Gender and Humour: Beyond a Joke

Jennifer Hay

Submitted to the Victoria University of Wellington

in ful
lment of the requirements for the degree of
Master of Arts in Linguistics.

June 13, 1995

Abstract

This thesis investigates the interaction between gender and humour in spontaneous New Zealand English. Tapes were collected of men and women in single
sex and mixed groups and instances of humour were identi
ed and transcribed.
The taping situation proved more comfortable for the female groups and the
mixed groups than for the male groups. Male groups tend to be more task-based,
and so often treated the taping as a task to be completed. The tape-recorder
was the primary reason for their conversation, whereas for other groups, the
tape-recorder was incidental to the interaction.
A taxonomy of types of humour was constructed. Log-linear modelling showed
that both gender and group composition a ect the type of humour used. The
most statistically signi
cant results were that women are more likely to use observational humour than men, insults and roleplay are more likely to occur in
single sex conversations and humour involving quotes or vulgarity seldom occurs
in mixed interaction and is more often used by men than by women. Other
trends indicated that anecdotes and fantasy humour are more likely to occur in
mixed interaction than single sex interaction, and men tend to use more roleplay
and wordplay than women whereas women are more likely to use jocular insults.
Most of the humour in the corpus was fresh humour, although men were shown
more likely to use external source humour than women. This supports prior
research which indicates men's humour is likely to be less context bound than
women's and that they are more likely than women to discuss things they have
seen or read.
An analysis of the functions of humour showed women more likely to use humour
to create solidarity than men. In particular, women often used humour to share
personal information about themselves. When using humour for solidarity based
purposes men were more likely to capitalise on shared experiences or highlight
similarities. Power based strategies such as controlling or fostering conict occurred extremely rarely. Men were more likely than women to use humour only
for the general function of increasing status and solidarity and performing positive work on their personal identity. Men were more likely to use humour to
cope with a situational problem, whereas females more often used general coping
strategies which enabled them to cope with problems not speci
c to the immedi-

ate situation. The results showed that, contrary to literature indicating teasing is
a predominantly male activity, men and women tease both in a powerful manner
and to create or maintain solidarity, although this strategy is largely restricted
to single sex groups.
The humour used by the females was more likely to be about a topic involving people than men's humour. The men joked more about work, computers,
television shows, movies or books, and alcohol.
Speakers were much more likely to use humour focussed on a same-sex group
member when in single sex groups than in mixed groups. In mixed interaction
speakers maintained gender boundaries by focussing humour on group members
of the opposite sex.
In mixed groups men contributed more humour than women. This supports
claims that, as joking has been regarded a masculine activity, women are reserved
with their humour when in mixed groups.
Analysis of how speakers present themselves through humour can give insight
into the value systems of the group. The humour in this corpus indicates that
the ability to laugh at oneself and to be open and honest are valued more by
female groups than male groups, and wit and \coolness" are traits valued more
by males.
The
nal aspect of humour investigated is that of humour support. It is shown
that the issue of humour support is much more complicated than many researchers claim. There are numerous support strategies available to any audience, some of which provide stronger support than others. The most appropriate
strategies will vary with context. Some types of humour do not require any support at all, and there are many reasons why an audience may choose not to
support an attempt at humour.
The results presented in this thesis are intended as ground work on which to
base further qualitative investigation of how gendered identities are constructed
via humour.

Acknowledgements
In Which Jen Admits that She had a Little Bit Of Help

This thesis could not have been completed without the time, help, advice and
energy of numerous people.
I am extremely indebted to those friends who succumbed to my bullying and
taped their conversations to contribute to my corpus. Thanks are also due to
Anita Easton, for providing me with several pre-transcribed conversations, and
to the Wellington Corpus of Spoken New Zealand English for allowing me access
to their data, and lending me a transcribing machine.
This thesis has bene
tted greatly from the advice and comments of several seminar and conference audiences. I would like to thank Miriam Meyerho and
Maria Stubbe for providing references, comments, enthusiasm and support. Useful references were also provided by Chris Lane, Laurie Bauer, Gary Johnson,
David Crabbe and Robert Sigley. Alan Tam converted my graphs to postscript,
and Annie McGregor provided papers, references, enthusiasm and good humour
when it was most needed. I'd also like to thank Alex Heatley for his thorough
attempt at retrieving my literature review from the back-up tapes!
Sally McConnell-Ginet provided much needed enlightenment as to the structure
of my function taxonomy, a willing ear, and useful advice and comments.
Many thanks are due to the sta of the linguistics department at Victoria, who
provided much support and encouragement. I would also like to thank Alastair
Gray for his advice and guidance on the statistics in this thesis.
Financially I am indebted to the BNZ and Victoria University for their generous
scholarships, the internal grants committee for covering thesis-related expenses
and the Linguistic Society of New Zealand for funding my trip to the conference
in Lincoln. I'd also like to thank Liz Pearce for the chance to accumulate some
teaching experience, Allan Bell for providing tapes on which to practise my
transcribing, and Janet Holmes and Allan Bell for giving me an excuse to spend
a few days in the sun.
I am eternally grateful to Mrs Wasmuth, my sixth form English teacher, who
provided boundless inspiration, enthusiasm and encouragement. Had it not been

for her class, I may well have been
nishing an MSc in computer science right
now. I can not express my gratitude enough!
I'm grateful to the many people who regularly stopped by the oce to chat,
providing a welcome distraction and relief. In particular such thanks are due to
Gary Johnson and Allan Bell.
For co ee, support, meals, chats, and laughter, I would like to thank Bernadette
Vine, Claire Oram, Andrew Chick, Linton Miller, Judi Lapsley Miller, David
Tulloch, Jenny Freeman, Lindy Smith, and Phil Plasma.
In particular I am indebted to Aaron, resident computer consultant, translator,
advisor, proofreader, friend, chau eur and chef-extraordinaire.
I am extremely grateful to my wonderful mother who patiently proofread the
draft of my thesis. Bernadette Vine and Gary Johnson also proofread parts of
this thesis, and Bernadette checked through my many tables and sacri
ced some
sleep to help with the last minute rush! Chris Lane provided useful comments
on the
nal draft.
My greatest debt is to my supervisor Janet Holmes. Her support and enthusiasm
for this project were constants I could always rely on. I would like to thank her
for providing both valuable guidance and the freedom to wander where my data
and whims lead me, for setting tight deadlines and understanding when I often
missed them, for
nding my work interesting even when I found it most boring,
and for being a mentor and a friend.
Finally, I would like to dedicate this thesis to my Father, who would have been
proud.

Contents
1 Introduction

2 Review of humour research

2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8

Introduction : : : : : : : : : : : : : : : : : : :
De
ning Humour : : : : : : : : : : : : : : : :
Theories of humour : : : : : : : : : : : : : : :
Humour research using elicitation or surveys :
Research on conversational joking : : : : : : :
Literature on types of humour : : : : : : : : :
The functions of humour : : : : : : : : : : : :
Conclusion : : : : : : : : : : : : : : : : : : :

: : : : : : : : : : :
: : : : : : : : : : :
: : : : : : : : : : :
: : : : : : : : : : :
: : : : : : : : : : :
: : : : : : : : : : :
: : : : : : : : : : :
: : : : : : : : : : :

3 Gender, Language and Humour

3.1 Introduction : : : : : : : : : : : : : : : : : : : : : : :
3.2 Men and Women in Conversation : : : : : : : : : : :
3.3 Explanations of di erence : : : : : : : : : : : : : : :
3.3.1 Female De
cit : : : : : : : : : : : : : : : : :
3.3.2 Di erence or Dominance? An ongoing debate
3.4 Conceptualisations of Gender { Moving Forward : :
3.5 Gender and Humour : : : : : : : : : : : : : : : : : :
3.6 Conclusion : : : : : : : : : : : : : : : : : : : : : : :

3
4
7
8
8
10
12
14

15
: : : : : : :
: : : : : : :
: : : : : : :
: : : : : : :
: : : : : : :
: : : : : : :
: : : : : : :
: : : : : : :

15
15
18
18
18
20
24
32

CONTENTS

ii

4 Methodology

4.1 Introduction : : : : : : : : : : : : : : : : :
4.2 Control of Variables : : : : : : : : : : : :
4.2.1 Number of Speakers : : : : : : : :
4.2.2 Sex : : : : : : : : : : : : : : : : : :
4.2.3 Ethnicity : : : : : : : : : : : : : :
4.2.4 Age : : : : : : : : : : : : : : : : :
4.2.5 Education : : : : : : : : : : : : : :
4.2.6 Intimacy of Speakers : : : : : : : :
4.2.7 Setting : : : : : : : : : : : : : : :
4.2.8 Alcohol : : : : : : : : : : : : : : :
4.3 Data Collection : : : : : : : : : : : : : : :
4.3.1 Speakers' knowledge of the project
4.4 Selection of examples : : : : : : : : : : : :
4.5 Methodological problems : : : : : : : : : :

35
: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :

5 Statistical methods

5.1 Log-linear modelling : : : : : : : : : : : : : : : : : : : : : : :

5.2 Clustering e ects : : : : : : : : : : : : : : : : : : : : : : : : :
5.2.1 Clustering due to individual behaviour : : : : : : : : :
5.2.2 Clustering due to conversational ow : : : : : : : : : :
5.2.3 Clustering due to complex examples : : : : : : : : : :
5.3 A statistical check on clustering : : : : : : : : : : : : : : : : :
5.4 The suitability of statistics for analysing conversational data

51
: :
: :
: :
: :
: :
: :
: :

6 Types of humour

6.1 Introduction : : : : : : :
6.2 The Taxonomy of Types
6.3 De
ning the Types : : :
6.3.1 Anecdote : : : :

35
35
35
36
36
36
37
37
38
38
38
39
40
41
51
59
59
59
60
60
60

63
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :

ii

63
65
65
65

CONTENTS

iii

6.3.2 Fantasy : : : : :
6.3.3 Insult : : : : : :
6.3.4 Irony : : : : : : :
6.3.5 Jokes : : : : : :
6.3.6 Observational : :
6.3.7 Quote : : : : : :
6.3.8 Roleplay : : : : :
6.3.9 Self Deprecation
6.3.10 Vulgarity : : : :
6.3.11 Wordplay : : : :
6.3.12 Other : : : : : :
6.4 Results : : : : : : : : : :
6.5 Discussion : : : : : : : :
6.6 Conclusion : : : : : : :

: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : :

7 Fresh Humour and Seconds Humour

7.1 Background
7.2 Results : : :
7.3 Discussion :

: : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : : : : : : : : : :

8 The Functions of Humour

8.1 Introduction : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
8.2 The Framework : : : : : : : : : : : : : : : : : : : : : : : : : : : :
8.3 Solidarity : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
8.3.1 To share : : : : : : : : : : : : : : : : : : : : : : : : : : : :
8.3.2 To highlight similarities or capitalise on shared experiences
8.3.3 To clarify and maintain boundaries : : : : : : : : : : : : :
8.3.4 To tease (S) : : : : : : : : : : : : : : : : : : : : : : : : : :
8.4 Power : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
8.4.1 To foster conict : : : : : : : : : : : : : : : : : : : : : : :
iii

68
70
71
73
74
75
76
78
78
79
81
82
84
89

91

91
95
96

97

97
97
99
100
100
102
102
103
104

CONTENTS

iv

8.5

8.6
8.7
8.8

8.9

8.10

8.4.2 To control : : : : : : : : : : : : : : :
8.4.3 To challenge and set boundaries : :
8.4.4 To tease (P) : : : : : : : : : : : : : :
Psychological Functions : : : : : : : : : : :
8.5.1 To defend : : : : : : : : : : : : : : :
8.5.2 To cope with a situational problem :
8.5.3 To cope with a general problem : : :
Results : : : : : : : : : : : : : : : : : : : : :
Overall results : : : : : : : : : : : : : : : :
Results of individual strategies : : : : : : :
8.8.1 Results of solidarity based strategies
8.8.2 Results of power based strategies : :
8.8.3 Results of psychological functions : :
Discussion : : : : : : : : : : : : : : : : : : :
8.9.1 The General Function : : : : : : : :
8.9.2 Solidarity : : : : : : : : : : : : : : :
8.9.3 Power : : : : : : : : : : : : : : : : :
8.9.4 Psychological Functions : : : : : : :
Conclusion : : : : : : : : : : : : : : : : : :

: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :
: : : : : : : : : : : :

9 Focus and Topic

9.1 Introduction : : : : : : : : : : : :
9.2 Focus : : : : : : : : : : : : : : :
9.2.1 Classi
cation and Results
9.2.2 Discussion : : : : : : : : :
9.3 Topic : : : : : : : : : : : : : : : :
9.3.1 Discussion : : : : : : : : :
9.4 Conclusion : : : : : : : : : : : :

iv

106
107
108
108
109
109
110
111
111
114
114
115
117
118
118
119
122
123
123

125
: : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : :

125
125
126
128
129
132
136

CONTENTS

10 Amount of Humour

10.1 Introduction : : : : : : : : : : : : : : : : : : : : :
10.2 Quantifying \Amount" of humour : : : : : : : :
10.2.1 Counting instances of humour : : : : : : :
10.2.2 Calculating relative humour contribution
10.3 Results : : : : : : : : : : : : : : : : : : : : : : : :
10.4 Discussion : : : : : : : : : : : : : : : : : : : : : :

137
: : : : : : : : :
: : : : : : : : :
: : : : : : : : :
: : : : : : : : :
: : : : : : : : :
: : : : : : : : :

11 Social and Personal Identity

11.1
11.2
11.3
11.4
11.5
11.6

Background : : : : : : : :
Ability to laugh at oneself
Wit/Cleverness : : : : : :
Openness/Honesty : : : :
Coolness : : : : : : : : : :
Conclusion : : : : : : : :

137
139
139
139
140
143

145
: : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : :
: : : : : : : : : : : : : : : : : : : : : :

12 Laughter and humour support strategies

12.1
12.2
12.3
12.4

Introduction : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
Laughter as humour support : : : : : : : : : : : : : : : : : : :
Counting laughter : : : : : : : : : : : : : : : : : : : : : : : : :
Humour support strategies : : : : : : : : : : : : : : : : : : : : :
12.4.1 Contributing more humour : : : : : : : : : : : : : : : :
12.4.2 Echo : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
12.4.3 O er sympathy or contradict self-deprecating humour :
12.4.4 Overlap and heightened involvement in the conversation
12.4.5 Non-verbal : : : : : : : : : : : : : : : : : : : : : : : : :
12.5 When explicit support is not needed : : : : : : : : : : : : : : :
12.5.1 The humour is a support strategy itself : : : : : : : : :
12.5.2 Irony : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
12.6 Failed humour : : : : : : : : : : : : : : : : : : : : : : : : : : :
v

145
147
148
149
151
153

155
:
:
:
:
:
:
:
:
:
:
:
:
:

155
155
157
162
162
166
167
169
169
170
170
170
171

CONTENTS

vi

12.6.1 Insucient contextualisation : : : : : : : : : : : : : : : : 171

12.6.2 Being too late or reviving \dead" humour : : : : : : : : : 172
12.6.3 Assuming too much background knowledge : : : : : : : : 173
12.6.4 Misjudging relation between speaker and audience : : : : 175
12.6.5 Negatively teasing someone present : : : : : : : : : : : : : 176
12.6.6 Trying to gain membership of exclusive sub-group : : : : 177
12.6.7 Disrupting serious conversation : : : : : : : : : : : : : : : 178
12.6.8 Portraying oneself inappropriately for one's status or gender179
12.6.9 Other unsupported humour : : : : : : : : : : : : : : : : : 181
12.7 Conclusion : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 181

13 Conclusion

183

A Key to Transcription Conventions

189

Bibliography

193

vi

List of Figures
8.1 The functions of humour :

: : : : : : : : : : : : : : : : : : : : : :

10.1 Humour contribution by individuals in mixed groups : :

10.2 Humour contribution by individuals in single sex groups

vii

: : : : :
: : : : :

98
142
142

List of Tables
4.1 Pre-collected data available

: : : : : : : : : : : : : : : : : : : : :

39

5.1 Example model: Observed frequencies : : : : : : : : : : : : : : :

5.2 Example model: Predicted frequencies according to the independence model : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
5.3 Example model: SAS analysis for independence model : : : : : :
5.4 Example model: SAS analysis of model showing interaction between all variables : : : : : : : : : : : : : : : : : : : : : : : : : :
5.5 Example model: SAS analysis of model of main e ects for hair :
5.6 Example model: Predicted frequencies for model of main e ects
for hair : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :
5.7 Example model: Predicted frequencies for model of main e ects
for hair { collapsed : : : : : : : : : : : : : : : : : : : : : : : : : :
5.8 Example model: Conditional probabilities : : : : : : : : : : : : :

53

6.1 Probability
gures for types of humour :

6.2 Odds ratios for types of humour : : : : :

82
83

7.1 Probabilities for \freshness" of humour :

7.2 Odds ratios for \freshness" of humour :
8.1
8.2
8.3
8.4

: : : : : : : : : : : : : :
: : : : : : : : : : : : : :

: : : : : : : : : : : : : :
: : : : : : : : : : : : : :

Results of possible models for overall function

Probability
gures for overall function : : : :
Odds ratios for overall function : : : : : : : :
Probabilities for solidarity based strategies : :
ix

: : : : : : : : : : :
: : : : : : : : : : :
: : : : : : : : : : :
: : : : : : : : : : :

53
54
55
56
57
58
58

95
95
112
112
113
114

CONTENTS

x
8.5
8.6
8.7
8.8
8.9

Odds ratios for solidarity based strategies

Probabilities for power based strategies :
Odds ratios for power based strategies : :
Probabilities for psychological functions :
Odds ratios for psychological functions : :

9.1 Probability
gures for focus of humour :

9.2 Odds ratios for focus of humour : : : : :
9.3 Percentage of humour on di erent topics

: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :
: : : : : : : : : : : : :

: : : : : : : : : : : : : :
: : : : : : : : : : : : : :
: : : : : : : : : : : : : :

10.1 Number of high and low contributors of humour

12.1 Percentage of humour supported by laughter

115
116
116
117
117
127
127
131

: : : : : : : : :

141

: : : : : : : : : : :

159

Chapter 1

Introduction
Interest in language and gender research has exploded in the last few decades.
Researchers have examined the relationship between language and gender using
numerous methodologies and with regard to a diverse set of variables. Humour
research has also become increasingly proli
c, to the extent that there is now
a regular journal devoted to the subject. The intersection between these
elds,
however, is relatively unexplored. There are many myths and stereotypes postulating gender di erences in use and appreciation of humour, but there is a
conspicuous lack of studies investigating how men and women use humour in
spontaneous, natural, spoken English. This thesis is an initial step towards
lling this gap.
In chapter 2 I briey review relevant literature in humour research, giving an indication of past
ndings and areas which would bene
t from closer investigation.
Chapter 3 outlines recent developments in the area of gender and language, and
goes on to discuss research which has dealt speci
cally with gender and humour.
The speci
c methodology employed in this study is described in chapter 4. The
design of the project, the data collection, the variables controlled, and resulting
methodological problems are addressed.
The statistical technique with which I have chosen to analyse several aspects of
my data may be unfamiliar to some readers. Chapter 5 outlines the reasoning
behind log-linear modelling and illustrates the technique with a simple example.
The discussion is designed to equip readers with the background necessary to
1

CHAPTER 1. INTRODUCTION

adequately interpret the results presented in the following chapters.

Chapter 6 outlines a taxonomy for categorising types of conversational humour.
The taxonomy was used to classify my data, and the results are presented and
discussed. In chapter 7 I introduce and illustrate the notion of freshness, and
discuss relevant gender di erences.
I also constructed a taxonomy of humour functions and strategies. This was
applied to the data. Chapter 8 describes and exempli
es the taxonomy and
outlines and discusses the results.
In chapter 9 I discuss the focus of the humour, and indicate the range and
distribution of topics. Chapter 10 outlines the proportion of humour contributed
by men and women in the recorded conversations.
Chapter 11 is a discussion of personal identity, and illustrates how humour, like
all interaction, provides an opportunity for the speaker to display certain characteristics. An analysis of how speakers portray themselves reveals information
about value systems. Signi
cant gender di erences in the representation of certain traits indicate that individuals are identifying as male or female, or more
precisely, what it means to identify as male or female.
Chapter 12 is a discussion of humour support strategies. I show that the issue of
humour support is much more complex than is generally assumed, and describe
a range of humour support strategies, and possible reasons why humour may go
unsupported.
I conclude in chapter 13 with a summary of the main
ndings and suggestions
for further research.

Chapter 2

Review of humour research

2.1 Introduction
The literature on humour is vast, and draws on a number of disciplines. Scholars
in psychology, sociology, anthropology, women's studies, communications and
management have investigated aspects of humour and joking, and linguists have
approached humour from several diverse angles. This discussion is, by necessity,
an extremely limited review of the huge amount of research available. It touches
briey on aspects of humour research which provide background information
relevant to this thesis. I will not touch on aspects of humour development,
the use of humour for therapy, cross-cultural issues, or the semantics of why a
joke is funny. I also omit discussion of literature on humorous written pieces
and people's reactions to and appreciation of such pieces, and have no room to
discuss the vast literature on comedy and stand-up routines.
In this chapter I limit discussion to literature which is directly relevant for the
topic of my thesis. I discuss problems in the de
nition of humour and related
terminological issues, and outline research which deals with spontaneous spoken
humour and is relevant to the analysis of types and functions of humour. Chapter
three outlines the current state of research and thinking in the area of language
and gender, discussing its potential impact on humour and gender research, and
concludes with a review of research which has combined the study of gender and
humour. This literature review is intended as an overview only. In the following
3

CHAPTER 2. REVIEW OF HUMOUR RESEARCH

chapters I introduce literature in more detail where appropriate. Reviews of the

literature on gender di erences in amount of speech, topic, focus and other areas
discussed in this thesis are provided in the relevant chapters.

2.2 Dening Humour

Very few researchers take care to de
ne what they mean by humour, or specify
the basis on which they selected their examples. Aspects and repercussions of
humour and the functions it can play are readily ascribed and discussed, and
entire theories are sometimes constructed, without a de
nition in sight.
Those researchers that do de
ne their terms or make explicit the criteria by which
they select examples adopt varying approaches and emphases. Comparison of
de
nitions highlights contradictions and inconsistencies in this area. The issue
is further complicated by terminological overlap and confusion.
De
nitions tend to focus on either speaker intention or audience interpretation.
Berger (1976) de
nes humour as \a speci
c type of communication that establishes an incongruent relationship or meaning and is presented in a way that
causes laughter." The fact that Berger includes laughter as part of his de
nition
reects the fact that he holds the audience's interpretation to be important in
the de
nition of an event.
Winick (1976) concentrates on speaker intention. His de
nition of a joke is \any
type of communication that has a witty or funny intent that is known in advance
by the teller."
Unsuccessful attempts at humour would be included in Winick's de
nition, but
not in Berger's. De
nitions focussing on audience response are much easier to
apply than those which attempt to establish speaker intention. A researcher can
never tell from recorded data alone the exact nature of a speaker's intention. It is
easier to monitor the audience's reaction, and particularly easy if you maintain,
as does Berger, that the incident must be met with laughter. This is, of course,
complicated by the fact that laughter can indicate many things of which positive
response to humour is only one (see Poyatos 1993).
Pizzini (1991) uses the same criterion as Winick to select her examples:
4

2.2. DEFINING HUMOUR

I have analysed those interactions from which the actors intended to

elicit a laugh or a smile. Thus I have disregarded what the researchers
considered as humorous, but which may not have been intended as
such. This has cast light on the situated nature of humour, as we see
in the examples of humorous remarks: the jokes recorded here do not
make the reader laugh, not only because they are sometimes tasteless
and unamusing, but because the reader (and writer) is alien to the
particular social group and situation in which the joke was expressed.
(Pizzini 1991:479-480)
Pizzini states that she is alien to the social group involved, and so she can have no
intuitive sense of the sorts of things they
nd humorous. She disregards what the
researchers think is funny, but which might not have been intended as humorous.
There can be no sure way of establishing speaker intention, so on what basis is she
selecting her data? She is most likely analysing those interactions that do elicit a
laugh or smile, and is perhaps guided by context and clues in the speaker's voice
which indicate that they are not serious. When joking, a speaker often adopts
a laughing or smiling voice, and they may speak more rapidly than normal,
using exaggeration and a wider pitch range, or otherwise indicate through their
intonation that their utterance is intended as humorous.
Martineau (1972:114) incorporates both elements in his de
nition. He states
\humour is conceived generically to be any communicative instance which is
perceived as humorous by any of the interacting parties". This de
nition, while
more comprehensive than many others, seems rather circular.
It can be seen that, though the de
nition of humour is possible, its application
to the selection of examples from recorded data is far from problem free. There
is inevitably some subjectivity involved, and the researcher must place some sort
of interpretation on the data. Researchers will no doubt concentrate on one
aspect depending on what they are interested in. If a researcher is interested
in how much humour is supported, then speaker intention is clearly the main
criteria for selection. Other approaches may require knowledge of all events an
audience found funny. The important thing is to make explicit which criteria are
being used, so that the reader may be sure exactly what is meant when the word
\humour" is used in any given piece of research.
5

CHAPTER 2. REVIEW OF HUMOUR RESEARCH

It is important to be aware of overlap between the terms wit, humour and jokes.
Duncan (1984) chooses to use the terms joke and humour interchangeably. Long
and Graesser (1988) de
ne humour in the wider sense of Martineau (1972), incorporating both interpretation and intention. They de
ne jokes as things said
to deliberately provoke amusement.
Humor is anything done or said purposefully or inadvertently, that
is found to be comical or amusing. In contrast, jokes are de
ned as
anything done or said to deliberately provoke amusement. (Long and
Graesser 1988:37)
So it would seem that they use the term \joke" in a way similar to Winick (1976).
They then continue though,
Jokes are also context free and self contained in the sense that they
can be told in many conversational contexts. Wit will be de
ned as
anything deliberately said that provokes amusement in a speci
c conversational context (ie context bound). (Long and Graesser 1988:37)
So their de
nition of joke is restricted to anything said to deliberately provoke
amusement, and which is context free. This would apply to canned jokes { the
term given to a learned joke such as a narrative joke or a riddle. Some other
types of humour would also
t into this de
nition though most other types of
humour are context bound. Context bound intentional humour is labelled wit by
Long and Graesser. So they are taking an umbrella de
nition of humour { their
de
nition is the same as Martineau (1972). Together Long and Graesser's \joke"
and \wit" come together under a category labelled \joke" by Winick (1976)
and \humour" by Pizzini (1992). \Humour", by Berger's (1976) de
nition, will
include some of Long and Graesser's \jokes" and \wit", and some of what would
just be in their general humour category, and exclude some examples from all
of these categories. Research which de
nes humour in the same way as Pizzini
(1992), claiming the speaker's intention is paramount, often taxonomizes the
examples into various subsets of types of humour. Two of these categories could
well be jokes and wit. In such a context the labels have de
nitions which are
6

2.3. THEORIES OF HUMOUR

more restrictive still. Sherzer (1985) for example, de
nes jokes as discourse units
consisting of two parts, the set up and the punchline.
It is obvious that there is some confusion and disagreement as to what the various
terms should refer to. This makes it all the more important for researchers to
de
ne their terms carefully to avoid misunderstanding. The terminology is a
minor problem compared to others facing us in humour research and all that is
really needed is some agreement. Until such time, however, it is important that
terms and criteria for selection of examples are very clearly documented for each
study.

2.3 Theories of humour

There are many diverse theories of humour. Most of these can be loosely categorised into one of three categories: superiority theories, incongruity theories
and relief theories.
Superiority theories have as their basic premise the assumption that laughter is
the triumph of one person over other people (Keith-Spiegel 1972).
Incongruity theories emphasise the cognitive processes involved in perceiving humour. Suls (1972) proposed a model for the appreciation of jokes and cartoons.
This involves a two stage process which relies on the generation and discon
rmation of a listener's expectations. Raskin (1985) proposes a similar theory of
humour, claiming that jokes are compatible with two di erent scripts that are in
some way opposite. A joke begins by being compatible with one script, and then
a script-switch trigger occurs which is inconsistent with the original script. The
trigger is usually the punch line. The listener then searches for an alternative,
more compatible script. The humour lies in the overlap between the two scripts.
Relief or arousal theories incorporate the belief that laughter is the release of
repressed energy (Freud 1905).
How such theories of humour interrelate and function is a complex question. I
will not be touching on the issue of why something is considered funny in this
thesis. Instead I will be concentrating on actual spoken humour, and how it
occurs and functions in every day conversation.
7

CHAPTER 2. REVIEW OF HUMOUR RESEARCH

2.4 Humour research using elicitation or surveys

Only recently have researchers begun to study spontaneous spoken humour. Past
studies have been based on questionnaires or surveys (Crawford and Gressley
1991, Hampes 1992, Vitulli and Barbin 1991, Duncan 1984, Neuliep 1991 and
others) and elicitation (Fink and Walker 1977).
The methodology of some early studies is highly questionable. Fink and Walker
(1977) looked at humorous responses to embarrassment. The methodology employed in this study is somewhat dubious for both linguistic and ethical reasons. Sixty male subjects were put through an embarrassing interaction with
an experimenter by phone. They were asked to talk about issues such as sexual
attractiveness and appearance, and to describe the most embarrassing situation
they had experienced. The same interviewer was used throughout, but how she
introduced herself was varied. She introduced herself either as a professor, an
undergraduate or a high school student. This was so the variable of status could
be examined. In fact there were no signi
cant di erences in humour used with
regard to status, although signi
cantly more laughter occurred between people of
relatively equal status than between people with unequal status. Of course, these
are perceived status di erences rather than actual ones. The status is determined
by the way in which the interviewer introduced herself, but her conversational
style remains the same. This cannot be an exact simulation of conversation
between equals/unequals, although the authors seem to treat it as such.
The many questionnaire studies vary in their approach. Some ask for reactions
to certain instances of humour (Cox, Read and van Auken 1990), or ask for
descriptions of people the respondents perceive as particularly funny and about
their own use of humour (Crawford and Gressley 1991).
The questionnaire based studies of interest to this project are those which relate
to humour and gender. These are discussed in chapter 3.

2.5 Research on conversational joking

The most interesting and relevant humour research is that which is relatively
recent, and deals with recorded natural conversation.
8

2.5. RESEARCH ON CONVERSATIONAL JOKING

Davies (1984) investigated natural conversation, scripts, talk shows, interviews

and several other sources. She describes three joking styles used by speakers
of American English. The
rst consists mainly of insults, but only recognised
strengths must be attacked. This requires careful monitoring of reactions. In the
second style, speakers build on what others have done. Speakers demonstrate an
ability and willingness to participate in the joint e ort. It must be kept impersonal, and laughing and latching are common. Characteristic of this style are
phonetic play, lexical repetition and semantic linking between utterances. The
third style is an intimate positive politeness style in which the speaker displays
an understanding of what the other person is feeling/thinking. It uses allusion to
shared symbols and incongruity, gives the person a chance to express themselves
further and is usually accompanied by small smiles and laugh particles.
Norrick (1993:2) claims that in order to fully understand how joking can simultaneously express aggression and build rapport, we need to view joke-telling,
punning and teasing in relation to power, solidarity and distance and in light of
the principles of politeness and cooperation. He points out that the majority of
conversational joking grows from preceding talk, with much playing on it directly.
Norrick's book discusses di erent forms of humour, and gives some general examples of various joking styles. It is useful both as a signi
cant contribution to
research on situational joking, and an indication of how much we still have to
do!
Chiaro (1992) discusses jokes in depth. Her analysis is primarily based on jokes
and incidents drawn from personal memory, although some is taken from joke
collections, and the most interesting from recordings of dinner parties. She emphasises that for an audience to \get" a joke, there needs to be shared knowledge.
If a joke is too culturally oriented, then it does not travel successfully. A hearer
must have linguistic, sociocultural and poetic competence to understand a joke.
Chapter
ve of Chiaro's book is the most useful chapter. Here she looks at wordplay in action, claiming that if someone decides to be witty, something in the
context in which the conversation is taking place must have triggered this. This
prompt is most likely an element of the language or social context around the
conversation. We can joke \at any given moment" with people with whom we
are on friendly terms, but for those with whom we are less acquainted, play is
reserved for conventionally relaxed situations such as informal lunches or parties.
9

10

CHAPTER 2. REVIEW OF HUMOUR RESEARCH

Chiaro discusses various joking styles and the implications of these. She also looks
at the features that distinguish serious and humorous discourse. This leads her to
the conclusion that the two labels are opposite ends of a continuum, and the area
in the middle could be de
ned either by neither or by both labels. The distinction
is not a clear-cut one. Some features which do tend to distinguish humorous from
serious discourse are choice of words, use of euphemisms, intonation and informal
discourse markers, gesture and comic expressions, exaggeration and interaction
with the audience, and a casual and chatty style. When analysing my data I
used many of these features to help identify and interpret instances of humour
(see chapter 4).

2.6 Literature on types of humour

In this section I will briey review some of the taxonomies for classifying types of
humour proposed by humour researchers. The crucial criterion here is the form
the humour takes. There are many such taxonomies, and listing them all would
not be particularly useful. Instead I pick out a few and through these identify
some key categories. Many taxonomies are intended only for categorising canned
or formulaic jokes.
Monro (1953) provides an outline of what he regards to be the traditional classes
of humour: a) any breach of the usual order of events, b) any forbidden breach
of the usual order of events, c) indecency, d) importing into one situation what
belongs in another, e) anything masquerading as something it's not, f) wordplay,
g) nonsense, h) small misfortunes, i) want of knowledge or skill, j) veiled insults.
This seems to be more or less a list of things that people
nd funny. These are
mostly the topics of the humour, rather than its actual form. Categories like
\wordplay", and \veiled insults" though, do describe the form of the humour.
Zijderveld (1983) describes humour as the exploitation of institutionalised meanings, and breaks down types of humour into exploitation of either language (e.g.
puns, spoonerisms), logic (wit, elephant jokes), emotions (black humour) or the
activities of everyday life (parody, understatement).
Feigelson's (1989) taxonomy of humour among employees in a factory is as follows:
10

2.6. LITERATURE ON TYPES OF HUMOUR

11

1. Puns
2. Goo
ng o (slapstick)
3. Jokes/anecdotes

Humorous self-ridicule
Bawdy jokes (sexual or racial basis)
Industry jokes

4. Teasing

Teasing to get things done

Bantering { the great leveller

Categories such as industry jokes, are context speci
c. Puns are a relevant
category. These would fall into a broader category of wordplay, as described in
Norrick (1993: chap.3). Wordplay can include not only puns, but also types of
humour such as spoonerisms, allusion, hyperbole and metaphor.
Feigelson has put jokes and anecdotes together as one category, although many
researchers choose to keep these distinct. Telling a joke is di erent from telling
an anecdote. An anecdote will be more personal, whereas telling a joke is more
of a performance.
Mitchell (1985) makes a further distinction between narrative jokes and question
and answer jokes.
Teasing is an important category. There is some confusion, though, as to whether
it is a type of humour, or a strategy, or function. It is rather di erent than other
\types" identi
ed in that a tease can not be formally identi
ed by any criterion.
In my data, humour which served to tease someone present appeared in the form
of fantasy humour, insults, wordplay, anecdotes and roleplay, and so is clearly
on a level di erent from a purely formal taxonomy. I regard tease as a strategy
which can take any number of forms, and which can be used to create either
solidarity or power. This category is further discussed in chapter 8.
There are two further categories identi
ed in the literature which do not seem to

t into Feigelson's framework. One is sarcasm as discussed by Norrick (1993:73).
11

12

CHAPTER 2. REVIEW OF HUMOUR RESEARCH

And
nally, Morreall (1983) provides a comprehensive taxonomy of humour,

based on the concept of incongruity. Most of his categories could be slotted into
one of those discussed above, with the exception of mimicry.
In chapter 6 I outline a taxonomy developed for this thesis, which draws on the
literature discussed above and is further modi
ed to suit the data being analysed.

