Académique Documents
Professionnel Documents
Culture Documents
Jennifer Hay
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 benetted greatly from the advice and comments of several sem-
inar and conference audiences. I would like to thank Miriam Meyerho and
Maria Stubbe for providing references, comments, enthusiasm and support. Use-
ful 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 sacriced 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 1
2 Review of humour research 3
2.1 Introduction : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 3
2.2 Dening Humour : : : : : : : : : : : : : : : : : : : : : : : : : : : 4
2.3 Theories of humour : : : : : : : : : : : : : : : : : : : : : : : : : : 7
2.4 Humour research using elicitation or surveys : : : : : : : : : : : : 8
2.5 Research on conversational joking : : : : : : : : : : : : : : : : : : 8
2.6 Literature on types of humour : : : : : : : : : : : : : : : : : : : : 10
2.7 The functions of humour : : : : : : : : : : : : : : : : : : : : : : : 12
2.8 Conclusion : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 14
i
ii CONTENTS
4 Methodology 35
4.1 Introduction : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 35
4.2 Control of Variables : : : : : : : : : : : : : : : : : : : : : : : : : 35
4.2.1 Number of Speakers : : : : : : : : : : : : : : : : : : : : : 35
4.2.2 Sex : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 36
4.2.3 Ethnicity : : : : : : : : : : : : : : : : : : : : : : : : : : : 36
4.2.4 Age : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 36
4.2.5 Education : : : : : : : : : : : : : : : : : : : : : : : : : : : 37
4.2.6 Intimacy of Speakers : : : : : : : : : : : : : : : : : : : : : 37
4.2.7 Setting : : : : : : : : : : : : : : : : : : : : : : : : : : : : 38
4.2.8 Alcohol : : : : : : : : : : : : : : : : : : : : : : : : : : : : 38
4.3 Data Collection : : : : : : : : : : : : : : : : : : : : : : : : : : : : 38
4.3.1 Speakers' knowledge of the project : : : : : : : : : : : : : 39
4.4 Selection of examples : : : : : : : : : : : : : : : : : : : : : : : : : 40
4.5 Methodological problems : : : : : : : : : : : : : : : : : : : : : : : 41
5 Statistical methods 51
5.1 Log-linear modelling : : : : : : : : : : : : : : : : : : : : : : : : : 51
5.2 Clustering e ects : : : : : : : : : : : : : : : : : : : : : : : : : : : 59
5.2.1 Clustering due to individual behaviour : : : : : : : : : : : 59
5.2.2 Clustering due to conversational ow : : : : : : : : : : : : 59
5.2.3 Clustering due to complex examples : : : : : : : : : : : : 60
5.3 A statistical check on clustering : : : : : : : : : : : : : : : : : : : 60
5.4 The suitability of statistics for analysing conversational data : : 60
6 Types of humour 63
6.1 Introduction : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 63
6.2 The Taxonomy of Types : : : : : : : : : : : : : : : : : : : : : : : 65
6.3 Dening the Types : : : : : : : : : : : : : : : : : : : : : : : : : : 65
6.3.1 Anecdote : : : : : : : : : : : : : : : : : : : : : : : : : : : 65
ii
CONTENTS iii
6.3.2 Fantasy : : : : : : : : : : : : : : : : : : : : : : : : : : : : 68
6.3.3 Insult : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 70
6.3.4 Irony : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 71
6.3.5 Jokes : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 73
6.3.6 Observational : : : : : : : : : : : : : : : : : : : : : : : : : 74
6.3.7 Quote : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 75
6.3.8 Roleplay : : : : : : : : : : : : : : : : : : : : : : : : : : : : 76
6.3.9 Self Deprecation : : : : : : : : : : : : : : : : : : : : : : : 78
6.3.10 Vulgarity : : : : : : : : : : : : : : : : : : : : : : : : : : : 78
6.3.11 Wordplay : : : : : : : : : : : : : : : : : : : : : : : : : : : 79
6.3.12 Other : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 81
6.4 Results : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 82
6.5 Discussion : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 84
6.6 Conclusion : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 89
iii
iv CONTENTS
iv
CONTENTS v
v
vi CONTENTS
13 Conclusion 183
A Key to Transcription Conventions 189
Bibliography 193
vi
List of Figures
8.1 The functions of humour : : : : : : : : : : : : : : : : : : : : : : : 98
