1 -

The Language of Advertising: a Pragmatic Approach

Thesis submitted for the degree of

Doctor of Philosophy

in the University of London

(School of Oriental and African Studies)

by
Keiko Tanaka

1989
 ProQuest Number: 11010337

All rights reserved

INFORMATION TO ALL USERS

The quality of this reproduction is d e p e n d e n t upon the quality of the copy subm itted.

In the unlikely e v e n t that the a u thor did not send a c o m p le te m anuscript

and there are missing pages, these will be noted. Also, if m aterial had to be rem oved,
a n o te will ind ica te the deletion.

uest
ProQuest 11010337

Published by ProQuest LLC(2018). C opyright of the Dissertation is held by the Author.

All rights reserved.

This work is protected against unauthorized copying under Title 17, United States C o d e
M icroform Edition © ProQuest LLC.

ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 4 8 1 0 6 - 1346
•$>
 - 2 -

Abstract

The language used in advertising has been the subject of studies

in different diciplines, but surprisingly little has been done in

linguistics. The main purpose of my study is to give an adequate

analysis of the language of written advertising in the U.K. and in

Japan, within the framework of pragmatics, and to explain how

communication occurs between the advertiser and audience. I

consider what communication is and how it is achieved, and

investigate aspects of communication prominent in the language of

advertising.

The first chapter is a survey of the literature, covering

structuralist, semiotic and linguistic aproaches to the language of

advertising. The second chapter is a discussion of pragmatic

theories. It is argued that Relevance Theory (Sperber and Vilson

1986a) provides the best basis for explaining the comprehension of

utterances, including advertising. A study of puns forms the

subject of the third chapter. As a trigger for processing which

does not necessarily add to the informative content of an utterance,

puns provide a potential problem for Relevance Theory. But I argue

that Relevance Theory sheds light on the variety of ways in which

puns function in advertising. In the fourth chapter, I investigate

another potential problem for Relevance Theory, posed by the

language of advertising, that of partial suppression of the

speaker's intentions. The fifth chapter focuses on the projection

of the image of women in advertising.

 - 3 -

Table of contents

Page

Abstract

Acknowledgements

Chapter 1: Some problems of meaning in advertising

1.1. Analysis ol meaning in advertising

1.2. Some structuralist approaches

1.3. Some semi otic approaches

1.4. Some linguistic approaches

1.5. Conclusion

Chapter 2 - M eaning.and.. c q h u p u nidat ion

2.1. The basic concepts of communication 49

2.1.1. Grice's approach and its critique 50

2.1.2. Sperber and Wilson's approach 56

 .4

2.1.2.1. Manifestness and cognitive environment 59

2.1.2.2. Ostensive-inferential communication 62

2.2. Indirect communication 69

2.2.1. Grice's approach 69

2.2.2. Other pragmatists' approaches 74

2.2.3. Sperber and Wilson's approach 79

2.2.3.1. The deductive inference system 81

2.2.3.2. Relevance 85

2.2.3.3. How context is determined 93

2.2.3.4. Descriptive and interpretive use 101

Chapter 3: The pun in advertising

3.1. Introduction 106

3.2. The interpretation of the pun 111

3.3. Ambiguity and the pun 141

3.4. Puns and metaphors. 149

3.5. Conclusion 159

Chapter 4; Covert Communication

4.1. Introduction 163

4.2. Sperber and Wilson's ostensive communication 164

4.3. Non-ostensive communication inadvertising 167

4.4. Communication and ostension 186

4.5. Weak implicatures and non-ostensive communication 202

4.6. Covert communication 214

4.7. Conclusion 220

Chapter 5: Images of women in advertising

4.1. Introduction 222

4.2. Word meaning and concept 223

4.3. Intelligence 234

4.3. Individualism 246

4.5. Feminist 262

4.6. Conclusion 269

Epilogue 272

Notes 274

Bibliography 275
 - 6 -

Acknowledgements

My gratitude goes firstly to Professor Ruth Kempson, my

supervisor, for all her useful comments and advice. She kindly kept

me on as her supervisee and continued to help me for my last year,

in spite of the fact that she was on sabbatical. My special

gratitude goes to two other people who gave me strong academic and

moral support: Professor Brian Moeran and Dr. Deidre Wilson.

Before Professor Kempson took me on as a supervisee, Professor

Moeran was the only person in the world to find my work interesting

and worth-while, and his interest and support kept me going. I

sincerely hope that this thesis has not proven his judgement wrong.

Dr. Wilson gave me many helpful comments and much encouragement,

especially when I was in the later stages of writing, despite the

fact that she was on sabbatical for most of that period. Without

help and encouragement from these people, I could not have finished

my thesis, and I count it a great priviledge to have been working

with them. If I remain unenlightened, the fault is entirely mine.

I am also extremely grateful to my late grandmother, Hatsue

Akutagawa. Without her financial support, I could not even have

started a Ph.D. I deeply regret having been unable to finish my

thesis before her death in March 1988. In addition, I must express

my thanks to the Great Britain-Sasakawa Foundation for giving me a

grant, which helped me to pay the fees for my final term.

 - 7 -

Mr. Patrick Hanks taught me the joy of studying linguistics,

and indeed the joy of studying in general, which I failed to

experience throughout sixteen years of the Japanese educational

system. Had I not met him at University of Essex, where I did my

post-graduate diploma and M.A., I would not have experienced such

joy and would not even have dreamed of becoming an academic. I

stand in his debt.

My thanks go to Yoko Hino for supplying me with Japanese

magazines essential for my research, and to William Tanaka for

proof-reading my thesis, collecting examples of British

advertisements, and general encouragement.

Last but not least, my heart-felt gratitude goes to my family,

especially to my two mentors: my mother, Kazuko Tanaka, and my aunt,

Takako Takizawa, for their manifold support and encouragement. The

family who encouraged and helped me both financially and mentally

through my Ph.D. were all women - grandmother, mother and aunts - to

whom higher education was denied, whose husband and brothers went to

university, and who were forced to play a domestic role for lack of

opportunities. I am extremely grateful that they generously

supported me and gave me an opportunity which they never had.

 - 8 -

Chapter 1: Some Problems of Meaning in Advertising

1.1. Analysis of meaning in advertising

There has long been considerable interest in the means used by

advertisers to convey a desired message to their audience. In this

opening chapter, I wish to survey selected parts of this literature,

in order to assess how the problems raised by the language of

advertising have been seen and approached by other authors, and what

are the major strengths and weaknesses of these approaches. More

specifically, I am interested in analysing in this chapter how other

authors have dealt with the following basic questions: How is

communication achieved? In particular, how is advertising as a form

of communication carried out? How does the advertiser convey a

message to his audience by what he says?

In Section 2, I shall examine two books by structuralists, and

in Section 3, I shall consider works by two semioticians. Although

semiotics is a form of structuralism, it concentrates more narrowly

on systems of signs. Section 4 will look at various studies of the

language of advertising within the domain of linguistics.

It is my contention that the literature to date has not paid

enough attention to the context within which audience read and

process advertisements, and has relied too much on the supposed

structures of symbols in the text itself. This in turn explains my

own recourse to Relevance Theory as the most appropriate tool for

the study of the language of advertising, as I attempt to

demonstrate in the remaining chapters of the thesis.

1.2. Some structuralist approaches

Some attempts to analyse advertisements within the framework of

structuralism have tended to obscure the problems, rather than to

clarify them, notably the two books published by Leymore and Millun

in 1975. However, they have had considerable influence on thinking

about the language of advertising through their major claim that

representations in advertising are reducible to an underlying

structure. Leymore attempts to analyse advertisements in terms of

binary oppositions, and Millun in terms of classifications.

Leymore attemps to derive from advertisements 'apparent

characteristics' (1975:22), and from them what she calls 'exhaustive

common denominateors' (1975:43), which are both reduced to binary

oppositions and regarded as the structure of the system. For

example, in the analysis of the advertisements for butter, she

argues that their argument on margarine is based on the following

equations:
 - 1 0 -

butter : margarine - dear cheap

concord : protest

content : discontent

care negligence -

love hate

(Leymore 1975:43)

where *: ' means 'is to' and '-' means 'like'. The exhaustive common

denominators which she derives from binary oppositions above are:

peace : war - butter : margarine

(Leymore 1975:43)

Crucial to her analysis is an assumption that all relations in

a universe of discourse can be reduced to binary oppositions. She

argues that 'it is possible... to reduce all advertising phenomena

into a binary structure, which is at once exhaustive and complete'

(1975:127). Furthermore, she claims that a binary structure is

based upon psychological reality. She argues:

the essential activity of the human mind is

 - 1 1 -

classificatory and its most fundamental rules may be

reduced to those of contradictions and permutations -

in other words, to binary oppositions* ... the innate

structuring ability of the human mind can be

expressed in the form of binary oppositions and,

consequently, the binary oppositions with the

highest generalizing force are, in the last resort,

the invariants or the universals of the human mind.

(Leymore 1975:10-1)

The claim that the human mind categorises in terms of binary

oppositions is not new (Trubetzkoy 1939). However, the argument

against it is just as old. In fact, Trubetzkoy's original

phonological representation was in part based on gradual

oppositions, which, it has been argued, defy the notion of binary

opposition (Hyman 1975).

But what is the status of binary oppositions? It is unclear

whether they are physical properties, names of concepts or abstract

entities. Nor is it clear whether they belong to the conscious

experience of anyone, and whether they vary from individual to

individual or are shared by different persons. As for the claim

that they are 'exhaustive', firstly, it is not clear how she derives

these oppositions, and secondly, it is hard to see how any

opposition can be claimed as the 'correct' one. Vhere does

 « 32 -

atomisation stop? And how can correct oppositions be identified?

For example, why is the opposition to 'concord' 'protest', rather

than 'discord', or, why do we have 'love' and 'care', but not 'like'

or 'affection'? It is not plausible to argue that when we see

advertisements, we also see these oppositions.

Leymore does discuss two reservations which have been expressed

about the binary structure. One of them is that it is not clear

that all human thought is as a rule dichotomised in binary

oppositions. The other is that the structure is two-valued and does

not accommodate for gradable features. However, in support of the

binary structure, she argues that in all such cases, the sense of

opposition stems from the fact that the terms are discussed with

respect to some 'implicit norm' (1975:7). Vhat is not discussed is

what she means by 'implicit norm', a term which remains vague and

undefined.

Even if we accepted that there were such things as binary

oppositions in respect to some 'implicit norm', there is still doubt

as to whether binary oppositions could be applied to everything.

For example, let us consider colour terms. It could be argued that

black/white is a binary opposition, but what about red and brown?

It might be accepted that black is incompatible with white, red, and

brown, in the sense that there is no such colour as whitish black,

reddish black, or brownish black. But the colour term 'reddish

brown' is acceptable: red and brown can co-exist with each other.

How can we tell which of these colours is the 'norm'? Leymore does
 - 13 -

not provide any device to identify them. It is not clear if it is

of any use to classify everything in terms of binary oppositions.

Even if we accept that classificatory activity is the essential

activity of the human mind, it does not mean that its underlying

principles are reducible to binary oppositions.

There are two possible positive points to be retained from

Leymore's approach, although neither is without flaws. Firstly,

contrast can be expressed in an economical way through the use of

binary oppositions. It helps to draw attention to differences

between two items. But this should not make one forget the

importance of similarities. Secondly, Leymore's analysis might help

in testing the presence or absence of certain values in a society.

There might be some advantage in examining advertisements in terms

of whether certain values are manipulated. This said, values are

culture-dependent and it is not clear if there is any such thing as

'universal' (1975:11) values, as she claims.

Millun proposes to connect the role of advertising and the role

of women by analysing women's images in advertising in terms of a

system of classification, which he explains as follows:

...the categories that are in fact derived, stem

directly from what have already been seen to be

central concerns and should not be derived from some

theoretical superimposed structure. The categories

 - u -

and classifications should emerge from the material

rather than be imposed upon it. Each

classification should be the result of previous

thought and investigation, and tailored to bring out

those aspects felt to be the most significant.

(Millun 1975:25, author's italics)

Now, let us examine some of his classification systems. The

following are types of expression of men and women in advertising

proposed by him:

Types of men's expression Types of women's expression

thoughtful soft/introverted

self-reliant cool/level

seductive seductive

narcissistic

carefree carefree

kittenlike

paternal maternal

practical practical

comic comic

catalogue catalogue
 - 15 -

(Millun 1975:97-8)

If we compare the men's list to the women's list, it seems that

'thoughtful' and 'self-reliant* replace 'soft/introverted* and

'cool', respectively. One begins to wonder whether this is because

these categories simply 'emerged' from the material, or because

someone is called 'thoughtful' if he is a man but if she is a woman

she is described as 'introverted'. Furthermore, there seem to be

two types of expression missing from the men's list, which are on

the women's list, namely 'narcissistic' and 'kittenlike'. It is not

clear what is meant by 'kittenlike' and how it is different from,

for example, 'comic'. As to 'narcissistic', the reason why there is

no such type for men is surely not because men are incapable of

being 'narcissistic'; after all, the word comes from a male name.

There seem to be two possible explanations for this: one is that

when a man is 'narcissistic', he is described as something else,

'self-reliant', for example; the other is that men are

'narcissistic' all the time, anyway.

These classifications are not only culture-dependent. They

also depend upon the time, place and purpose or interest of the

research. They cannot be accepted as properties which are central

to the structure.

Millun calls such categories 'emerged from the material*

(1975:25) and assumes that they are in some sense directly

 - 16 -

observable. This is in effect an attempt at a thoroughly empiricist

position, which suffers from the central contradiction inherent in

an empiricist view of language. In so far as they are correct, these

categories are not independently identifiable but resolutely

cognitive. Thus, although he argues that the categories and

classifications should 'emerge from the material' rather than 'be

imposed upon it' (1975:25), they do not really 'emerge' from

advertisements: it is merely that one thinks that they do. In

effect, he admits this, for he adds that categories should be the

result of 'previous thought' (1975:25) and they are designed to

bring out aspects which are 'felt' (1975:25) to be significant.

The categories and classifications Millun is suggesting as

somehow value-free are in fact influenced by the time, place and

purpose of the research. Thus, in defining his method of analysis,

Millun starts by suggesting that one needs to note the presence or

absence of four elements: the product, the props, the setting and

the actors (1975:88), without any explanation or justification as to

why there are four elements and why those four. He finds it

'remarkable' (1975:136) that there should be a close relationship

between setting and actors. I would argue that settings and actors

are both part of his fifth element -'arrangement', and that it is

only natural that one should find that they are closely related.

Thus, Millun's classifications are marked by a high degree of

arbitrariness. On the one hand, they depend on the purposes,

interests and bias of the researcher. On the other hand, their

relationship to each other cannot be adequately explained.

 - 17 -

When Millun applies his system of classification to

advertisements, he is bound to find some correlation between certain

categories, by virtue of the nature of the categories. For example,

a 'sophisticated woman' tends to be found in a setting with 'non-

everyday moods', rather than a 'family setting with everyday moods',

partly because the actress is arranged to match the settings, or

vice versa, and partly because he calls a woman found in that kind

of setting sophisticated. His sets of categories describing men and

women's expressions are far from self-evident: they reveal a

particular view of men and women, upon which his analysis is based,

rather than images of men and women in advertising. In an attempt

to examine the ways in which women are presented in advertising,

Millun endorses stereotypical views of women, rather than

questioning them. His analysis is a categorisation of images of

women according to criteria which are accepted in the respective

society. In Chapter 5, I shall return to the analysis of images of

women in advertising, and attempt to show how it can be approached

by examining how certain words are used to describe women in

advertising.

1.3. Some semiotic approaches

If Leymore and Millun fail to provide an adequate explanation

as to how communication in advertising is achieved and how

 - 18 -

advertising messages are understood, a more convincing system of

meaning is presented by Williamson (1978).

Williamson starts off by analysing a Goodyear tyre

advertisement, which has an illustration showing a jetty with a car

on it. Apparently, the car is being tested for its braking

performance. She argues that, on the 'manifest* (1978:19) level,

the jetty signifies the test of braking power and connotes 'risk',

but on the 'latent' (1978:19) level, it signifies tyre because of

their similarity in appearance, and connotes 'safety'. She goes on

to argue:

...this transference of significance does not exist

as completed in the ad, but requires us to wake the

connection: it is nowhere stated that the tyre is as

strong as the jetty, therefore this meaning does not exist

until we complete the transference ourselves.

(Williamson 1978:19, author's italics)

Williamson nearly stumbles on the fact that advertising

messages are not fully encoded and that interpreting advertisements

takes more than just decoding; they need the audience to make

appropriate connections. All semiotic aproaches are based upon the

assumption that communication is achieved by simply encoding a

 - 19 -

message, and this assumption is precisely the defect they suffer

from. It is a defect shared by Barthes, whose approach is surveyed

below.

The problems of the semiotic approach have been argued by

Sperber and Wilson (1986a) at length. Let me highlight the points

they make which are valid in this discussion. They argue that a

generative grammar is a code; what is encoded is phonetic

representations of sentences and what is decoded is semantic

representations of sentences. However, a sentence is used in

different ways to convey different messages, as in (1):

(1) He is charming.

There are possible situations in which (1) is used to mean (2) or

(3):

(2) Paul is charming.

(3) Bob is charming.

Furthermore, sentence (1) can be used with a certain tone of voice

to mean (4) or (5):

 - 20 -

(4) Paul is not very pleasant.

(5) Bob is not very pleasant.

Thus, a sentence can be used with various linguistic and non-

linguistic properties, such as a tone of voice, who the speaker and

the hearer are, the time and the place of the occasion, and so on.

That is, the same sentence can be used as different utterances to

have different interpretations. The semantic representation of

sentence (1) can be decoded and be analysed in terms of a generative

grammar. However, a generative grammar cannot determine whether

'he' refers to Paul or Bob. Nor can it account for how (1) can be

interpreted as (4) or (5). It might provide the hearer of (1) with

the information that 'he' refers to a single male third person, or

that (1) suggests a certain attitude of the speaker, that is, that

he is being ironical, rather than sincere. However, the hearer is

still left with the gap between linguistic representations of the

sentence and the interpretations that the different utterances are

used to convey. In each situation, there are always other

possibilities, and the hearer has to select a referent or an

interpretation among others.

Thus, utterance interpretation cannot be fully accounted for in

terms of semantic rules as part of grammar. As we have seen in

examples (1) - (5), the semantic representation of the sentence does

 - 21 -

not provide sufficient information for the hearer to achieve a

complete interpretation of an utterance in context.

While suggesting that interpreting advertising messages is not

merely a matter of decoding, Williamson resolves this dilemma by

presenting a system of meaning. She argues that the 'transference'

is 'based on the fact that the first object (jetty) has a

significance to be transferred' (1978:19, author's italics):

...the advertisement does not create meaning

initially but invites us to make a transaction where

it is passed from one thing to another. A system of

meaning must already exist in which jetties are seen

as strong, and this system is exterior to the ad -

which simply refers to it, using one of its components as

a carrier of value (in the case of (the example above),

strength, durability) i.e. as a currency.

(Williamson 1978:19, author's italics)

She calls this mechanism 'referent system' (1978:19) and goes on to

define 'currency* as follows:

Currency is something which is representing a value

 - 22 -

and in its interchangeability with other things,

gives them their 'value', too.

(Williamson 1978:20)

Pateman (1983) rejects her arguments as unsound. For a

currency which allows the jetty-strength connection to exist, there

have to be an infinite number of such systems. Williamson herself

notes that the jetty represents risk as well as strength. She may

be right in saying that the jetty in the advertisement represents

both risk and strength. However, she does not explain how an

audience will know which system is valid, either the jetty-risk one

or the jetty-strength one, or yet another one. This is especially

problematic when, according to Williamson, the jetty-risk system is

used on the 'manifest' level and the jetty-stength one on the

'latent' level. How is an audience to know which system is to be

used when? She points out herself that there is nothing in the

advertisement written to the effect that the jetty stands for

something. How is an audience supposed to select these two valid

systems out of all the possibilities and use them at the right

level? What criteria does an audience have in choosing the correct

currency? These are the questions Williamson conveniently neglects.

And these are the questions for which an adequate theory of

communication must provide an answer.

 - 23 -

Williamson is aware that, for communication to succeed, the

audience must be involved in doing the work of processing and that

there is a gap between the message which is obtained by decoding and

the message the audience actually recovers. She argues that the gap

is filled by the audience's knowledge, but she goes on to assert

that this knowledge-base is itself rule-governed:

To fill in gaps we must know what to fill in, to

decipher and solve problems we must know the rules of

the game. Advertisement clearly produces knowledge

. . . but this knowledge is always produced from

something already known, that acts as a guarantee,

in its anteriority, for the ' truth' in the ad itself.

(Williamson 1978:99, author's italics)

According to Williamson, 'the assumption of pre-existing bodies

of knowledge' (1978:100) allows the reference system to work.

However, the question is: what criteria does the audience use in

order to choose relevant information out of the whole range of

knowledge they have? She adds that her concern is precisely which

reference to pre-existing bodies of knowledge is used in

advertising. It is worth investigating what ideologies advertisers

employ in their advertisements. From this perspective, her analysis

is stimulating and raises a central question. However, she is

 - 24 -

unable to answer the crucial question which she raises: how are

advertisements understood?

'Rhetoric of the image' by Barthes (1977) is claimed by Dyer

(1982:224) to be a major essay on semiotics and its application to

the analysis of an advertising message. Barthes studies the

interrelationships between the image and the advertising message,

using an advertisement for pasta called Panzani as an example. This

is perhaps the text which gets us closest to an answer to our

crucial question, although it still suffers from defects of a

structuralist interpretation.

Barthes argues that there are three kinds of message: the

linguistic message, the coded iconic message and the non-coded

iconic message. The linguistic message consists of the caption and

the labels. Barthes points out that there exist two levels of

interpretation of the linguistic message, namely, denotational and

connotational. In the example, the advertised product name denotes

the pasta, and connotates 'Italianicity'. Putting aside the

linguistic message, Barthes argues that we are left with the pure

image. This is divided into two categories, which are the iconic

equivalents of connotation and denotation. The former is also

categorised as a coded, symbolic, and cultural message, while the

latter is said to be the non-coded, perceptual and literal message.

However, the denotation-connotation distinction is not clear-

cut neither at the linguistic nor at the iconic level. Perceptual

 - 25

information is not independent of cultural knowledge. As Sperber

and Wilson (1986a:65) point out, a distinction between 'central'

thought processes and 'input', 'perceptual' or 'peripheral'

processes is assumed in current cognitive psychology: so-called

input systems are said to transform information from sensory

representations ihfe conceptual representations, all of which are in

the same format. The information provided by such input systems,

however, underdetermines the information derived from them. The

central process integrates information derived from the perceptual

system with information stored in memory to determine what is

actually perceived. Thus, even the processing of information by

perceptual organs is affected by one's cultural knowledge.

Barthes mentions 'anthropological' (1977:36) knowledge of a

tomato, a string-bag, a packet of pasta, and so on, but once a red

and round object is recognised as a 'tomato', rather than as an

'imperfect ball' or a 'red vegetable', perceptual knowledge is no

longer devoid of cultural knowledge about 'tomatoes'. Moreover, as

a string-bag and a packet of pasta are cultural objects, it can

hardly be seen as valid or useful to separate their 'perceptual'

values from their 'cultural' values.

Barthes argues that if all cultural signs are removed from the

image, one continues to 'read' (1977:35) the image, and that one

obtains the non-coded, perceptual message. This is a statement

which is hardly acceptable. Mot only can cultural knowledge not be

separated from perceptual information, but it is common-sensical to

 - 26 -

argue that it is perception which comes before knowledge: it is a

perceptU-.al stimulus which comes to one's cognitive system first.

This initial stimulus triggers the recuperation of stored

information and therefore calls upon knowledge, rather than

knowledge coming to one's mind first and giving way to the late

comer, namely, perception. One has to perceive a round and red

object first in order to identify it as a 'tomato', before gaining

messages such as ' Italianicity' , or freshness.

The analysis of the advertisement in question concludes that

there are four discontinuous connotational signs: (a) a suggestion

of a return from the market implies the freshness of the products

and domestic preparation; (b) the colours used in the poster mean

' Italianicity' ; (c) the serried collection of different objects

suggests the idea of a total culinary service; (d) the composition

of the objects implies the image of the still life painting.

Barthes emphasises the discontinuity of these signs (1977:34-5).

However, this 'discontinuity' becomes unconvincing if one looks

at these signs. How can the freshness of the products and the

domestic preparation be categorised as one message, while the idea

of a culinary service forms a distinctive category? The division

seems arbitrary, in much the same way as the classifications of

other authors discussed above. Moreover, it is possible for an

audience to put more effort into deciphering the advertisement and

to derive further messages, for example, the idea of harvest,

abundance, and so forth.

 - 27 -

Ve have considered in examples (1) to (5) that decoding is not

quite enough for communication. Then, the question to be asked

would be how the audience achieves an adequate understanding of an

advertisement, while there exist a potentially infinite number of

different messages carried by a finite number of signals in it.

Barthes seems to be aware of the problem, but he attempts to

resolve it by arguing that the linguistic message has a function

vis-a-vis the iconic message, a function which he calls 'anchorage'.

By this, he means the selection of the intended message:

... all images are polysemous . .. (which) poses a

question of meaning and this question always comes

through as a dysfunction even if this dysfunction is

'recuperated by society as a tragic ... or a poetic

game ... (The linguistic message) helps (an

audience) to choose the correct level of perception. ..

(Barthes 1977:39, author's italics)

It sounds convenient, but unfortunately it is not realistic.

It is a truism that linguistic message is also polysemous. However

much we would like them to be clear and straightforward, linguistic

codes are not devoid of ambiguity. Reference assignment,

 - 28 -

disambiguation and enrichment are only part of the normal process of

utterance interpretation (Sperber and Wilson 1986a:185), as in the

following example:

<6) It is strange.

(7) The food is hot.

(8) Come back early in the morning.

'It' in (6) would have to be assigned to an appropriate referent.

The word 'hot' can mean either 'having a high temperature' or

'spicy', and the audience of (7) would have to decide which the word

means in the context. 'Early' in (8) is vague and the understanding

of (8) would not be completed untill the audiance has enriched

information as to how early the speaker of (8) means.

If we were to accept that communication is a matter of

decoding, and if we wished to account for interpretation of <6) to

(8), for example, we would have to accept that there are rules of

pragmatic interpretation and add them onto rules of semantic

interpretation. However, it is not possible to write general rules

to account for most aspects of utterance interpretation, and it

would be necessary to have an infinite number of rules.

 - 29 -

Sperber and Wilson argue (1986a:27) that verbal communication

involves more than a single form of communication. It involves

linguistic encoding and decoding, but there is a gap between the

linguistically encoded message of a sentence and what the speaker

means, or what the hearer understands. They argue that verbal

communication also involves inferential processes. Their claim is

that in verbal communication, the code and inferential inodes of

communication are combined together and enable people to achieve a

more sophisticated level of communication which neither model can

provide on its own.

Let us now consider the inferential model of communication.

The input of a decoding process is a signal and the output is a

message, which is paired to the signal by an underlying code. The

input of an inferential process is a set of premises and the output

is a set of conclusions, which follow logically from the premises,

as in the following examples:

(9) Input: An amber light.

Output: Be cautious.

(10) Input: If it is a Sunday, the shop will be closed.

It is a Sunday.
 - 30 -

Output: The shop is closed.

However, the question is how the premises used in the process

are determined. The answer lies in finding premises which match the

intentions of the speaker and determining how the hearer can come to

recognise them. It was Grice (1957) who first raised the

significance of the publication and recognition of intentions in

communication, and various pragmatists have since developed

inferential models of communication. The inferential model is

formulated around the speaker's intention. It states that

communication is achieved by the speaker providing evidence of his

intentions and the hearer inferring his intentions from the evidence

(Sperber and Wilson 1986:24). 1

Another problematic feature of Barthes' account is his claim

that language is the prime example of a semiological system (Culler

1983:73), and that it is possible to talk of explicit, discontinuous

messages, as in the example of the four messages he claims to derive

from the pasta advertisement. Sperber and Wilson (1986a:66) argue

against this assumption, saying that the kind of explicit

communication that can be achieved by the use of language is not a

typical but a limiting case. For Sperber and Wilson, the goal of

communication is the improvement of the individual's knowledge of

the world, and the communicator does not convey a finite number of

explicit and precise messages. Instead, he provides his evidence

for a number of conclusions, which become more accessible to the

 - 31 -

audience to varying degrees as a result. While it is right to

argue that linguistic communication can achieve a degree of

precision and complexicity unattainable by non-linguistic means of

communication (Sperber and Wilson 1986a:174), linguistic

communication does not function by communicating a limited number of

explicit and strong messages.

Barthes' semiotic analysis also fails to provide an adequate

account of the degree of strength of a conveyed message, or of the

indeterminacy of meaning. It is questionable whether advertisers

regard the polysemous nature of both linguistic and non-linguistic

messages as the 'tragedy' which Barthes did (1977:39). On the

contrary, as I argue extensively in Chapter 3, the advantage of

punning, which is frequently used in advertising, lies in its

multiplicity of meanings, for it can be eye-catching and it can

sustain an audience's attention longer. Nor does it matter much

whether a particular message is conveyed or not, even if an

advertiser has a set of assumptions which he hopes his advertisement

might convey. As long as an advertisement catches an audience's

attention, sustains it for a time, creates some response and remains

in the audience's memory, the advertisement is said to be

successful. The recovery of a message other than that intended is

far from being regarded as redundant or as a failure.

Thus, Barthes' analysis is incapable of explaining the

following case. A Winston advertisement shows a wok pushed into a

black forest gateau with the following caption:

 - 32 -

(11) Ve're not allowed to tell you anything about

Winston cigarettes, so here's a wok in the Black

Forest.

The appreciation of this advertisement involves more than de­

coding the linguistic and iconic message in the advertisement. A

wok in the Black Forest is absurd and it is irrelevant to

cigarettes. And this is all intentional. The advertiser expects

this caption to attract the audience's attention and sustain it,

hopefully longer than usual, as the caption is quite absurd and

comes as a surprise. Possibly, the pun might attract extra

attention, too. The advertiser hopes that it is memorable; due to

its absurdity. Even better, the audience might like it, as it might

appeal to their sense of humour. Thus, the advertiser could

possibly create some favorable feelings among the audience.

1.4. Some linguistic approaches

The next work on advertising to be surveyed is an article

called 'How is understanding an advertisement possible?' by Pateman

(1983), which appeared in Language, Image, Kedia edited by Davis et

al. It is a criticism of Barthes and Williamson, and it favours a

pragmatic approach. Pateman argues that semioticians take for

 - 33 -

granted 'important conditions of possibility of the routine

accomplishment' (1983:187). By 'the routine accomplishment', he

means not only audience's linguistic knowledge, but also assumptions

about the communicator's intention, the principles of conversation,

activity type, point or purpose, and so on.

First, he introduces a theory of script-based understanding

(Schank and Abelson 1977); then he notes that advertisements are

'rarely identified in isolation and retrospectively but rather they

are identified in a context where they have been anticipated'

(1983:188, author's italics), and argues that unless an

advertisement is identified as an advertisement it would be

'strictly impossible for us to understand...it' (1983:189). His

argument is adequate to the extent that advertisements tend to

appear where they are anticipated, and knowledge that something is

an advertisement helps the audience to understand it.

However, Pateman's thesis is inadequate for a general

understanding of the language of advertising. Although it may be

rare, a text can be identified as an advertisement without any prior

knowledge that it is one. It does sometimes happen that one reads a

text, regarding it as an article and finds out as one reads it that

it is in fact an advertisement. The information that something is

an advertisement is not a prerequisite for understanding it.

Secondly, Pateman argues that knowledge about what kind of

thing fills a slot can be used in analysing a particular object

 - 34 -

which fills a slot. He borrows the notion of 'activity type' from

Levinson (1978), which is defined as follows:

I take the notion of an activity type to refer to a

fuzzy category whose focal-members are goal-defined,

socially constituted, bounded events with constraints

on participants, setting and so on, but above all on

the kinds of allowable contribution. Paradigm

examples would be teaching, a job interview, a jural

interrogation, a football game, a task in a workshop,

a dinner party, and so on.

(Levinson 1979:368)

Pateman adds advertising to this list. He argues that participants

in different activity types are prepared with minimal knowledge

about an activity in which they are engaged, which includes the

purpose of the activity. He goes on to say that in the case of

advertising the purpose is to sell products, and that without this

knowledge an advertisement is not understandable.

However, there are considerable problems with this formulation.

For example, the goals of a dinner party are by no means clear.

Being sociable and not offending anyone come to little more than a

vague and arbitrary set of 'dos and don'ts'. Nor does Levinson
 - 35 -

prove his point by referring to a court case in which an alleged

rape victim had just admitted that she had been sexually involved

with two men before; at this point the defence lawyer said to

her:'And you are seventeen and a half?', implying that a girl of

seventeen who has already slept with two men was not a woman of a

good repute (1978:380-1). This is not a convincing argument to

support the notion of activity type. For such a statement does not

have to come from a lawyer in a criminal court in order to suggest

this meaning. This example may suggest something about a lawyer's

tactics, but it does not explain what it is supposed to explain,

that is, how people understand what they understand.

Sperber and Wilson (1987:742) argue that, while recognising the

importance of goals, purposes, plans,' and so on, there is no

defining them, and that it is impossible to show how they are

selected or constructed, and, once selected, how they affect

comprehension. Their claim is that there is one sole criterion used

in comprehension which is called relevance, which is to be discussed

at length in Chapter 2, and they define relevance in a context and

to an individual. They go on to argue that given a definition of

relevance in a context, and the way the context is constructed,

assumptions about the goals and purposes of the participants in a

conversation form part of the context, and that therefore the notion

of relevance is not incompatible with our intuitions about goals and

purposes.
 - 36 -

Pateman's category of goal-defined activity is just as vacuous

when applied to advertising. The purpose of an advertisement is to

make the audience want to buy a certain product, when they would not

buy it but for the advertisement. If advertisers are to depend

entirely upon their audience's recognition of their goal, their

chances of success will not be high. Advertising cannot be seen as

goal-bound in a social sense, for the goal of an advertisement is

not shared by the audience. In effect, advertising is socially

'goal-divided' rather than 'goal defined', inasmuch as the different

parties involved are striving for different social goals. The

ultimate purpose of advertising is to cause a change in the thought

and behaviour of the audience against their will. The advertiser

wishes to change the minds of uninterested persons in his audience

and make them buy his product. How this can be achieved is the

topic of Chapter 4.

Thus, Pateman's argument against semiotic approaches in favour

of a pragmatic approach is sound, and his intuitions to support the

notion of goals and purposes are valid. However, his analysis

suffers from the defects of the theory it is based upon. A more

adequate pragmatic theory is needed in order to achieve his purposes

of analysing how advertisements are understood. I shall argue in

Chapter 2 that Relevance Theory provides just that.

Until now, we have been considering non-linguists' contributions

to the study of the language of advertising, even though people like

 - 37 -

Pateman use linguistic theory. It is now necessary to cover the

writings of linguists.

Leech (1963) is regarded as a 'classic' in this field: as Dyer

(1982:225) remarks, Leech provides a comprehensive study of

advertising style. However, his work is descriptive and stylistic,

and it is not an attempt to explain how advertising as communication

is achieved. It was written before work on pragmatics developed.