2.7 The functions of humour

In this section I briey outline the literature dealing with functions of humour
and identify some recurring themes.
Martineau (1972) discusses three functions of humour: consensus, conict and
control. The term consensus refers to the reduction of social distance. The
function of such humour is to initiate and solidify the development of social
relationships. It therefore encompasses functions such as integration into a group,
and creating solidarity. Conict humour introduces or fosters conict in a group.
Ridicule is a form of humour that can e ectively introduce conict. The term
control refers to the control of others. Humour is used to express grievances and
to draw people's attention to their mistakes.
The control function is also discussed by Collinson (1988). He discusses humour
on the shop oor, identifying three main functions { to resist boredom, to conform
and to control others. Collinson's conform reects the consensus category created
by Martineau. To resist boredom is a function not mentioned by Martineau. This
will be the function of a large number of instances, although probably not often
the sole reason. A broader term is to amuse or to entertain.
Pogrebin and Poole (1988) also identify three functions of humour. The
rst
is exploration, or probing. Humour allows us to test the attitudes and beliefs
of others in a non-threatening, o -record manner. In this way we can clarify
boundaries and standards. Humour is also used as a coping strategy. It helps
to normalise crises and deal with circumstances beyond our control. The third
function Pogrebin and Poole comment on is that of solidarity.
An aspect of solidarity is also investigated by Hampes (1992), who uses questionnaires to establish a correlation between humour and intimacy, concluding
12

2.7. THE FUNCTIONS OF HUMOUR

13

\humor apparently helps people succeed in intimate relationships because it allows them to handle the stress within those relationships" (Hampes 1992:127).
Fink and Walker (1977) identify humour as a face-saving device in an embarrassing situation. This is another example of the coping function of humour.
The exploration and coping functions of humour are also discussed by Linstead
(1985), who points out that humour often performs a boundary function. He
refers to Davies (1982) work on ethnic jokes. By making fun of peripheral groups
we clarify boundaries. These boundaries can be social, geographic or moral. We
can establish acceptable standards, and ensure that there is consensus amongst
the group as to what these standards are.
In discussing self-directed humour, Ziv (1984) identi
es the following four functions:

Rede
ning the social hierarchy by higher status individuals in order to

create solidarity among group members of di ering social status

Protecting the self by identifying a weakness before anyone else does

Sharing similarities between self and others

Coping with weaknesses by making light of them

Ervin-Tripp and Lampert (1992) refer to these functions as equalizing, defending,

sharing and coping.
The categories I have mentioned are those identi
ed in a small sample of literature. The most comprehensive review available is Graham, Papa and Brooks
(1992), who derive 24 functions of humour from the literature. These are listed,
along with authors who have discussed these functions (pg 167-168). The functions are drawn from 31 di erent works. This review was particularly helpful
in the construction of my working taxonomy of functions of humour, which is
outlined in chapter 8.

13

CHAPTER 2. REVIEW OF HUMOUR RESEARCH

14

2.8 Conclusion
In this chapter I have briey discussed some relevant literature on humour and
joking. It can be seen that the terminology in this area is still somewhat confused,
and that there is a need for much more research. As Graham, Papa and Brookes
(1992:177) point out:
We are lacking a substantial body of research that focuses on the use
of humour in conversational settings. Such research is necessary for
the development of a single, uni
ed functional model of humour.
Research has become increasingly advanced in many areas of humour research,
but research on spontaneous spoken humour is still a rarity. This is no doubt
partly due to the diculties involved in collecting appropriate data and identifying and analysing spoken humour.
In the next chapter I review relevant work in the area of language and gender,
and discuss developments in humour and gender research.

14

Chapter 3

Gender, Language and

Humour
Nothing spoils a romance so much as sense of humour in the woman,
or the want of it in the man. (Oscar Wilde)

3.1 Introduction
This chapter outlines developments in research on gender and its relation to
language and humour. Section 3.2 describes some di erences that have been
claimed to exist between the speech of men and women. In section 3.3 I discuss
some of the explanations proposed for these di erences, and in section 3.4 I
address in more detail recent thinking on the notion of gender. Section 3.5
describes research which has dealt speci
cally with humour and gender.

3.2 Men and Women in Conversation

In 1975 Robin Lako hypothesised some di erences in the speech of men and
women. She described her impressions of the features of \women's language".
These features involved expression of uncertainty, avoidance of strong expressions
of feeling, and subject matter deemed \trivial to the real world". This sparked
15

16

CHAPTER 3. GENDER, LANGUAGE AND HUMOUR

a urry of action in language and gender research and so, while much disputed,
provided a valuable impetus and focus for further research.
The hypotheses have been tested in a number of ways. Many studies concentrate on more formal speech or speech collected in arti
cial environments (O'Barr
and Atkins 1980, Preisler 1986 among others). This is problematic because
Lako 's hypotheses related explicitly to speech in informal situations. The various methodologies adopted by researchers have also meant that much research
is non-comparable and results are sometimes conicting.
Many researchers tested the hypotheses with counts of the forms Lako identi
ed as being typical of women's language (see Philips 1980 for review). Rather
than being particularly useful such research may actually conceal facts. Hedges,
which Lako claimed reect uncertainty, can also be used for politeness purposes
(Holmes 1984b, 1986). Pure hedge counts are therefore relatively uninformative.
Researchers have now started to look at the functions that such forms play within
the context of the conversation. Many have found that women use forms for facilitative, expressive or a ective functions more often than men (e.g. Holmes
1984b Cameron et al 1988).
Spender (1980) noted that people have inaccurate perceptions of how much men
and women contribute, particularly in formal settings. People tend to have the
perception that women have contributed and talked much more than is actually
the case. Studies of contributions in formal contexts have shown that men contribute more than women do (Swacker 1979, Edelsky 1981, Holmes 1988, Swann
1988, Bashiruddin, Edge and Hughes-Pelegrin 1990).
It has been claimed that women use minimal responses more than men (Maltz
and Borker 1983), are more likely to acknowledge what the other speaker has
said (Fishman 1977), and ask more questions (Maltz and Borker 1983). Men are
more likely to use disruptive interruptions (Zimmerman and West 1975, Stubbe
1991).
There is also an indication that men and women tend to talk about di erent
topics. Kiper (1987) found that women teachers talked about home and family
and male teachers discussed work and recreation. Studies relying on self-report
data revealed that women discuss their feelings and relationships whereas men
compared knowledge and experience and described their competitive exploits
16

3.2. MEN AND WOMEN IN CONVERSATION

17

(Haas and Sherman 1982, Aries and Johnson 1983, Aries 1976).
Lako (1975) was concerned that women do not seem to bond as much as men:
We would like, that is, not only for men to accept women as integral
parts of their groups, but for women to be able to group with other
women as men do with men...I think that a start is being made, in
women's groups, to overcome this tendency of women not to bond.
(Lako 1975:78)
Despite Lako 's concerns, research has indicated that women seem to value
community and co-operate in communication more than men. Makri-Tsilipakou
points out:
...men can be seen as mostly opting for privacy, separateness, independence and non-involvement. In contrast, women (...) seem to be
essentially opting for community, rapport and involvement. (MakriTsilipakou 1991:84)
There is still a relative paucity of research in informal settings, and men's talk in
informal contexts has been less studied than women's. Studies of women's talk
have generally found the subjects to be collaborative and supportive. Tannen
(1990a) claims that where women seek agreement, men tend to challenge each
other.
The women in Johnstone's (1993) study told stories about communities, whereas
the stories told by men were of acting alone, or of competition. The women emphasised the power of interdependence and community, whereas men emphasised
the power of the individual in opposition to others.
Researchers investigating accommodation have found that females tend to accommodate to the interlocutor more than males (Mulac et al 1988, Bilous and
Kraus 1988).
The literature on gender di erences in language is huge, and covers many di erent
aspects of speech. Many researchers conclude that there are signi
cant gender
di erences in conversational style. Probably the most consistent
nding is that
women tend to be supportive in their conversational style, and men competitive
17

18

CHAPTER 3. GENDER, LANGUAGE AND HUMOUR

(see Aries 1976, Edelsky 1981, Fishman 1983, Maltz and Borker 1983, Coates
1986, Preisler 1986, among others).

3.3 Explanations of dierence

In the previous section I outlined a small number of the many di erences that
have been claimed to exist in the speech of men and women. Several theories
have arisen in an attempt to explain these di erences.

3.3.1 Female Decit

One of the earliest explanations for apparent gender di erences in language usage
was that of female de
cit. A much cited example is Jespersen (1922), who, in
a book about language, included a chapter entitled \The Woman". Women's
speech was seen as inferior and di erent to men's, and so a deviation from the
norm. One can infer from Jesperson's writings that he considers this di erence to
be biological. Lako (1975) in her much criticised but very inuential paper, also
took the view that women's language is inferior. She regards socialisation as the
main explanation for this, and describes women's speech as reecting insecurity
and hesitance. She also regards male as norm, and notes that \when we leave the
lexicon and venture into syntax, we
nd that syntactically too, women's speech
is peculiar".
There is no evidence for the theory of female de
cit. On the contrary, there is
some evidence that females possess superior linguistic skills (see Chambers 1992
for a summary of research). Evidence presented by Lako is anecdotal, and could
be interpreted in a number of ways. Pure superiority judgements are necessarily
extremely subjective, and o er no explanation for the observed di erences.

3.3.2 Dierence or Dominance? An ongoing debate

For a long time the two main interpretations were that of culture and power.
Those taking the culture line (Maltz and Borker 1983, Tannen 1990a), claim
that men and women use language di erently because they are socialised into
18

3.3. EXPLANATIONS OF DIFFERENCE

19

di erent cultures. From childhood, boys and girls play in di erent groups and
in di erent ways. Girls' play emphasises solidarity, and teaches them to use
language to create and maintain friendships, to criticise others in acceptable
ways, and to interpret accurately the speech of other girls. The groups boys
play in, however, are hierarchical and competitive. Through this play, boys learn
to assert a position of dominance, attract and maintain an audience and assert
oneself when other speakers have the oor. It is clear that the culture into which
women are socialised is very much solidarity based, whereas there is an emphasis
in men's culture on things which reinforce power and status. Maltz and Borker
and Tannen, do not seem to regard this as signi
cant, although Maltz and Borker
do admit that \power di erentials may make some contribution". The culture
theory is
ne as far as it goes. But it really seems to miss the issue at hand.
Why is it that men and women are socialised di erently, and { more importantly
{ why the speci
c di erences that are observed? Eckert and McConnell-Ginet
are critical of this approach, claiming it:
...seems to suppose that people ignore all but the interactional possibilities predominant in their own gender-speci
c subcultures and
make no real interactional choices, simply acting as passive sponges
who soak up gendered identities. (Eckert and McConnell-Ginet 1992:466)
Some theorists prefer a power-based explanation and are highly critical of the different but equal cultures approach (Fishman 1983, Henley and Kramarae 1991).
A purely power based approach claims men's conversational dominance, and differences reected in speech functions can be explained by men's superior social
power.
Cameron (1992)
nds dominance and di erence both unsatisfactory explanations, and points out that what one believes is not so much a function of linguistic theory, but a reection of a political position. She also observes that our
theoretical position will predispose us to
nd di erent things in our data, and
that theorists from the two camps focus on di erent data. Researchers subscribing to the dominance theory most often concentrate on mixed sex interaction,
whereas for those following the two cultures account, single sex conversation is
likely to hold the most interest.
19

20

CHAPTER 3. GENDER, LANGUAGE AND HUMOUR

Both are extreme views, and almost any piece of data or source of miscommunication could be analysed as solely power or solely cultural di erences. Both
explanations have an element of truth, but, individually, they are simplistic. It
must be true that men and women are socialised into di erent styles and rules
of conversation. Parents interpret behaviour of male and female children differently (Condry and Condry 1976) and also interact with them very di erently
(Lewis 1972, Thoman, Leiderman and Olson 1972, Lewis and Freedle 1973). The
question that must be asked is why these particular patterns? The answer must
somehow reect men's power. The di erent speech styles clearly lead to males
claiming a dominant role in conversation. The di erent socialisation patterns,
the di erent \cultures" are caused by, reect and reinforce men's dominant position in society. Uchida (1992) points out that there must be other intervening
and contributing factors. Race, age, occupation, education, context and numerous other factors will a ect the types and causes of miscommunication likely to
arise between individuals and groups. It is important not to condense or ignore
these factors in search of a tidy all-encompassing explanatory theory. Clearly
socialisation and male dominance play a primary role in cross-sex miscommunication, but they will often not be the only variables involved. And as Maltz and
Borker (1982) point out, most theories of miscommunication are based on the
behaviour of whites. We can not necessarily expect to
nd observed patterns nor
proposed explanations generalisable to other cultures.

3.4 Conceptualisations of Gender { Moving Forward

The discussion up until this point has presented the concept of gender as relatively straight-forward. Sex is a biological category, and gender is regarded as
a social construct. Almost all linguistic research, however, has regarded gender
as primarily a binary division. The term gender is used to acknowledge that the
di erences observed are not absolute and biological, but rather are socialised,
yet the categorisation of people's gender is based very rigidly on their biological
sex.
West and Zimmerman (1987) distinguish between sex, sex-category and gender.
The
rst two of these are rather easier to describe than the latter. Sex is a
biological category based on the genitalia one is born with. A sex-category is
20

3.4. CONCEPTUALISATIONS OF GENDER { MOVING FORWARD

21

where you are placed by society. At birth, sex-category will be based on one's sex,
but it is possible to claim membership of a sex-category without possessing the
expected genitals. The concept of gender is much more tricky. Most theorists
agree that gender is a social construct { something inuenced and shaped by
one's environment. There has been a theoretical shift away from gender as a
binary construct. Many see it more as a continuum, though I have not yet seen
this applied in any practical sense to sociolinguistic research.
Smith (1985) claims one's gender identity can be marked on a scale ranging between the stereotypically feminine and the stereotypically masculine. Among
the variables claimed to contribute to gender are sexual orientation, masculinity/femininity, stereotypes and archetypes, social roles (e.g. family sphere or
social sphere), and power/resource distribution. He describes two methods for
measuring masculinity and femininity. As Smith points out, this approach is
problematic as it assumes that stereotypes reect di erences.
Eckert (1990) regards gender as a set of traits, which she claims are the result of
men's and women's place in society at a given time. Chambers (1992) sees gender
as a social role, claiming gender-based variability in language is a reection of
men's and women's mobility in the society.
There is, however, a shift away from the perception of gender as a set of traits.
The question has become not whether these sets of traits can be sensibly divided
into two categories, or whether they belong on a continuum, but something much
more fundamental - what is gender.
West and Fenstermaker point out that treatment of gender as poles of masculinity
and femininity is problematic.
...the bifurcation of gender into femininity and masculinity e ectively
reduces gender to sex (cf Gerson 1985), while the treatment of gender
as a role obscures the work involved in producing gender in everyday
activities. (West and Fenstermaker 1993:151)
Theory, then, has shifted to a conception of gender as something mobile and
dynamic. West and Zimmerman (1987) claim that, in interaction, we \do gender". Gender is the activity of interacting within the framework of conceptions
of attitudes and activities appropriate to one's sex-category. They claim gender
21

22

CHAPTER 3. GENDER, LANGUAGE AND HUMOUR

activities emerge from, and reinforce claims to membership in a sex-category.

So there is not a male gender identity and a female gender identity, but rather
a myriad of possible ways of identifying as male or female, and we reect and
construct these identities in interaction constantly. Subtle shifts will occur in
our gender identity as we shift from interaction to interaction, from interlocutor
to interlocutor. At certain times our gender identity will be much more salient
than at others.
Penelope Eckert and Sally McConnell-Ginet point out that there is much scope
for gender variation within sex-categories. Di erences among men and among
women are also important aspects of gender. Tomboys, goody-goodies etc are
all categories of girls and women \whose mutual di erences are part of their
construction of themselves and each other as gendered beings." (Eckert and
McConnell-Ginet 1992:470)
West and Fenstermaker touch on this important point as well:
doing gender does not require heterosexual situations....Hence we may

nd some of the most extreme versions of essential womanly and
manly natures in those settings that are reserved for members of a
single sex category, for example, locker rooms or beauty salons. (West
and Fenstermaker 1993:158)
West and Fenstermaker (1993:156) note that gender is a situated accomplishment: \the local management of conduct in relation to normative conceptions of
appropriate attitudes and activities for particular sex categories."
Doing gender, then, is not necessarily striving to
t your identity to a particular
normative ideal, but it is holding behaviour accountable in relation to these ideals.
In short, persons engaged in virtually any activity can hold themselves accountable as women or as men. Their membership in one
or the other sex category can be invoked to legitimate or discredit
their performance... to reiterate, doing gender does not always mean
living up to normative conceptions of femininity or masculinity, what
it means is rendering action accountable in these terms." (West and
Fenstermaker 1993:157)
22

3.4. CONCEPTUALISATIONS OF GENDER { MOVING FORWARD

23

This conception of gender as a situated accomplishment, of an ongoing, dynamic

process, clearly has implications for approaches to the study of language and
gender. Cameron (1992) is strongly critical of the approach often used in sociolinguistic studies of male and female behaviour.
Merely to say that `men do x and women do y' is inherently problematic: it
ts comfortably into a tradition of prescriptive and sexist
comment about what \normal" masculinity and femininity entail.
(Cameron 1992:21)
Instead she suggests an alternative viewpoint, which reects more the concept
that interaction constantly shapes and reects our gender identity.
..we ought to say \that's how people are under pressure to behave
in order to realize certain ideas about what men and women should
be like { to construct themselves as REAL or NORMAL men and
women." NOT \how do men and women behave?" but \how do
certain modes of behaviour make people into men and women?"
(Cameron 1992:22)
The conceptualisation of gender as situated and dynamic is important for language and gender studies. It means a move away from large form-counting
exercises, and towards studies of individuals' behaviour in certain contexts. We
need to study how individuals construct an identity, and \do gender".
Penelope Eckert and Sally McConnell-Ginet adopt Lave and Wenger's (1991)
concept of Communities of Practice.
A community of practice is an aggregate of people who come together
around mutual engagement in a particular endeavour. Ways of doing things, ways of talking, beliefs, values, peer relations { in short,
practices { emerge in the course of this mutual endeavour. (Eckert
and McConnell-Ginet 1992:464)
The investigation of communities of practice will reveal much about methods of
identity formation, and help re
ne the conceptualisation of the notion of gender.
23

24

CHAPTER 3. GENDER, LANGUAGE AND HUMOUR

Eckert and McConnell-Ginet regard this as the direction in which we should

move. They note that correlations between linguistic forms and sex are useful,
but only in that they indicate areas where future investigation may potentially
reveal the practices entering into gender dynamics within a community.
In many areas this ground work has now been laid, and researchers are equipped
to move on to detailed investigation of various aspects of gender dynamics. Humour research is an area where there are few ground studies. The literature indicates some trends, but much of it relies on introspection or arti
cially elicited
data. In this study I investigate di erences between men and women statistically
with regard to a number of aspects of humour usage. I do not intend to produce
any startling or categorical conclusions about how men and women use humour,
but rather provide the ground work which has been lacking in this area up until
now. I hope to point to some interesting trends, and identify areas of humour
and gender research which could reward close examination. In chapter 11 I look
more closely at several aspects of one's personal identity, and more speci
cally,
one's gender identity, and how they can be shaped and reected in humour. In
chapter 12 I investigate more closely issues of humour support and reasons why
humour might fail. The next section discusses research which has already been
done in the area of humour and gender.

3.5 Gender and Humour

Women have traditionally been considered humourless beings. Much literature
portrays females as unable to produce or even appreciate humour.
In her inuential but much disputed paper Lako (1975) lists the forms comprising \women's language". One of her observations is \Women don't tell jokes"
(Lako 1975:56).
It is axiomatic in middle-class American society that,
rst, women
can't tell jokes - they are bound to ruin the punchline, they mix up
the order of things and so on. Moreover, they don't \get" jokes. In
short, women have no sense of humor." (Lako 1975:56)
Some researchers have attempted a justi
cation for the \fact" that women don't
24

3.5. GENDER AND HUMOUR

25

use humour. Freud (1905) claimed women do not need a sense of humour because
they have fewer strong feelings to repress. Grotjahn (1957) suggested that women
do not tell jokes because joke-telling is an aggressive act.
Lizbeth Goodman, in a paper primarily about stand-up comedy, points out:
There is a lingering perception that women are not best suited to
telling jokes but rather to being the punchlines. (Goodman 1992:286)
This perception has applied, not only to joke-telling, but to humour in general.
The attitude is slowly changing as researchers begin the process of collecting and
documenting humour used by both men and women in a variety of contexts.
Kramarae (1987) points out that men and women have di erent perceptions of
the world and so consequently probably have di erent joking interests. Society
is such that women have to work within the social symbols of the dominant
group, so it is more likely that women will recognise the joking interests of males
than vice-versa. Kramarae believes this is the basis of the common assertion
that women have no sense of humour. In short, women have to understand male
humour, men do not have to understand women's. This is reinforced by Jenkins
(1985) who notes this asymmetry:
I wondered why it was that when a man tells a joke and women don't
laugh, we are told we have no sense of humor, but when a woman tells
a joke and men don't laugh, we are told we are not funny. (Jenkins
1985:135)
Many researchers have pointed out that humour is an inherently powerful act.
In order to gain acceptance as a \true" woman, it is therefore unacceptable for
women to display humour in mixed company. Marlowe (1989) observes:
When women produce and present humor they reverse conventional
social situations by putting themselves in the foreground, threatening
the most basic social gender arrangements. (Marlowe 1989:150)
Women are said to have a sense of humour, not if they produce humour, but if
they respond to and appreciate it (Coser 1960, McGhee 1979b, Barreca 1991).
25

26

CHAPTER 3. GENDER, LANGUAGE AND HUMOUR

For these reasons, some researchers have found that women use humour when by
themselves, and tend to avoid it in mixed groups (Coser 1960, Goodman 1992).
Women just do not attempt to be humorous in a mixed group setting and the reason seems to be that women are neither expected,
nor trained, to joke in this culture....it seems reasonable to propose
that attempting a witty remark is often an intrusive, disturbing and
aggressive act, and within this culture, probably unacceptable for a
female. (Pollio and Edgerly 1976: 225)
Grotjahn (1957) issues a warning to female readers.
The woman of today....had better not show her wit too obviously if
she is young and intelligent, for she will scare the contemporary male,
who is easily frightened in his masculinity. (Grotjahn 1957:52)
Crawford (1989) points out that many studies of gender and humour have involved bias. Many concentrate on humour occurring in the public sphere. This
is clearly easier to collect than private, spontaneous joking, but, as the public
sphere is generally the domain of males, observed women's humour is unlikely
to be typical. There has also been a lot of research concentrating on responses
to set piece jokes. Canned jokes have been shown to be a more typically male
form of humour (Jenkins 1985, Goodwin 1982), and so predictably, many results
show men more appreciative of the jokes than the women. This often leads to
the conclusion that women have less of a sense of humour.
Unfortunately poor methodology has even recently lead some researchers to mistakenly conclude that impressions of humourless women are substantiated by
fact. Cox, Read and van Auken (1990) looked at gender di erences in communicating job-related humour. The study is triggered by the concern that:
Abundant literature indicates the importance of humour in the workplace. Yet it is also proposed by some authors that women lack (or
do not make use of) humour when communicating in professional
activities. (Cox, Read and van Auken 1990:287)
26

3.5. GENDER AND HUMOUR

27

This is an interesting area, and one which clearly warrants some empirical investigation. However the work documented in this paper has some serious aws. The
data is collected by administering a questionnaire to 156 male and 106 female
undergraduate business students. Most students probably have no experience
in the work place, so using them to establish gender di erences in job-related
humour is highly dubious.
The questionnaire contained 15 hypothetical situations in which the respondent's
colleague is placed in a potentially embarrassing situation in a job-related circumstance. Each situation was followed by three reactions, one related to ignoral,
one to humour and one to helpfulness. For each one, the informant had to rate
how likely it was that they would react in that way. The one example given was
a colleague who accidently dropped a lot of papers, and the humour response
was: \I would tease him about being a master paper shuer". Students were
asked to rate such statements from 1 (strongly agree) to 7 (strongly disagree).
All this actually tests is whether respondents have the same sense of humour
as whoever wrote the questions. If a respondent rated the humour question
low, then this was taken to mean they would not use humour in that particular
situation. All it really means is that they would not use the statement provided.
If it is the case that men and women have di erent joking styles, then this
methodology will be heavily biased towards the gender of the author of the
questions.
The results showed that men used the humour response signi
cantly more, and
women were signi
cantly more likely to react helpfully. Cox, Read and van
Auken therefore conclude:
This study seems to verify what most of the non-empirical literature
has hypothesized about women's use of humor, namely, that humor
is less a part of the female's communications pattern. (Cox, Read
and van Auken 1990:293)
Even if the methodology did not lead us to question the validity of the results,
this study could in no way lead us to such a dramatic conclusion. Firstly, this
questionnaire is concerned with one type of humour { humour at the expense of
others. This is a small sub-class of the many types of humour, and so we cannot
conclude on the basis of this study that \humor is less a part of the female's
27

28

CHAPTER 3. GENDER, LANGUAGE AND HUMOUR

communication pattern." Most aspects of humour have not been discussed let
alone examined in this paper. Secondly, the study set out to look at job-related
humour, but this conclusion seems to be making claims about communication in
general.
Crawford and Gressley (1991) also used a questionnaire to elicit their data. Their

ndings are both less spectacular and more credible than those of Cox, Read and
van Auken. They administered a two-part questionnaire to 72 males and 131 females (aged 16-84). The
rst section of the questionnaire asked participants to
describe someone they knew who had an outstanding sense of humour, and the
second part asked them to rate how much they themselves participate in various
humour-related activities (e.g. do you enjoy ethnic humour). The descriptions
were analysed to establish which of the following dimensions were mentioned:
hostility (cutting, sarcastic, humour at others' expense), jokes, creativity (spontaneous, can think on feet), real life (stories and anecdotes), caring (to alleviate
anger or tension) or none of the above. In describing someone they knew, men
used the creativity dimension signi
cantly more than women. Sixty-two of 92
females wrote about males, and 41 of 49 males wrote about males. When writing about males, respondents were signi
cantly more likely to use the creativity
dimension. Males reported they enjoyed both slapstick and hostile humour (e.g.
racist or sexist humour) more often than females, and they reported more use of
formulaic joking, whereas females reported more anecdotal humour.
Vitulli and Barbin (1991) administered questionnaires to establish whether subjects thought humour was a male oriented activity or a female oriented activity,
and to see if there was a perceived di erence in reaction to humour depending on
whether the initiator was male or female. They also investigated the possibility
of age di erences in such perceptions.
They refer to McGhee's (1979a) \egalitarian hypothesis" which predicts that
the sex-role convergence in recent decades will lead to more equal \sharing" of
humour appreciation by the sexes. Vitulli and Tyler (1988) claimed that sex
di erences in humour appreciation converge between high school and college,
with high school level males still
nding aggressive humour funnier than females
do. Vitulli and Barbin found a number of signi
cant results. Men more than
women thought humour to be a male oriented behaviour, and 8th grade and
high school males more than college males thought humour was a male oriented
28

3.5. GENDER AND HUMOUR

29

behaviour. Females in 8th grade more than females in college thought that
humour was a female oriented behaviour. The
nal signi
cant result was that
females at college level thought that responses to humour di ered depending on
the sex of the initiator, more than 8th grade females did.
Kottho (1986) deals with spontaneous conversational joking, and also proposes
a gender di erence. She provides three very interesting hypotheses based on
transcripts taken from various discourse analysis literature. The hypotheses are:

Men more often than women joke at the cost of others.

Women joke about themselves and their experiences. For them joking is a
means of establishing common ground and intimacy.

Women actively encourage the success of the speaker by providing support

through laughter. Men do this less frequently, especially when the speaker
is a woman.

She states that for women and other oppressed groups, the safest joking method
is to make oneself the butt of the joke. This allows the audience amusement at
the speaker's expense and also, the laughter provides an acceptable outlet for
aggression. She stresses that women's ability to laugh at themselves should be
considered positive.
Kottho also refers to an unpublished masters thesis, Dreher (1982). Dreher
studied four conversations and found that women laughed more than men, and
that both men and women laughed more at men than at women.
Pizzini (1991) maintains that humorous remarks are one way in which patients
are interrupted and silenced. Investigating four men and four women gynaecologists with 40 patients, Pizzini found that there was a prevalence of female
gynaecologists using humour, but that there was a markedly high proportion of
humour by males used to stop patients from talking. Male gynaecologists tended
to use humour to maintain the asymmetry in possession of scienti
c knowledge.
Female gynaecologists used it to cut down discussions not considered relevant.
Ervin-Tripp and Lampert (1992) also investigated humour in naturally occurring situations and found men more likely to initiate a humorous key. Women
maintained a humorous key across participants so there was a larger amount
29

30

CHAPTER 3. GENDER, LANGUAGE AND HUMOUR

of humour elicitation by women. Women were more collaborative in their humour. They often used duets in wordplay, and their self directed humour in
single sex interaction was more likely than men's to be built on someone else's
remark. Ervin Tripp and Lampert call this stacked humour. When men used
self directed humour it was more novel and less collaborative. It also tended
to be more exaggerated or clearly false, giving a performance quality to men's
humour. Women volunteered real stories about themselves to resolve and heal
old embarrassments or to build togetherness by revealing shared experiences.
Men often used wisecracks about themselves, usually following something they
had done wrong, to defend themselves by making light of the situation. Women
used humour to share and cope, men used it to defend and equalise. In mixed
groups humour dealt with power tensions, built solidarity and displaced anger
to di erent persons.
Di erences have also been observed in men's and women's joke-telling habits.
Mitchell (1985) collected 1507 jokes, noting who told them, and in what context,
and categorised them according to form, character, setting and theme. This
resulted in a large taxonomy of characters and themes. The sex of the teller
and the composition of the audience had signi
cant inuence on the choice of
joke. Men told a higher percentage of openly aggressive jokes than women, and
seemed to enjoy competitive joke-telling sessions. Women rarely participated in
these sessions. Men were also more likely to use jokes to deride someone they
did not like, and to tell jokes they thought might be o ensive to some members
of their audience.
Women preferred to tell jokes in their homes to small groups of close friends,
whereas men enjoyed telling jokes to larger audiences in public places. Men
told a higher percentage of obscene, religious, ethnic, racial, death and drinking
jokes, whereas women told a higher percentage of absurd, morbid, Polack and
authority jokes, and used more word-play. Mitchell found that men told more
narrative jokes, and women more question and answer jokes. Joke-telling by men
was much more performance-based.
Jenkins (1985) also notes that male humour tends to be more performance-based
than women's humour. Jenkins observes that men's humour is characterised as
self-aggrandizing one-upmanship. They more often use formulaic jokes which are
markedly separate from the surrounding discourse and which involve a perfor30

3.5. GENDER AND HUMOUR

31

mance. This establishes them as credible performers and gives them an audience.
Women tend to rely on the context more in the creation of their humour, and
use it in a way that is supportive and healing.
Goodman (1992) points out that joke telling is less typical for female comedians
than males. Female comedians tend to prefer narrative comedy, in which humour
is di used over the situation rather than contained in a single phrase. She quotes
Helen Lederer describing gender di erences she perceives in the use of humour
by professional comedians.
When a man tells a formalized joke I tend to switch o because it's
quite authoritarian: you have to listen in order to get the payo , the
punchline, and then you have to laugh. It's quite strict and inexible.
It's far more interesting for me to ramble on, hopefully hitting the
right targets, certainly with a through-line, and certainly with an
end, but not in the same formalized way. I would rather just sit and
hope that it's funny. (Helen Lederer as quoted in Goodman 1992:295)
Kaufman (1991) attempts to characterise feminist humour. She also notes a
preference for spontaneity rather than for formulaic humour.
(Feminists') preferences are toward spontaneous wit, amusing real-life
anecdotes and other forms of humour that are participatory. Jokes
involve tellers and listeners, the teller is the active one at the centre of attention, and the listeners are relatively passive...spontaneous
human interaction is largely absent. (Kaufman 1991:248)
She points out that witty remarks contribute to the dialogue, whereas jokes tend
to disrupt and distract from it.
Crawford (1989) administered questionnaires designed to elicit the types of humour used by women, and their perceptions about gender di erences in the use
of humour, and the types of humour they valued in others. She found that the
types of humour the women di erentially attributed to themselves were the same
as the types they valued in others. Namely, anecdotes about one's own, and one's
friends' personal experiences. Crawford says of women's humour:
31

32

CHAPTER 3. GENDER, LANGUAGE AND HUMOUR

it involves not only creative spontaneity but connectedness and compassion it invites self-disclosure and reciprocal sharing of perspectives it is dependent on the immediate social context. (Crawford
1989:160)

3.6 Conclusion
In the last few decades we have witnessed an explosion of research in language and
gender. This work has become increasingly sophisticated and intricate. The generally consistent
nding is that women tend to be the more supportive speakers.
They emphasise interaction, solidarity and community, whereas men emphasise
status and competition. Jenkins (1985) characterises the general trend as follows:
Men in their groups seem to be saying \I'm great. I'm great too.
Gee we're a great bunch of guys." In contrast, women seem to be
saying \Did this ever happen to you? Yeah. Oh good, I'm not crazy"
(Jenkins 1985:137)
Recently theorists have become uncomfortable with such sweeping statements
and many have doubts of the wisdom of polarizing the genders so distinctly. We
should be careful to keep in mind that such claims are at best, large generalisations. They are useful primarily in that they provide pointers to interesting
phenomena and provide a base for further, qualitative research.
Given that there is a tendency for women to be more supportive in their interactions, we might expect this to have implications for women's use of humour.
Early research into humour and gender was unbalanced, or based primarily on
speculation, and claimed that women did not use humour at all, or that they
used humour much less than men. This claim is largely a result of the fact that
pressure is put on women not to joke in mixed sex interaction. Joking is regarded as a particularly masculine behaviour. This means that much of women's
humour occurs in single-sex groups.
Only recently has research targeted humour as it occurs in spontaneous conversation. Emerging results suggest that women's humour is, in fact, more supportive
than men's. Men seem to use more formulaic jokes and use performance-based
32

3.6. CONCLUSION

33

humour to claim an audience and gain status. Women's humour is more contextdependent, involves their audience, and is more supportive and personal.

33

Chapter 4

Methodology
\Supposing we hit him by mistake?" said Pooh anxiously. \Or supposing you missed him by mistake" said Eeyore. \Think of all the
possibilities, Piglet, before you settle down to enjoy yourselves"1

4.1 Introduction
This chapter discusses the methodology employed in this study. Section 4.2 outlines the variables I attempted to control during my taping, section 4.3 describes
my method of data collection, and 4.4 describes the criteria used in the selection
of examples and the resulting corpus. Finally, in section 4.5 I describe some outcomes of the taping which indicate that the methodology used may be slightly
biased.

4.2 Control of Variables

4.2.1 Number of Speakers
Beattie (1981) found that the number of speakers present can inuence behaviour. I decided to analyse conversations with four speakers. The dynamics of
1

Introductory quotes in this and subsequent chapters courtesy of A.A. Milne

35

CHAPTER 4. METHODOLOGY

36

dyadic interaction are in many ways quite di erent, and with regards to humour,
more limited. Larger groups allow for the possibility of victimisation or sidetaking within the group for humorous e ect. Also, as pointed out by Preisler
(1986), small groups can be more fruitful for analysis than dyads because members can more easily relax, and hide in the conversation if they do not feel they
want to contribute. The more relaxed the speakers are, the more natural the
data will be.

4.2.2 Sex
As gender is central to my analysis, sex clearly had to be carefully controlled.
I wanted to analyse both mixed sex and single sex interaction. I therefore decided my corpus should include 18 conversations { six conversations between
four females, six conversations between four males, and six conversations with
two female speakers and two male speakers.

4.2.3 Ethnicity
Davies (1990) carefully documents the use of ethnic jokes in a variety of cultures,
and Ziv (1988) describes some di ering national styles of humour. Chiaro (1992)
also points out that appreciation of jokes usually requires, among other things,
shared cultural backgrounds and assumptions. Although there has been no research done on di erences between Maori and Pakeha use of humour, it is highly
possible that such di erences do exist. This is clearly an interesting avenue for
future research. Because of possible cultural di erences I chose to restrict my
sample to Pakeha speakers. All of the speakers self-identi
ed as Pakeha except
one, who classed himself as Maori/European. This speaker identi
ed with both
ethnic groups and so was not excluded from the sample.

