Peyman G. P. Sabet, Grace Q. Zhang (Auth.) - Communicating Through Vague Language - A Comparative Study of L1 and L2 Speakers-Palgrave Macmillan UK (2015)

Communicating through Vague Language
Also by Peyman Sabet

15 M.A. PARALLEL TESTS IN TEFL FOR EXAMINATIONS
A Book for M.A. Candidates in TEFL (co-authored)
EFFECTIVE READING FOR COMPREHENSION (co-authored)
Also by Grace Zhang

ELASTIC LANGUAGE
How and Why We Stretch Our Words
CRITICAL PERSPECTIVES ON LANGUAGE EDUCATION: AUSTRALIA AND
THE ASIA PACIFIC (co-edited)
NEGOTIATING WITH VAGUE LANGUAGE
A Chinese Perspective (co-authored)
USING CHINESE SYNONYMS
REQUEST STRATEGIES
A Comparative Study in Mandarin Chinese and Korean (co-authored)
FUZZY SEMANTICS (2nd edition)
Communicating through
Vague Language
A Comparative Study of L1 and
L2 Speakers
Peyman G.P. Sabet and Grace Q. Zhang

Curtin University, Australia
© Peyman G.P. Sabet and Grace Q. Zhang 2015
Softcover reprint of the hardcover 1st edition 2015 978-1-137-48637-0
All rights reserved. No reproduction, copy or transmission of this
publication may be made without written permission.
No portion of this publication may be reproduced, copied or transmitted
save with written permission or in accordance with the provisions of the
Copyright, Designs and Patents Act 1988, or under the terms of any licence
permitting limited copying issued by the Copyright Licensing Agency,
Saffron House, 6–10 Kirby Street, London EC1N 8TS.
Any person who does any unauthorized act in relation to this publication
may be liable to criminal prosecution and civil claims for damages.
The authors have asserted their rights to be identified as the authors of this
work in accordance with the Copyright, Designs and Patents Act 1988.
First published 2015 by
PALGRAVE MACMILLAN
Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited,
registered in England, company number 785998, of Houndmills, Basingstoke,
Hampshire RG21 6XS.
Palgrave Macmillan in the US is a division of St Martin’s Press LLC,
175 Fifth Avenue, New York, NY 10010.
Palgrave Macmillan is the global academic imprint of the above companies
and has companies and representatives throughout the world.
Palgrave® and Macmillan® are registered trademarks in the United States,
the United Kingdom, Europe and other countries.
ISBN 978-1-349-55661-8 ISBN 978-1-137-48638-7 (eBook)
DOI 10.1057/9781137486387
This book is printed on paper suitable for recycling and made from fully
managed and sustained forest sources. Logging, pulping and manufacturing
processes are expected to conform to the environmental regulations of the
country of origin.
A catalogue record for this book is available from the British Library.
Library of Congress Cataloging-in-Publication Data
Sabet, Peyman G.P., 1976–
Communicating through vague language : a comparative study of L1 and L2 speakers /
Peyman G.P. Sabet, Curtin University, Australia; Grace Q. Zhang, Curtin University,
Australia.
pages cm
Summary: “Vague language refers to expressions with unspecified meaning (for
instance, ‘I kind of want that job’), and is an important but often overlooked part of
linguistic communication. This book is a comparative study of vague language based
on naturally occurring data of a rare combination: L1 (American) and L2 (Chinese and
Persian) speakers in academic settings. The findings indicate that L2 learners have
diverse and culturally specific needs for vague language, and generally use vague
words in a more concentrated fashion compared with L1s. The book explores the
interconnection and continuum between the linguistic realizations of a vague term
and the functions it serves, and how ‘elasticity’ underpins the frequency and functions
of vague language. The implication is that vague language may be integrated into
the curriculum of English language teaching as part of L2 speakers’ communicative
competence. This book will be of great interest to researchers in the fields of
intercultural education and applied linguistics”— Provided by publisher.
1. Semantics. 2. Vagueness (Philosophy) 3. Ambiguity. 4. Language and
languages—Philosophy. I. Zhang, Grace Qiao, author. II. Title.
P325.S123 2015
401'.43—dc23 2015018336
Typeset by MPS Limited, Chennai, India.

Contents
List of Figures and Tables viii

Acknowledgments x
Transcription Conventions xi
List of Abbreviations xii
1 Introduction 1
1.1 Purpose of the study 2
1.2 Organization 3
2 Theoretical Foundations 4
2.1 Vague language: what is it and why do we need it? 5
2.2 Vague language in different settings 11
2.3 Vague language and education 15
2.3.1 Vague language in the classroom 15
2.3.2 Vague language in language learning and teaching 17
2.3.3 Vague language and learner language 19
2.4 Theoretical frameworks 23
2.4.1 Vague language and the cooperative principle 23
2.4.2 Vague language and Relevance Theory 26
2.4.3 The elasticity of vague language 29
2.5 Concluding remarks 31
3 Methodology 33
3.1 Three approaches 33
3.1.1 Quantitative approach 33
3.1.2 Qualitative approach 35
3.1.3 Mixed methods approach 38
3.2 Naturally occurring data 41
3.3 Data 42
3.4 Data analysis 44
4 Lexical Analysis of Vague Language 47
4.1 Subjectivizers 47
4.1.1 I think 48
4.1.2 I guess 53
v
vi Contents
4.1.3 I don’t know 54

4.1.4 I believe 55
4.2 Possibility indicators 58
4.2.1 Maybe 59
4.2.2 May 63
4.2.3 Might 65
4.2.4 Probably and possible 66
4.3 Vague quantifiers 69
4.3.1 Some 71
4.3.2 Much 74
4.3.3 Many 75
4.3.4 A lot of 77
4.3.5 Most 77
4.3.6 (A) few 79
4.3.7 A little 80
4.3.8 Lots of 80
4.3.9 A lot 80
4.3.10 Majority 81
4.4 Vague intensifiers 84
4.4.1 Really 85
4.4.2 Very 88
4.4.3 Actually 92
4.4.4 So 94
4.4.5 Too 95
4.4.6 Quite 96
4.5 Placeholders 99
4.5.1 Something 99
4.5.2 Things 102
4.5.3 Thing 104
4.5.4 Anything 106
4.5.5 Someone 107
4.5.6 Somebody 108
4.5.7 Anybody 109
5 Pragmatic Functions of Vague Language 113
5.1 Mitigation 114
5.1.1 Self-protection 115
5.1.2 Politeness 118
5.1.3 Downtoning 123
Contents vii
5.2 Right amount of information 126

5.2.1 Quantification 126
5.2.2 Emphasizing 129
5.2.3 Possibility 132
5.2.4 Uncertainty 136
5.3 Structural function 138
5.3.1 Repair 139
5.3.2 Hesitation 143
5.3.3 Turn management 146
6 General Discussion 151
6.1 Overall frequency distribution 151
6.2 Clustering patterns 153
6.3 Density of vague language and concentrated distribution 160
6.4 Influence of first language 162
6.5 Influence of cultural protocols 169
6.6 Impact of lack of language proficiency 173
6.7 Impact of cognitive processing focus 175
6.8 Different communicative approaches 178
6.9 The manifestation of elasticity of vague language 181
6.9.1 Linguistic elasticity 182
6.9.2 Pragmatic elasticity 186
6.9.3 Versatility between realizations and functions 187
7 Conclusions and Implications 189
7.1 Conclusions 190
7.2 Limitations 194
7.3 Implications 194
7.4 Further research 196
Appendix I: Consent Form for the Director and Teachers 198
Appendix II: Consent Form for Participants 199
References 200
Subject Index 209
Author Index 211

List of Figures and Tables
Figures
4.1 Frequency distribution of subjectivizers 56

4.2 Percentage distribution of subjectivizers 57
4.3 Frequency distribution of possibility indicators 68
4.4 Percentage distribution of possibility indicators 69
4.5 Frequency distribution of vague quantifiers 82
4.6 Percentage distribution of vague quantifiers 83
4.7 Frequency distribution of vague intensifiers 97
4.8 Percentage distribution of vague intensifiers 98
4.9 Frequency distribution of placeholders 109
4.10 Percentage distribution of placeholders 110
6.1 Positions of the three groups in the use
of five lexical categories 154
6.2 Overall frequency continuum of possibility indicators 185
6.3 Elastic continuum of maybe 185
6.4 Overall frequency continuum of placeholders 186
6.5 Elastic continuum of things 186
Tables
3.1 Differences between quantitative and qualitative methods 35

3.2 Vague language lexical categories 44
3.3 Vague language pragmatic function categories 45
4.1 Distribution of subjectivizers 48
4.2 I think positions and clusters 52
4.3 Distribution of possibility indicators 58
4.4 Maybe positions and clusters 62
4.5 May clusters 64
viii
List of Figures and Tables ix
4.6 Might clusters 65

4.7 Distribution of vague quantifiers 70
4.8 Some clusters 71
4.9 Much clusters 74
4.10 Many clusters 76
4.11 Distribution of vague intensifiers 84
4.12 Really clusters 86
4.13 Very clusters 89
4.14 Distribution of placeholders 99
4.15 Something clusters 100
4.16 Things clusters 103
4.17 Thing clusters 105
4.18 Anything clusters 106
4.19 Frequency of lexical categories of vague language 111
6.1 Overall distributions of vague expressions 152
6.2 Ratio of vague expressions to total word count 161
Acknowledgments
We would like to express our sincere gratitude to Chris Conlan, the

anonymous reviewers and the final clearance reader; their valuable
feedback has been incorporated into this book. We thank Rebecca
Brennan and Libby Forrest at Palgrave Macmillan, Geetha Williams,
Portia Winters and anyone else involved in this project, for their help
with the production of this book. We are truly grateful to Margaret
Johnson for her skilful assistance with the editing.
Without the willingness of participants in the data collection, this
book would not have been possible. We thank wholeheartedly those
who participated in the data collection in Iran and China, and the
organiser Hongbian Feng.
We acknowledge, with gratitude, the research grant (AAPI and
humanities) and scholarships awarded by Curtin University, and the use
of the Michigan Corpus of Academic Spoken English.
Finally, we thank our parents and spouses for all the selfless support
and never-ending love they have given us.
x
Transcription Conventions
[ S1:] Overlap
<SU-f, SU-m> Unknown speaker with gender identified
<P: 04> Pause of 4 seconds
… Pause of 2–3 seconds
<LAUGH> Current speaker laughs
<S8 LAUGH> Speaker 8 laughs
<SS LAUGH> Two or more speakers laugh
(xx) Indecipherable
xi
List of Abbreviations
CANCODE Cambridge and Nottingham Corpus of

Discourse in English
CSLE Chinese-speaking learner of English
CTFV Contrary-to-face-value
EFL English as a foreign language
ELT English language teaching
ERIC Education Resources Information Centre
ESL English as a second language
LIBELCASE Limerick-Belfast Corpus of Academic Spoken English
L1s L1 speakers
MICASE Michigan Corpus of Academic Spoken English
PSLE Persian-speaking learner of English
SOV Subject-object-verb
VOA Voice of America
xii
1
Introduction
Vague language (VL) is an integral part of language and has an essential

role in effective communication. In this study, vague language is defined
as inexplicit expressions, which are used elastically to enrich commu-
nication. Vagueness has been considered a demerit by some, who judge
it an undesirable phenomenon and a negative feature of language. For
them, meaning is considered the core component of communication,
irrespective of the speakers’ intentions and contextual factors (Aristotle
1946; Plato 1914), and any kind of communication breakdown is a fail-
ure in the person’s ability to create a connection between the right word
and the right meaning (Lakoff & Johnson 1980). Others view vagueness
differently, and regard the appropriate use of vague language as part of
a speaker’s communicative competence (Channell 1994).
The study of vague language has gained popularity since Channell’s
(1994) semantic and pragmatic study. In recent years there has been
a paradigm shift from philosophy to linguistics in this field, and con-
sequently vague language use has come to be commonly perceived as
an integral component of language. This is obvious in Cutting’s (2007)
assertion that ‘VL is a central feature of daily language in use, both
spoken and written’ (p. 3). The most recent development is by Zhang
(2011), who proposes and develops the important concept of elasticity
in vague language, pointing to a new direction in research.
With research interest in vague language on the rise, this feature of
natural language has found its way into language teaching. Cutting
(2007) states, ‘Since the mid-1990s, a limited number of applied lin-
guistics and methodology books have begun to contain a discussion of
possible teaching techniques to raise students’ awareness of VL’ (p. 236).
As this makes clear, the quantity of books on the subject is few and the
focus restricted to awareness-raising during instruction. This study aims
1
2 Communicating through Vague Language
to give depth to the role of vague language in language learning and

teaching: the sources from which vague language originates in English
language teaching (ELT), how it can be positioned to provide language
learners with an extremely effective communication tool, and com-
mon functions that language learners use it for in communication. It
also sheds light on the appropriate criteria to be used as the basis for
consciousness-raising on vague language in ELT.
The scope of vague language is contained in the area of pragmatic
competence in language teaching. The small but growing body of lit-
erature indicates that instruction on pragmatic competence has proven
remarkably effective in language teaching (Liddicoat & Crozet 2001;
Niezgoda & Röver 2001; Ohta 2001). The present study adopts a vague
language perspective in cross-cultural and interlanguage pragmatics.
The naturally occurring data of this study comprise the classroom inter-
actions of three groups of speakers of English: L1s (American English),
Chinese-speaking learners of English (the Chinese) and Persian-speaking
learners of English (the Persian speakers). There are approximately 50,000
words from each group, making a total of 150,000 words for the data. The
L1 data are selected from transcripts of tutorials and small lectures on social
topics from the Michigan Corpus of Academic Spoken English (MICASE).
The Chinese data are a transcript of the video-recorded classroom interac-
tions of upper-intermediate to advanced level learners of English in China,
and the data for the Persian speakers are similar, but video-recorded in Iran.
The data are analyzed on two levels: lexical, to investigate the frequency ,
position of occurrence, and clusters of five vague items; and functional, to
examine the pragmatic properties of vague language.
1.1 Purpose of the study
Attention has been, and is still being, drawn to vague language use in set-
tings such as work-related interactions (Koester 2007), healthcare (Adolphs,
Atkins & Harvey 2007), and courtrooms (Cotterill 2007), but suffers a
dearth of research in academic contexts. While Ruzaitė (2007) carried out
a useful study in educational settings, the scope of her research is limited
to quantifiers and approximators. This research is one of the few investi-
gations of vague language in English language learning classes with stu-
dents from two vastly different socio-cultural and linguistic backgrounds,
Chinese and Persian-speaking learners of English, compared with L1s.1
1
In this study L1 refers to American English speakers, L2 to Chinese and Persian
learners of English. The equivalent terms NS (native speaker) and NNS (non-
native speaker) are kept if used in the work of other researchers.
Introduction 3
Learners of English tend to use vague language at higher or lower

