986K^

In the wider sense Altenberg (1990:4)calls

continuous collocations and in a stricter sense he calls

significant collocations. The Firthian notion of

collocation is thus a more extensive lexical concept


than recurrent word combination. The methodological and

practical problems^it raises are also considerable.

192
Since it is essentially a probabilistic - and virtually

open ended - phenomenon it is .well suited—for, a large -

scale corpus study.

Lehrer (1974:176) points out that the main

criticism against the lexical approach to co-occurrence

is that it does not explain anything. One lexical item

is found to collocate with a second item and does not

with a third but no explaination is given. Collocations

and sets are treated as if combinatorial processes of a

language were arbitrary.

4.3.2. The semantic Position

The semantic position is exemplified by Katz

and Fordor (1963), Katz and Postal (1964), Weinreich

(1966), Fillmore (1968), MacCawley (1968) and others.

They claim that co-occurrence restrictions are the result

of the meaning of the lexical items and that

collocations are reflections of this factT^ Selection

restrictions are to be stated in terms of what semantic

features (primes, markers) may occur together.

Katz and postal (1964:15) observe : Each

reading in the dictionary entry for a lexical item must

contain a selection restriction, i.e. a formally


expressed necessary and sufficient condition for that

reading to combine with others. Thus, the selection

restriction attached to a reading determines the

193
combinations with the readings of other lexical items

into which that reading can enter when a projection rule

is applied. A selection restriction must be stated for

each grammatical construction that an item can enter

into. Hence each verb would have selection restrictions

for subject, direct object, indirect object, instrument,

location, direction manner etc. if the verb can appear

in such constructions. Most of the examples given are

obvious and uncomplicated, such as sleep, which requires

a subject with the feature [Amimate], or hit, which

requires the feature [physical object] for a direct

object. The fact that pretty is not normally applicable

to males but could be stated as part of the dictionary


■j

entry for pretty. However, it was also apparent that

when the proper context is unspecified as to [+ Male],

the word pretty itself specifies it as [-Male].

Weinreich (1966:429) proposes transfer

features in place of selection restrictions. Weinreich

calls [-Male] in the case of pretty a transfer feature

and symbolizes it by angular brackets. He observes the

example : M(a,b —> ) and N (c,d) are formulated

dictionary entries, and M + N is a nesting construction;

then the meaning of M + N is represented by (a,b —>c,d).


But we find that when N (c,d) is constructed with M

(a,b—>), a semantic feature W appears which clusters

with (c,d). We may represent this W as a transfer

feature of M as follows :

194
Given : M (a,b—> <w>) ; N (c,d)

Then : M + N (a,b—> c,d,w).

An example of a transfer feature would be the feature

C + Time ] in the preposition during or the postposition

ago; that is to say, whatever word is constructed with

during or ago has a feature of [ + Time ] transferred to

it. The nesting word may already have an inherent

feature identical with the transferred one (e.g. during

[ + Time ] the day [ + Time ]); it may be unmarked with

respect to the feature (e.g. during [ + Time ] it) or it

may contain a contradictory inherent feature (e.g.

during [ + Time ] the wall [ - Time ] ).

Transfer features operate differently in

Weinreich's conception of a semantic part of a

generative grammar from the selection restrictions of

Katz, Fodor and Postal, but the content of these

entities would be similar. Lehrer (1974:177) quotes the

following two sentences from Fillmore (1968:381) to

illustrate the difference between the two views :

1) Peter touched"the dog.

2) Peter broke the dog.

Both touch and break requires a direct object selection

restriction of [Physical object], break requires, in


addition, the information that the object be rigid;

hence we can assign the selection restriction [Rigid] to

break. There is no reason for assigning the feature

[Rigid] to the lexical item dog. In Katz's theory break

195
the dog would receive no interpretation, and it would be

judged anomalous since there is no match between the

verbal selection restriction and the noun features. In

Weinreich's system, however, the feature [Rigid] would

be transferred to dog, and one of the semantic

implications of the sentence would be that the dog being

talked about was rigid. As Fillmore points out,

sentence (2) is acceptable and intelligible, but it

requires an interpretation in which the dog is either

frozen or is not a dog but rather some rigid physical

object in the shape of a dog. Fillmore refers to these

relations as certain specific understandings that verb

imposes on their associated noun phrases.

The transfer features of the Weinreich's

theory correspond to Chomsky's selectional features; the

difference lies in the fact that whereas Chomsky's

grammar merely ascertains whether the selectional

features of the verb correspond to the inherent features

of the nouns in its environment, Weinreich's theory

functions more activity - by transferring the feature

from the verb to the nouns.

Palmer (1876:97) observes three kinds of

collocational resrlotions. First, some are based wholly


on the meaning of the item as in the unlikely green cow.