10.1 Humour contribution by individuals in mixed groups : : : : : : : 142
10.2 Humour contribution by individuals in single sex groups : : : : : 142
vii
List of Tables
4.1 Pre-collected data available : : : : : : : : : : : : : : : : : : : : : 39
5.1 Example model: Observed frequencies : : : : : : : : : : : : : : : 53
5.2 Example model: Predicted frequencies according to the indepen-
dence model : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 53
5.3 Example model: SAS analysis for independence model : : : : : : 54
5.4 Example model: SAS analysis of model showing interaction be-
tween all variables : : : : : : : : : : : : : : : : : : : : : : : : : : 55
5.5 Example model: SAS analysis of model of main e ects for hair : 56
5.6 Example model: Predicted frequencies for model of main e ects
for hair : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 57
5.7 Example model: Predicted frequencies for model of main e ects
for hair { collapsed : : : : : : : : : : : : : : : : : : : : : : : : : : 58
5.8 Example model: Conditional probabilities : : : : : : : : : : : : : 58
6.1 Probability gures for types of humour : : : : : : : : : : : : : : : 82
6.2 Odds ratios for types of humour : : : : : : : : : : : : : : : : : : : 83
7.1 Probabilities for \freshness" of humour : : : : : : : : : : : : : : : 95
7.2 Odds ratios for \freshness" of humour : : : : : : : : : : : : : : : 95
8.1 Results of possible models for overall function : : : : : : : : : : : 112
8.2 Probability gures for overall function : : : : : : : : : : : : : : : 112
8.3 Odds ratios for overall function : : : : : : : : : : : : : : : : : : : 113
8.4 Probabilities for solidarity based strategies : : : : : : : : : : : : : 114
ix
x CONTENTS
x
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 prolic, 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 pos-
tulating 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 ll-
ing this gap.
In chapter 2 I briey review relevant literature in humour research, giving an in-
dication of past ndings and areas which would benet from closer investigation.
Chapter 3 outlines recent developments in the area of gender and language, and
goes on to discuss research which has dealt specically with gender and humour.
The specic 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
2 CHAPTER 1. INTRODUCTION
2
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 denition 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
4 CHAPTER 2. REVIEW OF HUMOUR RESEARCH
4
2.2. DEFINING HUMOUR 5
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 denition. He states
\humour is conceived generically to be any communicative instance which is
perceived as humorous by any of the interacting parties". This denition, while
more comprehensive than many others, seems rather circular.
It can be seen that, though the denition 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
6 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) dene humour in the wider sense of Martineau (1972), in-
corporating both interpretation and intention. They dene jokes as things said
to deliberately provoke amusement.
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 dened as
anything deliberately said that provokes amusement in a specic con-
versational context (ie context bound). (Long and Graesser 1988:37)
6
2.3. THEORIES OF HUMOUR 7
more restrictive still. Sherzer (1985) for example, denes 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
dene 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.
7
8 CHAPTER 2. REVIEW OF HUMOUR RESEARCH
8
2.5. RESEARCH ON CONVERSATIONAL JOKING 9
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 dened 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).
10
2.6. LITERATURE ON TYPES OF HUMOUR 11
1. Puns
2. Goong 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 specic. 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" identied in that a tease can not be formally identied 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 identied 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
12
2.7. THE FUNCTIONS OF HUMOUR 13
13
14 CHAPTER 2. REVIEW OF HUMOUR RESEARCH
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:
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 specically with humour and gender.