Since then, linguists turned their attention to the nature of

communication, whereas Leech merely provides an account of the style

and range of constructions used.

The object of Leech's work is to describe British advertising

language in a linguistic framework, but it is not of great interest

to my study because of its descriptive nature. For example, he

presents a list of the adjectives most frequently used in television

advertising, which includes 'new', 'good/better/best', 'free',

'fresh', 'delicious', and so on. However, it could be argued that

they are all predictable. Firstly, they fall into a category of

most frequently used adjectives in English in general. Secondly,

given that food and detergents are two of the product categories on

which most is spent by advertisers, it can be expected that a

considerable proportion of the materials studied was in these

categories, and it is thus not surprising that words such as 'fresh'

and 'delicious' should get onto the list. Moreover, advertisers

often use 'free' referring to incentives used in their promotion.

Thus, although his findings are interesting and useful from a

 - AA -

descriptive point of view, they do not offer much of a contribution

to understanding how communication takes place in advertising.

The latest relevant work is by two linguists, Vestergaard and

Schroder (1985). They argue that advertisements tend to take a

certain 'behavioural normalcy' for granted (Vestergaard and Schroder

1985:141). Although the notion of 'behavioural normalcy' needs more

explanation, what they mean is that advertisers take a certain

behaviour or attitude as the norm without explicitly saying so, as

in the following example, which is a caption for a Dr. White's Panty

Pads advertisement:

(12) Is there anything you can't wear?

(Vestergaard and Schroder 1985:143)

This advertisement is emphasising that the pads are thin and don't

show, and the question presumably is whether there is anything one

can't wear with the kind of pad in question. Vestergaard and

Schroder argue that since the question does not strike one as

meaningless, it must imply that with other towels there are things

you cannot wear. They have a point in that advertisements treat

certain assumptions as valid without explicitly stating them.

 - 39 -

However, this process is not exclusive to advertising, and it

is involved in a whole range of utterance interpretation, as will be

shown later. Where Vestaargard and Schroder go really wrong is when

they argue that this is a 'semantic process of ... imposition of a

behavioural normalcy'. It is not one's semantic knowledge about the

words in (12) which reveals an underlying assumption that there are

things one cannot wear with other pads. This problem cannnot be

treated at a semantic level: it is a pragmatic one.

The last work to be surveyed here is The Language of Television

Advertising by Geis (1982). The purpose of the study is to describe

how language is used in American television advertising, and how the

language of advertising is understood by audiences in terms of

pragmatic theory.

Geis argues that the advertiser should be held responsible for

non-idiosyncratic inferences drawn by an audience, as well as what

his advertisement asserts and what its assertions entail, as

ordinary people cannnot be expected to distinguish between valid and

invalid inferences (1982:33). He takes a Gricean approach and

argues that pragmatic theory must be dependent on semantic theory

(1982:34). His assumption is that the hearer draws inferences from

an utterance with the aid of his semantic knowledge and Grice's

cooperative principle:

A sentence S conversationally implies a proposition P

 - 40 -

in a given conversation if P can be "calculated"

given

a. the literal meaning of S,

b. general principles governing conversation,

c. the context of the conversation

d. background knowledge shared by speaker and hearer.

(Geis 1982:30)

Problems of Gricean analysis have been discussed by Sperber and

Wilson, and will be explained in Chapter 2. Here, it is sufficient

to point out that Geis has not suggested how 'the context of the

conversation' is determined. Moreover, the notion of mutual

knowledge has been discussed and dismissed as psychologically

unreal. (Clark and Marshall 1981, Sperber and Wilson 1986a) It

would not be possible to pin down an advertiser on these notions

which are impossible to establish. An advertiser could deny any

accusation by saying that he was not aware that his viewers shared

that particular knowledge. Geis argues that advertisers should be

held responsible for the conversational implicature of what they

say. While this is in itself a reasonable and interesting proposal,

Geis' actual analysis is not a succesful basis for implementing any

such restriction.

Geis bases his analysis upon Grice's notion of cooperation

between participants in a conversation, and presents six maxims,

 - 41 -

rather than Grice's four, incorporating maxims taken from Grice,

Boer and Lycan (1975), as well as his own. In effect, Grice's

Maxims of Quantity and Quality are each broken down into two maxims,

the Maxims of Strength and Parsimony, and the Maxims of Truth

and Evidence, respectively. They are defined as follows:

The Maxim of Strength: Say no less than is necessary.

The Maxim of Parsimony: Say no more than is necessary

The Maxim of Truth: Do not say what you believe to be

false.

The Maxim of Evidence: Do not say that for which you

adequate evidence.

The Maxim of Relevance: Be relevant.

The Maxim of Clarity: Avoid obscurity of expression.

(Geis 1982:31)

Sperber and Wilson (1981) have argued that Grice's maxims are

not all independently necessary and that they may be reduced to a

single principle, which is the principle of relevance. Geis himself

is aware of problems which arise from inter-relations among maxims.

He gives as an example an advertisement for Aftate, which is a

deodorant for feet:

 - 42 -

(13) Aftate for Athlete's foot, with a medication that

kills athlete's foot fungus on contact.

(Geis 1982:55)

Geis argues that it is implied that Aftate kills athlete's foot

fungus on contact, and that the Maxim of Relevance is responsible

for this implicature. However, taking into consideration the Maxim

of Strength, and the fact that advertisers make the strongest claim

they can in their advertisement, the fact that it is never actually

claimed that this medication kills athlete's foot fungus on contact

implies that Aftate may not in fact kill athlete's foot fungus on

contact. He adds:

In general, the Maxims of Strength and Relevance can

give rise to quite different implicatures ........ in

general consumers are much more likely to go with the

Maxim of Relevance than with the Maxim of Strength,

for in such cases the latter requires much more

sophisticated reasoning than does the former.

(Geis 1982:55-6)
 - 43 -

He is suggesting that there is a hierarchy among the maxims and

the Maxim of Relevance is superior to that of Strength. However,

advertisers cannot be held responsible for different implicatures

which are argued to be derived from different maxims. Geis argues

that 'in general' audience are 'much more likely to' go with the

Maxim of Relevance. But this cannot be used for determining what

advertisers are responsible for, as the claim is too weak and not

specific enough. Thus, again, Geis' approach suffers from the

general defects of Gricean pragmatics, that is vagueness and

arbitrariness. Sperber and Vi Ison offer an account which supersedes

Geis' considerations and which does not suffer from the problems

inherent in a Gricean approach,

Geis' major finding is how weak claims made by advertisers are,

as in the following example:

<14) ...a remarkable nasal spray that lasts and lasts

up to 12 continuous hours.

(Geis 1982:3)

(14) claims that the nasal spray in question remains effective for

twelve hours at most, which is weaker than just saying that it lasts

for twelve hours. Based on the assumption that advertisers will

make the strongest claims that they can possibly defend, Geis
 - 44 -

concludes (1982:4) that the advertiser of the nasal spray cannot

justify the stronger claim. He goes on to argue that judging from

the fact that weak claims as such are commonly used in advertising,

they must have a stronger impact on the audience than their literal

strength would indicate.

Geis presents scales to measure the strength of claims, for

example, strength of probability and modal verbs. He gives an

imaginary situation in which John Jones is known by 100 women and is

interested in knowing how many of those women like him, and presents

a list of sentences:

(15) a. Every one of the women like John Jones.

b. Most of the women like J.J,

c. Many of the women like J.J.

d. Some of the women like J.J.

e. Few of the women like J.J.

f. None of the women like J.J.

(Geis 1982:63)
 - 45 -

and he proposes a method of ranking relative strength as follows:

Given a scale S (say 10 point scale plus 0) and a

comparison class C formed with respect to lexical

element E, assign those members of C that have

absolute values to the appropriate points on S and

assign all other members of C to S, giving equal

space on S to each member.

(Geis 1982:64)

He argues that given the above method, sentences (15a)-(15e)

would be mapped into the following scale:

0 1 2 3 4 5 6____ 2 S 2___ 1SL

none few some many most all

(Geis 1982:65)

and he argues that 'Many X's are P' would be interpreted as claiming

that something between 50% and 75% of X have property P. (1982:65)

 - 46 -

This definition is too specific and unrealistic at the same

time. His argument sounds acceptable in this particular context.

However, it is possible to imagine situations in which this

measurement does not work, depending on the number of subjects

involved. The question is how far beyond the given context does

this work? It might be argued that it applies to the case of

ninety-nine women, instead of one hundred. Or it might apply to the

case of ninety-eight women, and so on. However, it would not apply

if the number of women is three: even if nearly 70% of the three

women liked John Jones, it could not be claimed that many women

liked John Jones. Where is the cut-off point, if there is any? The

scale pays no consideration to context. A scale that is valid in an

imaginary context but not in others is of little use to an analysis

of advertisements.

Thus Geis' analysis suffers in part from the general defect of

formal pragmatics. It attempts to explain utterance comprehension

by adding rules of pragmatic interpretation, which are unrealistic

and not comprehensive, and it does not account for the role of

context in utterance comprehension.

Geis considers problems in utterance interpretation and argues

that they should be dealt within the framework of pragmatics.

However, his analysis shares the problems of Gricean and formal

pragmatics, and remains too vague and arbitrary. Furthermore, his

proposed solutions, such as scales for probability terms and modal

verbs are inadequate, for he pays no attention to the context

 - 4*7 -

against which an utterance is processed. It is necessary to

investigate these problems in terms of a more adequate pragmatic

analysis, which provides an adequate account for the notion of

context.

1.5 Conclusion

The lack of any account of context has indeed been the

fundamental weakness of all the writings on the language of

advertising to date. They have failed to take into consideration

the context in which an advertisement is interpreted.

In this chapter, I have argued, against semiotic approaches,

that understanding advertisements is not merely a matter of

decoding, and that we must consider inferential processes which are

involved in utterance interpretation. I have further argued that

the interpretation of advertisements is best approached from a

pragmatic point of view, and that we need an account which

integrates the notion of context. However, existing pragmatic

approaches suffer from the general defects of Gricean pragmatics,

namely, vagueness and arbitrariness, or of formal pragmatics, that

is unrealistic rigidity. In short, existing approaches to

advertising share the defects of the theories of communication on

which they are based.

 - 48 -

In the next chapter, I shall argue that Relevance Theory by

Sperber and Wilson offers a principled account of how an utterance

is interpreted by the hearer against the context. It is a theory

which supersedes all the intuitions and partial approaches surveyed

in this chapter. Relevance Theory is flexible and comprehensive at

the same time, and it provides a better theory of communication than

any of those surveyed in this chapter, and therefore it offers an

effective framework for the interpretation of advertisements.

 - 49 -

Chapter 2: Meaning and Communication

2.1. The basic concept of communication

In Chapter 1, I have examined works by various social

scientists, including linguists, on the analysis of meaning in

advertising. I have argued that structuralist and semiotic

approaches are not very fruitful, and that problems of meaning

cannot be fully accounted for by semantic analysis. It would seem

that the topic is best approached from a pragmatic viewpoint. In

this chapter, I shall consider how some pragmatists have approached

the problems of utterance interpretation and how best we can analyse

the interpretation of advertisements.

Section 2.1 will examine Grice's analysis and an alternative

approach by Sperber and Wilson. In Section 2.1.1., I shall look at

Grice's approach to meaning and communication, and in Section

2.1.2., I shall consider Sperber and Wilson's criticism of Grice's

approach, Section 2.1.2. will be further broken down into smaller

sections, in order to study Sperber and Wilson's analysis

systematically. Section 2.1.2.1. will look at cognitive

environment. Section 2.1.2.2. will focus on the notion of

manifestness. Ostensive-inferential communication will be defined

and explained in Section 2.1.2.3. Section 2.2 will discuss indirect

communication and various approaches to analysing it.

 - 50 -

2.1.1. Grice's approach and its critique

Grice tries to draw a distinction between two kinds of meaning

as in the following examples:

<1) Those spots mean measles.

<2) Those three rings of the bell (of the bus) mean

that the bus is full.

(Grice 1957:377)

Grice calls the first type of meaning 'natural meaning', and

the second 'non-natural meaning'. Hepresents (3) as the

formulation for natural meaning and (4a) and (4b)as those for non­

natural meaning:

(3) A means to do so-and-so (by x).

(4) a. A means something by x.

b. A means by x that.
 - 51 -

(Grice 1957:378-9)

where A is a human agent.

The point is that, in regard to (1), there is no asking 'Who

means it?'. However, in the case of (2>, it is possible to ask 'Who

means it?' or 'What does he mean by it?'. The key difference is

that, in the case of non-natural meaning, there is some intended

cause of belief.

Grice proceeds to a consideration of different stimuli, which

he calls 'utterances', such as an artist drawing a picture, a

conductor ringing a bell, and a police officer waving to a driver,

all of which are non-linguistic acts, but arguably genuine acts of

communication. The word 'utterance' is confusing here, so following

Sperber and Wilson, I shall use the term 'stimulus' instead.

Grice argues that the recognition by the audience of the

intention behind a stimulus is crucial. For example, suppose that A

left B's handkerchief near the scene of a murder in order to induce

a detective to believe that B was the murderer. In this case, Grice

claims, one could not say that the handkerchief has a non-natural

meaning. Nor can one argue that A, by leaving the handkerchief,

meant as a non-natural meaning that B was the murderer, in that A

specifically does not intend his audience to know his intention.

 - 52 -

Another concern Grice has as to the definition of non-natural

meaning is the distinction between non-natural meaning and

'deliberately and openly letting someone know' or 'getting someone

to t h ink'. He explicates this distinction with the following

examples:

(5) a. Herod presents Salome with the head of St.

John the Baptist on a charger.

b. Feeling faint, a child lets its mother see

how pale it is (hoping that she may draw her

own conclusions and help).

c. A man leaves the china his daughter has

broken lying around for his wife to see.

(Grice 1957:382)

Grice argues that, although here we seem to have cases which

satisfy conditions so far given for non-natural meaning, yet (5a) to

(5c) are not cases of non-natural meaning. He contrasts 'telling*

with 'deliberately and openly letting someone know' or 'getting

someone to think' , and says that the latter two cases are not
 - 53 -

'telling' or non-natural meaning. The question is: is this

distinction valid?

Grice gives further examples:

(6) a. A shows Mr. X a photograph of Mr. Y displaying

undue familiarity to Mrs. X.

b. A draws a picture of Mr. Y behaving in this

manner and shows it to Mr. X.

(Grice 1957:382)

According to Grice, (6a) is not a case of non-natural meaning,

but <6b) is. The reasons are that in the case of (6a), Mr. X's

recognition of A's intention to make him believe that there is

something between Mr. Y and Mrs. X is irrelevant to the production

of this effect by the photograph. However, in the case of (6b), it

will make a difference to the effect of A's drawing of Mrs. X,

whether or not he takes A to be intending to inform him about Mrs.

X, and not to be just doodling or trying to produce a work of art.

It is worth noting that this distinction between photography

and drawing has been made by Barthes in his analysis of advertising.

He gives the following reasons to support this distinction: (a) to

 - 54 -

produce an object or a scene in a drawing requires a set of rule-

governed transpositions; (b) the operation of the drawing

immediately necessitates a certain division between the significant

and the insignificant; (c) drawing demands an apprenticeship

(1977:43).

However, all these points in fact apply to photography as much

as to drawing. It is not the case that the reproduction of an

object in a photograph is carried out without human operation; it

takes more than simply pressing a shutter. It includes some

manipulation, such as the intended selection of objects and

focusing. It is particularly interesting that Barthes draws this

distinction in his discussion of advertising. In advertising,

everything, whether a photograph or drawing, is carefully and

skilfully controlled.

Thus, the distinction between intended non-natural meaning and

intended meaning, as reflected in the distinction between

photography and drawing which Grice is trying to make is not as

clear-cut as he claims it to be. Even in the case of (6a), Mr. X's

recognition of A's intention to make him believe that there is

something between Mr. Y. and Mrs. X can be crucial. For example, if

A leaves the photograph in question lying around by accident and it

is found by Mr. X among other photographs of many people enjoying

themselves at a party, Mr. X may not suspect anything between his

wife and Mr. Y.

 - 55 -

As to (5a) - (5c), too, it is difficult to observe the

distinction between 'telling' and 'deliberately and openly letting

someone know' or 'getting someone to think'. The difficulty partly

6tems from the fact that Grice does not make clear what is

communicated by the respective communicators of (5a) - (5c). For

example, it might be accepted that in (5a) Herod was 'deliberately

and openly letting' Salome know that John the Baptist was dead.

However, one might be tempted to say that Herod was 'telling' that

he had kept his promise to her. It could be argued that (5a) - (5c)

are both 'telling' and 'deliberately and openly letting someone

know' a whole range of things.

Indeed, Sperber and Wilson argue (1986a:53) that there is no

clear-cut distinction between the two, and that there is a continuum

of cases between 'deliberately and openly letting someone know' or

'getting someone to think', where strong direct evidence is

provided, and 'telling', where all the evidence is indirect.

As to what is communicated, Grice argues (1957:386) that only

what may be called the primary intention of a communicator has a

bearing on the non-natural meaning of a stimulus. For example, he

argues that if a man intends to get a woman to do something by

giving her some information, it cannot be regarded as relevant to

the non-natural meaning of his stimulus to describe what he intends

her to do. His argument is valid in that the communicator's

informing her of something is an act of communication, but his

getting her to do something as a result is not. However, if the

 - 56 -

former is called the primary intention, can the latter be called the

secondary intention? Are there layers of intentions? Is there a

clear-cut boundary between them? These points are not elaborated by

Grice, but they are by Sperber and Wilson, who distinguish between

what they call informative and communicative intentions.

2.1.2. Sperber and Wilson's approach

Grice's analysis of non-natural meaning can be reformulated in

terms of three sub-intent ions on the part of speaker: the speaker

utters x, intending to inform the hearer of y; in so doing, the

speaker intends the hearer to recognise his intention to inform her

of y; and the speaker intends the hearer's recognition of his

intention to inform to play a part in her comprehension process.

Let us suppose that the speaker thought that he studied hard

the night before the utterance and wanted to communicate the fact by

saying (7):

(7) I studied hard last night.

 - 57 -

The speaker will be said to have succeeded in communication

once his intention to inform the hearer of the fact that he studied

hard the previous night is recognised by the hearer. There are

situations in which the hearer recognises that the speaker intends

to inform her of a fact, but nevertheless she does not believe it.

For example, she may have seen him at a party the night before, and

therefore she may not be convinced that he studied hard on the night

in question. Thus, it is possible that the speaker's intention to

inform is recognised without this intention being fulfilled. When

the speaker's intention to inform is recognised and yet fails to be

fulfilled, his intention of making his intention to inform a part of

the hearer's reasoning also fails. Then, only his intention of

making his intention to inform recognised will be fulfilled. Both

his intention to inform and his intention to make the hearer*

recognition of his intention to inform a part of her comprehension

process will not be fulfilled. However, even then, the speaker will

have succeeded in communication; what he will have failed to do is

to convince his hearer.

Sperber and Wilson argue that if intention to inform does not

need to be fulfilled for communication to succeed, as I have just

shown, this intention to inform cannot be described as an intention

to communicate. They therefore call this intention to inform the

'informative intention'. On the other hand, the speaker's intention

to make his intention to inform recognised by the hearer, according

to Sperber and Wilson, is the true intention to communicate, and

they call it the 'communicative intention'. By definition, when the

 - 58 -

speaker's intention to inform is not fulfilled, the speaker's

intention that the recognition by the hearer of his intention to

inform should play a part in her comprehension process is not

fulfilled either. Since an intention to inform need not be

fulfilled for the purposes of successful communication, an intention

that the hearer's recognition of the speaker's intention to inform

should function as part of her comprehension process cannot be

necessary either. Sperber and Wilson define these two layers of

intentions as follows:

Informative intention: to inform the audience of

something.

Communicative intention: to inform the audience of one's

informative intention.

(Sperber and Wilson 1986a:29)

As I have shown above, an informative intention can be

recognised without being fulfilled. Once an informative intention

gets recognised, the consequent communicative intention is

fulfilled. However, the reverse is not necessarily true. A

communicative intention can be fulfilled without the corresponding

informative intention being fulfilled: the hearer may recover the

information intended by the speaker without believing it.

 - 59 -

This distinction between informative and communicative

intentions will turn out to be of great significance in

characterising the covert transmission of information and attitudes

typical of many advertisements. It will be central to my discussion

of 'covert forms of communication' (Chapter 4), and it is important

to clarify here just what is at issue.

These complex intentions all depend on the notions of

manifestness, cognitive environment, ostention, and ostensive-

inferential communication. I shall therefore take these four

notions in turn.

2.1.2.1. Manifestness and cognitive environment

When an assumption is accessible to an individual, and there is

more evidence for it than against it, Sperber and Wilson call it

'manifest', and they define cognitive environment in terms of

manifestness as follows:

A fact is manifest to an individual at a given time if and

only if he is capable at that time of representing it

mentally and accepting its representation as true or

probably true.
 - 60 -

(Sperber and Wilson 1986a:39)

A cognitive environment of an individual is a set of facts

that are manifest to him.

(Sperber and Wilson 1986a:39)

The notion of manifestness can be extended from facts to all

assumptions. To be manifest is to be either perceptible or

inferable. Manifestness is also a matter of degree. An assumption

can be weakly or strongly manifest to a certain individual.

Sperber and Wilson argue that the communicator's informative

intention is better described as an intention to modify the

cognitive environment of the addressee, rather than as an intention

directly to modify the thoughts of the addressee. They add that the

actual cognitive effects of a modification of the cognitive

environment are only partly predictable.

A communicator may form a certain design on his audience.

However, to what extent he has control over his addressee is

questionable, and is certainly a matter of degree. He may not have

much control over her actual thoughts, but he may have more control

over her cognitive environment. It is plausible that he should aim

more at altering her cognitive environment than her actual thoughts.

 - 61 -

There are questions related to this, which have been posed

above: what is communicated?; what is communication? Communication,

according to Sperber and Wilson, is a modification of the audience's

cognitive environment. It is not meant to create a thought or a

belief in the audience. Nor is it meant to transfer a certain piece

of information to the audience's mind. The communicator

communicates something by making a set of assumptions accessible to

the audience, and providing evidence for them, i.e. making them

manifest, or more manifest. What is communicated is a set of

assumptions, rather than a specific piece of information. The

communicator intends to alter the audience's cognitive environment

by adding a set of assumptions to that environment or making them

more accessible.

Sperber and Wilson also suggest that communication is a matter

of degree. On the one hand, when the communicator intends to make a

particular assumption strongly manifest to the audience, then that

assumption is strongly communicated. On the other hand, when the

communicator's intention is marginally to increase the manifestness

of a wider range of assumptions, then these assumptions are weakly

communicated. There is no cut-off point between strongly

communicated assumptions and weakly communicated assumptions, nor

between what is communicated and what is not communicated. Rather,

there is a range of assumptions, some of them strongly communicated

and some of them weakly so.

 - 62 -

Suppose that I ran into a friend, who asked me what I was going

to do on the weekend, and I pointed to the books which I was

carrying. By doing so, I would intend to modify her cognitive

environment and to draw her attention to the books I was carrying.

The fact that I was carrying books would be made more manifest to

her, although she might have noticed them already. Among the things

I would be making manifest to her is the fact that I was behaving in

this particular way. Let us suppose that my friend concluded that

my behaviour was deliberate, that I was not just trying to balance

what I was carrying, and that I was making this gesture in order to

attract her attention to the books. Then my behaviour can be said

to have made manifest to her the fact that I intended to make some

particular assumptions manifest to her. It is behaviour which makes

manifest an intention to make something manifest, and it is called

by Sperber and Vilson 'ostensive behaviour', or 'ostention'

(1986a:49).

2.1.2.2. Ostensive-inferential communication

Sperber and Vilson argue that a certain type of human

intentional communication is a case of ostention. It is this

ostensive communication that they are interested in and that they

propose an account for.

 - 63 -

Ostention reveals two layers of information: first, there is

the information which has been pointed out; secondly, there is the

information that the first layer of information has been pointed out

intentionally. In the case above, my friend might or might not have

noticed that I was carrying books. My ostensive behaviour would be

drawing her attention to them in an obviously intentional way,

making her conclude that there was some relevant information to be

obtained from the fact that I was drawing her attention to the books

which I was carrying.

Sperber and Vilson argue that all human beings automatically

aim at the most efficient information processing possible, and that

this efficiency can be assessed by the standard of 'relevance'. The

notion of what is relevant will be explicated later. Here, it

suffices to note that when the processing of new information gives

rise to a multiplication effect, it is called 'relevant' (1986a:48).

A multiplication effect is defined as the process whereby when new

information is added to old information, they yield new information.

Ostensive behaviour provides evidence of one's thoughts. It

succeeds in doing so because it implies a guarantee of relevance.

Vhen I ostensively pointed out my books to my friend, my behaviour

gave her a guarantee of relevance, that there was something worth

her attention, and this made manifest the intention behind my

behaviour.
 - 64 -

Sperber and Vilson argue that the term 'communication' can

legitimately be applied to all cases of ostention. They also argue

that communication is inferential. Thus, they treat ostensive

communication, inferential communication, and ostensive-inferential

communication as being all the same thing.Ostensive-inferential

communication consists in making manifest to an audience one's

intention to make manifest a basic layer of information, and

therefore it can be described in terms of an informative and a

communicative intention.

Here is a reformulation of the notions of informative and

communicative intentions, taking into account the concepts of

raanifestness:

Informative intention: to make manifest or more manifest

to the audience a set of assumptions.

(Sperber and Vilson 1986a:58)

Communicative intention: to make it mutually manifest to

audience and communicator that the communicator has this

informative intention.

(Sperber and Vilson 1986a:61)

 - 65 -

It is worth noting that communicative intention is defined in

terms of mutual manifestness. The notion of mutual manifestness

needs explanation. A cognitive environment shared by two or more

people, to whom it is manifest that they share it, is called by

Sperber and Vilson a mutual cognitive environment. In a mutual

cognitive environment, for every manifest assumption, the fact that

it is manifest to the people who share the environment is also

manifest. Every manifest assumption in a mutual cognitive

environment is called mutually manifest. Mutual manifestness is

weaker than mutual knowledge. It does not suffer from the

psychological objections to mutual knowledge discussed by Clark and

Marshall (1981) and Sperber and Vilson (1986a). And yet, the notion

provides a sense of overtness to Sperber and Vilson*s notion of

ostensive communication. For, when the speaker makes an assumption

mutually manifest to his hearer and to himself, he is making

manifest to her not only the assumption itself, but also the fact

that it is manifest to him as well.

To Sperber and Vilson, communication must to some extent be

overt. However, one might ask: what difference, if any, does it

make whether the communicative intention is mutually manifest to the

communicator and audience or not? Can this be a criterion for

distinguishing one variety of communication from other forms of

information transmission? This is an issue to which I shall return

in Chapter 4.
 - 66 -

Sperber and Vilson argue that mere informing alters the

cognitive environment of the audience, but ostensive communication

alters the mutual cognitive environment of the audience and the

communicator. Their claim is that, although mutual manifestness may

not have significant cognitive importance, it is of crucial social

importance. For, a change in the mutual cognitive 'Environment of

the communicator and the audience is a change in their possibilities

of interaction. (1986a:61-2) Let us look at the following examples:

(8) Could you give me a hand with my luggage?

Suppose that Paul said (8) to Kay as he was carrying luggage.

Kay would recover that by saying (8), Paul was asking her to help

him with his luggage. She might decide to help him, and by doing

so, comply with his request. Or, she might tell him that she was in

a hurry, and by doing so, she might refuse to comply with his

request. Either way, by ostensively communicating (8), Paul will be

faced with Kay's response, which has social consequences: in the

former case, he would feel an obligation to her for complying with

his request; in the latter case, he would be disappointed by her

refusal. On the other hand, if Paul did not say (8) and did not

ostensively communicate his request, but communicated to Kay that he

was in need of help, for example by showing his struggle with the

luggage, it would be a different situation. Even if Kay voluntarily

offered to help him, Paul would not be under the same kind of
 - 67 -

obligation to her as when she responded to his request. Or, if she

ignored him, it would spare Paul the loss of face involved in her

open refusal to comply with his request. Either way, Paul would be

free of the social consequences which would be caused by his

ostensively communicating his request to her.

Thus, ostensive communication, that is communication by making

the speaker's informative intention mutually manifest to the hearer

and himself, involves some social consequences. By making his

informative intention mutually manifest, the speaker is

simultaneously making it mutually manifest that the fulfilment of

his informative intention is in the hands of the hearer. Whether or

not Kay offered to help Paul, her reaction would cause some

alteration in their relationship. Ostensive communication can cause

social consequences, which mere information transmission does not.

This social aspect of ostensive communication is significant in

the study of advertising. The distinction between informative and

communicative intentions plays an important role in characterising

the covert form of informative transmission and attitudes

characteristic of much advertising, which will be the subject of

Chapter 4.

Finally, Sperber and Vilson define ostensive-inferential

communication as follows:
 - 68 -

Ostensive-inferential communication: the communicator

produces a stimulus which makes it mutually manifest to

communicator and audience that the communicator intends,

by means of this stimulus, to make manifest or more

manifest to the audience a set of assumptions <I).

(Sperber and Wilson 1986a:63)

 - 69 -

2.2 Indirect communication

I have so far addressed the problem of communication by

considering instances of information directly communicated.

However, a major problem in pragmatics is the phenomenon of indirect

communication, which Grice calls conversational implicatures.

In Section 2.2.1., I shall first examine Grice's approach to

indirect communication. In Section 2.2.2., I shall consider

accounts proposed by other pragmatists. Section 2.2.3 will consider

an alternative approach by Sperber and Vilson.

2.2.1. Grice's approach

In his 'Logic and conversation' (1975), Grice gives an example

of two people,A and B, having a conversation about a mutual friend

C. A asks B how C is getting onin his new job, and B replies, 'Oh

quite well, I think; he likes his colleagues, and he hasn't been to

prison y e t . ' Grice explains, 'I think it is clear that whatever B

implied, suggested, meant, e t c . , in this example, is distinct from

what B said...' (1975:43)

 - 70 -

This example is an attempt to draw attention to those aspects

of meaning which are not semantically determined, that is, so-called

'conversational implicature' (1975:45), taken to be opposed to

'conventional implicature', which is semantically determined.

Conversational implicature is worked out from the meaning of

the sentence uttered, together with the context, on the basis of the

assumption that communication is governed by the principle of

cooperation. More specifically, the assumption is that the speaker

has observed certain general principles of communication. For

Grice, the crucial distinction between conventional implicature and

conversational implicature is that the latter is calculable:

The presence of a conversational implicature must

be capable of being worked out; for even if it can in

fact be intuitively grasped, unless the intuition is

replaceable by an argument, the implicature (if

present at all) will not count as a conversational

implicature; it will be a conventional implicature.

(Grice 1975:50)

Grice then proposes a formulation of how this calculation might

go, as follows:
 - 71 -

(9) a. He has said that p.

b. There is no reason to suppose that he is not

observing the maxims.

c. He could not be doing this unless he thought that

q-

d. He knows (and knows that I know that he knows)

that I can see that the supposition that he thinks

that q is required.

e. He had done nothing to stop me thinking that q.

f. He intends me to think, or is at least willing to

allow me to think, that q.

g. And so, he has implicated that q.

(Grice 1975:50)

Sperber and Vilson argue that this is not a logical argument;

it is not even clear which of (9a) to (9g) are meant to be premises

and which conclusions. In (9c), the content of the implicature is

introduced for the first time, but this is by no means deducible

from (9a) and (9b). (9c) has to be either an independent premise

 72

itself, or derivable from (9a) and (9b) with some supplementary

premises, which remain to be specified. Thus, (9) does not offer an

adequate account of the working out of conversational implicatures.

This is also the case in the following example:

(10) She: Would you go and get some flowers?

He: I'm just going to Covent Garden.

How can the conversational implicatures of his utterance be

recovered? This is the kind of question a pragmatist aims to

answer. As it has been argued, they are not recoverable from his

utterance and Grice's general communication principles alone.

If it was used in a context containing assumption (11), then

his utterance would be treated as conveying proposition (12):

(11) They sell flowers in Covent Garden.

(12) He will be able to get some flowers.

 73 -

On the other hand, if it was used in a context containing assumption

(13), then his utterance would be interpreted as communication (14):

(13) They do not sell flowers in Covent Garden.

(14) He will not get any flowers.

Grice presents a list of sources of data upon which the hearer

will rely to work out conversational implicatures as follows:

(15) (i) the conventional meaning of the word used,

together with the identity of any

references that may be involved;

(ii) the Cooperative Principle and its maxims;

(iii) the context, linguistic or otherwise, of

the utterance;

(iv) other items of background knowledge;

(v) the fact (or supposed fact) that all

relevant items falling under the previous

headings are available to both

 - 74 -

participants and both participants know or

assume this to be the case.

(Grice 1975:50)

(11) and (13) are alternative contextual assumptions, which,

together with other assumptions, permit the deduction of conclusions

(12) and (14), respectively.

However, there is nothing in Grice's account to determine where

the required premises come from, or exactly how they interact to

determine the full import of indirect reply (12). In the next

section, we shall examine what other pragmatists have to offer as to

these questions.

2.2.2. Other pragmatists' approaches

Many pragmatists have no clear theory of non-demonstrative

reasoning, except a shared conviction that non-demonstrative

reasoning cannot even contain a deduction as one of its sub-parts.

As Sperber and Wilson point out, other pragmatists do not offer a

positive alternative. Take Bach and Harnish, for example:

 - 75 -

Our empirical thinking in general is rife with

generalizations and inference principles that we are

not conscious of when we use them, if we are

conscious of them at all. It would take us well

beyond present-day cognitive psychology to speculate

on the details of any of this. Whatever these

processes are, whatever activates them whatever

principles or strategies are involved, they work, and

work well.

(Bach and Harnish 1979:93)

Bach and Harnish argue that the form of inference by which

implicatures are recovered 'is not deductive but what might be

called an inference to a plausible explanation' (1979:92-3).

Brown and Yule take a similar line:

It may be the case that we are capable of deriving a

specific conclusion ... from specific premises . . .

via deductive inference, but we are rarely asked to

do so in the everyday discourse we encounter .... We

are more likely to operate with a rather loose form

of inferencing...
 - 76 -

(Brown and Yule 1983:33-4)

Brown and Yule present the following as a counter-example to

deductive reasoning in comprehension:

(16) ... in the kitchen there was a huge dresser and

when anyone went in . . . the hats and coats were

all dumped on this dresser

(Brown and Yule 1983:34)

They argue that deductive reasoning could not lead the audience

to infer that the hats and coats mentioned in (16) belong to

visitors of the house in which there is a dresser in the kitchen.

They are right that implicated premises are not deduced - only

implicated conclusions are deduced, as Sperber and Vilson make

clear. Implicated premises are retrieved or derived from memory,

and confirmed via consideration of relevance.

Leech claims that all implicatures are 'probabilistic*

(1983:30) and the process by which they are recovered is 'not a

formalised deductive logic, but an informal rational problem-solving

strategy' (1983:31).
 - 77 -

Levinson argues that implicatures are 'quite unlike logical

inferences, and cannot directly be modelled in terms of some

semantic relation like entailment' (1983:116). He uses the

following example, in which there is an assumption that in order to

get the lavish subsidy under some scheme one must have three cows.