levels and in different forms from L1s; and the ways in which they
mobilize it is also different (Cheng 2007; Cheng & Warren 1999). The
central research question of this study is: what are the different levels of
frequencies and forms of vague language used between L1 and L2 speak-
ers, the strategic moves, and the contributing cultural and linguistic
factors? Four research questions are addressed:
a. How is vague language realized among L1s, the Chinese and the
Persian speakers?
b. How frequently is vague language used, and what are the most flu-
ently used lexical items? How does frequency differ among the three
groups?
c. How and why is vague language strategically mobilized? What dis-
crepancies are discerned among the three groups?
d. What cultural and linguistic factors underlie interlinguistic and
intercultural diversity in vague language use?
The objectives of this study are to explore vague language realization

in terms of its diverse forms and frequencies across L1s, Chinese and
Persian speakers, in order to analyze its pragmatic functions and the
impact of the underpinning cultural and linguistic factors (such as first-
language transfer). The findings of this study will have implications for
developing an understanding of the effective use of vague language by
learners of English.
1.2 Organization
This book consists of seven chapters structured as follows: Chapter 2

provides a theoretical foundation through a review of previous studies
in the field. Chapter 3 presents the approach and methodology selected
for this study. The results obtained from the three data sets are pre-
sented in Chapter 4. Chapter 5 presents a functional analysis of vague
language, and a discussion of the results occurs in Chapter 6. Chapter 7
draws the conclusion and considers the implications of the study.
2
Theoretical Foundations
Vague language, as an increasingly explored phenomenon in language,

is gaining a long overdue position as a device to express imprecision
in academic discourse. The nature of academic discourse, contrary to
common belief, requires vague expressions, and teachers and students
tend to make extensive use of them to allow them to express degrees
of certainty about the strength of a statement (see, for example, Ruzaitė
2007).
There is sufficient evidence that imprecision—or vagueness—is an
integral part of academic discourse. Should it be omitted, communica-
tion would be adversely affected, because ‘Mastery of vague language
(both active and passive) is one measure of communicative competence
in a foreign or second language, particularly those aspects termed as
“strategic competence” and “sociolinguistic competence”’ (Cheng &
Warren 2001, p. 98). Tarnyikova declares ‘though relevant arguments
are mostly based on the precision of their wording, vague language
needs not necessarily be an “enemy” of sound argumentation, since
the deliberate refrain from being vague might result in a precise but less
polite or impolite interaction’ (2009, p. 129).
Research on vagueness shows that it is a feature of natural language,
serving various functions in communication. Channell (1994) believes
vague language cannot be assumed to be the exception rather than the
rule. Tarnyikova (2009, p. 119) considers vagueness strategies and mani-
festations of vague language to be ‘partly universal but to a considerable
degree language- and culture-specific’. Ruzaitė (2004, p. 220) asserts that
‘vagueness cannot and should not be avoided, since over precision can
lead to communicative breakdowns’.
Vague language, in a broad sense, is assumed to be more frequently
employed in spoken than written language (Biber, Johansson, Leech,
4
Theoretical Foundations 5
Conrad & Finegan 1999). The reasons that support this claim are that
in spoken discourse, interlocutors share contextual clues such as facial
expressions, which may not exist in written discourse, and less preci-
sion is required in spoken forms than written discourse (Cook 1989).
Speakers have access to intonation (Brazil 1997), which can help them
clarify what they mean by what they say. None of these reasons for
using vague language in speech implies that in written mode it should
be ignored or neglected (Hyland 1998). Myers (1996) points out that
research into academic discourse demonstrates that writers too use
vague language.
Vague language is elastic. Zhang (2011) states that its elasticity allows
it to stretch in many directions, as required by the interlocutor for
effective communication. The elasticity of vague language is a func-
tion of its lack of specific connotation, so that its interpretation relies
on context and communicative purpose. This allows language users to
make strategic use of vague language to mitigate, emphasize, evade and
so on (Zhang 2011).
2.1 Vague language: what is it and why do we need it?
Early work on vague language is associated with Bertrand Russell (1923),

who views vagueness from a philosophical point of view and argues that
vagueness or precision are nothing beyond representation: ‘Apart from
representation, whether cognitive or mechanical, there can be no such
thing as vagueness or precision; things are what they are, and there is an
end of it. Nothing is more or less what it is, or to a certain extent possessed
of the properties which it possesses’ (p. 85). He considers vagueness a
concept that may be applied to any kind of representation, such as a pho-
tograph: ‘a representation is vague when the relation of the representing
system to the represented system is not one–one, but one–many’ (p. 89).
After Russell, Zadeh (1965) addresses the question of whether con-
cepts in natural language are yes-or-no types or more-or-less types,
challenging the tenet of classical set theory which assumes that an ele-
ment either belongs to a set or it does not: for instance, in ‘John is old’,
according to classical theory, John is either old or he is not: being old
is not a relative concept. This perspective implies a clear-cut boundary
that makes concepts distinct in terms of their truthfulness, which is far
from reality in practice. Zadeh (1965) developed an alternative known
as fuzzy set theory, expressing a view that counts category membership
as a matter of degree rather than a clear-cut issue. Instead of being in the
set or not, an individual is in the set to a degree. Lakoff (1972, p. 458)
supports Zadeh’s theory, as ‘Clearly any attempt to limit truth condi-

tions for natural language sentences to true, false and “nonsense” will
distort natural language concepts by portraying them as having sharply
defined rather than fuzzily defined boundaries’.
There seems to be a hierarchical ranking to the truthfulness of a prop-
osition. Heider (1973) offers the idea of a distinction existing between
the central membership of a category and peripheral members, using
birdiness (defined typically as flying creatures with feathers) as an exam-
ple. A hierarchal order was established:
robins
eagles
chickens, ducks, geese
penguins, pelicans
bats
In this, some kinds of bird are more birdy than other kinds: robins are
regarded as typical of birds; eagles are less typical than robins; chickens,
ducks and geese less typical than eagles; penguins and pelicans less typi-
cal than chickens, ducks and geese, and bats are counted as hardly a
bird at all. Such a hierarchy is in line with the ‘prototype theory’ (Rosch
1973). The meaning of a vague term, such as birdiness, could be appro-
priately represented by a ranking exemplified by a prototype.
Vague language is defined by Carter and McCarthy (2006, p. 928) as
‘words or phrases with very general meanings which deliberately refer
to people and things in a non-specific, imprecise way’. Williamson
(1994, p. 4869) states that ‘Used as a technical term, “vague” is not
pejorative. Indeed, vagueness is a desirable feature of natural languages.
Vague words often suffice for the purpose in hand, and too much
precision can lead to time wasting and inflexibility’. In the same way,
vague language in this study is used without any negative connotation;
instead it is considered to be an important part of everyday language.
Various terms similar to vague language have been used in linguistic
studies, the most common being indirectness and inexplicitness. These
are not the same. Cheng and Warren (2003) propose a classification in
an attempt to clarify the confusion caused by similar terms, arguing that
‘indirect language’ involves an inferential process through which mean-
ing is created, while the hearer has access to language and context. This
term embodies paradigms such as conversational implicatures (Grice
1975), illocutionary acts (Austin 1962), indirect speech acts (Searle 1968)
and pre-sequences (Levinson 1983). Inexplicitness refers to cases such as
items of reference, that and it, which cannot stand on their own. They
are independent of context, but once used in a specific context they gain
a certain meaning: in other words, meaning is created through ‘joint
construction’ (Cheng & Warren 2003, p. 397) by the participants in the
context in which the words are used. Substitution, deixis, and reference
all fall into the category of inexplicitness. Inexplicitness is considered
a characteristic of a native speaker’s conversation, ‘the degree to which
linguistic behaviour is reliant on context to convey meaning’ (Cheng &
Warren 1999, p. 295). It emerges when a speaker chooses to use ellipsis
and substitution, deictic conversation and reference, relying on context
to convey the intended meaning. Vagueness differs from both indirect-
ness and inexplicitness in that even when used within a specific context,
its property of vagueness is retained. Vague language remains vague in
context, rather than becoming precise. As a linguistic phenomenon,
it is associated with fuzziness, imprecision, and indefiniteness (Ruzaitė
2007; Zhang 1998). Janicki (2002) opts for the phrase ‘incomprehensible
language’ as a broad term which embodies vague language, defining it as
‘words, expressions, formulations, idioms, texts, and the like which are
easy to misunderstand, which are hard to understand, or not possible to
understand at all’ (p. 215), and claims that ‘incomprehensible language’
appears consistently rather than sporadically in conversation. However,
this concept is not equivalent to vague language, which does not disrupt
ease of communication.
Channell (1994) presents cogent evidence that in order for communi-
cation to be effective and successful, speakers need to use vague words
and expressions at an appropriate level: in other words, they need to
be appropriately inexplicit. An important element in what constitutes
a speaker’s communicative competence is the use of vague language,
which is contextually appropriate and understandable. Channell also
states that vagueness in language is not a matter of badness or goodness,
but one of appropriateness, serving a variety of purposes:
giving the right amount of information,

deliberately withholding information,
using language persuasively,
displaying power,
being polite,
self-protecting,
establishing informality,
filling in lexical gaps and missing information
(Channell 1994, p. 194).
Not all vague words are equally vague: the boundaries of conceptual cat-
egories manifested through vague words are vague to different degrees,
with some more vague or more context-dependent than others (Ruzaitė
2007).
Basing their classification on degree of vagueness, linguists have
proposed different categories of vague language. An early version
presented by Crystal and Davy (1975) described it as 1. placeholders,
2. summarizing lexical items, 3. vague generic terms and collective nouns,
4. approximate quantities, 5. words with suffixes. A more recent classi-
fication proposed by Channell (1994) offered 1. quantifiers, 2. approxi-
mators, 3. placeholders, 4. vague references to categories. As with other
terms, there is some overlap and discrepancy in the terms offered by
linguists: what is called ‘vague references to categories’ by Channell is
called ‘general extenders’ by Overstreet (1995) and ‘summarising lexical
items’ by Crystal and Davy (1975), while Stenström (1944) uses the term
hedges to refer to approximators.
Vague language functions as a hedging strategy in making a claim.
Prince, Frader and Bosk (1982) state that ‘plausibility shields’ (p. 90), as
they call vague expressions, function to protect speakers from a full or
personal commitment regarding the truth condition of an utterance.
Similarly, Brown and Levinson (1987) claim that an application of
vague language can maintain the negative face of the interactants and
smooth the progress of a difficult conversation. Myers (1989) argues
that to be on the safe side when making claims with regard to new
research findings, authors of scientific writing employ hedges to indi-
cate a potential lack of certainty. Erev, Wallston and Neal (1991) likewise
find that vague communication in tasks which demand cooperation
between group members reinforces the sense of cooperation. Hamilton
and Minoe (1998, p. 6) maintain that ‘imprecise language can facilitate
a polite exchange between source and receiver. A precisely worded mes-
sage might come across as too personal, threatening a receiver’s self-
esteem. Vague language allows the preservation of face’. Along the same
line, Metsä-Ketelä (2006, p. 123) argues that ‘when it comes to interac-
tion, vague language functions as a marker of politeness and unreserved
atmosphere’. Given these many necessary functions, the investigation
of vagueness is as important as the study of preciseness.
There have been two studies of how vague language originates in
communication. The first focuses on language itself. Ullman (1962,
p. 118) refers to a number of factors of the origin of vague language
including ‘(a) generic character of words; (b) meaning is never homo-
geneous (that is, it is context-bound); (c) lack of clear-cut boundaries
in the non-linguistic world; (d) lack of familiarity with what the words
stand for’. He assumes of factor (a) that a word refers to a broad term,
one that is not a single entity but a class of items or events that have
some elements in common. In (b), meaning needs to be interpreted
with reference to context: that is, the context specifies meaning. In (c),
the world that the word represents is vague by nature. To consider how
this works, ‘ask oneself when a hill becomes large enough to qualify as
a mountain, or at what precise age a girl starts to be correctly referred
to as a woman’ (Channell 1994, p. 7). Factor (d) refers to uncertainty of
what is being talked about.
The second focus takes a psychological perspective. Deese (1974)
maintains that vagueness exists in the structure of ideas rather than
in the language system, arising from the ideas that express language
rather than from the language itself. Crystal and Davy (1975, p. 11)
put forward four reasons for vagueness, or ‘lack of precision’ as they
call it:
a. memory loss—the speaker forgets the correct word;