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 concen-
trate on more formal speech or speech collected in articial 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 var-
ious 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 fa-
cilitative, 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 con-
tribute 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:
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 em-
phasised 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 ac-
commodate 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 signicant 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).
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 signicant, 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 specic 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 pos-
sibilities predominant in their own gender-specic 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 dif-
ferent but equal cultures approach (Fishman 1983, Henley and Kramarae 1991).
A purely power based approach claims men's conversational dominance, and dif-
ferences reected in speech functions can be explained by men's superior social
power.
Cameron (1992) nds dominance and di erence both unsatisfactory explana-
tions, and points out that what one believes is not so much a function of lin-
guistic 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 subscrib-
ing 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 miscommu-
nication 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 dif-
ferently (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 po-
sition in society. Uchida (1992) points out that there must be other intervening
and contributing factors. Race, age, occupation, education, context and numer-
ous 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 miscommuni-
cation, 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.
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 be-
tween the stereotypically feminine and the stereotypically masculine. Among
the variables claimed to contribute to gender are sexual orientation, masculin-
ity/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.
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)
Some researchers have attempted a justication 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:
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:
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).
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 in-
volved 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 mis-
takenly conclude that impressions of humourless women are substantiated by
fact. Cox, Read and van Auken (1990) looked at gender di erences in commu-
nicating job-related humour. The study is triggered by the concern that:
26
3.5. GENDER AND HUMOUR 27
This is an interesting area, and one which clearly warrants some empirical investi-
gation. 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 cir-
cumstance. 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 signicantly more, and
women were signicantly 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 fe-
males (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 (spon-
taneous, 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 signicantly more than women. Sixty-two of 92
females wrote about males, and 41 of 49 males wrote about males. When writ-
ing about males, respondents were signicantly 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 sub-
jects 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 signicant 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 signicant 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 gynae-
cologists 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 scientic knowledge.
Female gynaecologists used it to cut down discussions not considered relevant.
Ervin-Tripp and Lampert (1992) also investigated humour in naturally occur-
ring 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
30
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.
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 hu-
mour 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
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 gen-
erally 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)
humour to claim an audience and gain status. Women's humour is more context-
dependent, involves their audience, and is more supportive and personal.
33
Chapter 4
Methodology
\Supposing we hit him by mistake?" said Pooh anxiously. \Or sup-
posing 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 out-
lines 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 out-
comes of the taping which indicate that the methodology used may be slightly
biased.
35
36 CHAPTER 4. METHODOLOGY
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 side-
taking 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 mem-
bers 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 de-
cided 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 re-
search 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-identied as Pakeha except
one, who classed himself as Maori/European. This speaker identied 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.
37
38 CHAPTER 4. METHODOLOGY
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 signicantly
increase the naturalness of the conversation, and make the speaker feel as relaxed
as possible.
38
4.3. DATA COLLECTION 39
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 knowl-
edge 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 appro-
priate opportunities arose. This technique allowed me to ll the gaps identied
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.
39
40 CHAPTER 4. METHODOLOGY
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 Cor-
pus 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.
41
42 CHAPTER 4. METHODOLOGY
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 tape-
recorder, 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]\\
they have signed leads to more coping humour { wordplay based on the word
\corpus."
(2)
All: /laugh]\\
CM: ( )
AM: yeah
43
44 CHAPTER 4. METHODOLOGY
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)
(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)
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 chil-
dren were left to talk. The boys exhibited considerable discomfort with the
situation, whereas the girls had no problem nding something to talk about.
45
46 CHAPTER 4. METHODOLOGY
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)
All: laugh]
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 47
DM: yawns]
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
47
48 CHAPTER 4. METHODOLOGY
PM: yeha]ah
The examples discussed in this section are representative of the comments di-
rected 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)
CM: effectively
AM: ah
PM: mm mm tut]
All: laugh]
48
4.5. METHODOLOGICAL PROBLEMS 49
All: laugh]
49
Chapter 5
Statistical methods
Now it happened that Kanga had felt rather motherly that morning,
and Wanting to Count Things...