The inspector asks John's neighbour the following question:

(17) a. I: Has John really got the requisite number of

cows?

b. N: Oh sure, he's got three cows all right.

(Levinson 1983:116)

Levinson points out that the neighbour's reply does not commit

him to the implicature ordinarily associated with (18a), namely,

(18b), by Grice's maxim of quantity:

(18) a. John has three cows.

b. John has only three cows and no more.

 - 78 -

The survey of studies on inference system by other pragmatists

has been presented by Blakemore (1987). She points out that in the

background to Levinson's rejection of deduction there is an

underlying distinction betwen context-dependent inferences and

context-independent inferences. She further argues that inferences

which hinge on the meanings of the words are deductive, while

inferences which depend on contextual information used as premises

are inductive, but 'inductive' inference of this type can be reduced

to deductive inference, as Sperber and Vilson show.

Sperber and Vilson disagree with Levinson as to where the

premises which are used in the recovery of implicatures come from,

and argue that this pragmatist follows directly Grice's

characterisation of implicature, with all its associated vagueness

and defects.

Let us look at example (10) to (14) again. Neither (13) nor

(14) are deducible from the content of (10) alone. Moreover, if

they were deducible from (10b) they would not be implicatures in

Grice's sense, for, according to him, 'the truth of a conversational

implicature is not required by the truth of what is said' (1975:58).

If this is a defining feature of implicature, no implicature will be

deducible from the explicit content of an utterance alone. However,

to say that (12) and (14) are not directly deducible from (10b) is

not to say that deductive reasoning does not play a significant role

in their derivation.
 - 79 -

2.2.3. Sperber and Wilson's approach

How, then, can the process of recovering implicatures be

accounted for? Sperber and Vilson argue that the principle

underlying the recovery of implicature is identical to the principle

of recovering the speaker's intention.

The process of inferential comprehension is a combination of

hypothesis formation and logical deduction. In principle, it is not

fully determinable. Even under the best of circumstances,

communication may fail. The speaker's informative intention may

neither be decoded nor deduced. All the hearer can do is to

formulate a hypothesis based on the speaker's ostensive stimuli.

Her hypothesis may later be confirmed through the following

exchange. However, there does not exist such a thing as a proof of

the speaker's informative intention.

Contrary to all other pragmatists, Sperber and Vilson argue

that such deduction processes, as shown by examples (10) to (14),

play a central role in the recovery of implicatures. Indeed, this

follows directly from their theory of cognition, which is an

expansion of Fodor's theory of the mind (1983). Their claim is that

in processing information, people aim to bring about the greatest

improvement of their overall representation of the world for the

least cost in processing.

 - 80 -

A non-demonstrative inference process has free access to

conceptual memory. They argue that it is a 'central thought

process' (1986a:66), which involves forming a hypothesis on the

basis of the input delivered from various perceptual and linguistic

systems and confirming it against background assumptions stored in

memory.

This distinction between 'central' processes and 'input',

'perceptual' or 'peripheral' processes is attributed to Fodor <1983

and 1985). Fodor argues that the central systems are beyond our

investigation for two reasons: (a) The function of the central

system is the fixation of beliefs, about which little is known; <b)

The range of facts or beliefs which are involved in the fixation of

further beliefs is unlimited. He compares the psychology of central

processes with the philosophy of confirming scientific hypotheses.

Sperber and Wilson reject this comparison. They see the

central thought process of utterance interpretation as more typical

of central cognitive processes than is the forming and confirming of

scientific hypotheses, which operates on a different time-scale.

The latter requires an enormous amount of time and effort, while

ordinary utterance comprehension is instantaneous. Very little time

and effort is spent on interpreting an utterance, for example (10b).

Sperber and Vilson also argue that, although the central

processes are not fully determinable in principle, as argued by

Fodor, at least some of them are highly constrained in practice.

 - 61 -

Their claim is that the goal of cognitive activities is the

maximisation of relevance, that is, to get the greatest contextual

effects with the least processing effort, and that this counts as a

single criterion in the process of utterance interpretation.

Thus, Sperber and Vilson argue that conversational implicatures

can be characterised as a partly deductive process. I shall examine

their analysis systematically. In Section 2.2.3.1, I shall look at

the deductive inference system. Section 2.2.3.2 will consider the

notion of relevance. The determination of context will be the

subject of Section 2.2.3.3. In Section 2.2.3.4., I shall discuss

the notions of descriptive use and interpretive use.

2.2.3.1. The deductive inference system

A deductive inference is a formal operation which takes

propositions as premises and yields propositions as conclusions.

Given a set of deductive rules, and given a set of premises, the

deductive conclusions do not vary. A set of conclusions are

automatically generated. They need not therefore be stored

separately. Thus, a deductive system would provide a significant

economy of storage.
 - 82 -

According to Sperber and Vilson, propositional representations

are stored as factual descriptions of the world, which are called

•factual assumptions' (1986a:74). They are entertained with greater

or lesser confidence, depending on the form of their acquisition.

They argue that a conclusion derived via deduction inherits its

strength from the strength of the premises used in the derivation

processes.

A deductive system would provide not only a tool for the

working out of new assumptions and their addition to an existing

system, but also a device to monitor the accuracy of any conclusions

derived by the system. Since the strength of these conclusions

depends on the strength of the premises used in their derivation,

inference rules can be used to assess the accuracy of the

assumptions entertained by testing them against one another.

Sperber and Wilson's hypothesis is that the human deductive

device yields only non-trivial conclusions, which they define as

follows:

Non-trivial logical implication

A set of assumptions (P) logically and non-trivially

implies an assumption Q if and only if, when (P) is

the set of initial theses in a derivation involving

only elimination rules, Q belongs to the set of

final theses.
 - S3 -

(Sperber and Vilson 1986a:97)

Sperber and Vilson call those rules which take only one

proposition as input 'analytic rules', and the implications which

are derived entirely by such rules 'analytic implications'. Rules

which take two separate propositions as input are called 'synthetic

rules', and any implication which is not analytic is a 'synthetic

implication' (1986a:104).

The distinction between analytic implications and synthetic

ones is crucial. The analytic implications of a given assumption

are intrinsic to it, and they are recoverable as long as the

assumption itself is recoverable, simply by reprocessing it through

the deductive device. On the other hand, synthetic implications are

not intrinsic to any single member of the set of the assumptions

from which they are derived. If humans are interested in improving

their overall representation of the world, they must therefore be

interested in recovering as many synthetic implications as possible

from any set of assumptions they are currently processing, before

putting them away into separate storage. Analytic implications, by

contrast, are only worth recovering as a way to recovering further

synthetic implications.

Assumptions which are taken into the memory of the deductive

device may come from the following four sources: (a) perception; (b)

linguistic decoding; (c) encyclopaedic memory; and (d) the result of

 - 8+ -

a deductive process. Here, we are concerned with the effect of

newly presented information, in particular of assumptions derived

via the linguistic input system, on old information which is drawn

from an existing representation of the world. In other words, we

are concerned with the effect of deductions, in which the set of

initial theses placed in the memory of the deductive device can be

partitioned into two subsets, new information and old information.

Sperber and Vilson call this effect 'contextual implication', which

is defined as follows:

Contextual implication

A set of assumptions (P> contextually implies an

assumption Q in the context <C> if and only if;

(i) the union of P and <C> non-trivially implies Q,

(ii) (P> does not non-trivially imply Q, and

(iii) (C) does not non-trivially imply Q.

(Sperber and Vilson 1986a:107-8)

A contextual implication is new information: it cannot be derived

from <P> or <C) alone. It is derivable only from the union of old

and new information.

A central function of the deductive device is thus to derive

instantaneously and automatically, and therefore unconsciously, the

contextual implications of newly presented information in a context

of old information.

2.2.3.2. Relevance

To modify and improve a context is to have some effect on that

context. Sperber and Vilson call such an effect a 'contextual

effect' (1986a:108). An utterance can have contextual effects in a

context in one of three ways. The first one is new information,

together with old assumptions, yielding information which is not

derivable from the new information or the old assumptions alone.

The second one is new information strengthening old assumptions.

And the last one is new information contradicting, and therefore

weakening or leading to the abandonment of old assumptions.

Assumptions placed in the memory of the deductive device come with

varying degrees of strength, and hence a deduction may result in

contradicting weak assumptions, which are then abandoned.

It can be recalled that our basic assumption has been that all

humans automatically aim at the most efficient information

processing possible, and that there is a single criterion for

efficiency, which is 'relevance'. 'Relevance' here does not mean

exactly the same thing as the English word 'relevance'. It is a

universal concept, and natural languages need not lexicalise it.

 - &6 -

However, Sperber and Vilson argue that scientific psychology needs a

concept which is close enough to the ordinary language notion of

’relevance'. They also argue (1986a:119) that there is an important

psychological property which the ordinary notion of 'relevance'

roughly approximates, and hence it is appropriate to use the term in

a related, more technical sense.

The notion of contextual effect helps to describe two essential

properties of utterance comprehension: first, comprehension involves

the joint processing of a set of assumptions; and second, in that

set some assumptions stand out as newly presented information which

are being processed in the context of old information.

Sperber and Vilson define 'relevance' as follows:

Relevance

An assumption is relevant in a context if and only if

it has some contextual effect in that context.

(Sperber and Vilson 1986a:122)

This definition satisfies the intuition that the notion of

relevance must involve the context in which an assumption is

relevant.
 - 87 -

It is worth emphasising that Sperber and Vilson define

relevance purely in terms of the contextual effect. There is an

intuitive question as to whether humans process a stimulus, striving

for only more contextual effects. The reward humans aim at when

they spend time and energy in processing a stimulus might include

other kinds of advantages which are not analysable in terms of

contextual effects. This is the question I shall come back to in my

discussion of puns and covert forms of information transmission in

Chapter 3 and 4 respectively.

The definition of relevance is insufficient without mentioning

that relevance is a matter of degree. Intuitively, it is clear that

the greater the contextual effects of a newly presented item of

information, the more relevant it is. But this is not yet

satisfactory. For, it might mean that, to this end, humans would

continue to process a newly conceived phenomenon, combining it with

an endless stock of information, in an attempt to see if it improves

their representation of the world. This is where the concern over

processing effort is so important. Sperber and Vilson define

relevance in terms of contextual effect in combination with

processing effort:

Relevance of a phenomenon

Extent condition 1: a phenomenon is relevant to an

individual to the extent that the contextual effects

achieved when it is optimally processed are large.

 - 66 -

Extent condition 2: a phenomenon is relevant to an

individual to the extent that the effort required to

process it optimally is small.

(Sperber and Vilson 1986a:153)

Thus, other things being equal, the greater contextual effects the

greater the relevance, and the smaller processing effort the greater

the relevance. Let us consider a context which contains assumption

(19):

(19) If Tom is going to the party, I will go, too.

Suppose that in this context, (20) was posed:

(20) Is Tom going to the party?

Now, let us assess the effects (21a) - (21c) would bring:

(21) a. Tom is going to the party.

 - 89 -

b. John is going to the party.

c. Tom is going to the party, and John is going to

the party too.

(21a) and <21b> have the same syntactic structure, and

intuitively they would require approximately the same processing

effort. However, in the context, (21b) would not have contextual

effects which (21a) would carry, in other words, (21b) would not

bear the relevance (21a) would have. Thus, in the context, (21a)

would be more relevant than (21c). On the other hand, (21c) would

yield the same relevance which (21a) has, but it clearly requires

more processing effort than (21a). Hence, (21a) would be more

relevant than (21c).

Earlier, we defined communication in terms of the

communicator's ostensive behaviour. In order to succeed in

communication, the communicator must attract the audience's

attention. An act of ostension can be described as a request for

attention. Now, someone who asks you to behave in a certain way,

either physically or cognitively, suggests that there is a good

reason to assume that you might benefit from complying with his

request. This suggestion may be made in bad faith, but it cannot be

wholly cancelled. If a request is made, the requester must have

assumed that the requestee would have some motive for complying with

it .
 - 90 -

There is a significant difference between being exposed to an

ostensive stimulus directed at oneself and being exposed to another

kind of stimulus. For example, there is considerable difference

between someone shouting at you, 'Vatch out!' and you overhearing

someone reading out a passage including such an utterance. With an

ostensive stimulus, the addressee can not only hope that the

stimulus may bear some relevant information for herself, but also

has a precise expectation that the stimulus is intended to be

relevant to her.

Thus, Sperber and Vilson argue that in ostensive-inferential

communication, the communicator necessarily communicates that the

stimulus he uses is relevant to the audience. In other words, an

act of ostensive communication automatically communicates what they

call a 'presumption of relevance' (1986a: 158). By this they mean

that what is communicated, to the best of the communicator's

knowledge, is that the ostensive stimulus is relevant enough to be

worth the addressee's attention.

As we have seen, the relevance of a stimulus is determined by

two factors, namely, contextual effects and processing effort.

Sperber and Vilson argue that, on the effect side, the presumption

is that the level of effects achievable is never less than what is

needed to make the stimulus worth processing. On the effort side,

the level of effort required is never more than what is needed to

achieve these effects. On the effect side, what is communicated

must be relevant enough to make the stimulus worth processing, hence

 - 91 -

the presumption is a matter of adequacy. It is worth emphasising

here that the presumption of relevance is not a presumption of

maximal relevance. On the processing effort side, unless the

communicator chooses the most relevant stimulus to make his

informative intention mutually manifest, he may be confronted by the

addressee's refusal to make the required processing effort. It

would be in the interest of both the communicator and the addressee

that the stimulus chosen requires the least processing effort to

achieve the intended effects. Thus, on the processing effort side,

the presumption is one of more than adequacy.

This heavy emphasis on minimising processing cost will face an

apparent counter-example in punning. I shall discuss this point in

the analysis of puns in Chapter 3.

The level of relevance that will be presumed to exist takes

into account the interest of both the communicator and the

addressee. Such a level is called a level of 'optimal relevance'

(1986a:158). Accordingly, the 'presumption of optimal relevance' is

defined as follows:

Presumption of optimal relevance

(a) The set of assumptions <11 which the communicator

, intends to make manifest to the addressee is

relevant enough to make it worth the addressee's

while to process the ostensive stimulus.

 - 92 -

<b> The ostensive stimulus is the most relevant one

the communicator could have used to communicate

<I> .

(Sperber and Vilson 1986a:158)

And the principle of relevance as follows:

Principle of relevance

Every act of ostensive communication communicates the

presumption of its own optimal relevance.

(Sperber and Vilson 1986a:158)

Now, the fact that an ostensive stimulus carries a guarantee of

optimal relevance does not necessarily mean that it is optimally

relevant to the hearer. The speaker may be mistaken or he may be in

bad faith. For example, I may say to you that they are showing

Richard II at the nearby theatre and find out that Richard II is

over and that they are now showing Richard III. In this case, the

information is irrelevant to you. Then, my utterance does not

comply with the presumption of optimal relevance. However, it will

be consistent with the principle of relevance, so long as I

rationally thought it would be optimally relevant to you.

 - 93 -

Thus, saying that an ostensive communication bears the

presumption of relevance only means that the intended interpretation

of an utterance is consistent with the principle of relevance, if a

rational speaker might have expected to communicate an adequate

range of contextual effects at the least possible processing effort.

It is this consistency with the principle of relevance which acts as

the sole criterion in utterance comprehension.

Sperber and Wilson's claim is that the principle of relevance

provides an adequate explanation of the role of contextual

assumptions in all aspects of utterance interpretation. In their

view, the responsibility for successful communication is not shared

equally by the communicator and the addressee. It is left to the

communicator to make correct assumptions about the codes and

contextual information that the addressee will have accessible and

will be likely to use in the comprehension process. The

responsibility for avoiding misunderstanding also lies with the

communicator. All that the addressee has to do is to recover the

interpretation which is consistent with the principle of relevance.

2.2.3.3. How context is determined

As shown above, relevance has been defined in terms of the

contextual effect which an assumption yields in a given context. The

 - 94 -

question to be asked, then, is how the context is determined. It is

a question which must be answered in any adequate account of

utterance comprehension.

In much of the literature, it is a common assumption that the

context for the comprehension of a given utterance is provided, and

is not a matter of choice. It is also assumed that the context is

not only determined at any given point of a verbal exchange, but

also that it is determined in advance of the comprehension process.

The assumption is that what the hearer does is to combine the new,

that is, what is explicitly expressed, and the old, that is, the

context which is already present in her mind at the start of the

communication. The view that context is fixed in advance is

independent of the mutual knowledge framework. It is just a view

that many semanticists and pragmatists happen to hold. Sperber and

Vilson argue that the selection of an adequate context is part of

the interpretation process, and must therefore be accounted for by a

pragmatic theory.

In cognition, the individual aims to maximise relevance.

Consideration of optimal relevance only comes in at the level of

communication, when the hearer is trying tofind out what

information is being offered. Here, the criterion isconsistency

with the principle of optimal relevance.

The hearer of an utterance has available a set of potential

contexts from which an actual context must be chosen. A context

 - 95 -

consists of assumptions from different sources, such as long-term

memory, short-term memory, perception, and so on. It does not mean

that any arbitrary subset of the total set of assumptions available

might become a context. Sperber and Vilson claim (1986a:138) that

the organisation of the individual's encyclopaedic memory, and the

mental activity in which she is engaged, limit the class of

potential contexts from which an actual context can be chosen at any

given time.

For example, it is generally accepted that encyclopaedic

information in long-term memory is organised into some kind of

units. They have been discussed in the literature under such terms

as 'schema', 'frame', 'scenario' and 'prototype'. Here, I shall

follow Sperber and Vilson and use the term 'chunk'. Chunks may

themselves be grouped into larger chunks, and contain smaller

chunks. It seems plausible to argue that the smallest units that

can be transferred from encyclopaedic memory to the memory of the

deductive device are chunks rather than individual assumptions. For

instance, it might be impossible to remember and add to the context

the information that an elephant has large ears without also

remembering and adding that it has a trunk. Also, it seems that not

all encyclopaedic information is equally accessible at any given

time. For instance, it seems plausible to assume that the

encyclopaedic entry which has just been accessed in the most recent

utterance comprehension is more accessible than any other.

 - 96 -

Let us assume that there is a small, Immediately accessible

context, fixed in advance, and consisting of the proposition which

has most recently been processed, together with its contextual

implications, and any assumptions used in deriving it. When new

information is received, it will be processed in this most immediate

context.

Now let us consider some examples. If the initial context is

(22) and the proposition expressed by the utterance is (23), some

degree of relevance is immediately achieved:

(22) If the opera is by Mozart, Peter will go.

(23) The opera is by Mozart.

What would happen, then, if the proposition expressed is (24a),

(24b) or (24c), instead?

(24) a. The opera is by the composer you have just

mentioned.

b. The opera is 'The Magic Flute'.

c. The opera is this (meaning the music which is

 - 97 -

being played at the time of the utterance).

Unless the initial context is extended in some way, no degree of

relevance can be achieved in the cases of (24a) to (24c).

If the goal of processing is to find an interpretation

consistent with the principle of relevance, i.e. which a rational

speaker might have thought would be optimally relevant to the

hearer, the hearer will be forced to add to the initial context (22)

further information, which will be remembered from earlier

discourse, as in the case of (24a), or recovered from encyclopaedic

memory, as in the case of (24b), or derived from perception, as in

the case of (24c). The goal will be to find premises which will

combine with the old assumptions and yield adequate contextual

implications in return for the minimum processing effort.

As these examples have shown, the accessibility of potential

contexts is partly determined by the content of the proposition

being processed. The content may direct the hearer's attention to

the preceding discourse, or it may direct it to encyclopaedic

information, or to the physical environment, as we have seen in

examples (24a), (24b) and (24c), respectively. There will be cases

where more complex extension is required and simultaneous or

sequential extensions of the context in a variety of different

directions will take place. There is no principled limit on the

number of extensions that may be needed to establish the relevance

 - 9S -

of a given proposition. It is only the hearer's capacity for

extending the context which will practically limit the number and

complexity of extensions involved for communication to be

successful.

General procedures of context selection are cognitive, and

geared to maximising relevance. Questions about the intended

context are answered by reference to the criterion of consistency

with the principle of relevance.

It is a basic assumption in most of the literature that the

comprehension process is as follows: firstly, the context is

determined; secondly, the interpretation process takes place;

lastly, relevance is assessed. Relevance is regarded as a by­

product of the comprehension process.

However, Sperber and Vilson argue that from a psychological

point of view, this is highly implausible. From a purely logical

point of view, any of the hearer's assumptions could be used in the

interpretation of the utterance. However, the constraints on

utterance interpretations are not only logical, but also

psychological. Humans are not in the business of simply assessing

the relevance of new information. Their interest is to process

information as efficiently as possible, that is, they try to obtain

from each new item of information as great a contextual effect as

possible for the smallest possible processing effort. The

assessment of relevance is not the goal of the comprehension

 - 99 -

process, but only a means to an end, the end being to optimise the

relevance of any information which is being processed.

Thus, Sperber and Wilson's account of the determination of

context is contrary to the pre-existing view in the literature: it

is not that the context is determined, and then relevance is

assessed; rather, humans expect that the assumption which is being

processed is relevant, and they try to select a context in which

their expectation is justified. In ostensive-inferential

communication, relevance is guaranteed. It is relevance which is

treated as given, and context which is treated as a variable.

Let us consider the following example and see how the speaker,

with specific expectation about how his utterance will be relevant,

directs his hearer towards an appropriate context:

(25) a. Kay: Would you like to go to Tosca?

b. Paul: I'm not keen on Puccini.

Here, Paul has not answered Kay's question directly, he has

only implied an answer. Kay has to recover the implicatures, or

some oi them, of the utterance. She must supply certain premises,

either by retrieving them from her memory or by constructing them by

developing assumption schemas retrieved from memory. Consistency

 - loo -

with the principle of relevance will provide an adequate criterion.

All she has to do is to supply premises which a rational speaker

might have thought would lead to an interpretation which yields

optimal relevance. One of the premises which Kay should be able to

supply is:

(26) Tosca was composed by Puccini.

By processing Paul's reply (25b) against a context which contains

assumption (26), Kay should derive the contextual implication (27):

(27) Paul would not like to go to Tosca.

Moreover, it is hard to see how Paul could have expected his reply

to be optimally relevant without being processed in this way.

Sperber and Wilson distinguish (1986a:194-5) two kinds of

implicatures: implicated premises, such as (26), and implicated

conclusions, such as (27), and claim that all implicatures fall into

one of these two categories.

They also argue that the kind of implicatures discussed above

share two properties. Firstly, they are fully determinate. Paul

 - 101 -

expects Kay to supply not merely something like premise (26) and

conclusion (27), but a premise and a conclusion with just this

logical content. Secondly, Paulis entirely responsible for their

truth. Even if Kay did not know that Tosca was composed by Puccini

prior to Paul's reply (25b), Paul has strongly encouraged Kay to

supply assumption (26) by uttering (25b). He is just as responsible

for the truth of (26) and (27) as if he had asserted them directly.

As Sperber and Wilson argue, this example shows that mutual

knowledge is not a prerequisite for successful communication. Paul

does not need to know if Kay has the information in (26), he would

not even have to wonder whether it is shared by Kay. Kay would come

to have this knowledge as a result of interpreting the utterance, if

not before. By following the criterion of consistency with the

principle of relevance, she will have to supply premise (26) and

then to deduce conclusion (27). Paul may expect Kay to supply this

premise, not because he has ground for thinking that it is already

highly accessible to her, but rather because his utterance has made

it accessible to her. In other words, by producing the implicit

answer (25b), he has constrained her choice of context and directed

her towards a particular interpretation.

2.2.3.4. Descriptive use and interpretive use

 - 102 -

P
So far, the analysis has been based on the assumjJ:ion that an

utterance is used to represent things in virtue of its propositional

form being true of that state of affairs. Let us consider the

following exchange:

(28) a. Kay: Vhat did Mary say?

b. Paul: She is tired.

How, Mary might have said to Paul (29):

(29) I am tired.

Then, (28b) would used in virtue of its propositional form being

true of that state of affairs.

Sperber and Wilson argue that there is another way in which an

utterance can be used, that is, an utterance can be used in virtue

of resembling some phenomenon, rather than in virtue of its

proposition being true. For example, Mary might have said (30):

(30) Tsukare-ta.
 - 103 -

tire past

I am tired.

(30) is a Japanese translation of (29). It would be used because it

resembles Mary's utterance (29), as it has the same semantic

structure.

Instead, Mary might have said to Paul (31)

(31) I am absolutely over-worked and could do with a good

rest.

Then, again, (29) would be used because of its resemblance to

Mary's utterance. It would be a summary. Even though (29) has a

different proposition to Mary's actual utterance, they share some

logical properties and some contextual effects.

Or else, Mary might have said things which did not make much

sense, which could be attributed to the fact that she was tired. In

this case, (29) would share with Mary's utterance neither its

logical and semantic properties, nor its contextual effects. It

would be used, because it resembled the state Mary was in.

 - 104 -

Thus, an utterance can be used to represent things in two ways.

On the one hand, it can be used to represent some state of affairs,

in virtue of its proposition being true of that state of affairs.

On the other hand, it can represent some other representation which

also has a proposition in virtue of some resemblance between the two

propositions. Sperber and Wilson name (1986a:228-9) the former

descriptive use and the latter interpretive use.

Any two phenomena resemble each other at least in some way.

There is a question of how close the propositional forms of two

representations must be, in order that an interpretive use be

acceptable. Sperber and Wilson claim that the criterion is the

principle of relevance.

This notion of interpretive use of language, that is,

representation in virtue of resemblance, is crucial to the analysis

of rhetorical devices, such as metaphor and irony. It also plays an

important role in the analysis of how language is employed to

present images of women in advertising, which will be the subject of

Chapter 5.

2.3. Summary
 - 105 -

In this chapter, I have argued that Sperber and Wilson's

Relevance Theory provides the most comprehensive account of

utterance interpretation. The framework is based on the notion of

ostention, that is, the communicator's intention to communicate and

to publicise this intention. It centres around the principle of

relevance, which claims that an ostentive stimulus creates a

presumption of optimal relevance. The task of the audience in this

kind of communication is to process the communicator's utterance

against background information and derive an interpretation which is

consistent with the principle of relevance. The correct

interpretation is the one that meets this criterion.

In the following chapters, I shall apply Relevance Theory to

the analysis of advertisements, with a particular focus on wordplay

such as puns, manipulative aspect of advertising, and images of

women in advertisements.
 - 106 -

Chapter 3: The Pun in Advertising

3.1. Introduction

Punning is one of the most common forms of speech play,

although it has long been regarded as having a somewhat low status.

Lionel Duisit describes puns as the 'least literary' (1978, cited

from Redfern 1984:4) of figures, while Dryden calls them 'the lowest

and most grovelling kind of wit' (1926:237, cited Redfern 1984:4).

Culler quotes Pope's dismissal of the pun, saying 'he that would pun

would pick a pocket' (1988:4).

However, not everybody is in the business of condemning and

dismissing the pun. Nor has it disappeared and been forgotten.

Sherzer argues that today puns are 'most often considered to be

humorous in intention, inappropriate for serious discourse'

(1985:215). They are often used as 'the witty comebacks of

conversation stoppers and the punch line of jokes' (1985:214).

Sherzer goes on to make the important observation that puns are

'highly appropriate for advertising' (1985:215). He does not

develop this throw-away remark any further, and he does not explain

how and why puns might be 'appropriate' for advertising. But it is

indeed striking how ubiquitous puns are in British advertising, and

how popular they also are among Japanese advertisers.

 - 108 -

(1) a. We're literally about to open.

b. Literally the finest store in Europe.

c. Book now for Christmas.

d. High brows raised here.

e. Browsers Welcome. (High brows and low brows)

f. Over 5 miles of books.

And they're all way over your head.

g. Materially supplied for seats of learning.

h. If you think this station's deep

You should see our poetry department.

i. Go to Dillon's. And be transported.

It looks as though people of letters in London appreciate the

pun, for these examples have been selected from the captions used in

a campaign run by Dillon's Bookstore after it was taken over by a

large multinational corporation. According to their advertising

agent (personal communication), Dillon's regard themselves as

catering to the intellectual population not only of London, but

 - 107 -

There is a suggestion that even people in advertising share the

literary critics' low opinion of puns, although this is open to

doubt. Redfern (1984) reports that when he wrote to twenty of the

largest international advertising agencies to find out the status of

wordplay in advertising, the most common reaction was to claim that

it was out of date to pun in advertisements. His explanation for

this is that the agencies were reflecting the age-old embarrassment

connected with puns, and upholding claims to the dignity of the

profession. Redfern quotes Hopkins, who is an ex-ad man:

Frivolity has no place in advertising. Nor has humour.

Spending money is usually a serious business... People do

not buy from clowns.

(Hopkins 1927, quoted from Redfern 1984:130).

As Redfern points out, this is contradicted by McDonald's

advertisements, which make use of a clown, although the situation

and the trend may have changed since the days of Hopkins.

Let us look at the following examples from an advertising

campaign for a prestigious bookstore, which would appear to indicate

that advertisers attribute a high status to puns:

y<? <<
,ar
 - 109 -

also of Britain, and possibly of Europe. They thought that the pun

would appeal to their target audience because of its wit and humour.

Apparently, the campaign has been very successful, Dillon's are

happy, and this advertising campaign based on the pun continues for

the moment.

A quick observation of advertisements, say at a tube station or

in a magazine, would prove that the Dillon's campaign is not an

exception. Use of the pun in British advertising is rife. This is

perhaps not so much the case in Japan, but the pun is still a

prominent form of wordplay found in Japanese advertisements. Even

though I do not have more precise information on the statistical

frequency of the occurence of puns in British and Japanese

advertising, there can be no doubt that puns are frequently employed

and are not considered inferior advertising strategy.

The reasons for the discrepancy between the information

provided to Redfern and the realities of the advertising world need

not detain me here, for the focus of this study is how puns function

in advertising. Redfern's research may be out of date (it was

published in 1984) or what we are observing might be the usual

difference between principles and practice. It is sufficient for my

purposes to point out that punning is a much used rhetorical device

in advertising both in the U.K. and in Japan. Vhat this chapter

aims at is to investigate how the device can be analysed in terms of

the interpretation process involved. My assumption is simply that

puns clearly offer advertisers advantages which they exploit.

 - no -

The assumed low status of puns may explain why attempts to

investigate them within an academic framework have not been

abundant. This is especially striking if one compares work on puns

to that done on other rhetorical devices such as metaphor. As there

has not been any study on puns carried out from the viewpoint of how

they are understood by the hearer, I shall not present a survey of

the literature.

Nor do I intend to present any taxonomy of puns. Culler

argues that of the studies which attempt to define and classify

puns, 'the results have never met with much success' (1988:4).

Redfern declines to categorise by quoting Mahood, saying, 'Naming

the parts does not show us what makes the gun go o f f (1984:5). On

the assumption that the assignment of an utterance to this

particular form of wordplay, namely the pun, or to a particular type

of pun, is not part of what is communicated and does not play a

necessary role in comprehension, neither definition nor

classification of the pun make any significant contribution to the

purposes of my study.

What makes the pun interesting to my study herejie that it

appears to be problematic for Relevance Theory. It is plausible to

argue that a pun requires greater processing effort than an

utterance which is not a pun. Then, is the extra processing effort

justifiable?. If it is so, how? If it is not, is punning a counter­

example to Relevance Theory? These are the questions which I shall

investigate in this chapter.

 In Section 3.2, I shall attempt to analyse puns in advertising

in the framework of Relevance Theory. Section 3.3 will consider

ambiguity in the pun. Although a pun has more than one possible

interpretation, it may not necessarily strike its audiance as being

ambiguous. I shall try to discuss why this is so. The comparison

between puns and metaphors will be the subject of Section 3.4. Both

a pun and a metaphor provide more than one possible interpretation.

So what is the difference between them? This is the question I

shall attempt to answer. In Section 3.5, I shall try to summarise

the advantages of using puns in advertising.

3.2. The interpretation of the pun

The process of interpreting a pun appears to be related to the

advantages it offers advertisers. The work by Redfern, mentioned

above, includes a chapter on puns in advertising, in which he

explains the utility of puns in advertising as follows:

Advertising space is costly. Economy is essential, and

puns are highly economical (two meanings for the price of

one word or phrase), and are in fact much more of a

labour-saving device than many of the products they seek

to promote
 - 112 -

(Redfern 1984:131).

Redfern's claims can be assessed in the light of the

development of a pragmatic theory of relevance by Sperber and Vi Ison

(1986a), which has provided an exciting new framework for analysing

the comprehension of puns. As I have shown above, Sperber and

Wilson argue that there is a single, necessary and sufficient

criterion which the hearer uses in interpreting the speaker's

utterance. This is consistency with the principle of relevance,

which I repeat here for convenience:

Principle of relevance

Every act of ostensive communication communicates the

presumption of its own optimal relevance.

(Sperber and Wilson 1986a:158).

When the communicator ostensively attracts the addressee's

attention, a presumption is created that the ostensive stimulus is

worth the addressee's attention. It is in the interest of the

addressee that the stimulus is the most effective one the

communicator could have chosen, that is, the most relevant one.

However, this may not always be in the communicator's interest.

What is in the communicator's intrest is to make sure that he has

 - 313 -

put the addressee to no unnecessary effort in achieving the intended

effects. If the addressee is in doubt that this has been done and

suspects that the communicator has deliberately chosen an obscure

stimulus and caused her an unnecessary processing effort, she might

doubt the communicator's genuine intention to communicate and refuse

to make an effort to process the stimulus produced. Here the

communicator's interest coincides with that of the addressee. It is

in his interest to be understood and therefore to make it as easy as

possible for the adressee to understand him. The stimulus he

produces must be the most relevant one he could have used to achieve

the intended effects.

All this does not seem to apply to a pun, for it demands extra

processing effort. A pun triggers an alternative interpretation by

its phonetic similarities. The addressee seems to be put to

unnecessary processing effort into deriving an intended

interpretation. Contrary to what Redfern says, a pun appears to be

uneconomical. Or, is the extra processing effort imposed upon by a

pun compensated for by extra contextual effects? I shall examine

this point in this section.

Let us look at some examples of puns. The first one is from an

advertisement for London Transport which appeared in the London tube

stations and trains in 1981:

(2) Less bread. No jam.

 - 114 -

The criterion of consistency with the principle of relevance

helps the addressee recover its intended interpretation. Reading

the words 'bread' and 'jam', she will most probably take the words

to mean 'bread' and 'jam' as food. Those will be the most

accessible interpretations for most people. 'Bread and Jam' are

stereotypical foods; whether separately or together, 'bread' and

'jam' will be thought of as food. However, those interpretations

will have to be rejected, as inconsistent with the fact that it is

an advertisement, that it is for London Transport, and that it is

found in an underground station or train. Having rejected the

interpretations which will come to her mind first, the addressee

would hopefully remember that 'bread' can mean 'money' and 'jam' can

mean 'traffic jam', and recover the proposition, 'less money, no

traffic j a m * .Having done this, she will have to resolve some

further indeterminacies, such as what costs less money, costs whom

less money, costs less money than what, and so on. Following the

criterion of consistency with the principle of relevance and

considering that it is an advertisement for London Transport, and

that it is meant to promote London Transport, something like the

following should be recovered:

(3) If you travel by London Transport it will cost you

less than travelling by car, and you will suffer no

traffic jams, unlike when travelling by car.