b. the language has no suitable exact word or the speaker does not
know it;
c. the subject of the conversation is not such that it requires precision,
and an approximation or characterization will do;
d. the choice of a vague item is deliberate to maintain the atmosphere.
Jucker, Smith and Lüdge (2003, p. 1765) believe that the most obvious
reason for vague language use is ‘uncertainty at the time of speaking.
Sometimes speakers lack information about a given quantity, quality
or identity. They, therefore, cannot be more precise even if they want
to’. Cutting (2007, pp. 3–16) claims that if speakers are exhausted or in
a hurry so that they can’t find the right word, a vague expression may
yield implications that are contextually more relevant for the hearer
than the exact words are.
Regardless of the lack of an agreed definition, various classifications
of vague language have been proposed. Walsh, O’Keeffe and McCarthy
(2008) divide vague categories into lexical and non-lexical types. Lexical
categories, or what Channell (1994, p. 123) calls ‘common categories’,
are items which have a graded structure and which can have a proto-
type, as we have seen with ‘bird’. Non-lexical categories, ‘vague cat-
egory markers’ (VCM) (Walsh et al. 2008) or ‘vague category identifiers’
(Channell 1994), refer to ad hoc items, which are the by-products of
interaction. Examples include vague tags such as (cloth) and that kind of
thing, (money) and things like that, which allow the listener to infer what
is meant by the speaker.
All languages, whether solely spoken or comprising both spoken and
written modes, use a variety of components to express vagueness, and
those used in speaking outnumber those used in writing (Metsä-Ketelä
2006). The rich verbal vagueness arises from the existence of semantic
vagueness, and also from concepts lacking clear definitions, thereby
necessitating imprecision. Take the distinction between a hill and a
mountain as an example: since there does not exist a clear-cut bor-
derline splitting these two concepts, a distinction between when a sea
becomes an ocean or a kitten becomes a cat seems highly unlikely. The
second factor contributing to imprecision in language is the existence of
phenomena such as metaphor, ellipsis, euphemism and pronoun refer-
ences (Metsä-Ketelä 2006). These two factors are two examples among
many other contributing factors for vagueness in language.
Precision can sometimes create inappropriateness, which a vague
expression can avoid (Jucker et al. 2003). Tannen (1989) claims an
increase of precision may have adverse effects. In other words, inap-
propriate use of detail can be boring, as can be witnessed when young
people listen to old people give more information about their daily rou-
tine than the hearers need to know. This can become insulting if used
in criticism. On the other hand, a precise statement can sometimes be
fuzzier than a vague statement. In his psychologically oriented study on
precision and vagueness in what he calls a preciseness paradox, Teigen
(1990) concludes that precise language suits any circumstance involv-
ing discussion in the past or present tense, but vague language is more
appropriate for any kind of future prediction. Past and present involve
more precision, but using precise language for the future can prompt
skepticism. For example, ‘Tomorrow at two o’clock the mountain will
erupt and 5,000 people will be killed’ vs. ‘That mountain is going to
explode any day now’. The latter statement is more believable than the
former one, because the latter resorts to vague language rather than pre-
cise language. Similarly, Moxey and Sanford (1993) state that ‘it would
appear that if one is looking for reasons to have faith in a proposition,
then specificity suggests expertise, which in turn meets that criterion. In
contrast, if one is looking for reasons to be sceptical, then precision may
signal suspicion’ (p. 16). However, Teigen rejects the trade-off between
confidence and skepticism, arguing that some features that consolidate
confidence can also contribute to doubts. It seems that the relation-
ship between confidence and precision or doubt and imprecision is not
straightforward, rather multi-faceted.
Vague language is primarily a strategic phenomenon with important

pragmatic functions (Channell 1994; Ruzaitė 2007); as will be demon-
strated in this study by the way it is pragmatically employed in aca-
demic settings.
2.2 Vague language in different settings
Vague language plays different roles in different settings. Focusing

on vague language across different spoken settings such as academic
discourse, business discourse, conversation and public discourse in
intercultural contexts, Cheng’s (2007) study reveals that the discourse
type rather than the speaker group determines the form and frequency
of vague language use. For instance, it is assumed that the legal system
requires maximum precision in its context, but in research on vague
language use in forensic situations, Cotterill (2007) finds that even
within this one context the role it plays can differ from one position
to another. For example, conducting an examination-in-chief, a bar-
rister may resort to vague language to suggest that he has not received
a precise enough account or sufficient details of the case from the wit-
ness. However, when a witness uses vague language, a cross-examiner
may call the witness’s credibility into question; use it to ‘represent an
opportunity for confrontation, since vagueness may be seen to stem
from witness failings in memory, expression or integrity in the eyes
of the cross-examiner. Exploitation of any of these shortcomings may
pay dividends in the destruction of the witness’s evidential credibility’
(Cotterill 2007, p. 112). Lakoff (1990) also points out that legal contexts
demand vague language, and that to some extent even laws need to be
ambiguous, as it is impossible to foresee all future potential applications
in as yet unimagined different contexts: thus, he implies, the use of
vague language reinforces the flexibility of laws.
A similar situation is found in medical settings. Adolphs et al. (2007)
state that vague language is frequently used by physicians and nurses
to provide patients with a clear and true description of their illnesses.
As an example, in a professional–patient consultation, a doctor may
turn to vague language in order to adjust abstruse language to suit a
non-specialist patient’s knowledge. Prince et al. (1982) note that in
medical settings the existence of vague language originates from a need
to express uncertainty: when physicians use it, it ‘demonstrates a schol-
arly orderliness in their representation of knowledge’ (Adolphs et al.
2007, p. 64). In other words, while talking of diagnoses and prognoses,
physicians need an inherent degree of uncertainty in their statements to
indicate that medical science does not yet have a thorough understand-
ing of such diseases. This demonstrates that the use of vague language in
such cases is an appropriate acknowledgment of their limitations,
rather than an indication of undesirable imperfection. Adolphs et al.’s
(2007, p. 74) analysis of vague language in UK’s National Health Service
direct phone-ins shows that it also mitigates the force of requests for
embarrassing information, or orders to undergo unpleasant regimens
of treatment. Since the consultations are conducted on the phone, the
discourse involves a higher level of vague language use, than in face-
to-face patient–nurse interactions, which demand less. Vague language
helps the invisible participants feel socially closer and, while giving
the patient a clear idea of ‘the serious nature’ of the topic, helps keep
the atmosphere relaxed. As Adolph et al.’s findings illustrate, vague lan-
guage serves specific and significant roles, evidence that it has special
communicative purposes in medical settings.
Factors such as speakers’ linguistic and cultural background also play
determining roles. In a study of native and non-native speakers, Cheng
and Warren (2001) find that rather than creating confusion and misun-
derstanding, vague language can, in a broad sense, enhance friendliness
and reinforce a ‘cooperative tone of exchange’, and create informality
in conversation. It also has ‘more specific uses, normally classification,
compensating for a lack of vocabulary (as an accommodation strategy
and as an avoidance strategy), compensating for a lack of knowledge,
politeness and finally “self-protection”’ (p. 86). An instance of accom-
modation strategy is when native speakers adjust their language to a
non-native speaker’s level by using short sentences, simple structures
and commonly used words: ‘varying the degree of specificity is one way
in which NS accommodate NNS, and whether this means using more or
less vague language will depend on contextual factors such as the NS’s
perception of the NNS’s linguistic ability’ (p. 94). This demonstrates
that vague language can be used to create a friendly atmosphere when
it is needed.
Another favorable function of vague language is that by enhancing
mutual tolerance between interlocutors, it makes the audience an active
participant in the process of communication. It, therefore, serves as a
hearer-involvement device (Ruzaitė 2007), using vague expressions to
urge active and attentive participation from the interlocutor to con-
struct the meaning of the message.
The study of vague language has a wide application in various lan-
guages. There have been important works based on Chinese language.
Wu’s (1979) seminal work introduced Zadeh’s (1965) fuzzy set theory
into Chinese linguistics. Wu argues that the concept of membership can

best represent vague expressions such as ‘tallness’ and ‘middle-aged’.
Chen and Wu (2002) investigate vagueness from the perspective of cog-
nitive linguistics, stating that linguistic categories are vague and should
not be made more precise than they really are. Zhang (1998–2004)
focuses on semantic and pragmatic vagueness in Chinese, arguing that
as vague language has no clear-cut referential boundary, it cannot be
contextually resolved. She adds that vague expressions play important
roles in communication.
There are a few studies on vague language in Persian. An investigation
of the frequency and grammatical distribution of general extenders (GE)
by Parvaresh, Tavangar and Eslami Rasekh (2010) finds that adjunctive
general extenders are more frequently used in Persian than disjunc-
tive general extenders: in other words, Persian speakers prefer general
extenders beginning with and to ones beginning with or. This is in line
with Cheshire’s (2007) finding on native British English, but in contrast
with Overstreet’s (2005) finding of native American English. Another
finding by Parvaresh et al. was that disjunctive general extenders in
Persian were less likely to occur after prepositional phrases, and that
‘both Persian and English disjunctive general extenders show smaller
variability of forms compared with their adjunctive counterparts’ (2010,
p. 33). By contrast, while Persian speakers demonstrated a tendency to
use general extenders both clause-finally and clause-internally, they
appeared in English in clause-final position only.
Another study, by Parvaresh, Tavangar, Eslami Rasekh and Izadi
(2012), considers Persian EFL learners’ use of vague language with
special attention to general extenders. They find that the use of clause-
internal general extenders in English results from the SOV order in their
L1 (Persian). This finding also reveals a transfer of a vague language
category from L1 to L2: ‘Non-native speakers defied the pattern con-
junction + noun phrase/determiner phrase + (like that) in such unique
general extenders as “and and and” and “and this and that”. This might
be attributed to transfer from Persian’ (p. 277). They reported a new
function of this vague category in the Persian corpus but missing in the
non-native speaker data: a general extender used by an interlocutor to
express outrage at what another interlocutor had mentioned. Unlike
native speakers of English who attached intensifying effects to general
extenders, the Persian speakers did not assign this function to the same
category of vague language, either in their L1 or in English as an L2. The
EFL group’s dominant use of disjunctive general extenders was the result
of uncertainty in word choice, which is a case that occurs with a very low
frequency in the Persian corpus. This present study employs a similar

participant group (Persian speakers) to that in Parvaresh et al.’s (2012)
study, and the addition of two other groups, the Chinese and L1s of
English, enables a wider comparison of how groups use vague language.
Two other Persian studies are of significance. Parvaresh and Tayebi
(2014) explore vague language structure and pragmatic use in Persian,
revealing some vague expressions unique to the language: ‘rhyming
words’, ‘replacing expressions’, and ‘affective completers’. A classifica-
tory model features cooperative functions (enough information pro-
vided, speaker’s own stance oriented or social engagement oriented)
and non-cooperative functions (not enough information provided).
Beighmohammadi (2003) investigates intensifiers in written Persian
across three different domains: science, social science and Teaching
English as a Foreign Language (TEFL). His data reveal that general
extenders occur twice as frequently in social science as in science and
TEFL; the explanation for this being that social science writers may
resort to discursive and rhetorical strategies when presenting what they
find, while the others focus more on reporting facts. Abdollahzadeh
(2003) finds no significant difference in the use of hedges between
Iranian and Anglo-American writers in his investigation of interper-
sonal metadiscourses, such as the discussion and conclusion sections
of ELT papers. Although these findings all compare Iranian language
users with L1s of English, there is an explicit need to take into account
inter-cultural and cross-linguistic factors in identifying the reasons for
culturally specific vague language phenomena, which is the focus of
this current study.
Other studies have been conducted of the similarities and differences
of vague language between two varieties of a language. In research on
approximators (for instance about 20) between American English (AE)
and British English (BE), Ruzaitė (2007) finds that there exist quantita-
tive and qualitative differences between the two varieties of English in
terms of approximators use. British English uses approximators much
more frequently than American English, implying that British English
speakers tend to be more vague. She suggests that American English
could be seen as too straightforward by British English speakers, and
British English speakers might be seen as too evasive by American
English speakers. Specifically, British English speakers used about fre-
quently, while around, approximately, and roughly were more popular
with American English speakers. This demonstrates how speakers
of two varieties of the same language used vague language differ-
ently. Ruzaitė concludes that even though English is spoken in both
these countries, cultural differences may cause discrepancies in their

patterns of vague language use.
Cultural and language differences seem to be a determining factor,
and more work is needed to explore the use of vague language in rela-
tion to cultural and linguistic backgrounds. This current research exam-
ines these factors in a bid to reveal previously unidentified tendencies
in vague language communication.
2.3 Vague language and education
The vague language works relating to education is reviewed in three

parts: vague language in the classroom, vague language in language
teaching and learning, vague language and learner language.
2.3.1 Vague language in the classroom

Several works explore the role of vague language in classroom settings.
Rowland (2007) finds that mathematics, which is concerned with
absolute precision by its very nature, also involves vague language use.
Participants in maths classes use hedges for making predictions and
generalizations, indicating that even in talking about facts, speakers
need to resort to vague expressions.
Adopting Walsh’s (2006) framework for classroom interaction (Self-
Evaluation of Teacher Talk), Walsh et al. (2008) examine vague category
markers such as exemplar + vague tag in academic spoken English
contexts. A classroom interaction framework comprises four modes:
managerial, material, skills and systems, and classroom context; a
mode is defined by Walsh et al. as a ‘classroom micro-context that has a
clearly defined pedagogic goal and distinctive interaction features deter-
mined largely by a teacher’s use of language’ (2008, pp. 62–63). This
cross-corpus study revealed that compared with casual conversation,
academic discourse involves less use of vague category markers, and
that those used have distinctive functions. For example, in managerial
mode the stage occurring at the commencement of each lesson consists
of one clause by the teacher and no student turn-taking. Frequent rep-
etition and the ‘handover’ to students come about in the end of this
stage. In the material mode, vague category markers are employed by
the teacher to ‘help expedite the start up phase of a lesson or activity,
since they can provide shortcuts that mark information or concepts
that can be taken as given, shared or unproblematic, they very quickly
set up what is common ground and facilitate a speedy handing over to
the task phase of the lesson’ (2008, p. 26). This stage encompasses the
teaching material or input used, where students answer questions and

their comprehension is checked. Walsh et al. are not able to identify
vague category markers due to the limited language used at this stage.
In skills and systems mode the interlocutors are involved in interaction
on the ‘core subject’ of the lesson; the goal is to familiarize learners with
skills and concepts new to them and provide them with appropriate
feedback. This stage is characterized by tightly controlled discourse and
the teacher’s frequent use of display questions that lead to responses by
students and evaluations by teachers. Vague category markers at this
phase serve as
two-way portals. For the teacher, they can open a door to what is
likely shared knowledge for this phase of the lesson and create a
shared space around this commonage. For the learner, they open a
door to a space where it is safe to take risks. Tentative propositions
can be marked using VCMs and loss of face is avoided. In this mode,
they engage cooperative listenership on the part of peers which also
facilitates learning.
(Walsh et al. 2008, p. 26)
In the last stage, the classroom context mode, the local context deter-
mines the management of turns and topics and there are abundant
opportunities for communication, which allows students as much time
and space for interaction as they need. The teacher is mainly a listener
and promotes interaction. The goal here is for students to extend dia-
logue and discussion, so they are given the chance to express themselves
and participate in the discussion and give long responses. A vague cat-
egory marker at this stage behaves similarly to the way it functions in
daily conversation, since ‘it acts as an “involvement device” ensuring
listener participation and prompting equity and understanding’ (Walsh
et al. 2008, p. 25). While Walsh et al.’s study is narrow in scope as only
vague category markers are investigated, it sheds light on the role of
vague language in educational settings in that L2 learners’ vague lan-
guage use is similar to the use in day-to-day conversations.
Although it seems that English for Special Purposes includes scientific
writings made up of a series of objective statements regarding facts,
vague expressions are frequent in scientific journals and play significant
roles in academic writing (Hyland 1996, 1998). In academic discourse,
writers may be able to express a proposition with more precision by
using vague language, as quantifying the world precisely is almost
impossible. To present information as accurately as possible, a writer
tries to keep fact and interpretation balanced. Vague language is an

instrument to present uncertain scientific claims with caution, ‘So writ-
ers often say “X may cause Y” rather than “X causes Y” to specify the
actual state of knowledge on the subject’ (Hyland 1996, p. 478).
Vague language also allows room for the possibly negative conse-
quences of being proven wrong. In order to avoid direct responsibility
for statements they make, academic writers use vague language to offer
speculations. Writers of scientific articles need to keep their writings
open to further developments. The work by Myers (1996) on strategic
vagueness in academic writing suggests an overall picture of the use of
vague language use:
a. Vagueness in statements of results allows them to be compared to

results from slightly different conditions.
b. Vagueness in treatment of numbers not relevant to the argument
guides the reader on the preferred path.
c. Vagueness in articulation between results and implications allows
the text to be assimilated to future developments. (Myers 1996,
p. 12)
A good command of vague language is thus a feature of a proficient L2

reader and writer in academic discourse, which indicates the need for
vague language teaching in English language classes which are often
complementary to General English programs.
There is a trade-off between the level of explicitness and the lexical
density of a conversation, so that the less the degree of explicitness, the
lower lexical density (fewer words) it involves. Just as the lexical density
of a university tutorial tends to be greater than that of a kindergarten
class, the level of explicitness in the language of an academic lecture
may be higher than that of a naturally occurring conversation. Cheng
and Warren (1999) show that non-native speakers of English use higher
levels of explicitness compared with native speakers: in other words,
non-native speakers employ more lexical density in their conversations.
Whilst most studies of vague language were based on everyday con-
versations, few explore its use in academic settings such as this study
does. Lack of research on vague language used by the Chinese and
Persian learners of English makes this study valuable.
2.3.2 Vague language in language learning and teaching