4.2.4 Age
Age is another variable which may cause di erences in the type of humour used.
Hill (1987) found that behaviour patterns of people in the Malinche community
changed as they got older. Women in particular tended much more towards
36

4.2. CONTROL OF VARIABLES

37

rowdy wit, outspokenness and blasphemous joking once they were in their forties
than they would have in younger days. In order to minimise the e ect of age, I
ensured all of the speakers on my tapes were aged between 18 and 35.

4.2.5 Education
I tried to ensure that all speakers had some form of tertiary education. In the case
of the corpus data, this was not always possible, as the background information
sheets do not explicitly ask for this information. In all of the conversations
though, at least three of the speakers have had some form of tertiary education,
and the other one probably has. The reason for analysing this particular group of
subjects is simply that this was the group most accessible to me. It is unfortunate
that I could not target non-students, but it would have been impractical. My
initial source of data was the Wellington Corpus of Spoken New Zealand English
which, not surprisingly, includes a lot of extracts with student speakers. For
the remainder of my data I called on my friends, who are mainly students or
ex-students.

4.2.6 Intimacy of Speakers

I took account of how well the speakers knew each other. Again, for the corpus
data this information was not always readily available. In the corpus tapes used,
however, the speakers seem to know each other reasonably well, and references
to past events establish that they had certainly met before the recording took
place. For each of the conversations collected, the speakers were all good friends,
and in most cases, had known the other speakers for some years. How well you
know the audience will clearly a ect interaction style, and so this is an important
variable (Wolfson 1988, Brown and Levinson 1987). In an ideal study, detailed
information would be collected from each of the speakers and a network score
devised. This would be a lengthy and delicate process. This study involves 18
groups and 72 speakers, and so such investigation would be too time consuming.
There will no doubt be variation in the intimacy of the speakers across my tapes,
but all of the conversations can be described as discourse between good friends.

37

CHAPTER 4. METHODOLOGY

38

4.2.7 Setting
I wanted to keep the setting as relaxed as possible. Most of the conversations
were recorded in a relaxed environment in someone's home. There were a couple
of conversations collected outside of the home, but these were in settings in which
the speakers were comfortable, and where they often met and talked.

4.2.8 Alcohol
Some of the speakers had one or two drinks during the course of the taping. It
seems very likely that alcohol will a ect the type of humour produced, and so
I did not want the level of consumption to be very high. I excluded one tape
because the speakers seemed to be a ected by their prior alcohol consumption.
On the other hand, if a group of friends usually meet to chat over a beer, then
this is the most natural situation in which to tape. As long as they did not drink
too much, in some cases I thought having a beer in their hand would signi
cantly
increase the naturalness of the conversation, and make the speaker feel as relaxed
as possible.

4.3 Data Collection

I was looking for conversations between four Pakeha speakers. The composition
had to be either single sex, or two male and two female speakers, and fall into the
categories outlined above. My plan was to draw my data from the Wellington
Corpus of Spoken New Zealand English, collecting data to
ll any remaining
gaps. I also had access to the Linguistics Library at Victoria University and a
set of tapes collected by Anita Easton for her paper \Talk and Laughter in New
Zealand Women's and Men's Speech" (Easton 1994). A search of the Corpus
and Linguistics Library provided the data shown in table 4.1.
There are clearly gaps here, and these were rather larger than I had expected.
I contacted members of my immediate circle of friends and asked them if they
could help me with the data collection for my thesis. I told them that recordings
of groups of four people chatting together were needed, described the necessary
constraints, and asked them if they could think of a group of people that would
38

4.3. DATA COLLECTION

39

FEMALE MALE MIXED

CORPUS
2
1
3
LIBRARY
1
EASTON
1
2
3
1
6
Table 4.1: Pre-collected data available
not mind helping me. This method has a number of advantages. Firstly, all of
the speakers either knew me, or were approached by someone who knew me, and
so were more comfortable with the situation than if they had been approached by
a total stranger. Secondly, knowing most of the speakers on the tapes made the
conversations easier to transcribe, and also enabled me to use my inside knowledge in the analysis. Two groups consisted of good friends who were invited to
my at for the taping. I left the room for an hour while the taping took place.
The rest of the gaps were
lled by friends who recorded tapes for me when appropriate opportunities arose. This technique allowed me to
ll the gaps identi
ed
above. The
nal corpus consisted of four recordings from the Wellington Corpus
of Spoken New Zealand English, one from the Victoria University of Wellington
Linguistics Library, three collected by Anita Easton and ten collected by me.

4.3.1 Speakers' knowledge of the project

The speakers on the Corpus, Library and Easton tapes were clearly not a ected
by knowledge of my project as they were recorded before it started. All of the
speakers on the tapes I collected were told that the data would be used for my
thesis. They were told I was investigating gender di erences in a speci
c variable.
After the taping I gave a more complete description of the project. All speakers
signed the Corpus Background Information Sheet. The signature follows the
statement:
I give permission for the recording of my voice to be included in
a corpus of New Zealand English to be used for linguistic research
purposes.
39

CHAPTER 4. METHODOLOGY

40

Drawing on my immediate circle of friends made it dicult to avoid including

people with whom I had already discussed my project. One tape in the all
male section includes a speaker who was aware of the fact that I was analysing
humour. This does not seem to have a ected his behaviour in any signi
cant
way. There are two tapes in the single sex female section which have one speaker
in common. I had told her about my project, and thought she was aware of the
topic during the recording, but a comment she made later revealed that she had
become confused and thought I was looking only at jocular abuse, a topic studied
in my honours year. Apart from these two speakers, none of the speakers knew
I was studying humour when the taping took place.

4.4 Selection of examples

From each tape I took a twenty minute excerpt. The longer the speakers have
been recording, the more likely they are to be relaxed and natural. I therefore
took the last twenty minutes of recording unless there was reason not to. In
one tape, for example, the speakers are drinking and so are sober at the beginning of the taping but not at the end. There is also a tape which has frequent
interruptions towards the end, and so an earlier 20 minutes was selected.
In de
ning criteria for selection of examples and in the analysis, I was interested
in the intention of the speaker. Because one of the things I was investigating
was whether the humour of males and females was equally supported, this automatically precluded a de
nition based on audience response. Tannen (1993:166)
points out that the true intention of any utterance can not be established from the
examination of linguistic form alone. I acknowledge this entirely, and I therefore
analyse each example carefully and independently, taking into account a number of factors. In identifying and coding the examples I used numerous clues
to determine the speaker's intention. I drew on my knowledge of the speakers
and the groups as a whole, and also on knowledge of the groups gleaned from
the remaining non-analysed tape. I relied heavily on context to determine the
speaker's intention, and also took into account the audience's response. The
audience formed part of the group as a whole and so probably shared a similar
sense of what is funny with the speaker, so if something appeared to be meant
humorously, then an amused audience would provide evidence in support of this.
40

4.5. METHODOLOGICAL PROBLEMS

41

The speaker's tone of voice was also important. Sudden changes in pace or pitch,
a laughing or smiling voice and other verbal clues were taken into consideration
(see Crystal 1969).
I transcribed all instances of humour, and sucient preceding dialogue to provide
adequate context. In most cases this meant transcribing the entire 20 minutes,
with very few breaks. This was a time-consuming process, and one I had not
counted on, having planned to take most of my data from the Wellington Corpus of Spoken New Zealand English. The transcription conventions used are
described in Appendix A. This resulted in a corpus of 815 examples. Of these,
333 examples were from single sex male groups, 216 from single sex female groups,
163 from males in mixed groups, and 103 from females in mixed groups.

4.5 Methodological problems

When collecting my data, some groups were much easier to arrange than others.
Mixed groups were relatively easy to
nd. There were a number of these in the
corpus, and I collected a few myself, so that there were ten tapes from which to
choose the best six. In general, social groups made up of two women and two
men seem to be quite common.
There were three all female tapes available from the corpus, and I supplemented
these with three more. This was not as easy as
nding mixed groups who were
prepared to talk for me, but nonetheless, proved manageable.
There was only one conversation between four males in the corpus, leaving
ve
to collect. This proved a dicult task. One male I approached commented that
he could not remember the last time he sat around with three other men and
chatted for half an hour. \Men tend to come in twos and threes" he said. If
this is true, then it indicates I have chosen a grouping more \natural" for female
and mixed groups than for males, which could potentially distort the results.
Also, I could have gathered this data much more quickly if I were able to tape
in the pub. Sitting at home with three friends may not be a natural, or even
comfortable situation for some males. Nevertheless, I did manage to produce a
collection of six tapes of groups of four males who were natural social groups.
All groups seem to be relaxed with the situation, and chat normally. However
41

CHAPTER 4. METHODOLOGY

42

there is one clue that the situation was less comfortable, or perhaps just more
unusual, for the single sex male groups than for the other groups. In the male
conversations there is a disproportionately high amount of humour revolving
around the presence of the tape recorder. Every one of the all male groups
draws some humour from the presence of the tape recorder. In none of the
all female conversations, and in only one of the mixed conversations does this
happen. In all there are 39 instances of humour in the male conversations which
comment on, or are directed at the tape recorder, or the taping situation. This
is in contrast with one instance in the mixed groups, and none in the female
groups.
The examples indicate that the males are generally more conscious of the taperecorder, and so their conversation may be less natural. This could also have
something to do with the fact that I am a female. The speakers know that I will
be listening to the tape, and so are possibly monitoring their speech accordingly.
Awareness of the fact that someone else will be listening to the tape is evident
in a number of examples. In (1) EM's description is accompanied by a visual
aid. PM realises that I will not be able to interpret this fully, and so helpfully
\translates" for the tape.
(1)
EM:

my end up conversation was like

(2 secs)

All:

laugh] 2 secs

PM:

shall we describe what evan did //he had=

All:
PM:

/laugh]\\
=him in a head-lock\

In (2) PM remembers that the tape is running, following a particularly crude

comment. This draws attention to the tape, and what exactly it is that will
be done with it. The fact that they do not really understand the declaration
42

4.5. METHODOLOGICAL PROBLEMS

43

they have signed leads to more coping humour { wordplay based on the word
\corpus."
(2)
PM:

deep voice]: is this tape for general

//publication i hope not:\

All:

/laugh]\\

AM:

ha ha] oh well too bad if it is

CM:

what does it say on the bottom of that form?

ha ha]

PM:

um something like i consent for this to be

CM:

PM:

i give permission for this for the recording of

my voice to be included in a corpus of new

zealand english to be used for linguistics
research purposes
BM:

wow we've got really big words

PM:

begs the question what's a corpus

AM:

yeah ha ha]

PM:

and i always thought it was maybe like a dead

latin

AM:

yeah

CM:

ha ha]

43

CHAPTER 4. METHODOLOGY

44
AM:

or dead company + corpus

When encouraging my friends to make these tapes for me, the males seemed
more concerned with what it is that they were to \do." My female friends did
not seem to
nd it so strange that I wanted to tape their conversations. Related
to this, some humour in the male conversations stems from not knowing if they
are behaving appropriately, or producing the type of data that I need. The
extract in (3) follows a period of time during which the conversation had split
into two small conversations. TM, aware of what they are there for, comments
that this will make things dicult for me to transcribe. CM then latches on to
this and generates some fantasy humour.
(3)
CM:

we had our own conversation bugger off h]

quite quite a cool conversation

TM:

making it difficult for jen to tran- transcribe

it all later=/

MM:
CM:

/=oh it's her //p-\

/oh\\ good does
anyone know any foreign lingo h]

(4) occurs after a group has been playing computer games for ten minutes or so.
PM questions whether this is going to be useful to me. In giving the groups the
tape-recorder I stressed that I wanted just normal conversation. \What do you
want us to talk about" was a common question, to which I replied, just whatever
they would normally talk about. I stressed that they should just relax. In (4)
speakers discuss the most appropriate behaviour under the circumstances. BM's
comment that if they were not playing games or drinking the situation would
not be \normal", supports the idea that sitting around at home and chatting is
behaviour more familiar to women than to men.
44

4.5. METHODOLOGICAL PROBLEMS

45

(4)
(talking about computer game)
PM:

i don't know if we should be i don't know if we

should be playing games making this recording
for jen

MM:

i don't think we're meant to be drinking but

then we wouldn't be doing our normal things

CM:

ha ha]

AM:

thing is if we didn't have the games on we'd

probably be sitting here going //um\

BM:

/er um\\ what

now

CM:

yeah yeah shame about the titanic

PM:

oh no

All:

laugh]

These last few examples indicate that making this tape is seen as a task that
the men have to complete. They negotiate appropriate ways of completing the
task, and also, in the examples I am about to discuss, look forward to completing
the task. The mixed and female groups on the other hand, are merely enjoying
themselves and chatting normally, the tape-recorder is incidental to the activity,
rather than the cause of it. This reects
ndings by other researchers that, in
general, men tend to be more task oriented, and women tend to focus more on
interaction (see Parsons and Bales 1955, Soskin and John 1963).
Tannen (1990b) analysed interactions that occurred when dyads of school children were left to talk. The boys exhibited considerable discomfort with the
situation, whereas the girls had no problem
nding something to talk about.
45

CHAPTER 4. METHODOLOGY

46
She notes that:

The second-grade boys exhibit extreme discomfort in the situation of

sitting in a room with nothing to do but talk. Much of their talk is
about
nding something to do. (Tannen 1990b:75)
Research by Lever (1976) and Goodwin (1982a) showed that just sitting and
talking was an activity frequently chosen by girls, whereas boys more often play
games outside.
The extract in (5) is taken from a conversation recorded in a university oce.
I had left the tape recorder with the four males, the arrangement being that
I would come back around four and collect it, and we would all go and play
monopoly. This example shows that they are keen to get the taping over with,
and get on with the game.
(5)
CM:

i thought she was going to rescue us at four

YOU'RE LATE

All:

laugh]

CM:

yells]: come on:

MM:

we should start with the credits now yeah

NM:

this conversation HAS featured=/

TM:

/=ha ha ha]

Similarly, in (6) CM investigates the tape to see how far through it has run.
(6)
CM:

miles to go

46

4.5. METHODOLOGICAL PROBLEMS

DM:

yawns]

CM:

must be a ninety minute tape

47

MM+EM:laugh]
D:

bugger

M:

ha ha ha]

There are also quite a few fantasy sequences revolving around the tape-recorder,
and what they could say or do that would be particularly witty. (7) is typical of
these.
(7)
EM:

yeah when we'll rewind it back and it she'll

play it back and it goes right from one
conversation ha ha] //and then this other one\

All:
PM:

/laugh] 2 secs\\
or we could go through and edit it later all the
//swear words\

All:

/laugh] 3 secs\\

GM:

yeah ha ha]=/

EM:
DM:

/=yeah
or we could put swear words in it //huh ha]\

EM:

/ha ha]\\
and then put intellectual + words

47

CHAPTER 4. METHODOLOGY

48
PM:

yeha]ah

The examples discussed in this section are representative of the comments directed at the tape-recorder. All participants were happy to record conversations,
but the data suggests that the male groups regarded this more as a task to be
completed than did the other groups. This indicates that the setting was in
some ways less natural for the male groups, and they were more aware of the
tape-recorder, probably inuencing the topics and style of speech. That I am
female may also have inuenced their behaviour. Some speakers addressed me
directly during the course of the conversation. In (8) the speakers realise that
the topic of conversation is not the most suitable given that it is being taped,
and \censor" the recording, addressing me directly.
(8)
BM:

what so it's just a bunch of sheilas prancing

round naked

PM:

yeah basically

CM:

effectively

BM:

i saw enough of that on bloody saturday

AM:

ah

PM:

mm mm tut]

AM:

//what with the meditations h]\

BM:

/loud]: due to censorship\\ we're going to

turn the tape recorder off

All:

laugh]

BM:

loud]: HA HA HA that was good:

48

4.5. METHODOLOGICAL PROBLEMS

AM:

ha ha ha]

PM:

ha ha ha ha ha ha ha ha] oh god

BM:

loud]: sorry i had to cut you out of that:

All:

laugh]

49

In other extracts I am addressed by name, indicating that the tape-recorder is regarded as a
fth party to the conversation, and more importantly a female party.
This brings into question the extent to which this data can be representative of
conversation which would occur in a true all male situation, to which a female
would never be party. Solving this methodological problem without breaching
ethical considerations would be a dicult task.
In considering my results, one should also keep in mind that in general, the all
male groups seem to be less comfortable with the situation than the other groups,
and so their conversation may be less natural. It is dicult to see how one could
resolve this problem without resorting to incomparable data. This is an issue
that researchers will have to take into account, or at least be aware of, when
designing data collection techniques.
Chapter 5 describes the statistical techniques used in this thesis and the following
chapters discuss the analysis of the data, and present and discuss the results.

49

Chapter 5

Statistical methods
Now it happened that Kanga had felt rather motherly that morning,
and Wanting to Count Things...

5.1 Log-linear modelling

In this chapter I discuss log-linear modelling, the statistical technique with which
I have chosen to investigate my data. Much of the information in this discussion
is drawn from the introductory chapters of Gilbert (1981), which provides an
introduction to log-linear modelling for social sciences. Useful books for indepth discussion of the techniques and mathematical grounding of log-linear
modelling are Christensen (1990), and Kennedy (1983). It is a relatively complex
procedure, and here I briey discuss the ideas behind it, and equip readers with
the background necessary to follow the discussion in the following chapters. I
will not delve deeply into the mathematical processes involved in the technique.
I refer readers interested in following up this discussion to the books mentioned
above.
When observed data is presented in table format, it is often possible to identify
patterns from simple inspection. There may be many other patterns in the data,
however, which are not as easy to recognise. The detection of these patterns
requires more sophisticated statistical techniques and analysis. It is possible
to
nd a large number of patterns in any data table, and so the best possible
51

52

CHAPTER 5. STATISTICAL METHODS

approach is to begin with one's prior theoretical knowledge to hypothesise which

social processes may be involved in any given piece of data, and test to see if the
data reects these processes. One simple model, for example, might be that the
functions of humour and the sex of the speaker are directly related.
When one tests a model, an alternative world is created, which is theoretically
identical in all respects to the real world, and shows the relationships speci
ed
in the model. In e ect, this world is the world that would exist if the model
were true. The imaginary world can then be compared to the \real world" as
reected in the observed data. If they are indistinguishable, then there is evidence
for concluding that the model is correct.
The process is not 100 percent foolproof { it would be foolish to expect this
when working with real life data. Even if a model is correct, it is inevitably
a simpli
ed representation of the world. There are many possible contributing
factors which could inuence any given variable. In social research especially, a
model can represent only a subset of these { with luck, the more important ones.
The use of log-linear models, then, is a form of exploratory data analysis. Unlike
more traditional approaches it does not require that analysts have carefully formulated hypotheses to be tested, but, given some notion of the form of suitable
models, allows the in-depth exploration of possible patterns in the data. A hypothesis generally deals with just one relationship, whereas a model may involve
a complex set of relationships.
To illustrate how log-linear modelling works, I will briey discuss an example
of a simple model. In this example a model is used to determine whether hair
colour and eye colour a ect the likelihood that a person will be sunburnt if
exposed to the sun for a given period of time. The data is totally
ctitious, and
deliberately contrived to be simple. Most models deal with data which is much
less straight-forward than this. The observed data is shown in table 5.1.
In log-linear analysis we want to
nd the simplest model which can adequately
explain the observed data. As a
rst step then, we might investigate whether
an independence model can hold, i.e. whether a model showing no relationship
between burning and eye or hair colour is sucient. In order to do this a table
needs to be constructed using an independence model, i.e. a table of \imaginary
data" with which we can compare our observed frequencies. This table should
52

5.1. LOG-LINEAR MODELLING

53

burnt not burnt

blue hair, pink eyes
450
200
green hair, purple eyes 300
350
blue hair, purple eyes
490
160
green hair, pink eyes
320
330
Table 5.1: Example model: Observed frequencies
be the same in all respects as the observed data, but should show no association
between the explanatory variables (i.e. hair and eye colour) and the response
variable.
We can calculate the required frequencies by using the following formula:
model cell entry=(product of corresponding marginals)/table total.
So the predicted frequency for blue hair, pink eyes and burnt is:
(450+200)*(450+300+490+320)/(450+300+490+320+200+350+160+330)=
390.
The results of these calculations are shown in table 5.2.
burnt not burnt
blue hair, pink eyes
390
260
green hair, purple eyes 390
260
blue hair, purple eyes
390
260
green hair, pink eyes
390
260
Table 5.2: Example model: Predicted frequencies according to the independence
model
It is clear that the predicted and observed data tables di er, so there must be
some association present between at least one of the explanatory variables and
burning.
53

CHAPTER 5. STATISTICAL METHODS

54

MAXIMUM-LIKELIHOOD ANALYSIS-OF-VARIANCE TABLE

Source
DF
Chi-Square
Prob
-------------------------------------------------HAIR
1
0.00
1.0000
BURNT
1
102.59
0.0000
EYES
1
0.00
1.0000
EYES*HAIR
1
0.00
1.0000
LIKELIHOOD RATIO

173.58

0.0000

Table 5.3: Example model: SAS analysis for independence model

We can use a computer to generate these predicted frequencies and provide an
objective measure of their similarity. The SAS Maximum likelihood analysis of
variance for the independence model is shown as table 5.3. Here DF stands for
degrees of freedom, and Prob stands for probability.
We would be prepared to accept this model if the chi-square1 value for the
likelihood ratio were under the signi
cance level. The acceptable level for 3
degrees of freedom2 to the 0.05 level is 7.82. The chi-square for the independence
model therefore shows that the model is far from acceptable for explaining the
data. In assessing the suitability of a model we also investigate the probability

gure, which in this case is zero. If the chi-square
gure were at an acceptable
level, then the closer the probability was to one, the stronger the model would be.
The closer the chi-square to zero, and the closer the probability to one, the better
the model. Note that in the independence model I have included the interaction
between EYES and HAIR. It is possible that there may be a correlation between
eye colour and hair colour. We are not interested in any such correlation. For
our purpose eye and hair colour are explanatory variables. We do not want any
uneven distributions in either eye or hair colour to inuence the model. We must
The chi-square compares observed frequencies with those we would expect on the basis of
the model.
2
The number of degrees of freedom is the number of values of the variable which are free to
vary. This is equal to the number of cells less one for each independent linear restriction placed
on the observed cell counts. The signicance level for a chi-square gure varies according to
degrees of freedom.
1

54

5.1. LOG-LINEAR MODELLING

55

therefore
x the hair-eye con
guration (see Kennedy 1987:125). Note that in the
examples I will be discussing in the next chapters, speaker sex and audience
composition are the explanatory variables. The association between these is
therefore included in all of the models.
The independence model is inappropriate so in order to make the model
t the
data we clearly need to incorporate some sort of interaction. Calculation of
model frequencies is less straight forward in non-independence models. There is
no formula, but the frequencies can be calculated using a procedure known as
iterative proportional scaling. A
rst guess is made at a solution. The frequencies
are repeatedly scaled and improved until all three marginals match the marginals
from the original data, either exactly, or so closely that they might as well be
exact. This would clearly be an exhausting process by hand, but computer tools
make it relatively simple. I have used SAS for all of my statistical analysis in
this project.
Next we might investigate the model which shows interaction between all of the
variables. Table 5.4 shows the maximum likelihood analysis of variance table for
this model.
MAXIMUM-LIKELIHOOD ANALYSIS-OF-VARIANCE TABLE
Source
DF
Chi-Square
Prob
-------------------------------------------------BURNT
1
108.67
0.0000
EYES
1
0.03
0.8645
HAIR
1
6.98
0.0082
HAIR*BURNT
1
159.89
0.0000
EYES*BURNT
1
0.68
0.4082
EYES*HAIR
1
0.04
0.8354
LIKELIHOOD RATIO

6.71

0.0096

Table 5.4: Example model: SAS analysis of model showing interaction between
all variables
This model
ts the data better than the independence model. The likelihood
ratio is 6.71. This is not below the chi-square
gure for 1 degree of freedom at
the 0.05 level which is 3.84, but it is certainly closer than the independence model
55

56

CHAPTER 5. STATISTICAL METHODS

was. The probability
gure for the model is only 0.0096, and so the model is far
from ideal. The task now is to simplify the model. If any of the factors can be
removed without signi
cantly increasing the likelihood ratio, then the simpli
ed
model will be a better model. In log-linear modelling we aim to
nd the simplest
model possible which can adequately explain the data.
To decide which variables to remove from the data, we examine the
gures shown
in the Analysis of Variance table. The chi-square
gure for a given variable
indicates the extent to which that variable contributes to the model. We therefore
want to remove any variables which show particularly low chi-square
gures,
particularly if the probability for these
gures is reasonably high. Eyes * hair,
for example, shows a low chi-square with a high probability
gure, indicating
that there is not a signi
cant interaction between the colour of one's eyes and
the colour of one's hair in this model. Remember though, that these are the
explanatory variables, and so this interaction has to remain in the model. The
other variables which seem worth rejecting at this stage are eyes * burnt, and
eyes. If eyes * burnt is not in the model, this indicates there is no relationship
between the colour of one's eyes and whether or not one gets burnt. The SAS
output for this model is shown in table 5.5
MAXIMUM-LIKELIHOOD ANALYSIS-OF-VARIANCE TABLE
Source
DF
Chi-Square
Prob
-------------------------------------------------HAIR
1
6.98
0.0083
BURNT
1
108.65
0.0000
EYES*HAIR
1
0.00
1.0000
HAIR*BURNT
1
159.86
0.0000
LIKELIHOOD RATIO

7.39

0.0605

Table 5.5: Example model: SAS analysis of model of main e ects for hair
This is an acceptable model. The likelihood ratio falls under the chi-square

gure for 3 degrees of freedom to the 0.05 level, and the probability is much
higher than in the previous models. We want to check that this model is not
statistically di erent from the non-simpli
ed model. If it were then the fuller
56

5.1. LOG-LINEAR MODELLING

57

model would be the better one. To check whether the di erence between the
models is signi
cant we look at the di erence between the two chi-square
gures
and check its signi
cance to x degrees of freedom, where x is the di erence
between the degrees of freedom in each of the two models. So for this example,
we subtract 6.71 from 7.39 and check the signi
cance of this to 2 (3-1) degrees of
freedom. This calculation shows that the fuller model is not signi
cantly better
than the model of main e ects for hair, and so we can accept the simpler model.
The fact that eye colour is not included in the model means that variation in the
observed data can be adequately explained by one's hair colour. Eye colour has
no e ect. The model shows main e ects for hair colour.
Now we know that there is an association between whether someone is burnt
and their hair colour. Clearly we are interested in the nature of this association.
We therefore need to measure the strength of association between variables.
To do this, we
rst compute the conditional probabilities given the predicted
frequencies. Having accepted the above model of main e ects for hair colour, we
can look at the predicted frequencies for each cell.
burnt not burnt
blue hair, pink eyes
470
180
green hair, purple eyes 310
340
blue hair, purple eyes
470
180
green hair, pink eyes
310
340
Table 5.6: Example model: Predicted frequencies for model of main e ects for
hair
If we examine these frequencies we can see that eye colour is having no e ect on
whether the person is burnt or not, the
gures are based entirely on hair colour.
This is expected, as the model of best
t was one that showed no association
between burning and eye colour.
Table 5.6 can therefore be collapsed as shown in table 5.7. The conditional
probabilities can be calculated from this table. The probability of a person getting burnt, given they have blue hair is 470/(180+470) = 0.72. The conditional
probability
gures are shown in table 5.8.
57

CHAPTER 5. STATISTICAL METHODS

58

burnt not burnt

blue hair
470
180
green hair 310
340
Table 5.7: Example model: Predicted frequencies for model of main e ects for
hair { collapsed

blue hair
green hair

burnt not burnt

0.72
0.28
0.48
0.52

Table 5.8: Example model: Conditional probabilities

These give us an indication of what is going on. A useful way in which to
investigate these results is to examine the ratio of conditional odds, or odds ratio
for short. We
rst calculate the odds of a person getting burnt for each of the
two hair colours. The odds are calculated by p/1-p. The odds of someone with
blue hair getting burnt are 0.72/0.28 = 2.57:1. Note that the denominator in this
equation is the
gure given for not burnt. The response variable will often have
more than two realisations, and so the probability of an event not happening is
generally calculated by subtracting the probability of it happening from 1. The
odds of someone with green hair getting burnt is less than one. 0.48/0.52 = 0.92.
Now if there were no association between burning and hair colour we would
expect the ratio of these two odds to be equal to one. A comparison of the
respective conditional odds gives a measure of the strength of the association. In
this case, the odds ratio is 2.57/0.92 = 2.79. We can therefore say that the odds
of someone with blue hair getting burnt are approximately 2.79 greater than the
odds of someone with green hair getting burnt.
This is the measure of association I will use to describe most of the relationships
present in my data. I also present probability tables so the reader can observe
the source of these ratios. As a measure of association, the odds ratio has both
advantages and weaknesses. The main advantage is that the resulting
gure is
completely independent of sample size. This allows for considerable comparability with other studies using a similar method. It is also una ected by unequal
58

5.2. CLUSTERING EFFECTS

59

marginal distributions { there are several measures of association which are very
sensitive to skewed marginal distributions. The main disadvantage of the odds
ratio is that it is not a symmetric measure. It is anchored around 1, so that a
ratio of one shows independence. As the association becomes more intense, the
odds ratio approaches either 0 or in
nity. There are ways to make the measure
symmetric around a point which represents no association. This can be done
using Yule's Q for 2*2 tables, or by using log odds rather than the actual odds
ratio. For the purposes of this study, I have tried to keep the statistics as simple
as possible. This is, after all, a linguistics thesis, and not a statistics one! For
both simplicity's sake, and the sake of the reader with minimal statistics background, I have decided to use the simple odds ratio as the measure of association.
In examples for which the odds ratios are particularly high, I will reiterate this
point.

5.2 Clustering eects

Log-linear modelling is ideal for independent data. Unfortunately the data I am
analysing is far from independent. There are three types of clustering involved.

5.2.1 Clustering due to individual behaviour

Men and women are individuals. Women as a group, and men, as a group, may
tend towards certain humour types or strategies. The behaviour of individual
men and women will clearly not always be entirely representative of the group.
Individuals may prefer a certain type of humour, e.g. be predisposed to wordplay.
The more instances of humour collected from any one speaker, the greater the
chance that the data will be distorted due to clustering e ects. I have 72 speakers
in my sample, and have collected several examples from each speaker. The data,
then, is not entirely independent.

5.2.2 Clustering due to conversational ow

Clustering of humour may occur within conversations. If someone starts a fantasy
routine, for example, it is highly possible that this is the direction that humour
59

60

CHAPTER 5. STATISTICAL METHODS

in that conversation will follow for some time. Sparring matches and banter
encourage similar types of humour to occur in clusters within conversations. My
examples are drawn from 18 conversations. A number of examples were taken
from each conversation and so in this sense also, the examples cannot be said to
be entirely independent of each other.

5.2.3 Clustering due to complex examples

Some examples were coded for more than one type, or more than one function.
These were very few, and so I decided to treat them as two examples for the purposes of log-linear modelling. The numbers were so small as to have no signi
cant
e ect on the models, except in one case. This is described in section 8.8.2.

5.3 A statistical check on clustering

One way of accounting for possible clustering e ects is described by Manly
(1992:253). A heterogeneity factor is calculated by dividing the chi-square value
for the
tted model by its degrees of freedom. For each predicted frequency in
the model, we divide the estimated standard error by the standard error multiplied by the square root of the heterogeneity
gure. That is, we calculate SE
Estimate/SE* where SE* = square root of chi-square/DF.
If the resulting
gure is less than two, then this indicates that clustering may
have a ected the model. I performed this adjustment on all of my models and
none of them showed any indication of possible distortion due to clustering.

5.4 The suitability of statistics for analysing conversational data

Conversations are real and dynamic. The performance of statistical analysis on
such data can be regarded as at best indicative. I regard the log-linear modelling
and results discussed in the statistical sections of this thesis simply as tools for
indicating areas of research which may reward in-depth, qualitative analysis.
Conversational data is best suited to qualitative investigation. In chapter 12 I
60

5.4. THE SUITABILITY OF STATISTICS FOR ANALYSING

CONVERSATIONAL DATA

61

illustrate the dangers of being led by numbers rather than being led by your
data. In my analysis I have followed my data in the creation of the taxonomies,
rather than constricting and limiting the categories and trying to force my data
into them. Statistical results are interesting indications, but the most fruitful
and interesting research will be that which goes beyond numbers, and on to
detailed examination of the data itself. Humour research is a relatively new
eld
and one in which there is much research still to be done. Qualitative analysis
of all of the aspects of humour discussed in the following chapters is beyond the
scope of this thesis. The analysis is restricted to statistical analysis, and in-depth
investigation of only a few key areas. Log-linear modelling is a useful tool for
highlighting areas where future research may prove fruitful.

61

Chapter 6

Types of humour
Even at the very bottom of the river I didn't stop to say to myself \Is
this a Hearty Joke or is it the Merest Accident?" I just oated to the
surface and said to myself \it is wet."

6.1 Introduction
In the next six chapters I will discuss the analysis of a number of aspects of my
data. In this chapter I develop a taxonomy for investigating types of humour,
and discuss the results of analysing my data using this framework. In chapter 7,
humour is further categorised according to its freshness. Chapter 8 presents a
taxonomy for functions of humour and strategies used to ful
ll them. In chapter 9 I discuss the focus and topic of the humour in my corpus. Chapter 10
discusses the relative amount of humour contributed by men and women. In
chapter 11 I investigate several aspects of personal identity and how they can be
portrayed and shaped via humour. Finally, in chapter 12, I discuss humour support strategies and show that the issue of humour support is much more complex
than is generally assumed. I describe a range of humour support strategies and
possible reasons why humour may go unsupported.
These aspects of humour are discussed in separate chapters for convenience and
coherence, although they are not entirely independent of one another. This
chapter will concentrate on types of humour.
63

64

CHAPTER 6. TYPES OF HUMOUR

For almost every paper one reads on types or forms of humour there is a new
taxonomy. We are in dire need of a uni
ed theory and a thorough, well regarded
taxonomy. Comparison of results is complicated by di erent and overlapping
sets of categories. Unfortunately I am not going to break the pattern and adopt
someone else's taxonomy in my analysis. Many of the taxonomies are speci
cally
designed for certain contexts for example, the classroom (Neuliep 1991), the
workplace (Vinton 1989), or to form parts of questionnaires (Graham, Papa and
Brookes 1992). Those that claim to be full taxonomies do not provide adequate
coverage of my data.
It is unfortunate to be adding yet another taxonomy to the literature, but perhaps
this too can be seen as suited to a particular context. It is designed for classifying
instances of humour occurring in small, natural friendship groups. I drew on
the huge body of literature (see section 2.6) and uni
ed the discussions and
classi
cations into a small taxonomy of types. This taxonomy was then modi
ed
after an initial examination of my data. Some categories were unnecessary, some
were collapsed, and some categories were expanded.
There are three categories in my taxonomy that do not seem to be mentioned
in past discussions of types of humour. I added the categories fantasy humour
(see section 6.3.2), observational humour (see 6.3.6) and quotes (see 6.3.7). It is
interesting that so many categories which are central to my taxonomy are not
identi
ed in any other research, and equally interesting that many categories
frequently discussed have proved inappropriate or unnecessary for dealing with
my data.
That my data requires its own taxonomy reects a number of factors. Firstly
I am dealing with young New Zealanders. Until now there has been little or
no work done on New Zealand humour. Di erent nationalities have their own
particular sense and brand of humour (Davies 1984, Ziv 1988) and so the fact
that I am working with New Zealanders may be part of the reason taxonomies
constructed overseas were not appropriate. Also, to my knowledge, this the
rst
taxonomy which has been constructed in parallel with close scrutinisation of
recordings of natural, spoken conversations between friends, rather than through
arti
cially elicited data, participant observation, or introspection.

64

6.2. THE TAXONOMY OF TYPES

65

6.2 The Taxonomy of Types

The taxonomy contains the following categories:
1. ANECDOTES
2. FANTASY
3. INSULT
4. IRONY
5. JOKES
6. OBSERVATIONAL
7. QUOTE
8. ROLEPLAY
9. SELF-DEPRECATION
10. VULGARITY
11. WORDPLAY
12. OTHER
Some examples were identi
ed as belonging to more than one of these categories.
In order to simplify the statistics, examples coded in two categories were treated
as two examples. There were not many of these examples and so this did not
signi
cantly a ect the results (see section 5.2.3).