 - 115 -

Although this may seem ironical when the addressee reads it, being

jammed in a train, it is unlikely that it was intended by the

advertiser to be so.

Now, it scarcely seems economical to use the pun, contrary to

Redfern's arguement. In terms of the processing effort it costs the

addressee to recover proposition (3), it can hardly be said that it

was economically expressed. The addressee will have to reject the

first accessible interpretations of 'bread' and 'jam', and search in

her memory for more relevant interpretations. 'Money', for example,

may not be a highly accessible interpretation of the word 'bread'.

In any case, the advertiser could easily have used 'money' and

'traffic jam', and he must have been aware that these

interpretations are not the first ones to be recovered by his

addressee from thewords actually employed. He is deliberately

causing her an extra processing effort.

If apun was uneconomical and caused the addressee unnecessary

processing effort, it would breach the principle of relevance.

Then, punning would be a counter-example to Relevance Theory. Does

Relevance Theory fail to account for punning? Does the pun in (2)

put the addressee to unjustifiable processing effort?

Although a pun may seem to contradict the principle of

relevance at first sight, it does not require more processing effort

than necessary, and Relevance Theory can account for it. In fact,

it is precisely the advertiser's aim to get the addressee involved

 136 -

in the advertisement. First of all, an advertisement must attract

the addressee's attention. The caption will do just that because it

was deliberately put there, but it seems so irrelevant at first

sight. It will make her think, 'Vhat does that mean?' It will be

more successful in attracting the addressee's attention than a

caption like, 'Take the tube. It is cheaper...', which she may

entirely ignore. The purpose of this caption is not to convey a

novel idea: there are not many new things to say about the tube.

When the ultimate message is so obvious, namely, 'Take the tube', it

may well be more appealing for the addressee when it reads initially

as if it was about 'food', rather than London Transport.

Another point to note is that at first sight, the caption seems

'negative', which is unusual for an advertisement: from our

experience we are used to advertisements always mentioning good

things only. This caption is different, for it is saying less or

none of desirable things, namely, 'bread' and 'Jam'.'Jam',

especially, has a meaning of something desirable, as in the

expression 'jam tomorrow'. Thus the caption strikes the addressee

by sounding negative and makes her wonder what it is about.

Following on from this is the fact that since the caption costs

the audience an extra processing effort, it sustains their attention

for longer. This means that the caption may be remembered for

longer. It is not a straightforward message, and it causes some

thinking. Some of the audience may like it, because they think it

is 'clever', or because it is 'unusual'. Some may not recover the

 - 137 -

message immediately, and then they may think about it for a while

until they finally get it, or they may ask someone else what it

means. Some of the audience may decide that they are not very keen

on it, for it is so 'obscure'. These different reactions will all

be welcomed by the advertiser. An advertiser would like his

audience to look at his advertisement, think about it, and react to

it. Even if the reaction is negative, advertisers consider that any

reaction is better than no reaction.

From the point of view of the addressee, it may seem at first

sight that there is no extra reward to outweigh the extra processing

effort caused by the pun. In recovering the intended message of the

caption, the addressee will automatically process the 'food'

interpretation of the words 'bread' and 'jam', for these will

normally be the most accessible interpretations, which will trigger

the recovery of some encyclopaedic knowledge about food. However,

they do not give rise to contextual effects in themselves. These

interpretations are intended by the advertiser to be recovered by

the addressee, but they are not the intended interpretations.

Rather, they are intended to be ultimately rejected in the search

for the intended interpretations.

However, an extra reward which the addressee will get is the

pleasure of solving the pun. Solving a pun can give a kind of

intellectual satisfaction to many people. And the resulting

pleasure may have been intended by the advertiser, and give rise to

a number of contextual effects. For example, the audience could

 - 218 -

have self-congratulatory thoughts, thoughts about the congeniality

of the advertiser, and so on. So the addressee does get some reward

- but at the price of recovering the advertiser's intended message,

which she may otherwise have discounted or ignored.

There is also the point that if the pun had not been present,

the addressee would have paid no attention to the advertisement, and

hence it would have achieved no effects. So the effort needed to

process the pun is still the minimum the advertiser was justified in

demanding, given that he wanted to achieve the effects he did.

Hence, the effort demanded is consistent with the principle of

relevance.

According to Relevance Theory, the ostensive stimulus is the

c
most relevant, i.e. most economjal, one the communicator could have

used to achieve the intended effects. Does this apply to the case

here? The answer must be yes. The caption is the most economical

one the advertiser could have used to achieve the intended effects,

for, without going through the whole process, the addressee would

not get the message at the appropriate strength; she would not even

pay attention to it if the caption was something like, 'Take the

tube...'

Another aspect of Redfern's claim of the economy required in

advertising is his characterisation of the pun as 'two meanings for

the price of one word or phrase' . Are there really two meanings

communicated in the London Transport advertisement? Is this

 - 119 -

generally the case? Surely, it varies. Redfern does not clarify

what he means by 'two meanings'. As far as the meaning which the

communicator has intended his audience to retain is concerned, there

is only one meaning. The advertiser has intended 'bread' to mean

'money' and 'jam' to mean 'traffic jam'. The 'food' interpretations

are intended to be triggered, but they are also intended to be

abandoned. They are not intended interpretations as such, and that

is precisely why the audience will continue to search for other

interpretations. If the addresseee was to take these 'food'

interpretations to be the only ones intended by the advertiser, she

would stop processing and would not search for other

interpretations.

It has been pointed out by Moeran (personal communication) that

in the U.K.,food is often used to promote technology. 'Apple

Computer', 'Apricot' and 'Peach Software', for example, use food

names as their brand names. Packard, in his classic The Hidden

Fersuaders, quotes some social science research, saying that food is

'loaded with hidden meanings'. (1981:87). It is possible that food

names have some appeal to an audience, as it gives humans life.

The use of food names in the promotion of technology, or any

products other than food, is not practised in Japanese advertising.

On the contrary, it has been observed that new brand of snacks are

often given names which have no apparent association with food

(Kookoku Hiyoo, December, 1984). Some of the examples are 'Suzuki-

san' (Mr. Suzuki), 'Sato-san' (Mr. Sato), 'Choonan' (the first son),
 - 120 -

and 'Choojo' (the first daughter) for crispy snacks, 'Yama no

Donguri* (the acorn in the mountain) for chocolate snacks, 'Kiri no

Uki-fune' (the floating boat in the mist' , for chocolate. The point

may be worth investigating for cultural studies, although I shall

not pursue it in this study.

Let us now look at another example, which is acaption for

Papillon underwear. It reads as follows:

(4) The last thing you'll want to be seen in.

This caption is accompanied by an illustration, showing a woman in

brassiere and panties. She has a white jumper around her neck and

shoulders, which she is pulling with her hands in front of her

throat. It seems that she is taking it off.

Reading the caption, (4), would immediately force the addressee

to recover the following interpretation:

(5) You would not want to be seen in Papillon underwear.

This would gives the addressee a context which includes the

following assumption:
 - 121 -

(6) One does not want to be seen in something because it

does not look nice.

Considering that it is an advertisement for underwear, and that

advertisements normally say positive things about the product being

advertised, it is quite unlikely that the advertiser has intended to

communicate (5) to his addressee. This element of surprise would

contribute to drawing the addressee's attention to the

advertisement. The advertiser has intended the addressee to reject

(5). In fact, underneath the caption it reads:

(7) Much too pretty to hide.

On the assumption that (4) and (7) are somehow connected, the

addressee might derive (8):

(8) The last thing you'll want to be seen in, because it

is much too pretty to hide.

Then, the addressee would have access to another interpretation of

(4). The advertiser has intended the audience to recover (5), but
 - 122 -

then reject it, and eventually recover the alternative

interpretation (9):

(9) The Papillon underwear is the final piece of clothing

you want to be seen in as you undress.

If (9) is processed in a context which includes (7),it would

give rise to further contextual effects, including (10):

(10) You want to be seen in Papillon underwear because it

is so pretty.

It seems that examples (2) and (4) belong to the same category

of puns. They trigger an interpretation, which is intended by the

speaker, but also intended to be rejected, so that the addressee

recovers an alternative interpretation, which is the one the speaker

has intended to communicate. They have more than one possible

interpretation. However, the speaker has intended to communicate

only one of them. The rest are intended to be triggered, only to be

eventually rejected. The immediately accessible interpretation,

which is later rejected, is irrelevant in the case of (2), and

'negative' in the case of (4). These properties help the

advertisements catch the audience's attention. They also contribute

to the increase in strength of the messages and the increase in

memorability. Moreover, the audience get some satisfaction at

solving a pun, which would give rise to a weak contextual effect.

There are more examples in this category, such as two captions

for a Milton Keynes advertising campaign:

(11) a. Mind your own business. Move it to Milton Keynes,

b. You're welcome to London. Milton Keynes.

(11a) was found at the Milton Keynes railway station. (lib)

was found on a billboard which can be seen from the train

approaching London Victoria Station. In both cases, the first

accessible interpretation of the first part of the captions would be

rejected in favour of an alternative interpretation, which is the

intended one.

The extent to which puns do or do not contribute to conveying

the intended message varies from pun to pun. In the next example,

unlike the London Transport advertisement, the processing of the

advertisement as a pun does contribute to the overall message

conveyed, despite the fact that the pun itself does not even provide

an alternative propositional content. In 1977, the year of the

 - 124 -

Queen's Silver Jubilee, Guinness presented an advertisement which

reads as follows:

(12) We've poured throughout her reign.

In order to identify the proposition expressed by the above

utterance, disambiguation of the word 'pour' and reference

assignment to the two pronouns, namely, 'we' and 'her', must be

undertaken. Consistency with the principle of relevance provides an

adequate criterion. 'Pour' can mean either 'flow* or 'serve

liquid'. Considering the fact that it is an advertisement for beer,

the 'serve liquid1 interpretation will be selected. From our

experience, we know that the word 'we' in advertisements can refer

to many things; an advertiser, audience, characters in it, the

product advertised, and so on. Here, though, because of the verb

which follows, namely, 'pour', it will be assumed that 'we' mean the

Guiness company, which has been serving beer. The second pronoun

will be more easily assigned. Because it is followed by the word

'reign', and because it is the Queen's Silver Jubilee year and one

hears about it all the time in the media, it will be interpreted as

'the Queen's'. Thus, the following proposition will be recovered:

(13) Ve, the Guinness company, have served beer for as

long as the Queen has been on the throne.

 - 125 -

However, the addressee could not help thinking of 'pour' in the

sense of 'flow' and 'reign' as 'rain'. Although it is spelled 'r-e-

i-g-n', and not 'r-a-i-n', among the homophones, that is, the words

with the same phonetic properties, such as 'rain', 'reign' and

'rein', 'rain' will probably be the most accessible one to the

addressee. Reading 'reign', the addressee will think of 'rain', and

they will be encouraged to do so by the word 'pour'. The advertiser

further encourages his addressee along this line by deliberately

omitting the word 'beer' after the verb. He could have said:

(14) Ve've poured beer throughout her reign.

but that would have invited the interpretation 'pouring rain' much

less effectively.

The 'pouring rain' interpretation is only triggered; it cannot

be claimed to be explicitly expressed, for the spelling used is

'reign' and not 'rain'. However, the image of 'pouring rain' will

be within easy access of the addressee, because of the kind of

weather she is used to, so the effect would be achieved without much

difficulty. And once 'pouring rain' is triggered, it will give the

addressee access to her encyclopaedic knowledge about rain, which

will include that it is typical British weather, that it is typical

Queen's weather, for it tends to rain on her big occasions, and so

on. So, the association between 'beer' and 'rain' may be added to
 • 326 -

the context in which the advertisement is finally processed, as an

implicated assumption.

In Japan, 'beer' has the image of a 'summer drink', as opposed

to sake (rice wine), which is traditionally regarded as a 'winter

drink' . Beer is consumed cool and is advertised more before and

during summer. When there is a bad or shorter summer due to an

extended rainy season, the sales figures drop. However, in the

U.K., where most beer is consumed lukewarm, the situation is

different. It is a 'year round' drink, just as rain is a 'year

round* weather. In Japan, beer is always advertised as a drink for

hot and sunny weather. But in the U.K., it is reasonable that the

advertiser should want to associate it with rain. The advertisement

came out in the summer of 1977, when there was an unusually wet and

bad summer. Guinness has always presented an image of a 'national

drink*. It is surely not a coincidence that its advertiser has made

an association with 'pouring rain', which is one of the prominent

national characteristics of the country. So also is joking about

the weather. Each of these additional contextual effects serves to

enhance the desired effect of the advertisement that Guinness be

seen as an integral part of Britain's national heritage, even though

it could be argued that Guinness is in reality an Irish beer.

In this example, then, the pun is used to provide access to

additional implicated contextual assumptions, which do give rise to

intended contextual effects when the intended interpretation is

eventually recovered and processed. The pun is not merely an

 - 127 -

attention-getting device, but it actually contributes to the

eventual intended interpretation.

Let us look at a Japanese example. The following is a caption

for an airline advertisement, promoting its Okinawa flight, Okinawa

being islands in the far south, known as a resort area:

(15) Oo Kli NAa VAh

The capital letters correspond to larger characters in the

Japanese original. The underlined letters are in katakana, which is

a syllabic script used for things like onomatopoeias, loan words,

exclamations, and so on, while the rest are in hirakana, which is

the unmarked form of syllabic script.

By picking up the larger letters only, it reads 'o-ki-na-wa',

that is, 'Okinawa', the destination for the promoted flight.

Reading all the letters, it reads 'oo-kii-naa-wah', where ookii

means 'big', naa an exclamation particle, and wah in katakana an

exclamation like 'wow'. Thus, the two different interpretations

will be recovered:

(16) Okinawa.
 - 128 -

(17) How big! Vow!

The question which (17) will raise is: what is big?

Considering that it is an advertisement promoting an Okinawa

holiday, it is likely that it refers to the Okinawa Islands. For

the Japanese, who live in an over-crowded country, space is a major

advantage, and it is reasonable for the advertiser to want to stress

it. On the other hand, the audience could not help noticing that

the girl in the illustration, in her bikini, is rather well-endowed,

and wondering if ookii (big) refers to her breasts.

There are at least two possible candidates for the object

described as ookii (big). However, the advertiser's intention in

regard to them does not seem to be equally manifest. Among my

several Japanese-speaking informants, some wondered if the 'breasts'

one was a coincidence, and some did not question the intentionality

of it at all. Some of the second group did not even think of the

possibility of Okinawa being referred to, and incidentally, the

second group were exclusively men. Considering that it is an

advertisement, which is a carefully prepared work, where what

Chomsky calls 'performance error' is unlikely, it is probable that

the 'breasts' interpretation is invited intentionally, rather than

accidentally.

Thus, the following interpretation might be derived:

 - 129 -

(18) In the Okinawa Islands, there are a lot of beautiful

girls with big breasts.

However, I would hesitate to say that the advertiser has

intended to convey this on the basis of making his intention

mutually manifest. Thus, the communicator's intention to convey an

interpretation is sometimes made mutually manifest, and in others it

is not. This question of the publication of the speaker's intention

will be the topic of Chapter 4.

Nudity is shown to a greater extent in advertising and in mass

media in general in Japan than in the U.K.. Even so, to show a

well-endowed girl in a bikini and explicitly say, 'How big her

breasts a re !' will be regarded as irrelevant and too vulgar for an

advertisement for an airline company. The advertiser's intention to

refer to the girl is not made as manifest as it could have been; it

may not become mutually manifest at all. It is vague, which leaves

some room for the advertiser to deny the intention if necessary.

This example, then, may be regarded, unlike the two previous

ones, as a case of deliberate equivocation, that is, unresolvable

ambiguity. Or it may be that the pun is used only as an attention-

getting device, with the 'space' interpretation being the only one

mdcle mutually manifest, and perhaps additional contextual effects

about the sort of people you will find in Okinawa also being

communicated.
 - 130 -

Examples (12) and (15) seem to be different from examples (2)

(4), (11a) and (lib) which, it could be argued, form another

category. (2) and (4) force the addressee to recover

interpretations which she eventually rejects as irrelevant. The

extra processing effort caused by the unintended interpretations is

compensated for by the increased strength of the messages and their

increased memorability, or by the addressee's pleasure at solving

the puns. In the cases of (12) and (15), the unintended

interpretations provide access to encyclopaedic information, which

is actually used in processing the intended interpretations, and

thus contribute to additional contextual effects.

So far, I have analysed two categories of puns. The first is

that which triggers a seemingly most accessible interpretation,

which is eventually rejected as irrelevant and not contribute to the

processing of the intended interpretation. In the second category,

a pun also triggers an interpretation which is to be rejected as

unintended. But this interpretation contributes to the processing

of the intended interpretation by providing access to encyclopaedic

information, against which the intended interpretation is processed

and yields contextual effects.

There is a third category, in which a pun triggers two or more

interpretations, neither of which is intended to be rejected by the

speaker. Let us look at the following caption, which is a caption

for a Mazda advertisement:

 - 131 -

(19) The perfect car for a long drive.

The advertisement shows a Mazda car parked on a lengthy drive

which leads to a mansion. The word 'drive' can mean 'a car ride',

and on this interpretation the audience would derive (20):

(20) The perfect car for a long car ride.

Given that it is an advertisement for a car, (20) would yield a

number of contextual effects, which include (21) and (22):

(21) The car is for people who have to go on long car

rides.

(22) The car functions well for a long drive.

However, 'drive' can mean 'driveway* and the illustration

showing a long driveway would encourage this interpretation. Thus,

the addressee is encouraged to derive (23) from (19) as well:

(23) The perfect car for people who have a long driveway.
 - 332 -

The illustration also shows a mansion to which the driveway leads,

and it would encourage the addressee to extend her search for

assumptions about the kind of life style which may be suggested by

it. Thus, the addressee might derive further contextual effects,

such as (24) and (25):

(24) The car is perfect for people who have a long drive­

way and a mansion.

(25) The car is perfect for people who enjoy a comfortable

1 if e .

In this case, both (20) and (23) yield substantial contextual

effects and it would not be clear to the addressee which is the

interpretation intended by the speaker. It is manifest to both the

addressee and the advertiser that both interpretations are relevant

enough to be worth the addressee's attention, and both were intended

to be triggered. One reasonable inference for the addressee to draw

is that both were intended to be retained.

The next example is an advertisement for women's underwear:

(26) Next to myself, I like Vedonis.

 - 133 -

where 'Vedonis' is a brand name. The audience would probably derive

interpretation (27) as the most accessible one:

(27) After myself, I like Vedonis.

However, the addressee would not be able to help thinking that

it is a strange thing to say about underwear. For, underwear is not

an obvious thing that one likes after oneself. It is more common to

talk about, for example, one's mother.

'Next to' can be interpreted either physically or more

abstractly. Considering that it is an advertisement for underwear,

which one wears next to the skin, the addressee would also find a

physical interpretation also possible, and thus recover proposition

(27):

(28) Next to my skin, I like wearing Vedonis underwear.

This would give rise to further contextual effects, such as (29):

(29) Vedonis underwear feels good on one's skin.

 - 134 -

However, there is no denying that the communicator has chosen

an utterance which causes his addressee extra processing effort by

using (26), when he could easily have said, for example, (30)

instead:

(30) Next to my skin, I like Vedonis.

Moreover, it would be more common to say (31), rather than (26), if

one intends to communicate physical contact to one's skin:

(31) Next to me, I like Vedonis.

The interpretation would therefore seem to be inconsistent with the

principle of relevance, since the intended effects could have been

more economically achieved. Perhaps, the speaker has also intended

the more abstract interpretation (27), which I repeat here for

convenience:

(27) Alter myself, I like Vedonis.

 - 135 -

Underwear is not exactly meant to receive public exposure.

Unlike other outfits, which can be chosen on the ground of their

appeal to the spectators, underwear has to be chosen on the ground

that it appeals to the wearer. Interpretation (27) is based on that

assumption. Furthermore, the advertiser would like his addressee to

choose underwear on the ground that the wearer likes it, rather than

that it is functional. He is trying to encourage his addressee to

buy underwear which she likes, not Just something sufficient, even

though she may not be able to show it off to others. That seems to

be the reason why the advertiser would want to tickle the

addressee's narcissism.

The brand name 'Vedonis' can be analysed as a combination of

'Venus' and 'Adonis', which are both names of mythological figures

famous for their beauty. In Greek mythology, Adonis was loved by

Aphrodite for his beauty, and Aphrodite herself was the goddes of

beauty, corresponding to Venus in Roman mythology. Hence the more

abstract interpretation may further suggest (32):

(32) I like Venus and Adonis for their beauty after

myself.

Vhen I tried this hypothesis with native English speakers, the

only person to come up with Venus and Adonis was an anthropologist,

whose research interests include the language of advertising, and I

 - 136 -

did wander if we were reading too much into it. Vhat has happened

is, perhaps, that due to an unusually high level of attention which

the anthropologist and I shared, we put more than the usual

processing effort into the interpretation, and we extended our

search in memory to a greater degree than a general audience would

have done. It could be argued that the advertiser could not have

expected the audience all to be familiar with Greek and Roman

mythologies, or have such a high level of attention. However, it is

unlikely that it was a coincidence. The advertiser could not have

expected it to be unnoticed by everybody in his _ audience,

either.

Obviously, it was not the advertiser but the company which

chose the name 'Vedonis' . This choice may have been made for its

association with Venus, Adonis and beauty, and a context of this

type might be automatically accessed by at least some readers of the

advertisement. This will in turn provide access to implicated

assumption (33):

(33) Vedonis underwear is so beautiful that some people

might like it after themselves.

and this assumption can be combined with the more physical

interpretation (28) to implicate the following conclusion:

 - 137 -

(34) I like Vedonis not only for its function, but also

for its exceptional beauty.

Or, it could be-argued that the advertiser exploited the fact

that the brand name is a combination of Venus and Adonis, and used

it to further tickle the addressee's narcissism by suggesting (35):

(35) I like Venus and Adonis for their beauty, but only

after myself.

The advertiser has chosen the caption, being aware that it will

make the addressee go through both abstract and physical

interpretations (27) and (28). Yet, in choosing utterance (26), he

is consistent with the principle of relevance. For, if he had

chosen an utterance such as (30) or (31), which are more

straightforward, the addressee would not have recovered (27) and

would have missed any additional implications to do with the

exceptional beauty of the underwear. Without going through the

process of having to reach the physical interpretation after the

abstract one, the addressee would not pay as much attention to the

caption, either, and therefore would not get the message. In

deriving the interpretation intended by the speaker, the addressee

is encouraged to process both the abstract and physical

interpretations. Then she will be automatically given access to

 - 138 -

contextual assumptions needed for both interpretations, which

include the functional qualities of underwear, narcissism and the

mythologies. Thus, the relevance of (26) is established by

recovering a wide range of weak contextual effects, arising from two

distinct interpretations.

Thus, (19) and (26) are neither like (2), (4), (11a) and (lib),

nor like (12) and (15), in that the communicator has intended both

interpretations to be recovered and retained. In the cases of (2)

and (4), the communicator has intended only one interpretation to be

retained and the rest to be rejected as irrelevant. In the cases of

(12) and (15), the communicator has intended only one

interpretation. Nevertheless, the unintended interpretation is to

be recovered to give access to encyclopaedic knowledge, which is to

be used in processing the intended interpretations, so that

additional contextual effects can be achieved. In the case of (19)

and (26), neither of the two sets of possible interpretations, that

is (20) and (23), and (27) and (28), respectively, is intended by

the communicator to be rejected by the addressee. The communicator

intends his addressee to process and retain both, together with

their effects. Thus, (19) and (26) are not only ambiguous, but

equivocal, i.e. unresolvably ambiguous.

Redfern's comment that puns provide 'two meanings for the price

of one word or phrase' is valid in these cases. The word 'drive'

means both 'car ride' and 'drive way', and they are both intended by

the communicator, unlike the first two cases. The phrase 'next to'
 - 139 -

can be understood both in the physical and in more abstract senses.

In the cases of the first two types, there is only one intended

interpretation. However, more than one interpretation is triggered,

and gives access to various kinds of encyclopaedic information,

which can be used in processing the intended interpretation. In the

case of the third type, both interpretations are intended by the

communicator.

It has been argued that puns can be categorised in terms of

hierarchy among possible interpretations they provide. However,

this is not to say that they are distinctive categories. I would

argue that the boundaries are not clear-cut and are a matter of

degree. It could be argued that these examples occur along a

spectrum, with examples (2), (4), (11a) and (lib) belonging to one

end of the spectrum, and example (19) and (26) belonging to the

other. Other examples fall somewhere in between.

I have thus argued that Sperber and Wilson's Relevance Theory

(1986a) sheds light on a variety of ways in which puns function in

advertising:

1. They attract the addressee's attention; the extra processing

effort involved may be said to be the price the advertiser has

to pay to get his message noticed at all. Without going

through the whole process, the addressee might ignore the

message, and hence no effect would be achieved.

 - 140 -

2. The extra processing effort demanded is compensated for by the

increased strength of the message conveyed, or by increased

memorability. An opinion which the addressee might discount as

obviously of little credibility is strengthened and possibly

remembered due to the extra processing effort involved.

3. There are extra contextual effects based on the addressee's

pleasure and satisfaction at having solved the pun: these may

affect the addressee's attitude to the advertiser and

ultimately to the product advertised.

4. The extra unintended, or uncommunicated, interpretation

provides access to encyclopaedic information which is used in

processing the intended interpretation, and thus gives rise to

additional contextual effects.

In all these cases, the advertisement achieves optimal

relevance despite the extra processing effort, because it is still

the most economical way of achieving the full range of intended

contextual effects.
 - 141 -

3.3. Ambiguity and the pun

I have argued above that a pun presents more than one possible

interpretation and that the hearer selects the interpretation which

is consistent with the principle of relevance, which should be the

one intended by the speaker. Then, does a hearer of a pun see it as

ambiguous? In this section, I investigate the relationship between

the pun and ambiguity.

Leach argues that a pun 'forbids us to recognise that the sound

pattern is ambiguous* (1964:25), without explaining how the hearer

is 'forbidden' from seeing a pun as ambiguous. Whatever his reasons

may be, I would argue that his comment is unsound. The audience

must see the ambiguity of the pun in order to recognise it as a pun.

Attridge argues that a pun is not just ambiguity, but it is

'ambiguity unashamed of itself' (his italics) (1988:141). He points

out that in spite of its polysemous nature, language works well

enough because a context acts as a disambiguating device:

In place of a context designed to suppress latent

ambiguity, the pun is the product of a context

I
deliberately constructed to enforce an ambiguity, to

render impossible the choice between meanings, to

i
 leave the reader or hearer endlessly oscillating in

semantic space. (his italics)

(Attridge 1988:141)

There is an underlying assumption in his argument that a

context is determined prior to an utterance. However, this

assumption is false, as argued by Sperber and Wilson (1986a) and

discussed in Chapter 2. Attridge goes on to argue that a context,

which normally helps disambiguation, enforces ambiguity in the case

of a pun, althou^khe does not explicate how a context 'enforces' the

ambiguity of a pun.

Leach and Attridge seem to contradict each other. On the one

hand, Leach argues that, although a pun is ambiguous, the audience

does not recognise it as ambiguous, because there is something about

a pun which prevents the audience from recognising it as ambiguous.

On the other hand, Attridge's claim is that a pun is not only

ambiguous - and presumably that the audience would recognise it as

ambiguous - but also that its context is constructed in such a way

that the ambiguity of the pun is enforced.

First of all, it is necessary toclarify what is meant by

'ambiguous' here. Utterances which are strictly speaking

indeterminate are so ubiquitous in ordinary communication that it is

difficult to find utterances which do not require any

 - 143 -

disambiguation, reference assignment and enrichment. Let us look at

the following examples.

(36) a. Kay went to the bank.

b. Paul has got his umbrella.

c. I will see you soon.

The word 'bank' in (36a) can refer to either a financial

institution or a river bank and needs disambiguation. The umbrella

mentioned in (36b) can be either Paul's or some other male human

being's, and 'his' has to be assigned to an appropriate reference.

The hearer of (36c) would have to enrich it by deciding how soon the

word 'soon' means, whether it is a matter of a few minutes, hours,

weeks, and so on.

However, hearers of (36a) - (36c) do not necessarily regard

these utterances as ambiguous. Nor do the speakers usually intend

these utterances to be ambiguous. They may be linguistically

ambiguous - in fact, Sperber and Wilson would treat (36a) as

ambiguous, and (36b) and (36c) as ambivalent or vague - but the

ambiguity is almost invariably resolvable in context. Indeed,

unless the ambiguity is resolved the utterance will not succeed as a

vehicle for communication. What the hearer has to do in processing

utterances, such as (36a) - (36c), is to identify the propositions

which meet the criterion of consistency with the principle of

relevance.

However, there are cases in which the hearer is not able to

identify the interpretation intended by the speaker. These

utterances are then interpreted as ambiguous, and the hearer, by

asking, 'Which bank?', 'Whose umbrella?', 'How soon?' and so on, has

to seek clarification before being able to recover the speaker's

intention. Such unrecoverable ambiguity is called equivocation.

Sperber and Wilson argue that the success of communication

depends on the hearer's recovery of the speaker's intended meaning

of the utterance (not merely on her recognition of its linguistic

meaning) (1986a:23). Communication can succeed, even when there is

more than one possible interpretation of the utterance. In

principle, it can succeed with various types of indeterminacy of the

utterance, so long as the speaker's intended interpretation is

recoverable. Communication fails when ambiguity is unresolvable,

for unless there is such resolution, the hearer will not be able to

determine which intentions the speaker had. It becomes equivocal

when the hearer is unable to disambiguate it and hence assign it a

single intended interpretation.

Puns stand in striking contrast to this because conscious

recognition of ambiguity as part of the utterance process is

essential to processing the utterance as a pun. However, even in

 - H5 -

these cases, there is a single communicative goal which a speaker is

trying to achieve in using a pun, and this is what the hearer has to

recognise. In some cases, this goal is transparent. Although two

or more interpretations are intentionally triggered by the speaker,

only one is intended to be retained, as we have seen examples (2),

(4), (12) and (15). Let us look at example (11a), which is a

caption for an advertisement promoting Milton Keynes, found at the

railway station. I repeat it here for convenience:

(11) a. Mind your own business. Move it to Milton Keynes.

This is accompanied by a line underneath, saying:

(37) Curious to find out why so many top British companies

are moving to Milton Keynes? Ring (the number).

The first sentence of (11a) is a pun. The expression 'mind

your own business' can mean (38):

(38) Do not concern yourself with other people's affairs.

 - 146 -

However, this does not yield enough relevance to outweigh the

addressee's processing effort, as the addressee would have no idea

what other people's affairs it is referring to. It would probably

help to attract the addressee's attention, for the expression 'mind

your own business' tends to be used as a harsh statement.

The addressee would then be encouraged to search for

alternative interpretations in her memory, and she would hopefully

remember that 'mind' can mean 'look after' and that 'Business' can

mean 'firm', 'company' and so forth. Furthermore, the addressee

would probably know that it is an advertisement, because of the kind

of place in which it is found. (37) would help her to realise that

it is an advertisement for Milton Keynes. She would then derive

(39):

(39) Look after your own company.

Together with the latter part of the caption, the addressee would

derive further contextual effects, such as (40):

(40) You can improve your company prospects by moving to

Milton Keynes.
 - 147 -

Thus, (39) would yield enough relevance to make the addressee's

processing effort worth while.

The addressee would recover (38), but reject it, as it does not

yield enough relevance to make it worth her while to process it.

Then she would recover (39), which is consistent with the principle

of relevance. It is manifest to both the addressee and the

advertiser that the advertiser intends her to notice both (38) and

(39). However, it is also made mutually manifest to both the

addressee and the advertiser that (38) is to be rejected and that

(39) is the interpretation which is intended by the advertiser.

Ambiguity is not tolerable unless the communicator could have meant

both. But the communicator could only have meant both if both

contribute however indirectly to the single overall intended effect.

Example (11a) belongs to the first category of puns, as we have

discussed above. There is only one interpretation that is intended

by the speaker, namely (39). The other interpretation (38), which

the audience would be likely to recover first, is intended by the

speaker only to be triggered but eventually rejected as irrelevant.

Leach's comment that a pun forbids the audience from

recognising it as ambiguous is acceptable insofar as the ambiguity

of a pun is usually resolvable. The audience of a pun would be able

to identify the intended interpretation, which is the key to

successful communication. Moreover, the audience would appreciate

 - 148 -

the ambiguity of the pun as essential for its success, and would not

see ambiguity as failure.

However, what about the third category of puns, such as (19)

and (26)? They have more than one interpretation intended by the

speaker, both of which are intended to be retained, unlike the first

two types of puns. I repeat the example for convenience:

(19) The perfect car for a long drive.

As discussed above, it has the following interpretations.

(22) The car functions well for a lengthy drive.

(25) The car is perfect for people who enjoy a comfortable

life.

However, despite the independence of (22) and (25) from each

other, they can combine the additional premise that people who have

a long drive are the sort of people who, drive long distances. So,

despite the essential recognition and retention of the two triggered

interpretations, there is a single overall message conveyed to which

 - 149 -

the two interpretations jointly contribute. Thus, despite the

ambiguity caused by the two interpretations, the pun does not lead

to communication breakdown.

In this section, I have considered the relationship between the

pun and ambiguity. It has been argued that there are different

degrees and types of ambiguity in utterances in general. It has

also been argued that there are different types of puns. In some

cases, the ambiguity of the pun is resolvable, as in (11a), and in

some cases, there is genuine equivocation, i.e. unresolvable

ambiguity, as in (19). In either case, it is essential for a pun to

be successful that the addressee should recover more than one

interpretation. Two or more interpretations were intended by the

speaker, whether some of them are only intended to be triggered and

eventually rejected, or whether they are all intended to be

retained. It could be argued that the essence of the pun lies in

ambiguity.

3.4. Puns and metaphors

It has been argued that interpreting a pun involves recovering

one or more interpretations, after rejecting the most accessible

one. A metaphor similarly involves the recovery of an

interpretation, apparently via some literal interpretation which is

 - 150 -

discarded. Vhat then is the difference between pun and metaphor?

This is the question I shall investigate in this section.

First, let us consider the following caption from an

advertisement for British Rail:

(41) Will you end up at this station?

Underneath this caption is a picture of the blue light typically

seen outside a police station, with the word police on it, and a

second caption urging the audience to get the right ticket before

travelling.

The word 'station' is a pun, used to trigger two separate

interpretations, i.e. the train station and the police station.

Then, (41) makes the audience process the following interpretations:

(42) Will you end up at this train station?

(43) Will you end up at the police station?

 - 151 -

Both interpretations (42) and (43) are intentionally triggered

by the speaker. However.it is mutually manifest that he intended

(42) to be rejected in favour of (43). (43) is the one that the

communicator intended to communicate.

Thus, (41) offers more than one interpretation, namely, (42)

and (43), even though (42) is intended by the speaker to be derived

and then rejected by the addressee. In processing (41), the

addressee would have to search in her memory information about the

word 'station' and derive more than one interpretation, i.e. (42)

and (43) .