A learner’s discourse is often different from that of a native speaker’s,
and this discrepancy may result from factors such as the impact of the
vocabulary of the first language, differences in conversational rules

in the first language, lack of access to the required word in the target
language, or cross-cultural misunderstanding. Any of these may lead to
unnaturalness in the learner’s discourse, in that the learner is either too
precise or too vague; but they do not necessarily cause communicative
breakdown (Cheng & Warren 2001). Cheng and Warren identify dif-
ferences in vague language use by native and non-native speakers in
their data, but the differences are not remarkable. This indicates that
it may be premature to claim that native and non-native speakers
suffer misunderstandings caused by differences in their language use.
Sociolinguistic competence, according to Bachman (1990), concerns
appropriateness of function in terms of context. It deals with variations
in dialect or register, naturalness, and being able to interpret cultural
references and figures of speech. Naturalness is a characteristic of vague
language, and discourse without an appropriate degree of vagueness
lacks naturalness—which explains why the writing or speech of a com-
petent L2 speaker can seem pedantic and unsatisfactory to L1s, even
when it displays high linguistic accuracy. Bachman (1990, p. 86) points
out that ‘it is this very interaction between the various competencies
and the language use context that characterizes communicative lan-
guage use’.
Pragmatic proficiency refers to appropriateness of language function
or use. Not following pragmatic norms can result in an L2 speaker
appearing rude or offensive (Nikula 1996). Thomas (1995) suggests that
what he calls modifying devices, for instance, softeners and straighteners
of pragmatic force, create vagueness in what the speakers say, and that
their frequency demonstrates that they are communicatively significant
for both L1 and L2 speakers. Nikula (1996, p. 20) considers expres-
sions such as I suppose, probably, or sort of are ‘mitigating and reducing
the force of utterances’ which can serve different purposes: modifying
devices such as as it were and I presume are used in highly formal situa-
tions. Vague language is also associated with expressing politeness and
formality. James (1983, p. 201) maintains that expressions such as sort
of, or whatever, and you know ‘contribute to a certain informality of style
and intimacy of relationship’.
In line with developing language learners’ sociolinguistic competence
in second or foreign language learning, most studies in pragmatic com-
petence focus on such domains as politeness (Brown & Levinson 1987),
request, apologizing (Bardovi-Harlig & Griffin 2005), complimenting
(Holmes 1986; Holmes & Brown 1987; Pomerantz 1978; Wolfson 1981),
and making suggestions (Alcon 2005). Not much research considers the
role of vague language in improving learners’ sociolinguistic compe-

tence and pragmatic competence. The findings of this research contrib-
ute, conceptually and empirically, to the development of the body of
knowledge required to analyze learners’ needs in language pedagogy,
and in the design of instructional materials.
Several studies of pragmatic competence address the significance of
vague language as a crucial building block in an L2’s successful com-
munication. Works from cognitive and social perspectives (Alcon &
Martinez-Flor 2008, p. 6) pay much attention to the cognitive devel-
opment of an individual’s pragmatic competence, while others focus
on how social interactions lay the foundations of an individual’s
pragmatic competence. Cheng and Warren (2001) argue that vague
language is a component of strategic competence; a consideration of
Communicative Language Teaching. In the growing literature, there
is still little on vague language in language learning and teaching,
although some researchers such as Cutting (2007) mention it as part
of the implications of their findings. The present study compares L2 s’
patterns of vague language use as a feature of pragmatic competence
in EFL settings against L1s’ patterns, to determine participants’ vague
language competence; the result is used as evidence to illustrate how
vague language can help L2 learners to pragmatically manoeuvre in
communication.
2.3.3 Vague language and learner language

Based on a corpora of French-speaking advanced EFL learners in com-
parison with L1s of British English, De Cock et al.’s (1998) study reveals
that vague tags, such as and everything and or something, are considerably
underused1 by EFL learners, as much as four times less. EFL learners also
underuse vague expressions such as sort of and kind of. There are also
cases of overuse: they strikingly overuse and so on, ten times more often
than L1s. De Cock et al. attribute advanced learners’ inability to use
vague expressions appropriately to three causes, including ‘systematic
differences in the way vagueness is expressed in their French mother
tongue and in English; shortfall in teaching (the use of vague language
1
This study has a different view in terms of ‘underuse’ or ‘overuse’. The position
of this study is that as far as the frequency of language use is concerned, the
norm of L1s is not necessarily the norm that L2s have to adhere to. For various
reasons L2 speakers may use certain types of expression more or less, and this
is a legitimate discrepancy between L1 and L2 speakers rather than a matter of
‘overuse’ or ‘underuse’. For a detailed discussion of this issue, please see Zhang
and Sabet (in press).
in the classroom may be stigmatized); and finally, lack of contact with

native speakers, a particular problem for EFL learners’ (p. 78). These
findings clash with that of Cheng and Warren (2001), who conclude
that native and non-native speakers show no significant differences
in terms of their patterns of vague language use, but are supported by
Drave’s (2002) finding that there are discrepancies in the ways in which
native and non-native speakers employ it in English.
Drawing on vocabulary frequencies in advanced learner English from
different countries, Ringbom (1998) finds that learners with almost the
same cultural and educational backgrounds show consistencies in their
English language vocabulary use which differ from native speakers’
language (p. 49). He claims that ‘learner language is vague and stereo-
typed’ (p. 49), but does not identify the source of this vagueness. Non-
numerical quantifiers such as more, all, other, some and very are overused
by learners, but many and any are overused by native speakers. Two
general vague words, people and thing, are highly overused by advanced
learners of English. Contrary to the views of Channell (1994) and
Cutting (2006), Ringbom claims that ‘The limited vocabulary that
advanced learners have in comparison with native speakers is a main
reason for the general impression of learner language as dull, repetitive
and unimaginative, with many undeveloped themes’ (p. 50), and that
these features ‘are less due to errors than to an insufficient and impre-
cise, though not necessarily erroneous, use of the resources available in
English’ (p. 51). Some of Ringbom’s (1998) judgments seem somewhat
premature, since they are based merely on the quantitative evaluation
of frequency rate and do not consider the qualitative dimensions of
vague language use in context, such as the functions of vague expres-
sions or the structures of such terms; advanced learners might have
outperformed L1s in expressing the degree of certainty or strength of a
claim. Even if Ringbom proves correct that advanced EFL learners’ lan-
guage is dull, repetitive and unimaginative, with many undeveloped
themes, it could be because the learners were selected from France,
Spain, Finland, Sweden, the Netherlands and Germany: countries
culturally and educationally similar to one another. This factor, rather
than their ‘limited vocabulary’, might have been the reason for their
overuse. In contrast to Ringbom’s findings, Nikula (1996) demonstrates
underuse of vague expressions by Finnish L2 speakers of English, and
reports that such expressions as more or less, kind of, and stuff like that,
and everything are less commonly used by L2 speaking subjects in com-
parison with L1s.
Metsä-Ketelä (2006) investigates the use of vague language by non-

native speakers in the context of English as a lingua franca (a third lan-
guage used by two different language speakers as a ‘bridge’), compared
with native speakers of English. The study of more or less shows it as the
most frequent vague marker in Metsä-Ketelä’s corpora. The expression
is of a relatively high frequency in academic lingua franca English, espe-
cially in monologues such as presentations and lectures.
In terms of function, non-native speakers show three prominent
functions in their use of more or less: minimizing, comparing similari-
ties, and approximating quantities. The minimizing ‘more or less is used
in a similar manner to simply, only, or just, and its purpose is to indi-
cate that the concept is either small in scale or that it is not adequate’
(Metsä-Ketelä 2006, p. 135). This function is unique to non-native
speakers. The comparative frequently serves in the lingua franca corpus
to compare concepts or entities. The third function, not very frequent,
is found only in non-native speakers’ data, as a device to approximate
quantities or express generalizations.
Metsä-Ketelä states that although L2s use ‘deviates from the standard
or native use of the expression, it does not seem to cause any confu-
sion in the interaction’, indicating that some unconventional uses of
vague language by L2s ‘supports the view that lingua franca speakers
can come up with innovative ways of using the language and negotiate
new meanings for old words. It also suggests that cooperativeness and
the will to understand each other play a crucial role in lingua franca
English and therefore the unorthodox use of language does not neces-
sarily result in communication breakdown’ (p. 141). This study sup-
ports Metsä-Ketelä’s view that L1s and L2s can be different in the use
of English, as long as both parties manage to communicate successfully.
In her later work of 2012, based on a wider scope of vague expressions,
Metsä-Ketelä finds that vague expressions are employed almost twice as
frequently by lingua franca speakers as by L1s. General extenders occur
commonly when there is similarity between interlocutors in terms of
status at university. Vague metadiscourse particles are densely located
in doctoral defenses ‘where speaker roles are clearly assigned and hier-
archical’ (p. 280).
The ability that most language learners appear to have problems
with, especially in writing, is expressing appropriate degrees of doubt
and certainty (Hyland & Milton 1997). This is a frequently used
strategy to distinguish between facts and opinions. Studies of learn-
ing in EFL/ESL contexts reveal that depending on their linguistic and
cultural backgrounds, language learners display differences in express-

ing degrees of probability. For example, Chinese-speaking L2 learn-
ers of English opt for a more direct and authoritative tone as well as
for stronger modals than L1s of English (Hu, Brown & Brown 1982).
Allison’s (1995) research reveals inappropriately strong assertions by
ESL writers in Hong Kong, and ‘These problems persist for L2 writers at
post-graduate level where PhD supervisors are often required to coun-
sel the need for appropriate degrees of qualification and confidence
in expressing claims’ (Dudley-Evans 1991, p. 47). Hyland (1998) also
finds that language behavior principles and patterns for exposition and
argumentation are culture-specific and cause differences in students’
writing in English.
Hyland and Milton (1997) focus on hedges used in examination
scripts by L1 British school leavers and Cantonese-speaking English
learners, and find that the Chinese use only half as many hedges as
their British counterparts in similar circumstances. According to Holmes
(1988) and Hyland (1994), one major reason for students’ problems with
the use of hedges and uncertainty markers is the lack of attention and
the misrepresentation of such expressions in pedagogical materials for
ESL classes. Hyland (1998) states that ‘students require an understand-
ing of hedges not only as text-based items, but also as discourse-based
strategies, showing how they relate to the writer’s overall text plan’
(p. 235): that is, language learners need to gain mastery in expressing
degrees of certainty.
Wu, Wang and Cai’s (2010) examination of I think used by Chinese
EFL learners reveals that compared to L1s of English, L2 speakers use
I think far more often. There are similarities in the functions of I think
between the two groups, such as downgrading, marking deliberation,
taking and holding turns, and delaying. The Chinese group in their
study was found to have attached other functions to I think; they used
it to ‘signal conclusions’ and to refer to listing, collocated with so and
firstly.
Research findings in cross-cultural rhetoric show that people with dif-
ferent languages and cultures tend to adhere to their own language and
culture (Connor 1996; Soter 1988), and ‘Such differences can make
non-native speakers vulnerable to the risk of violating communicative
norms as their writing may appear as too direct, running the risk of
being considered as either brusque or dogmatic, or as too tentative, and
therefore seen as equivocal, diffident or naive’ (Hyland & Milton 1997,
p. 186). This particularly applies to the use of vague language: if L2s can-
not handle it in an appropriate way, their communications may suffer.
In research on the use of expressions of doubt and uncertainty by L1s

and L2 users, Hyland and Milton (1997) find that despite the fact that
the two groups made extensive use of some modal verbs and adverbs,
the L2 students turned out to have problems with ‘the manipulation
of certainty and affect in academic writing’ (p. 201). Research (Biber
et al. 1999; Brazil 1997; Cook 1989; Hyland & Milton 1997; Myers 1996)
demonstrates the influential role of vague language in both written
and spoken modes, as well as indicating that it has characteristic fea-
tures in terms of frequency and function in either mode. For example,
discourse intonation is unique to the spoken mode, which can support
vague language use, and the appropriate use of informal language in
spoken discourse can demand a high frequency of vague language
(Brazil 1997).
Channell states that ‘It is often noticed by teachers that English of
advanced students, while grammatically, phonologically, and lexically
correct, may sound rather bookish and pedantic to a native speaker.
This results in part from an inability to include appropriate vague
expressions’ (1994, p. 21). Her recommendation is that vague expres-
sions be incorporated into the curriculum of EFL classrooms; she even
takes a step further by suggesting incorporating vague expressions in
native speakers’ higher education curriculum.
2.4 Theoretical frameworks
Vague language plays an indispensable part in communication.