6.3 Dening the Types

6.3.1 Anecdote
An anecdote is a story which the speaker perceives to be amusing. It is not necessarily drawn out, but must impart information. This will most often be about
65

CHAPTER 6. TYPES OF HUMOUR

66

the experiences or actions of either the speaker or someone they are acquainted
with. (9) is an example of a short anecdote.
(9)
PM:

but you know those you know my old three

girh]ls barbara lisa and the other one + um i
supposed to be going out with them on
Wednesdayhay]=/

DM:

/=ha ha ha ha]=/

PM:

/=don't tell
//mark\

DM:

/huh] //ha ha]\

In my taxonomy, brief descriptions of humorous episodes experienced by all

present will also be classed as anecdotes. In such examples the speaker is triggering the memory of a previous humorous incident, so the group can laugh at
it and enjoy the humour again. These examples will be grouped together with
the more typical anecdotes in this part of the taxonomy, but they will be distinguished in terms of their \freshness." Examples of anecdotes about events or
experiences shared by the audience are examples of recycled humour (see chapter 7). Anecdotes can sometimes be long and drawn out, and may contribute
more to the overall humour than other types of humour. It is desirable that this
is reected in the
gure given for the speaker's contribution of humour to the
conversation. A narrative may typically consist of several funny points, or several
sub-plots within the one story. In my data, each separate funny incident which
the speaker describes is counted as an anecdote in its own right. In general,
the more funny components a story has, the longer it takes to tell, so this goes
part way to solving the problem of apportioning sucient weight to anecdotes.
A story about a person who was distracted about something and walked into
a lamp-post on the way to the shop would be one anecdote. A story about a
person who walked into a lamp-post, then went to the shop and knocked over
a big pile of cans, and
nally went to pay and discovered they had no money,
66

6.3. DEFINING THE TYPES

67

would be three anecdotes. The narrative consists of three related stories, each
of which is considered funny in its own right. This is one area in which it is particularly important to observe speaker intentions. There is always a temptation
to code everything that the audience laughs at as an instance of humour. It is
much more dicult to establish the distinct parts of the anecdote, or comments
that the speaker thinks are funny. This problem is discussed in sections 2.2
and 4.4. Example (10) is a narrative sequence which, for the purposes of this
study, includes two anecdotes.
(10)
CM:

h] oh i came off a couple of times and it was

actually the last time that i rode it i was
driving around looking at cars that //were in=

All:
CM:

/laugh]\\
=the paper\ and i you know went and saw one and
i came off and it was a hot day and i hit some
diesel or something on the road and i was going
around a corner (

++
CM:

//so i i determined\

DM:

/(

MM:

ha ha ha]=/

CM:

)\\

/=well i did i determined i'm going

to buy a car TODAY=/

EM:
CM:

/=ha ha]=/
/=and i DID i bought

67

anec]

CHAPTER 6. TYPES OF HUMOUR

68

the next one h] i went //too\

EM:

/coughs]\\ ha ha]=/

CM:

/=it

anec]

was probably a STUPID decision h]=/

MM:

/=ha ha]

CM:

but i've since sold it //to my sister h]\=

DM:
CM:

/not only\\
=evil laugh]

The
rst anecdote is a story about why CM gave up motor-biking. He was out
looking at cars and fell o his bike, so decided that he would de
nitely buy a car
that day. The second anecdote tells what he has done with the bike since. They
are related, but the humour in each of them is drawn from completely separate
sources - the
rst, an incident on the bike, and the second, CM successfully
duping his sister. The two anecdotes have di erent functions, and the speaker
portrays himself di erently in each of them, so splitting them into two anecdotes
eliminates problems which would arise in pinpointing functions. If we did not
distinguish between parts of a narrative sequence in this way, then an entire
conversation could conceivably be interpreted as consisting of a single humorous
instance - an anecdote, when in fact one speaker had spent twenty minutes
describing funny things that had happened to them on an overseas trip.

6.3.2 Fantasy
Fantasy is the construction of humorous, imaginary scenarios or events. This
is usually a collaborative activity, in which the participants jointly construct
a possible (or impossible) series of events. This is perhaps best clari
ed by
exempli
cation. In (11), TM suggests a reason why people have not frequented
a restaurant, despite rave reviews.
68

6.3. DEFINING THE TYPES

69

(11)
DM:

/it's just\\ along um + down that that

way you know can't remember what it used to be
called but yeah no i read a couple of reviews
on it over a various time span and they've both
been sort of saying why haven't more people
discovered this wonderful place inhales]
+

TM:

cause the cook greased them up that day=/

fantasy]

The important point is that all examples of fantasy will involve the construction
of imaginary circumstances or happenings. Example (12) is another example in
this category. The speakers are joking about what they could do on the tape.
(12)
EM:

yeah when we'll rewind it back and it she'll

fantasy]

play it back and it goes right from one

conversation ha ha] //and then this other one\
All:
PM:

/laugh] 2 secs\\
or we could go through and edit it later all the

fantasy]

//swear words\
All:

/laugh] 3 secs\\

GM:

yeah ha ha]=/

EM:
DM:

/=yeah
or we could put swear words in it //huh ha]\

69

fantasy]

CHAPTER 6. TYPES OF HUMOUR

70
EM:

/ha ha]\\
and then put intellectual + words

PM:

fantasy]

yeha]ah

As in the above extract, fantasy humour will typically involve a lot of collaborative humour.

6.3.3 Insult
An insult is a remark that puts someone down, or ascribes a negative characteristic to them. There are two distinct types of humour which come under this
heading. One is jocular abuse, in which the speaker jokingly insults a member
of the audience. The other is when someone absent is insulted. The insult here
is likely to be genuine, and the humour stems from the unexpectedness of the
statement, which in most circumstances would be unacceptable.
(13) is an example of an absent person being insulted.
(13)
DM:

/=she's older than your

insult]

mother
GM:

she just she's just like a couple of beacons

insult]

sh]ort of likh]e sayh]ing take me //ha ha]\

DM:

/huh h]\\

(14) is an example of jocular abuse. Someone present is insulted for humorous

e ect. The insult is jocular, in that it is not intended to o end. On the contrary,
it highlights similarities and serves to maintain solidarity. DF is also living with
her partner and so when she criticises CF for doing the same, although the
humour takes the form of an insult, CF knows it is not meant literally.
70

6.3. DEFINING THE TYPES

71

(14)
DF:

i usually just um turn off the electric blanket

BF:

yeah well i did

CF:

i don't i roll over alex onto the cold side

(//soho]\

AF:
DF:

/oh ha]\\
well chris that //just shows that\ you're a=

BF:
DF:

/good on you\\
=wanton //woman\

insult]

6.3.4 Irony
This category includes examples of both irony and sarcasm. De
nitions of irony
are generally based on one of two criteria: saying the opposite of what you mean,
or saying something di erent from what you mean (Haverkate 1990:77). There
are problems with both these de
nitions (Roy 1978 as cited in Tannen 1984),
and as Tannen (1984:131) points out \to arrive at a satisfying de
nition of irony
would require a major study in itself." The de
nition I have used is a exible
compromise between the two positions. If the speaker does not mean their words
to be taken literally, and in saying them, is implying the opposite, or something
with a markedly di erent meaning, then it is classed as irony. Knowledge of the
speakers and the context will be important in identifying irony. When listening
to speakers I do not know, it is necessary to rely on the surrounding context,
paralinguistic clues, and the reaction of the audience. Some examples of insults
will also be irony, but they will not be placed in this category. This category
is intended solely for verbal irony, and will not include ironic situations. A
description of an ironic situation would be classed as an anecdote. If the humour
stems from some form of situational irony which is co-occurrent with the extract,
71

CHAPTER 6. TYPES OF HUMOUR

72

the humour will be an example of observational humour. (15) is an example of

irony.
(15)
CM:

well yeah when i c- when i came out of the

interview and saw YOU the next person to go in
i thought oh NO dan's going for THIS job jees i
can't have this one=/

MM:

/=ha ha ha //ha ha ha ha]\

DM:

/well i'm glad\

irony]

you think so //highly of me\\

All:
CM:

/laugh]\\
no that's not true at all there's the next day
.....

DM's comment in this extract is not meant literally. There are a number of clues
to this. The
rst is, quite simply, that if the comment were meant literally, then
it would not make sense in this context. The second clue is a very sarcastic tone
of voice, which clearly does not come across in a transcript. DM uses a rise-fall
intonation { a tone frequently used to reinforce irony or sarcasm (Cruttenden
1986). Finally there is CM's reaction \no not at all." CM is not denying the
literal sense of DM's words, but rather the implied meaning, and he goes on to
explain the real reason he did not take the job. (16) is another example of irony,
involving the same speakers.
(16)
CM:

yes well i mean it was i guess fairly shortly

after that that i gave up on motor-biking

MM:

ha ha ha ha ha]=/

72

6.3. DEFINING THE TYPES

DM:

73

/=well i wonder why=/

CM:

irony]

/=ha ha]
well it wasn't //so much that but\...

This is similar to example (15). DM's words \well I wonder why" are not meant
literally, but instead imply that the reason is obvious. Again, CM understands
the irony, and denies the intended meaning of the words. He goes on to explain
that it was not \that," i.e. it was not for the reason that DM had implied, but
for another reason all together.

6.3.5 Jokes
I use the term jokes to refer speci
cally to canned jokes. These are chunks
of humour whose basic form has been memorised. Canned jokes will have a
punchline, or some point at which an incongruity is resolved (Raskin 1985).
They often have a standardised form. Narrative jokes and question and answer
jokes are subcategories of this. A narrative joke is a joke that takes the form of a
story. Jokes such as those beginning \there was an Englishman an Irishman and
a Scotsman...." are typical examples of narrative jokes. Question and answer
jokes are jokes in which the punchline takes the form of an answer to a question.
\Why did the chicken cross the road" jokes are question and answer jokes. I had
originally intended to code for the various subcategories of jokes, but the very
small number of jokes occurring in my data make this pointless. Example (17)
is an example of a joke taken from my corpus.
(17)
BM:

well it's a bit bit like that er joke about

what's the difference between a hedgehog and a
range-rover

PM:

yeah h] range-rover's got pricks on the

//inside\

AM:

/inside\\

73

CHAPTER 6. TYPES OF HUMOUR

74
BM:

ha ha]

6.3.6 Observational
This category was born of the fact that a lot of the examples originally categorised
as \Other" seemed quite similar. Humour in this category consists of quips or
comments about the environment, the events occurring at the time, or about
the previous person's words. In this sense it is \observational" - the speaker is
making an observation about something funny, or making a witty observation.
In (18) LM makes an observation about RF's comment, which he
nds funny.
(18)
RF:

JUST been reading the lord of the rings myself

frodo and sam are PRETTY CHUMmy too

All:

laugh] 2sec

LM:

ha ha ha] oh my god //it's sort of a\ looking=

SF:
LM:

/ha ha ha ha]\\
=for sexual deviance and //degradation of the=

SF+RF:
LM:

/laugh]\\

=lord of the rings\=/

In (19) DM observes that the group has run out of chips. The group is at my
at, and they know that I am going to be listening to the tape. The comment is
therefore funny because it is an observation that would usually go unsaid. DM
rustles and crunches the new packet right beside the tape-recorder, and MM
makes the observation that they are now left with a bag of crunched up chips.

74

6.3. DEFINING THE TYPES

75

(19)
DM:

silly voice]: we need more chips:

All:

laugh]

CM:

they have actually supplied us with some

DM:

thank you

rustles right by tape recorder]

obs]

other]

EM:

ha ha ha ha ha ha]

MM:

we now have these laughs]: crinkly:

obs]

6.3.7 Quote
A quote is a line taken from a television show or a movie, usually a comedy.
It depends very much on the group as to whether this type of humour is used.
Some groups quote extensively, and have memorised whole routines which they
launch into given the slightest prompt. This is a very performance based type
of humour. The speaker establishes solidarity with members who can recognise
and laugh at the quote. They also gain a certain amount of prestige for being
able to memorise the quote and deliver it satisfactorily. Appreciation of the
humour requires a certain amount of in-group knowledge. Example (20) follows
discussion about which episodes of the Muppets people had seen. LM mentions
that he has seen the Muppets episode in which Steve Martin is guest star. DM
then quotes some lines from the Steve Martin episode, pauses slightly, then says
\
ve
ve
ve," a quote from LA Story, a Steve Martin movie. This humour is
entirely for LM's bene
t, as he has recently seen the Muppets episode, and he
and DM saw LA Story together.
(20)
LM:

/=i got i got

75

CHAPTER 6. TYPES OF HUMOUR

76
to see steve martin
TM:

you need one of those g codes that ha huh]

DM:

exhales] quotes]: he's a rambling guy:

LM:

nh nh] yeah

DM:

quotes]: you loved rambling guy ha ha]

you'll LOVE juggling guy:
+

DM:

tut] quotes]: five five five: h]

LM:

nh h huh]

6.3.8 Roleplay
Roleplay is the adoption of another voice or personality for humorous e ect. The
speaker steps into someone else's shoes. Roleplay could be quite speci
c mimicry
of a particular person, or just the general adoption of a stereotypical voice or
attitude. This is very much performance-based humour the speaker is acting
for their audience. This type of humour, like fantasy humour, has received very
little attention in the literature, although both are relatively common in spoken
discourse. Morreal (1983) identi
es mimicry as a type of humour, although this
is restricted to the mimicry of a speci
c person. Example (21) is an example of
the mimicry of a speci
c person's actions on a particular occasion. By imitating
the person, the speakers ridicule him.
(21)
GM:

dave and him were having a good old time weren't

they=/

76

6.3. DEFINING THE TYPES

EM:

77

/=fuck and they were really

getting into it they thought it was
excellent=/

GM:

/=ha ha ha ha ha] what about those

women that were //(

EM:

)\

/what like\\ imitates dave]:

roleplay]

oh YEAH mate YEAH oh yeah:

GM:

imitates dave]: fuck i'm going to waste you today:

All:

imitate dave for 1 sec]

EM:

imitates dave]: YEAH you know where i'm coming

roleplay]

roleplay]

from:
GM:

yeah //he was so full of BULLshit talk\

In (22) the speakers are discussing an incident in which a customer at a restaurant

was accidentally served dish-washing liquid. VF places herself in the shoes of the
person who did this.
(22)
VF:

oh chevy's managed to do one of the more major

impressive fuck ups though

SF:

yeah yeah ha ha] dishwashh]ing lihi]quidh]

//oh gross\

VF:

/loud]: OH\\ i wonder if this is REALLY

dish//washing liquid\ let me taste //it\=

SF:

/ha ha ha ha]\\

/ha ha=

77

roleplay]

CHAPTER 6. TYPES OF HUMOUR

78
VF:

=EVERYBODY have a try //ha ha]\ excuse me=

SF:

=ha]\\

VF:

=madam //voc] would youho] carehe] fh]or a=

All:
VF:

/ha ha]\\

/laugh]\\
=bh]it\....

6.3.9 Self Deprecation

Self deprecation is an insult directed at oneself. In (23) DF has farted. She
herself comments that the fart smells bad. This is a defence strategy. If she
points this out herself, then any insults directed at her because of the fart will
hold much less force.
(23)
BF:

DAYna

DF:

shh]it this is bh]ad man

All:

laugh] 5 secs

6.3.10 Vulgarity
It is clear that it is possible for examples in all of the categories to be crass. Jokes,
wordplay, insults etc. can all have an aspect of vulgarity. For these cases, the
vulgar aspect would be picked up in the classi
cation of topics, and so the extra
category would be unnecessary. There are some examples however, in which the
sole source of the humour is its crassness. Toilet humour and sexual humour will
be typical instances of vulgarity. It is clearly desirable to be able to distinguish
between toilet humour and humour about sex, but this distinction will be dealt
with in the analysis of topics. Putting both in one category of humour can
78

6.3. DEFINING THE TYPES

79

therefore be justi
ed. The humour in both cases derives from the fact that the
speakers are breaking some sort of taboo. (24) is an example of vulgarity.
(24)
GM:

/=yeah oh i had a

vulgarity]

fantastic crap in there this morning=/

DM:

/=//ha ha=

EM:

/ha ha=

DM:

=ha]\

EM:

=ohh] ha //ha ha ha ha]\\

6.3.11 Wordplay
Wordplay is any humorous statement in which the humour derives from the
meanings, sounds or ambiguities of words. The most typical instance would be
a pun. This may involve a speaker deliberately punning, or a hearer identifying
an ambiguity in the speaker's speech and exploiting it for humorous e ect. (25)
is an example of wordplay.
(25)
MM:
NM:

associative databases=/
/=deductive=/

MM:

/=deDUCtive
databases //that's the ones yeah\

NM:

/yeah yeah\\ well //there's\

TM:

/well i de-\\ d- i

79

wordplay]

CHAPTER 6. TYPES OF HUMOUR

80
deduced that
MM:

groans]

In (26) DF identi
es an ambiguity in CF's comment, and pretends to mistake

CF's intended meaning of pulse = heart beat for pulse = legume. CF plays along
with DF's deliberate misinterpretation, and together they fool BF.
(26)
CF:

i mean i've got bad feeling in my hands anyway

BF:

have you

CF:

like i can never feel pulses or stuff like like

you know

DF:

pulses what like beans? like beans? you mean

BF:

NO

DF:

pulses you mean //kidney beans\ and the like

CF:
CF:
BF:

wordplay]

/yeah\\
and lentils=/
/=oh does she h]=/

CF:

/=i find it really

hard to feel lentils

Wordplay is not con
ned to puns. It may involve exploiting similarities or differences between words in a variety of ways. (27) includes two examples of
wordplay.
80

6.3. DEFINING THE TYPES

81

(27)
DM:

h]h] there's one of them's called the non

born king gwen keeps saying it's the non
bonking king //h]huh]h]\ non bonkh]ing

GM:

wordplay]

/uninterested]: oh yeah:\\

GM:

yeah

DM:

the thing is that once you hear it it's

like you know like that //really useless=

GM:

/singing to tape=

DM:

=song you know\ climb every woman=

GM:

=in background]: it's right to know:\\

DM:

=song=/

wordplay]

The
rst example derives its humorous e ect from running several together. In
the second example, the words of a song \climb every mountain" are altered for
humorous e ect.

6.3.12 Other
There is also an \other" category, for humour which does not slot into any of the
ten main categories of humour. A taxonomy of types of humour which claimed
to encompass every possible example would be either incredibly huge or contain
particularly general categories.

81

CHAPTER 6. TYPES OF HUMOUR

82

6.4 Results
In this section I discuss the results of applying the taxonomy of types of humour
to my data. The model showing interaction between type of humour and both
sex and group composition returns a 7.49 chi square with 11 degrees of freedom
and 0.7580 probability. This is a good
t, and is signi
cantly better than the
independence model and either of the models showing main e ects for just one of
the variables. Table 6.1 shows the results of this section of my research in terms
of conditional probabilities for speakers in each of the four groups. Here the

rst letter stands for the sex of the speaker (M/F) and the second for the group
composition (C=cross sex, S=single sex). The probability
gure given for FC
anecdote is therefore the probability that a speaker will tell an anecdote given
they are a female speaker in a mixed sex group. So the probabilities of each of
the four columns should add to one. In fact, they add to between 0.99 and 1.01
due to rounding.
FC
anec 0.43
fant 0.28
insult 0.04
irony 0.06
joke 0.00
obs
0.09
quote 0.00
role
0.03
self
0.00
vulg 0.00
word 0.05
other 0.03

MC
0.39
0.31
0.03
0.06
0.00
0.05
0.00
0.05
0.00
0.00
0.07
0.03

FS
0.33
0.23
0.09
0.07
0.00
0.11
0.01
0.06
0.01
0.01
0.04
0.04

MS
0.29
0.25
0.06
0.07
0.00
0.06
0.03
0.10
0.01
0.02
0.06
0.05

Table 6.1: Probability
gures for types of humour

Anecdotes and fantasy are by far the most frequently occurring types of humour
for all four groups. The odds ratios for types of humour are given in table 6.2.
The two groups being compared are joined by a hyphen, with the group more
likely to use that particular type of humour
rst. So I only ever
ll in one of
FC-MC and MC-FC, unless the odds ratio is one. I have marked with an asterisk
82

6.4. RESULTS

83

positions where I could not possibly place a value, because one of the values was
zero, making the calculation of an odds ratio impossible.
anec
fant
insul
irony
joke
obs
quote
role
self
vulg
word
other

FC-MC MC-FC
1.18
{
{
1.15
1.35
{
1.00
1.00
*
*
1.88
{
*
*
{
1.70
*
*
*
*
{
1.43
1.00
1.00

FC-FS FS-FC
1.53
{
1.30
{
{
2.37
{
1.18
*
*
{
1.25
*
*
{
2.06
*
*
*
*
1.26
{
{
1.35

MC-MS MS-MC
1.57
{
1.35
{
{
2.06
{
1.18
*
*
{
1.21
*
*
{
2.11
*
*
*
*
1.18
{
{
1.70

FS-MS MS-FS
1.21
{
{
1.12
1.55
{
1.00
1.00
*
*
1.94
{
{
3.06
{
1.74
1.00
1.00
{
2.02
{
1.53
{
1.26

Table 6.2: Odds ratios for types of humour

The women use anecdotal humour more often than men, although the di erence
is not as pronounced as might have been expected. Both sexes use anecdotal
humour slightly more in mixed sex conversations than single sex conversations.
Fantasy humour is used slightly more by males, and slightly more in mixed sex
than in single sex conversations.
The odds of a speaker using a jocular insult in single sex interaction are more than
twice the odds than for in mixed interaction, and in both types of interaction,
women are slightly more likely than men to use insults.
The amount of irony used by men and women in both group compositions is
more or less equal. There is an indication that irony may be slightly more likely
to occur in single sex conversations, but the di erence is very small.
Only one joke was told and this was in a single sex male group. The model
therefore predicts the probability of a joke occurring in any of the groups to be
close to zero.
The odds of a female using observational humour are almost twice as high as
the odds of a male using it, and there is a slight indication that observational
humour may be more likely in single sex conversations than mixed conversations.
Quotes are more likely in single sex male groups than any other. The percentage
83

84

CHAPTER 6. TYPES OF HUMOUR

of quotes occurring overall in the corpus is very small. The model predicts that
they occur in single sex and not mixed sex conversation.
Males are more likely than females to use roleplay humour, and for both sexes
the odds of using roleplay humour are more than twice as high in single sex
groups than in mixed groups.
Self deprecation was used only in single-sex groups. The numbers are very small,
but the model predicts both men and women equally likely to use this type of
humour.
Similarly, there was a small amount of vulgarity humour. This too, occurred only
in single sex groups - and was more likely in male groups than female groups.
Finally, wordplay seems more likely to be used by men than by women over both
group compositions. Both sexes uses slightly more wordplay in mixed groups
than single sex groups.

6.5 Discussion
Crawford and Gressley (1991) elicited subjects' impressions on various matters,
including the types of humour they use. Their results led them to conclude that
males use more formulaic joking, whereas females use more anecdotal humour.
Ervin-Tripp and Lampert (1992) also note that women in their corpus of spoken
humour volunteered real stories about themselves, whereas men were much less
likely to do so.
Personal anecdotes are less aggressive than forms of humour such as canned jokes
or wordplay, because they do not impose an understanding test on the listener
(Norrick 1994). The anecdote provides an amusing story which invites listeners to
laugh and participate by adding comments of their own. As female speakers are
generally more supportive and value intimacy and involvement in conversation,
it is not surprising that women have been shown to utilise this form of humour
more than men.
That females use more anecdotal humour than males is in part supported by
the results of this study. Women use slightly more anecdotes than men. Men
increase their use of anecdotes in mixed interaction, possibly accommodating to
84

6.5. DISCUSSION

85

the women's use of this type of humour. It is interesting to note that the women
as well as the men in this study increased their use of anecdotes when in mixed
sex interaction. That the percentage of anecdotes for both sexes increases in
mixed sex interaction may be a reection of the fact that there are certain types
of humour which seem to be preferentially used in single sex company rather than
mixed { i.e., in this study, insults, self deprecation and vulgarity. The fact that
these are less likely to occur in mixed interaction may account for the fact that
the odds of an example of humour being an anecdote increase in mixed groups.
That men are slightly more likely to use fantasy humour in mixed sex interaction may have something to do with the responsive audience provided for them.
Fantasy humour is often a joint display, where speakers bounce o each other
and jointly build up a hypothetical scenario. In chapter 12 I discuss support
mechanisms for humour. It has been claimed that women are more \supportive"
of humour, although this claim refers almost exclusively to humour support via
laughter. Men often support each other in other ways, one of which is to continue
on the humour sequence, and to spar with each other. It is sometimes the case,
then, that men will begin a fantasy routine in a mixed group. The women will
usually support this with laughter, leaving the men free to carry on the routine,
sparring with each other and delighting in the laughter their audience is providing. This is a performance of sorts. In a single sex group, males are less likely
to be given ongoing support via laughter, but instead all four speakers will want
to contribute to the sequence. This leads to short lived chaos and amusement,
but the fantasy is less of a performance, more of a confusion, and hence shorter
fantasy routines result. This seems a likely reason why men may capitalise on
the laughter o ered in mixed sex groups, bounce quips o each other and enjoy lengthy fantasy sequences, and support from their female audience. Women
are slightly more likely to use fantasy humour in mixed groups than single sex
groups. This could be accommodation for men's preference for fantasy humour.
The results show that jocular insults are signi
cantly more likely to occur in
single sex interaction than mixed interaction. Such humour is usually employed
to tease someone present. The distribution of insults is therefore very similar
to the distribution of teasing humour. I discuss this and possible explanations
for the fact that insults and teasing occur more in single sex than in mixed sex
groups in section 8.9.2.
85

86

CHAPTER 6. TYPES OF HUMOUR

Interestingly, women seem slightly more likely than men to use insults as a form
of humour. Coates (1986:153) claims that insults, along with shouting, namecalling and threats, are part of male aggressiveness, whereas women try to avoid
such displays,
nding them unpleasant and often interpret them as being meant
personally. It may be true that women avoid genuine insults, but joking insults, such as those identi
ed in this study, seem to be just as much a part of
females' repertoire and banter as males', if not more so. It was only recently that
researchers identi
ed such behaviour in female groups (see Eder 1990), after a
long lived assumption that it was a male only activity. This prevented any sort
of comparative research { why investigate gender di erences in an activity only
indulged in by males? This study indicates that, while conversing in single sex
groups, females are just as likely as males to use jocular abuse. A qualitative
study into the nature of such abuse in male and female conversations would no
doubt prove fascinating and fruitful.
There are no signi
cant di erences between the di erent groups' use of irony.
That only one joke occurred in my data is particularly interesting considering the
large amount of humour research concentrating on joke telling. The hypothesis
put forward by several researchers, including Lako (1975), that men tell more
jokes, and tell them better than women can unfortunately be neither con
rmed
nor denied on the basis of my data. The one joke that did occur was in an allmale conversation, but this hardly tells us much. Perhaps the main information
to take from this is that jokes seldom occur in relaxed conversational English,
and, in this corpus, form 0.1 percent of the total humour. It is time, perhaps, to
give the other 99.9 percent its fair share of the attention!
Observational humour is used more by women than by men. Jenkins (1985)
noted that women's humour tends to be more context bound, whereas men's is
more performance-based and often transferable from one context to another.
Indeed, performance oriented types of humour are used more often by the men
in my sample than the women. The use of quotes is one example of this. That
quotes are more likely to occur in male conversations than female conversations
could also be a reection of the
nding that men are more likely to talk about
television shows, books or movies they have seen (see section 9.3). Men tend to
use external source humour more than women (chapter 7).
86

6.5. DISCUSSION

87

Roleplay is also a very performance oriented type of humour. Therefore it is

unsurprising that this is more likely to be used by males than by females. It is
interesting that roleplay is signi
cantly more likely to occur in single sex groups
than mixed groups. The success of roleplay humour relies very heavily on group
knowledge. Members agree on who is an appropriate target for imitation or
ridicule, and successful roleplay relies on an audience's recognition of the performance. The single sex groups in my corpus tend to have more group history {
many are old school friends, and so they have more material and inclination for
roleplay. Also, as roleplay is very performance-based and on-record, the loss of
face involved if it fails will be much larger than for other types of humour. This
could be a contributing reason why both men and women are more likely to use
roleplay humour in single sex groups. These may provide safer environments.
Self deprecation is used equally by men and women, but only in single sex groups.
It is important to remember that this is used as a purely formal category. The
category includes only insults directed at oneself. It is possible to put oneself
down indirectly through other types of humour, telling an anecdote for example.
There are surprisingly few examples in this formal self deprecation category.
Much has been made of the \fact" that women use a lot of self deprecating
humour. In fact, there is very little evidence to support this, and most claims
can be traced back to Levine (1976 - as cited in Neitz (1980)). Levine studied
female comics and found them most likely to disparage themselves. Several other
studies have led to the claim that women use self-disparaging humour more than
men. These are largely based on response to humour and humour preferences,
rather than actual production (Zillman and Stocking 1976, Zillman and Cantor
1976). There is, in fact, very little evidence to suggest that this pattern is also
present in spontaneous humour occurring in friendship groups. Jenkins (1985)
suggests that this myth may have been helped by misinterpretation of women's
humour. Women tell a lot of anecdotes about embarrassing things they have
done, or other anecdotes which could be seen as portraying them negatively,
and a non-observant analyst may class such humour as self deprecating. Jenkins
points out that this is mistaken.
That women tell jokes or laugh at themselves is negative only in the
competitive atmosphere of men where the intent is to show yourself
o to the best advantage. (Jenkins 1985:138)
87

88

CHAPTER 6. TYPES OF HUMOUR

Such humour, then, is not self deprecating, but self healing. If a bad experience
is turned into a funny story that can be shared, everyone feels better. Whether
or not a researcher classes an instance of humour as self deprecating may depend
largely on their value system, on their knowledge of the value system of the
speaker, and on the range of functions they attribute to self deprecating humour.
In some cases they may be misled. In chapter 11 I claim that a speaker will almost
never portray themselves negatively. Humour which researchers may claim is self
deprecating may actually be serving to portray an open and sharing identity, or
to heal and cope with problems. I therefore regard the categorisation of self
deprecation as a function of humour as inherently problematic. It can more
accurately be regarded as a type of humour, which can serve any number of
functions.
So the fact that so few of my examples fall into the category \self deprecation,"
and that women in my sample did not use this more than men, is largely due
to the fact that I used the category as a purely formal description of a type of
humour which may serve any of a number of functions, rather than the more
encompassing, inherently awed approach of treating it as a function itself.
That vulgarity occurred only in single sex groups reects a general perception of
gender divisions within society. Men do not want to be disrespectful, or insult
any women present. This attitude is often observable in groups in which men are
scolded for using vulgarity when there are \ladies in the room." It is considered
unladylike to use vulgarity, and so women avoid this in the presence of men.
Both sexes avoid the use of vulgarity humour in mixed interaction, although
both indulge in it in single sex groups. This is similar to the situation with
insults and teasing humour. The same pattern was observed by Folb (1980).
Finally, wordplay is more likely to be used by men than by women over both
group compositions. This is a slight trend, but not signi
cant. Contrary to
Holmes' claim that \there is no such thing as a female punster"(Holmes 1864 as
cited in Redfern 1984) this type of humour was used by both men and women.
Wordplay is a performance-based type of humour, and often has an element of
competition. Norrick (1994) points out that punning disrupts ongoing interaction. Puns can interrupt and redirect conversation, and oblige participants to
disrupt the ow of interaction in order to acknowledge the pun. In contrast
with more narrative types of humour, wordplay directly challenges the hearer
88

6.6. CONCLUSION

89

by testing them. There are certain circumstances in which wordplay may serve
to create cohesion, particularly if banter occurs, through the demonstration of
shared background knowledge and understanding and laughing together. Sherzer
(1985) notes that puns can function both disjunctively and cohesively. Wordplay
is also an excellent means of displaying wit, a quality valued in male groups (see
chapter 11). So, being performance-based, and often competitive or disruptive, it
makes sense that wordplay is a form of humour used more by males than females.

6.6 Conclusion
Very little work has been done on gender di erences in the use of types of humour.
There is therefore a minimal body of work with which to compare the results of
this study. This research has identi
ed patterns in the use of various humour
types, and so provides a more solid base for future research and hypothesis
building. Statistically, the most signi
cant results are that insults and roleplay
are more likely to occur in single sex conversation, women are more likely to
use observational humour than men, and humour involving quotes or vulgarity
is unlikely to occur in mixed conversations, and is more often used by men than
women.
Other trends show anecdotes and fantasy more likely to occur in mixed conversations than single sex conversations, men more likely to use roleplay and wordplay
than women, and women more likely to use insults.
The taxonomy used to categorise the data was constructed in parallel with close
analysis of the data. This distinguishes it from most taxonomies of humour
types, which are largely based on questionnaire data, introspection, or other less
reliable techniques. It is also the
rst taxonomy to be constructed speci
cally
for categorising New Zealand humour.
In the following chapter I discuss the further categorisation of humour according
to its \freshness."

89

Chapter 7

Fresh Humour and Seconds

Humour
7.1 Background
When coding my examples I made the distinction between fresh humour and
seconds humour. Fresh humour is humour that is created on the spot. Seconds
humour is humour that originated elsewhere and is being repeated in the extract
under analysis. Seconds humour is further classi
ed into three subsets. The

rst is external source humour (Neuliep 1991:350). External source humour is
humour that is derived from elsewhere, perhaps a humorous historical event,
or something funny that happened in a movie. Relating something funny that
a friend said would be an example of external source humour. Jokes are also
external source, unless it is clear that the joke is of the speaker's own design.
Example (28) is an example of external source humour.
(28)
DM:

h]h] there's one of them's called the non

born king gwen keeps saying it's the non
bonking king //h]huh]h]\ non bonkh]ing

The boundary is not always clear cut, and the situations I have described as
being external source humour may on the surface seem very arbitrary. What
91

CHAPTER 7. FRESH HUMOUR AND SECONDS HUMOUR

92

about anecdotes, for example, describing something funny someone else did?
If someone saying something funny is external source humour, then one would
expect someone doing something funny to also be external source humour. The
crucial point with regards the speech or action of another individual, is whether
they were trying to be funny or not. If my friend said something particularly
witty, and I relate this in a later discussion, then I am using external source
humour. If my friend said something exceedingly stupid, and I repeat it and
laugh about it in a later discussion, this is fresh humour. In the
rst example,
the humour comes from my friend and I am just repeating it. In the second
example, the humour comes from me. I have imposed humour on the situation.
The same applies for actions. Relating how someone executed a cunning practical
joke is external source humour. Relating how someone walked into a lamp-post
is not.
The second subset of seconds humour is a type of humour I have labelled recycled
humour. When humour is centred around an event also experienced by members
of the audience, then the humour may function quite di erently than when the
knowledge is unique to the speaker. If the audience was present when the event
occurred, an anecdote may be less of a story, and more of a reminiscence, reminding the audience of the humour of the situation. If the experience was novel to
the speaker, then they are telling a story for the amusement of the audience, and
its humour will have to be clearly explained. When the experience is a shared
one, the humour will consist mainly of key phrases, just enough to trigger the
event in the audience's mind. An example of recycled humour is shown in (29).
Four males are discussing a night on the town.
(29)
GM:

dave and him were having a good old time weren't

they=/

EM:

/=fuck and they were really

getting into it they though it was
excellent=/

GM:

/=ha ha ha ha ha] what about those

women that were //(

)\

92

7.1. BACKGROUND
EM:

93
/what like\\ imitates dave]:

oh YEAH mate YEAH oh yeah:

GM:

imitates dave]: fuck i'm going to waste you today:

All:

imitate dave for 1 sec]

EM:

imitates dave]: YEAH you know where i'm coming

from:

GM:

yeah //he was so full of BULLshit talk\

The third category of seconds humour is joint story-telling. This is when some
members of the group are reminiscing about a humorous event, while jointly
relating the story to the other participants. This will have characteristics of
both fresh humour (relaying a story for the amusement of others) and recycled
humour (reminiscing about a shared experience, probably creating solidarity). It
therefore requires its own category. This category is speci
cally for instances in
which two or three members of the group are relating something humorous. It
does not include jointly constructed fantasy, punning or any fresh humour. Joint
story-telling is a subset of seconds humour.
Example (30) is an example of joint story-telling. AM and BM are relating the
events in a movie they had seen to the rest of the group. It is also an example
of recycled humour.
(30)
AM:

/=and then like

later on in the movie when there's two of them
left=/

BM:
AM:

/=yeah=/
/=and one guy's um + you know they
rob this um store this gas station because the

93

94

CHAPTER 7. FRESH HUMOUR AND SECONDS HUMOUR

real mental psycho guy g- um just kills just
laughs]: strangles: the guy behind the
counter=/

CM+DM:

/=laugh]=/

BM:

/=yeah they were stealing like

um what was it=/

AM:

/=yeah they laughs]: //were=

BM:

/bags of=

AM:

=stealing the pet food:\

BM:

=lollies\\ //and shit like this\

ALL:
BM:

/laugh]\\
kill them we want the fucking fruit bursts=/

AM:
/=laughs] and then he strangled him to death=/
DM+CM:
/=laugh]=/
AM:
BM:

/=bugger it=/
/=laugh]=/

AM:

/=high pitched]:
shit you killed him:=/

94

7.2. RESULTS

95

7.2 Results
The best log-linear model for explaining the freshness of the humour shows main
e ects for both the sex of the speaker and the group composition. This model

ts extremely well, with a likelihood ratio of 0.05 with 3 degrees of freedom, and
a probability of 0.9972
Table 7.1 shows the resulting probabilities. Recall that the
rst letter indicates
the sex of the speaker, and the second the group composition, where S stands
for single sex, and C cross-sex interaction.
external
fresh
recycled
joint

FC
0.04
0.96
0.00
0.00

MC
0.07
0.93
0.00
0.00

FS
0.04
0.93
0.03
0.00

MS
0.07
0.86
0.03
0.04

Table 7.1: Probabilities for \freshness" of humour

It can be seen that fresh humour is used far more than any of the other categories.
All groups use a small amount of external source humour, recycled humour occurs
only in single sex conversations, and joint storytelling only in all male groups.
The odds ratios are shown in table 7.2.
external
fresh
recycled
joint

FC-MC MC-FC FC-FS FS-FC MC-MS MS-MC FS-MS MS-FS

1.81
1.00
1.00
1.00
1.00
1.81
1.81
1.81
2.16
2.16
*
*
*
*
*
*
1.00
1.00
*
*
*
*
*
*
*
*

Table 7.2: Odds ratios for \freshness" of humour

The odds of men using external source humour are 1.81 times higher than the
odds of women using it, but group composition and use of external source humour
are independent. In single sex groups fresh humour is more likely to be used by
females than males. Males increase their use of fresh humour when in single
sex conversations. Recycled humour occurs only in single sex groups and is
independent of the sex of the speaker. There is very little joint story-telling,
and most of this occurs in one particular all-male conversation. It only occurs in
95

96

CHAPTER 7. FRESH HUMOUR AND SECONDS HUMOUR

all-male conversations and so odds ratios can not be calculated.