Does this make (41) a metaphor? It has been established above

that puns involve the triggering of two or more interpretations,

with the addressee in some cases rejecting one of these, leaving

just one interpretation as the intended one. On the other hand,

metaphors similarly involve the recovery of an interpretation,

apparently via some literal interpretation which is eventually

discarded. There is no reason why the word 'station' should not be

used as a metaphor. If (41) is not a metaphor, what then is the

difference between a pun and a metaphor? This is the question which

I shall investigate in this section.

First, I shall consider Sperber and Wilson's approach to the

analysis of metaphor. Having done this, I shall compare puns and

metaphors and consider what the differences between them are.

 - 152 -

Crucial to Sperber and Wilson's analysis of metaphor (see

1986a) is the notion of interpretive use, which has been discussed

in Chapter 2. Metaphor involves an interpretive relation between

the proposition expressed by an utterance and the proposition of the

thought it represents, perhaps the thought of the speaker. The two

propositions resemble each other, that is they partly share

contextual effects. However, unlike descriptive use, the

proposition of an utterance need not be true of the state of affairs

it is used to represent.

According to Sperber and Wilson, the relevance of a metaphor is

established by recovering an array of contextual implications

(1986a: 236). The range and the strength of these implications are a

matter of degree. On the one hand, in the case of a standardised

metaphor, the addressee would be encouraged to recover a narrow

range of strong implications. On the other hand, in the case of a

novel metaphor, the addressee would be forced to look for a wide

range of weak implications. The following is an example of a fairly

creative metaphor:

(44) He was just a station in my life.

The addressee would have to search through her encyclopaedic

knowledge of 'station' for a number of assumptions about the word.

For, it is false to say that a person is a station and therefore it

 - 153 -

is not possible to derive relevance from a strictly literal construal

t

of (44). The addressee would then hopefully remember that a

station is where one stops for a short time, that it is where only

temporary stops are made, that the stops which one makes at the

station are not important in the journey, and so on. When (44) is

processed against this context, which contains such information, it

will yield a fairly wide range of weak implications as follows:

(45) a. He is just somebody with whom I had a short

affair.

b. The relationship with him was meant to be a casual

affair.

c. The affair with him was not very significant in my

life.

However, the speaker could not have intended to communicate

(45a) - (45c) alone. Had he done so, he would have used an

utterance, such as (46), rather than (44):

(46) He was just someone with whom I had a short, casual

and insignificant affair.

 - 354 -

For (46) would require less processing effort than (44). There are

other contextual effects intended by (44) to outweigh the processing

effort. Perhaps it could be the degree of casualness, which is more

than the addressee might imagine otherwise. Or, the lack of

committment on the part of the speaker might be beyond the

addressee's imagination. Perhaps, the speaker intended to suggest

other characteristics of a station, for instance that a station is

never a destination in itself, and that it is only en route to

somewhere else, which the speaker could not have quite communicated

with (46). Thus, there are contextual implications intended by the

speaker of (44), which could not have been communicated by (46), and

an extra processing effort caused by (44) would be compensated by

them. The relevance of (44) is achieved by inferring weak

contextual implications which are derived around the encyclopaedic

information about 'station'.

Note that in the account of this example, there has been no

mention of first recovering the literal meaning and then the

metaphorical interpretation. Rather, the use of the word 'station',

with its evocation of the contextual implications of being intended

for a short and temporary stop was sufficient to trigger the single

interpretation.

Puns, however, by definition involve processing more than one

interpretation, although nothing precludes one of these being a

metaphorical interpretation. In the case of (41), the addressee

would be given access to two seperate schemata, namely, the train

 - 155 -

station and the police station, even though the access comes from a

single word 'station'. Furthermore, these two schemata have a

different status. The access to 'the train station* is triggered by

the speaker, but it would not lead to the intended message, whereas

'the police station' is the one intended by the speaker. The

speaker has intended to make (42) mutually manifest to the addressee

and himself, but he has also intended it to be rejected by the

addressee, so that she would go on to look for the intended

interpretation that 'station' means the police station.

Thus, the relevance of the pun is achieved by rejecting the

seemingly most accessible interpretation and finding the intended

interpretation. The addressee would have to go though more than one

r

schema triggered by a word, and the success of a pun depends on

having access to those different schemata. The effect of a pun

involves the contrast between the two schemata, that is, that of the

train station and that of the police station in the case of the

example above.

Interpreting a metaphor involes a single schema. Unlike puns,

where the effect comes from the contrast of two or more

interpretations, the effect of a metaphor lies in the condensation,

in the fact that a single expression can offer a wide range of weak

implicatures. This is one of the major differences between a pun

and a metaphor.
 - 156 -

As we have seen above, metaphor is based on resemblance between

the proposition of the utterance used and the proposition of the

speaker's thought. The two propositions share some contextual

effects. However, there is no clear definition as to which

contextual effects they share. In example (44), the speaker intends

to communicate something along the lines of (46), but it cannot be

paraphrased, partly because paraphrasing loses something, and partly

because the message is not so definite. Thus, a metaphor conveys an

indeterminate thought, that is, the speaker intends to communicate a

range of implicatures, rather than a fixed set of them, and the

communication would succeed when the hearer has recovered some of

the inplicatures within the range.

Moreover, in the case of a pun, none of the examples in our

previous section involved indeterminacy of communicative effect.

For example, the speaker of (41) intended to make manifest (42) and

(43), even though he intended (42) to be rejected eventually, and he

did not intend the addressee to look for further interpretations. A

pun necessarily triggers two or more interpretations, which contrast

with each other, but these interpretations are definite, unlike

metaphors which convey an indeterminate thought.

Furthermore, a metaphor communicates a wide range of weak

implicatures, while a pun may communicate as little as a single

implicature, even though additional interpretations may be

triggered, and this intended implicature can be strong. It must be

added that the strength of implicatures communicated by a metaphor

 - 157 -

varies: in the case of a standardized metaphor, the speaker intends

to communicate a narrower range of stronger contextual effects,

whereas in the case of a novel metaphor a wider range of weaker

contextual effects are intended. However, in general, the

contextual effects of a metaphor are relatively weakly communicated.

Thus, the pun and the metaphor are completely different, even

though processing a pun and a metaphor both involve a search for

relevance through a single expression. They are mutually

independent of each other.

However, it must be pointed out that the pun and the metaphor

are not mutually exclusive of each other. That is to say that an

expression can be both a pun and a metaphor at the same time, and in

fact, many puns are also metaphors, as in example (47):

(47) Ve colour pictures. Not facts.

(47) is a caption of an advertisement for The Mirror, and it is

a pun and a metaphor at the same time. The addressee would

eventually recover the following interpretation:

(48) Ve (literally) colour pictures, but we do not

(metaphorically) colour facts.

 - 158 -

Given that it is an advertisement for a newspaper, (48) would have

implications, such as (49) and (50):

(49) Ve paint pictures, but we do not influence the facts.

(50) The Mirror has coloured pictures, and the facts are

presented without distortion.

The 'colour the facts' metaphor would give rise to a wide range of

contexual effects, such as the way the facts are transformed

according to how they are presented, the degree to which they

change, the similarities and contrast between colouring pictures and

transforming iacts, and so on. The relevance of (43) is established

by recovering literal and abstract interpretation of the word

'colour', and a wide range of contextual effects is achieved from

searching for resemblances between painting objects and affecting

facts.

The main purpose of this section has been to compare the pun

with the metaphor. I have argued that, whereas the relevance of

metaphors is established by recovering an indeterminate range of

weak contextual implications which are derived by searching memory

around a single schema, the relevance of puns is achieved by

inferring two or more determinate interpretations which come from

 - 150 -

two or more contrasting schemata. The pun and the metaphor are

mutually independent of each other, although they are not mutually

exclusive of each other.

3.5. Conclusion

Sperber and Wilson say (1986a:217) that style is the

relationship. They argue that the style which the speaker chooses

reveals what kind of relationship he envisages between himself and

his addressee:

From the style of a communication it is possible to infer

such things as what the speaker takes to be the hearer's

cognitive capacities and level of attention, how much help

or guidance she is prepared to give him in processing her

utterance, the degree of complicity between them, their

emotional closeness or distance. In other words, a

speaker not only aims to enlarge the mutual cognitive

environment she shares with the hearer; she also assumes a

certain degree of mutuality, which is indicated, and

sometimes communicated by her style.

(Sperber and Wilson 1986a:217-8)

 - 160 -

For the advertiser, it is crucial to attract the addressee's

attention. For this reason, advertisers take advantage of various

linguistic and non-linguistic devices to attract attention. As I

have argued, the pun is one of them. It functions as an attention-

getting device, enabling the advertiser to make seemingly negative

or implausible statements, which would be highly relevant if true,

for example, 'less bread', 'no j a m 1, 'mind your own business' and so

on.

This might help to raise the addressee's level of attention,

making her more alert, and hence more prepared to work out the

effect of the advertisement. Moreover, as the addressee has to

reject the most obvious interpretation and infer the intended

interpretation, it causes her an extra processing effort. This

might help to sustain her attention for longer. Moreover, it might

lead to higher memorability. These aspects would help to involve

the addressee in the advertisement.

The wit and humour of the pun might also contribute to

congenial feelings towards the advertiser. In spite of what

literary critics and advertising agencies have said, there is some

indication that the general public do enjoy puns. Deciphering a pun

might create in the addressee a self-congratulatory thought. This

pleasure would help the addressee form favourable feelings about the

advertiser. Having said that the advertiser cannot expect a high

level of attention from the addressee, a large number of

advertisements using puns have been found in places such as stations

 - 163 -

and trains, where the advertiser has not only a captive audience but

also a bored audience, who might appreciate entertainment. By

offering an amusing pun at a tube station, the advertiser is

providing the addressee with some entertainment which she wants.

There is, perhaps, more to this humorous and aesthetic effect

of the pun than Relevance Theory is capable of analysing, as it now

stands. Relevance is defined and measured in terms of contextual

effects and processing effort. However, it is possible that a pun

provides a kind of humorous pleasure and aesthetic effect which are

not fully analysable in terms of contextual effect alone. For

example, in the case of the airline advertisement, the advertiser

uses a pun to refer to the girl's breasts. Now, there is a problem

in accepting that the message is relevant; it is possible to argue

that the message is not relevant but provides some aesthetic

pleasure to the addressee. Then, one could argue that the pun ^

raises a problem for Relevance Theory in its present stage of ^

elaboration. Further research on humour and aesthetics in the

Relevance framework is awaited. I shall return to this from a

different viewpoint, using the same airline advertisment as an

example, in Chapter 4.

As I have shown above, the relevance of the pun is generally

established by recovering two or more interpretations. The pun

presents more than one interpretation, even though only one

interpretation is intended by the speaker. This would give the

addressee access to different schemata, from which she would derive

 - 262 -

various contextual implications. It has also been argued that some

puns are equivocal. However, it is clear that there is nothing

equivocal about the ultimate message advertisements are conveying,

namely, 'Buy this product!1.

flow, the advertiser does not aim to send his ultimate message

by directly communicating this single message: rather, he would like

to communicate an 'impression' or 'impressions' of the product or

the company. In the Relevance framework, an impression is seen as a

change in one's cognitive environment, which results from relatively

minor alterations in the manifestness of a large number of

assumptions, instead of major alterations of a few assumptions

(Sperber and Wilson 1986a:59). The pun could help achieve this

effect. For example, in the case of the Mazda car advertisement,

the pun creates two separate impressions: one is the fact that it is

a pun, hence witty, thus creating a feeling of complicity and

cleverness in the audience; the other, through the 'long driveway'

interpretation, allows access to assumptions about the lifestyle of

the car's drivers, generating a certain impression of the car

itself.

Thus, a pun can provide the advertiser with a way of

communicating two or more impressions; one by the very fact that it

is a pun, and the other (s) by its intended interpretation(s). If

his advertisement manages to create an impression in his addressee,

he would consider the advertisement successful. This must explain

why there are so many puns in advertisements.

 - 163 -

Chapter 4: Covert Communication

4.1. Introduction

Crystal and Davy (1983:222) have argued that two of the main

functions of the language of advertising are to inform and to

persuade. However, these two functions are not of equal importance:

there is a hierarchy between them. The advertiser does not inform

for the sake of improving his viewers' knowledge about the world; he

informs in order to persuade them to buy the product which he is

advertising. He would be content if he managed to persuade them but

failed to inform them of something (if that was possible), but not

vice versa. It is certainly the ultimate goal of the advertiser to

persuade his viewers to buy a product. In this chapter, I

investigate how the process of communication is manipulated in

language to achieve this goal and I try to show how Relevance Theory

provides a useful framework for the analysis.

In Section 4.2, I examine in detail Sperber and Wilson's

account of ostensive communication. In Section 4.3, I discuss

examples in advertising which appear not to satisfy the requirements

of ostensive communication, thus providing apparent counter-examples

to their account of communication. Section 3.4 discusses the notion

of ostention and the role it plays in communication. In Section

4.5, I examine weak implicatures and how they differ from covert
 - 164 -

communication. Finally, in Section 4.6, I show how Relevance Theory

can analyse covert forms of communication.

4.2. Sperber and Wilson's ostensive communication

Sperber and Wilson's theory of utterance interpretation is

based upon the definition of 'ostensive-inferential communication'

repeated here for convenience:

Ostensive-inferential communication: the communicator

produces a stimulus which makes it mutually manifest to

communicator and audience that the communicator intends,

by means of this stimulus, to make manifest or more

manifest to the audience a set of assumptions <I).

(Sperber and Wilson 1986a:63)

Information is manifest to an individual at a certain time if

the individual is capable at that time of representing it and

accepting that representation as true or probably true. Information

is mutually manifest to two individuals if it is manifest to both,

and it is manifest to both that it is manifest to both, and so on

 - 165 -

indefinitely. Moreover, manifestness is a matter of degree: the

more likely the individual is to construct a certain representation,

the more manifest it is to him.

According to this analysis, ostensive communication involves

two layers of intention; a first-order intention to inform the

hearer of something, which Sperber and Vilson call the informative

intention:

Informative intention: to make manifest or more manifest

to the audience a set of assumptions -(I).

(Sperber and Vilson 1986a:58)

and a second-order intention, the communicative intention to make

that informative intention mutually manifest:

Communicative intention: to make it mutually manifest to

audience and communicator that the communicator has this

informative intention.

(Sperber and Vilson 1986a:61)

 - 166 -

There is thus in Sperber and Wilson's definition of ostensive

communication a requirement of overtness, to the effect that the

speaker's informative intention must be made mutually manifest.

This requirement of overtness is crucial for Sperber and Wilson's

theory. It is the assumption that the informative intention is

being made mutually manifest which helps the hearer to recover the

intended message - that is, the set of assumptions <I) which the

speaker intended to make manifest, or more manifest, to the hearer.

Only if the speaker's informative intention becomes mutually

manifest does ostensive communication take place.

When this requirement of overtness is not fulfilled, that is,

when the informative intention is not made mutually manifest,

Sperber and Vilson believe that this is not a genuine instance of

ostensive communication. They give the following example of 'non-

ostensive' non-verbal communication:

Suppose, for instance, that Mary wants Peter to mend

her broken hair-drier, but does not want to ask him

openly. Vhat she does is begin to take her hair-drier to

pieces and leave the pieces lying around as if she were in

the process of mending it. She does not expect Peter to

be taken in by this staging; in fact, if he really

believed that she was in the process of mending her hair­

drier herself, he would probably not interfere. She does

expect him to be clever enough to work out that this is a

 - 167 -

staging intended to inform him of the fact that she needs

some help with her hair-drier. However, she does not

expect him to reason along just these lines. Since she is

not really asking, if Peter fails to help, it will not

really count as a refusal either.

(Sperber and Vilson 1986a:30)

Sperber and Vilson argue (1986a:30) that there is 'an intuitive

reluctance to say that Mary "meant" that she wanted Peter's help, or

that she was "communicating" with Peter in the sense we are trying

to characterise', and that this reluctance is 'well-founded' and

comes from the fact that Mary's second-order intention to have her

first-order informative intention recognised is hidden from Peter.

I shall investigate cases of verbal communication which, it

will be argued, are parallel to Sperber and Wilson's example of non­

verbal, non-ostensive communication. The language used in

advertising often exhibits examples of 'covert forms of information

transmission' (Sperber and Vilson 1986a:30) .

4.3. Non-ostensive communication in advertising

 - 168 -

Let us look at the example of All Nippon Airlines again. I

repeat it for thesake of convenience: (

(1) Oo Kli NAa VAh.

As I have pointed out in Chapter 3, there are the following two

possible readings:

(2) Okinawa

(3) How big! Wow!

The question which (3) raises is: what is big? The problem

here is that there are at least two possible candidates for the

object described as ookii (big). That is, the Okinawa Islands,

which are the destination of the flight promoted in the

advertisement, and the big breasts of the girl shown in the

illustration. The second interpretation would lead to (4):

(4) In those islands, there are a lot of beautiful girls

with big breasts.

 - 169 -

However, though the advertiser may have intended (4) to be

manifest to the audince, he surely did not intend (4) to be manifest

on the basis of making this informative intention mutually manifest.

Thus, the communicator's intention to convey an interpretation is

sometimes made mutually manifest, and in others it is not.

Nudity is shown to a greater extent in advertising and in mass

media in general in Japan than in the U.K. Even so, to show a well-

endowed girl in a bikini and explicitly say, 'Vhat big breasts she

has!1 would be regarded as irrelevant and vulgar in an advertisement

for an airline company. Since there is an alternative

interpretation, the advertiser's intention to refer to the girl is

not made as manifest as it could have been; it may not become

mutually manifest at all. It is vague, which leaves some room for

the advertiser to deny the intention if necessary.

There are some more examples of this sort. (5) is the caption

of an advertisement for men's toiletries, the brand name of which is

Tech 21, for which I have provided the translation in (6):

(5) Otoko wa seinoo.

man topic capacity

(6) For a man, it is capacity (which is desirable).

 - 170 -

Considering that it is an advertisement, the following premise must

be added to the context:

(7) The product advertised helps the buyer attain what is

desirable.

When assumptions (6) and (7) are combined, the conclusion in (8) can

be recovered by the audience:

(8) The product helps the buyer attain his capacity.

The word seinoo (capacity) refers to functional capacity and is

usually used to describe machinery, such as a car, a camera, etc.

It is not a common collocation to use the word for a description of

people. However, the brand name, Tech 21, has some overtones of

machinery, and it may be the company's principle, not the

advertiser's, to advertise it in association with machinery.

Perhaps, after all, it is not just women who are treated as sex

machines in advertising.

Two Chinese characters are used to transcribe the word seinoo,

each with its own meaning, that is, sei (sex) and noo (ability).

Thus, the following interpretation is possible:

 - 171 -

(9) Otoko wa sei noo.

man topic sex ability

which can be translated into English as follows:

(10) For a man, it is sexual ability (which is desirable).

This interpretation is reinforced by the illustration, which shows

the face and torso of a young Western man with long hair (arms not

shown).

When (10) is processed in a context which contains assumptions

(7) and (8), the conclusion in (11) can be recovered:

(11) The product helps the buyer attain his sexual

ability.

However, it could hardly be argued that the advertiser has

communicated (9), and consequently (10) and (11), by making mutually

manifest his intention to do so. Seinoo (capacity) is one word,

even though it is transcribed in two separate Chinese characters and

therefore can be broken down into two words, namely, sei (sex) and
 - 172-

noo (ability). The advertiser could not be said to have overtly

communicated (9), (10) and (11).

It is not surprising that an advertiser, promoting men's

toiletries, should want to associate the product with sexuality,

especially when one recalls that homosexuals consume far more

toiletries than heterosexuals, and are therefore a potential target

group for advertisers. In fact, there are in the illustration some

overtones of a picture from a 'gay' magazine. However, at the same

time, it is reasonable that the advertiser should not want

explicitly to make an association with sexuality, for that might be

regarded as vulgar and distasteful, particularly as men's toiletries

are often purchased by women as a gift, and such women might be put

off by an obvious association between toiletries and sexual ability.

The next examples are both captions, advertising the same

miniature television set:

(12) Ki-tai mono o gamansuru to ningen

watch-want to thing o b j . forbear if personality

ga chijimimasu.

topic shrink

If you go without watching what you want to watch,

your personality will shrink.

 - 173 -

(13) Mi-tai mono o gamansuru to jinsei

watch-want to thing ob j. forbear if life

kuraku narimasu.

depressing become

If you go without watching what you want to watch,

your life will be depressing.

These captions are accompanied by somewhat similar illustrations.

C12) is shown with three young Western people, two girls dominating

the picture, intimately holding each other. A boy is holding one of

the girls from behind with one hand, and he has a miniature

television set in the other, with the faint suggestion of an orgy.

Caption (13) is illustrated by two Western girls intimately close to

one another, with one girl dominant and holding the other from

behind. The girl who is being held is playing the piano, on the

keys of which is placed a miniature television set. There are strong

overtones of a lesbian relationship, and a sense of forbidden

romance.

There is also some text attached to these captions. (14) and

(15) accompany (12) and (13), respectively:

 - 174 -

(14) Jinrui batten no rekishi, kore sunawachi kookishin

juusoku no rekishi. 'Are mo mitai, kore mo mitai' no

yajiuma-konjoo koso, ningen-seichoo no gendooryoku-na

no de arimasu. . .

(15) Are-kore to, mamanaranu-koto no ooi kono yononaka.

Semete, mi-tai mono gurai jiyuu ni mire-nakya, mani

ga tanoshimi de ikite-iru no ka wakara-naku narimasu

yo ne. . .

I have provided the following translations for these two texts:

(16) The history of the development of mankind is the

history of the satisfaction of curiosity. 'Want to

see this. Want to see that...' The curiosity which

drives a mob is the source of energy which has

driven mankind to grow this far ...

(17) Can't do this, can't do that... There are many

things forbidden in this world. What's the point of

living unless we can at least watch what we want to

when we want to ?. ..
 - 175 -

The caption and text are closely related. (12) and (14) both

encourage curiosity. (12) says that if you give up your curiosity

you will 'shrink', and (14) says that it is curiosity which helps

people grow. There is a contrast between chijimu (shrink) and

seichoo (growth). The word chijimu (shrink) is normally used to

describe an inanimate object, and it is unusual for people to be

said to 'shrink'. The word has a slight hint of the lack of

erection. Thus, (12) may be interpreted as follows:

(18) If you refrain from watching what you want to watch,

your penis will shrink.

or, to the extent that the illustration suggests an orgy;

(19) If you refrain from having an orgy, your penis will

shrink.

Both (13) and (15) are about taboos; (13) says that life is

depressing when one goes without the forbidden, and (15) asks what

is the point of living if we are to go without what is forbidden.

With the illustration suggesting a lesbian relationship, which is

taboo, the following interpretation is possible:

 - 176 -

(20) If you go without a lesbian relationship, your life

will be depressing.

Moreover, there is a caption, written in English, found in both

advertisements:

(21) Anytime OK! Everywhere OK!

Vhat is OK anytime and everywhere is ambivalent. Considering that

it is an advertisement for a miniature television set and that the

caption is apparently talking about not being able to watch

something, the following interpretation is possible:

(22) With our television set you can watch television

anywhere at any time.

On the other hand, in a context in which (18) and (19), or (20) are

present, (23) and (24) may be recovered, respectively.

(23) You can have an orgy anywhere at any time.

 - 177 -

(24) You can have lesbian sex anywhere at any time.

However, there is an intuitive reluctance to say that the

advertiser has overtly communicated (18), (19), (20), (22), (23) and

(24) by means of (12), (13), (14), (15), and (21). (23) and (24)

are inconsistent with the fact that they are advertisements and that

they are advertising a miniature television set. They are

advertising neither an orgy nor a lesbian relationship, and there

seems to be no reason why an advertisement for a television set

should be encouraging an orgy or a lesbian relationship, which are

not only irrelevant to the advertisement, but are also taboo in

Japanese society. Even though it is known that sex is used in

advertising in order to catch the attention and provide stimulation

(Key 1973), the discussion goes onto a 'subliminal* level, and it

does not apply to ostensive communication. It cannot be argued that

the advertiser has intended to inform his audience of (18), (19),

(20), (22), (23) and (24) by making mutually manifest to himself and

his audience his intention to do so.

In some forms of communication, such as the one we have just

seen, although the speaker has an informative intention - he intends

to make certain information manifest - he does not intend to make

that informative intention mutually manifest to his hearer and

himself. In advertising, and indeed in communication in general,

the success of the communication, i.e. the recovery by the hearer of

the intended information, is not always achieved by making the

informative intention mutually manifest. In some cases, as I shall

show, the fulfilment of the speaker's informative intention can

actually best be achieved if the speaker's communicative intention,

i.e. his intention to make his informative intention mutually

manifest, is concealed from his hearer.

The cases just described would be considered by Sperber and

Wilson as cases of 'information transmission* but not ones of

'ostensive communication' governed by the principle of relevance.

Sperber and Wilson argue (1986a:2) that there is no general theory

of communication, and they propose a theory specifically to account

for ostensive communication.

According to Sperber and Wilson (1986a:158), the principle of

relevance only applies to ostensive communication, where the

speaker, by his linguistic behaviour, intends to affect the mutual

cognitive environment of himself and the hearer, i.e. the set of

assumptions mutually manifest to both the hearer and the speaker.

When the mutual cognitive environment of the hearer and the speaker

is affected, i.e. when the speaker's informative intention is made

mutually manifest, further assumptions are made mutually manifest to

both interlocutors. As a result, their social relationship is

affected, in particular their possibility of future linguistic or

non-linguistic interaction.

In a case of information transmission, on the other hand, the

mutual cognitive environment of the speaker and the hearer is not

 - 179 -

affected. Information transmission only affects the cognitive

environment of the audience. A set of assumptions (I) become

manifest to the hearer; the speaker's informative intention may

become manifest, too, but it does not become mutually manifest to

the hearer and the speaker. As a result, no new social relation is

created between the speaker and the hearer, that is, their mutual

cognitive environment is not affected. This may itself be a reason

for the advertiser to engage in these forms of communication. I

shall call these forms of communication 'covert communication', as

opposed to ostensive communication.

The advertisement captions, discussed above, can be explained

along these lines. The advertiser of the Japanese airline intended

to make manifest to the audience assumption (4), but he has avoided

a possible accusation of pornography or a protest from feminists by

not making that informative intention mutually manifest to himself

and to his audience. Similarly, the advertiser of Tech 21 intended

to make manifest to the audience assumptions (10) and (11), but has

avoided negative reaction to explicit statements about sexuality.

The advertiser of the miniature television set intended to make

manifest to the audience assumptions (15), (16) and (17), but he has

not made his informative intention mutually manifest to the audience

and himself, for those assumptions are apparently inconsistent with

the principle of relevance and they may be offensive to viewers.

 - 180 -

The advertiser of the television set seems to have calculated

that his advertisements would not be too offensive to his audience,

as they both appeared in a Japanese magazine equivalent to the

supplement to The News of the World. It is worth noting at this

point that a different advertisement for the same television set

appeared in a magazine equivalent to the supplement to The Sunday

Times around the same time in 1986, showing four people: a middle-

aged woman, apparently a housewife, a schoolboy with a baseball cap,

a young woman with a magazine, and a young boy, apparently a

student. Apart from the young woman holding her magazine and

looking at the camera/ audience, everybody else is watching the

miniature television set, with caption (25):

(25) Chikagoro, tsuukin-densha ga shumi desu.

recently commuting-train topic hobby is

Recently the commuting-train has become my hobby.

Because everybody, except for the girl looking at the audience, is

watching television, a further interpretation can be derived:

(26) Recently, watching television on the commuter train

has become my hobby.

 - 181 -

There is no trace of sexual suggestion either in the writing

or in the picture. The English caption (21) has been replaced by

the following Japanese caption:

(27) Itsu-demo doko-demo TV taimu.

when-even where-even TV time

TV-time, anytime, anywhere.

where 'TV' is written in the English alphabet, and taimu (time) in

katakana characters, which are used for foreign loan words. From

this, interpretation (28) can be derived:

(28) You can watch television anywhere at any time.

In (27), the topic, television time, is specified and there is

no other possible topic. In the earlier advertisements, the topic

in the English caption is not specified and it is left ambiguous as

to what 'O.K.' might refer to. Considering that it is an

advertisement for a miniature television set, watching television is

an obvious candidate for being 'O.K.', but the pictures suggest an

orgy and a lesbian relationship. Thus, the audience are left with

two possibilities.
 - 182 -

These advertisers intended to affect the cognitive environment

oi their audience by making them recover assumptions (4), (10),

(11), (18), (19), (20), (22), (23) and (24), and yet they avoided

any modification of the mutual cognitive environment of the audience

and themselves by failing to make this intention mutually manifest

to themselves and to the audience. By this sort of manipulation,

the advertiser has avoided the social consequences of making his

intention mutually manifest to his audience and himself, and he has

allowed himself the possibility of denying having had that intention

in the first place.

A cognitive environment is a set of assumptions that are

manifest to an individual. A mutual cognitive environment is a set

of assumptions that are mutually manifest to two or more

individuals. Sharing a mutual cognitive environment is the basis

for cooperation between 'equal' interlocutors.

Mutual manifestness may be of little cognitive importance,

but it is of crucial social importance. A change in the

mutual cognitive environment of two people is a change in

their possibility of interaction.

(Sperber and Wilson 1986a:61-62)

However, in many cases of verbal exchange in advertising, and in

communication in general, although the speaker intends to transmit

some information and intends to make his addressee recognise that

information, and perhaps the intention behind it, he does not want

to affect the mutual cognitive environment that he shares with his

addressee, precisely because he wants to avoid the social

implications that such a modification of the mutual cognitive

environment necessarily brings with it.

In the Okinawa case above, the speaker does not make his

informative intention mutually manifest to himself and the addressee

because of the possible social consequences. The advertiser intends

to inform the addressee of (4), but making his informative intention

mutually manifest may have unwanted side effects, such as provoking

accusations of pornography or sexism. The fulfilment of his

informative intention, that is, the recovery of the intended

information by the addressee, is not helped by his obscuring his

informative intention. However, the picture of the well-endowed

girl in a bikini is sufficient to help the addressee to get the

message. That is why, even though some of my informants were not

sure if the association between the word ookii (big) and the girl's

breasts was intended by the advertiser, the advertiser's intention

was achieved, because these audience could not help making the

connection between the two.

In the men's toiletries advertisement, the advertiser does not

make his informative intention mutually manifest to himself and his

 - 184 -

addressee, because he wishes to avoid being accused of vulgarity and

bad taste, or of encouraging homosexuality. The advertiser intends

to inform the addressee of (10) and (11), but wishes to avoid the

unwelcome social consequences it may bring to make this intention

mutually manifest to the addressee and himself. He hopes that the

addressee will be able to recover (10) and (11) on their own, with

the help of the illustration.

In the case of the miniature television set advertisements in

the Japanese equivalent of The News of the World, the advertiser

intends to inform the viewer of (18), (19), (20), (22), (23) and

(.24), but does not intend to make his informative intention mutually

manifest, in case doing so may cause some unwanted reaction, such as

accusations of pornography or dislike of the advertiser. He has

chosen a Japanese equivalent of The News of the World for his

advertisements, as the readers of such a paper are unlikely to have

negative reactions to these advertisements, but he is still careful

not to make his informative intention mutually manifest and not to

take entire responsibility for these interpretations.

The addressee who recovers interpretations (18), (19), (20),

(22), (23) and (24) has to take a larger share of the responsibility

than usual, being helped by the illustrations. When these

interpretations are recovered without the addressee responding

negatively to them, it may even create a feeling of solidarity with

the advertiser for being daring and interesting. For the Japanese

equivalent of The Sunday Times, the advertiser has chosen a

 - 185 -

completely different version, for he is aware that the readers are

more likely to find the advertisements in the other magazine

unacceptable. The readers of the Japanese equivalent of The Sunday

Times would not have felt solidarity with an advertiser showing

sexually suggestive illustration. They would have been dismayed, if

not appalled.

As we have seen in this section, there are examples in

advertising, in which advertisers attempt to communicate something

without making their informative intention mutually manifest to the

audience and themselves. In contrast, ostensive communication,

which is defined in terms of the principle of relevance, requires

some overtness: the speaker's informative intention must be made

mutually manifest to the audience and the speaker himself.

Is covert communication no longer a case of communication? Is

it a deviation from the norm not to have the speaker's informative

intention made mutually manifest? The issue here is not just a

terminological one. As was pointed out above, Sperber and Vilson

maintain that there can be no general theory of communication, and

propose a theory to account for a specific type of communication,

i.e. ostensive-inierential communication, which is defined in terms

of the principle of relevance. The interesting questions are: how

can one explain covert communication or information transmission?

Can Relevance Theory provide a useful framework for this process,

and if so how? These are the questions I shall consider in the

following sections.
 - 186 -

4.4. Communication and ostension

Sperber and Wilson argue that it is not hard to see how the

fulfilment of the communicative intention, that is, the

communicator's intention to make it mutually manifest to both the

audience and himself that he intends to inform the audience of

something, can lead to the fulfilment of his informative intention:

...the realisation that a trustworthy communicator intends

to make you believe something is an excellent reason for

believing it.

(Sperber and Wilson 1986a:163)

As it stands, the argument is sound. However, it raises the

quesion: who is a trustworthy communicator? It is idealistic and

unrealistic to assume that communication always takes place between

a trustworthy communicator and a trusting addressee. Let us look at

Sperber and Wilson's argument again:

There are situations in which the mere fact that an

intention is recognised may lead to its fulfilment.

 Suppose that Mary intends to please Peter. If Peter

becomes aware of her intention to please him, this may in

itself be enough to please him. Similarly, when the

inmates of a prison recognise their warden's intention to

make them fear him, this may be enough in itself to make

them fear him. There is one type of intention for which

this possibility, rather than being exceptional, is

regularly exploited: intentions to inform are quite

generally fulfilled by being made recognisable.

(Sperber and Wilson 1986a:21-22)

Although it is not stated explicitly, in Sperber and Wilson's

example, Mary ar\d Peter are presented as a couple living together,

apparently on equal terms, and are portrayed as a trustworthy

communicator and a trusting addressee. In this case, the fact that

the communicator intends to make the addressee believe something is

an excellent reason for the addressee to believe it. There is

underlying 'cooperation' between the interlocutors which justifies

the addressee in believing something, on the grounds that the

communicator intends to make the addressee believe it. In the case

of a prison warder and inmates, the warder is in a position where he

could make the prisoners' life difficult, if he were to set his mind

to it. Again, this is a case where, for the addressee to believe

something, it is sufficient that the communicator should intend to

communicate it. The communicator is clearly in possession of a

 - 188 -

social power which enables him to achieve his intention over his

addressee, regardless of the addressee's reaction.

Let us look at some examples. Suppose that Peter has said to

Mary (29):

(29) I'll make a meal because you are tired.

In this case, the mere fact that Mary has intended to make Peter

believe (.29) is an excellent reason for her to believe it.

Or, let us suppose a prison warder has said (30) to his

prisoners:

(30) You'd better be cooperative, or you may find life

here unpleasant.

It is a good enough reason for the prisoners to believe (30) that he

has intended to make them believe it.