McCarthy (1998) claims that vague expressions make important con-
tributions to naturalness and to the informal, convergent tenor of
everyday talk. Jucker et al. (2003, p. 1766) state that vague expressions
are ‘not just a poor substitute for a precise expression. Rather, they
often convey meaning that is different from and more relevant than a
precise expression’. There are three relevant theoretical frameworks to
this study: the cooperative principle (CP, Grice 1975), relevance theory
(RT, Sperber & Wilson 1995) and the elasticity of vague language (Zhang
2011), discussed in detail as follows.
2.4.1 Vague language and the cooperative principle

Vague language is mainly examined from a pragmatic perspective,
which relates to the cooperative principle proposed by Grice (1975),
and particularly the notion of conversational implicature. Grice asserts
that successful communication in any conversational context is the
result of the communicator’s adherence to an underlying principle to
be as cooperative as possible. The idea is to ‘Make your conversational

contribution such as is required, at the stage at which it occurs, by the
accepted purpose or direction of the talk exchange’ (Grice 1989, p. 26).
The principle is manifested in four maxims: quantity (make your con-
tribution as informative as required, do not make it more informative
than is required), quality (be truthful, namely don’t say what seems to
be false and that which you lack evidence for), relevance (make your
contribution relevant) and manner (be clear, brief and orderly, avoiding
obscurity and ambiguity).
Huang (2007) argues that the use of hedges in conversations indicates
that speakers are subconsciously aware of the existence of these maxims
and attempt to observe and adhere to the cooperative principle. He
gives examples corresponding to Grice’s four maxims:
a. Quality:
As far as I know, …
b. Quantity:
As you probably already know, …
c. Relation:
Oh, by the way, …
d. Manner:
This may be a bit tedious, but … (pp. 26–27)
Grice states that conversational implicature is a result of flouting the

maxims. Huang (2007) presents the following example (p. 28):
Relation:
John: What’s the time?
Mary: The museum hasn’t opened yet.
[Implicature]: It’s at least before whenever the museum normally
opens.
This seems to flout the Maxim of Manner, in that it does not provide
a clear, straightforward answer to the question. Huang argues that
the Maxim of Relevance here leads to conversational implicature.
Mary’s answer is relevant, assuming Mary is cooperatively answering
John’s question; so ‘we can infer that while Mary is not in a position
to provide a straightforward answer, nevertheless she thinks that the
museum’s not being open yet might help John to get a partial answer’
(Huang 2007, p. 29).
When a speaker ostensibly flouts maxims, the hearer may either

realize that the cooperative principle has been violated or assume that
the seeming lack of cooperation marks the speaker’s attempt to follow
the Principle at a deeper level (Huang 2007): that is, the addressee may
realize a responsibility inferring the message beyond the words. For
example,
Maxim of Quality:
Abdullah: Mumar is engaged to three different girls.
Susan: Oh, he’s a real monogamist.
[conversational implicature] Mumar is no monogamist at all.
The listener will know that the statement is not true, and assumes that
the maxim of quality is being violated to maintain the cooperative
principle. The hearer will also tend to assume that the speaker means
something quite different from what has actually been said, something
beyond the literal meaning of the statement: ‘the ironic meaning’
(Huang 2007). Vague language use does not necessarily flout the maxim
of quantity because it is often used for succinctness, saying no more and
no less than the context requires, as often there is no need to say more
(Zhang 2005).
Conversational implicature is classically a type of figurative language
(Sperber & Wilson 1986a, p. 155), in that the speaker violates the coop-
erative principle on the assumption that the hearer is able to understand
the implied meaning, because ‘When the literal interpretation is inap-
propriate, the appropriate figurative interpretation somehow comes to
the hearer’s mind’. Channell (1994, p. 33) mentions that when she is
asked about the time to be home from work and she does not know, the
most truthful answer would be ‘about six o’clock’ (a vague expression).
She assumes that from this answer her hearer will infer that she cannot
say exactly: that is, she is trying to be as truthful as possible (Maxim of
Quality) and is, to the best of her ability; but her answer violates the
maxim of manner in that she does not clearly address the question.
Grice’s framework has met with some criticism. Davies (2007, p. 2310)
has concerns that when speakers flout conversational maxims and cre-
ate conversational implicatures, conveying their intentions in a veiled
manner, it makes the task of the hearer more difficult. It would seem
that the cooperative principle is not about making the task of the hearer
straightforward; it can be quite the reverse. However, Davies’s concern
is a non-issue in the case of vague language use, as a not-so-precise
utterance creates no barriers to hearers; in fact, neither speakers nor

hearers are likely to notice it in communication. Language users are
able to communicate in vague language successfully without difficulties
(Channell 1994; Zhang 2011).
Cutting (2008) comments on the lack of cultural considerations in
Grice’s framework. She argues that maxims of conversations are cultur-
ally specific, and that language users from different countries and com-
munities may have different ways of adhering to the maxims. Cutting
demonstrates that a maxim flouted in one culture is counted as strictly
observed in another. For example, in terms of maxim of quantity, ‘How
are you?’ in the United States is followed by a reply such as ‘fine’, but
in another culture respondents may be expected to refer to their actual
state of health. When speaker A says, ‘We’ll call you in about two weeks’
to speaker B and fails to do so, this is regarded as flouting the maxim of
quality in Britain because A didn’t tell the truth; but in other countries
it is taken as a way of indirectly stating ‘We are not interested in you’
(p. 40). Cutting also claims that it is impossible to establish clear-cut
boundaries between the four maxims, and there could be more than
one in operation.
There is interconnection between the conversation maxims/conversa-
tion implicature in Grice’s framework and vague language phenomenon
in this study, and the concepts of the cooperative principle are relevant
and useful in the discussion of vague language behaviors and data
analysis of this study.
2.4.2 Vague language and Relevance Theory

As an alternative approach to Gricean pragmatics, Relevance Theory
(Sperber & Wilson 1986b/1995) states that human cognition is involved
in maximizing relevance in communication. While the cooperative
principle is based on usage principles of communication, relevance
theory is related to cognitive principles (Levinson 1989). The point of
departure for this theory is not the socially-acquired cooperative prin-
ciple followed by communicators, but human cognition. Underlying
relevance theory is the notion of relevance: that humans make the least
processing effort to achieve the maximum positive cognitive effect in
communication. ‘Various pragmatic theories appeal to complex sets
of rules, maxims, or conventions to explain how this linguistic under-
determination is contextually overcome. We claim that the principle of
relevance is enough on its own to explain how linguistic structure and
background knowledge interact to determine verbal communication’
(Sperber & Wilson 1986a, p. 161). Unlike the cooperative principle, the
relevance theory principles are not like rules to be followed and obeyed
in communication, but ‘are an automatic reflex of the human men-
tal capacity that works without the communicators having any overt
knowledge of it’ (Huang 2007, p. 202).
In the framework of relevance theory, pragmatics is regarded as a
single notion of relevance, manifested in two principles: the cognitive
principle of relevance and the communicative principle of relevance.
As visualized by Sperber and Wilson (1995), relevance is a measure
consisting of cognitive effect and processing effort. Cognitive effect
refers to the interaction of a new input and a set of assumptions already
existing in a cognitive system, and processing effort addresses the effort
spent by that cognitive system to produce an appropriate interpretation
of incoming information ‘geared to the maximization of relevance’
(Sperber & Wilson 1995, p. 252). The relevance of an input to the
person is a matter of balance between cognitive effects (benefit) and
processing efforts (cost):
a. Other things being equal, the greater the positive effects achieved
by processing an input, the greater the relevance of the input to the
individual at that time.
b. Other things being equal, the greater the processing effort expended,
the lower the relevance of the input to the individual at that time.
(p. 252)
The optimum is to achieve the greatest cognitive effect using the least
processing effort. As for the communicative principle of relevance
(ostensive-inferential communication), it is not true that in the process
of communication we are absorbing every possible input and scan-
ning it for relevance, because this would make communication quite
difficult. Relevance theory assumes that communication contains two
kinds of information: that which the speaker wishes to transmit, and
that which conveys the speaker’s intent to inform the audience of this.
In other words, ostension and inference are the two poles of communi-
cation. Ostension informs the communicator’s point of view and infer-
ence informs the audience’s perspective; the communicative principle
of relevance is met when the ostensive stimulus is relevant enough to
be worth the audience’s processing effort, and is the most compatible
with the communicator’s abilities and preferences (Sperber & Wilson
1995, p. 254).
Cheng and Warren (2001, p. 93) assert that ‘Since vague language
seems to be easier to process and makes fewer demands on the listener,
it is probably also the case that a speaker may choose to use a greater
amount of vague language to make the discourse easier for the hearer(s)
to understand’. A speaker will skip technical words or ‘specialized lan-
guage’ that the listener doesn’t know and employ a simplified language
instead.
Jucker et al. (2003), who adopt relevance theory in their study, state
that vague language enables speakers to maintain fluency when some
information is not immediately available to them. When this happens,
the speakers ‘may then decide that the processing cost of accessing it,
and the cost to fluency, are not warranted in terms of any benefits to
be gained by precision. However, speakers may choose vague expres-
sions even when they could have stated their utterances more precisely.
A vague utterance may be more efficient in the sense that it yields the
same contextual assumptions for lower processing cost’ (Jucker et al.
2003, p. 1765).
Zhang (2005) argues that relevance theory rightly supports the non-
numerical approach to vague language, but that the numerically and
semantically oriented approach has its own merits. She argues that
Zadeh’s (1965) fuzzy set theory and relevance theory have convergent
and divergent principles: both give priority to optimality, relying on
what is best or most suitable in a situation, but the former adopts a
quantitative approach, emphasizing mainly the semantic aspect of
meaning with numerical values, whereas relevance theory insists on a
cognitively oriented approach to interpretation, which is useful because
‘sometimes we don’t know or cannot agree on the exact numerical
value for fuzzy expressions’ (p. 80). More importantly, there are occa-
sions when we know the numerical information but prefer to use vague
expressions for such reasons as withholding information or safeguard-
ing ourselves. Zhang asserts that what determines the realization of
optimal relevance is not the option for fuzzy or non-fuzzy forms of
language, but the communicator’s perception of the relevance of the
utterance in question.
A similar concept to vague language proposed in relevance theory is
loose talk. Sperber and Wilson (1986a) maintain that loose talk is a type
of non-literal use ‘based on resemblance relations among representa-
tions, and involve interpretive rather than descriptive dimensions of
language use’ (p. 164). When someone loosely understands a proposi-
tion or concept, this does not mean that the concept or proposition is
vague; nor does it indicate that the proposition comes with a guarantee
of approximate truth. ‘Instead, certain of its logical and contextual
implications are taken to be accompanied by regular guarantees of
truth, whereas others are simply ignored. Thus the truth–conditional

relation between propositions and the state of affairs they represent
remains unaltered: what varies is how closely the proposition expressed
is taken to represent the speaker’s thought’ (p. 164). Obviously, loose
talk adheres to the relevance principle; similarly the use of vague lan-
guage serves the purpose of obtaining maximum positive effect using
the least processing effort.
Zhang (2005) points out that relevance theory offers only a limited
explanation of how contextual effects and processing efforts can be
measured objectively, or how they can be compared with each other.
Cutting (2008, p. 42) brings ‘cultural and social dimensions’ once again
to attention and states that, like the cooperative principle, relevance
theory falls short of observing the influence of such factors as ‘age, gen-
der, status and nationality’. She notes that each country or culture pos-
sesses unique ways of abiding by or demonstrating any maxims. Franken
(1997) questions the foundation of Sperber and Wilson’s account of
vagueness and approximative utterances as cases of loose talk, asserting
that there is no reason to put them in this category. Vagueness origi-
nates from vague concepts—but there is no such thing as approximate
concepts (p. 150). Sperber and Wilson state that concepts, while they
can be loosely represented, have precise and well-defined boundaries,
but Franken argues that some concepts are ineffable: that is, they lack
as lexical, logical or encyclopaedic entry; some concepts are vague by
nature (1997, p. 145).
The position of this study is that the use of vague language serves the
purpose of obtaining maximum positive effects using the least process-
ing effort: that is, it is in accordance with the principles of relevance
theory, rather than against them. Similarly, Jucker et al. (2003, p. 1742)
state that ‘Vague expressions may guide listeners to find the best match
between the utterance and the intended meaning’. While relevance
theory mainly takes a cognitive approach to the study of meaning, it
offers useful insights into vague language study; however, both the coop-
erative principle and relevance theory can explain the vague language
phenomenon in some way, even though they have not been developed
specifically for vague language communication, unlike Zhang’s work
on elasticity.
2.4.3 The elasticity of vague language

Zhang’s (2011) work is one of the few attempts to provide an overarch-
ing conceptualization of vague language, which has previously been
lacking in research. The elasticity of vague language refers to vague
language as a versatile strategy at interlocutors’ disposal for effective

communication. To elucidate the mechanism of elasticity, Zhang uses
the metaphor of a slingshot: the rubber band is stretched and aimed by
the user, and the stone is released to hit the target. She describes this
as a three-stage process comprising ‘stretch, aim/adjust, and release/
hit’ (p. 579). The rubber is stretchable an infinite number of times and
to different degrees, as required to hit the target: ‘When the target is
close, simple, or clear, the result may be accurate and certain. When the
target is far, complicated, or unclear, the result may be less accurate and
certain’ (p. 579). Zhang maintains that vague language can be stretched
in any direction and is as elastic as needed to suit the context in ques-
tion. She describes three ‘directions’ to demonstrate how elasticity can
be realized in communication:
This is very important. (upward)
This is a bit embarrassing. (downward)
There are about 20 students in the classroom. (horizontal)
The word very functions as a booster to stretch the strength of utterance

claim upward; a bit is a hedge, pulling the strength of the utterance
downward. The approximator about modifies the utterance’s meaning
horizontally around the number. These examples of vague language in
use show that it is elastic and stretchable in various directions to serve
the needs of communication.
To describe the strategic manipulation of vague language, Zhang
coins a new term, ‘vague work’, which is ‘a way of vague-ing language to
fit a situation’ (2011, p. 573). Focusing on the dynamic nature of vague
language, Zhang believes that its interpretation ‘depends on contex-
tual and communicative purpose’ (p. 578). The theoretical framework
has been developed around a main maxim, four specific sub-maxims
and three characteristics of vague language. The main maxim assumes
that language can be elastically stretched in discursive negotiations to
enhance communication. The four specific vague language elasticity
maxims are:
a. Go just-right: provide the right amount of information (for example,