51
52 CHAPTER 5. STATISTICAL METHODS
52
5.1. LOG-LINEAR MODELLING 53
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:
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.
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
54 CHAPTER 5. STATISTICAL METHODS
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.
54
5.1. LOG-LINEAR MODELLING 55
therefore x the hair-eye conguration (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
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 signicantly increasing the likelihood ratio, then the simplied
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 signicant 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
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-simplied 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 signicant we look at the di erence between the two chi-square gures
and check its signicance 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 signicance of this to 2 (3-1) degrees of
freedom. This calculation shows that the fuller model is not signicantly 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.
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 get-
ting burnt, given they have blue hair is 470/(180+470) = 0.72. The conditional
probability gures are shown in table 5.8.
57
58 CHAPTER 5. STATISTICAL METHODS
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 innity. 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 back-
ground, 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.
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.
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 fulll them. In chap-
ter 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 sup-
port 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 unied 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 specically
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 unied the discussions and
classications into a small taxonomy of types. This taxonomy was then modied
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
identied 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
articially elicited data, participant observation, or introspection.
64
6.2. THE TAXONOMY OF TYPES 65
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 identied 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
signicantly a ect the results (see section 5.2.3).
65
66 CHAPTER 6. TYPES OF HUMOUR
the experiences or actions of either the speaker or someone they are acquainted
with. (9) is an example of a short anecdote.
(9)
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 par-
ticularly 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)
All: /laugh]\\
CM: =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 ( )
++
DM: /( )\\
67
68 CHAPTER 6. TYPES OF HUMOUR
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 denitely 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 claried by
exemplication. In (11), TM suggests a reason why people have not frequented
a restaurant, despite rave reviews.
68
6.3. DEFINING THE TYPES 69
(11)
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
69
70 CHAPTER 6. TYPES OF HUMOUR
PM: yeha]ah
As in the above extract, fantasy humour will typically involve a lot of collabora-
tive humour.
6.3.3 Insult
An insult is a remark that puts someone down, or ascribes a negative charac-
teristic 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)
70
6.3. DEFINING THE TYPES 71
(14)
6.3.4 Irony
This category includes examples of both irony and sarcasm. Denitions 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 denitions (Roy 1978 as cited in Tannen 1984),
and as Tannen (1984:131) points out \to arrive at a satisfying denition of irony
would require a major study in itself." The denition 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
72 CHAPTER 6. TYPES OF HUMOUR
All: /laugh]\\
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
72
6.3. DEFINING THE TYPES 73
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 specically 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)
AM: /inside\\
73
74 CHAPTER 6. TYPES OF HUMOUR
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)
SF+RF: /laugh]\\
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)
All: laugh]
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 benet, as he has recently seen the Muppets episode, and he
and DM saw LA Story together.
(20)
75
76 CHAPTER 6. TYPES OF HUMOUR
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 specic 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) identies mimicry as a type of humour, although this
is restricted to the mimicry of a specic person. Example (21) is an example of
the mimicry of a specic 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 77
GM: imitates dave]: fuck i'm going to waste you today: roleplay]
EM: imitates dave]: YEAH you know where i'm coming roleplay]
from:
77
78 CHAPTER 6. TYPES OF HUMOUR
All: /laugh]\\
VF: =bh]it\....