However, this does not mean to say that these cases are the

norm and that anything that does not work in the same way is a

deviation from normal communication. The point which Sperber and

 - 169 -

Vilson make, as illustrated in these examples, is that the strength

with which the communicated assumptions will be entertained is

commensurate with the hearer's trust in the speaker. If the hearer

does not trust the speaker, she would not believe what he says; if

she trusts him a little, she will believe it a little; and if she

has a great trust in him, she will believe it to a great extent.

Moreover, the speaker can not only satisfy the hearer's expectations

he can also exploit them.

However, as Sperber and Vilson are aware, there are cases in

which the assumptions about a trustworthy speaker and a trusting

hearer do not hold. For example, when a double-glazing salesman

intends to make his customer believe that she would benefit from his

services, it is unlikely that the customer would regard the fact

that the salesman intends to make her believe it as a good reason to

believe it. Similarly, neither side in a court case, whether

plaintiff or defendant, would accept as a good reason to believe

something the mere fact that the communicator intends to make the

audience believe it. Suppose that (31) has been said by a double-

glazing salesman, and (32) by a defendant in a court:

(31) You would find the room much quieter.

(32) I am innocent.
 - 190 -

The customer and the jury would not regard it as a good reason to

believe (31) and (32), respectively, that the salesman and the

defendant had intended to make them believe these statements. It is

obvious that (31) and (32) have been said for the respective speaker

to achieve his own goals. Each speaker might have said (31) or (32)

even if these statements were inconsistent with the truth, that is,

not worth the hearer's while to believe.

These cases, where it is not a sufficient reason for the

addressee to believe something that the communicator intends to make

her believe it, are not exceptional. The kinds of communication

which take place in advertising, political debate, and academic

discussion, for example, all seem to belong to this category.

Indeed, even in our daily social conversation, these cases are rife.

Suppose someone is trying to impress you, the realisation that the

communicator intends to make you believe something would not be a

good reason for you to believe it. On the contrary, it is likely

that it would prove to be an excellent reason for you not to believe

it.

As Sperber and Vilson show, the hearer can recover the set of

assumptions (I) intended by the speaker without actually believing

them. As argued above, the strength which the hearer assigns to

these assumptions depends on her confidence in the speaker. All

this applies to ostensive-inferential communication. The hearer has

to firstly recognise the informative intention, and secondly decide

whether she believes what she is being told. The success of

 - 191 -

communication has been defined as the hearer recovering the

speaker's informative intention, and not as the speaker making the

hearer believe something.

Communication can take place between interlocutors who trust

each other, hold this trust to be mutual, and strive towards the

same goal, that is, the enlargement of the mutual cognitive

environment, as in the case of Peter trying to please Mary. It is

mutually manifest that there is social cooperation between them.

Communication may also occur between socially unequal

interlocutors, as in the case of the prison warder frightening his

inmates. The warder possesses a power over the inmates which

enables him to do things to them regardless of their will, or even

against their will. The inmates cannot but take the warder's

communicative intention seriously for their own interest. They are

forced to cooperate with the warder.

The case of the double-glazing salesman persuading his customer

is different again from either category just mentioned, except that

it has in common with the case of the prison warder and his

prisoners the fact that communication takes place between unequal

interlocutors. Here the addressee does not necessarily regard the

communicator as trustworthy, and the communicator intends to inform

the addressee of something without her trusting him, or even despite

her distrusting him. It is mutually manifest that the intended

effect of the communication is for the communicator's benefit, and

 - 192 -

not necessarily for the addressee's, even though the communicator

may claim otherwise. It is the addressee who has a power over the

communicator. The communicator cannot afford to rely on the

addressee's cooperation.

This is not to argue that the hearer's trust and co-operation

are necessary for ostensive communication. Vhen it exists, it is

just one contextual factor among others. However, it does help the

fixation of belief, which is presumably why people engage in

communication. The point is that there is a social aspect to be

taken into consideration when studying communication.

It must be pointed out here that the language of advertising is

a language of unequal interlocutors, unequal in the sense that the

speaker and the hearer are not on equal terms. They are not a

trustworthy speaker and a trusting hearer, who cooperate with each

other, and strive for the same end. It is like the type of

communication which takes place between a double-glazing salesman

and his customer, or a defendant and the jury, rather than that

between a prison warder and his prisoners. It is manifest to both

the advertiser and his addressee that he is saying something because

he wants her to buy the product or service for his own benefit, and jy-y^^c

not because he genuinely believes that she would benefit from the

product or the service. It is not a good reason for the addressee

to believe something that the advertiser has intended to make her

believe it. The advertiser must make his addressee believe

 1 r»o _

something without her trusting in him, or, indeed, despite her

distrusting him.

At this point, it is worth distinguishing 'cooperation' at the

social level from 'cooperation' at what one can call 'the cognitive

level'. Cooperation at the cognitive level is always present when

communication of any kind takes place, and it is a prerequisite for

it to happen. Even when war is declared against an enemy, the enemy

pays attention to the stimulus the communicator is using, the enemy

processes it, and then communication becomes successful. The two

sides may not be cooperative with each other socially, but they have

cooperation at the cognitive level.

There are cases of communication in which there is cooperation

between the communicator and the addressee, not only at the

cognitive level, but also at 'the social level', as in the case of

Peter offering Mary to make a meal. Sometimes a communicator can

rely on the fact that he is regarded by the addressee as a

trustworthy communicator, that she is a trusting addressee, and that

it is an excellent reason for her to believe something that he

intends to make her believe.

Sometimes a communicator can depend on the adressee's

cooperation at both cognitive and social levels, but not intend to

bear responsibility for the social consequences which result from

the communication, as we have seen in the case of Mary getting Peter

to mend her hair-drier. The communicator can rely on both .cognitive

 - 194 -

and social cooperation from the addressee, even if the communicator

does not reciprocate by being willing to modify the mutual cognitive

environment, as in this case.

Furthermore, there can also be a communicator who does not

expect cooperation at the social level. He has to aim at the

intended effect by means of his stimulus, and not by means of the

addressee's willingness to cooperate. The double-glazing salesman

is an example of this category. All these communicators have their

addressees' cooperation at the cognitive level, or otherwise

communication does not happen at all, but it is cooperation at the

social level which varies depending on the situation.

This distinction between the cognitive and the social level in

cooperation between communicator and addressee is crucial to my

analysis of advertising. The advertiser does not take for granted

that he has his addressee's cooperation at either level. Firstly,

he works at drawing the addressee's attention and thus at obtaining

her cooperation at the cognitive level, by using various stimuli,

not only linguistic but also audiovisual. Secondly, even when he

has managed to gain the addressee's cognitive cooperation, he cannot

expect her cooperation at the social level. It is mutually manifest

to the advertiser and to the addressee that they are not equal

partners engaged in communication and striving to reach the same

goal. The addressee would not treat the fact that the advertiser

intends to make her believe something as a good reason for

believing it. On the contrary, she might even regard it as an

 - 195 -

excellent reason to disbelieve it. Thecommunicator has to resort

to other ways of making his addressee believe his message.

In ostensive communication, the speaker provides a guarantee of

optimal relevance, and asks for the hearer's cooperation at a

cognitive level. However, as Sperber and Vilson point out, the

speaker needs not live up to the guarantee of relevance: he may give

it in bad faith. So, the cognitive co-operation of the speaker is

not required even for ostensive communication to take place.

Relevance Theory can provide a useful framework to analyse

ostensive communication in terms of the two layers of intentions,

namely, the informative intention and the communicative intention,

and the notion of mutual manifestness. I will repeat here the

definitions of informative and communicative intentions for the sake

of convenience:

Informative intention: to make manifest or more manifest

to the audience a set of assumptions II) .

(Sperber and Vilson 1986a:58)

Communicative intention: to make it mutually manifest to

audience and communicator that the communicator has this

informative intention.
 - 196 -

(Sperber and Vilson 1986a:61)

The goal of communication can be described as the fulfilment of

the communicator's informative intention. However, communication

may succeed without the communicator's informative intention being

fulfilled, for it is possible that the addressee disbelieves or

disagrees with assumptions which the communicator intends to make

manifest or more manifest to her. The communicator's informative

intention can be recognised without its being fulfilled. All the

communicator can do is to hope that the addressee will believe what

is intended to be made manifest or more manifest to her by the

communicator's stimulus.

In ostensive communication, the communicator hopes to increase

his chances of success in communication, that is, of his informative

intention being fulfilled, by means of his communicative Intention,

i.e. by making his informative intention mutually manifest to both

the addressee and himself. He hopes that the fact that he intends

to inform his addressee of something will be a good enough reason

for her to believe it.

Vhat happens, then, after the communicator's informative

intention has been made mutually manifest, is described by Sperber

and Vilson as, follows:

 - 197 -

By making her informative intention mutually manifest,

the communicator creates the following situation: it

becomes mutually manifest that the fulfilment of her

informative intention is, so to speak, in the hands of

the audience; if the assumptions that she intends to

make manifest to the audience become manifest, then she

is successful; if the audience refuses to accept these

assumptions as true or probably true, then she has

failed in her informative intention.

(Sperber and Vilson 1986a:62)

Whether the communicator's informative intention is fulfilled or

not, in ostensive-communication, the communicator's communicative

intention is, by definition, fulfilled. This then alters the

mutual cognitive environment of the communicator and the addressee,

regardless of the fulfilment of the communicator's informative

intention. In other words, the mutual cognitive environment is

affected whether or not the communicator achieves the fulfil-ment of

his informative intention.

Vhat is crucially involved in ostensive communication is the

alteration of the mutual cognitive environment between the

communicator and the addressee, and this inevitably causes the

alteration of the social relationship between them. By engaging in

ostensive communication, the communicator leaves the fulfilment of

his informative intention 'in the hands of' the addressee, as

Sperber and Vilson put it (1986a:62). The addressee can accept it

or reject it. Either way, this creates, as a result, some change in

the relationship between the communicator and the addressee.

I have examined situations, such as the case of Mary getting

Peter to mend her hail— drier, in which the communicator makes his

informative intention manifest to the addressee, without making it

mutually manifest to both the addressee and the communicator

himself. This can be put in the following way: the communicator may

intend to have his informative intention fulfilled without reference

to his communicative intention. In all the relevant cases I have

referred to above, the communicator uses other means of

communication to help, such as the display of a hair-drier being set

apart, or the illustrations used by the advertisers. Perhaps, these

non-linguistic stimuli act as an aid to the fulfilment of his

informative intention, which could be helped with the communicator's

communicative intention in the case of ostensive communication.

Here, the communicator aims at the fulfilment of his informative

intention without having a communicative intention of the type

described above. He intends to make his informative intention

manifest to the addressee, in other words, to have his informative

intention recognised by the addressee, but not to make it mutually

manifest to the addressee and himself: no communicative intention is

formed. His motivation for doing this is to avoid any alteration of

the mutual cognitive environment of the addressee and himself, and

 - 109 -

thus any social consequences the alteration may bring with it.

Having recognised the communicator's informative intention to make a

set of assumptions manifest, the addressee is given a choice of

accepting these assumptions as true or probably true. However,

because the communicator's informative intention is not made

mutually manifest, whether the addressee accepts these assumptions

as true or not, the mutual cognitive environment of the communicator

and the addressee stays unchanged. Consequently, the communicator

is saved from any social implications that may result if his

informative intention is made mutually manifest to the addressee and

himself.

In the case of the double-glazing salesman persuading his

customer, the communicator intends to fulfil his informative

intention without making this intention manifest to his addressee.

In other words, the communicator may intend to make manifest or more

manifest a set of assumptions, concealing his informative intention.

In this case, the reason for the communicator to engage in such

communication is that he is aware that revealing his informative

intention to the addressee may have adverse effects on. the

fulfilment of his informative intention.

The examples of the advertisements discussed above which

contain sexual stimuli seem to belong to a yet another type. Having

his informative intention mutually manifest to the viewer and the

communicator himself may bring unwanted social implications, as the

advertiser of the Okinawa advertisement may face protest from

 - 200 -

feminists. Or, it may have adverse effects on the achievement of

his aim, as viewers of advertisements in general may distrust

advertisers. Moreover, it may not be crucial for the communicator

to make the addressee believe a set of assumptions to be true. As

long as concepts of sex are triggered, the addressee will entertain

them, as humans are inclined to be drawn to them, and the ac^essee

will presumably form favourable assumptions and associate them with

the advertisement, and consequently with the product advertised in

it.

In the case of ostensive communication, it is hoped that the

recognition by the addressee of the communicator's informative

intention should lead to the fulfilment of his informative

intention. The cases of non-ostensive communication are different

and achieved without the communicator's communicative intention

formed, that is without the communicator's informative intention

mutually manifest. In the case of Mary getting Peter to mend her

hair-drier, the recognition by the addressee of the communicator's

informative intention would help with the fulfilment of this

intention. However, the social implications which the modification

of the mutual cognitive environment between the communicator and the

addressee may bring must be avoided, and therefore the

communicator's informative intention is not made mutually manifest.

The double-glazing salesman does not make his informative intention

mutually manifest, as doing so may interfere with the fulfilment of

this intention. In the examples of advertisements containing sexual

stimuli as discussed above, the communicator merely intends to

 - 201 -

trigger concepts of sex, and it is not crucial for him whether the

addressee should believe his message to be true.

In this section, I have examined the notion of ostension and

the role it plays in communication. Ostension is based on two

layers of intentions, namely, the informative intention and the

communicative intention. The fulfillment of the communicative

intention, i.e. the recognition of the informative intention, can

lead to the fulfillment of this informative intention, which is

presumably the communicator's ultimate goal. However, for the

communicative intention to successfully lead to the fulfillment of

the informative intention, there has to be some kind ofcooperation

between the communicator and the addressee. I have considered the

significance of cooperation between the communicator and the

addressee, and distinguished between cognitive cooperation and

social cooperation. For communication to take place at all,

cooperation at the cognitive level is a prerequisite.

Although social cooperation is not a requirement for

communication to occur, it can play a crucial role in the fixation

of beliefs, and hence in the fulfillment of the communicator's

informative intention, that is the intention to make a set of

assumptions manifest to the addressee. However, there are cases in

which the communicator has to make his addressee believe something

without having her social cooperation. In these cases, the

communicator may resort to covert communication. He intends to make

his informative intention manifest without forming any communicative

 OAO _

intention, i.e. intention to make his informative intention mutually

manifest to both the addressee and himself. Or, the communicator

may engage in covert communication, not necessarily because he does

not have the addressee's social cooperation, but possibly because he

wishes to avoid the social consequences which engaging in ostensive

communication may cause.

Communication does not always occur between a trustworthy

communicator and a trusting addressee, or between socially equal

participants. It may bring adverse effects for a communicator to

engage in ostensive communication, making his informative intention

mutually manifest to the addressee and himself. Or, it may have

social implications, which the communicator does not welcome. Thus,

the communicator may not form any communicative intention, or he may

conceal part or all of his communicative intention. These would not

be cases of ostensive communication; they are cases of covert

communication. Sperber and Wilson's notions of informative and

communicative intentions provide an adequate framework to analyse

the difference between ostensive communication and covert

communication.

4.5. Weak implicatures and non-ostensive communication

 Before I try to analyse covert forms of communication, it will

become important to distinguish weak implicatures, which are

nontheless ostensively communicated, from covert forms of

communication, which are non-ostensive. In this section, I shall

examine weak implicatures which conform to Sperber and Wilson's

notion of ostensive communication and therefore can be analysed in

terms of the principle of relevance.

In indirect but ostensive communication, some implicatures are

strongly backed by the speaker, whereas some are very weakly

implicated and, therefore the hearer has to take a great

responsibility if she is to believe them. Nevertheless, these weak

implicatures are ostensively communicated. It is worth recalling

here that in ostensive communication the communicator's informative

intention must not only be made manifest to the addresse, but must

also be made mutually manifest to both the addressee and the

communicator himself. This overt aspect of communication is crucial

to Sperber and Wilson's account of ostensive-inferential

communication. Weak implicatures, which are ostensively

communicated, however weakly communicated they may be, comply with

this requirement of overtness. They must be distinguished from

covert forms of communication, which are non-ostensive and therefore

do not conform to this overtness.

In Sperber and Wilson's account of ostensive communication, the

communicator's informative intention is not an intention directly to

modify the addressee's thoughts. What he does intend instead is to

modify the cognitive environment of the addressee, by making a set

of assumptions manifest or more manifest to her. Let us suppose

that Paul has said (33) to Kay:

(33) I went to Glyndebourne last Saturday.

Now, where is the hearer supposed to look for the relevance of (33)?

The explicatures of (33), that is, assumptions obtained by

development of the logical forms encoded by (33), include the

following assumption:

(34) Paul has said that he went to Glyndebourne on the

previous Saturday.

It can be said that Paul has explicitly communicated (34). It is

also plausible to argue that he has expected Kay to supply premise

(35) and derive (36):

(35) Glyndebourne is known for its opera company.

(36) Paul went to the opera at Glyndebourne.

 - 205 -

The question is, what more, if anything, has Paul implicitly

communicated? Are there further assumptions which are members of

(I), that is, assumptions which Paul has intended to make manifest

to Kay? Kay could provide premises (37)-(40), for example:

(37) Tickets for Glyndebourne are expensive.

(38) People who can buy expensive opera tickets are

wealthy.

(39) People who go to the opera are cultured.

(40) People who go to Glyndebourne belong to a certain

social class.

Processing (33) in a context which includes (37) - (40), Kay might

derive (41) - (43):

(41) Paul is wealthy.

(42) Paul is cultured.

(43) Paul belongs to a certain social class.

 - 206 -

However weak these implicatures may be, in many circumstances

Paul would overtly communicate (41)-(43), even if he did not intend

Kay to derive these particular conclusions. It cannot be argued

that he strongly backs them up; she takes the responsibility to a

degree for supplying these particular premises and conclusions. On

the other hand, Paul could have made a certain assumption about

Kay's knowledge and guessed that she would supply just these

premises and conclusions.

Thus, it could be that Paul has said (33) in order to impress

Kay with his financial and social background, that he is wealthy,

cultured, and belongs to a certain social class. He has probably

succeeded in making mutually manifest his intention to make these

assumptions manifest to her, hence implicated them. Vhat he cannot

be sure oi is that she will actually access and use just these

assumptions, of all those that were manifest. That is what makes

them weak implicatures in many circumstances. Because they are

weakly implicated, he is dependent on the hearer's ability and

willingness to supply these premises and derive the intended

conclusions. But he could try to raise his success rate by making a

right assumption about Kay, about what information she is likely to

have accessible.

Thus, the communicator can achieve successful communication

through weak implicatures. Even though he cannot be sure of

success, he can make a right assumption about his addressee that she
is likely to access and use certain assumptions in interpreting his

utterance, and aim at high probability.

It has been argued that there is a sole criterion by which the

hearer identifies assumptions ostensively communicated by the

utterance, namely, consistency with the principle of relevance.

Thus, (33) opens up a number of possibilities of implicature. The

hearer supplies premises which are accessible in her cognitive

environment, and then derives implicatures consistent with the

principle oi relevance. These are implicatures which a rational

speaker might have thought would yield enough effects to make her

processing effort worth while, and put her to no unjustiable

processing effort in achieving these effects.

The question here is, how farshould Kay seek for

possibilities? In other words, how far should she keep adding

premises to the context and recovering further conclusions? For

example,should Kay add (44) the context and derive (45)?:

(44) People who go to Glyndebourne do notlike hunting,

shooting and fishiing.

(45) Paul does not like hunting, shooting and fishing.

 - 208 -

It is very doubtful that Paul has given his hearer Kay any

encouragement to supply the premise in (44) and derive (45). The

conclusion to be drawn from these examples is that there may be no

cut-off point between assumptions strongly backed by the

communicator, i.e. implicatures, and assumptions derived from the

utterance, but only on the addressee's sole responsibility, i.e.

non-iraplicated contextual implications.

It can be argued in the example above that (35) is a strongly

implicated premise and (36) a strongly implicated conclusion, which

are strongly backed by the speaker. In contrast, (44) is a very

weakly implicated premise and (45) a very weakly implicated

conclusion. Though there is some backing from the speaker, the

backing is very weak, and the hearer has to derive them on her own

responsibility. Nevertheless, both strong and weak implicatures are

ostensively communicated, and therefore are consistent with the

principle of relevance.

On the one hand, from the hearer's point of view, there are

actually two criteria which she uses in deciding how far she is to

investigate. Firstly, the hearer uses the criterion of consistency

with the principle of relevance. Secondly, in deciding how far she

should go in providing premises, adding them to the context, and

deriving conclusions, she uses the criterion governing her own

cognitive activities, namely, the search for maximal relevance.

That is, she goes as far as she finds it relevant to go. She would

not go beyond the point where her processing effort outweighs the
 - 209 -

effects she gets out of the derived conclusions. If the effect

obtained from the derived assumptions is weighed against processing

effort, there will be a point at which it is not worth going any

further. On the other hand, from the speaker's point of view, the

sole criterion is consistency with the principle of relevance.

It has been argued that the communicator can achieve successful

communication through weak implicatures. Even if he cannot be

absolutely sure of success, he can rely on a high degree of

probability that his addressee will access and use certain

assumptions in interpreting his utterance. However, because each

individual has a differen-t cognitive environment, and different

cognitive preferences and abilities, what is worth the processing

effort depends on the individual. Let us suppose that Paul has said

(46) :

(46) I like Beethoven.

In uttering (46), Paul has in mind symphonies and concertos by

Beethoven, as many people might have. On the other hand, his

hearer, Kay, has just seen the opera Fidelio, and has (47)

immediately accessible in her mind:

(47) Beethoven composed Fidelio.

 - 210 -

Thus, she adds premise (47) to the context and derives (48):

(48) Paul likes Fidelio.

However, it is quite possible that Paul did not have Fidelio in

mind on uttering (46) and did not expect Kay to derive (48). Indeed,

Paul may not even like Fidelio at all. But since Kay has just seen

Fidelio, she finds it the most accessible piece by Beethoven.

Providing premise (47) thus does not cause her much effort. From

Kay's point of view, (48) is a conclusion easily derived from Paul's

utterance (46). From Paul's point of view, he has certainly not

strongly encouraged his hearer to derive (48); he only very weakly

implicated it. Thus, as mentioned above, in the case of a weak

implicature, the hearer must take considerable responsibility if she

is going to treat it as true.

The communicator can take advantage of this indeterminacy, as

we shall see in the examples below. Since there is no clear

boundary between assumptions implicated by the communicator and non­

implicated assumptions derived on the addressee's responsibility,

and since the addressee is always to some extent responsible for

weakly implicated assumptions she derives, there is no fool-proof

way of determining which assumptions are weakly implicated by the

communicator and which are merely non-implicated premises and

conclusions: that is, there is no way of proving which assumptions

 - 221 -

are weakly, ostensively communicated, and which are not ostensively

communicated at all.

Let us look at the following exchange between Malcom Turnbull,

the lawyer who acted for Peter Wright, and Sir Robert Armstrong, who

was the main witness, representing the British government, in the

Spycatcher trial. During the trial, seeing Armstrong nervous and

rattled by the suggestion that he had told an untruth in order to

protect the sources of operations of M15, Turnbull ploughed on. He

turned to a letter, which was written by Sir Robert Armstrong to

William Armstrong, chairman of the book's publishers, on 23 March

1981, which reads as follows:

I have seen the extracts in the Daily Mail today from

Chapman Pincher's forthcoming book, Their Trade is

Treachery. The prime minister is in my judgment likely to

come under pressure to make some statement on the matters

with which Mr. Pincher is dealing ... I should like to be

able to put her in a position where she could make a

statement this Thursday (26th March), if she could wish to

do so. I should therefore be very grateful if you would

be willing to make one or (preferably) two copies of the

book available to me as soon as possible ...

(From 'The wily colonial boy versus the upper

 - 232 -

class Brit' in the 15th March, 1987 issue of

The Sunday Times)

Sir Robert agreed that the letter was designed to give the

impression that the gove^nent did not have a copy of the book,

whereas it had somehow obtained the page proofs some six weeks

previously. He was communicating the following implicature:

(.49) The government did not have a copy of Their Trade is

Treachery.

This was followed by the exchange below:

T: I put it to you that this letter contains an untruth.

A: It does not say that we have already got a copy of the

book, that is quite true.

T: So it contains a lie?

A: As one person said, it is perhaps being economical with

the truth.
 - 213 -

(Edited extracts from 'The wily colonial boy

versus the upper class Brit' in the 15th March,

1987 issue of The Sunday Times,

where 'T.' stands for Turnbull and 'A.' for

Armstrong.)

Armstrong did intend to communicate, even though by

implication, (.49). It was, perhaps, indirectly and weakly

communicated, nevertheless, it was ostensively communicated by

Armstrong making his informative intention mutually manifest to the

publisher and himself.

Thus, in this case, the speaker was trying to take advantage of

the fact that he was communicating certain assumptions by

implicature, rather than explicature, and he was denying his backing

for these assumptions, despite the fact that he had ostensively

communicated them. These cases are not exceptional; implicatures

are often denied by speakers who ostensively communicated them but

want to shift the responsibility for them to their audience.

The examples I have considered in this section involve weak

implicatures. However weakly they may be implicated, they are

ostensively communicated, and therefore can be analysed in terms of

the principle of relevance. They all conform to the requirement of

overtness, i.e. the communicator intends to inform his audience of

 - 214 -

something by making his informative intention mutually manifest to

the audience and himself.

In the next section, I shall go back to examples of covert

forms of communication, where this overt aspect of communication is

lacking, and I shall consider how they can be analysed.

4.6. Covert communication

In Section 4.3, I examined examples of covert communication, in

b
which the speaker does not formAcommunicative intention, that is, he

does not intend to make his informative intention mutually manifest,

for example, advertisement captions (1), (5), (12) and (13). These

are not cases of ostensive communication, and therefore the

principle of relevance does not apply to them, for it applies only

to cases of ostensive communication. In Section 4.5, I have also

considered examples of weak implicature, which are ostensively

communicated, and therefore can be accounted for by the principle of

relevance.

Now, can Relevance Theory account for covert communication?

Sperber and Wilson's notions of informative intention and

communicative intentions provide an adequate framework to define

covert communication. On the one hand, in ostensive communication,

 - 235 -

the speaker aims at the fulfillment of his informative intention by

fulfilling his communicative intention, that is, his intention to

make it mutually manifest to the addressee and himself that he has

this informative intention. On the other hand, in covert

communication, the speaker intends to achieve fulfillment of his

informative intention without forming any communicative intention,

i.e. without intending to make his informative intention mutually

manifest to the addressee and himself. In these cases, the speaker

intends to affect the cognitive environment of his addressee by

making her recover certain assumptions, yet he avoids modification of

the mutual cognitive environment of the addressee and himself by not

making this intention mutually manifest to his addressee and

himself. The difference between ostensive communication and covert

communication has been described as follows:

Ostensive communication: an overt form of

communication where there is, on the part of the speaker,

an intention to alter the mutual cognitive environment of

the speaker and the hearer.

Covert communication: a case of communication where the

intention of the speaker is to alter the cognitive

environment of the hearer, i.e. to make a set of

assumptions more manifest to her, without making this

intention mutually manifest.

 - 216 -

(Rencherif and Tanaka 1987)

Relevance Theory deals not only with a specified type of

communication, that is, ostensive-inferential communication, but

also with cognition. I argue that the notion of relevance can be

extended to account for non-ostensive stimuli, and that

considerations of relevance offer an appropriate explanation of how

covert forms of communication are carried out.

In non-ostensive communication, the hearer does not have the

speaker's guarantee of optimal relevance, which offers a criterion

for deciding what the speaker intends to communicate, and thus

5
directs the hearer's attention to certain assumptions. However, the

principle of relevance is not the only factor involved in human

cognition. Human cognition is organised in such a way that it is

perhaps more susceptible to certain stimuli than others.

According to Sperber and Wilson, human cognition is designed to

pick out relevant stimuli, and to process them in the most efficient

way. The notion of relevance proposed by Sperber and Wilson is not

only valid in ostensive-inferential communication, but describes a

general cognitive tendency determining which of the available

stimuli are likely to be processed, and in what way.

It can be argued that covert communication exploits this human

susceptibility to certain stimuli. The communicator intends to

 - 217 -

communicate something without being overt about it, that is, not by

making his informative intention mutually manifest to the addressee

and himself, but by relying on the addressee noticing certain

things, given the way her general cognitive system is organised. Ve

have considered examples in which sex is used as a stimulus which

draws the audience's attention to something. In Chapter 3, it was

suggested that perhaps food was one of the stimuli which appeal to

human cognition as relevant.

In example (1), the advertiser intends to convey (4) without

making the informative intention behind it mutually manifest. He

can expect to succeed in conveying (4), even though he does not make

his informative intention mutually manifest, because he knows that

his audience will probably notice the girl's breasts. The audience

are likely to notice the breasts, not because the advertiser made it

mutually manifest that he intended them to notice it, but because

their mind is organised in such a way that it is susceptible to such

stimili, that these stimuli are relevant to them.

In (5), the suggestion of sex is carried out through both the

Chinese character which can mean 'sex' on its own and the

illustration, showing a naked torso of a young man. This

information transmission is achieved, because the audience are

likely to pay attention to these phenomena, even though the

advertiser does not make his informative intention mutually manifest

to them and himself.

 - 218 -

The advertiser of (12) and (13) succeeds in making some

assumptions about sex manifest or more manifest to his audience

without having to make his informative intention mutually manifest

to them and himself, and therefore without giving his guarantee of

relevance. The slightest hint of an orgy and a lesbian relationship

draws the audience's attention, because their cognitive system is

organised in such a way that it is more susceptible to this kind of

information. Quite generally, covert communication manipulates

triggers of the sort which the human mind cannot avoid responding

to.

Despite the prima facia inconsistency of some aspects of

communication in advertising with Relevance Theory, the central

concepts of the theory in fact make it possible to provide an

explanation oi these very aspects. Thus, we have used the concepts

of mutual cognitive environment, mutual manifestness, informative

and communicative intentions to bring out the distinction between

equal and unequal partners in acts of communication.

Sharing a mutual cognitive environment is the basis for the

cooperation and coordination which underlie all genuine acts of

communication between 'equal' partners, where the speaker is sincere

and the hearer trusts him, and where they both strive towards the

same end, that is, maximising the relevance of all new information

in order to modify their knowledge of the world.

 - 219 -

However, in unequal partnerships, such as in advertising, the

communicator, though intending to transmit some information and

intending to make the audience recognise that information, does not

wish to affect the mutual cognitive environment that he shares with

his audience, because he wants to avoid the social consequences that

such modification of their mutual cognitive environment necessarily

brings with it.

 - 220 -

4.6. Conclusion

The notion of ostention is crucial to Sperber and Wilson's

analysis of ostensive-inferential communication. Ostensive

communication must have the overtness of the speaker intending to

alter the mutual cognitive environment shared by the hearer and

himself.

The main purpose of this chapter has been to consider examples

oi discourse which do not conform to ostention. In this type of

communication, covert communication, the speaker intends to alter

the cognitive environment oi the hearer but does not intend to make

this intention mutually manifest to the hearer and himself and thus

alter their mutual cognitive environment. The reason for this may

be that the speaker believes that revealing his informative

intention would have an adverse effect on the fulfilment of his

informative intention. Or, it may be that the speaker wishes to

avoid the social consequences which might ensue from engaging in

ostensive communication.

JL
Now, does Relevance Theory offer^contribution to analysing this

type of communication, which does not confarm to ostensive

communication? I have argued that the notion of informative and

communicative intention provides an adequate definition of covert

communication and distinguishes it from ostensive communication.

 - 22 1 -

The notion of relevance is defined in terms of contextual

effect and processing effort: everything being equal, the more

contextual effect, the more relevant; everything being equal, the

less processing effort, the more relevant. The hearer processes a

stimulus and derives contextual effects as a reward. Contextual

effects are derived through inference.

In covert communication, however, the hearer does not get her

reward entirely through inference. Nor does the speaker rely on the

hearer to receive enough contextual effects to compensate for her

processing effort. The speaker exploits the fact that humans are

susceptible to certain stimuli, such as sex and food. In

advertising, for example, the advertiser includes a sexual

illustration in the advertisement, which acts as a reward to

compensate for the audience's attention.

It could thus be argued that there are elements in covert

communication, such as the fact that humans are suceptible to

certain stimuli, which are not fully accounted for by Relevance

Theory as it stands. The theory provides a useful framework to

define covert communication and delineates the position of such

elements, but it does not explain how they work. Further research

is awaited to explain how factors such as human suceptibility and

emotions function in communication and how they interact with the

principle oi relevance.
 - 222 -

Chapter 5: Images of Women in Advertising

5.1. Introduction

This chapter discusses some specific examples of the language

used in contemporary advertisements in Japan and the U.K. which

target young women. The projection of the image of women in

advertising has been the subject of a vast literature in various

disciplines, such as linguistics, anthropology, and sociology. My

aim in this chapter is to show how assumptions of Relevance Theory

can be used as the basis for a detailed analysis of how individual

concepts are manipulated in advertising. The focus here is on words

which are frequently used in relation to women in advertisements.

'Intelligence', 'individualism' and 'feminism' have been chosen to

illustrate this process. I am concerned to reveal what these words

actually mean within the particular medium of advertising, and what

values they present, by analysing how they are used in a number of

concrete cases.

Vhat follows is not a systematic sociological study of the

manipulation oi Japanese or British women by advertisers, but rather

an attempt to see how studying the language used by advertisers may

indicate some of the techniques of manipulation. Moreover, in doing

so, I attempt to show the position of women in the respective

societies.
 - 223 -

The examples have been culled from Japanese monthly magazines

for young women, J.J., Can. Can, More, With and Cosmopolitan, between

September 1984 and December 1987 as well as the British equivalents,

such as Cosmopolitan, Vogue, Elle, and New Women. It is worth

noting that these Japanese magazines all have English titles. As

will be seen later on, a certain cultural dependence on the Vest

informs much of the interpretation of these words.

Section 5.2 is the discussion of theoretical framework. After

a brief examination oi how word meaning has been analysed in various

linguistic approaches, I shall argue that Relevance Theory provides

a principled account for the analysis of the words in question. In

the sections which follow, namely, Sections 5.3, 5.4 and 5.5, I

shall describe the particular examples 'intelligent*,

'individualism' and 'feminist' and their use in advertising.

5.2. Word meaning and concepts

There are a number of different linguistic approaches to word

meaning. Among them, there seems to be a general agreement that the

'meaning' of a word can be analysed in terras of the associated

concept. The classical view is that the meaning of a word is a set

of semantic features or a complex concept. This approach is called

componential analysis and claims that word meaning is provided by a

 - 224 -

definition which expresses the necessary and sufficient conditions.

According to this approach, the word 'bachelor' would mean a d u l t a n d

male and human and unmarried. This analysis suffers from various

defects, however, one of which is that it is not plausible that a

total and exhaustive account of a particular word can be provided by

a componential analysis of the word.

Based on this criticism, some linguists have adopted an

approach, called meaning postulates, as an alternative to

componential analysis. This approach views the meaning of a word as

a simple concept with meaning postulates or inference rules

attached. In this approach, the word 'bachelor' would mean

BACHELOR, plus inference rules showing that propositions about

bachelors entail propositions about unmarried adult human males.