Secondly some are based on range a word may be used with

a whole set of words that have some semantic features in

common. This accounts for the unlikeliness of the

196
rhododendron passed away and equally of the pretty boy

(Pretty being used with words denoting females).

Thirdly, some restrictions are collocational in the

strictest sense, involving neither meaning nor range, as

addled with eggs and brain and rancid with butter and
bacon.

Quirk et.al(1985) apply semantic restrictions

to the word classes. For example adjectives. The

oddity of * The music is too green is explained by the

requirement that green should be in a copular

relationship with a concrete noun (unless the noun is a

superordinate term as in The colour (I like best) is

green or unless green has the metaphorical sense


immature.

Leech (1981:138) refers to the earlier

attempts to account for selection restrictions dealt

with them as conditions for syntactic co-occurrence

(Chomsky 1965), or at least as co-occurrence conditions

defines on syntactic units such as nouns and verbs (Katz

and Fodor, 1963; Weinreich 1966). But, there are two

good reasons why selection restrictions are to be

defined not syntactically, but semantically, in terms of

units such as arguments and predicates:


1) If two words are synonymous, their selection

restrictions are the same. Thus assuming that frighten

and scare have the same sense (although they are


stylistically different), one and the same condition

197
explains why The idea frightned the girl and The idea

scared the girl are acceptable, while * the girl

frightned the idea and * The girl scared the idea are

equally unacceptable.

2) If two expressions are converses (e.g. own and

belong to), they have the same selection restrictions,

except that these restrictions apply in the reverse

order. It is the same restriction which explains why

The man owned a fortune and A fortune belonged to the

man both make good sense, while * A fortune owned the

man and * The man belonged to a fortune are both

nonsensical.

In both these cases, restrictions that would

require two separate statements on the syntactic level

can be brought together on the semantic level. So the

simpler and more explainatory account is that which

deals with selection restrictions semantically.

Lehrer (1974:180) quotes the problem

concerning selection restrictions raised by McCawley

(1970). McCawley argues that selection restrictions do


not have independent status in semantic analysis but are

predictable from the meanings of lexical items.

Otherwise, lexical items could have identical meanings,


but differ only with respect to selection restrictions.

One way of considering the collocations of a


word is as the lexical realization of the context of
situation. By looking at the linguistic contexts of

198
words we can often distinguish between different

meanings. Nida (1975:163) for instance, discusses the

use of chair in :

1) he sat in a chair

2) the baby's high chair

3) the chair of philosopy

4) has accepted a university chair

5) chairman of the meeting

6) will chair the meeting

7) the electric chair

8) condemned to the chair

These are clearly in pairs, giving four different

meanings of the word. But this does not so much

establish, as illustrate, difference of meaning.

Dictionaries, especially the larger ones, quite rightly

make considerable use of this kind of contextualization.

Thus, it may be said, though it has been

argued that all collocations are determined by the

meaning of the words, the linguistic context and


sometimes other contextual considerations, attribute

more specific meaning to words in particular

collocations.

4.3.3. Lexico-Semantic Position

The lexical and semantic positions on lexical

co-occurrence have been discussed. These are two

opposing hypotheses on the co-occurrence (or

collocation) of words. As Lehrer (1974:173) observes,

199
although the two positions conflict, they may both be

partially true ___ that is, each may be true for

different parts of the vocabulary and it is also

possible to hold a view somewhere between the two.

Lehrer (1974:183) argues that certain verbs

collocate with certain sets of nouns, although no

explaination is offered for this occurrence. For

example, bake but not roast would collocate not only

with fish . but with anything that was a fishpike,

pompano, cod, sole etc. Although there would not be

anything about the meaning of bake or roast that could

explain this collocation of bake and fish, at least

there would be some attempt to explain the collocation

of pike with bake but not with roast.

Another mixed position would be that although

the contraints on word combinations are determined by

their meaning, including selection restrictions, there

are still cases where certain combinations are

arbitrarily restricted. Thus, in our study selection

restrictions play a dominant part of the analysis of the

Adjective-Noun collocations. We can expect either that

the members of the field will have the same selection

restrictions and differ with respect to other features,

or we may expect them to have more or less the same

features and differ with respect to selection

restrictions.

200
4.3.4. Idiom Principle

Idioms have usually been defined as a set of

words whose meaning is not predicted from the parts.

Kick the bucket die and red herring something to draw

attention from an issue have meanings quite different

from those of the parts, and there is no motivation for

adding a meaning to red that appears only in the context

of herring and another meaning for herring that occurs

only in the context of red. This device would be

impossible in the case of kick the bucket because it is

difficult to see how to distribute the meaning of die in

such a way that some goes to kick and some to bucket.

Thus, the meaning of the resultant combination is

opaque. So Idioms involves collocation of a special

kind.