7.3 Discussion
The most interesting result found from investigating this aspect of humour is
the fact that men are more likely to use external source humour than women.
This can be explained by several factors. First, Kiper (1987) suggests that men
are more likely than women to discuss things that they have seen or read. This
was backed up by my data (see chapter 9). Movies or books are a common
resource for external source humour. Secondly, Jenkins (1985) points out that
men's humour tends to be less context-bound than women's. Their humour can
be stored, and performed in any number of contexts. The
nding that men use
more external source humour than women supports this claim.
In chapter 8 I move on to discuss a taxonomy for categorising the function of a
given instance of humour, and discussing the distribution of these functions in
my data.

96

Chapter 8

The Functions of Humour

8.1 Introduction
In chapter 6 I de
ned and discussed the di erent types of humour. Each type
of humour can be used for strategies which can ful
ll a number of functions. In
ful
lling the function of the humour, the speaker constructs certain aspects of
their identity. In this chapter I will outline a classi
cation system for functions
of humour and discuss the results of using this taxonomy to analyse my data.
The taxonomy grew from the literature outlined in section 2.7. I began with the
functions discussed there, and further developed a taxonomy which incorporated
the most relevant of these and seemed to
t the natural categories into which
my data fell.
Graham, Papa and Brookes (1992) review the literature and produce a list of
24 documented functions. All of the functions they identify can be broadly
categorised into one of the categories in the taxonomy I have developed.

8.2 The Framework

This framework assumes that every attempt at humour is an attempt to both
express solidarity with the audience and construct a position of respect and status
within the group. Tannen (1993:167) points out that, although solidarity and
97

CHAPTER 8. THE FUNCTIONS OF HUMOUR

98

General

Functions

Psychological

Solidarity

Defend

Power

Cope

Situational

General

Strategies

Share

Highlight/
Capitalise

BoundS

TeaseS

Other

Conflict

Control

BoundP

TeaseP

Figure 8.1: The functions of humour

power may on
rst consideration appear to be opposites, each also entails the
other. This is particularly true of humour. There is a sense in which whenever
you attempt humour and it succeeds, your status within the group is positively
a ected. You have amused the audience and so illustrated that you share with
them a common idea of what is funny. This serves to create or maintain solidarity.
Some instances of humour will have this general function, and no other, beyond
the creation of a positive self-identity.
Figure 8.1 shows the structure of the taxonomy developed for this section of my
research. The general function as described above is at the head of the tree. All
humour will serve this function. Those examples which can not be classi
ed further as serving primarily solidarity based, power based or psychological functions
will simply be categorised as serving the general function. Categories which are
potential
nal-level labels are shaded in grey. The next level of the taxonomy
shows overall functions which examples could ful
ll. Humour will normally serve
primarily solidarity based, power based or psychological needs.
98

8.3. SOLIDARITY

99

The psychological category contains the functions \to defend" and \to cope".
Coping humour is further categorised into two more speci
c functions. I do not
identify speci
c strategies through which the psychological functions could be
ful
lled. Strategies such as putting oneself down before someone else does, or
making light of a serious situation would serve to ful
ll psychological functions.
As the psychological categories are already quite speci
c at the function level,
strategy level categories would be particularly detailed and contain very few
examples. I therefore felt it sucient to code such examples only by the speci
c
function they served. The function of forming or maintaining solidarity, however,
is much broader. There are several recurrent strategies that speakers use in order
to ful
ll this function. The same is true of humour which functions speci
cally
to exert power.
The diagram makes the interrelation of the di erent levels, strategies and functions appear relatively straight-forward. This is an attempt at representing
graphically the taxonomy I am using. The distinction between strategy and
function is less distinct than I implied in the preceding paragraph. Strategies
can be seen as more precise descriptions of functions. One could go on forever
postulating various levels of such a tree, and the application of labels to each of
the levels, while convenient, is somewhat arti
cial. The use of the tree to represent the taxonomy also implies that each example can be identi
ed as
tting
into one and only one of the categories. In fact, it is possible for an instance of
humour to simultaneously ful
l several of the identi
ed functions. One example
could serve both psychological and solidarity functions.
In the following sections I de
ne and exemplify each of the categories in the
taxonomy.

8.3 Solidarity
Many instances of humour serve to create solidarity within the group or between
particular members of the group. This section of the taxonomy identi
es a
number of primary strategies used to create solidarity and consensus. As there
are countless strategies which one could potentially use, there are some examples
which do not
t into the main categories identi
ed. Such examples are labelled
99

CHAPTER 8. THE FUNCTIONS OF HUMOUR

100

simply as \other" within this category.

8.3.1 To share
This category covers humour which reveals something about the speaker, or lets
the audience know them better. A lot of anecdotes
t into this category. The
speaker lets the audience know them better, and so positively a ects solidarity.
Sharing sensitive information also indicates a speaker's trust of the audience,
and so can enhance solidarity. In (31) RF shares a memory from her childhood.
(31)
SF:
RF:

/they were\\ great

i LIKED my poncho + except it had little holes
about the size of my fingers so i'd go to reach
for something voc: xunk h]] right through
poncho and and be stopped you know +

SF:

ha h ha]

LM:

oh deah]r

RF:

but other than that ha] it was warm and you

could wear it over anything

8.3.2 To highlight similarities or capitalise on shared experiences

This was originally two categories, but they often overlap, and the boundary
between them is not always clear. As the strategies are similar, I decided to
make them just one category. Ziv (1984) de
nes one of the functions of humour
as \sharing similarities between self and others". This category is for humour
which identi
es or celebrates shared ideas, shared interests and other similarities
between speakers. Also in this category I include references to and reminiscences
about shared experiences. Example (32) includes humour which refers to shared
experiences, and humour which highlights similarities. The group is reminiscing
100

8.3. SOLIDARITY

101

about a course they did which was particularly tough, and during which they
regularly stayed up all night in the university computer labs. MM recalls that
he drank an incredible amount of coke on one particular night before they all
had an assignment due, which has put him o coke ever since. He is capitalising
on shared experiences. CM's comment, at the end of the extract is a reference
to a Monty Python skit. The group has been sitting round exchanging horror
stories about the course, and CM's quip implies that they sound like the men in
this skit, who try to outdo each other with hard luck stories. They all recognise
the reference, indicating they share an appreciation of this particular skit, and
highlighting another similarity between the members of the group.
(32)
CM:

/yeah\\ that's it's a it's it's an

//experience\

MM:

/something\\ you'd want to do once +

CM:

just cause it's quite //quite\

TM:

/ruin\\ your body by

ingesting all that coke=/

MM:

/=mm i still can't drink

coke like i used after that //episode\ i think=

TM:
MM:

/h ha]\\
=i drank about eight cans of coke and four
cookie time biscuits all on one night and i
didn't feel quite rh]ight ever since=/

CM:

/=god
that's NOTHING when I did three oh nine

All:

laugh] 1 sec

101

CHAPTER 8. THE FUNCTIONS OF HUMOUR

102

8.3.3 To clarify and maintain boundaries

Humour can be used to reinforce norms and values and make explicit the boundaries of acceptability. It can also clarify who belongs in di erent groups. Making
fun of outsiders serves a boundary function. If the humour reinforces readily
accepted and agreed upon standards, then it will increase solidarity. Note that
there is a similar function within the power set of functions. Boundary humour
can also be used to impose boundaries, or to clarify boundaries by ridiculing a
member of the group who has unwittingly over-stepped the boundaries of acceptability. For convenience, I have used the label boundS for boundary humour
which maintains solidarity, and boundP for boundary humour which increases
or reinforces the speaker's power. An example of boundP humour is given in
section 8.4.3. BoundS humour is humour which attempts to clarify boundaries,
or supports boundaries which are already clearly established.
Example (33) is an example of boundary humour. An outsider is made fun of.
All agree that this person is not a member of their group, and that she has some
undesirable characteristics, and so the humour reinforces solidarity.
(33)
NF:

i saw tessa davies in the on the train like

JF:

UGH

SF:

ha ha ha]

JF:

what a grotter

8.3.4 To tease (S)

This is another strategy which can function in two ways. Some teasing reinforces
solidarity, some maintains the power of the teaser. As with boundary humour, I
therefore distinguish between teaseS and teaseP. An example and discussion of
teasing which expresses power is given in section 8.4.4
Teasing can reinforce solidarity if it occurs within what Radcli e-Brown (1952)
102

8.4. POWER

103

terms a \Joking Relationship". Within such a relationship individuals routinely

tease and insult each other. This serves a number of functions, and is primarily
a strategy for expressing solidarity. In (34) a simple, meaningful look is sucient
for JF to realise she is being teased. Evan had a crush on JF, but NF is the
only member of the group that knows this. She has regularly teased JF about
it. When the name comes up in the conversation, NF looks at JF, and JF
immediately reacts with laughter. NF is teasing JF about something which is
secret, and about which they often joke. It therefore clearly serves to maintain
solidarity between them.
(34)
SF:

evan really changed eh

TF:

shit yeah mega

SF:

he went really arty

TF:

he and yeah he went to new plymouth and then he

came back and now oh i don't know where he is
now

JF&NF: laugh]
SF:

what's this going on

JF:

laughs]: nothing + just good humour:

A tease will also reinforce solidarity if it is about something that is clearly false
or trivial.

8.4 Power
Far fewer examples fall into this category. As my data consists of natural friendship groups, it is not surprising that a large percentage of the instances of humour
103

CHAPTER 8. THE FUNCTIONS OF HUMOUR

104

served to create and maintain solidarity. There are at least four strategies which
can be used to maintain or create power: fostering conict, controlling, teasing
and creating boundaries.

8.4.1 To foster conict

As the label suggests, humour in this category introduces or fosters conict within
the group. This is one of three broad functions Martineau (1972) describes in his
model of the social functions of humour. It may purposefully demean a member
of the audience, or transmit an aggressive message. As the groups I am studying
are friendship groups who often come together to chat, it is not surprising that
this strategy is used very little in my data. Of 815 examples, only one uses
this strategy. This is shown as example (35). BM and DF both know BM's
relation who is the focus of the discussion. BM tells how she had surgery, and
DF disagrees with BM's assessment that the surgery was mainly cosmetic. Her
comment is framed in such a way as to foster conict. BM is belittled and the
rest of the audience laugh.
(35)
BM:

/=my cousin's wife um ++ has just

spent thousands and thousands i don't know how
much but it's something in the order of ten
thousand dollars=/

AF:
BM:

/=mm
on having her + teeth straightened up //she=

DF:

/but=

BM:

=had aMAZing-\

DF:

=her whole jaw\\ was receding + eh it was

//(still moving) back\

104

8.4. POWER
BM:

105

/her jaw (was)\\ carry on back=/

AF:

/=mm

DF:

//she couldn't\ eat properly any more //eh\

BM:

/and er\\

/she\\

got to the stage- yeah she couldn't eat

properly
AF:

mm major

BM:

but //it\\ was a co- it was basically a=

DF:
BM:

/yeah\\
=cosmetic thing though //+ i\ mean there=

AF:
BM:

/yeah\\
=wasn't any-=/

DF:

/=//challenging tone]: well=

AF:

/(was it quite)\\

DF:

=it was\ partly 'cause she couldn't eat:

AF:

//but that's what but that's what=

DF:

/laughs] i think she's (

CM:

/laughs]\\

105

)\\

CHAPTER 8. THE FUNCTIONS OF HUMOUR

106
BM:

/laughs]\\

AF:

=orthodontics is\ i mean //you know\=

8.4.2 To control
Any humour intended to inuence the behaviour of the audience would be an example of control. This function is identi
ed by a number of researchers, including
Martineau (1972), Collinson (1988) and Graham, Papa and Brooks (1992).
Most examples of boundP humour would be examples of the speaker trying to
control or restrict the audience's behaviour in some way, but I have not included
those examples in this category.
I found only one example of non-boundary humour which was intended to inuence the behaviour of the audience. This, again, is likely to be a reection of the
type of data I am investigating. One would expect to
nd much more controlling humour in the workplace, or some other hierarchical environment, and it is
not surprising that this seldom occurs in friendship groups. The one example is
shown in example (36). This is from one of the tapes collected by the Wellington
Corpus of Spoken New Zealand English. One person present was collecting the
tape, and keen that the conversation be as natural as possible. Therefore when
the tape is mentioned, she uses humour to indicate that this topic should not be
talked about.
(36)
pour wine]
BM:

the sound of wine pouring always sounds good on

tape

DF:

what tape ha]

BM:

nh ha]

106

8.4. POWER

107

8.4.3 To challenge and set boundaries

This category has already been discussed to some extent in section 8.3.3. The
examples in this category challenge existing boundaries, attempt to set new ones,
or create or maintain boundaries by making an example of someone present.
Example (37) is an extract from a conversation between geology students. SF's
comment in (37) clearly draws a boundary between acceptable and unacceptable applications of geology. This is not a readily established boundary within
the group, and BF has already expressed an interest in the type of geology SF
condemns.
(37)
BF:

i like petroleum geology i think it's cool=/

AF:

/=do
you=/

BF:
AF:
BF:

/=mm //that's what i'd like to do\ if i do=

/(

)\\

=anything in geology=/

SF:

/=far out=/

AF:

/=it's where the

money is=/

BF:

/=i'm just a (sucker for it)=/

AF:

/=really=/

CF:
/=yeah=/

107

CHAPTER 8. THE FUNCTIONS OF HUMOUR

108
SF:

/=mhm i'm not interested in money i more

interested in the research side i could never
do //coal and that\

BF:

/oh i want to\\ make big bikkies=/

AF:

/=ha ha]

CF:

//like me i just want to marry\ ha]=

SF:

/raping and pillaging the land ha ha]\\

8.4.4 To tease (P)

Teases which attack personal details, or seem to make genuine criticisms serve
to increase or maintain the speaker's power. This will often overlap with the
boundary category. In (38) LM teases DT for forgetting to set the video for
a programme they both wanted to watch. LM was annoyed about it, and has
teased DT quite a bit about it since it happened.
(38)
LM:

i'd love to see the john cleese one as well

DT:

yeah yeah //yes i set the video wrong give me=

LM:
DT:

/nh nh nh nh ha\\
=a break\

8.5 Psychological Functions

Defending and coping humour serve psychological functions. These categories
are discussed in the following sections.
108

8.5. PSYCHOLOGICAL FUNCTIONS

109

8.5.1 To defend
This label is used by Ervin-Tripp and Lampert (1992) for Ziv's (1984) function:
\protecting the self by identifying a weakness before anyone else does." I use the
label to apply to any humour which is used to protect oneself. One example may
be to avoid revealing personal information about oneself. In (39) WF is insecure
about the trie she has made, and makes excuses to protect herself from any
criticism that may be forthcoming.
(39)
PM:

it was a nice trifle

WF:

normally //yeah well it was nice\

TF:
WF:

/what's this\\ WAS shit

laughs] but it sort of the problem is i just
kept adding more and more and i only had
certain size bowls so i couldn't balance it all
up and i didn't have enough ingredients to just
keep adding //ha huh]\

8.5.2 To cope with a situational problem

The coping function is identi
ed by a number of researchers, including Pogrebin
and Poole (1988), Fink and Walker (1977) and Ziv (1984). I have divided humour
which is used as a coping device into two further categories { coping with a
situational problem and coping with a general problem. The
rst, coping with a
situational problem, includes any humour which is used to cope with a problem
arising in the course of the conversation. The problem could range from a social
ga e of some sort to a pot boiling over, the main point is that it arises, and must
be coped with during the conversation. (40) is a typical example of this type
of humour, in which the speakers use humour to cope with the presence of the
tape-recorder.
109

CHAPTER 8. THE FUNCTIONS OF HUMOUR

110
(40)
CM:

i thought she was going to rescue us at four

YOU'RE LATE

All:

laugh]

CM:

yells]: come on:

MM:

we should start with the credits now yeah

NM:

this conversation HAS featured=/

TM:

/=ha ha ha]

The issue of humour relating to the tape-recorder is dealt with in detail in section 4.5.

8.5.3 To cope with a general problem

This category covers humour used to cope with more general problems such as
sickness or death. Joking about something gory, scary or depressing will often
be an example of coping humour. One way of capturing the essential di erence
between situational and general coping, is that the
rst copes with problems
we need to get through to survive the conversation, and the second copes with
problems we need to get through to survive life, or a period of our life. In (41)
humour is used to cope with a problem TF has. She has employed tradespeople
who made an initial appearance, but have recently disappeared into thin air.
(41)
TF:

they obviously thought that that i looked like

the type that wouldn't make them //ring\ up i=

WF:

/oh\\

110

8.6. RESULTS

111

TF:

=# r # d i iMAgine

LF:

should ring the i # r # d up if they don't come

back

TF:

h]if they dh]on't come back h]we just

woh]n't PAY //them anything\

8.6 Results
I
tted log-linear models,
rstly to investigate the distribution of the four main
functions of humour, and then to investigate more closely behaviour with regard
to all of the strategies and the speci
c psychological functions. I outline the
overall distribution of the di erent functions in section 8.7 and then move on to
describe the model which
ts each of the strategies separately.
The function group into which most instances of humour fell was the creation
and maintenance of solidarity. This is hardly surprising considering my groups
were speci
cally chosen because they were natural friendship groups. One might
expect more power based humour to occur in work-related groups, or humour
occurring amongst strangers.

8.7 Overall results

The
tted model for the distribution of general, solidarity, power and psychological functions showed main e ects for both the sex of the speaker and group
composition. The degrees of freedom, chi-square and probability
gures for the
various options are shown in table 8.1.
Remember that group composition and speaker sex are explanatory variables
and so the 12 interaction is kept in the model at all times. Model 1 shows main
e ects for both speaker sex and group composition. Model two says that sex
a ects the function but the function is independent of group composition given
sex. The third model is the same as the second with group composition and sex
reversed, and the fourth model is saying that function is independent both of
111

CHAPTER 8. THE FUNCTIONS OF HUMOUR

112

1=group composition, 2=sex, 3=function

Model no.

Model

1
2
3
4

12 13 23
12 23
12 13
12

DF

chi-sq

prob.

3
6
6
9

7.93
34.77
48.45
75.27

0.0475
0.0000
0.0000
0.0000

Table 8.1: Results of possible models for overall function

speaker sex and group composition. Model one is clearly the best model { it has
the lowest chi-square and the highest probability of the four. We can clarify that
it is signi
cantly better than the other options by comparing it with the closest
model { number two. We can subtract 7.93 from 34.77 and check the signi
cance
level for this to three degrees of freedom (we reach three by subtracting the DF
for model one from the DF for model 2). This calculation shows that model one
is signi
cantly better than the other models { to the .001 level.
The probabilities predicted by this model are shown in table 8.2. These probabilities are calculated as conditional probabilities. So, for example, the
gure
for the general function, is conditional on the sex of the speaker and the group
composition. This is done to allow for di erences in the cell sizes for each of the
groups. The probabilities for each of the four groups should therefore add to
one, although there is some slight variation due to rounding.
General
Solid
Power
Psych

FC
0.25
0.59
0.10
0.06

MC
0.44
0.36
0.13
0.07

FS
0.13
0.65
0.12
0.11

MS
0.26
0.45
0.17
0.13

Table 8.2: Probability
gures for overall function

We can spot some general trends from investigating the probabilities, but calculation of the odds ratios reveals the exact nature of these trends. For each
112

8.7. OVERALL RESULTS

113

function I have calculated four odds ratios. These show the relation between the
behaviour of females and males in both single sex and mixed sex conversations,
and the di erence between single sex and mixed sex conversations for both females and males. The odds ratios for the four overall functions are shown in
table 8.3
General
Solid
Power
Psych

FC-MC MC-FC
2.36
2.56
1.34
1.18

FC-FS FS-FC
2.23
1.29
1.23
1.94

MC-MS MS-MC
2.24
1.45
1.37
1.99

FS-MS MS-FS
2.35
2.27
1.50
1.21

Table 8.3: Odds ratios for overall function

The MC-FC and MS-FS ratios are virtually equal for the general function, as are
the ratios for FC-FS and MC-MS. We can therefore collapse the table and say
the odds of men using only the general function are approximately 2.35 times
higher than for women, and that the odds of only the general function being used
are roughly 2.23 times higher in mixed conversations than in single sex ones.
The odds of females using humour for a solidarity function are 2.56 times higher
than males when in a mixed group, and 2.27 times higher in single sex groups.
Solidarity maintaining humour is also slightly more likely to be used in single
sex than in mixed sex groups.
The odds ratios for the power function do not provide such conclusive evidence.
The ratios show a number of slight trends. Males tend to use power based
strategies more often than females, and these strategies occur slightly more in
single sex than in mixed sex groups.
Males seem more likely to use humour for psychological functions, though we
should not put too much weight on this as the di erence is only slight. More
conclusive is the fact that humour is more likely to be used for this function in
single sex conversations than mixed conversations.

113

114

CHAPTER 8. THE FUNCTIONS OF HUMOUR

8.8 Results of individual strategies

I adopted the same model for looking at the distribution of the individual strategies as I did for the overall functions. When applied to the individual strategies,
the model of main e ects for speaker sex and group composition is signi
cantly
better than any other variations on the model, although the model is marginally
o reaching signi
cance itself. This is primarily because the expanded contingency table is much sparser than the table in which all the functions are collapsed
into four overall categories. As the collapsed tables still held the same data, it
is safe to assume that the correct model for the individual strategies is the same
as for the grouped categories. In the following sections I outline the results for
each of the individual strategies.

8.8.1 Results of solidarity based strategies

Table 8.4 shows the probability
gures for each of the strategies used to create or
maintain solidarity. Clarifying and maintaining boundaries is the strategy most
frequently used to ful
ll this function.
FC MC FS MS
share
0.09 0.01 0.15 0.02
highlight 0.08 0.12 0.09 0.16
boundS 0.31 0.19 0.23 0.15
teaseS
0.06 0.05 0.11 0.12
other
0.06 0.01 0.07 0.01
Table 8.4: Probabilities for solidarity based strategies
In table 8.5 I present the odds ratios for solidarity based strategies.
The results for sharing humour are conclusive. There is no doubt that women
are much more likely to use humour in this way than men. The model predicts
the odds of females using sharing humour to be 9.79 times higher than the
odds of males using sharing humour in mixed conversations, and in single sex
conversations the odds ratio for females to males is 8.65:1. The
gures also show
both sexes more likely to use sharing humour in single sex than in mixed sex
114

8.8. RESULTS OF INDIVIDUAL STRATEGIES

share
highlight
boundS
teaseS
other

FC-MC MC-FC
9.79
1.57
1.92
1.21
6.32
-

FC-FS FS-FC
1.78
1.14
1.50
1.94
1.18

MC-MS MS-MC
2.02
1.40
1.33
2.59
1.00
1.00

115
FS-MS MS-FS
8.65
1.93
1.69
1.10
7.45
-

Table 8.5: Odds ratios for solidarity based strategies

interaction.
Highlighting similarities and capitalising on shared experiences are strategies
more likely to be used by men than women. There is also a very slight indication that they are more likely to be used in single sex rather than mixed sex
conversations.
None of the odds ratios for boundS reach two, and so the patterns shown can
only be regarded as possible trends. The
gures show that maintenance and
clari
cation of boundaries is more likely to be used by women than men, and in
mixed conversations rather than single sex conversations.
Teasing in a manner that enforces solidarity is more likely to occur in single sex
conversations than mixed sex conversations. This is particularly true for male
speakers.
The odds ratios show women 6-7 times more likely to use an alternative solidarity building strategy than men. As the particular nature of these strategies is
varied, this category is not so interesting by itself, but plays an important part
in contributing to the overall consideration of the solidarity function as discussed
in section 8.7

8.8.2 Results of power based strategies

Table 8.6 shows the strategies that serve to maintain or create power, and the
probabilities that a speaker in each of the groups use these strategies.
When calculating the log-linear models, examples which employed more than
one strategy were coded for both. I then checked the models using a calculation
discussed in section 5.3 which adjusts for resulting clustering e ects. In most
115

CHAPTER 8. THE FUNCTIONS OF HUMOUR

116

FC
boundP 0.07
teaseP 0.07
other
0.02

MC FS MS
0.06 0.09 0.06
0.06 0.12 0.11
0.01 0.00 0.00

Table 8.6: Probabilities for power based strategies

cases this did not a ect things signi
cantly, as only a very small percentage of
examples were coded for more than one strategy. A notable exception to this,
however, was the co-occurrence of power based boundary marking and teasing.
These two are somewhat related, and sucient examples were coded as both of
these to a ect the model. To bypass this problem I created a separate category
for those examples coded as both tease-P and boundary-P. To illustrate the
trends in the data clearly, I have re-separated these categories for the following
discussion. The probability
gure predicted by the model for examples which
both powerfully tease and create boundaries have been added to the boundary-P
category, and also to the tease-P category so that we can independently consider
the probability of each of these occurring. This means that the probability
gures
presented for all of the strategies add to a little over one in each group. This is
because part of the probability
gures for boundP and teaseP overlap.
The odds ratios for each of the strategies that are used to create or maintain
power are given in table 8.7. No odds ratios are given for the categories of foster
conict or control. There was only one example in each of these categories and
so nothing meaningful can be said about the distribution. The most interesting
result here is the fact that so few examples fell into these categories.
boundP
teaseP

FC-MC MC-FC
1.18
1.18
-

FC-FS FS-FC
1.31
1.81

MC-MS MS-MC
1.00
1.00
1.94

FS-MS MS-FS
1.55
1.10
-

Table 8.7: Odds ratios for power based strategies

The ratios for creating or challenging boundaries are not particularly striking.
The women seem to use humour in this way slightly more than the men, both
in single sex and mixed interaction. Women use it slightly more in single sex
groups than mixed groups, but group composition does not a ect the extent to
116

8.8. RESULTS OF INDIVIDUAL STRATEGIES

117

which men use humour to create or challenge boundaries.

The ratios show that there is a tendency for powerful teases to occur in single
sex conversations more often than mixed sex conversations. The model shows
women slightly more likely to use humour in this way than men, though this

gure is far from conclusive.

8.8.3 Results of psychological functions

The probabilities that a speaker will use humour to cope or defend are shown in
table 8.8.
FC MC FS
defend 0.02 0.02 0.03
cope-gen 0.02 0.01 0.04
cope-sit 0.02 0.04 0.03

MS
0.03
0.03
0.07

Table 8.8: Probabilities for psychological functions

In table 8.9 I present the odds ratios for these functions.
defend
cope-gen
cope-sit

FC-MC MC-FC
1.00
1.00
2.02
2.04

FC-FS FS-FC
1.52
2.04
1.52

MC-MS MS-MC
1.52
3.06
1.81

FS-MS MS-FS
1.00
1.00
1.35
2.43

Table 8.9: Odds ratios for psychological functions

The
gures show no notable di erence in the use of defending humour between
men and women or mixed and single sex conversations. The one trend indicated
by the
gures is a slight tendency for defending humour to occur more in single
than in mixed sex conversations.
The probabilities given for humour which copes with a general problem are fairly
small, and so the resulting odds ratios may be disproportionately high. They
do indicate some clear trends however. This type of humour is more likely to
occur in single sex conversations than in mixed sex conversations, and females
are more likely than males to use humour in this way.
117

118

CHAPTER 8. THE FUNCTIONS OF HUMOUR

The strongest trend involving situational coping is that the odds of males using
situational coping humour are more than twice as high as for women, across both
group compositions. This is largely due to the fact that the all male groups are
less comfortable with the context as discussed in chapter 4. There also seems to
be a slight trend for situational coping humour to occur more in single sex than
in mixed sex groups.

8.9 Discussion
That I chose to investigate natural friendship groups has no doubt inuenced
the types of functions for which humour is used in my data. There are very
few examples of the particularly negative categories { to control, and to create
conict. These functions have been identi
ed as common functions for humour
in the literature (Martineau 1972, Collinson 1988), but it seems very plausible
that the informal friendly context is the reason there are few instances in my
sample.

8.9.1 The General Function

Achieving the general function described in 8.2 is more often the only purpose
of the humour produced by men. Women are more likely to also use the humour
for some further function, such as to create solidarity through sharing.
That the general function is used signi
cantly more by men than by women
is probably related to the fact that appearing witty is more central to a male
personal identity than a female identity (see chapter 11). A short quip or oneliner, then, performs positive work on a male personal identity. In general, this
ability seems not so important for women, and so when they use humour, it tends
to be performing a further identi
able function beyond the general function of
increasing solidarity and power, and positively a ecting personal identity.
It could be argued that, because I am a female I was unable to interpret much of
the male humour, and so more male humour than female humour ended up in this
general category, which could be seen in some ways as an \other" category. This
is a possibility, but I tried to safeguard against this by checking my intuitions
118

8.9. DISCUSSION

119

with several male \interpreters".

That males more often use humour for the sole purpose of impressing, appearing funny, or creating a positive personal identity is in part supported by past

ndings.
Ervin-Tripp and Lampert (1992) observed that men's humour consists largely
of ip wisecracks and that this tendency further increased in mixed sex groups.
Most short wisecracks in my sample fell into this general function. They are
typical of a type of humour designed to elevate status and solidarity within the
group and work on personal identity, without performing any further function.
In my data, too, this type of humour increased in mixed sex groups.

8.9.2 Solidarity
The solidarity ratio shows exactly what one would expect, given the literature.
The odds of women using humour to create or maintain solidarity are more than
twice as high as men.
The results for sharing humour show conclusively that women use this strategy
much more than men do. This supports Ervin-Tripp and Lampert's (1992) claim
that women use humour to share. Johnson and Aries (1983) is one of several
studies which found women much more likely to self-disclose than men. Cozby
(1973) provides a comprehensive review of early work in this area. Some of the
studies outlined in this review show that women self-disclose more than men,
while some found no di erence. No studies, however, found men more likely to
self-disclose than women.
Komarovsky (1962) found the men in her study unlikely to share or disclose
personal information.
The phrase \incapacity to share" aims to convey a certain view about
the men's articulateness. The ideal of masculinity into which they
were socialised inhibits expressiveness both directly, with its emphasis
on reserve, and indirectly, by identifying personal interchange with
the feminine role. (Komarovsky 1962:151)
My results show males more likely than females to use humour to highlight
119

120

CHAPTER 8. THE FUNCTIONS OF HUMOUR

similarities or capitalise on shared experiences.

Many of these examples consist of joint reminiscences about previously shared
activities, or of reminding each other of jokes from movies, or quoting lines from
favourite books. In this way speakers express similar tastes, and capitalise on
time they have shared together. So while a group of women may share new,
personal information about themselves in order to maintain solidarity with their
friendship group, men may be more likely to relive old times to achieve the same
end. Johnson and Aries (1983) found that men reminisced more than women
and talked about hobbies and shared activities.
This could relate to the interaction based/task based distinction noted by Parsons and Bales (1955) and discussed in chapter 4. Interaction seems crucial
to female friendship groups, whereas activities are more important in all male
groups (Goodwin 1982a). Douvan and Adelson (1966) point to the fact that females are socialised to place value on interpersonal relationships, whereas males
are taught to value activities and achievement. It therefore makes sense that
while women create solidarity by sharing intimate information, men remind each
other of shared activities. Brotherton and Penman (1977) noted that women
tended to talk about relationships, whereas men would talk more about tasks.
This is consistent with the
nding that men create solidarity by reminiscing more
about previous shared activities than women do.
Females are slightly more likely than males to use humour to maintain or clarify
boundaries, particularly in mixed groups. This strategy serves to increase group
solidarity, often by identifying an outgroup.
Solidarity based teases are used more often in single sex than in mixed sex groups.
This is particularly true of men. The odds of a male using a solidarity based
tease are 2.59 times greater in single sex than in mixed sex interaction.
Most literature on teasing is restricted to teasing amongst males (see Radcli eBrown 1952, Loudon 1970, Kuiper 1991). Some studies show teasing in mixed
groups, but this is of a particularly sexual nature (Spradley and Mann 1975,
Whitehead 1976, Parkin 1979). Only recently have researchers found that teasing
also occurs in all female groups (Eder 1990, Eder 1993).
Limbrick (1991) looked at use of expletives by New Zealand men and women.
Both groups used roughly the same amount of expletives when in single sex
120

8.9. DISCUSSION

121

groups. In mixed groups, however, the males decreased their use of expletives
by a substantial amount. Limbrick interpreted this as a desire not to o end and
accommodation to the stereotype of females' lesser expletive usage.
Folb (1980) found that the Black American girls she studied used the vernacular
and swear words only when out of earshot of males and of adults.
When I was privy to all female conversation, I found that the quantity
of talk, joking, boasting, argument, cursing and even shooting the
dozens rivalled male expressive behaviour (Folb 1980:195)
Much of the teasing humour in my corpus takes the form of jocular insults. Folb
points out that such behaviour is not \lady-like" and so regarded as inappropriate
behaviour to display to boys, or to adults. Similarly, the boys in her study toned
down their vernacular usage among young women as to do otherwise would be
disrespectful.
So both men and women engage in jocular abuse and teasing activities, though
they do this much more often in single sex groups than mixed groups. It seems
likely that the reasons this behaviour is restricted in mixed conversation di er for
both men and women, but both reect the gender stereotyping and expectation
of \appropriate" gender-speci
c behaviour.
The results of this research show that teasing is an activity indulged in equally by
men and women. It tends to occur much more in single sex interaction, perhaps
also because these are the groups in which the most intimate friendships are
made. Teasing and sparring can develop a sense of \comrade-ship" and joviality
within groups. Because this tends to occur most in single sex groups, many
seem to believe that it is restricted to just one gender. My topic sparked much
debate amongst my friends, and many gave me \insights" into what they believed
were the essentials about humour. A male friend explained to me that his best
friends were those he insulted regularly and with great vigour. His wife did not
understand this, and he most certainly could not insult her in the same manner!
He wondered why it was that men could enjoy this type of humour and solidarity,
whereas women did not seem to be able to. The answer seems to be that women
do, in fact, use humour to tease to the same extent as men. But both men and
women do this most in single sex groups rather than mixed sex groups.
121

CHAPTER 8. THE FUNCTIONS OF HUMOUR

122

It is important to note that this is a speci
c type of humour which is usually

restricted to certain friendship groups. One either has a \joking relationship" established, or one does not. Joking relationships were
rst described by Radcli eBrown:
What is meant by the term joking relationship is a relation between
two persons in which one is by custom permitted, and in some cases
required to tease or make fun of the other, who in return is required
to take no o ence (Radcli e-Brown 1952:90).
If such a relationship exists, teasing humour will be prevalent, if it does not, there
will most likely be no humour that serves this purpose. The examples of teasing
in this corpus, then, are not equally distributed amongst all the conversations.