This approach does not suffer from the assumption about exhaustive

decomposition of the meaning of a word into an integral number of

so-called universal components, nor is it based on any assumptions

about atomic components. It has considerable advantages over

componential analysis.

Componential analysis presupposes that the meaning of a word

is firmly fixed. But some linguists stress the lack of clear-cut

boundaries. The alternative approach might be called the fuzzy

meaning approach. It is based on the assumption that the meaning of

a word is an inherently fuzzy concept and its main claim is that it

is impossible to assign a firm definition to any word. What is

claimed to iollow from this is that there is no point in asking

 - 225 -

whether a certain container that seems to fall somewhere between

being a cup, a mug, a bowl and a vase is really any one of them.

Rather, these concepts are fuzzy, in such a way that any one of them

can accommodate our container as a marginal case, even though its

level of so-called cuppiness, mugginess, and so on, is low. The

advantage claimed for the notion of fuzzy concepts over that of

well-defined classificatory concepts is that it offers an account of

a wider range of data than the latter can possibly cover.

Sperber and Wilson agree with Fodor, and in fact many others,

in that they see the advantages of meaning postulates over

componential analysis (.see Sperber and Wilson 1986a). They argue

tko,
that the meaning of^ majority of words cannot be decomposed into more

primitive concepts, and that words such as 'bachelor* , to which

componential analysis provides a convincing analysis, are

exceptional (1986a: 91). They point out that one only has to think

of a word such as 'yellow' to realise the difficulty of analysing

all words in terms of componential analysis.

Incidentally, the meaning postulates approach is compatible

with the fuzzy meaning approach. The meaning of a word can be a

fuzzy concept and can be analysed in terms of meaning postulates.

Sperber and Wilson's main claim is that there is nota single format

which maps different words on to their respective meanings and that

different words may have meanings of different formats (1986a:90).

What seems to follow from this suggestion is the possibility that

some words are classi- ficatory and some are fuzzy.

 - 226 -

However, this distinction between classi— ficatory words and

fuzzy words does not concern me here. The point I am making in this

chapter goes well beyond the question of whether 'intelligent' and

so on are fuzzy concepts, although Sperber and Wilson would treat

them as fuzzy (Wilson, personal communication; Sperber, workshop

presentation, Essex University, June 1989). My point is that in the

advertisements I examine, words, such as 'intelligent', are being

used in ways which clearly go beyond their standard meanings,

whether these meanings are classificatory or fuzzy. To say that a

concept is fuzzy is not to say that it can be stretched indefinitely

far: the word 'cup' may be reasonably applied to various objects, as

Labov has shown (see 1973), but to call a computer a 'cup', for

example, would be unacceptable. I shall show with examples that

the concepts which are dealt with in this chapter are being

stretched beyond what most people would consider normal limits.

What concerns me in this chapter is not normal fuzziness of meaning,

but ovei— extensions of meaning.

Having accepted that some or all simple concepts may be fuzzy,

Sperber and Wilson argue that existing arguments for fuzziness are

inadequate. There is a pitfall: the baldness paradox (see Sperber

and Wilson 1986b). To summarize their argument, one is led into

this paradox by accepting, first, that a man with no hair is bald,

and secondly, that if a man with no hair is bald then a man with one

hair is bald,, and then, via the general principle that if a man with

(n) hair is bald then a man with (n + 1) hair is bald, to the

conclusion that a man with a full head of hair is bald.

 - 227 -

Sperber and Wilson offer a solution to this paradox. It starts

off by accepting that 'bald' is classificatory concept after all.

The underlying assumption is that existing accounts for fuzziness

have not shown fuzziness at all. A classificatory concept offers a

way of avoiding the baldness paradox. 'Bald', they claim, has a

necessary and sufficient condition: having no hair. It follows from

this that to describe a man with one hair as bald is strictly

speaking false. However, they show, with examples, that many

utterances which are strictly speaking false are nonetheless

pragmatically appropriate. The use of 'bald' to apply to a man with

very little hair would fall into this category. I outline their

solution below.

Sperber and Wilson suggest that what is often analysed as

literal use of a fuzzy concept might instead be analysed as loose

use of a classificatory concept (see Sperber and Wilson 1986b). In

the example above, 'bald* may well have clear-cut boundaries and

still be used loosely, but acceptably, to refer to a person who

falls outside these boundaries.

In this approach, the notions of descriptive use and

interpretive use are crucial. As I have outlined in Chapter 2, an

utterance can be used to represent a state of affairs which makes it

true, namely, descriptive use; or an utterance can be used to

represent another utterance or thought which it resembles in

meaning, i.e. interpretive use. Utterance (1), for example, may be

descriptively used to represent a certain state of affairs, or

 - 226 -

interpretively used to represent another thought or utterance which

it resembles in meaning.

(1) I'm starving.

A proposition, in isolation from any context, has a number of

analytic implications. However, propositions are entertained not in

isolation but in a context of background assumptions. When a

proposition P is processed in a context <C>, P may yield a number of

contextual implications. As shown in Chapter 2, a contextual

implication of P in the context <C> is a proposition implied neither

by P alone, nor by <C> alone, but by the combination of P and (C).

According to Sperber and Wilson, two representations with P and Q as

their propositional content interpretively resemble one another in a

context {C> to the extent that they share their analytic and

contextual implications in the context {C}. Thus, by definition,

interpretive resemblance is context-dependent.

In normal circumstances, a descriptive use of (1) would be

false, for the speaker is not really starving. However, the truth

of an utterance is not necessary for successful communication. As

discussed in Chapter 1, Sperber and Wilson argue that there is no

requirement for a maxim of truthfulness. They further argue that

the sole criterion that the hearer uses in utterance comprehension

is a presumption of relevance, that is, the speaker, in choosing his

 - 229 -

utterance, has caused his hearer no unjustifiable processing effort

in conveying his message. Utterances are used to represent the

speaker's thoughts, and the assumption is that the speaker will

choose the most economical means of conveying the thought in

question.

Now, Sperber and Wilson call an utterance 'literal' (1986a:233)

if it has the same proposition as the thought it is used to express.

That is to say that all the implications of the proposition

expressed are endorsed by the speaker. When only some of the

implications of the proposition expressed are endorsed by the

speaker Sperber and Wilson call an utterance 'loose' (1986a:234).

I argue that the notion of loose use of language, provides an

adequate framework for the ways in which words such as

'intelligence' are used in advertising.

Sperber and Wilson argue that loose use of language is rife in

ordinary communication, in fact they argue that cases where an

utterance is used strictly literally are a limiting case (see

Sperber and Wilson 1986a, 1986b). One may recall that in Relevance

Theory, no maxim of truthfulness is required. The only

consideration that the speaker has is considerations of relevance,

which is defined in terms of contextual effect and processing

effort. The speaker will aim at optimal relevance, rather than at

literal truth. Often the most economical means of communicating the

speaker's thought is not a literal interpretation of it: more often

than not, it is a loose one. From the hearer's point of view,

 - 230 -

unless there is a specific reason for her to believe that the

utterance is literal, she would assume that some, and not

necessarily all, of the implications of the proposition expressed

by the utterance are communicated by the speaker.

This matches our intuition. For example, my flat is in the

building which is located exactly opposite the main entrance of the

British Museum. All my visitors are informed of this in advance,

but every single one of them points out on their arrival, with a

tone oi surprise, that my flat is exactly opposite the British

Museum. This seems to confirm that the hearer does not take for

granted that an utterance is literal, and that all the implications

of the proposition expressed are endorsed by the speaker, and

assumes that only some of the implications are communicated. My

hearers follow this, despite my effort to convey that ray utterance

is literal.

Thus, utterance cl), which isstrictly speaking false, may be

good enough for pragmatic purposes, and may indeed be better, than a

strictly speaking true one, such as (.2), for (1) shares many

contextual implications with (2), and yet is easier to process than

(2), and thus causes the hearer less processing effort:

(2) I'm extremely hungry and want to eat immediately.

 - 231 -

The question here is: how, of all the implications of the

proposition expressed by the utterance, does the hearer identify the

intended subset? Sperber and Wilson argue that a presumption of

optimal relevance provides the sole criterion. When an utterance is

made by the speaker, the hearer assumes that it is intended as an

interpretation of one of the speaker's thoughts. She processes the

utterance and computes those implications which may be relevant to

her. She continues to do so as long as it is consistent with the

principle oi relevance, that is, so long as she gets enough

contextual effects to make her processing effort worth while. This

process is done in order of accessibility. It is plausible that the

literal interpretation is exceptional, rather than the norm, for the

literal interpretation may not necessarily be the most accessible

o ne .

Sperber and Wilson's approach deals with loose use without

abandoning truth-conditional semantics. When a proposition or

concept is loosely used, it is not necessarily the case that it is a

vague proposition or concept: it is not that a guarantee of

approximate truth is given to the proposition expressed, for no

guarantee of truth is given to this proposition at all in the first

place. Instead, only some of its logical and contextual

implications are taken to be accompanied by a regular guarantee of

truth, whereas others are simply ignored.

There is a point at which looseness becomes unacceptable. This

is context-dependent (see Sperber and Wilson 1986a, 1986b). To call

a man with one hair bald is just as false as calling a man with a

full head of hair bald. The difference between them is not the fact

that one is true and the other false, but the fact that one is an

acceptable loose use because many of its logical and contextual

implications are true, whereas the other is unacceptable since the

hearer would be able to derive from it virtually no true descriptive

information about the state of affairs it purports to represent.

Thus, the notion of loose use offers a better alternative to

the fuzzy meaning approach, which supersedes assumptions of fuzzy

meaning but does not share its defects.

It has been mentioned above that the meaning of a word is

provided by the associated concept. According to Sperber and Wilson

(1986a:86), a concept holds three distinctive types of information,

namely, logical, encyclopaedic and lexical. Firstly, the logical

entry for a concept is a set of deductive rules. These deductive

rules apply to logical forms of which the concept is a constituent,

and thus determine its analytic implications. For example, the

logical entry for the concept 'elephant' would indicate that it is

an animal. Secondly, the encyclopaedic entry for a concept consists

of information about the extension of the concept. This type of

information contributes to contextual implications of the concept.

The encyclopaedic entry for the concept 'elephant' would include

assumptions about its tusks and trunk. Finally, the lexical entry

of a concept is information about its counter-part in the natural

language. The lexical entry of the concept 'elephant' would contain

 - 233 -

that it is expressed by the word 'elephant' in English. The

structure oi a concept analysed in terms of these three types of

information will be crucial to the explanation of over-extended ways

in which words may be used, in one of the cases leading to a shift

in word meaning, as I shall show in the following sections.

Relevance Theory provides a comprehensive framework for

analysing a variety of ways in which words may be used. In this

section, I have examined different linguistic approaches to word

meaning. I have argued that the notion of fuzzy concept offers an

apparently attractive analysis of many words. However, it has been

argued by Sperber and Wilson that the existing arguments for fuzzy

concepts are not satisfactory: they do not show real 'fuzziness'. I

have discussed the alternative approach by Sperber and Wilson which

centres around the notion of loose use. It has also been accepted

that word meaning can be analysed in terms of its associated

concept. The structure of a concept has been discussed within

Sperber and Wilson's framework.

In the following sections, 5.3, 5.4 and 5.5, I shall examine

how a selection of words and concepts are used in Japanese and

British advertisements. In doing so, I hope to explore the ways in

which women are presented in advertising in the respective

societies.
 - 234 -

5.3. Intelligence

In this section, the following examples are designed to show

how the words chisei (intelligence) and chiteki (intelligent) are

used in Japanese advertising. In advertisements found in the

magazines studied, and in recent advertisements in general, these

words have become ubiquitous. It appears that these words are used

in Japanese advertisements to mean something like 'sophisticated',

'elegant', 'fashionable', and so on, to indicate a desirable quality

about women, in particular, about their appearance. This, however,

is not an indication of a fuzzy concept of 'intelligence'. Rather,

I argue that it is a straightforward case of loose use. Let us

examine the following:

(4) Chiteki-ae joohin-na sJbiruku burause.

intelligent noble silk blouse

(Tokyo Blouse)

An intelligent and noble silk blouse.

(5) Chotto ereganio-de cbotto interijensu-na inshoo...

a bit elegant a bit intelligent impression

 - 235 -

<Paco Rabanne)

The impression of being a bit elegant and a bit intel­

ligent. . .

(6) Intelligence. (written in English)

Sore ga kimi no utsukushisa.

that sub. you of beauty

Kite-iru fuku ni chisel o kanjiru.

wear-ing clothes in intelligence o b j . feel

Tazta ippon no kuchibeni kara mo...

only one oi lipstick from even

(Kanebo)

Intelligence, (in English)

That is (the secret of) your beauty.

Your intelligence is seen in your clothes. Even in

the lipstick you wear.

(7) Chisei to y a s e i .San Rooran no ganchiku.

intelligence and wildness Saint Laurent of suggestion

(Yves Saint Laurent)

 Intelligence and wildness. Saint Laurent's

suggestion.

These are only a lew among many examples of advertisements

using chisei (intelligence). They are all for women's fashion and

they all suggest that women should express their intelligence

through their clothes, lipsticks, handbags, and so on. This

suggestion is in direct contrast with the fact that in Japanese

society, intelligence is regarded as a desirable quality in a man,

but not in a woman, and that it is regarded as desirable that women

should not be as intelligent or well-educated as men (see Hosoi

1986, Smith 1987). Can we thus take this phenomenon as a sign that

Japanese society is changing, and that now intelligent women are

accepted and encouraged? This may be a premature conclusion. It is

obvious in the examples above that chisei (intelligence) in these

advertisements is a superficial quality, a sort of quality which

clothes, cosmetics, and accesories can give women. It seems to be

synonymous with 'elegance', or 'femininity', or 'sophistication',

rather than 'brightness', 'cleverness' and so on as a thesaurus

would tell us. It is used for appearances, rather than for one's

mental state.

Words such as chiteki and chisei are usually, though not

exclusively, used for women in advertising, that is, they describe

women in advertisements which are targeted at women. It may be that

women are being seen as a chiteki gender, chiteki in a limited,

superficial sense. And this is all desirable, even though women are

still not encouraged to exercise their mental ability to the full.

Perhaps women are accepted as chiteki in their own, small ways, so

long as they do not cross the border and invade men's fields and

threaten them. Example (8) provides support for this

suggestion:

(8) Egao mo onna no chisei kashira.

smiling face also woman of intelligence it appears

(Narisu Cosmetics)

It appears that a smiling face is also women's

intel1igence.

In (6), the caption reiers to smiling face as 'women's

intelligence', implying that there is a distinction between

intelligence for women and that for men, and that a smile belongs to

the former, not the latter. It might be that 'intelligence' as we

have observed it in the examples above is not 'intelligence' in an

ordinary, general sense, but in a marked sense, i.e. intelligence

for and only for women.

 - 238 -

It has been argued above that 'intelligence' in these examples

is synonymous with 'elegance', 'femininity' and 'sophistication'.

To be sure, 'sophistication' entails some mental quality; after all,

it is derived from the Greek meaning 'wisdom'. But again the word

is used here in a limited sense. It manifests itself in one's

outfits and make-up. It might be that it is desirable for a woman

to be intelligent in her choice of clothes and cosmetics. It is not

that a woman is encouraged to be intelligent in a general sense, nor

is it that she is encouraged to be silly. It is virtuous for a

woman to be intelligent, but only in a restricted way. Japanese

women are expected to rule the home, and it is they who are

responsible for its day-to-day operation, the care of the children

and the management of the household budget. They are encouraged to

exercise their brain in their jobs, as managers of households,

mothers and consumers.

It is possible to argue that Japanese society has long valued

women's mental qualities, albeit in limited domains, for an ideal

woman has been described since the Meiji era as ryoo-sai ken-bo,

that is, 'good wife, wise mother' (see Smith 1987:7). One could

argue that a chiteki (intelligent) woman as depicted in contemporary

advertising is a kind of modernisation of this Meiji slogan.

It is worth noting at this point that in Japanese society women

are the prime consumers not only of commodities but also of art and

culture (Moeran 1983:101). It is possible that these two facts are

connected. Advertisements seem to suggest that a chiteki woman is

 - 239 -

someone who wears smart clothes and decent make-up, smiles sweetly

and goes to art exhibitions. Here is an example to illustrate this

point:

(9) Bijutsukan e iku josei ga fuete-imasu.

art gallery to go women s ub j. increas-ing

Purachina o tsukeru josei ga fuete-imasu.

platinum o b j . wear women subj. increas-ing

(Platinum Promotion Forum)

The number oi women going to art galleries is

increasing.

The number oi women wearing platinum is increasing.

These images of women seem to parallel the kind of education

received by women in contemporary Japan (see Hosoi 1986, Smith

1987). Before the war, there was an enormous difference between the

education received by men and women. After 1945, however,

educational opportunities for women were improved. In fact, looking

at the figures for boys and girls attending senior high school and

college or university, it looks as though Japanese women are

enjoying equal opportunities. But if one investigates the kind and

quality of education received by men and women, one realises that

 - 240 -

there is a significant difference: women tend to go to junior

colleges, which are regarded as 'a modern version of the old schools

for brides' (Smith 1987:11) and study so called women’s subjects,

such as home economics, education, language and literature.

It has been argued that the meaning of the word chisei

(intelligence) is restricted to 'intelligence' which is to be

manifested in limited domains. These are traditionally categorised

as women's domains, such as fashion, the household, and the

consumption of commodities and culture. Sometimes the mental

quality of the word even seems to be repressed altogether, leaving

the word to mean simply some kind of desirability and appeal to men.

It is revealing to look at a feature carried by J.J. (October,

1986). It is entitled, 'Suggestions for intelligent elegance'

(Chiteki eregansu no teian), and includes the following suggestions

(my translation):

(10) a. In choosing a dress. . . (remember that) 50/i of the

y
creation oi an impre^ion oi intelligence depends

on your neck-line. An appropriate round-neck and

golden buttons are the key to success.

b. The main trend of this year's suits is towards

those which have a tightly shaped waist. ...a

tight waist-line leads to an expression of

 - 241 -

intelligence.

c. You cannot omit black, for it is a front runner

for an intelligent-looking colour.

(10a) - (10c) give us a good idea of what an 'intelligent'

woman should look like: she wears a black dress with a roundneck and

a tight waist-line, with golden buttons. There is no reason why a

roundneck should be more 'intelligent' than a v-neck, for example.

Nor is it self-evident why a small waist, the colour black, and

golden buttons should be associated with 'intelligence'. The point

is that all these features are in fashion at present. Moreover,

this ensemble conforms to what Japanese men find desirable in women.

I shall come back to this point.

Let us look at another example:

(11) Chiteki eregansu. Jetta.

intelligent elegance Jetta

Chiteki-na machi o chiteki-na josei to

intelligent town obj. woman with

Chiteki-ni hashirimasu,

intelligent drive
 - 242 -

(Volkswagen)

Intelligent elegance. Jetta.

It drives through an intelligent town with an

intelligent woman in an intelligent manner.

Whatever 'intelligent elegance' may mean, clearly the word

chiteki here means something different from the English word

'intelligent', or even from what the Japanese word chiteki, is

supposed to mean. In Japanese, let alone in English, 'intelligent

town' and 'The car drives in an intelligent manner' are

pragmatically unacceptable, and 'intelligent elegance' pragmatically

questionable. Only 'intelligent woman' is well-formed. However,

what does it mean to say that a woman is chiteki (intelligent; here?

Because the same word is used for a car and a town, which cannot

have the quality oibeing 'intelligent', it is dubious whether the

advertiser meant to describe the woman along these lines. The word

is used to express some desirable quality, but not one exactly to do

with intellect. It is some sort of pleasant quality for a woman to

have, such as being 'fashionable' or 'sophisticated'.

It is worth noting that although the car advertised is German,

the background to the advertisement is a European-style, red brick

building, with a sign indicating that it is a French restaurant.

Apparently, a French restaurant is seen as 'intelligent'.

Possibly, the word chiteki is synonymous with 'fashionable', since

 - 243 -

that is how France is regarded by Japanese women. Or, it might be

that German technology and French cooking are seen as chiteki assets

of the respective countries.

The usage oi the word chiteki (intelligent) as seen above can

be accounted for in terms of loose use. The word is used not in

virtue oi truth-conditions, but in virtue of resemblance. There is

some resemblance between the content the advertisers wish to

communicate and what the word chiteki expresses. I assume, as

pointed out above, that word meaning can be analysed in terms of

the associated concept or concepts. I also assume, following

Sperber and Wilson (1986a), that a concept consists of subsets of

implications. Some of the implications of the concept associated

with the word chiteki (intelligent), which have to do with

desirability and good presentation, have been retained, while

others, which have to do with mental ability, have been suppressed.

However, it is the latter implications which are, in normal

circumstances, more accessible than the former. Vhat we have seen

can be analysed as loose use of the word, and loose to a large

extent it is. But the communicator can trust the addressee to

achieve the intended interpretation of the word, that is, the

implications to do with desirability and presentation, rather than

those to do with mental ability, because of the context in which the

word is used.. Our cognitive system is organised in such a way that

it would just reject inconsistency or contradiction. In processing

(11), lor example, it would cause too much friction in the context
 - 244 -

to interpret the word as meaning 'mentally excellent', and there is

no reason to believe that the word is being used to refer to some

mental property, some of the normal implications are simply

suppressed.

The images of women we have seen in the examples above all

conform to a stereotype of Japanese women; presentable and a capable

house-keeper but not intellectual or academic. Here, I rely on the

notion of stereotype, which needs to be clarified in further

research. But it seems plausible to argue that this stereotype

helps the audience process these examples above by giving the

audience access to an adequate context in which they should be

processed.

It has been argued here that the words chiteki/ chisei

(intelligent/intelligence) are used in Japanese advertising for

women not to imply much about their mental quality, but rather to

imply desirability to men. (12) is a clear example of this:

(12) Kono aki no shuryuu wa yahari

this autumn of main stream s u b j . as expected

chiteki-de sekushii.

intelligent-and sexy

(Robe)
 - 245 -

The main trend this autumn, as you may have guessed,

is to be intelligent and sexy.

Can.Can,, a rival magazine of J.J. (mentioned above), carried a

feature entitled 'the Declaration to become a "desirable woman"'

('Ji o n n a ' e no henshin sengen) (December 1987). The suggestions

made there almost completely overlap with those in J.J., for

'intelligent elegance', including the colour black, a small waist

and golden buttons. The feature also recommends visits to art

galleries, which supports the point made above. Furthermore, one

of the three cars listed as 'Cars for "desirable woman"' was Jetta,

which is mentioned above, for the very reason of being 'intelligent-

looking'. Thus, 'intelligence' and desirability, that is, appeal to

men, seem to be virtually interchangeable. Moreover, the

implications which are carried by both attributes, on the face of

it, appear much more related to desirability than to intelligence.

I have argued that in Japanese advertising words such as

chiteki (intelligent) and chisei (intelligence) are used to mean

some desirable quality for women, which is to do with their

appearance and possibly to do with their capability as a house­

keeper, rather than with their mental quality. I have also argued

that this is achieved via the loose use of the words. More

specifically, I have argued that, assuming that a concept consists

of subsets of implications, one subset of the implications of the

concept 'intelligent', which have to do with mental quality of a

 - 2 46 -

person, is suppressed, and the rest simply retained. Thus, the

words mean some desirable attribute for women, without the drawbacks

of being associated with assumptions about mental excellency, which

are regarded as undesirable for women in Japanese society.

This is a clear case of intentional loose use of the words

'intelligent* and 'intelligence'. In Japanese advertising they are

used to imply sophistication, femininity, and elegance, which are

regarded as desirab-le for women, which are, in turn, desirable

selling points from the advertiser's point of view. In the next

section, I shall examine another type of loose use.

5.4. Individualism

It is known that in Japan there has been a strong ideological

emphasis on the importance of the group over the individual, and the

necessity for the individual to subordinate her interests to those

of the primary group to which she belongs. However, there has been

an increasing number of advertisements which place an emphasis on

the importance of kosei (individuality). Does this mean that

Japanese group ideology is seriously threatened, a point raised by

Moeran (1983:105, 1984:262)? Does it, in other words, mean that

Japanese conceptual structures are shifting? I shall argue that an

analysis based on Relevance Theory reveals how the word is being

 - 2 47 -

used on the contrary to buttress such group ideology. Let us

examine some examples.

(13) Vatashi wa interia-dezainaa. ...Ronjin no

I s u b j . interior designer Longines of

konkuesto wa... chisei-bi dokusoosei ni

Conquesto subj. intelligence-beauty originality of

aiure, kiwadatte kosei-teki.

full strikingly individualistic

(Longines)

I am an interior designer. ...Longines'' Conquest is

lull oi^/intelligent-beauty and originality, and is

strikingly individualistic.

Here, it is worth noting that the watch is described in terms of

chisei-be (intel1igent-beauty).

What is 'individualistic' about the Longines watch remains to

beexplained. The illustration shows a young European-looking

woman,apparently an interior designer, in black and white on the

left, and a pair of Longines watches in full colour on the right.

 - 248 -

It is worth noting that the woman is engaged in a so-called katakana

profession, that is, a profession which is described in katakana

script, which is the marked form of writing used for Western loan

words. Katakana professions, such as kopii raitaa (copywriter),

sutairisuto, and so on, due to their Western flavour, are regarded

as fashionable and therefore desirable. The design of the watches

is not dissimilar to that of the classical Rolex 'oyster', which is

regarded as prestigious and are in fashion in Japan at the time.

The clue to what is meant by the word kosei-teki may be found

in the way in which the word is used in the same magazine, With

(July 1967) (My translation):

(14) a. This year's popular colour, green, plays an

important role in emphasising individuality...

b. You should wear a vivid-coloured polo-neck shirt

to emphasise your individuality.

c. Your fringe should be cut short and

'individualistic'...

d. The key to making you look fresh and

individualistic is to have your hair in an off-

the-face style.
 - 249 -

(14a) suggests that the colour green provides you with

individuality, for it is the colour in fashion. In the same issue,

it is mentioned that green is the most fashionable colour in Paris

at the present moment. (14b) widens the selection, as it suggests

any vivid colour. (14c) and (14d) are suggestions about hair style;

the former recommends a short fringe and the latter an off-the-face

style, which is recommended every summer, since it is only sensible

to keep one's hair short and off one's face to survive the hot and

humid summer in Japan.

What (14a) - (14d) are suggesting is that wearing a certain

colour or having a certain hair style makes one kosei-teki

(individualistic). It is not a matter of having your own style, but

of doing things which are supposed to be 'individualistic', which

include using a colour which is in fashion at the present time. An

'individualistic' woman of 1987 wears a vivid-coloured polo-neck

shirt, carries a green handbag, has her hair up and has a short

fringe. Using a colour because it is in fashion, or because it is

popular in Paris, is hardly an individualistic thing to do. Vivid

colours have always been popular in summer, rather than winter,

because of their association with the tropics. It has been

recommended every summer that a woman should wear her hair short,

or, if long, put it up, for commonsensical reasons. The word kosei

in (14a) - (14d) implies 'fashionable', rather than

'individualistic'. It is used to mean doing things which are in

fashion and which, therefore, other people are doing, rather than

doing things one's own way.

 - 250 -

This interpretation of the word koselteki as doing things which

are fashionable is not limited to advertising. A survey in 1972

(see Suzuki 1975) asked the question 'Do you think that you can

achieve greater individualism by folowing fashion?' 61.5% of women

and 48.2% of men answered positively to the question. Whereas only

slightly more women than men replied positively, it is interesting

to note that nearly twice as many men as women answered with a clear

no (41.9% vs. 22.9%), whereas a larger number of women were

undecided.

Let us look at another example:

(15) . . . Itaria no ii iro, ii katachi Guccini.

Italy of good colour good shape Guccini

Jin&ei o kosei-teki ni tanoshimu josei-tachi

life obj . individualistic in enjoy woman-pl.

no tame ni.

of sake for

(Guccini)

Good Italian colour and shape, Guccini. For women

who enjoy their life in an individual style.

 - 251 -

Individualism should be about accepting that every person is

different and that each person should be allowed to believe or do

whatever she thinks is right. But here in example (15), it means

doing a particular thing, namely, buying certain table-ware; this

comes from Italy, which is regarded as a country which produces high

fashion and kitchen ware. Here are some more examples:

(16) Jibun-rashisa ni kodawarimasu.

self-like to stick

(Hermes)

I stick to my own self.

C17) Shinayaka-ni watashi no tempo de ikite-yuku.

gracefully I of tempo at live-go

(Les Sportique)

I am going to live gracefully at my own pace.

Examples (15) - (17) promote European goods, all high-class and

expensive. Being oneself is using Italian table-ware and wearing

French suits. These advertisements apparently endorse the idea of

doing o n e ’s own thing, but in practice they advise buying European

 - 252 -

products which are expensive and have high status, and which are,

therefore, approved of by society. There is a suggestion that the

'individualism' seen in examples (15) - (17) is equivalent to

elitism: to be individualistic is to be able to afford expensive

European goods, such as Italian tableware and French outfits. Also,

there is a hint that 'doing your own thing' is doing things Western,

as in example (18):

(18) Kono natsu, jibun no kami o

this summer self of hair o b j .

Amerika no onna no ko no yoo ni

America of female of child like

Yooroppa no onna no ko no yoo ni

European of female of child like

Jibun de heaa-dezain shichao.

self by hair-design let's

(Benezel)

Let's design our own hair this summer. Just like

American and European girls.

 - 25 3 -

Thus, designing your own hair is doing what American and

European girls do. Here, 'doing your own thing' is acceptable and

desirable because, firstly, American and European girls are already

doing it, and secondly because America and Europe are regarded as

culturally superior by the Japanese.

As noted above, there is also a sense in which 'individualism'

is equivalent to elitism: doing things European or American is

desirable, for they are superior. Individualism is accessible only

to an elite group, who can afford to buy European or American goods.

Let us look at one more example:

(19) IVatashi no iro o motte-inai to, hazukashii.

I of colour o b j . have-not then embarrassing

It is embarrassing not have my own colour.

(Casio)

This advertisement is promoting identical wrist watches in

different colours. It is recommended to have one's own colour, for,

otherwise it will be embarrassing. Thus, having one's own colour is

a way of seeking social approval.

 - 254 -

Ve have examined advertisements which make use of the notion of

individualism. References are always superficial, about clothes,

accessories, hair style, and so on. Moreover, what is regarded as

individualistic in these advertisements is what is in fashion at the

time.Thus, being 'individualistic' means being fashionable.

Whereas individualism should be about each person acting in her

perticular way, in Japanese advertising it is used to mean a

specific thing, such as wearing a suit of a certain brand, having a

handbag of a certain colour, and having one's hair done in a certain

style. Considering that advertising is a form of mass communication

and is thus targeted at a mass audience, it will mean everybody

doing the same thing. What is described as 'individualistic' is

something which has already gained social approval, by being in

fashion, having high status, originating in the West, and so on.

I argue that Japanese groupism is far from being threatened by

individualism. What is happening here is an extension of word

meaning via loose use of the word. And yet it is different in kind

from the case of chiteki (intelligent) considered in the previous

section. In the case of the word chiteki (intelligent), the

implications oi the word about one's mental property are suppressed.

The transformation of the word 'individualistic' is achieved rather

more subtly. What is required is to add an extra premise to the

context. The philosophy of individualism is being able to do what

one wants; to this context, add the premise that whata woman wants

is what the rest of society wants; then, in this context, the

philosophy oi individualism becomes doing what the rest of society

wants.

Here, too, however, some of the implications of the concept

'individualimsm' are abandoned: the implications which have to do

with each person doing unique things are forgotten. This is due

to the fact that doing unique things does not go along with doing

what the rest of society wants, where there is a strong emphasis in

the society that one should behave as others do and that one should

not stand out. As the human mind is organised in such a way that it

rejects contradiction and inconsistency, it is part of a reasoning

process to eliminate implications to do with doing unique things

once it is accepted that 'individualism' is doing what the rest of

society wants. The assumption that advertising messages are

'positive' would also help this comprehension process: if

'individualism* is promoted in advertising, it must mean something

positive and desirable, so it is unlikely to mean something like

behaving differently from others and standing out in the crowd.

Thus, Japanese group ideology is rescued from supposed threats,

absorbing them into the conventional. The use of the word kosei-

teki (individualistic) in Japanese advertisements suggests a way of

resolving a clash oi cultures by apparently absorbing a notion of

the new culture, but interpreting it in the old context against

conventional premises. This conclusion is consistent with the

findings of Moeran (1984;262) that kosei is not equivalent to what

we know as 'individualism' in the West and that kosei is neatly

absorbed in ’Japaneseness' and thus the Japanese have been saved

from a clash with group ideology.

However, there is a question yet to be asked: how does the

notion of 'individualism' occur in Western advertising? Is its

usage consistent with what we know as 'individualism', that is that

each one of us acts in our own particular way? An immediate

reaction would be to suspect that individualism in this sense is not

promoted. For, if advertising is about selling things to a mass

audience, how can it coincide with the ideology that people should

be doing different things? The association between 'individualism'

and elitism can be traced in advertising in the Western press, as is

shown by an advertisement for Citizen Watch, which appeared in the

Sunday Times colour supplement (13th December, 1987).

The picture, in grey or neutral tones, shows a naked girl who

is said to be 'average' in height, hair colour, age and weight. The

only thing that does not make her average is her Citizen watch,

picked out in luminous gold on her wrist. The caption reads as

follows:

(20) There is no such thing as the average Citizen.

There is a pun on the word 'citizen': on the one hand, it is

the brand name, and on the other hand, it means a member oi society.
 - 257 -

As it is written with capitals, just like the brand name in the

advertisement, (21) would be recovered as the most accessible

interpretation:

(21) There is no such thing as the average Citizen watch.

By adding to the context assumption (22), the audience would

derive (23):

(22) An advertisement tends to say how superior the

advertised product is.

(23) A Citizen watch is better than average.

However, the alternative interpretation of the word, that is,

'a member of society* is also encouraged by the fact that the

advertisement shows the woman as well, rather than only showing a

watch, and the fact that her features are described as 'average'.

Thus, the following interpretation is also made accessible:

(24) There is no such thing as an average member of

society.
 - 258 -

The woman who, according to (20), is not 'average', is naked,

and there is no overt sign of any social group or class about her.

Moreover, she is described in the advertisement as 'average' in

height, hair colour, age and weight. Her only obvious

distinguishing feature is the watch she is wearing. Thus, the

wearing of a Citizen watch implies that she is a very special

person. Here, just as women in Japanese advertisements become

kosei-teki (individualistic) by weating an outfit of a particular

brand or having a handbag of a certain colour, an apparently average

woman in this British advertisement is made special by wearing a

certain brand of watch.

It is worth noting that, although the woman in the

advertisement is described as 'average' in various ways, the model

looks anything but 'average' with her beauty and mysterious looks.

This seems to suggest that it is not just that a Citizen watch makes

one special but that special people like this beautiful woman wear a

Citizen watch.

Thus, there is also a sign of elitism associated with

'individualism' here in a British advertisement: to be

'individualistic' means to be able to afford an expensive watch. It

is emphasised by the fact that the model, who is described as

'average', is strikingly beautiful and apparently belongs fo a

privileged group.
 - 259 -

The next example is a caption for Kotex Ferns Tampons, which

reads as follows:

(25) Ferns. Doing things your way,

which comes with the following text:

(26) Kotex Ferns Tampons are designed for today's woman.