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 classication 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 justied. 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)
DM: =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: associative databases=/
NM: /=deductive=/
MM: /=deDUCtive
databases //that's the ones yeah\
79
80 CHAPTER 6. TYPES OF HUMOUR
deduced that
MM: groans]
DF: pulses what like beans? like beans? you mean wordplay]
BF: NO
CF: /yeah\\
80
6.3. DEFINING THE TYPES 81
(27)
GM: yeah
DM: =song=/
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
82 CHAPTER 6. TYPES OF HUMOUR
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 signicantly 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 MC FS MS
anec 0.43 0.39 0.33 0.29
fant 0.28 0.31 0.23 0.25
insult 0.04 0.03 0.09 0.06
irony 0.06 0.06 0.07 0.07
joke 0.00 0.00 0.00 0.00
obs 0.09 0.05 0.11 0.06
quote 0.00 0.00 0.01 0.03
role 0.03 0.05 0.06 0.10
self 0.00 0.00 0.01 0.01
vulg 0.00 0.00 0.01 0.02
word 0.05 0.07 0.04 0.06
other 0.03 0.03 0.04 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.
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 interac-
tion 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 provid-
ing. 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 en-
joy 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 signicantly 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, name-
calling 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 in-
sults, such as those identied 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 identied 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 signicant 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 conrmed
nor denied on the basis of my data. The one joke that did occur was in an all-
male 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
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 signicant. 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 interac-
tion. 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 identied patterns in the use of various humour
types, and so provides a more solid base for future research and hypothesis
building. Statistically, the most signicant 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 conversa-
tions 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 specically
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 classied 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
92 CHAPTER 7. FRESH HUMOUR AND SECONDS HUMOUR
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, remind-
ing 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=/
92
7.1. BACKGROUND 93
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 specically 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: /=yeah=/
93
94 CHAPTER 7. FRESH HUMOUR AND SECONDS HUMOUR
CM+DM: /=laugh]=/
ALL: /laugh]\\
AM:
/=laughs] and then he strangled him to death=/
DM+CM:
/=laugh]=/
BM: /=laugh]=/
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.
FC MC FS MS
external 0.04 0.07 0.04 0.07
fresh 0.96 0.93 0.93 0.86
recycled 0.00 0.00 0.03 0.03
joint 0.00 0.00 0.00 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.
FC-MC MC-FC FC-FS FS-FC MC-MS MS-MC FS-MS MS-FS
external - 1.81 1.00 1.00 1.00 1.00 - 1.81
fresh 1.81 - 1.81 - 2.16 - 2.16 -
recycled * * * * * * 1.00 1.00
joint * * * * * * * *
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
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 dened and discussed the di erent types of humour. Each type
of humour can be used for strategies which can fulll a number of functions. In
fullling the function of the humour, the speaker constructs certain aspects of
their identity. In this chapter I will outline a classication 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.
97
98 CHAPTER 8. THE FUNCTIONS OF HUMOUR
General
Functions
Defend Cope
General Situational
Strategies
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 classied fur-
ther 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 fulll. 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 specic functions. I do not
identify specic strategies through which the psychological functions could be
fullled. Strategies such as putting oneself down before someone else does, or
making light of a serious situation would serve to fulll psychological functions.
As the psychological categories are already quite specic 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 specic
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 fulll this function. The same is true of humour which functions specically
to exert power.
The diagram makes the interrelation of the di erent levels, strategies and func-
tions 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 articial. The use of the tree to rep-
resent the taxonomy also implies that each example can be identied as tting
into one and only one of the categories. In fact, it is possible for an instance of
humour to simultaneously full several of the identied functions. One example
could serve both psychological and solidarity functions.
In the following sections I dene 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 identies 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 identied. Such examples are labelled
99
100 CHAPTER 8. THE FUNCTIONS OF HUMOUR
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: /they were\\ great
LM: oh deah]r
RF: but other than that ha] it was warm and you
could wear it over anything
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\
CM: /=god
that's NOTHING when I did three oh nine
101
102 CHAPTER 8. THE FUNCTIONS OF HUMOUR
JF: UGH
102
8.4. POWER 103
JF&NF: laugh]
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 friend-
ship groups, it is not surprising that a large percentage of the instances of humour
103
104 CHAPTER 8. THE FUNCTIONS OF HUMOUR
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.