8.9.3 Power
Much of the language and gender literature presents power and solidarity as dichotomous. Men, in general, are said to place importance on power and competition within conversation, and females prioritise the expression and maintenance
of solidarity (see Aries 1976, Edelsky 1981, Fishman 1983, Maltz and Borker
1983, Coates 1986, Preisler 1986, among others). Context is also important
here, with men's style used more often in public settings, and women's in private
interaction.
It is surprising then, that the di erence between men and women in the use
of power based functions is not huge. Men do use humour for power based
functions slightly more than women, but the
gures are far from conclusive. It
is particularly interesting that the odds ratio for male-female single sex groups
is greater than for male-female mixed groups. Given that men are said to exert
their power more often { especially over women, we would have expected quite
a large ratio here. Perhaps the distinction is so small because the groups are all
natural friendship groups. Status and power di erentials are minimal.
There are no striking di erences between men and women or single and mixed
conversations in the use of humour which creates or challenges boundaries. This
too is contrary to expectations, and is probably a reection of the fact that the
conversations occur amongst friends.
122

8.10. CONCLUSION

123

Powerful teases occur more in single sex conversations than mixed sex conversations. This is interesting as the literature indicates that these are more likely
to be used by men than by women. Both groups tend to \behave" more when
in the presence of the opposite sex. In section 8.9.2 I discussed solidarity based
teases. The distribution of the two types of tease are very similar.

8.9.4 Psychological Functions

Humour was more likely to be used for a psychological reason in single sex conversations than mixed conversations. A contributing factor will no doubt be the
large amount of situational coping used by males in single sex conversations as
discussed in chapter 4.
Coping humour tends to occur in single sex groups { females are more likely
than males to use humour to deal with a general problem, whereas men are more
likely to use it to cope with a situational problem. I discuss in the next section
the fact that studies have shown women more likely to self-disclose. We could
expect women to be more likely to discuss a general problem facing them at any
given time. It is therefore not surprising that they more often employ humour
to cope with such problems.
Ervin-Tripp and Lampert (1992) found women more likely to use humour as a
coping strategy, although the de
nition I have used of coping has much wider
scope than theirs, as they are interested speci
cally in self-directed humour.
Ervin-Tripp and Lampert also found men's humour more likely to be used to
defend than women's. Again, I have expanded the de
nition for my purposes,
which makes comparison dicult. They include only humour which protects the
self by identifying a weakness before anyone else does, whereas I include in this
category any humour which serves to protect the speaker. My results show no
signi
cant di erences in the use of defending humour.

8.10 Conclusion
The results of this study show that, as expected, women are more likely to use
humour to form or maintain solidarity than are men. In particular, women use
123

124

CHAPTER 8. THE FUNCTIONS OF HUMOUR

humour to share personal information about themselves. When using solidarity

based humour, men are more likely than women to capitalise on shared experiences or highlight similarities. Men are more likely to use humour solely for the
general function of increasing solidarity and status and performing positive work
on their personal identity. The fact that my data came from natural friendship
groups meant that particularly power based strategies such as controlling and fostering conict seldom occurred. The speakers in my corpus also used humour for
psychological purposes. Men more often used humour to cope with a situational
problem, whereas women were more likely to use general coping strategies which
helped cope with problems that were not speci
c to the immediate situation.
Despite literature to the contrary, this research has shown that women do, in
fact, indulge in teasing humour. Both men and women tease both in a powerful
manner, and to create solidarity, though this strategy is largely restricted to
single sex groups.
In the next chapter I investigate the topic and focus of the humour in my corpus.
Chapter 10 then discusses the relative humour contribution by men and women,
chapter 11 investgates aspects of personal identity formation via humour and

nally in chapter 12 I investigate humour support strategies and possible reasons
for failed humour.

124

Chapter 9

Focus and Topic

9.1 Introduction
In this chapter I discuss the topic of the humour, and on whom it is focussed.
The
rst part of the chapter concentrates on the focus. Topic is discussed in
section 9.3.

9.2 Focus
The focus of an instance of humour is the character or characters central to the
humour. The focus is generally the butt of the humour. Some types of humour
are untargeted. Wordplay, for example, is a type of humour that often has no
focus.
Mitchell (1985) divided the main characters in her collection of canned jokes into

fteen categories: Children, Fools, Animals, Women, Religious Figures, Political
Figures, Polacks, Negroes, Members of Ethnic Groups (excluding Polacks and
Negroes), Authority Figures, Deformed characters, Military Personnel, Female
Professionals, Male Professionals and Other. The Polack and Negro categories
are most likely to be replaced by Irish jokes in New Zealand English. I originally
intended to classify any jokes occurring in my data within this taxonomy in
order to provide comparison with Mitchell's data. As only one joke occurred in
125

CHAPTER 9. FOCUS AND TOPIC

126

my data, this was not necessary. This joke would have fallen into the \other"
category.
Ervin-Tripp and Lampert looked at focus in their corpus of spoken humour and
(1992) discuss two categories in which they found signi
cant gender di erences,
namely, the targeting of self or of absent people. Both men and women used
more outgroup put-downs when in mixed sex groups, and women also decreased
their use of self directed humour when in mixed groups.
In section 9.2.1 I discuss the classi
cation of my examples into focus categories
and outline the results.

9.2.1 Classication and Results

I classi
ed all of my examples into one of
ve categories.
1. Absent: Humour focussed on a person known to the speaker but absent
from the group.
2. Self: Self-directed humour.
3. Present(M): Humour focussed on a male member of the group.
4. Present(F): Humour focussed on a female member of the group.
5. Other: Other target, or untargeted.
I
tted the results to a log-linear model which showed main e ects for both
speaker sex and group composition. The model has a chi-square of 2.79, with 3
degrees of freedom and a probability of 0.4255. It is signi
cant, and signi
cantly
better than any possible simpler model. The probabilities predicted by the model
are shown in table 9.1. The asterisks show structural zeros in the model. It is
impossible to have humour directed at a male who is present in an all female
group, or for males in a single sex group to direct humour at a female member
of that group.
The odds ratios give a clear indication of the most signi
cant results. They are
shown in table 9.2.
126

9.2. FOCUS

127
FC
abs
0.25
self
0.16
present(m) 0.15
present(f) 0.03
other
0.40

MC
0.24
0.15
0.03
0.10
0.48

FS
0.18
0.20
*
0.30
0.32

MS
0.15
0.17
0.30
*
0.38

Table 9.1: Probability
gures for focus of humour

abs
self
pres(m)
pres(f)
other

FC-MC MC-FC
1.06
1.08
5.71
3.59
1.38

FC-FS FS-FC
1.52
1.31
*
- 13.86
1.42
-

MC-MS MS-MC
1.79
1.16
*
13.86
*
*
1.51
-

FS-MS MS-FS
1.24
1.22
*
*
*
*
1.30

Table 9.2: Odds ratios for focus of humour

The asterisks indicate positions for which it was impossible to calculate an odds
ratio because of structural zeros in the contingency table. We can, however,
calculate the ratio of the odds that someone present will be the focus in a single
sex male group and in a single sex female group. The probability in each case is
0.30, and so the odds ratio is one.
The most signi
cant and the most interesting results here concern humour focussed on someone present. The odds of someone using humour focussed on a
same sex group member are overwhelmingly higher in single sex groups than
mixed groups. Humour focussed on same sex group members drops right back in
mixed conversation, and instead, when humour is focussed on a group member,
they tend to be of the opposite sex to the speaker. The odds of a female in a
mixed conversation using a male group member as the focus of their humour are
5.71 times higher than a male in a mixed sex group focussing on a male member.
Similarly, the odds of a male in a mixed group using humour focussed on a female
member are 3.59 times higher than the odds of a female focussing on someone of
her own sex in a mixed conversation.
None of the other results show signi
cantly high odds ratios. Women are slightly
more likely than men to use humour focussed on someone who is absent, and
127

CHAPTER 9. FOCUS AND TOPIC

128

both men and women increase humour focussed on someone absent when in
mixed groups. Women use themselves as the focus of humour slightly more than
do men, and both men and women use humour in this way in single sex groups
slightly more than in mixed groups.

9.2.2 Discussion
These results do provide some support for Ervin-Tripp and Lampert's (1992)

nding that self-directed humour occurs more in single sex groups and is used
more by females. This result is not signi
cant in my data, however, and shows
only as a possible trend. Similarly their observation that humour focussing on
someone absent was more proli
c in mixed groups is supported by only a slight
trend in my data. It is important to note, however, that Ervin-Tripp and Lampert's examples included only humour which put down an outgroup, whereas my
category includes all humour focussing in some way on someone absent. These
results are therefore not completely comparable.
That speakers tend to use members of the opposite sex as the focus of their
humour, and avoid focussing humour on their own sex is very interesting. In
Hay (1994) I analysed jocular abuse occurring in a natural, mixed friendship
group consisting of eight members. I showed that abuse served, among other
things, to highlight and enforce boundaries within the group. The most salient
group boundary was that of gender. The vast majority of my examples involved
abuse between men and women. The examples showed a friendly animosity
between the two genders, and clearly served to maintain gender divisions.
In her paper \Gossip revisited:- Language in all female groups" Jenny Coates
observes:
Where the main point of relaxed informal conversation among equals
is the maintenance of good (equal) social relationships, one of the
main goals of mixed interaction is inevitably the maintenance of gender divisions of male - female inequality. (Coates 1988:120-21)
The fact that, when in mixed conversations, speakers resist using members of
their own sex as the focus of humour, even though they do so proli
cally in
128

9.3. TOPIC

129

single sex conversations, shows that gender clearly marks a group boundary in
mixed conversations.
In the following section I discuss some sex di erences in the topics of humour
used.

9.3 Topic
There is some evidence that men and women talk about di erent topics, particularly when in single sex groups. Aries (1976) conducted a laboratory investigation
of same sex pairs and found that females tended to talk about themselves, their
feelings, homes and relationships, whereas males were more likely to talk about
sports, amusements, competition and things that they had seen or read. Kiper
(1987) investigated topics discussed by teachers, and found that women teachers
tended to talk about matters relating to the home and family, whereas male
teachers were more likely to discuss work and recreation.
Self report data has found similar results (Haas and Sherman 1982, Aries and
Johnson 1983).
Le Masters (1975:103) studied life in a rural working class tavern, and estimated
that the four most common subjects discussed by men were sex, jobs, people and
sports, not necessarily in that order. Komarovsky (1962) interviewed blue-collar
husbands and wives about aspects of their relationships. When asked whether
there were topics they avoided talking about with their spouse, the women's most
common answer was \myself", and the men's was \the job". Both most often
gave the same reason for the restraint - their partner \wasn't interested". Men
also mentioned sports, cars and politics as topics they wouldn't usually discuss
with their wives because women are \naturally" not interested in such things
(Komarovsky 1962:141).
If conversational topics vary across di erently compositioned groups, then we
would expect that the topics of humour would follow a similar pattern. Very
little research has been done into what men and women tend to joke about.
Mitchell (1985) touched on this theme in her investigation of canned jokes. She
coded the form, character and theme of 1507 jokes. About a third of these were
129

CHAPTER 9. FOCUS AND TOPIC

130

elicited, and the rest occurred spontaneously. As my data consists of mainly

spontaneous, situational humour, the topics joked about will di er somewhat
from the categories in Mitchell (1985). Her discussion does, however, give an
indication of the type of thing men and women are likely to joke about.
The tendency was for both men and women to tell their most openly hostile or
tendentious jokes in single sex audiences, though men told more openly hostile
and aggressive jokes than women. The most obscene jokes, morbid jokes, Negro
jokes, drinking jokes, jokes about deformity and jokes about authority
gures
were told by males in single sex groups. Men's least openly hostile jokes are told
to mixed audiences. They decrease their use of sexual jokes, and jokes about
deformity and authority
gures, and tell higher percentages of absurd jokes and
wordplay jokes to mixed audiences.
Castell and Goldstein (1977) also found that females told aggressive and sexual
jokes only in single sex company, whereas males told them in both single sex and
mixed company.
I was interested in whether these sorts of patterns were reected in my corpus,
and whether distinct patterns could be identi
ed as to the types of things men
and women joke about in mixed and single sex groups.
I identi
ed 25 di erent topic areas into which my examples could be divided. A
few examples covered more than one topic, and these were coded for both. Statistically this is clearly a large number of categories. The spread of examples over
these categories is too thin to make any sort of statistical modelling or detailed
analysis fruitful. To collapse the categories, however, would be to omit a lot of
interesting detail, and so instead I present all the categories, with percentages of
the total number of examples collected from each of the four groups, and pick
out some of the more interesting examples for discussion. The percentages are
shown in table 9.3.
It is important to keep in mind that these percentages are indicative only, and
certainly do not provide conclusive proof of the types of things men and women
joke about. In any given conversation, speakers may latch on to one particular
topic, contributing several instances of humour on this topic to the corpus. Many
of the slight di erences in the
gures given, then, are due to the contribution of a
small number of conversations, rather than over arching trends. To establish for
130

9.3. TOPIC

131

FS
MS
FC
MC
TOPIC
no. % no. % no. % no. %
animal
17 8
0 0 18 17 17 10
child
4 2
0 0
2 2
5 3
comp
0 0 30 9
0 0
0 0
drink
1 0 33 10
3 3
6 4
drugs
0 0
7 2 10 10 10 6
food
27 13 24 7
3 3
1 1
gay
0 0
3 1
2 2
0 0
lang
14 6
0 0
6 6 21 13
money
6 3
0 0
1 1
1 1
music
0 0
2 1
2 2
4 2
opp-sex
29 13 27 8
0 0
0 0
other
34 16 38 11 12 12 17 10
person
16 7 10 3
8 8
9 6
phys.
8 4
4 1
7 7
7 4
prank
3 1
5 2
1 1
4 2
sex
2 1 11 3
2 2
4 2
show
9 4 34 10 15 15 25 15
illness
5 2
5 2
0 0
2 1
sport
2 1
5 2
4 4
4 2
star
11 5
0 0
2 2
6 4
tape
0 0 39 12
1 1
0 0
toilet
21 10
7 2
0 0
1 1
transport 2 1
8 2
0 0
0 0
weather
3 1
8 2
0 0
0 0
wed
0 0
0 0
7 7 12 7
work
5 2 30 9
4 4 17 10
Table 9.3: Percentage of humour on di erent topics

131

CHAPTER 9. FOCUS AND TOPIC

132

certain that men and women joke about di erent topics we would need to examine
each conversation carefully and independently, tracing the life and development
of certain topics within the conversations. This is beyond the scope of this thesis.
Here I intend to give only a rough overview of the types of topics occurring in my
thesis, and hope to highlight some possible di erences which could well bene
t
from in-depth qualitative investigation. In section 9.3.1 I address the topics I feel
would bene
t from discussion. I will comment on all topics for which there is a

ve percent or larger discrepancy between the lowest and the highest percentage,
and also touch on some topics for which we may have expected to
nd a di erence,
but none was shown.

9.3.1 Discussion
Topics preferentially discussed by women
Animals accounted for a surprisingly high percentage of humour in all group
compositions apart from in the single sex male groups, in which no humour based
on animals occurred. In the mixed groups, the women joked about animals more
than men. Much of the humour about animals is based on anecdotes about
family pets. As the literature has found women more likely to talk about home
and family, it is perhaps not surprising that they use humour based on animals
more frequently than do men.
Women also used humour about the opposite sex more than men did. Humour
was coded as being on the topic of the opposite sex if the opposite sex in general
was the topic of the humour, or if a speci
c person was the topic, whose sex was
clearly salient. Not surprisingly this topic only occurred in single sex groups.
This would often serve to maintain group solidarity by joking about an outgroup.
Women joked about the opposite sex almost twice as much as men did. This is
clearly a function of the
nding that women in general have been found to talk
more about their feelings and relationships than men, and women more openly
talk about personal topics.
Women also used more humour in the category I call \people". Examples in this
section revolve around a speci
c person. Either this person is of the same sex as
the speaker, or the sex of the person is not salient or important to the humour.
132

9.3. TOPIC

133

Women joked about people more than men did both in single sex and mixed sex
conversations.
In single sex groups, women joked about famous people much more than men.
Humour in this category made up nine percent of the total instances of humour
in the single sex female groups, whereas there was no such humour in the single
sex male groups. Men did use humour based on famous people when in mixed
groups, as discussed in section 9.3.1.

Topics preferentially discussed by men

Nine percent of the humour in single sex male groups revolved around computers. This contrasts sharply with the zero percent both in mixed groups and single
sex female groups. This can perhaps be partly explained by a slight bias in my
sample. Of the six all male groups, three included at least one computer scientist. Nevertheless, even in male groups with no computer scientists, some of the
humour revolved around computers or computer games. Talk about computers
can in most cases be related to talk about work, or about a speci
c (computerrelated) activity. As these are areas which have been identi
ed in the literature
as more often talked about by men, the result is less surprising than it
rst seems.
It is also interesting that computer talk does not occur in mixed conversations.
This indicates that computers may be one area of \restraint". Men may resist
talking about computers in the company of women, believing they will not be
interested. In many cases they are probably right! If Komarovsky's (1962) study
were repeated today, and amongst middle class subjects, I suspect computers
would be one of the topics listed by men as something they avoid talking about
with their partners.
Men also joke about work roughly seven percent more than women both in single
sex groups and mixed sex groups. This supports Kiper's (1987)
nding that men
were more likely to discuss work matters than women.
Almost ten percent of the humour in single sex male groups relates to alcohol.
This is in contrast with less than one percent in single sex female groups, and
3-4 percent of humour in mixed groups. This no doubt relates to the \mateship
culture" discussed in Pilkington (1992:58). Among certain groups of males it is
particularly cool to be a frequent and excessive user of alcohol. This is discussed
133

CHAPTER 9. FOCUS AND TOPIC

134

in 11.5. Phillips (1987) explains the history behind the kiwi male stereotype,
which he describes as caught between the Protestant middle class, and the tradition of physical strength, mateship, drinking and gambling. He describes the
pioneers in New Zealand, claiming:
Drinking was without doubt the most important and de
ning ritual
of the male community. (Phillips 1987:35)
Men used humour on the topic of a movie, television show or book more than
women, regardless of the group composition. This is consistent with Aries (1976)
observation that men tend to talk about things they have seen or read.
Finally, about 12 percent of all male humour revolved around the tape or the
taping process. One percent of the women's humour in mixed groups was on
this topic, and there were no instances from males in mixed groups, or from all
female groups. This related to the methodology of this study, and so is discussed
in section 4.5.

Mixed groups { accommodation and \safe" topics

Both men and women increased their use of humour on television shows, movies
or books when in mixed sex groups. Perhaps because this is a relatively \safe"
topic, and often relates to experiences the whole group has shared or can relate
to. Women's use of humour on this topic increases by over ten percent in mixed
sex interaction.
Language is another relatively safe topic. It is the topic of about six percent of
women's humour { regardless of the composition. In male single sex groups, none
of the humour was on this topic, but in mixed groups, it accounted for almost 13
percent of the men's humour. This may be related to the safeness of the topic,
and also be accommodation on the part of the men.
Men joked about people slightly more in mixed sex than single sex conversations.
This is possibly also the e ect of accommodation.
One would expect the distribution of humour relating to drugs to be similar to
that relating to alcohol. Interestingly this is not the case. Drugs are not joked
about by females in single sex conversations, but in mixed conversations women
134

9.3. TOPIC

135

joke about drugs almost 10 percent of the time { roughly four percent more than
men in mixed groups. Men, too, joke about drugs more in mixed conversations
than single sex conversations. Perhaps this is because the use of drugs is less
frequent than the use of alcohol, and so more something to \boast" about in
groups in which this is regarded as cool. This is not touched on at all in all
female conversations, perhaps because appearing \cool" is not an important, nor
even desirable trait in many all female groups. If, in mixed sex conversations,
the groups generally adopt the values of the males as those of the group as a
whole, and these become the values one should display to be regarded positively
within the group, this could explain women's increase in jokes about drugs when
in mixed groups. That humour revolving around drugs is more frequent than
humour relating to alcohol is probably a reection of the fact that drugs are more
\taboo" in our society than is alcohol.

Results distorted by individual conversations

Two of the topics have slightly distorted results.
The most toilet humour occurs in single sex female conversations. This is largely
because of one conversation. I included jokes about farting in this category, and
one of the single sex female conversations includes a lot of humour in which one
of the participants is teased because of her farts. This conversation has raised the
percentage, but if these particular examples were removed, then the percentages
for all of the groups for this category would be fairly low.
Around seven percent of the humour from both men and women in mixed groups
relates to weddings. There is no wedding humour in the single sex groups. This
statistic, again, is largely the result of one conversation. One group involved were
in the process of planning a wedding, and so the subject cropped up often. The
high
gures for wedding humour in mixed sex interaction are therefore distorted
and should be largely disregarded.

135

CHAPTER 9. FOCUS AND TOPIC

136

9.4 Conclusion
Much of the literature on topics of conversation is several decades old. It would
therefore not be surprising if similar studies today showed radically di erent
results. Particularly with the rise of feminism, and more women in the workplace,
we might expect the work/personal division apparent in conversational patterns
in the sixties and seventies to be disappearing. This is particularly true of the
people in my sample who are all well educated and almost all working. In
spite of this, the types of topics joked about by men and women in my sample
seem to follow a similar distribution to what would be expected on the basis of
the literature. Men's humour revolves more often than women's around work,
computers, television shows, movies or books, and alcohol. Humour occurring in
female conversations is more likely to involve people-oriented topics.
In this chapter I also showed that speakers are much more likely to use humour
based on a same-sex group member when in single sex than in mixed sex groups.
In mixed groups, gender group boundaries are maintained by the use of humour
focussed on group members of the opposite sex.
The next chapter discusses the amount of humour contributed by the men and
women in my corpus.

136

Chapter 10

Amount of Humour
10.1 Introduction
In this chapter I discuss the amount of humour contributed by men and women.
Section 10.2 discusses several approaches to this question and outlines the methodology I have used. The results are presented in 10.3 and discussed in section 10.4.
There is much evidence for the fact that men talk more than women in public
settings. In formal contexts such as conferences (Swacker 1979), meetings (Eakins
and Eakins 1979, Edelsky 1981), seminars (Holmes 1988) and television talk
shows (Appleman et al. 1987) men speak more often and for longer.
Holmes (1992:132) points out that:
In contexts where the primary function of talk is interpersonal or
social, women tend to contribute more. When the primary function
of talk is referential and focussed on information, men often talk
more.
Gilligan (1982) puts forward the explanation that men's domination of talking
time in formal contexts and women's in less formal contexts, is a result of the fact
that women are more solidarity-oriented, and men more concerned with status.
There is, in fact, mixed evidence as to whether women talk more in informal
settings. Stubbe (1991) found the young girls in her sample took a greater pro137

138

CHAPTER 10. AMOUNT OF HUMOUR

portion of the available talking time than boys in mixed sex informal dyads.
Easton (1994:15) found the opposite result. Men in mixed sex conversations
contributed 70 percent of the talk. Their turns were both longer and more frequent than women's turns. Duncan and Fiske (1977) used a laboratory study to
show that males spoke more often and held the oor for longer. Other laboratory studies have shown the distribution of talk to be evenly shared (Hirschman
1973 { published as Hirschman 1994, Bilous and Krauss 1988), and studies of
married couples also found symmetrical distribution (Kenkel 1963, Strodtbeck
1975). Some of the apparently conicting results here may be due to audience
size. There is an obligation for members of dyads to take part in the conversation, whereas in larger groups members can more easily hide and contribute
less.
Middleton and Moland (1959) observed that females were less likely to tell jokes
when both sexes were present than were males. Literature discussed in chapter
3 shows that humour is generally regarded a masculine activity and so deemed
inappropriate for females particularly in mixed company.
Edelsky (1981) showed how men and women contributed di erently to two different types of oor. F1 is a one-at-a-time oor, whereas F2 is a oor in which
several people talk at once and chip in ideas and comments. Edelsky found that
in meetings, men took longer turns in F1, and less in F2, occasionally even less
than the women. The women joked more in F2, and the men in F1. F1 occurred
much more than F2 which could partly explain the general
nding that men
dominate speaking in public settings. The \appropriate" format is F1, the oor
preferred by most men. An investigation of amount of humour contributed in
certain types of oor would have been very interesting, and this is certainly a
direction in which this sort of research should move. For the purposes of this
study, however, I was interested only in a general indication of who contributes
the most humour. The literature on humour would certainly suggest it is men
(see section 3.5). We might, on the other hand, expect contributions of humour
to reect overall contribution to the conversation, in which case the form of the
hypothesis would be a little less certain.
In section 10.2 I discuss the process of establishing the distribution of humour
within my data.
138

10.2. QUANTIFYING \AMOUNT" OF HUMOUR

139

10.2 Quantifying \Amount" of humour

10.2.1 Counting instances of humour
There are several di erent approaches to the problem of evaluating the amount
of humour contributed by an individual. The two most obvious approaches are to
count the number of humorous instances contributed, or to establish the overall
time spent being funny. The latter is relatively time consuming, and may not
give an accurate idea of total humour contribution.
A quick witty quip can inject as much humour into a conversation as a long
anecdote, and so I wanted them both to be regarded as equal contributions of
humour. I therefore chose to base the
gure for amount of humour contributed
on the number of examples volunteered, rather than the length of the examples.
Of course, anecdotes can incorporate several funny elements. They are sometimes
lengthy, and so it is desirable that, if there is a lot of humour in them, this is
reected in the
gure given for the speaker's contribution of humour to the
conversation. I tried as far as possible to identify separate humorous elements of
an anecdote. An anecdote which revolved around one funny point, and repeated
that point in several ways, was treated as one instance of humour. An anecdote
based on a series of humorous episodes was counted as more than one instance
of humour. In this way I was able to get a reasonably accurate
gure for the
number of instances each speaker contributed to each conversation.

10.2.2 Calculating relative humour contribution

The easiest, but least satisfying method of establishing men's and women's overall contribution of humour is to compare the number of examples I managed to
collect from each sex in the twenty minute extracts that I chose. This is relatively
uninformative because certain conversations will be more light-hearted and more
prone to humour than others. The topic and atmosphere will have a huge inuence on the amount of humour occurring in any given conversation. Comparing
conversations of equal length will be of little use, because the amount of humour
used will depend largely on topic and atmosphere. If a group of men use a lot of
humour in a 30 minute extract, but a group of women use very little in an equal
139

140

CHAPTER 10. AMOUNT OF HUMOUR

length conversation, this does not tell us much. Conversation \moods" can vary
wildly. The question of how humour is distributed is therefore tricky, and is best
looked at within conversations, rather than across the entire data base.
The relevant question is then, in any given conversation, do all participants have
an equal chance to be humorous? In looking at modals, Preisler (1986) uses a
technique drawn from early social dialectology research (see Romaine 1981). He
looks at the number of modals occurring as a proportion of the number of places
where they could potentially occur. This makes sense. The concept of occurrence/potential is in e ect operative if we look at an individual's contribution of
humour as a percentage of the total humour occurring in a conversation. Across
di erent conversations the \potential for humour" will vary, but if we concentrate
on individual conversations, it will be relatively controlled. Power, solidarity and
other factors will mean that in practice, participants in a conversation will
nd
di ering potential for humor, but at least they will all have the same material
with which to work. The mood of the conversation is a constant.
The amount of humour produced by an individual is compared with the total
humour produced by all interactants, and not with the number of utterances
produced by the speaker. The focus is on how humour functions in small groups,
and its distribution within those groups. I am not looking at the percentage of
people's utterances that are humorous. That is a very di erent question.
In investigating whether attempts at humour are equally distributed, the most
useful quantitative information will be the amount of humour used by each member of a given conversation as a percentage of the total humour used in that
conversation.

10.3 Results
I calculated the percentage of humour each speaker contributed to their conversation. This gave 24 percentage
gures for each of the single sex conversations,
and twelve each for males and females in mixed conversations. As all of the conversations are made up of four speakers, then if all speakers contributed equally
each of the percentages would be around 25 percent. I divided the percentages
into two groups { low contribution (under 25 percent) and high contribution (25
140

10.3. RESULTS

141

percent and over). The results are shown in table 10.1.

FS
MS
FC
MC

Low High
13
11
12
12
8
4
3
9

Table 10.1: Number of high and low contributors of humour

Table 10.1 shows that in single sex conversations, there are usually an equal
number of high contributors and low contributors. If we look at the mixed
groups there is a de
nite pattern. Men are most often high contributors and
women low contributors.
The numbers are too small for this to show up in log-linear modelling. In investigating the models, I was unable to reject the independence model as the
simplest, well-
tting model. These di erences then, while not signi
cant, show
a de
nite trend. The men contributed more humour to mixed sex conversations
than the women did. This can also be seen in
gure 10.1, which plots the percentages for individuals in mixed conversations from highest to lowest on a line
graph. The line which shows the distribution of the females' contribution runs
almost parallel to the males' line at roughly ten percent lower. When the same
graph is plotted for males and females in single sex conversations the lines follow
each other tightly (see
gure 10.2).

141

CHAPTER 10. AMOUNT OF HUMOUR

142

60
FC

percentage

50

MC

40

30

20

10

0
A

individual

Figure 10.1: Percentage humour contribution by individuals in mixed groups

60
MS
50

FS

percentage

40

30

20

10

0
P

individual

Figure 10.2: Percentage humour contribution by individuals in single sex groups

142

10.4. DISCUSSION

143

10.4 Discussion
The results show that, when in mixed groups, men tend to be more frequent
contributors of humour than women. In section 3.5 I discussed literature claiming
women have no sense of humour. More recently researchers have found that
women do, in fact, use humour, although often only in single sex groups. Humour
is seen as a powerful act, and so not appropriate for women to use in the company
of men. Regina Barreca reminisces:
We were taught that what we laughed at had to stay among ourselves
{it wasn't for public consumption. Our funny stories had to be kept
as secret as our Kotex pads, stu ed away in pretty cases to disguise
them and discussed only in all female groups. (Barreca 1991:196)
The results of this study show that women do use humour in mixed sex groups,
although not to the same extent as men. The pattern discussed in section 3.5
of men joking and performing, and women laughing and displaying appreciation
is therefore, to some extent, present. Men are expected to joke and perform
in mixed sex groups, whereas when women use humour in mixed groups they
threaten \the most basic social gender arrangements" (Marlowe 1989:150). Although women do contribute less humour in mixed groups than men, they still
contribute a signi
cant amount of humour. An optimistic interpretation could
be that this is a sign that these \gender arrangements" are in ux, and that it is
becoming increasingly acceptable for women to use humour in mixed company.

143

Chapter 11

Social and Personal Identity

11.1 Background
Tajfel (1982) discusses interpersonal and intergroup relations, considering them
opposite extremes of a continuum.
Interpersonal relations consist of \interaction between 2 or more individuals
which is very largely determined by their individual characteristics and the nature of the personal relations between them" (Tajfel 1979:401).
Intergroup relations, on the other hand, are \interactions which are largely determined by group memberships of the participants and very little { if at all {
by their personal relations or individual characteristics."
The \Social Identity Model" as described by Turner (1982) claims that members
of a group perceive themselves to be of the same social category.
This de
nition stresses that members of a social group often share
no more than a collective perception of their own unity and yet this
seems to be sucient for them to act as a group. (Turner 1982:15)
One's self concept is made up of one's social identity (sex, religion etc) and one's
personal identity (personal tastes, bodily attributes etc).
Everyone has a need for positive self-esteem, and thus for a positive social identity. This motivates a search for positive distinctiveness for one's own group in
145

146

CHAPTER 11. SOCIAL AND PERSONAL IDENTITY

terms of positively valued group characteristics.

The need for positive self-esteem will manifest itself in intragroup as well as
intergroup behaviour. To maintain a positive self-esteem, speakers may need to
di erentiate themselves from other members of the group. This may be achieved
by assigning positive attributes to themselves, and negative characteristics to the
other members of the group, or by assuming that they themselves are closer to
the normative ideal than the others.
These ideals will be di erent for di erent groups. When delivering humour, as
in all interaction, speakers are given the opportunity to portray themselves in a
certain way. It is in the speaker's interest to exploit this to their advantage, and
portray values or characteristics which are positively regarded within the group,
and which they, themselves, regard positively.
In all of my examples, the speaker will be displaying a sense of humour. Regardless of whether the humour is well received or not, it was the speaker's intention
to be funny. This is a characteristic widely valued in our society, so by being funny, the speaker is already displaying one desirable characteristic. If they
succeed they will therefore maintain a position of respect within the group and
increase liking by members of the group. When someone is funny they positively
a ect their status within the group, and solidarity with the group's members.
In addition to a sense of humour, there are many other characteristics that a
speaker could portray. A speaker will portray di erent characteristics depending
on the value system of the group. I have made the assumption that a speaker
will always portray themselves positively. If someone tells a funny story about
an embarrassing moment in their life then the ability to laugh at oneself and be
open and honest is likely to be more positively valued within that group than
always doing everything right.
For almost every instance of humour it is possible to look at the manner in which
the speaker presents themselves, and which aspects of their identity they are
emphasising and shaping. It may be their openness and honesty, their successful
lifestyle, their wit or their ability to laugh at themselves. It could be their
membership of the group, or their identity as a woman, a student or a mother.
It is important to note that I am not suggesting that speakers make a deliberate decision as to how they are going to portray themselves. No doubt this is
146

11.2. ABILITY TO LAUGH AT ONESELF

147

sometimes the case, but in most instances the decision will not be a conscious
one.
Di erences in how the speakers portray themselves through humour may allow
insight into the value systems of the di erent groups. Signi
cant gender di erences in certain contexts may also lead us to conclude that the individuals are
identifying as male or female, and more precisely what it means to identify as
male or female.
This is an interesting issue, and methodologically quite challenging.
In this chapter I examine a very small number of characteristics which speakers
may portray. This is a tiny selection of the huge range of traits that could potentially be displayed through humour. I show how in humour, as in all interaction,
one constantly creates and portrays a personality made up of a number of traits.
I point to what seem to be the main gender di erences occurring in my data,
and identify ways in which this line of research could be fruitfully pursued. It is
important to note that the
gures used for this part of my research are simple
percentages. They can be taken as indications of trends, but not much more can
be read into them. No account has been taken of possible clustering e ects.