He is about 20 years old)
b. Go general: speak in general terms (for example, Do you have any
convictions or anything?)
c. Go hypothetical: speak in hypothetical terms (for example, It could

be him)
d. Go subjective: speak in subjective terms (for example, I think she is
dishonest).
(Zhang 2011, p. 579)
These four maxims serve various pragmatic functions. They are not an
exhaustive list but should cover a majority of vague language uses. They
are interconnected by the fact that all can be used strategically to serve
functions that precise language is unable to perform. Each has a differ-
ent priority. For example, ‘go just-right’ stresses the approximated infor-
mation and ‘go subjective’ concentrates on the speaker’s attitudinal and
stance information. Zhang also states that vague language elasticity has
the following three major characteristics:
a. Interconnected patterns of strategic elasticity: between lexical items

of vague language realization, pragmatic functions, and the four
maxims. In particular, one type of lexical item tends to serve a cer-
tain type of function and conform certain maxims. For instance,
something is a general term, often performs a function of mitigating,
and observes the maxim of ‘go general’;
b. Determinant communicative purposes: The spontaneous communi-
cative purpose in a particular context directs the behavior of vague
language. How far the vague language is stretched and in what
direction is specified by the communication needs in a particular
context;
c. Versatile pragmatic strategies: vague language elasticity involves a
moving back and forth within the two poles of a continuum. This
kind of elasticity can range within ‘contrastive pragmatic functions’
such as ‘soft and tough, firm and flexible, cooperative and uncoop-
erative moves’ (2011, p. 579).
2.5 Concluding remarks
The review of literature in this chapter has looked into the different
paradigms of vague language research, which provide the required
theoretical foundations for this study. Vague language can be looked at
from different perspectives, including Grice’s cooperative principle and
Sperber and Wilson’s relevance theory. Zhang’s framework of elasticity
offers a more promising approach for this study.
Despite a growing body of literature on vague language, there are few

works considering it in an ELT setting. This research examines how the
elastic features of vague language serve the diverse needs of language learn-
ers, compared with L1s. It focuses on the manifestation of vague language
and its versatile roles in communication. This study provides a theoretical
exploration of vagueness as well as an empirical analysis of how vague
language is manipulated similarly as well as differently among L1 and L2
groups.
3
Methodology
The present study is a mixed methods research, combining both quanti-

tative and qualitative methods. There is integration and complementarity
between the two approaches.
3.1 Three approaches
Traditionally a quantitative approach dominated educational research,

but qualitative research has come to be accepted as well. Creswell (2008,
p. 46) states that ‘The development of the two approaches is not a case
of one approach replacing the other; instead, it reflects the addition of
qualitative inquiry to the traditional quantitative approach’. What is
interesting about these approaches is that neither can be purely applied
in a study (Firestone, 1987): in other words, it is hardly possible to
claim that a study is solely quantitative or solely qualitative. Creswell
(2008) claims that in any study, the researcher moves within a frame-
work which gives more weight to one approach than the other so that
‘A study tends to be more qualitative than quantitative or vice versa’
(Creswell 2009, p. 3); Reichardt and Cook (1979) add that, rather than
being either-or-neither, research lies along a continuum of qualitative
and quantitative approaches.
3.1.1 Quantitative approach

In quantitative research, ‘the researcher decides what to study; asks
specific, narrow questions; collects quantifiable data from participants;
analyses these numbers using statistics; and conducts the inquiry in an
unbiased, objective manner’ (Creswell 2008, p. 46). This kind of research
dates back to the late nineteenth century when it gained a remarkable
prominence in education that endured during most of the next century
33
(Travers 1992). The philosophy behind quantitative research is the

post-positivist paradigm: ‘the need to identify and assess the causes
that influence outcomes, such as found in experiments’ (Creswell 2009,
p. 7). The method is also referred to as ‘scientific method’ or ‘empirical
science’.
In quantitative research, measurement, assessment, numbers and
experimental research are common phenomena, permitting the con-
version of patterns into accurately quantifiable measures for study. The
most common designs associated with the quantitative approach are
experimental, correlational, and survey designs. Quantitative research
has three basic elements: statistics, test and measurement practices, and
research design (Creswell 2008). The statistical procedure is developed
as a result of ‘correlation analysis’, establishing relationships between
two or more variables. For example, the quantitative method can be
employed to measure individuals’ achievements, in the form of aptitude
tests, selection tests, or placement tests, to name a few. Early linguistic
research designs were simple surveys of relevant issues, which gradu-
ally gave way to experimental studies such as comparing performance
or attitudes in two groups. Later, researchers developed more complex
designs which let them study multiple groups and also administer
multiple tests.
Two of the major elements of quantitative research are the variables
and the hypothesis, which give meaning to measurement and statistics.
As Creswell (2008, p. 139) states, ‘In quantitative research, researchers
often test theories, broad explanations that predict the results from
relating variables … the investigator employs a closed-ended stance by
identifying variables and selecting instruments to collect data before
the study begins. Quantitative research questions and hypotheses do
not change during the study’. In Creswell’s view, quantitative research
is deductive. It has shortcomings: for example, it does not fit well with
the reality of social sciences, which engage with ‘everyday-life questions
and problems’ (Flick 2002); and scholars in education have argued that
the approach focuses on the ‘researcher’s view of education’ rather than
the ‘participant’s view’ (Creswell 2008); the traditional situations in
which experimental research is conducted tend to exclude the partici-
pants from their natural settings and place them in an artificial situa-
tion dissimilar to real-life situations.
Despite these inadequacies, this research is partly based on a quantita-
tive analysis. In particular, the frequency distribution of vague language
by groups of speakers is investigated, to present a general pattern of
participants’ vague language behaviors. The frequency analysis makes
Methodology 35
use of quantitative processes to identify potential differences between

groups. A statistical analysis employing the Chi-square test was also
conducted, to validate the significance of differences in the use of vague
language across the three groups.
3.1.2 Qualitative approach

Creswell (2009, p. 4) defines a qualitative approach as ‘a means for
exploring and understanding the meaning individuals or groups ascribe
to social or human problems’. In qualitative research, researchers act
as data collection instruments, directly involved in data collection,
for instance by observing participants’ behaviors and by interview-
ing them. There are protocols designed for collecting data, but the
researcher is responsible for carrying out the procedure. He or she can
also use multiple sources of data (Creswell 2009).
Table 3.1 shows that the quantitative method measures objective
facts through analyses such as variable and statistical tests. It values reli-
ability and the emergence of general patterns that are independent of
context. There is little role for researchers to play in the process of data
collection. A qualitative approach is quite the opposite: it investigates
social realities and culturally constructed meanings through interactive
data, subjected to thematic analysis. Researchers play a significant role
in data collection, and the outcomes are context-dependent.
The method of data collection for this present study is predominantly
qualitative. No pre-designed hypothesis was established to narrow the
direction of the study. The data to be used for this study are the texts
of interaction among the participants. The data analysis is in principle
accordance with conversation analysis (CA), a generally qualitative
Table 3.1 Differences between quantitative and qualitative methods
Quantitative Qualitative
measuring objective facts constituting social reality and cultural

meaning
variables interactive processes and events
reliability authenticity
generalization independent of context specific to particular context
statistical analysis thematic analysis
no significant role for researcher in researcher directly involved in the data
the data collection collection
Source: adapted from Neuman (1997).

approach employed to study talk in interaction (Sacks, Schegloff &

Jefferson 1974). Hutchby and Wooffitt define it as ‘the systematic analy-
sis of the talk produced in everyday situations of human interaction:
talk-in-interaction’ (2008, p. 14).
The premise behind conversation analysis is that there is more to
talk-in-interaction than simply analyzing conversation; in fact, conver-
sation analysis deals with how interlocutors understand and respond
to each other, and the central focus is on how sequences of actions are
produced (Hutchby & Wooffitt 1998). Despite its seemingly disordered
look, conversation follows a close order, with a ‘uniformity in structure’
constructed by the orderly ways adopted by the participants in interac-
tion. The structure of conversation is not inherently ordered, but is
constructed in the act of conversation (Liddicoat 2007; Wooffitt 2005).
The central rationale of conversation analysis is that ‘ordinary talk is
a highly organized, ordered phenomenon’ (Hutchby & Wooffitt 1998,
p. 14). Conversation analysis studies the sociolinguistic competen-
cies underlying the creation and interpretation of talk in organized
sequences of interaction. It focuses not only on language but on the
practical social accomplishments made through language (pragmatic
functions). Liddicoat (2007) defines conversation as the means of social-
izing, developing and maintaining relationships between people, point-
ing out that it involves more than exchanging linguistic codes; extra
features include ‘eye gaze and body posture, silence and the real world
context in which the talk is produced’ (p. 1).
One of the distinctive features of conversation analysis is the require-
ment of naturally occurring data. Conversation analysis involves an
in-depth analysis of real life interactions in order to determine how
activities are performed through utterances. Audio or video recordings
are particularly useful ways of collecting these data, as they can be
reviewed as often as required so that the correct level of accuracy can
be obtained. In addition to verbal information, conversation analysis also
pays attention to non-verbal activities as features of spoken language.
For example, it is possible to mark periods of overlap between two turns,
or gaps between words or turns, and even to show the point where a
speaker paused to breathe, on the understanding that these elements can
add depth to an analysis of human linguistic interactions.
Another feature of conversation analysis is associated with the way
data are analyzed. Rather than using numerical units such as percent-
ages and frequency, adjectives and adverbs such as commonly, overwhelm-
ingly, regularly, typically, and the like are used to analyze data (Liddicoat
2005; Schegloff 1993). The reason is that data in conversation analysis
Methodology 37
are a collection of actions; and each instance in the collection is from a

contextualized talk by a participant. ‘The study of collections is there-
fore the study of multiple examples, in which each next case demon-
strates the systematic commonalities which exist across participants
and contexts’ (Liddicoat 2007, p. 11). Because of this, one of the foci
of conversation analysis is the concept of turn-taking. This deals with
how interactants figure out when it will be their turn to talk or how
the other interactant will realize their co-participant is handing over.
Turn-taking is an accidental phenomenon in conversation, as no one
can anticipate how many turns will be taken, how long each will take
or how the turn sequence will be organized (Wooffitt 2005). Sacks et al.
(1974) have presented descriptions of systematic turn-taking in which
they tacitly rely on turn-taking components and a set of procedures for
turn allocation.
A predominant characteristic of conversation analysis is the use of
naturally recorded conversations as the basis for analysis, which means
it can deal with the activities people perform with their utterances in
real-life situations. An advantage of conversation analysis is that it
makes access to all the contributions to an interaction (for example,
accidental aspects) possible, and details that might seem irrelevant at
first glance may prove to be interactionally significant. Hutchby and
Woffitt summarize the methodological basis of conversation analysis
as follows: ‘Talk-in-interaction is systematically organized and deeply
ordered. The production of talk-in-interaction is methodical. The
analysis of talk-in-interaction should be based on naturally occurring
data. Analysis should not initially be constrained by prior theoretical
assumptions’. (1998, p. 23)
As with any other research method, conversation analysis is not free
of shortcomings. It fails to ‘address the kinds of topics which are cen-
tral to traditional sociological inquiry: for example, the manifestation
of power and inequality in social relationship and mobilization of dis-
advantage based on gender, ethnicity or class’ (Wooffitt 2005, p. 158).
Conversation analysis is narrow in scope; it is not able to ‘take account
of the essentially argumentative nature of everyday discourse, focusing
instead on the management of interpersonal harmony and accord …
its focus on the “technical” aspects of the sequential organization of
turn-taking means that it cannot address the wider historical, cultural
and political contexts and meanings which are invoked by and reflected
in the kinds of words and phrases we use in everyday communication’
(Wooffitt 2005, p. 154). However, Billig (1999) argues that conversation
analysis actually reflects a specific kind of social order, claiming that an
assumption of conversation analysis is that people have equal status in

the interaction under study.
This study analyzes vague language patterns in context. The data were
not collected in accordance with any predefined qualitative approach.
For this study, conversation analysis is an appropriate method to
investigate pragmatic functions of vague language in particular. The
approach can also look into how features such as pauses, overlaps
in speech, and other phenomena that occur naturally in conversation
can influence participants’ patterns of vague language use. Conversation
analysis prefers naturally recorded audio and video data, which are the
type of data collected in this study.
3.1.3 Mixed methods approach

As a quantitative analysis alone was insufficient to address the research
questions in this study, the qualitative analysis was carried out to
validate the quantitative analysis and explore the area of pragmatic
functions of vague language. The combination of qualitative and quan-
titative methods provides a full account of the use of vague language in
the classroom context.
Creswell (2009, p. 4) states that ‘Mixed methods research is an
approach to inquiry that combines or associates both qualitative and
quantitative forms. It involves philosophical assumptions, the use
of qualitative and quantitative approaches, and the mixing of both
approaches in a study’. Creswell notes that this approach needs the
researcher to have a good command of research skills, which go beyond
merely collecting quantitative or qualitative research data, to knowing
how to integrate and link them.
The mixed methods approach is the result of an evolutionary move-
ment in the development of research approach. As with quantitative
and qualitative approaches, this method has its own philosophical
underpinning, called pragmatism. Pragmatism relies on practices,
situations and outcomes rather than pre-established conditions; it is
concerned with what is most appropriate to solving a problem (Patton
1990). As Rossman and Wilson (1985) assert, the researcher’s main focus
in this approach is the research problem, and all different approaches
are considered in an attempt to gain the best understanding of the
problem; the researcher does not focus on one particular method, but
rather selects the most appropriate tools from every toolbox to find the
answer.
The merit of mixed methods research is that it employs both quan-
titative and qualitative methods, making up for the shortcomings of
Methodology 39
one by taking advantage of the strengths of the other. A quantitative

researcher may dislike the possibility of qualitative research being dis-
torted by researcher bias or interpretation, or the difficulty of general-
izing any findings to a large population because of the small number
of participants. A qualitative researcher may dislike the possibility of
quantitative research being too general without considering in detail
the contextual information. A mixed methods researcher may make use
of all kinds of data-collection tool and is able to present more compre-
hensive evidence for their study.
The rationale underlying the mixed-methods design is that the
strengths in one approach make up for weaknesses in the other. For
instance, the natural setting of a qualitative study can make up for
the artificial setting common in quantitative studies. Creswell (2008,
p 557) describes how triangulation works: ‘The researcher gathers both
quantitative and qualitative data, analyzes both datasets separately,
compares the results from the analysis of both datasets, and makes
an interpretation as to whether the results support or contradict each
other’. Both quantitative and qualitative data are collected simultane-
ously, and the researcher gives equal weight to quantitative and quali-
tative data. The results gained from the analysis are compared to see if
the two data sets indicate any similarities or differences. Morse (1991)
states that the aim of triangulation is to obtain data that are different
but complement each other regarding a selected topic.
Creswell (1994, p. 177) defines four types of mixed methods design:
Sequential studies: The researcher first conducts a qualitative phase of