Idioms from the semantic point of view,

however, is not only the syntactic restrictions, not

even the frozen phrases. In an idiom it should be

impossible to substitute one term with a near synonym

and retain the meaning. For example, kick the pail can

not possibly mean die nor can scarlet herring mean

something to draw the attention from an issue. Thus,

idioms that are freest syntactically do not necessarily

allow lexical substitution or vice versa,

e.g kick the bucket


* sock the bucket

* kick the pail

201
what is and what is not an idjsm is a matter of degree.

A great deal of individual variation is to be expected

on the acceptability of the phrases,

e.g build castles in the air

? build castles in the clouds

? build castles in the sky

? build palaces in the air

* build barns in the air

Even if all those sentences that deviate from the idom

are unacceptable, judgements on the degree of

unacceptability are likely to differ somewhat, but

certainly some pharses with lexical substitution are

worse than others. The operating factor here is the

degree of semantic motivation in the idiom.

A very common type of idpm in English is what

is usually called the phrasal verb,, the combination of

verb plus adverb of the kind make up, give in, put down.

Cowie and Mackin (1985 vol. 1:1 xv) draw a

distinotion between sets of collocates to which other

words can be added at will, and sets which actually

exhaust all the possibilities of choice open to the

speaker. The difference between the first, or open kind

of set and the second, or restricted kind of course not


clear cut but a gradual progression. They give an

example of the first kind as follows :


Come in : .... become fashionable. Subjects:long hair,

whiskers:short skirts, full sleeves......

202
There are certain sets which represent a

severely limited range of choice :

e.g. keep an/one's eye on:.... observes - and if

necessary take appropriate action; watch over.

Adjectives (which collocate): careful, professional,

sharp, watchful....

Cowie and Mackin warn the user that the set is

highly restricted:other adjectives are unlikely to

collocate with the head phrase. This set, like any

other sets does not exhaust the speakers options:he


might, for special effect, choose vigilant or

amateurish. But these are unusual choices.

Sinclair (1987) advances two different

principles of interpretation to explain the way in which

meaning arises from language text. One is the open

choice principle and the other is the idom principle.

Sinclair considers the role of collocation

within the framework of -these two principles.

Collocation, according to him, illustrates the idiom

principle. He agrues that words appear to be chosen in

pairs or groups and these are not necessarily adjacent,

on some occasions.

Words do not occur at random in a text, and

that the open choice principle does not provide

substantial enough restraints. The principle of idiom

can be seen in the apparently simultaneous choice of two

words, e.g. of course. This pharse operates

203
effectively as a single word, and the word space, which

is structurally bogus, may disappear in time, as we see

in maybe, anyway and another. However, the principle of

idiom is far more pervasive and elusive. Thus, many uses

of words and phrases attract other words in strong

collocaion; for example hard work, hard fact, hard

evidence.

Sinclair illustrates collocational pattern

with the word back, considering upward and downward

collocations,

e.g. A = back

Node A B Node

downward upward

collocate B A collocate

Downward collocation gives us a semantic

analysis of a word. He concludes that early predictions

of lexical structure were suitably cautious; there was

no reason to believe that the patterns of lexis should

map on to semantic structures. For one thing, lexis was

syntagmaic and semantics was paradigmatic; for

another lexis was limited to evidence of physical co­

occurrence whereas semantics was intuitive and

associative.

Leech (1981:141) states, many metaphorical uses

of language involve a violaion of selection

204
restrictions. These restrictions are frequently

violated in poetry and in other imaginative uses of

languages. Anamoly is the term used for semantic

oddness caused by the violation selection restrictions.

Thus, Christopher is killing phoneme is anamalous

because the meanings of the predicates kill and phoneme

cannot be combined in this way.

4.4. Conclusions

Componential analysis has made considerable

contribution to the development of semantics. Leech

tried to integrate componential analysis into the system

of predication analysis. But componential analysis has

its limitations. Predominantly, it merely relies on

literal meaning of the words. The associative meaning

has got an important place in actual use of language.

Shastri (1978) observes that siok and ill have an

overlapping meaning. He finds that these words are

componentially synonymous. But he differentiates the

meaning of these words very clearly by obesrving the

collocations.

The present investigator proposes to study

collocations in Indian English in general and Addective-


Noun collocation in particular. Naturally the

COLLOCATIONAL PROCESSING is used . But componential


analysis and predication analysis is also used as the

approach is lexico-semantic. Nevertheless, Idiom

205
principle would be useful in establishing collocations.

The lexico-semantic analysis of the Adjective-Noun is

concerned with collocation.

It is further made clear that the collocation

patterns are investigated by using very large samples of

texts: The Kolhapur Corpus and LOB Corpus.

206

You might also like