AF: /=mm
DF: /but=
104
8.4. POWER 105
AF: /=mm
AF: mm major
DF: /yeah\\
AF: /yeah\\
CM: /laughs]\\
105
106 CHAPTER 8. THE FUNCTIONS OF HUMOUR
BM: /laughs]\\
8.4.2 To control
Any humour intended to inuence the behaviour of the audience would be an ex-
ample of control. This function is identied 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 inu-
ence 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 control-
ling 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
106
8.4. POWER 107
AF: /=do
you=/
AF: /( )\\
AF: /=really=/
CF:
/=yeah=/
107
108 CHAPTER 8. THE FUNCTIONS OF HUMOUR
DT: yeah yeah //yes i set the video wrong give me=
DT: =a break\
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)
109
110 CHAPTER 8. THE FUNCTIONS OF HUMOUR
(40)
All: laugh]
The issue of humour relating to the tape-recorder is dealt with in detail in sec-
tion 4.5.
WF: /oh\\
110
8.6. RESULTS 111
TF: =# r # d i iMAgine
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 specic 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 specically 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.
1 12 13 23 3 7.93 0.0475
2 12 23 6 34.77 0.0000
3 12 13 6 48.45 0.0000
4 12 9 75.27 0.0000
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 signicantly 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 signicance
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 signicantly better than the other models { to the .001 level.
The probabilities predicted by this model are shown in table 8.2. These prob-
abilities 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.
FC MC FS MS
General 0.25 0.44 0.13 0.26
Solid 0.59 0.36 0.65 0.45
Power 0.10 0.13 0.12 0.17
Psych 0.06 0.07 0.11 0.13
Table 8.2: Probability gures for overall function
We can spot some general trends from investigating the probabilities, but cal-
culation 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 fe-
males and males. The odds ratios for the four overall functions are shown in
table 8.3
113
114 CHAPTER 8. THE FUNCTIONS OF HUMOUR
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 115
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 indi-
cation 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
clarication 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 solidar-
ity 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
115
116 CHAPTER 8. THE FUNCTIONS OF HUMOUR
FC MC FS MS
boundP 0.07 0.06 0.09 0.06
teaseP 0.07 0.06 0.12 0.11
other 0.02 0.01 0.00 0.00
Table 8.6: Probabilities for power based strategies
cases this did not a ect things signicantly, 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.
FC MC FS MS
defend 0.02 0.02 0.03 0.03
cope-gen 0.02 0.01 0.04 0.03
cope-sit 0.02 0.04 0.03 0.07
Table 8.8: Probabilities for psychological functions
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 identied 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.
118
8.9. DISCUSSION 119
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
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-specic 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
122 CHAPTER 8. THE FUNCTIONS OF HUMOUR
8.9.3 Power
Much of the language and gender literature presents power and solidarity as di-
chotomous. Men, in general, are said to place importance on power and competi-
tion 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 conver-
sations. 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.
8.10 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
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
126 CHAPTER 9. FOCUS AND TOPIC
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 signicant 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 classication of my examples into focus categories
and outline the results.
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 signicant, and signicantly
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 signicant results. They are
shown in table 9.2.
126
9.2. FOCUS 127
FC MC FS MS
abs 0.25 0.24 0.18 0.15
self 0.16 0.15 0.20 0.17
present(m) 0.15 0.03 * 0.30
present(f) 0.03 0.10 0.30 *
other 0.40 0.48 0.32 0.38
Table 9.1: Probability gures for focus of humour
127
128 CHAPTER 9. FOCUS AND TOPIC
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 signicant in my data, however, and shows
only as a possible trend. Similarly their observation that humour focussing on
someone absent was more prolic in mixed groups is supported by only a slight
trend in my data. It is important to note, however, that Ervin-Tripp and Lam-
pert'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:
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 prolically 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, particu-
larly 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
130 CHAPTER 9. FOCUS AND TOPIC
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
132 CHAPTER 9. FOCUS AND TOPIC
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 benet
from in-depth qualitative investigation. In section 9.3.1 I address the topics I feel
would benet 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 specic 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 specic 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.