a study and then a quantitative phase, or vice versa. The two are
separate.
Parallel/simultaneous studies: The researcher conducts the qualitative
and quantitative phases at the same time.
Equivalent status design: The researcher conducts the study using both
the quantitative and the qualitative approaches about equally to
understand the phenomenon under study.
Embedded design: Dominant and less dominant paradigm with a small
component of the overall study drawn from an alternative design.
This study falls under the first of these categories, with the quantita-
tive section dealing with lexical analysis in the form of the frequency
of occurrence of vague language. This was conducted before the quali-
tative study of the functional use of vague expressions across three
groups. However, quantitative and qualitative methods should not be
completely separate: at the very least the qualitative work in this study
should validate the findings of the quantitative analysis. This study
also fits in with the last type of design, as the qualitative analysis is
dominant, mainly because the data were collected through a qualitative
approach.
A concept closely associated with mixed methods approach is trian-
gulation. Creswell (2008, p. 553) describes this as an investigator being
able to ‘improve his investigation by collecting and integrating vari-
ous kinds of data on the same phenomenon’; the three points of the
triangle are ‘the two sources of the data and the phenomenon’. The
idea of triangulating data sources as a way to create convergence across
quantitative and qualitative methods was first developed by Jick (1979).
Merging the two and using the results to better understand a research
question.
Creswell (2008) claims that the strength of the triangulation design
is that it takes advantage of the strengths of each type of data collec-
tion. However, its popularity cannot disguise potential problems. As it
involves both quantitative and qualitative data, it demands hard work
and a high level of expertise in both methods. The researcher may
find that the results of one approach contradict the other, requiring
the collection of fresh data, which may be difficult to obtain. Another
challenge is how to integrate the two different data sets so that they are
comparable (Creswell 2008). In any case, integrating the results of very
different methods to answer research questions is challenging, and ‘like
the triangulation design, the simultaneous data collection of quantita-
tive and qualitative data may be labor intensive for a single researcher’
(Creswell 2008, p. 559).
The other mixed methods design is the embedded design. The simi-
larity between this and triangulation is that both involve concurrent
quantitative and qualitative data collection. The difference is that for
the embedded design, one form of data makes up the primary source
while the other counts as a supportive source. In other words, the
researcher gives more weight to one and counts the other as com-
plementary evidence (Creswell & Plano Clark 2007). In an embedded
design the researcher can adopt a one-phase or a two-phase approach,
using quantitative and qualitative data to answer different research
questions (Hanson et al. 2005). In this design one method is considered
a component of another, answering questions which the primary source
has failed to address adequately. Creswell and Clark (2007) state that
the strength of the embedded design is that it gives the researcher the
opportunity to take advantage of two methods in a single study. It can
Methodology 41
also be applicable to situations where the researcher is short of the time

or resources to conduct both kinds of data collection, when one is less
significant than the other.
Creswell and Plano Clark (2007) note that collecting and analyzing
both quantitative and qualitative data is highly demanding in terms
of time and resources. It requires complicated research procedures and
needs clear presentation. More often than not researchers come from a
background that relies on only one form of research, so mixed methods
research can be particularly challenging. The present research primar-
ily follows an embedded design in that quantitative analysis serves as
a support to the qualitative element, and is a relatively straightforward
procedure which does not cause much complication.
3.2 Naturally occurring data
There are two means of gathering natural data: compiling field notes of
real life data and tape-recording. With regard to the first means, Tran
(2006, p. 3) states, ‘In this ethnographic method, researchers observe
real-life interactions and take notes of natural data on the communi-
cative acts in focus’. The latter involves audio- or videotaping social
interactions as in conversation analysis, to capture communicative acts
in progress.
The most salient advantage of tape-recording is that the data is natu-
ral and represents discourse features. It frees the researcher from note-
taking, so there is no need to rely on memory or selective attention; as
a result the reliability of data collection rises. Tape-recording allows the
researcher to replay the data and improve transcription; more impor-
tantly, it preserves the sequence of talk.
Despite these advantages, this method has met with some criticism.
The degree of the researcher’s control of social variables concerning the
interlocutors is low. Beebe and Cummings (1996, p. 81) note that ‘many
studies of natural speech have not given us scientifically collected sam-
ples that represent the speech of any identifiable group of speakers.
They do not give us situational control’; thus, conversation analysis
may yield unsystematic data.
The researcher may run the risk of not getting enough data regarding
the communicative acts under investigation after recording authentic
interactions for a long period. Another disadvantage is that an audio
recording may harm the confidentiality of the respondents, if not han-
dled carefully. As their exact words and voices are recorded, they run the
risk of being identified or their secrets being disclosed. The participants
may also feel nervous if there are cameras around, and act unnaturally.
The remedy to this problem is to keep the recording device as unobtrusive
as possible. The equipment need not be concealed, but should be located
where it does not distract respondents and its presence is not confronting.
Vague language is an indispensable part of natural speech, so the
possibility of not finding vague expressions in interactions through
conversation analysis is relatively low. The data collected for this study
offered a rich resource of vague language, and the application of a sim-
plified version of the transcription conventions of conversation analysis
provided a good platform for its analysis.
3.3 Data
The data consists of three sets of interactive discussions in classrooms,

with one L1 group and two L2 groups, each with about 70 participants.
L1 data were selected from MICASE, and the two L2 sets are video
recordings of classroom interactions by Chinese and Persian learn-
ers with similar, upper-intermediate to advanced, proficiency levels.
The data consist of 150 thousand words from a total of 20 hours of
recording.
Purposive sampling for L2 data was employed: the researchers pre-
selected groups for data collection. A pilot study consisting of one hour
of data recording in Iran was checked for quality; after this, the data
was recorded. Obscurities such as unclear words were discussed with the
class teacher by telephone.
The Persian speaker data were mainly recorded by a director at an
English language center and partly by one of the authors of this book,
using a digital video camera. The Chinese data were video-recorded by
an associate who was thoroughly informed of the requirements of the
recording. The conventions used in the transcription of the L1 data
were adopted for the transcription of the two L2 data sets. The L1 data
transcription was ready-made from a corpus; the transcription of the
L2 data sets was undertaken by one of the authors of this book.
The main criterion forming the basis of comparison across the three
groups was the number of words involved, rather than the length of
the recording. The pace of speech was slower in the L2 groups, and this
would have distorted the comparability of the data had time been the
criterion. Terraschke (2008) has pointed out the difference in word
length from one language to another and one discourse to another, and
comparing the use of pragmatic devices in terms of time is not optimal.
The L2 video-recorded data were naturally occurring (as opposed to
Methodology 43
manipulated) and represented real discourse features (as opposed to arti-

ficial or controlled interaction). It covered not only language behaviors
but nonverbal activities, including teachers’ and students’ facial expres-
sions and body language.
The L1 data presents American English in academic contexts. The
reason for choosing American English was that the Chinese and
Iranian L2 groups used American English materials in their language
learning. MICASE includes conversations across a wide range of con-
texts, and classroom (academic) contexts selected for this study offered
51,403 words. The subcorpora selected were academic spoken inter-
actions on social issues occurring in classroom discourse, mainly of
tutorials, plus a few brief lectures. The L1 data were taken as the norm
for comparing the L2 data sets, including number of words, level and
convention of transcription, and turn codification.
Chinese data were collected from eight sessions of interaction
between students and their facilitators at a university in Shaanxi
province, Central China. All participants were upper-intermediate to
advanced learners of English. Like the L1 data, the Chinese data con-
sisted of recordings of interactive discussions on social issues, with
between eight to ten participants in each session. The word count for
the Chinese transcript was 51,263, drawn from seven hours of record-
ing. As all the teachers were L1s of English, the recording was arranged
to capture classroom discussions run by a facilitator who was an L2
learner, to neutralize any possibility that an L1 presence could distort
the naturalness of the language used by participants.
The Persian data consisted of seven hours of video-recorded interac-
tion between a teacher whose mother tongue was Persian, not English,
and upper-intermediate to advanced learners of English. The data were
collected at Azin-E-Mehr language school, located in the city of Lahijan,
Northern Iran. In line with the Chinese data, these data also excluded
formal teaching, which would center around the teacher’s talk and
reduce the amount of naturally occurring conversation and interaction
in class. The data comprised discussions of social topics. The transcript
of the Persian speakers’ data contains 51,344 words, drawn from eight
sessions, each with seven to 20 students. For anonymity, a few names
appearing in the data have been changed.
Great effort was made to keep the three data sets as comparable as
possible, including the same data size for each group, and topics for dis-
cussion in both L2 speaker groups were kept similar to those discussed
in the L1 data, all of social issues. L1 groups were led by L1 teachers
and L2 groups by L2 teachers and facilitators, to ensure the data sets
were maximally comparable and valid. These measures mean that the
reliability of the data collection in this study is high, and the use of
video-recording has provided the researchers with the opportunity to
improve the accuracy of transcription.
3.4 Data analysis
The data in this study was analyzed according to the categories shown
in Tables 3.2 and 3.3, which present forms of vague language realization.
As Table 3.2 shows, in this study vague language is analyzed in five
lexical categories: subjectivizers, possibility indicators, vague quanti-
fiers, vague intensifiers, and placeholders. The term ‘placeholder’
is borrowed from Channell (1994), but with much wider scope in
this study. The lexical investigation is at micro level, particularly
corresponding to the first two research questions (a. and b.) listed
in Section 1.1: the realization and frequency of vague language in
the data, and how its form and distribution differ among the three
groups.
Table 3.2 Vague language lexical categories
Level of Form (micro-level) Examples

analysis
Subjectivizers: I think, I guess, I believe

diminishing the assertive
or imposing tone
(Blum-Kulka et al. 1989).
Possibility indicators: maybe, may, might, probably,
expressing possibility possible
involved in a statement
(Prince et al. 1982)
Lexical level Vague quantifiers: some (of), much, many, a lot of,
non-numerical expressions most (of ), (a) few, a little, lots
used for referring to of, a lot, majority
non-precise quantities (Ruzaitė
2007, p. 41).
Vague intensifiers: really, very, actually, so, too,
intensifying the tone of a quite
speech (Zhang 2011, p. 574).
Placeholders: dummy nouns something, thing, things,
in place of item names someone, anything, somebody,
(Channell 1994, p. 164). anybody
Methodology 45
The lexical analysis was conducted using Wordsmith Tools (Mike Scott
2010 version 6.0, for example concordance function) for the Chinese
and Persian speakers and L1 data. This program was used to acquire
information regarding the types and frequency of vague expressions
used and to provide information on the most and the least used vague
expressions and vague language clusters (vague expressions + other
collocated words). A Chi-square test was applied to statistically examine
the significance of differences among the three groups.
While the lexical level analysis in Table 3.2 is a micro-study, the
pragmatic analysis set out in Table 3.3 is a macro-analysis. This study
investigates three major functions of vague language: mitigation, giving
Table 3.3 Vague language pragmatic function categories
Level of analysis Strategies Examples

(macro-level)
Mitigation: reducing - self-protection: protecting self

the risk of offending against being proven wrong later
others, being polite (Channell 1994)
and meeting the needs - politeness: avoiding or reducing
of face (Zhang 2011) conflict (Ruzaitė 2007)
- downtoning: softening the tone
of speech (Zhang 2011, p. 574)
- self-distancing: distancing self
from claims (Ruzaitė 2007)
Functional level Giving the right - approximation and quantifica-
amount of information: making an approximation or
tion: providing an expressing vague quantity (Zhang
appropriate amount of 2011, p. 574)
information (Channell - emphasizing: emphasizing with
1994) a strong tone (Cheng 2007; Zhang
2014)
- possibility: expressing uncertain
degrees of possibility (Prince et al.
1982)
Discourse manage- - repair: making corrections in
ment: Facilitating speaking
the structural flow of - hesitation: solving oral discourse
speech and conducting production problem (Khurshudyan
discourse management 2006)
(Ruzaitė 2007, p. 187) - turn management: helping the
interlocutors realize how and
when to take or hand over speak-
ing turns
the right amount of information, and discourse management. The prag-

matic study corresponds to research questions 3 and 4 listed in Section
1.1: how and why vague language is strategically used and the impacts
and diversities of cultural and linguistic factors on its use across the three
groups. This level of analysis involves investigation of the function and
possible motivation of vague language used by the three groups, to find
inter-language and cross-cultural similarities and discrepancies.
A rigorous study of vague language requires a comprehensive and

adequate theoretical foundation and analysis system. As Chapter 2 con-
firms, this study considers that vague language research links to Grice’s
cooperative principle, to relevance theory, and particularly to Zhang’s
notion of elasticity. To achieve an in-depth understanding of vague
language use, a multifaceted analysis is needed. This study conducted
its investigation using a mixed methods approach: a quantitative
lexical analysis and a qualitative function analysis. The lexical analy-
sis considered the frequency of vague expressions and their clusters,
including five categories (subjectivizers, possibility indicators, vague
quantifiers, vague intensifiers and placeholders). A Chi-square test was
applied to statistically examine the significance of differences in using
the categories among the three groups. The functional analysis was a
qualitative examination of the pragmatic and strategic properties of
vague language use by each group of participants. The quantitative and
qualitative analyses support and complement each other, providing a
rich account of vague language use in naturally-occurring interactions
in classroom settings.
4
Lexical Analysis of Vague Language
This chapter presents the lexical analysis of the five vague language
categories by examining the frequency of vague expressions and their
clusters. A cluster refers to the grouping of a vague expression with
adjacent expressions.
4.1 Subjectivizers
Subjectivizers are also called epistemic phrases (see, for instance,

Kärkkäinen 2010), and are markers of stance and attitude expressing
a speaker’s subjective opinion. Four subjectivizers are considered:
I think, I guess, I don’t know, and I believe. I guess is normally used in
informal situations, and I believe tends to be found in relatively formal
contexts. I don’t know carries a negative meaning, often used to indicate
a speaker’s uncertainty, evasive attitude, and the like. Ruzaitė (2007:
158) asserts that ‘Hedges with I convey the speaker’s stance and his/her
attempt to distance him/herself’.
As can be seen in Table 4.1, the three groups of participants demonstrate
differences in their use of subjectivizers. The Chinese were very different
from the other groups in their overall use of this category. While the
Persian speakers and L1s were found to be different by 77 occurrences,
the Chinese used this category 2.6 times more than the Persian speak-
ers and 3.6 times more than the L1s. The difference in the frequency
of use of all four subjectivizers is statistically meaningful, p < 0.01
(χ² = 177.915, d.f.6).
Similarly, individual subjectivizers were distributed unevenly. The differ-
ence seems minor between the Persian speakers and L1s, although there
was a trend for each item to be used more often by the Persian speakers,
apart from the least frequently used item, I believe, which shows an even
47
Table 4.1 Distribution of subjectivizers
Item L1 (n = 205) CSLE (n = 741) PSLE (n = 282)
I think 161 (79%) 732 (99%) 207 (73%)