11.2 Ability to laugh at oneself

It has been claimed that women put themselves down more than men, and portray themselves negatively. Johnstone (1993) examined stories told by men and
women in the Midwest, and noted that women's stories often centred on incidents in which they violated social norms, and were scared or embarrassed.
When men told stories about themselves, they were more likely to be stories of
deeds in which they were successful.
When going through my transcripts, I noted all examples in which the speakers
joked about something embarrassing they had done or made themselves the butt
of the joke. This included just over ten percent of the humour in all-female
groups, as opposed to 3.6 percent of the all-male humour. In the mixed groups
the di erence was less marked. 5.5 percent of the male humour in mixed groups
showed the speaker's ability to laugh at himself, and 6.8 percent of the female
humour in mixed groups showed this characteristic. Zillman and Stocking (1976)
147

148

CHAPTER 11. SOCIAL AND PERSONAL IDENTITY

found that women found jokes funnier if the teller was making fun of themselves.
This is a pattern also found by Ervin-Tripp and Lampert (1992). In their mixed
groups they found the women did more outgroup putdowns and used less self
directed humour. Why, if individuals need to create a positive personal identity,
do women use so much self directed humour? The answer is that, by laughing at
themselves, the women are creating a positive self identity. They are portraying
an ability to laugh at themselves, a trait which seems to be highly valued amongst
women. This also explains why women do this most in single sex groups. An
audience of women will value and appreciate this trait more highly than a male
or mixed audience.
(42) is an example in which the speaker portrays the ability to laugh at herself.
She talks about when she was at school, and how she was paranoid about getting
lost in Wellington. This is something which she can now laugh about.
(42)
TF:

i really enjoyed going to field work we were

all paranoid about getting lost in wellington +
well I was //ha ha]\

JF:

/ha ha]\\

11.3 Wit/Cleverness
In a lot of examples in my corpus, the speakers portray themselves as being
particularly witty or clever. Such humour often has a competitive streak, and
may trigger bouts of witty repartee, as speakers try to outdo each other. If
a speaker uses wordplay they are almost invariably displaying an ability to be
witty. In the following example, WF is discussing the advice she gets about
having children and TF makes the witty observation that this is just like junk
mail.
(43)
WF:

but some of it's incredibly useful you know so

148

11.4. OPENNESS/HONESTY

149

you get a hundred things you already know and

then someone will give you a really useful tip
TF:

sounds like circulars in the mailbox=/

WF:

/=yeahha
ha]

SF:

ha ha ha]

Ervin-Tripp and Lampert (1992) noted that male humour consists largely of ip
wisecracks, and that this tendency further increased in mixed sex groups. This
indicates that the ability to be witty, or \quick on one's feet" is highly valued
amongst male groups. A similar pattern emerged in my data. 41.44 percent
of the all-male examples in my corpus displayed this trait, as opposed to 30.28
percent in the all female groups. This margin increased slightly in the mixed
groups, with 44.17 percent for the males and 29.13 percent for the females.
It is interesting to speculate reasons why the men actually increase their use
of wit in mixed sex conversations, while women decrease their use of humour
in which they laugh at themselves. It is plausible to assume that the values
regarded as positive in mixed groups reect those valued by the men. The male
speakers may attempt to portray a particularly positive image when in mixed
company, and so exaggerate display of such highly regarded traits. Their wit is
possibly also supported more in mixed groups than in single sex groups, which
would provide further encouragement.

11.4 Openness/Honesty
If speakers display a willingness to share personal information about themselves,
they are seen as being open or honest. Ervin-Tripp and Lampert (1992) found
that over half the women in their sample produced self-revealing narratives for
female audiences, whereas only 16 percent of men did this for male audiences.
In (44) Tracy is telling a story about a boy she went to school with, who used
to crowd her personal space. She had a crush on a boy named Gav at that time
149

CHAPTER 11. SOCIAL AND PERSONAL IDENTITY

150

and comments that he was the only one she wanted in her personal space.
(44)
TF:
SF:

ha ha] //and i\ and like i used to have to=

/ha ha]\\

TF:

=stand like this so that he couldn't touch me

NF:

oh really=/

TF:

/=it was gr- i mean not touch me just

get cause he he was the same thing and he was
really like really into your inhales] in your
in your private space i just wanted gav there
no-one else ha ha]

All:

laugh] 1 sec

TF:

and he was the one that was furtherest away

tut] never mind

In this extract certain aspects of her identity are highlighted. Among other
things, she is displaying openness and honesty in discussing this personal information with her friends. This trait is present in just over ten percent of
the all-female examples. This contrasts strongly with the male conversations,
of which only 0.9 percent of the examples portray the speaker as being open.
This is clearly a trait valued more by women than by men. As with the ability
to laugh at oneself, the women portray themselves as open and honest less frequently when in mixed company. 4.85 percent of the female examples in mixed
groups portrayed this characteristic, and 1.84 percent of the all male examples.

150

11.5. COOLNESS

151

11.5 Coolness
This category was created for the examples in which speakers portray themselves
as being experienced with either sex, drugs or excessive amounts of alcohol.
In (45) BM and AM jointly tell an anecdote in which they are portrayed as
particularly \cool". The anecdote is part of an extended narrative about a pub
crawl they had been on, during which they had got riotously drunk.
(45)
BM:

/and it's really obnoxious\\

/=it's

great really ob- yeah just WAY too obnoxious

drinking games and this + manager came along no
no this oh the guy came along and said oh the
manager says you HAVE to leave //and so\
AM:

/no no\\ f- for

starters he said like um oh can you can you
just quiet quieten it down deep voice]: boys
ha ha ha: //and then\ after a while he got=

DM:
AM:
DM:

/nh nh nh]\\
=well septic //eh\
/yeah\\ //ha ha]\

AM:

/cause all the\\ other

locals //were getting\ SLUTTED=/

BM:

/h]JUST GO\\

AM:

/=h]yeah (everyone else) was leaving

DM:

//ha huh ha huh]

151

CHAPTER 11. SOCIAL AND PERSONAL IDENTITY

152
CM:

/ha ha ha]\\

BM:

they were just saying go fuck you know go and

then the guy says the manager says you HAVE to
go //someone just\ goes sings]: manager is a=

AM:
BM:

/but we just go\\

=wussie //a wussie a wussie:\ and they didn't=

AM:

/wh]e would have just said yells]:

BM:

=bring us back it got REALLY loud and then=

AM:

=FUCK OFF:\\

BM:

=stayed

ALL:

loud laughter] 3 secs

This anecdote tells us quite a bit about the value systems of the group. Drinking
games are cool, being rude to bar managers is admirable, annoying locals is fun
and general drunken and obnoxious behaviour is something to be proud of and
boast about. The speakers are portraying \coolness" a trait which, at least in
their immediate social circle, is very positive and a desirable characteristic. The
speaker is seen to have this characteristic in 8.41 percent of the all male humour,
which contrasts strongly with 1.83 percent of the all female humour. This is as
one would expect, and no doubt reects the \mateship" culture as discussed in
Phillips (1987) and Pilkington (1992). \Coolness" is part of being a \man". The
di erence is much smaller in the mixed sex groups. The trait appears in 6.75
percent of the male humour and 4.85 percent of the female humour. Both groups
are accommodating - the women perhaps trying to be \one of the boys" and the
men no doubt conscious that it is more important to appear cool for one's male
friends than in mixed company.
152

11.6. CONCLUSION

153

11.6 Conclusion
This chapter has shown that in humour, as in all interaction, we are constantly
working on the creation and maintenance of a positive personal identity. In
any interaction, including humour, there are numerous attitudes, values and
traits one could potentially display. In the very act of being funny, a speaker is
portraying themselves as a person with a sense of humour. I have discussed four
traits in this chapter { the ability to laugh at oneself, wit, openness and coolness.
The ability to laugh at oneself and openness or honesty seem to be traits valued
more by female groups than male groups. Coolness and wit, on the other hand,
are traits displayed more by males. In portraying himself as cool, it could then
be said that a male is \doing gender". West and Zimmerman describe how we
are \recruited" from birth to a certain gender.
The \recruitment" process involves not only the appropriation of gender ideals (by the valuation of those ideals as the proper way of behaving) but also gender identities that are important to individuals
and that they strive to maintain. Thus gender di erences, or the
sociocultural shaping of \essential male and female natures" achieve
the status of objective facts." (West and Zimmerman 1987:142)
A signi
cant part of one's personal identity, then, is one's gender identity. The
creation of a positive personal identity occurs constantly throughout all interaction, and in humour in particular. Ideals will vary considerably from person to
person and be shaped by a number of factors, not the least of which is gender.

153

Chapter 12

Laughter and humour support

strategies
\I don't see much sense in that" said Rabbit. \No" said Pooh humbly,
\There isn't. But there was going to be when I began it. It's just that
something happened to it on the way."

12.1 Introduction
In this chapter I discuss the issue of humour support. Section 12.2 reviews literature in this area, and section 12.3 describes the proportion of humour attempts
supported by laughter in my sample. I explain why assuming laughter to be
the sole humour support strategy is a misleading approach, and in section 12.4
describe and exemplify alternative humour support strategies. Section 12.5 describes types of humour which do not require explicit support, and section 12.6
outlines a number of reasons why some attempts at humour may fail.

12.2 Laughter as humour support

Most literature on humour, and on strategies for supporting it, assume that
laughter is the normal and most appropriate support for an attempt at humour.
155

156 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES

Norrick (1993:23) claims joking and laughter are an adjacency pair, and includes
this assumption in the criteria he uses to identify instances of humour for his
research. If a laugh or \aw" or snide comment follow a recognisable joking
structure, he claims \it seems reasonable to say the speaker was joking, teasing,
playing with words, being sarcastic, or something similar"(1993:8). Norrick insists on these explicit signs that something is funny for an extract to be included
in his corpus of humour. He admits that laughter can be used for purposes other
than to support humour, but does not consider that other support strategies may
be available, or even more appropriate for certain types of humour.
Kottho (1986:23) hypothesises:
Frauen werden aktiv f#ur die Gespr#achserfolge ihres Gegen#ubers. Mit
ihrem Lachen leisten sie Beziehungsarbeit. M#anner tun dies (vor
allem f#ur Frauen) weniger.
Women actively encourage the success of the speaker. They provide
support through laughter. Men do this less frequently (especially for
women)."
Again, Kottho 's hypothesis seems to imply that laughter is the primary means
for supporting humour. She does not consider possible alternative strategies.
There have been a number of studies which indicate that women do, indeed,
laugh more than men. Dreher (1983 as cited in Kottho 1986) studied four
conversations and found not only did women laugh more than men, but both
the men and the women laughed more in support of male speakers than female
speakers. This supports Kottho 's hypothesis. Bogaers (1993) found that more
laughter occurred in her all female conversation, as did Easton (1994). MakriTsilipakou (1994) found that women laughed more in mixed conversations than
men, which meant that it was usually men who found themselves at the receiving
end of aliative laughter.
This pattern is generally explained by male dominance and the di erent roles
males and females are expected to play in conversation.
Pizzini (1991:483) comments:
In our culture this (societal) structure requires that women be passive
156

12.3. COUNTING LAUGHTER

157

and receptive: men make the jokes and women laugh at them.
Barreca (1991:5) reminisces:
Nobody said we should giggle at his jokes only if we found them
funny we had to giggle at his jokes even when we thought they were
dull, insulting or dumb.
Kramarae (1987) observes a similar point.
Women are often put in the situation of having to choose between
laughing at jokes that they do not think are funny ... or risking
becoming an outsider in many female/male social groups.
This explanation no doubt has some truth. Much research has found that women
are generally more conversationally supportive than men (see chapter 3), and so
this would lead us to predict that women would be more supportive of humour
than men, even when they do not
nd the humour funny. The quotes above also
suggest that they must provide this support in order to be accepted.

12.3 Counting laughter

One of the variables I was interested in investigating was whether attempts at
humour were supported. As an initial approach to the question I followed the
literature in assuming that laughter is the primary means for humour support. I
therefore coded each example of humour as either supported by laughter, or not
supported by laughter but soon discovered this is a simplistic, possibly misleading
approach. In this section I will briey outline the results found using this method,
and then discuss the problems with this approach, and address some of the
complexities involved in quantifying humour support.
None of the models proposing a relationship between speaker sex and support
t
the data well. There does, however, appear to be a slight relationship between
group composition and support by laughter. The relevant analysis of variance
table is shown below:
157

158 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES

MAXIMUM-LIKELIHOOD ANALYSIS-OF-VARIANCE TABLE
Source

DF

Chi-Square

Prob

-------------------------------------------------COMPOSITION

79.26

0.0000

SEX

33.88

0.0000

SUPPORT

161.20

0.0000

COMPOSITION*SEX

0.03

0.8644

COMPOSITION*SUPPORT

2.48

0.1155

LIKELIHOOD RATIO

2.55

0.2790

This model
ts well. The likelihood ratio chi square
gure falls easily below
the 7.82 mark for 3 degrees of freedom to .05 signi
cance. The probability is
reasonably high. The chi-square
gure for group composition * support however,
is low, and so we should try taking this out and see what e ect this has on the
model. The analysis of variance table for the independence model is shown below:
MAXIMUM-LIKELIHOOD ANALYSIS-OF-VARIANCE TABLE
Source

DF

Chi-Square

Prob

-------------------------------------------------COMPOSITION

90.17

0.0000

SEX

33.88

0.0000

SUPPORT

176.11

0.0000

COMPOSITION*SEX

0.03

0.8644

LIKELIHOOD RATIO

5.08

0.1660

The di erence between the two models is not statistically signi
cant, and so,
as the independence model is the simpler of the two, this is the one that we
should accept. We can see that the data does show some link between group
composition and humour support, but that this relationship does not reach the
158

12.3. COUNTING LAUGHTER

159

level of statistical signi
cance. The model of main e ects for group composition
is only signi
cantly di erent from the independence model to the .20 level.
An examination of the collected data shows the nature of the trend indicated in
the model. Table 12.1 shows the percentages of humour in each group supported
by laughter. All the percentages are roughly the same except for the
gure for
men in mixed groups. Men in mixed groups are almost ten percent more likely
to have their humour supported than men or women in single sex groups, and
women in mixed sex groups.
%
F - single 71.76
M - single 73.27
F - mixed 72.82
M - mixed 80.98

no.
155
244
75
132

Table 12.1: Percentage of humour supported by laughter

I did not note by whom the laughter was supported, although one could hypothesise that, since the literature indicates women are more supportive than men,
the women in this study are supporting men's humour more than the men are
supporting the women's. If this is the case, then it is interesting to hypothesise
why women would be more supportive of men's humour than they are of humour
produced by other women, in either single sex or mixed groups. This question is
complicated greatly by many issues surrounding the notion of humour support.
Example (46) is an excerpt from one of my mixed conversations, which conforms
to the observations of the researchers discussed in section 12.2. The men joke, the
women laugh. The group has been discussing the words the parents used for their
private parts when they were young. The group is laughing because BM's parents
called this part of his body Colin, and this had lead to some embarrassment when
he later started scouts. A fantasy sequence hypothesising worse names follows.
It is initiated by GM who suggests that \car" would have been an even worse
name, and then the two males, GM and BM joke for some time about various
words and the confusion that could ensue. The two women are laughing almost
constantly, whereas the men laugh relatively little. Given the subject matter
one may be tempted to interpret the women's laughter as embarrassment rather
159

160 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES

than humour support. It is clear from the tape however, that the women are not
embarrassed by the subject. The group speaks openly and freely about a range
of sensitive issues during the taping, and in this excerpt the women show no sign
of embarrassment. They are merely enjoying the humour.
(46)
JF:

/=ha ha ha ha ha] that's

brilliant //ha ha]\ your h] parents were=

AF:

/ha ha ha]\\

JF:
GM:

=very cruel
well i mean you could have called it they could
have been really cruel and called it something
like a car

AF:

//ha ha ha ha]\

JF:

/ha ha ha ha]\\ you wouldn't have been-

BM:

or a television

AF:

ha ha ha]

GM:

hey have you heard how HUGE //they (

AF+JF:
GM:
BM:
All:

/laugh]\\
=//i mean you really SCREW s-\
/twenty four inch colour television\\

laugh]

160

)\=

12.3. COUNTING LAUGHTER

BM:

161

remote CONTROL twenty four inch colour

television

AF+JF:laugh loudly]
GM:

with with um s- stereo sph]eakers or one of

those silly things that tilts in different
directions

AF+JF:still laughing - right through GM's speech]

BM:

what are you going to do with black and white

portables

AF:

ha ha ha ha ha]

GM:

with cars they have hoods AND (

BM:

yeah

AF:

h h huh]

GM:

and as for television watchers

BM:

mm + television critics

It is clear that, at least in this example, the women are supporting the humour
with laughter, and the males are laughing less frequently. But can we conclude from this that the men are not supporting each other in their attempts
at humour? Surely by picking up on GM's wit, and pursuing it further BM is
supporting GM in his attempt at humour. He must think the quip is funny, or he
would not develop the line of humour himself. Together they spar, both competing, and jointly developing the theme, and in doing so expressing a commonality
in their sense of humour, solidarity and support.
To assume, then, that laughter is the sole means of supporting humour, is to
161

162 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES

obscure a great deal. There are a number of available support strategies, some
more explicit than others, and in some circumstances, laughter may not be the
most appropriate. The general
nding that women laugh more than men tells
us only that women laugh more. We can not generalise from this to claim that
women are more supportive of humour than men, without
rst examining alternative support strategies, and exploring the possibility that some of these may
be used more often by men than by women. It is beyond the scope of this thesis
to examine the distribution of humour support strategies. I aim only to highlight
some of the methods used to support humour and indicate the direction in which
I believe future research in this area should lead.

12.4 Humour support strategies

12.4.1 Contributing more humour
In the example discussed in the previous section, the men supported each other
by developing a theme and contributing more humour. The humorous frame
is maintained. To contribute to a humorous frame is to acknowledge that one
exists, and so acknowledge the previous speaker's humour. Example (47) is an
extract from a discussion between four males. There is very little laughter, and
yet a humorous frame is maintained throughout. Most of the humour consists of
insults, and there is a wordplay towards the end of the sequence. The speakers
are clearly enjoying themselves, and there is no indication in any of their voices
or reactions that they feel their humour was rejected or ignored. By maintaining
a humorous frame and sparring and bouncing humour around the men support
each other's humour. The men will not feel unsatis
ed after this routine, nor
feel that their humour has gone unsupported or unappreciated.
(47)
CM:

/=cause she didn't

want you there that's all i mean come on a MALE
in a kitchen (that can right rid of you)

TM:

if //(they had girls they would've had)\

162

12.4. HUMOUR SUPPORT STRATEGIES

NM:

163

/clarence didn't count ha]\\=/

MM:

/=no of course
not

TM:

he's not male he's barely human=/

MM:
TM:

/=yeah
i don't know how they get that much body odour
in a female but i guess it's possibh]le

MM:

he's not even human at ALL thank you very much

CM:

oh that's okay thank you //very much (

MM:

)\

/a bloody insult
//saying (

) is human\\

NM:

/got kling on aspirations remember\\

CM:

probably got kling on genes in him but we won't

go into that

MM:

not wearing my jeans

TM:

enough of that=/

NM:

/=yes

Sometimes maintaining the humorous frame, or playing along with a \gag" initiated by the
rst speaker can provide very solid support. In (48) DF pretends
to misunderstand CF's words for humorous e ect. BF does not immediately
pick up that DF is being funny, and so CF agrees with DF in a quite serious
163

164 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES

voice, and continues with the gag. CF could not sound more serious in her reply, yet this provides strong support for DF's humour. Her words indicate that
she appreciates DF's humour, and together they fool BF into thinking they are
serious.
(48)
CF:

i mean i've got bad feeling in my hands anyway

BF:

have you

CF:

like i can never feel pulses or stuff like like

you know

DF:

pulses what like beans? like beans? you mean

BF:

NO

DF:

pulses you mean //kidney beans\ and the like

CF:
CF:

/yeah\\
and lentils=/

BF:

/=oh does she h]=/

CF:

/=i find it really

hard to feel lentils

BF

ha ho]

Irony is a type of humour which often invites the audience to join in, and support
the speaker by maintaining the ironic tone. In example (49) PM expresses mock
disgust at having to spend time at Waipuna Lodge, all expenses paid, for his
work. AM and BM support the irony with more irony, o ering PM mock sympathy for his upcoming \ordeal". No laughter occurs, yet the humour is adequately
164

12.4. HUMOUR SUPPORT STRATEGIES

165

supported.
(49)
PM:

yeah but i i tut] absolutely disGUSTed i've

got to spend two days in waipuna lodge

AM:

//tut] oh mate\

BM:

/drawls]: aw:\\ how sad

Fantasy humour is often supported by more fantasy. Speakers will jointly construct long and involved scenarios. The funniest contributions will be explicitly
supported with laughter, but most are supported only by more fantasy. Speakers
usually incorporate or build on humour o ered by the other participants, and so
the humour has by no means failed. In (50) the speakers speculate about what
could have happened if PM had responded to the advance made by a model the
previous evening. They jointly construct the scenario.
(50)
PM:

um but yeah that could have been the one

could've been rich lived a life of sin

DM:

//she could have set me with the (

)\

GM:

/she could have she could have been\\ in the

women's weekly man oh i could have seen like
you know pat and maybe (1 sec of everyone
speaking at once)

PM:

unnamed friend yeah ha ha] yeah i was that

unnamed friend

GM

next larry forensky or whatever tough man REAL

man sort of thing

165

166 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES

DM

) make their song and everything ha] in the

charts //ha ha]\

GM:

/yeah]\\ un-=/

DM:

/=write a book=/

EM:

/=oh ho ha
ha]

12.4.2 Echo
Humour can be e ectively supported by echoing the words of the speaker. A
member of the audience will repeat the words in appreciation, often as if savouring the humour. In (51) AM repeats CM's words in a tone that indicates he
appreciates them and
nds them funny.
(51)
CM:

/=too many
brain cells in his beer vat now

AM:
CM:

yeah //in his beer vat yeah\

/laughs]\\

Example (52) is similar. RM is describing a vegetarian restaurant crawl. One of

the restaurants had unexpectedly changed their menu and begun to serve meats,
and so he had eaten
sh as part of the crawl.
(52)
TM:

fish? they don't serve f- fish do they=/

166

12.4. HUMOUR SUPPORT STRATEGIES

RM:

167
/=oh yeah

they serve all meats=/

TM:

/=oh have they sort of

//um\

DM:

/ha ha]\\=/

LM:

/=ha ha] imitates RM's intonation]:

all meats:

RM's humour is supported through laughter, but LM also supports the humour
by echoing RM's words.

12.4.3 Oer sympathy or contradict self-deprecating humour

For some humour, laughter could actually be an inappropriate response. If humour is used in troubles-talk, then laughing could indicate that you
nd the
speaker's unfortunate situation funny. The speaker can laugh at their own problems, but in general, the appropriate response to such humour seems to be an
o er of sympathy. In example (53) TF tells of a woman she met just before she
had her baby, and laughs at the fact that the woman was paranoid about little
things that were not going to be ready when her baby was born, which seemed
ridiculous in comparison with TF's half
nished house. Rather than laugh at
TF's situation, the others o er sympathy.
(53)
TF:

this woman um was saying to me just before he

was born now high voice]: oh we haven't got
the nursery ready: h]and yeah we h]haven't
done this and //we haven't done\ that i said=

WF:

/drawls]: oh god:\\

167

168 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES

TF:

=FUCK we've only got half a bloody house=/

BF:
/=yeah=/
WF:

/=yeah=/

Similarly in (54) CF is not con
dent about an upcoming judo tournament in

which she is on a team with a friend of hers who is very good. She jokes that
they will be a mixed ability pair. Rather than laugh, which would indicate that
she agreed the skill di erence was that large, JF assures CF that she will do
ne.
(54)
CF:

/mixed ability pair\\

laughs]

JF:

you will it'll be fine...

In (55) BM tells an anecdote about his ears, and AF and DF assure him that his
ears look okay.
(55)
BM:

/a hard\\ time //right\ Mum used=

DF:
BM:

/yeah\\
=to tell me when i was a child that if i was
born a generation ago they would've put a big
band around my head laughing]: (you know to
keep them there):=/

AF:

/=but your ears aren't that

bad

168

12.4. HUMOUR SUPPORT STRATEGIES

BM:

no no

DF:

you've got quite a big head //so it sort of

169

balances it\

If they had laughed at his anecdote, it may have indicated that they agreed that
his ears seemed big. DF's comment could be interpreted as an underhand insult,
but there is no indication in her voice that it is intended as such!

12.4.4 Overlap and heightened involvement in the conversation

A speaker can show enthusiasm and appreciation for another's humour by indicating excitement, by using overlap or other means of signalling general involvement in the conversation. In example (56) the group are speculating about
reasons why a friend acts strangely when with a particular woman. VF supports
RF's hypothesis by completing her sentence.
(56)
RF:

/cause\\ SHE'S gone + i bet you've

never been laid in your life or- or maybe she's
INterested in him and so he's trying to look

VF:
RF:

world/ly\\
/ex\\perienced

12.4.5 Non-verbal
It is also possible to support humour non-verbally { usually with a smile. Unfortunately this can not be observed in my taped data.

169

170 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES

12.5 When explicit support is not needed

There are some types of humour which do not always need explicit support. The
speaker does not necessarily expect the audience to respond in any way, and so
when there is no laughter or other form of support, the humour has not failed.

12.5.1 The humour is a support strategy itself

Some follow-up humour supports humour initiated by another person. When
humour is a support strategy itself, then it does not need explicit support from
the audience. Example (57) shows the end of an anecdote in which BM describes
his experiences of nitrous oxide from when he was at secondary school. DF's
humour directly supports BM's anecdote, and so does not need support itself.
(57)
BM:

cause it does i mean light- light- lightening

fixes nitrogen + not necessarily nitrous
oxh]ide but i mean i didn't know the
difference

DF:

ha ha ha ha] you didn't really CARE did you

ha]

CM:

remember you used to get those little capsules

DF:

//yeah\

BM:

/oh right\\

12.5.2 Irony
There are a number of examples of irony which are not supported in any way,
and for which the speaker does not seem to expect support. Norrick (1993:72)
suggests irony may be an unmarked form of talk for some speakers. It is true that
170

12.6. FAILED HUMOUR

171

some speakers use irony quite a lot, and irony can be di erent from other forms
of humour in that it can sometimes be a ippant way of expressing quite a serious
meaning. When speakers use irony, they do not always expect explicit support
from the audience. Example (58) occurs after an explanation about something
which the speaker had originally expected the audience to know. TM had asked
for clari
cation, and once a relatively obscure explanation was received said:
(58)
TM:

/=yes oh silly for not knowing

The other speaker then continued with his story. Support was not o ered, nor
expected. TM was merely using irony to make a point.

12.6 Failed humour

Having identi
ed the possible means of supporting humour, and noted circumstances in which humour does not seem to need support, it was then possible
to identify those instances of humour in my corpus which were not supported,
and so failed. For most of the cases, it was possible to identify a reason for the
humour's failure. In this section I discuss \mistakes" that are sometimes made
when attempting humour.

12.6.1 Insucient contextualisation

Zajdman (1991) carefully documents ways in which a canned joke can be introduced into discourse. Not only jokes, but all humour must ow from the
previous conversation, and in some way
t neatly in. If an attempt at humour
is not relevant, or somehow removed from the situation, it may go unsupported.
In example (59) SM has been telling an anecdote about a group of people from
the IHC who had come to see a play he was in. They had laughed a lot which,
SM suggests, is probably why the audience laughed so much that night as well.
GM then attempts some fantasy humour about taking such people to the bathroom. The humour does not ow from SM's story and, uncontextualised, seems
unfunny, and is not supported.
171

172 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES

(59)
SM:

=huh] it was reh]ally //funny ha]\=

HM:
SM:

/ha ha ha ha]\\
=thah]t's probably what the audience were
laughing not //us but\ yeah ha]

HM:
GM:

/sighs]\\
wouldn't like to take some of the group of a
few people out to the bathroom

SM then changes the topic entirely.

12.6.2 Being too late or reviving \dead" humour

Groups often latch onto formulas or develop a theme of humour which can result
in lengthy routines. Once the conversation moves on, though, the humour dies.
It can be revived, but again, only if in a particularly relevant context. In example (60) EM is only a few seconds late with his humour, but nevertheless, the
conversation has already moved on, and so the humour fails. MM says he has a
tape with sound e ects on, which would have been funny to play onto the tape
they were making. The conversation moves on to the fridge, from which CM has
just fetched a beer. EM then contributes some humour on the theme of sound
e ects. He is too late with this, so it is not supported.
(60)
MM:

/=i knew i should have brought my

tape recorder which has sounds to play back at
them

CM:

tell you what alan's fridge is a lot better

stocked than our one is ha]=/

172

12.6. FAILED HUMOUR

173

EM:
DM:
CM:

/=ha ha]
was=/
/=laughs]: was:=/

MM:
EM:

/=ha ha ha]
i can do some whistles] feedback ha ha]
inhales] clears throat]

CM:

looking at tape-recorder] miles to go

DM:

yawns]

EM is clearly conscious that his humour has failed. He inhales and clears his
throat to cover the silence and his embarrassment. Clearing the throat seems to
be a relatively common strategy for coping with failed humour.

12.6.3 Assuming too much background knowledge

Assuming knowledge of people
Appreciation of funny anecdotes about people generally requires knowledge of
the subjects of the humour. The type of humour which relies on such knowledge
generally overlaps with sections of gossip. Jane Pilkington (1994) points out that
gossip has a \limited sphere of interest", if the audience is not familiar with the
protagonists, the gossip will usually be unsuccessful. Many funny anecdotes rely
on this familiarity. In (61) SF tries to joke about what a strange couple Tessa
Davies and Tim Dapple make. The audience is not familiar with Tim Dapple,
SF makes no attempts to further identify him, and so her attempt at humour is
unsupported.
(61)
173

174 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES

SF:

tessa davies is going out with tim dapple

can you think of a BETTER couple

TF:
JF:

i //can't remember\ tim dapple but i know=

/who's he\\

TF:

=i've heard the name and i know him

SF:

both being total bloody weirdho]os

Assuming specialist knowledge

An audience clearly will not respond to humour which relies on specialist knowledge they do not have. Example (62) is an extract taken from one of the tapes I
used from the New Zealand Corpus of Spoken English. One of the speakers was
a linguist. When GM jokes that there are some words he would never use in an
essay, AF jokes that she would like to transcribe them and discuss them in an
essay. None of the others are linguists and so probably do not understand her
reference to \elision" or why AF
nds this funny. The humour therefore fails.
(62)
GM:

i don't know there are still some words that i

won't use in an essay that i use in every day
life like + bugger me with //a wooden spoon\ i=

AF:
GM:
BM:

/ha ha ha]\\
mean //i've yet to\ see an essay that you-=
/yeah i know but\\

GM:

=tut] to quote citizen ropespierre //bugger me\=

JF:

/ha ha ha]\\

174

12.6. FAILED HUMOUR

GM:

175

=with a wooden spoon send him to the guillotine

it's just i mean

AF:

i'd quite like to transcribe it and say this

was said fast h]and that there was elision
there h]

BM:

but i mean in in in a twenty one hundred word

essay or something it's not uncommon to only
use four hundred different words

AF:

mm

12.6.4 Misjudging relation between speaker and audience

Some humour requires a certain solidarity or understanding between the speaker
and the focus of the humour. Insults are an example of a type of humour which
requires a certain relation to exist. People who have a joking relationship will
often insult each other, with the understanding that it is all in play and serves
to further solidarity. Only a select group of people will have the right to tease a
particular person. If someone misjudges a relation, for example by assuming a
joking relationship where none has previously existed, then the humour is prone
to failure. In example (63) the group is trying to
gure out why JF had been
absent from a geography trip at school some years ago. Earlier in the conversation
JF had joked that whenever she goes away on holiday she seems to get sick,
often, she insinuated, because she was drunk. It is acceptable for JF herself to
joke about this, but when SF tries to tease her about it in this extract it goes
unsupported. There is no evidence that a joking relationship exists between JF
and SF, and SF appears to have over-stepped the mark, and assumed a relation
with JF which would allow her to tease her about such matters. This relation
does not exist, and the humour fails. Chiaro (1992:5) notes: \Tacit rules underlie
where, when and with whom it is permissible to joke."
(63)
175

176 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES

SF:

cause it was at COLlege=/

JF:

/=how come i didn't go=/

SF:
/=yeah it was in geography=/
JF:

/=i must have been

sick=/

SF:

/=you must have been sick you must have

been on holiday //or s-\ ha ha]

TF:

/maybe the sixth form and\\

seventh form joined or something cause I did it
in seventh form

12.6.5 Negatively teasing someone present

Power based teases often go unsupported. In negatively teasing someone present,
the speaker places the audience in a bind. They can either support the tease, and
thereby indicate to the butt that they agree with the speaker, or they can restrain
from supporting the humour, thereby supporting the focus of the tease. In
example (64) BM teases AM that he will be \killed" in an upcoming tournament
they are both involved in. The audience does not respond to the humour. To do
so would be to explicitly take sides.
(64)
AM:

/=doubles so we're going to have

a doubles um wee tournament with you and kim

BM:

laughs] we'll just fucken //kill you\

AM:

/we got\\ a hundred

and twenty two million

176

12.6. FAILED HUMOUR

CM:

high pitched]: what:

DM:

that's an extra ball eh

BM:

not bad

177

12.6.6 Trying to gain membership of exclusive sub-group

Within a group there will often be smaller exclusive sub-groups. Such groups
may have established joking routines. To successfully join such a routine when
one is not a group member is usually challenging. Miriam Meyerho (1994)
discusses this concept in terms of network ties.
The way interlocutors communicate with each other is by means of
shared network ties which create sociolinguistically meaningful links
between individuals' shared social identities. (Meyerho 1994)
To be able to successfully participate in a routine, one must have access to the
appropriate network tie. In example (65) DM and LM are quoting from their
favourite TV programme. This is a humour routine they often use together. AM
has seen the episode in question and tries to join in, recalling a funny part of
the program that he had liked. LM replies with a non-commital \yeah", and the
two continue with their routine. AM does not have access to the appropriate
\network tie" and so is not permitted to participate in the routine.
(65)
DM:

the powdered toastman er ren and stimpy quotes]:

quick man //cling tenaciously to my buttocks:\

LM:

/quotes]: cling tenaciously to my

buttocks:\\=/

TM:

/=okay fine

177

178 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES

DM:

quotes]: what? drawls]: dome:: //ha ha ha=

LM:

/h h h ha ha=

DM:

=ha ha ha ha ha]\

LM:

=ha h]\\

AM:

yes i quite like the goh]ing backwards bit

that fh]un=/

LM:
DM:

/=yeah
quotes]: wind intensity too great:=/

LM:

/=h
huh]=/

DM:

/=quotes]: can't hold on:

12.6.7 Disrupting serious conversation

Humour is often disruptive. Chiaro (1992:114) points out that \evaluation following the joke, whether verbal or non-verbal, will distract participants from
whatever discourse had preceded the witty interruption." This distraction is
usually tolerated, but sometimes the audience will withhold evaluation or support so that the serious conversation can continue without further interruptions.
In example (66) MM has an important reason for wanting to know whether the
person under discussion has
nished his thesis. CM's wordplay is therefore not
responded to and serious conversation continues.
(66)
MM:

oh is he f- officially finished?

TM:

i //don't know\

178

12.6. FAILED HUMOUR

NM:

179

/he's finished his-\\

CM:

yeh]ah one way or the other he's finishedh]

MM:

better talk to p- pat about what she wants me

to do then

12.6.8 Portraying oneself inappropriately for one's status or gender

There are certain expectations about what certain people can joke about, and
how we should portray ourselves. If we put ourselves across in a way that violates society's expectations or norms then we sometimes will not receive support.
In chapter 11 I discussed the formation and representation of aspects of one's
identity through humour. One of the things I showed in that discussion was
that men, more than women, tend to portray themselves as experienced with
sex, drugs or excessive amounts of alcohol. That men joke about this indicates
that within these groups at least some members consider these to be positive
traits for a man to have. Women tend to joke or boast about these things less
frequently. This may indicate that it is \cool" for a man to have experience in
such matters, but \uncool" for women. The following excerpt is from a mixed
conversation, and follows a discussion of dentists and methods of pain relief.
(67)
BM

remember the gas though

DF:

i never had anything like that

AF:

i had gas for the baby tooth=/

BM:
AF:

/=yeah
and yeah and then i had these injections for
these molars

179

180 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES

BM:

that was GREAT=/

CM:

/=mm=/

DF:
AF:

/=ha ha ha]
the gas i c- yeah i can remember i had yeah i
had these baby teeth pulled out in hamilton
cause that's where i went to school
//secondary\ school and i remember walking=

BM:

/yeah\\

AF:

down hamilton east ha] and blimmen dazed going

back to the convent you know ha] i was hmmm
ha ha] (

DF:

that's nitrous oxide eh=/

BM:

/=yeah=/

DF:
BM:

/=yeah
tut] i'd just been doing fifth form science
clears throat] and i came out and i said deep
voice] : lightening makes this stuff: ha]

All: laugh]
BM:

cause it does i mean light- light- lightening

fixes nitrogen + not necessarily nitrous
oxh]ide but i mean i didn't know the
difference

180

12.7. CONCLUSION
DF:

181

ha ha ha ha] you didn't really CARE did you

ha]

The group reminisces about experiences they had with nitrous oxide at the dentist. BM jokingly comments: \that was GREAT" at which DF laughs. AF then
tells an anecdote about walking back to the convent in a daze after having gas.
The audience does not support the humour in her anecdote, instead clarifying
a technical point. Finally, BM tells an anecdote about when he had it in
fth
form. This is met with a lot of laughter. The fact that it seems generally more
\cool" and acceptable for men to have experience with, and have enjoyed drugs
than women may provide an explanation for why BM's comments on this topic
are supported, but AF's anecdote is not.