133
134 CHAPTER 9. FOCUS AND TOPIC
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 tra-
dition of physical strength, mateship, drinking and gambling. He describes the
pioneers in New Zealand, claiming:
Drinking was without doubt the most important and dening 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.
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.
135
136 CHAPTER 9. FOCUS AND TOPIC
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 method-
ology 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 pro-
137
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 fre-
quent 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 labora-
tory 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 conver-
sation, 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 dif-
ferent 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
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 occur-
rence/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 mem-
ber 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 conver-
sation. 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 con-
versations 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
Low High
FS 13 11
MS 12 12
FC 8 4
MC 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 denite 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 in-
vestigating the models, I was unable to reject the independence model as the
simplest, well-tting model. These di erences then, while not signicant, show
a denite 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 per-
centages 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
142 CHAPTER 10. AMOUNT OF HUMOUR
60
FC
50
MC
40
percentage
30
20
10
0
A B C D E F G H I J K L
individual
60
MS
50 FS
40
percentage
30
20
10
0
O
A B C D E F G H I J K L M N P Q R S T U V W X
individual
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). Al-
though women do contribute less humour in mixed groups than men, they still
contribute a signicant 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 na-
ture of the personal relations between them" (Tajfel 1979:401).
Intergroup relations, on the other hand, are \interactions which are largely de-
termined 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 denition 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 iden-
tity. This motivates a search for positive distinctiveness for one's own group in
145
146 CHAPTER 11. SOCIAL AND PERSONAL IDENTITY
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. Signicant gender di er-
ences 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 poten-
tially 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.
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]\
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
WF: /=yeahha
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
150 CHAPTER 11. SOCIAL AND PERSONAL IDENTITY
and comments that he was the only one she wanted in her personal space.
(44)
NF: oh really=/
In this extract certain aspects of her identity are highlighted. Among other
things, she is displaying openness and honesty in discussing this personal in-
formation 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 fre-
quently 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)
151
152 CHAPTER 11. SOCIAL AND PERSONAL IDENTITY
BM: =stayed
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.
A signicant part of one's personal identity, then, is one's gender identity. The
creation of a positive personal identity occurs constantly throughout all interac-
tion, 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 liter-
ature 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 de-
scribes 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.
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 in-
sists 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.
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). Makri-
Tsilipakou (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.
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.
This model ts well. The likelihood ratio chi square gure falls easily below
the 7.82 mark for 3 degrees of freedom to .05 signicance. 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:
The di erence between the two models is not statistically signicant, 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 signicance. The model of main e ects for group composition
is only signicantly 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.
% no.
F - single 71.76 155
M - single 73.27 244
F - mixed 72.82 75
M - mixed 80.98 132
Table 12.1: Percentage of humour supported by laughter
I did not note by whom the laughter was supported, although one could hypoth-
esise 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)
BM: or a television
AF+JF: /laugh]\\
All: laugh]
160
12.3. COUNTING LAUGHTER 161
AF+JF:laugh loudly]
BM: yeah
AF: h h huh]
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 con-
clude 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 compet-
ing, 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 alter-
native 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.
162
12.4. HUMOUR SUPPORT STRATEGIES 163
MM: /=yeah
TM: 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
NM: /=yes
Sometimes maintaining the humorous frame, or playing along with a \gag" ini-
tiated 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 re-
ply, 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)
BF: NO
CF: /yeah\\
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 sympa-
thy for his upcoming \ordeal". No laughter occurs, yet the humour is adequately
164
12.4. HUMOUR SUPPORT STRATEGIES 165
supported.