The woman who chooses to do things her way.

Both the caption and the text apparently endorse the notion of

individual ism.

There are two kinds of illustrations found with the above

caption. One oi them shows a young woman dressed casually in a

shirt and trousers with conspicuous pink punk hair with an orange-

dotted, black bow. She has a dog which is in exactly the same pink

colour with the same bow on its head. In interpreting (25) and (26)

against this background, the audience would extend the context to

the immediate visual environment, and interpret this woman as 'the

woman who chooses to do things her way' . The audience would be

given access to a context in which women are oppressed and cannot

do things their way, and therefore do not have such unconventional

hair or pets. The context would also contain a premise such as the

following:

(27) Dyeing your hair pink is doing things your way.

The other illustration shows a tall office building. Out of

its many windows comes a typewriter, apparently thrown by someone

inside, breaking the window pane. The only way (25) and (26) could

be relevant would be to interpret them as describing a woman who

cannot be seen in the illustration but is throwing the typewriter

out oi the window. The audience would be encouraged to imagine a

situation in which a woman has been working over a typewriter but

has lost her temper and is throwing the typewriter out of a window.

The audience would also be encouraged to extend the context by

thinking of another situation in which a woman is repressed and not

able to show her temper and throw a typewriter out of a window; she

would be a woman who does not do things her way. The context chosen

for the interpretation oi the advertisement would include a premise

such as the following:

(28) Throwing your typewriter out of the window is doing

things your way.

 - 261 -

Here, individualism manifests itself in dyeing one's hair pink

and throwing a typewriter out of a window. However, these forms of

behaviour are not part oi mature and responsible adulthood. They

are trivial, irrelevant, irrational, and irresponsible.

Individualism might imply doing something that other people do not

do. The use of the concept in these advertisements has retained only

this implication: it shares only some implications of the concept,

and not others, such as being responsible for the consequences of

one's own behaviour.

The notion oi individualism here is reduced to that of doing

things which are childish and silly. The women portrayed in these

advertisements exercise their individuality by dyeing their hair

pirik and throwing a typewriter out of a window, rather than doing

something that is responsible and significant and yet unique. One

does not have to be outrageous in order to be individualistic.

However, in the advertisements, outrageousness is emphasised.

Thus, in Japanese advertisements, individualism means doing

what other people are doing, or elitism, which particulary manifests

itself in the posession of expensive Western goods. In the U.K.

also, individualism implies elitism, as we have seen in the Citizen

advertisement; or it implies doing something childish and

irrational. This last description fits a stereotype of women.

Thus, it can be argued that in the U.K., as well as in Japan, the

notion of individualism is being used to endorse stereotypes, rather

than reflecting real changes in women's status in the respective

 - 26? -

societies. The basis of the explanation of such apparently

inappropriate uses of the concept can be provided by context-

extension. Here again, as in the case of 'intelligence',

stereotypes of women help the audience process utterances which

involve the concept of individual is/" by providing easy access to a

context which includes certain assumptions about women, thus leading

to intended interpretations.

5.5. Feminism

In this section, I examine a case which cannot be explained in

terms oi loose use, but rather in terms of what Wilson would call

(personal communication) 'loose understanding'. One of the

situations where loose understanding may happen is where a word is

borrowed by one society from another, thus resulting in a change of

meaning.

First, let us look at the following example of a caption used

in a series of advertisements for Tokyo Gas, one of the city gas

companies. This series was so successful that it won the 1984 Asahi

Advertising Prize.
 - 263 -

(29) Toshi gasu-tte feminisuto ne.

city gas feminist tag-q

(Tokyo Gas)

City Gas is a feminist, isn't it?

Why is the gas company called a feminist? The answer may lie

in the pictures shown in the advertisements. One of them shows a

large pot boiling over on a gas cooker. Another depicts a woman

wearing an apron and holding a plate, doing the washing up. Yet

another shows the same woman's reflection in amirror,putting on

lipstick. She is not looking at herself, for her attention seems to

be diverted by something else. These advertisements are for a gas

cooker with a special sensor, which automatically stops the gas

supply when the tire is extinguished, a gas boiler with similar

sensor equipment, and an alarm for gas leaks. These are all new

apparatuses produced by the company. The company is described as

'feminist' on the grounds that it has come up with these instruments

to help women with their house work.

The use of the particle ne in (29) marks the utterance as

distinctively feminine. Thus, the audience is encouraged to imagine

a woman's voice praising the gas company for improving kitchen

equipment and thus helping women in their household chores.

 - 264 -

It is obvious that the loan word feminisuto in these

advertisements does not mean what the original English word means.

A Japanese audience would know what the word feminisuto means in

Japanese, whether or not they know what the word 'feminist' means in

English. However, it is not necessary for the audience to know the

meaning of the word in advance. If someone saw the advertisement

and read the description, she would realise that feminisuto does not

mean the same thing as the original English 'feminist'. Apparently,

it does not take very long for students doing Japanese at a British

university to learn, once they are in Japan, that a feminisuto is

someone who readily opens the door for women whereas a feminist

might get offended by such a male chauvinistic gesture. They learn

it not because of their previous knowledge of the English word, but

despite their knowledge about it, through the contexts in which the

Japanese word is used.

The word feminisuto is used in the advertisements to describe a

desirable quality. It is meant as a compliment to Tokyo Gas that

they should be called feminisuto. The company is being praised for

improving kitchen instruments, because they help women to perform

their household chores. It is chauvinism, rather than feminism, to

assume that a way of gaining women's praises is to improve kitchen

instruments. Despite the use of the word feminisuto, which

originated from English 'feminist', the attitude behind the

advertisements is sexist. These kitchen instruments are made for

the use of women. And some women do not even question why this is

so; they simply compliment the gas company by calling it feminisuto.

 - 265 -

For these women, housework is part of being a woman, just as putting

on lipstick is. According to Hosoi (1986:80), 89 per cent of women

accept that housework is the women's responsibility. The very

concept of a feminist might be foreign to them.

Vhat seems to have happened to the word is that when the

English word 'feminist' was absorbed into Japanese, it only retained

part of its meaning, namely, the part that involved being somehow

nice to women and being worth receiving women's praises. The part

that involved being nice to women by treating them equally was

ignored. From the point of view oi the original meaning of the

word, the Japanese usage was loose. And it would have been an

unacceptable loose use of the word within an English-speaking

community, for it contradicts many of the logical and contextual

implications of the word 'feminist', such as accepting women as

equal to men. When the word was introduced into Japanese to express

male chauvinistic behaviour, there was no clash between the way the

word was used and people's knowledge of the word, given that the

word did not exist in Japanese. Thus, it was a completely

acceptable use of the word. It could be argued that the Western

notion oi feminism was introduced into Japanese-speaking society in

a distorted form. Moreover, the Japanese meaning has ended up by

being more contrary than similar to the original meaning. It must

be noted that the word 'feminist' has a different meaning in

Japanese society in general, not just in Japanese advertising for

women.
 - 266 -

This meaning shift can be explained using Sperber and Wilson's

framework for word meaning. And so, it seems, can historical

meaning change and the meaning changes that arise during language

acquisition. Let us assume that most words express simple concepts,

which have logical entries which determine their analytic

implications, and encyclopaedic entries which contribute to their

contextual implications. When one first hears the word 'feminist',

for example, one opens a new conceptual address, and begins, in a

gradual way, to build its logical and encyclopaedic entries.

Possible changes of meaning may occur in at least the following

ways: (a) a hearer with radically different encyclopaedic

assumptions about the world may fail to notice some of the

contextual implications intended by the speaker, and recover other,

unintended ones which may become part of the encyclopaedic

stereotype associated with the concept; (b) something that for the

original speaker was an analytic implication, determined by a

logical entry, may be perceived as a contextual implication

determined by an encyclopaedic entry, and vice versa. In either

case, 'loose understanding' occurs, and as the result a change of

meaning happens.

In the case of 'feminist', I suspect that the first process

mentioned above happened. That is to say that a Japanese-speaking

hearer, who held drastically different encyclopaedic assumptions

about women and their position in society failed to understand a

subset of implications intended by the speaker, which have to do

with women being independent and having equal opportunities, and

only recovered other implications , which have to do with being nice

to women. This latter subset of implications may have been

unintended by the speaker, but they may have been relevant and

easily accessible to the hearer because of the kind of society she

was in.

Thus, in Japanese the loan word feminisuto means something

quite different from the original English word 'feminist'. The

Western notion of feminism was introduced into Japanese-speaking

society in a distorted form which was contrary to the original

meaning. the Japanese have managed to alter the meaning, not

necessarily by redefining it, but by altering its implications. It

could be argued that the distortion in the meaning of the word is an

indication of the fact that Japanese society was not prepared for

feminist concepts.

In this section, I have examined an example which involves a

shift of word meaning which may occur when a word is borrowed by one

language from another. In Japanese the word feminisuto (feminist)

is used in a way which is drastically different from the original.

I have argued that this shift can be explained in terms of what

Sperber and Wilson would call 'loose understanding', where a hearer

of a new word, due to different assumptions about the world in

general, fails to notice some of the contextual implications

intended by the speaker using the word, and only recovers other

implications unintended by the speaker, which become attached to the

concept. In this case, the word 'feminist' has retained

implications about being nice to women and inviting women's praises,

but has lost implications about treating women equal to men and

respecting women's rights. Wilson suggests (personal communication)

that loose interpretation is the source of much change in word

meaning: children understand adults loosely (i.e they recover some

analytic or contextual implications, but not others); foreigners

understand foreign words loosely, and so on. Relevance Theory

provides a comprehensive framework to analyse such changes in word

meaning.

I am aware that there is much more work to be done in this

area. However, I hope to have provided a sketch of how the change

of word meaning may be analysed in the new pragmatic framework.

This is no more than a suggestion for a whole new area of research.

5.6. Conclusion

In this chapter, I have considered some concrete examples of

advertisements using a small number of specific words and shown how

extension and shift of word meaning may be analysed in a framework

based on Relevance Theory by Sperber and Wilson (1986a). The words

which have been examined in this chapter are 'intelligent',

’individualistic1 and 'feminist'. I have deliberately selected an

extremely small data set to exemplify how different kinds of

extension of word meaning, which may result in the change of word

meaning, can be explained using the assumptions of the theory.

Section 5.3 discussed Japanese examples involving the word

'intelligent', and showed how the word is used as synonymous with

words such as 'feminine', 'sophisticated', and 'elegant', I have

argued that this can be analysed using the notion of 'loose use'

(Sperber and Wilson 1986a, 1986b), where some of contextual

implications associated with the word are abandoned, while others

are retained. Contextual implications about one's mental quality

are suppressed in the examples, whereas those about presentability

and desirability to men are emphasised.

Section 5.4 focused on 'individualism' and examined Japanese

and British examples. It was argued that in advertisements in both

societies 'individualism' is used as closely associated with

 - 270 -

eliti- sm, that is to say that those who can afford to buy special

expensive goods, often Western goods in the case of Japanese

examples, achieve 'individualism'. It was also argued that, rather

than reflecting true individualism, the word in the examples tends

to maintain stereotypes of women. My claim was that this may be

analysed as yet another type of loose use which involves the

addition of an extra premise to the context.

The loan word feminisuto (feminist) has been examined in

Section 5.5. Examples have shown that in Japan the word has a

drastically different meaning from the original. I have argued that

this can be analysed in terms of what Sperber and Wilson would call

(personal communication) 'loose understanding'. The analysis is

based on the assumption that some contextual implications associated

with the word intended by the speaker may be lost by the hearer

because she has assumptions, easily accessible, about the world

which are not shared by the speaker, and vice versa.

In my analysis of how these words are used in advertising, I

also demonstrated the ways in which women are portrayed. Contrary

to the frequent use of words such as 'intelligent' and

'individualistic', which might suggest new images of women, a close

examination reveals reinforcement of the traditional role models.

Women's 'intelligence' manifests itself in their choice of clothes

and cosmetics and their sweet smile. 'Individualism' turns out to

be a search for high fashion and social approval. A 'feminist' is

someone who provides sophisticated kitchen gadget and thus helps

 - 271 -

women to do what is regarded as their job, i.e. housework. Rather

than reflecting a real change in the attitude to women, these words

are used in ways which support and emphasise stereotypes of women.

The observation of the advertisements using words such as

'individualistic' and 'feminist' has shown ways in which a society

may absorb new notions from another society. Japanese society seems

to have coped with potential threats from the introduction of

concepts such as individualism and feminism by interpreting them

against the background of traditional values, or removing some of

the contextual implications attached to the concept which may

contradict conventional premises. Thus, Japanese group ideology and

male chauvinism have been rescued, and potential cultural clashes

have been resolved.

 - 272 -

EPILOGUE

The purpose of this thesis has been to evaluate some of the

basic assumptions of Relevance Theory and apply the theory to an

analysis of the language used by advertisers. In so doing, I hope

to have shed new light on some of the theoretical assumptions and to

have added to an understanding of the overall style of advertising

language.

The heterogeneous nature of the thesis is the result of an

attempt to analyse a recognisable style. I take style to be a

re.
jigsaw puzzle; a picture consisting of numerous pieces of diffent

shapes and colours. A particular style of writing is a complex

conglomerate of various aspects of language use. Chapters 3 - 5

have isolated for analysis some of the most striking

characteristics of one particular style. Thus, Chapter 3 examined

the use of the pun in advertising. This took us through

considerations of ambiguity and metaphor. The focus of Chapter 4

was audience manipulation via covert communication, as opposed to

ostensive communication which is the core of Sperber and Wilson's

analysis of communication. This gave us an opportunity to re­

examine fundamental properties of communication. Chapter 5

discussed how individual words can be used in extended ways, in one

case leading to a definitive shift in word meaning. In doing so, we

were faced with sociological considerations of the status of women.

 - 273 -

The apparently narrow analytical focus of these chapters in

this way has led us through a whole array of theoretical and

descriptive problems in semantics, pragmatics, rhetoric and

sociology. Despite the low respect which advertising is accorded in

the intellectual world, advertising has thus in a sense proved an

avenue toward understanding the human mind.

 - 2?4 -

Botes

1 Contrary to Sperber and Wilson's usage, I systematically consider

the speaker to be male and the hearer to be female, on the
grounds that advertisers are nearly always men and their targets
tend to be women.

2 Vowel sequences such as aa, ii, and oo are frequently considered

to be long vowels. A true long vowel, however, can only be
assigned a prominence, that is, high tone or stress, on its
initial element. But it is possible for the second element of
a so-called long vowel in Japanese to bear such prominence.
Thereiore, they should be written as sequences of two vowels.
 - 275 -

Bi bli o^rsphy

Aitchison, J. (1987), Words in the mind: An introduction to the

mental lexicon. Blackwell, Oxford.

Allwood, J. (1976), Linguistic communication as action and

cooperation: A study in pragmatics Gothenburg monographs in
linguistics 2. University of Gothenburg Press, Gothenburg.

Anon. (1984), 'Kookoku 1984' (Advertising 1984). In Amano, Y.

(ed. ) Kookoku Hihyoo (Advertising critiques) 67. Madora
Shuppan, Tokyo. p.6.
*>
Augarde, T. (1984), The Oxlord guide to word games. Oxford
University Press, Oxford.

Bach, K . , and Harnish R. (1979), Linguistic communication and

speech acts. MIT Press, Cambridge, Mass.

Baehr, H. (ed.) (1980), Women and media. Pergamon Press, Oxford.

Barthes, R. (1984a), Mythologies. Translated from the French 1954-

6 by Lavers, A. Granada, London.

Barthes, R. (1984b), 'Rhetoric of the image'. Translated from the

French 1964 by Heath, S. In Barthes 1984c. pp. 32-51.

Barthes, R. (1984c), Image music text. Translated and selected

from the French 1961-71, by Heath, S. Fontana, London.

Barthes, R. (1985), Elements of semiology. Translated from the

French 1964 by Lavers, A. and Smith, C. Hill and Vang, New York.

Eatsleer, J., Davis, T . , O'Rourke, R . , and Veedon, C. (1985),

Rewriting English: Cultural politics of gender and class.
Methuen, London.
fL
Beaugi^nde, R. de and Dressier, V. (1981), Introduction to text
linguistics. Longman, London.

Bencherif, S. and Tanaka, K. (1987), 'Covert forms of

communication' . Paper delivered to the Linguistic Association of
Great Britain, September 1987.

Berger, J. (1984), Ways of seeing. Penguin, London.

Bennett, J. (1976), Linguistic behaviour. Cambridge University

Press, Cambridge.

Blakemore, D. (1987), Semantic constraints on relevance. Blackwell,

Oxford.

^ , P. C(1 ^ ), lA^pxc-lp; rta, , or- * f~

S' (Ai-ke. . Cu.l.le.r, J. (eX ppv f^l ~
 - 276 -

Blonsky, M. (ed.) (1985), On signs: A semiotic reader. Blackwell,

Oxford.

Bolinger, D. (1980), Language - the loaded weapon: The use and

abuse of language today. Longman, London.

Brown, G . , and Yule, G. (1983), Discourse analysis. Cambridge

University Press, Cambridge.

Burchill, J. (1986), Girls of film.. Virgin, London.

Carston, R. (1985), 'A reanalysis of some "quantity implicatures"'.

Paper delivered to the Linguistic Association of Great Britain,
April, 1985.

Clark, H. (1982), 'The relevance of common ground: Comments on

Sperber and Wilson's paper'. In Smith, N. (ed.). pp. 124-131.

Clark, H. , and Carlson, T. (1981), 'Context for comprehension'. In

Long, J. and Baddeley, A., Attention and performance IX. L.
Erlbaum, Hillsdale, NJ. pp. 313-30.

Clark, H . , and Marshall, C. (1981), 'Definite reference and mutual

knowledge'. In Joshi, A., Webber, B. and Sag, I. (eds.),
Elements of discourse understanding. Cambridge University Press,
Cambridge. pp. 10-63.

Clark, H . , and Schunk, D. (1980), 'Polite responses to polite

requests.' Cognition 8.2. pp. 111-143.

Cole, P. (ed.) (1978), Syntax and semantics 9: Fragmatics.

Academic Press, New York.

Cole, P. (1981), Radical pragmatics. Academic Press, New York.

Cole, P. and Morgan, J. (eds.) (1975), Syntax and semantics 3:

Speech acts. Academic Press, New York.

Coltheart, M. (1984), 'Sensory memory: A tutorial review'. In

Bouma and Bouwhuis, (eds.), Attention and Performance X: Control
of language processes. Lawrence Erlbaum Associates, London,
pp. 257-285.

Coward, R. and Ellis, J. (1977), Language and materialism.

Routledge and Kegan Paul, London.

Crain, S. and Steedman, M. (1985), 'On not being led up the garden
path: The use of context by the psychological syntax processor',
in Dowty, D. , Karttunen, L. and Zwicky, A. (eds.), Natural
language parsing. Cambridge University Press. pp. 320-58.

Crompton, A. (1979), The craft of copywriting. Huchinson, London.

Crystal, D. and Davy, D. (1983), Investigating English style.

 - 27? -

Longman, Harlow.

Culler, J. <.1983), The pursuit of signs: Semiotics, literature,

deconstruction. Routledge and Kegan Paul, London.

Culler, J. (1983), Barthes. Fontana, London.

Culler, J. (ed.) (1988), On puns. Blackwell, Oxford.

Davis, H. and Valton, P. (eds.) (1984), Language, image, media.

Blackwell, Oxford.

Dijk, T. van. (1982), Text and context: Explorations in the

semantics and pragmatics of discourse. Longman, London.

Dijk, T. van. (ed.) (1985a), Handbook of discourse analysis, vol.

3: Discourse and dialogue. Academic Press, London.

Dijk, T. van. (ed.) (1985b), Handbook of discourse analysis, vol.

4: Discourse analysis in society. Academic Press, London.

Douglas, M. (1968), 'The social control of cognition: Some factors

in joke perception'. Man 3. pp. 361-376.

Dyer, G. (1983), Advertising as communication. Methuen, London.

Eco, U. (1983), The role of the reader. Hutchinson, London.

Fiske, J. (1984), Introduction to communication studies. Methuen,

London.

Fiske, J. and Hartley, J. (1978), Heading television. Methuen,

London.

Fodor, J.A. (1975), The language of thought. Harvester Press,

Sussex.

Fodor, J.A. (1981), Representations. Harvester Press, Sussex.

Fodor, J.A. (1983), The modularity of mind. MIT Press, Cambirdge,

Mass.

Fodor, J.A., Garrett, M . , Walker, E. and Parkes, C. (1980),

'Against definitions’. Cognition 6.3. pp. 263-367.

Fowler, R. (1985), 'Power'. In Dijk, T. van, (ed.), (vol.4). pp.

61-82.

Gazdar, G. (1979), Fragmatics: Implicature, presupposition and

logical f-orm. Academic Press, New York.

Gazdar, G. and Good, D. (1982), 'On a notion of relevance'. In

Smith, N. (ed.). pp. 88-100.
 - 278 -

Geis, M. (1982), The language of television advertising. Academic

Press, New York.

Giglioli, P. (ed.) (1972), Language and social context. Penguin,

London.

Goffman, E. (1976), Gender advertisements. Macmillan, London.

Goffman, E. (1981), Forms of talk. University of Pennsylvania

Press, Philadelphia.

Gordon, D. and Lakoff, G. (1975), 'Conversational postulates'. In

Cole P. and Morgan J. (eds.). pp, 83-126.

Green, G. and Morgan, J. (1981), 'Pragmatics, grammar and

discourse'. In Cole, P. (ed.). pp. 167-181.

Greene, J. (1986), Language undestanding: A cognitive approach.

Open University Presd, Milton Keynes.

Grice, H.P. (1957), 'Meaning'. Philosophical Review 66. pp. 377-

88.

Grice, H.P. (1968), 'Utterer's meaning and intentions'.

Philosophical Review 78. pp. 147-77.

Grice, H.P. (1975), 'Logic and coversation'. In Cole, P. and

Morgan, J. (eds.) pp. 41-58.

Grice, H.P. (1978), 'Further notes on logic and conversation'. In

Cole, P. (ed.) pp. 183-98.

Grice, H.P. (1981), 'Presupposition and conversational

implicature'. In Cole, P. pp. 183-98.

Grice, H.P. (1982), 'Meaning revisited'. In Smith, H. (ed.) pp.

223-43.

Guiraud, P. (1981), Semiology. Translated from the French 1971 by

Gross, G. Routledge and Kegan Paul, London.

Guraperz, J. (1982), Discourse strategies. Cambridge University

Press, Cambridge.

Hall, S., Hobson, D . , Lowe, A. and Wills, P.(eds.) (1984) Culture,

media, language. Hutchinson, London.

Hawkes, T. (1985), Structuralism and semiotics. Methuen, London.

Hebdige, D. (1979), Subculture: The meaning of style. Methuen,

London.

Heller, L. (1974), 'Toward a general typology ofthe pun'.

Language and Style 7. pp. 271-82.
 - 279 -

Hickey, Leo. (1984), 'Being positive about thenegative in

advertising'. International Journal of Advertising 3. pp. 369-
72.

Hill, J.'Language, culture, and world view'. In Newmeyer, F.

(ed.), Linguistics: The Cambridge Survey 4: Language: The socio­
cultural context. pp. 14-36.

Hoshii, I. (1986), The world of sex Vol. 1: Sexual equality,

Paul Norobury, Kent.

Hunt, A. (1983), The language of television: Uses and abuses.

Methuen, London.

Hyman, L. (1975), Phonology: Theory and analysis. Holt, Rinehart

and Vinston, New York.

Jakobson, R. (1960), 'Linguistics and poetics'. In Sebeok, T.

(ed.) pp. 350-377.

Jamieson, G. (1985), Communication and persuasion. Groom Helm,

London.

Johnson-Laird, P. (1982), 'Mutual ignorance: Comments on Clark and

Carlson's paper'. In Smith, N. (ed.) pp. 40-45.

Joshi, A., Webber, B. , and Sag, I. (eds.) (1981), Elements of

discourse understanding. Cambridge University Press, Cambridge.

Katz, J. and Fodor, J. (1963), 'The structure of a semantic

theory'. Language 3.2. *pp. 170-210.

Kelly, L. (1971), 'Punning and the linguistic sign'. Linguistics

66. pp. 5-11.

Kempson, R. (1977),Semantic theory. Cambridge University Press.

Cambridge.

Kempson, R. (1986), 'Ambiguity and the semantics-pragmatics

distinction'. In Travis (ed.) pp. 77-103.

Kess, J., Copeland, A. and Hoppe, R. (1984), 'Intentional ambiguity

as verbal sleight of hand in commercial advertising'. Grazer
Linguistische Studien 22. pp. 147-66.

Key, W. (1973), Subliminal seduction. Signet, New York.

Kitamura, H. , Yamaji, R. and Tabuki, H. (eds.) (1981), Kookoku

Kyatchifureezu (Advertising catchphrase). Tokyo.

Kress, G. (1985), 'Ideological structures in discourse'. In Dijk,

T. van (ed.) (vol.4) pp. 27-42.

Labov, W. (1973), 'The boundaries of words and their meanings'. In

 _ non .

Bailey, C. and Shuy, R. , (eds.), New ways of analyzing variation

in English. Gerogetown University Press, Washington, D.C. pp.
340-371.

Lakoff, G. (1972), 'Hedges: A study in meaning criteria and the

logic of fuzzy concepts'. Chicago Linguistics Society. pp.183-
228.

Lakoff, R. (1972), 'Language in context'. Language 48.4. pp. 907-

27.

Lakoff, R. (1973), 'Questionable answers and answerable questions'.

In Kachru, B. et al. (eds.), Issues in linguistics: Papers in
Honor of Henry and Eenee Kahane. University of Illinois Press,
Urbana. pp. 453-67.

Lakoff, R. (1975), Language and woman'a place. Harper and Row,

Hew York.

Lakoff, R. (1982), 'Persuasive discourse and ordinary conversation,

with examples from advertising'. In Tannen, D. (ed.), Analyzing
discourse: Text and talk. Gerogetowrn University Press,
Washington, D.C. pp. 25-42.

Leach, E. (1976), Culture and communication: The logic by which

symbols are connected. Cambridge University Press, Cambridge.

Leech, G. (1966), English in advertising: A linguistic study of

advertising in Great Britain. Longmans, London.

Leech, G. (1969), A linguistic guide to English poetry. Longman,

London.

Leech, G. (1981) (2nd edition) Semantics. Penguin, London.

Leech, G. (1963), Principles of pragmatics. Longman, London.

Leech, G. and Short, M. (eds.) (1984), Style in fiction: A

linguistic introduction to English fictional prose. Longman,
London.

Lehrer, A. (1983), Vine and conversation. Indiana University

Press, Bloomington.

Levinson, S. (1979), 'Activity types and language'. Linguistics

17. pp. 365-99.

Levinson, S. (1983), Pragmatics. Cambridge University Press,

Cambridge.

Lyons, J. (1979), Introduction to theoretical linguistics.

Cambridge University Press, Cambridge.

Lyons, J. (1977), Semantics. Cambridge University Press,

^ fc-. C ) , 1/ V t k r e p e (e g <’ C f a.*p«.cTS cf Cd-Te-^ e i'f < j-

iMacA ^ ('A (€-{.), tCtiiMS U St*dy »f

e. , ; AliT /VfeSS.
 Cambridge.

Lyons, J. (1981), Language, meaning and context. Fontana, London.

Margalit, A. (ed. ) (1976), Meaning and use. D. Reidel, London.

Marslen-WiIson, V. and Tyler, L.(1980), 'The temporal structure of

spoken language understanding'. Cognition 8.1. pp. 1-72.

Marslen-WiIson, W. and Tyler, L. (1981), 'Central processes in

speech understanding'. Philosophical transactions of the Royal
Society London. pp. 317-32.

Milium, T. (1975), Images of woman: advertising in woman's

magazines. Chatto and Windus, London.

Moeran, B. (1983), 'The language of Japanese tourism'. Annals of

Tourism Research 10. pp. 93-108.

Moeran, B. (1984), 'Individual, group and seishin: Japan's internal

cultural d eb ate'. Man 19. pp. 252-66.

McCawley, J. (1978), 'Conversational implicature and the lexicon'.

In Cole, P. (ed.) pp. 245-59.

Nash, W. (1985), The language of humour: Style and technique in

comic discourse. Longman, London.

Norrick, N. (1984), 'Stock conversational witticisms'. Journal of

pragmatics 8. pp. 195-209.

Ogden, C. and Richards, I. (1985), The meaning of meaning: A study

of the influence of language upon thought and the science of
symbolism. Ark, London.

Packard, V. (1961) (Edition for the 1980'S), The hidden persuaders.

Penguin, London.

Parret, Herman. (1983), Semiotics and pragmatics. John Benjamins,

Amsterdam.

Pateman, T. (1960), Language, truth and politics: Towards a radical

theory of communication. John Stroud, London.

Pateman, T. (1981), 'Communicating with computer programs'.

Language and Communication 1. pp. 3-12.

Pateman, T. (1983), 'How is understanding an advertisement

possible?' In Davis, H. and Walton, P. (eds.) pp. 187-204.

Pateman, T. (1987), Language in mind and language in society.

Clarendon Press, Oxford.

Pulman, S. (1963), Word meaning and belief. Croom Helm, London.

 - 282 -

Redfern, V. (1982), 'Guano of the mind: Puns in advertising'.

Language and Communication 2.3. pp. 269-76.

Rediern, V. (1985), Puns. Blackwell, Oxford.

Rein, D. (1982), The language of advertising and merchandising in

English. Regents Publishing Company, New York.

Rey, G. (1983), 'Concepts and stereotypes'. Cognition 15. pp.

237-62.

Roth, I. and Frisby, J. (1986), Perception and representation: A

Cognitive approach. Open University Press, Milton Keynes.

Rusiecki, J. (1985), Adjectives and comparison in English: A

semantic study. Longman, London.

Ruskin, V. (1985),Semantic mechanisms of humor. D. Reidel,

Dordrecht.

Sacks, H. (1974), 'An analysis of the course of a jok e’s telling in

conversation'. In Bawman, R. and Sherzer, J. (eds.), Exploration
in the Ethnography of Speaking. Cambridge University Press,
Cambridge. pp. 337-53.

Sadock, J. (1978), 'On testing for conversational implicature'. In

Cole, P. (ed.) pp. 281-97.

Saussure, F. de (1974), Course in general linguistics. Translated

irom the French 1916 by Baskin, V. Peter Owen, Oxford.

Schank, R.and Abelson, R. (1977), 'Scripts, plans and knowledge',

in Johnson-Laird P. and Vason, P. (eds.), Thinking: Readings in
cognitive science. Cambridge University Press. pp. 421-32.

Schiffer, S. (1972), Meaning. Clarendon Press, Oxford.

Schmidt, R. and Kess, J. (1986), Television advertising and

televangeiism: Discourse analysis of persuasive language. John
Benjamins, Amsterdam.

Schudson, M. (1984), Advertising, the uneasy persuasion: Its

dubious impact on American society. Basic Books, New York.

Searle, J. (ed. ) (1971), The philosophy of language. Oxford

University Press, Oxford.

Searle, J. (1976), 'Intentionality and the use of language'. In

Margalit, A. (ed.), Meaning and use. D. Reidel, London. pp.
181-97.

Searle, J. (1983), Intentionality. Cambridge University Press,

Cambridge.
 - 283 -

Sebeok, T. (ed.) Style in language. MIT Press, Cambridge, Mass.

Seidal, G. (1985), ‘Political discourse analysis'. In Dijk, T. van

(ed.) (vol.4) pp. 43-59.

Sherry, J. (1987), 'May your life be marvelous: English language

labelling and the semiotics of Japanese promotion' . Journal of
consumer research 14. pp. 174-88.

Shezer, J. (1978), 'Oh! That's a pun and I didn't mean it'.

Semiotica 22:3/4. pp. 335-50.

S h ^ e r , J. (1985), 'Puns and jokes'. In Dijk, T. van (ed. ) (vol. 3)

pp. 213-21.

Shimamori, M. (1984), Kookoku no naka no onna-tachi (Women in

advertising). Yamato Shoboo, Tokyo.

Sinclair, J. and Coulthard, R. (1975), Towards an analysis of

discourse: The English used by teachers and pupils. Oxford
University Press, Oxford.

Smith, N. (ed.) (1982), Mutual knowledge. Academic Press, London.

Smith, P. (1985), Language, the sexes and society. Blackwell,

Oxiord.

Smith, R. (1987), 'Gender inequality in contemporary Japan'.

Journal of Japanese Studies 13.1. pp. 1-25.

Spender, D. (1985) (2nd edition), Man made language. Routledge and

Kegan Paul, London,

Sperber, D. (1975), Rethinking symbolism. Cambridge University

Press, Cambridge.

Sperber, D . , and Wilson, D. (1981a), 'Irony and the use-mention

distinction'. In Cole, P. (ed.) pp. 295-318.

Sperber, D. and Wilson, D. (1981b), 'Pragmatics'. Cognition 10.4.

pp. 281-266.

Sperber, D. and Wilson, D. (1982), 'Mutual knowledge and relevance

in theories of comprehension'. In Smith, N. (ed.) pp. 61-131.

Sperber, D. and Wilson, D. (1986a), Relevance: communication and

cognition. Blackwell, Oxford.

Sperber, D. and Wilson, D. (1986b), 'Loose Talk'. The Aristotelian

Society vol. LXXXVI.

Sperber, D. and Wilson, D. (1987), 'Authors' response'. Behavioral

and Brain Sciences 10.4. pp. 736-54.
 - 284 -

Strawson, P. (1964), 'Intention and convention in speech acts'. In

Searle J. (ed.) pp. 23-38.

Tanaka, K. (1987), 'The pun in advertising: a pragmatic approach'.

Paper delivered to the Linguistic Association of Great Britain,
September 1987.

Tanaka, K. (forthcoming), '"Intelligent elegance": images of women

in Japanese advertising'. In Moeran, B . , Ben-Ari, E. and
Valentine, J., eds. Unwrapping Japan. Manchester University
Press, Manchester.

Travis, C. (ed.) (1986), Meaning and interpretation. Blackwell,

Oxford.

Vestergaard, T. and Schroder, K. (1985), The language of

advertising. Blackwell, Oxford.

Vatkins, J. (1959), The 100 greatest advertisements: Vho wrote then

and what they did. Dover, Mew York.

Vest, C. and Zimmerman, D. (1985), 'Gender, language, and

discourse. In Dijk, T. van (ed.) (vol.4) pp. 103-124.

Viddowson, P. (1982) (ed.), Re-reading English. Menthuen, London.

Williams, Raymond. (1981), Key words: A vocabulary of culture and

society. Fontana, London.

Williamson, Judith. (1983), Decoding advertisements: Ideology and

meaning in advertising. Marion Boyars, London.

Wilson, D . ,and Sperber, D. (1981), 'On Grice's theory of

conversation'. In Verth, P. (ed), Conversation and discourse.
Croom Helm, London. pp. 155-78.

Wilson, D . ,and Sperber, D. (1986), 'Inference and implicature'.

In Travis, C. (ed.) pp. 45-75.

Wright, R. (1975), 'Meaning-nn and conversational implicature'. In

Cole, P. and Morgan, J. (eds.) pp. 363-82.