12.6.9 Other unsupported humour

Having increased the range of strategies that count as support for humour, and
eliminating the examples for which the speaker made some sort of violation, there
are very few remaining examples which are unsupported. There are six examples
from the all male conversations, one example from the all female conversations
and three from the mixed { all three of which are examples of women using
humour. These numbers represent a very small proportion of the corpus. They
indicate an expected trend, but considering the di erent number of examples in
each of the categories, really tell us nothing concrete at all.

12.7 Conclusion
It can be seen that there is a great deal to remember when attempting humour.
To maximise the chances that one's attempt at humour is successful, one must
attempt to avoid the pitfalls outlined in the previous section. To risk humour is
to risk considerable loss of face. Humour and audiences are, of course, unpredictable. Even if one of the above violations occurs, the humour may still be
supported. And if none of the above things go wrong, other factors may interfere
and the humour remain unsupported.
The issue of who is more \supportive" of humour is therefore a more complicated
181

182 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES

one than many researchers claim. There are a number of support strategies
which can be employed, some of which provide stronger support than others.
The context will, to some extent, dictate the most appropriate support strategy,
and this will not always be laughter. It seems that women and men may favour
di erent strategies for supporting humour. They may also withhold support for
di erent reasons and to di erent extents. We could hypothesise that women
may be more forgiving of the violations discussed in the previous sections, and
so support the humour anyway. It is also possible that certain violations seem
more severe to one sex than the other. Research on humour support now needs
to move beyond counts of laughter occurrences. We need to closely examine
the distribution of humour support strategies, and also investigate the types of
humour men and women support most, and identify the reasons for which they
tend to withhold support.

182

Chapter 13

Conclusion
...luckily Owl kept his head and told us that the Opposite of an Introduction, my dear Pooh, was a Contradiction and, as he is very good
at long words, I am sure that that's what it is.

Humour research is a relatively new and diverse
eld spanning numerous disciplines. The terminology in this area is often confused, ill-de
ned and inconsistent. We are lacking uni
ed taxonomies and theories, and only very recently
have researchers begun to look at naturally occurring, spoken humour.
Research in most areas of language and gender is much more advanced. Researchers consistently
nd results which suggest that women's conversational
style is supportive and collaborative, whereas men tend to emphasise status and
competition in their interaction. There are several proposed explanations as to
the reasons behind this di erence, but any reasonable explanation must include
aspects of both socialisation and dominance. We should begin to move away from
the conceptualisation of gender as a static category. The most interesting and
fruitful work in language and gender is detailed and careful qualitative analysis.
There is now a foundation of initial quantitative work on which such analysis
can be based. Unfortunately studies in gender and humour research are sparse.
This research aimed to lay some ground work which would identify interesting
trends and point to areas which could reward closer, more detailed examination.
I examined some aspects of gender and humour more closely in the
nal chapters
of this thesis, but there are many areas discussed primarily quantitatively which
183

CHAPTER 13. CONCLUSION

184

could now bene
t from in-depth analysis.

A number of biases have lead to some rather bizarre claims in the gender and
humour literature. The most pervasive of these is that women have no sense
of humour. This has been proven resoundingly false. There are several reasons
why this claim is so prevalent. The
rst is that joking is not considered a \ladylike" activity. In mixed groups women have not been expected or encouraged
to use humour, although they are expected to support humour volunteered by
males. Their joking has therefore been largely restricted to single sex groups.
The second reason for this myth is that women's sense of humour, and their
joking style is in many respects di erent from men's. The literature suggests
that women prefer types of humour that are spontaneous, loosely structured,
and which involve their audience. They prefer to use humour in small groups.
Men's humour is more likely to have a performance element to it, and serve a
status-forming function. Women's humour has been shown to be healing and
supportive.
These claims are based on a small body of literature, largely based on questionnaires, anecdotal evidence and introspection. I was interested in investigating
naturally occurring humour in small friendship groups, and identifying possible
di erences in the type, function, focus, topic and amount of humour contributed.
The corpus used for this work was partly contributed by data collected for Easton
(1994), the Victoria University Linguistics Library and the Wellington Corpus
of Spoken New Zealand English. The rest of the conversations were collected by
my friends and their contacts. The resulting corpus consists of eighteen conversations. Six are between four females, six between four males, and six take place
in mixed groups of two males and two females. All of the speakers self-identify
as Pakeha, are between 18 and 35 years, have some form of tertiary education,
are in a relaxed, familiar environment, and in natural friendship groups.
The taping situation proved more comfortable and natural for the female groups
and the mixed groups than for the all male groups. Female friendship groups
spend more time in conversation, whereas male friendship groups are often taskbased. They meet to partake in a given activity. For the women, then, the
tape-recorder was incidental to the discussion, for the males, it was the reason
for it. The fact that
ve out of six of the all male groups knew they were
184

185
recording the discussion for me, a female, may have meant that they monitored
their speech more than the other groups. The tape-recorder is gendered and so
the discussions may not be entirely representative of all male conversations.
Several of the variables were analysed using a statistical technique called loglinear modelling. This was described in detail in chapter 5.
A working taxonomy of di erent types of spoken humour was constructed from
a survey of the literature. The taxonomy was then modi
ed to reect the types
of humour occurring in my data. Log-linear modelling showed that both speaker
sex and group composition a ect the types of humour used. The most signi
cant
results from this part of my research were:

women are more likely to use observational humour than men

insults and roleplay are more likely to occur in single sex conversations

humour involving quotes or vulgarity seldom occurs in mixed interaction

and is more often used by men than by women

There were also a number of trends which are interesting, but not statistically
signi
cant:

anecdotes and fantasy humour are more likely to occur in mixed interaction
than single sex groups

men tend to use more roleplay and wordplay than women

women are likely to use more jocular insults

I also analysed the \freshness" of each instance of humour. The best model
showed main e ects for both the sex of the speaker and the group composition.
The majority of the humour used was fresh humour. Men are more likely to use
external source humour than women. This supports prior research which suggests
men are more likely to discuss things they have seen or read than are women. It
has also been suggested that men's humour tends to be less context-bound than
women's. External source humour can be stored and performed without much
regard for context.
185

CHAPTER 13. CONCLUSION

186

A taxonomy of functions and humour strategies was constructed. The results

showed women more likely to use humour to create solidarity than men. In
particular the women often use humour to share personal information about
themselves. When using humour for solidarity based purposes, men are more
likely to capitalise on shared experiences or highlight similarities. Particularly
power based functions such as controlling or fostering conict occurred rarely in
my data. This is a reection of the fact that my recordings were all of natural
friendship groups. Men are more likely to use humour only for the general function of increasing status and solidarity and performing positive work on their
personal identity. Men are more likely to use humour to cope with a situational
problem, whereas females more often use general coping strategies which enable
them to cope with problems not speci
c to the immediate situation. The results
also showed that, contrary to literature indicating teasing is a predominantly
male activity, men and women tease both in a powerful manner and to create
or maintain solidarity, although this strategy is largely restricted to single sex
groups.
An investigation of the types of topics joked about con
rmed the consistent

ndings of the last few decades. Linguistic and anthropological research has
found that men talk more about work, activities or things they have seen or
read, whereas women tend to discuss personal matters, relationships, home and
family. The topics of humour in this study followed a similar distribution. The
humour used by females was much more likely to be about a topic involving
people. Men joked more about work, computers, T.V. shows, movies or books
and alcohol.
An analysis of the focus of humour showed that speakers are much more likely
to use humour focussed on a same-sex group member when in single sex groups
than mixed groups. In mixed interaction speakers maintain gender boundaries
by focussing humour on group members of the opposite sex.
In mixed sex groups, men contributed more humour than women. This supports
claims that, as joking has been regarded a masculine activity, women are reserved
with their humour when in mixed groups.
When using humour, as in all interaction, speakers perform work on their personal identity. Humour is an opportunity to portray oneself in a certain way,
186

187
and it is in the speakers' interest to exploit this by portraying characteristics
which they regard as positive, and which are positively regarded by the group.
In addition to a sense of humour, there are many characteristics a speaker could
display when using humour. Analysis of how speakers present themselves can
give insight into the value systems of the group. The examples in my corpus
indicate that the ability to laugh at oneself and to be open and honest are valued more by female groups than male groups, and wit and \coolness" are traits
valued more by males.
In chapter 12 I show that the issue of humour support is much more complicated
than many researchers claim. There are numerous support strategies available
to any audience, some of which provide stronger support than others. The most
appropriate strategies will vary with context. Some types of humour do not
require any explicit support at all, and there are many reasons why an audience may choose not to support an attempt at humour. Future research needs
to move beyond pure laughter counts which are not only simplistic, but also
misleading. We need to carefully investigate the distribution of various humour
support strategies, examine the types of humour men and women are most likely
to support, and the reasons for which they tend to withhold support.
This is a good example of the necessity of considering conversational data in context. Counting forms without consideration of their function and the dynamics
of surrounding discourse is dangerous and obscures much interesting, often crucial information. In this study I have tried to combine qualitative research with
some basic quantitative analysis. The quantitative analyses of various aspects of
humour provide a solid groundwork on which future, more detailed research can
be based.
One challenge facing those who wish to analyse recorded spontaneous conversation is the collection of unbiased, comparable data. It appears the most natural
settings and environments for groups of males may be quite di erent from those
most natural for many females. There seems no obvious solution to this problem.
Either the naturalness of the conversation in the male groups, or comparable settings must be sacri
ced. This problem may not be so marked in the analysis of
conversations between dyads. It is likely to be much more natural for two men
to sit at home and talk than a group of four.
187

CHAPTER 13. CONCLUSION

188

An analysis of humour occurring in dyads would provide valuable comparative

data. It is possible that the dynamics of humour in such conversations di er from
humour in small groups. Similarly comparative analysis of ethnic di erences, age
di erences, di erences in intimacy or educational background of the speakers, the
inuence of alcohol, and other such variables would be fruitful. Some of these
variables have been investigated by way of questionnaire or experimental work,
but there is virtually no such data available based on careful investigation of
naturally occurring spoken conversation.
The trends identi
ed in this thesis provide a basis for much further empirical
investigation and analysis. The statistical results are interesting in themselves,
but are important primarily in that they provide a starting point for careful
in-depth analysis of how individuals construct gendered identities via humour.

188

Appendix A

Key to Transcription
Conventions
With the exception of the more detailed system for transcribing laughter, the
conventions used in this thesis are based largely on those developed at Victoria University for the Wellington Corpus of Spoken New Zealand English (ACCENZ).
Speakers are labelled using an initial and the letter F or M to indicate their sex.

Transcription in doubt

()
Speech indecipherable
(hello) Transcriber's best guess at an unclear utterance

Intonation

?
Rising or question intonation
{
Incomplete or cut-o utterance
YES Capitals indicate emphatic stress

Paralinguistic and other non-verbal features

Descriptions of paralinguistic and non-verbal features are contained in square
brackets. If the feature is concurrent with speech, or describing speech, the
relevant speech is placed between colons, e.g:
189

APPENDIX A. KEY TO TRANSCRIPTION CONVENTIONS

190
AM:

sneezes]

BM:

silly voice]: you never can tell with bees:

Pauses

+ pause of up to one second

++ pause of up to two seconds

Simultaneous Speech and Latching

Simultaneous speech is contained in slashes, as in the following example:
AF:

remember the time when //we were at school and\

BF:

/what about when you wore that\\ green

hat

If someone's speech follows another's directly then latching is signalled as in the

following example:
AF:

i used to go to school and=/

BM:

/=you wore that green hat

A \=" signals speech continues from an earlier line:

AM:

i would go to school almost //every day\ wearing this=

BF:
AM:

/ha ha ha]\\
=bright green hat

190

191

Laughter

$h]
$huh]
$ha]
$nh]
hello$ho]
$laughs] 2 secs

laughing exhalation
laughing inhalation
voiced laugh particle
nasalised laugh particle
laughing repetition of syllable
used for prolonged laughter,
or for a group of people laughing.

191

Bibliography
Appleman, S., Heijerman, A., van Puijenbroek, M. and Schreuder, K. (1987).
How to take the oor without being oored. In Brouwer, D. and de Haan,
D., editors, Women's Language, Socialization and Self-image, pages 164{
175. Foris Publications, Dordrecht.
Aries, E. J. (1976). Interaction patterns and themes of male, female and mixed
groups. Small Group Interaction, 7(1):7{18.
Aries, E. J. and Johnson, F. L. (1983). Close friendship in adulthood: conversational content between same-sex friends. Sex Roles, 9(12):1183{96.
Barreca, R. (1991). They used to call me Snow White but I drifted. Viking, New
York.
Bashiruddin, A., Edge, J. and Hughes-Pelegrin, E. (1990). Who speaks in seminars? Status, culture and gender at Durham University. In Clark, R.,
Fairclough, N., Ivanic, R., McCleod, N., Thomas, J. and Meara, P., editor, Language and Power, pages 74{84. Centre for Information on Language
Teaching, London.
Beattie, G. W. (1981). Interruption in conversational interaction and its relation
to the sex and status of the interactants. Linguistics, 19:15{35.
Berger, A. (1976). Anatomy of the joke. Journal of Communication, 26:113{115.
Bilous, F. R. and Kraus, R. M. (1988). Dominance and accommodation in the
conversational behaviours of same- and mixed-gender dyads. Language and
Communication, 8(3/4):183{94.
Bogaers, I. E. W. M. (1993). Gender in job interviews: some implications of
verbal interactions of women and men. Working Papers in Language, Gender
and Sexism, 3(1):53{82.
Brotherton, P. and Penman, R. (1977). A comparison of some of the characteristics of male and female speech. Journal of Social Psychology, 103:129{146.
193

BIBLIOGRAPHY

194

Brown, P. and Levinson, S. (1987). Politeness: some universals in language

usage. Cambridge University Press, Cambridge, New York.
Cameron, D. (1992). Not gender di erences but the di erence gender makes {
explanation in research on sex and language. International Journal of the
Sociology of Language, 94:13{26.
Cameron, D., McAlindon F. and O'Leary, K. (1988). Lako in context. In
Coates, J. and Cameron, D., editors, Women in their speech communities,
pages 74{93. Longman, London.
Castell, P. J. and Goldstein, J. H. (1977). Social Occasions for Joking: a CrossCultural Study. In Chapman, A. J. and Foot, H. C., editors, It's a funny
thing, humour: International conference on humour and laughter, pages
193{197. Pergamon Press, Oxford.
Chambers, J. C. (1992). Linguistic correlates of gender and sex. English WorldWide, 13(2):173{218.
Chiaro, D. (1992). The Language of Jokes. Routledge.
Christensen, R. (1990). Log-Linear Models. Springer-Verlag New York Ltd, New
York.
Coates, J. (1986). Women, Men and Language: a sociolinguistic account of sex
dierences in language. Longman, London, New York.
Coates, J. (1988). Gossip revisited: language in all-female groups. In Coates,
J. and Cameron, D., editors, Women in their Speech Communities, pages
94{122. Longman, London.
Collinson, D. (1988). Engineering humour: Masculinity, joking and conict in
shop-oor relations. Organization Studies, 9(2):181{199.
Condry, J. and Condry, S. (1976). Sex di erences: a study of the eye of the
beholder. Child Development, 47:812{819.
Coser, R. (1960). Laughter among colleagues: A study of the functions of humor
among the sta of a mental hospital. Psychiatry, 23:81{95.
Cox, J.A., Read, J.L. and Van Auken, P.M. (1990). Male-female di erences in
communicating job-related humor: an exploratory study. Humor, 3(3):287{
295.
Cozby, P. C. (1973). Self-disclosure: A literature review. Psychological Bulletin,
79:73{91.
194

195
Crawford, M. (1989). Humour in Conversational Context: Beyond Biases in
the Study of Gender and Humour. In Unger, R. K., editor, Representations: Social Constructions of Gender, pages 155{166. Baywood Publishing
Company inc., Amityville, NY.
Crawford, M. and Gressley, D. (1991). Creativity, Caring and Context { women's
and men's accounts of humor preferences and practices. Psychology of
Women Quarterly, 15(2):217{231.
Cruttenden, A. (1986). Intonation. Cambridge University Press, Cambridge.
Davies, C. (1982). Ethnic jokes, moral values and social boundaries. British
Journal of Sociology, 33(3):383{403.
Davies, C. (1984). Joint joking: Improvisational humorous episodes in conversation. In Brugman, C. et al., editor, Proceedings of the Tenth Annual Meeting
of the Berkeley Linguistics Society, pages 360{71, Berkeley. Berkeley Ling
Soc.
Davies, C. (1990). Ethnic Humor around the world. Indiana University Press,
Bloomington & Indianapolis.
Douvan, E. and Adelson, J. (1966). The adolescent experience. Wiley, New York.
Dreher, W. (1982). Gespr#achsanalyse: Macht als Kategorie m#annlichen Interacktionsverh#altens. Sprecherwechsel und Lachen. Master's thesis, Berlin.
Duncan, S. and Fiske, D. W. (1977). Face to Face interaction: research, methods
and theory. Halsted Press, New York.
Duncan, W. J. (1984). Perceived humor and social network patterns in a sample
of task-oriented groups: a reexamination of prior research. Human Relations,
37(11):895{907.
Eakins, B. and Eakins, G. (1979). Verbal turn-taking and exchanges in faculty
dialogue. In Dubois, B. L. and Crouch, I., editors, The Sociology of the
Languages of American Women, pages 53{62. Trinity University, Texas.
Easton, A. (1994). Talk and laughter in New Zealand women's and men's speech.
Wellington Working Papers in Linguistics, (6):1{25.
Eckert, P. (1990). Cooperative competition in adolescent `girl talk'. Discourse
Processes, 13:91{122.
Eckert, P. and McConnell-Ginet, S. (1992). Think practically and look locally:
language and gender as community-based practice. Annual Review of Anthropology, 21:461{90.
195

BIBLIOGRAPHY

196

Edelsky, C. (1981). Who's got the oor? Language in Society, 10:383{421.

Eder, D. (1990). Serious and playful disputes: variation in conict talk among
female adolescents. In Grimshaw, A., editor, Conict Talk: Sociolinguistic Investigations of Arguments in Conversations, pages 67{84. Cambridge
University Press, Cambridge.
Eder, D. (1993). \Go Get Ya a French!": Romantic and Sexual Teasing Amongst
Adolescent Girls. In Tannen, D., editor, Gender and Conversational Interaction, pages 17{23. Oxford University Press, NY/Oxford.
Ervin-Tripp, S. and Lampert, M. D. (1992). Gender Di erences in the Construction of Humorous Talk. In Hall, K., Bucholtz, M., and Moonwomon,
B., editors, Locating Power, Proceedings of the second Berkeley Women and
Language Conference April 4 and 5 1992, volume 1, pages 108{117, California. Berkeley Women and Language Group.
Feigelson, S. (1989). Mixing mirth and management. Supervision, 50(11):6{8.
Fink, E. L. and Walker, B. A. (1977). Humorous responses to embarrassment.
Psychological Reports, 40:475{485.
Fishman, P. (1977). Interactional Shitwork. Heresies: a Feminist publication on
Arts and Politics, (2):99{101.
Fishman, P. (1983). Interaction: the work women do. In Thorne, B., Kramarae
C. and Henley, N., editor, Language Gender and Society, pages 89{101.
Newbury House, Rowley, Mass.
Folb, E. A. (1980). Runnin' down some lines: The language and culture of Black
teenagers. Harvard Univ. Press, Cambridge, Mass.
Freud, S. (1905). Der Witz und seine Beziehung zum Unbewussten.
Gerson, J. M. (1985). The Variability and Salience of Gender: Issues of Conceptualization and Measurement. Paper presented at annual meeting of the
American Sociological Association, Washington D.C.
Gilbert, G. (1981). Modelling Society: an introduction to log-linear analysis for
social researchers. G. Allen and Unwin, London.
Gilligan, C. (1982). In a Dierent Voice. Harvard University Press, London.
Goodman, L. (1992). Gender and humour. In Bonner, Goodman, Allen, Janes
and King, editor, Imagining Women: Cultural Representations and Gender.
Polity, Cambridge.
196

197
Goodwin, M. H. (1982). `Instigating' storytelling as social process. American
Ethnologist, 9(4):799{819.
Graham, E., Papa, M. and G. Brooks (1992). Functions of humour in conversation: Conceptualization and measurement. Western Journal of Communication, 56(2):161{183.
Grotjahn, M. (1957). Beyond Laughter: A Psychoanalytical Approach to Humor.
McGraw Hill.
Haas, M. and Sherman, M. A. (1982). Reported topics of conversation among
same-sex adults. Communication Quarterly, 30(4):332{42.
Hampes, W. P. (1992). Relation between humour and intimacy. Pyschological
Reports, 71:127{130.
Haverkate, H. (1990). A speech act analysis of irony. Journal of Pragmatics,
14:77{109.
Hay, J. (1994). Jocular abuse in mixed gender interaction. Wellington Working
Papers in Linguistics, (6):26{55.
Henley, N. M. and Kramarae, C. (1991). Gender, power and miscommunication.
In Coupland, N., Giles, H., and Wiemann, J. W., editors, Miscommunication
and Problematic Talk, pages 18{43. Sage, London.
Hill, J. H. (1987). Women's speech in modern Mexicano. In Philips, S., Steele, S.,
and Tanz, C., editors, Language, gender and sex in comparative perspective,
pages 121{160. Cambridge University Press, Cambridge.
Hirschman, L. (1994). Female-male di erences in conversational interaction.
Language in Society, 23(3):428{42.
Holmes, J. (1984). Women's language: a functional approach. General Linguistics, 24(3):174{78.
Holmes, J. (1986). Functions of \you know" in women and men's speech. Language in Society, 15(1):1{21.
Holmes, J. (1988). Sex di erences in seminar contributions. BAAL Newsletter,
31:33{41.
Holmes, J. (1992). Women's talk in public contexts. Discourse and Society,
3(2):131{150.
Holmes, O. (1864). Soundings from the Atlantic. Ticknor and Fields, Boston.
197

BIBLIOGRAPHY

198

Jenkins, M. (1985). What's so funny?: Joking among women. In Bremner,

S., Caskey, N. and Moonwomon, B., editor, Proceedings of the rst Berkeley Women and Language Conference, pages 135{151, California. Berkeley
Women and Language Group.
Jespersen, O. (1922). Language: Its Nature, Development and Origins. Allen
and Unwin, London.
Johnson, F. L. and Aries, E. J. (1983). Conversational patterns among lateadolescent, close friends: A study of same sex pairs. Journal of Genetic
Psychology, 142:225{238.
Johnstone, B. (1993). Community and Contest: Midwestern men and women creating their worlds in conversational storytelling. In Tannen, D., editor, Gender and Conversational Interaction, pages 62{80. Oxford University Press,
NY/Oxford.
Kaufman, G. (1991). Introduction: Humour and power. In Kaufman, G., editor,
In stitches: a patchwork of feminist humor and satire, pages vii{xii. Indiana
University Press, Bloomington, IND.
Keith-Spiegel, P. (1972). Early conceptions of humor: varieties and issues. In
Goldstein, J. and McGhee, P., editors, The psychology of humor, pages 4{39.
Academic Press, New York.
Kenkel, W. F. (1963). Observational studies of husband-wife interaction in family
decision-making. In Sussman, M., editor, Sourcebook in Marriage and the
Family, pages 144{56. Houghton Miin, Boston.
Kennedy, J. J. (1983). Analyzing qualitative data. Praeger Publishers, New York.
Kiper, P. S. (1987). Gender and topic. Language in Society, 16(4):543{547.
Komarovsky, M. (1962). Blue-collar marriage. Random House, New York.
Kottho , H. (1986). Scherzen und Lachen in Gesprachen von Frauen und
M#annern. Der Deutschunterricht: Beitrage zu seiner Praxis und Wissenschaftlichen Grundlegung, 38(3):16{28.
Kramarae, C. (1987). Women and Men Speaking, chapter 3, pages 52{63. Newbury house.
Kuiper, K. (1991). Sporting formulae in New Zealand English: two models of
male solidarity. In Cheshire, J., editor, English Around the World: Sociolinguistic Perspectives, chapter 13, pages 200{209. Cambridge University
Press, Cambridge.
198

199
Lako , R. (1975). Language and woman's place. Harper Colophon, New York.
Lave, J. and Wenger, E. (1991). Situated Learning: Legitimate Peripheral Participation. Cambridge University Press, Cambridge.
Le Masters, E. (1975). Blue collar aristocrats: lifestyles at a working-class tavern.
University of Wisconsin Press, Madison.
Lever, J. (1976). Sex di erences in the games children play. Social Problems,
23:478{487.
Levine, J. (1976). The feminine routine. Journal of Communication, 26:173{175.
Lewis, M. (1972). Parents and Children: Sex-role Development. School Review,
80:229{240.
Lewis, M. and Freedle, R. (1973). Mother-Infant Dyad: The Cradle of Meaning. In Pliner, P., Krames, L. and Alloyway, T., editor, Communcation and
Aect, Language and Thought. Academic Press, New York.
Linstead, S. (1985). Jokers wild: the importance of humour in the maintenance
of organizational culture. Sociological Review, 33(4):741{767.
Long, D. and Graesser, A. (1988). Wit and humor in discourse processing.
Discourse Processes: A Multidisciplinary Journal, 11(2):35{60.
Loudon, J. (1970). Teasing and Socialization on Tristan da Cunha. In Mayer,
P., editor, Socialization: The approach from Social Anthropology, pages 293{
331. Tavistock Publications Ltd, London.
Makri-Tsilipakou, M. (1991). Doing disagreement, the case of gender. Working
Papers in Language, Gender and Sexism, 1(1):57{84.
Makri-Tsilipakou, M. (1994). Laughing their way: Gender and conversational
mirth. Working papers in Language, Gender and Sexism, 4(1):15{50.
Maltz, D. N. and Borker, R. A. (1983). A cultural approach to male-female
miscommunication. In Gumperz, J. J., editor, Language and Social Identity,
pages 196{216. Cambridge University Press, Cambridge.
Manly, B. (1992). The design and analysis of research studies. Cambridge University Press, Cambridge.
Marlowe, L. (1989). A sense of humour. In Unger, R. K., editor, Representations: Social Constructions of Gender, pages 145{154. Baywood Publishing
Company inc., Amityville, NY.
199

BIBLIOGRAPHY

200

Martineau, W. (1972). A Model of the Social Functions of Humor. In Goldstein,

J. and McGhee, P. E., editors, The Psychology of Humor, chapter 5, pages
101{125. Academic Press, New York and London.
McGhee, P. (1979a). Humor: its origin and development. Freeman, San Francisco.
McGhee, P. (1979b). The Role of Laughter and Humor in Growing Up Female.
In Kopp, C., editor, Becoming Female: Perspectives on Development, chapter 7. Plenum Press, New York.
Meyerho , M. (1994). Rethinking `gender' as a sociolinguistic variable. University of Pennsylvania Working Papers in Linguistics, 1(1).
Middleton, R. and Moland, J. (1959). Humor in negro and white subcultures:
A study of jokes among university students. American Sociological Review,
24:61{69.
Mitchell, C. (1985). Some di erences in male and female joke telling. In Jordan,
R. and Kalcik, S., editors, Women's Folklore, Women's Culture, pages 163{
186. U. of Pennsylvania Press.
Monro, D. (1953). Argument of Laughter. Melbourne University Press, Melbourne.
Morreall, J. (1983). Taking Laughter Seriously. State University of New York,
Albany.
Mulac, A, Wiemann, J.M., Widenmann S.J. and Gibson, T.W. (1988).
Male/female language di erences and e ects in same sex and mixed-sex
dyads: the gender-linked language e ect. Communication Monographs,
55(4):315{35.
Neitz, M. J. (1980). Humor, hierarchy and the changing status of women. Psychiatry, 43(3):211{223.
Neuliep, J. (1991). An examination of the content of high school teachers' humor
in the classroom and the development of an inductively derived taxonomy
of classroom humor. Communication Education, 40(4):343{355.
Norrick, N. (1993). Conversational Joking: humor in everyday talk. Indiana
University Press, Bloomington, Indianapolis.
Norrick, N. R. (1994). Involvement and joking in conversation. Journal of Pragmatics, 22:409{430.
200

201
O'Barr, W. M. and Atkins, B. K. (1980). `Women's language' or `powerless
language'? In McConnell-Ginet, S., Borker R. and Furman, N., editor,
Women and Language in Literature and Society, pages 93{110. Praeger,
New York.
Parkin, D. (1979). The creativity of abuse. Man, (15):45{64.
Parsons, T. and Bales, R. (1955). Family, Socialization and Interaction Processes.
Free Press, New York.
Philips, S. U. (1980). Sex di erences and language. Annual Review of Anthropology, 9:523{544.
Phillips, J. (1987). A man's country? The image of the Pakeha male { A history.
Penguin Books (NZ) Ltd, Auckland.
Pilkington, J. (1992). Don't try to make out that I'm nice! The di erent strategies women and men use when gossiping. Wellington Working Papers in
Linguistics, 5:37{60.
Pilkington, J. (1994). Women, Men and Gossip: What's the Story? Master's
thesis, Victoria University of Wellington, New Zealand.
Pizzini, F. (1991). Communication hierarchies in humour: Gender di erences in
the obstetrical/gynaecological setting. Discourse in Society, 2(4):477{488.
Pogrebin, M. and Poole, E. (1988). Humor in the brie
ng room. A study of the
strategic uses of humor among police. Journal of Contemporary Ethnography, 17(2):183{210.
Pollio, H. and Edgerly, J. (1976). Comedians and comic style. In Chapman, A.
and Foot, H., editors, Humour and laughter: Theory, research and applications. Wiley, London.
Poyatos, F. (1993). Paralanguage: a linguistic and interdisciplinary approach to
interactive speech and sound. J. Benjamins, Amsterdam.
Preisler, B. (1986). Linguistic Sex roles in conversation. Mouton de Gruyter,
Berlin.
Radcli e-Brown, A. (1952 (1940)). Structure and Function in Primitive Society,
chapter 4 - On Joking Relationships. Cohen and West Ltd, London.
Raskin, V. (1985). Semantic Mechanisms of humor. D. Reidel Pub. Co.,
Dortrecht Boston.
Redfern, W. D. (1984). Puns. Basil Blackwell Publisher, Oxford.
201

BIBLIOGRAPHY

202

Romaine, S. (1981). The status of variable rules in sociolinguistic theory. Journal

of Linguistics, 17:1{178.
Roy, A. M. (1978). Irony in Conversation. PhD thesis, U. of Michigan.
Sherzer, J. (1985). Puns and jokes. In Van Dijk, T. A., editor, Handbook of
Discourse Analysis, volume 3, chapter 15. Academic Press Inc. (London)
Ltd., London.
Smith, P. M. (1985). Language, the Sexes and Society. Blackwell, Oxford.
Soskin, W. F. and John, V. P. (1963). The study of spontaneous talk. In Barker,
R., editor, The Stream of Behaviour. Appleton-Century Crofts, New York.
Spradley, J. and Mann, B. (1975). The Cocktail Waitress. Wiley, London and
New York.
Strodtbeck, F. L. (1975). Husband-wife interaction over revealed di erences.
American Sociological Review, 16:468{473.
Stubbe, M. (1991). Talking at cross-purposes? The e ect of gender on New
Zealand primary schoolchildren's interaction strategies in pair discussions.
Master's thesis, Victoria University of Wellington, Wellington, New Zealand.
Suls, J. M. (1972). A two-stage model for the appreciation of jokes and cartoons: An information-processing analysis. In Goldstein, J. and McGhee,
P., editors, The psychology of humor. Academic, New York.
Swacker, M. (1979). Women's verbal behaviour at learned and professional conferences. In Dubois, B. L. and Crouch, I., editors, The Sociology of the
Languages of American Women, pages 155{160. Trinity University, Texas.
Swann, J. (1988). Talk control: an illustration from the classroom of problems
in analyzing male dominance in conversation. In Coates, J. and Cameron,
D., editors, Women in their speech communities. Longman.
Tajfel, H. (1982). Social Psychology of Intergroup Relations. Annual Review of
Psychology, 33:1{39.
Tannen, D. (1984). Conversational Style: Analyzing Talk Amongst Friends.
Ablex Publishing Corporation, Norwood, New Jersey.
Tannen, D. (1990a). You Just Don't Understand: Women and Men in Conversation. William Morrow, New York.
Tannen, D. (1990b). Gender di erences in topical coherence: Creating involvement in best friends' talk. Discourse Processes, 13:73{90.
202

203
Tannen, D. (1993). The Relativity of Linguistic Strategies: Rethinking Power
and Solidarity in Gender and Dominance. In Tannen, D., editor, Gender
and Conversational Interaction, pages 165{188. Oxford University Press,
New York/Oxford.
Thoman, E.B., Leiderman, P.H. and Olson, J.P. (1972). Neonate-Mother interaction during breastfeeding. Developmental Psychology, 6:110{118.
Turner, J. C. (1982). Towards a cognitive rede
nition of the social group. In
Tajfel, H., editor, Social Identity and Intergroup Relations, chapter 1, pages
15{40. Cambridge University Press.
Uchida, A. (1992). When `di erence' is `dominance': a critique of the `anti-powerbased' cultural approach to sex di erences. Language in Society, 21(4):547{
568.
Vinton, K. L. (1989). Humor in the workplace: it is more than telling jokes.
Small Group Behaviour, 20(2):151{166.
Vitulli, W. and Tyler, K. (1988). Sex-related attitudes toward humour among
high-school and college students. Psychological Reports, 63:616{618.
Vitulli, W. F. and Barbin, J. M. (1991). Humour-value assessment as a function
of sex, age and education. Psychological Reports, 69:1155{1164.
West, C. and Fenstermaker, S. (1993). Power, inequality and the accomplishment
of gender: an ethnomethodological view. In England, P., editor, Theory on
Gender/Feminism on Theory., chapter 8. Aldine de Gruyter, New York.
West, C. and Zimmerman, D. H. (1987). Doing Gender. Gender and Society,
1(2):123{151.
Whitehead, A. (1976). Sexual antagonism in Herefordshire. In Barker, D. and
Allen, S., editors, Dependence and Exploitation in Work and Marriage,
pages 169{203. Longman, London.
Winick, C. (1976). The social context of humour. Journal of Communication,
26:124{128.
Wolfson, N. (1988). The Bulge: A Theory of speech behaviour and social distance. In Fire, J., editor, Second Language Discourse: A Textbook of Current
Research, pages 21{38. Ablex Publishing Corporation, Norwood.
Zajdman, A. (1991). Contextualization of canned jokes in discourse. Humor,
4(1):23{40.
203

BIBLIOGRAPHY

204

Zijderveld, A. C. (1983). The sociology of humour and laughter. Current Sociology, 31(3):1{100.
Zillman, D. and Cantor, J. (1976). A Disposition Theory of Humor and Mirth. In
Chapman, A. and Foot, H., editors, Humor and Laughter: Theory, Research,
and Applications. Wiley.
Zillman, D. and Stocking, S. (1976). Putdown humour. Journal of Communication, 26(154-163).
Zimmerman, D. and West, C. (1975). Sex roles, interruptions and silences in
conversation. In Thorne, B. and Henley, N., editors, Language and Sex:
Dierence and Dominance, pages 105{129. Newbury House, Rowley, Mass.
Ziv, A. (1984). Personality and sense of humor. Springer, New York.
Ziv, A. (1988). National Styles of Humor. Greenwood Press, Inc, Connecticut,
USA.

204