(49)
Fantasy humour is often supported by more fantasy. Speakers will jointly con-
struct 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)
165
166 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES
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 savour-
ing the humour. In (51) AM repeats CM's words in a tone that indicates he
appreciates them and nds them funny.
(51)
CM: /laughs]\\
166
12.4. HUMOUR SUPPORT STRATEGIES 167
RM's humour is supported through laughter, but LM also supports the humour
by echoing RM's words.
167
168 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES
BF:
/=yeah=/
WF: /=yeah=/
In (55) BM tells an anecdote about his ears, and AF and DF assure him that his
ears look okay.
(55)
DF: /yeah\\
168
12.4. HUMOUR SUPPORT STRATEGIES 169
BM: no no
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!
VF: world/ly\\
RF: /ex\\perienced
12.4.5 Non-verbal
It is also possible to support humour non-verbally { usually with a smile. Un-
fortunately this can not be observed in my taped data.
169
170 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES
DF: //yeah\
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 clarication, and once a relatively obscure explanation was received said:
(58)
The other speaker then continued with his story. Support was not o ered, nor
expected. TM was merely using irony to make a point.
(59)
HM: /sighs]\\
172
12.6. FAILED HUMOUR 173
DM: was=/
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.
173
174 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES
174
12.6. FAILED HUMOUR 175
AF: mm
175
176 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES
SF:
/=yeah it was in geography=/
176
12.6. FAILED HUMOUR 177
177
178 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES
LM: /=yeah
LM: /=h
huh]=/
178
12.6. FAILED HUMOUR 179
BM: /=yeah
179
180 CHAPTER 12. LAUGHTER AND HUMOUR SUPPORT STRATEGIES
CM: /=mm=/
BM: /yeah\\
BM: /=yeah=/
DF: /=yeah
All: laugh]
180
12.7. CONCLUSION 181
The group reminisces about experiences they had with nitrous oxide at the den-
tist. 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.
12.7 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, unpre-
dictable. 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 Intro-
duction, 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 dis-
ciplines. The terminology in this area is often confused, ill-dened and incon-
sistent. We are lacking unied 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. Re-
searchers 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
184 CHAPTER 13. CONCLUSION
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 log-
linear 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 modied 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 signicant
results from this part of my research were:
There were also a number of trends which are interesting, but not statistically
signicant:
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
186 CHAPTER 13. CONCLUSION
186
187
187
188 CHAPTER 13. CONCLUSION
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 Victo-
ria University for the Wellington Corpus of Spoken New Zealand English (AC-
CENZ).
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
189
190 APPENDIX A. KEY TO TRANSCRIPTION CONVENTIONS
AM: sneezes]
Pauses
+ pause of up to one second
++ pause of up to two seconds
190
191
Laughter
$h] laughing exhalation
$huh] laughing inhalation
$ha] voiced laugh particle
$nh] nasalised laugh particle
hello$ho] laughing repetition of syllable
$laughs] 2 secs 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: conversa-
tional 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 sem-
inars? Status, culture and gender at Durham University. In Clark, R.,
Fairclough, N., Ivanic, R., McCleod, N., Thomas, J. and Meara, P., edi-
tor, 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 character-
istics of male and female speech. Journal of Social Psychology, 103:129{146.
193
194 BIBLIOGRAPHY
194
195
195
196 BIBLIOGRAPHY
196
197
197
198 BIBLIOGRAPHY
198
199
Lako , R. (1975). Language and woman's place. Harper Colophon, New York.
Lave, J. and Wenger, E. (1991). Situated Learning: Legitimate Peripheral Par-
ticipation. 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 Mean-
ing. 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 Uni-
versity Press, Cambridge.
Marlowe, L. (1989). A sense of humour. In Unger, R. K., editor, Representa-
tions: Social Constructions of Gender, pages 145{154. Baywood Publishing
Company inc., Amityville, NY.
199
200 BIBLIOGRAPHY
200
201
203
204 BIBLIOGRAPHY
204