I guess 23 (11%) 1 (0%) 41 (15%)
I don’t know 13 (6%) 5 (1%) 26 (9%)
I believe 8 (4%) 3 (0%) 8 (3%)
Total 205 (100%) 741 (100%) 282 (100%)
Note: All tables are ranked according to the frequency of the item in L1 interaction, and this
table is adapted from Zhang and Sabet (in press: 7).
distribution. The Chinese were again very different. They unevenly used
the four subjectivizers, concentrating disproportionately on I think (732
tokens, 99 per cent of the total subjectivizers used). The remaining three
items combined made up only 1 per cent, indicating an extremely focused
preference. It seems that the Chinese group required nothing but I think
to express a stance or attitude. While L1s and the Persian speakers also
preferred I think, they demonstrated similarity in the ranking order of the
other subjectivizers, which the Chinese barely used.
4.1.1 I think1
As Table 4.1 shows, the most remarkable difference in the use of indi-
vidual subjectivizers among the three groups emerges in the most
frequently used expression, I think. The Chinese showed a strong prefer-
ence for it, while the other two groups used it less consistently. The L1s
used it four and a half times (161 tokens) less and the Persian speakers
three and a half times (207 tokens) less than the Chinese (732 tokens).
The heavy use of I think may be associated with L2 learners’ inadequate
language proficiency and their discourse management, such as delay-
ing, searching for words and so on.
The three groups differed in positioning I think in a clause, par-
ticularly in clause-initial position and as a turn-initiating device. The
Chinese used I think in clause-initial position, 620 (85 per cent) times,
the Persian speakers 165 (80 per cent) times, and the L1s 122 (76 per
cent) times. The Chinese preferred I think in clause-initial position
roughly five times more than L1s, and four times more than the Persian
speakers, who used it about 35 per cent more than the L1s did. I think
in the turn-initial position was used 256 times by the Chinese, 89
by the Persian speakers, and 52 by the L1s. The L2 groups both used
1
A version of the data and some data analysis in this section has been published
in Zhang and Sabet (in press).
Lexical Analysis of Vague Language 49
more I think in clause-initial and turn-initial positions, indicating that

in addition to its other functions, I think is used by L2 speakers at the
beginning of utterances as a form of discourse management. This sup-
ports the claims that one of the structural functions I think performs
in initial position is to keep the speech smoothly flowing, particularly
in L2 discourses (Kaltenböck 2010; Kärkkäinen 2003; Stenström 1995).
The occurrence of I think in clause-final position also demonstrates
differences among the groups. It is found most frequently in the
Chinese data (37 times, 5 per cent), then in the Persian data (15 times,
7 per cent), and least in the L1 data (five times, 3 per cent). The Persian
speakers used the most clause-final position. When I think is positioned
at the end of a clause, it conveys an afterthought. The data shows that
I think in this position performed tentative and mitigating functions,
supporting claims made in Aijmer (1997), Conrad and Biber (2000),
Kaltenböck (2013), Simon-Vandenbergen (2000), Zhang (2014), and
Zhang and Sabet (in press).
A variety of clusters of I think emerged in the data, giving a clearer
picture of the linguistic behaviors of I think in context. The cluster I think
I was used by the Chinese most frequently (68 times, 9 per cent), by the
L1s 13 times (8 per cent), and by the Persian speakers the least (four times,
2 per cent). The Chinese and L1 groups showed more inclination to
use this cluster in classroom interaction, but the Persian group tended
to replace I think I with I think we. An interesting trend emerges in the
use of I think we: the Chinese and Persian participants used the cluster 45
(6 per cent) and 12 (6 per cent) times respectively, but the L1s used it
only three times (2 per cent). This is the reverse of the distribution of
I think I, which the Persian speakers used the least and the L1 group the
most. The contrastive singular I and plural we were preferred by different
groups, with the Persian speakers leaning toward the plural we and the
L1 group favoring the singular I. The extract below indicates this at work.
(4.1): This is a discussion between two Chinese participants over two
turns. They are talking about what they would like to be in the future.
S1: If I have a chance, if I have a chance, I think I can, huh; I want to

be a French interpreter. I think we can cooperate. (Ch: 4: 2542)
S3: Yes, yes. One of my, one of my close friends, her sister is a French
interpreter…. (Ch: 4: 255)
In (4.1), the Chinese speaker used I think we to include the hearer S3 in

the conversation. One reason for the Persian group to prefer I think we
2
Ch = Chinese data, 4 = session number, 254 = speaking turn count, S1 = Speaker 1.
over I think I is that they tended to pay more attention to the hearer
and were less authoritative. The L1s preferred I think I over I think we,
indicating that they tended to be more self-oriented and assertive.
The data reveal two opposing clusters involving I think: I think that
is vs. I think that + subject. I think that in this study is divided into two
categories according to the function of that: the first is where that serves
as the subject (pronoun) of the sentence, followed by is as the verb; the
second is where that serves as a conjunction. The first usage is demon-
strated below.
(4.2): This is a discussion between two L1 participants over two turns.
They are talking about whether newspapers should be granted access to
people’s photos.
S2: … be on the net or something like that, [S1: mhm] and so so

people are gonna go see it and I think that is a gross invasion of
privacy, [S8: mhm] to have your pictures of your, like if I was dead
and I had a autopsy. (L1: 1:49)
S5: But th- they they said exclusively though that’s not their expressed
intent for. (L1: 1:50)
In (4.2) that is a subject. The Persian speakers used this cluster only
twice (1 per cent), the Chinese 16 times (2 per cent), and the L1 group
19 times (12 per cent), a much higher percentage than either L2 group.
In this case, the two L2 groups were similar to each other but different
from the L1, providing a case of L1 vs. L2.
The second use is illustrated in (4.3), where the function of that in
I think that shifts from a subject to a conjunction introducing a clause.
(4.3): This is a discussion between two Persian-speaking participants
over two turns. They are talking about problems that people experience.
S5: … Actually by ‘we’ I mean all the people living in the world. You
know I think that people in the world suffer from spiritual crisis.
(P: 4: 48)
S1: There might be a special crisis in the world. What is happening in
Iran? (P: 4:49)
In this excerpt, that brings in a clause in the form of I think that + subject,
performing a different role from that of I think that is. The Persian speak-
ers used this pattern 16 times (8 per cent), more than either the Chinese
(eight times, 1 per cent) or the L1s (four times, 2 per cent). When the
function of that shifted, the preferences of the Persian speakers and L1s
swapped, so that the Persian speakers used it most and the L1s least.
Another cluster that emerged in the data is I think + negative sen-

tence. While the L1s did not use this form, the other groups did, the
Persian speakers 19 times (9 per cent), and the Chinese 25 times (3 per
cent). The L1s used the alternative I don’t think instead. The follow-
ing two excerpts illustrate the contrastive forms employed by L1 and
L2 groups.
(4.4): This is a discussion between two Persian-speaking participants
over two turns. They are talking about how differently respect is inter-
preted in different cultures.
S7: Ok, first of all, I must say the culture. (P: 6:442)
S3: I think it is not cultural. I think it is not cultural, whereas the L1s
prefers. (P: 6:443)

They are talking about how the press does not take measures to prevent
an incident.
S2: The press has already put heat on ’em they’re putting heat on
themselves. I don’t think it’s necessary. (L1: 1:61)
S1: Well they’re obviously not having enough heat put on them
because it keeps happening. [S2: I think (xx)] I mean for at least it’s
possible. Yes? (L1: 1:62)
The I don’t think pattern occurred 22 times (14 per cent) in the L1 data
and 19 times in the Chinese (3 per cent), while the Persian speakers made
use of this cluster six times (3 per cent). An interesting phenomenon is
that the Chinese used both patterns (I don’t think plus I think + not) at
similar rates, making any attempt to explain the patterns in relation to
L1 and L2 distinction murky (see Zhang & Sabet (in press) for a more
detailed discussion of the two patterns).
The occurrence of the conjunction but followed by I think expresses
a contrast. The Chinese data offered 33 occurrences (5 per cent), more
than the Persian speakers (23, 11 per cent). The least use of I think but
was among the L1s, seven times (4 per cent). Another cluster of I think
is discourse marker + I think: for example, I mean I think. No such clus-
ter appeared in the two L2 data at all, but the L1s used it seven times
(4 per cent). The other cluster is you know I think. This time the L1 group
did not use the cluster, while the Persian speakers used it nine times
(4 per cent) and the Chinese twice. The Persian speakers did not use
I mean I think, but used you know I think; the L1s did not use you know
I think but used I mean I think. These two discourse marker + I think
clusters were used for discourse management, for instance to repair the
speaker’s speech. You know I think also performed other functions, such
as drawing the attention of the hearer, eliciting the hearer’s agreement,
and creating informality. The Chinese were not inclined to use either
expression. The evidence of the use of these clusters suggests that the
L1 and Persian groups tend to go opposite ways, while the Chinese take
the middle ground.
In this study, for tables where a total does not appear in the last row,
one item may be classified under more than one cluster. Thus the total
of a column may sum to more than 100 per cent.
The results in Table 4.2 show two patterns emerging: L1s vs. L2s, and
L1 vs. the Persian speakers, with the Chinese in the middle. L1s and L2s
preferred opposite I think clusters, as shown in items 1, 2, 3, 5, 8, 10,
and 14. For example, in item 14, L1s used I mean I think seven times
(4 per cent), but no use occurred in any of the L2 conversations. On the
other hand, both L2 groups used I think (item 1) in clause-initial posi-
tion more than the L1s. Another interesting phenomenon is that the
L1s contrast with the Persian speakers, as in items 6, 7, 9, and 13. For
example, the cluster in item 7, ‘I think that is’, was used most by the L1
Table 4.2 I think positions and clusters
Item L1 (n=161) CSLE (n=732) PSLE (n=207)

(% of 161) (% of 732) (% of 207)
1. I think (clause-initial) 122 (76%) 620 (85%) 165 (80%)

2. I think (turn-initial) 52 (32%) 256 (35%) 89 (43%)
3. I think (clause-final) 5 (3%) 37 (5%) 15 (7%)
4. I think I 13 (8%) 68 (9%) 4 (2%)
5. I think we 3 (2%) 45 (6%) 12 (6%)
6. I think that + subject 4 (2%) 8 (1%) 16 (8%)
7. I think that is 19 (12%) 16 (2%) 2 (1%)
8. I think + negative sentence 0 25 (3%) 19 (9%)
9. I don’t think 22 (14%) 19 (3%) 6 (3%)
10. But I think 7 (4%) 33 (5%) 23 (11%)
11. I think we should 0 12 (2%) 1 (0%)
12. I think I will 0 12 (2%) 0
13. You know I think 0 2 (0%) 9 (4%)
14. I mean I think 7 (4%) 0 0
group and least by the Persian. In item 13, the situation is reversed. In
both cases the Chinese group sit in the middle, using each particular
item less than one group but more than the other. The remaining three
cases in Table 4.2 (items 4, 11, and 12) were used most by the Chinese
group.
Table 4.2 illustrates that the L2 groups preferred to put I think in the
beginning of a clause or a speaking turn more than the L1 group, and
also to put I think in the end of a clause where it takes a mitigating func-
tion. The L1 group used more I think I but less I think we; the Persian
speakers did the reverse. I think we may suggest a sense of collaboration.
The L1 group solely used I don’t think, while the Persian speakers and
to a lesser degree the Chinese showed a strong preference for I think +
negative sentence—a less authoritative replacement for I don’t think.
Based on the evidence of the use patterns of items 3, 4, 5, 8, and 9,
it appears that the Persian group was more tentative and cooperative,
and less authoritative, than the L1 group. The evidence points to the
Chinese group as having elements of both traits: more certain than the
Persian speakers but less affirmative than the L1s. The Persian group
also preferred the formal usage of I think that + subject more than the L1
and Chinese groups, but both L2 groups stayed away from the structure
I think that is. This is supported by other evidence: the Persian speakers
used more but I think than the L1 group to gently move to a contrastive
argument; and used more you know I think to create intimacy before
expressing an opposing viewpoint. The Persian speakers stand out in
their use of I think clusters, in that they tended to use them to express
disagreement or contrast without confrontation (see Chapter 6 and
Zhang & Sabet (in press) for more discussion of the issues raised in this
paragraph).
Similarities as well as discrepancies were revealed among the three
groups, but with more distinctive differences between L1 and L2, par-
ticularly between the L1 and Persian speakers.
4.1.2 I guess
Compared with I think, I guess seems to be more informal and to
suggest a greater degree of tentativeness. This was the Persian and L1
speakers’ second most frequently used subjectivizer (41 tokens, 15 per
cent; 23 tokens, 11 per cent) in this study. It was also the least common
for the Chinese, with a frequency of one: a huge contrast with their
preference for I think.
I guess was used by the Persian speakers in clause-initial position 15 times
(37 per cent), far ahead of the L1s who used it five times (22 per cent).
It was used in clause-final position seven times (17 per cent) by the Persian
speakers, and only twice by the L1, so the Persian speakers again used it
most; the Chinese provided only one occurrence. The use of I guess after
conjunctions for both the Persian speakers and L1s were almost equal (five
for L1, six for Persian), demonstrating a similarity in combining I guess
with conjunctions; however, the overall distribution of I guess reveals that
substantial intergroup disagreement exists in terms of both frequency and
pattern of use.
4.1.3 I don’t know

As shown in Table 4.1, I don’t know was the third most frequently occur-
ring subjectivizer in the Persian speakers and L1 data, and the second
most commonly used item by the Chinese. I don’t know can serve three
different functions in communication: as a literal term (the speaker lit-
erally does not know), as a shield, and as a discourse marker (such as a
filler). The second pragmatic function, the shield, is the focus of this
study; the third function is also relevant in that vague language plays
an important role in discourse management.
They are talking about how to browse a particular search engine.
S1: …. now if you still wanna enter in, I don’t know maybe this
is gonna, be what set six set seven set set eight, I don’t know
where it’s gonna end, but let’s say it ends at set thirteen? And ….
(L1: 3:53)
S11: Where would you put the parenthesis in the second line?
(L1: 3:54)
In (4.6), the first instance of I don’t know seems a shield, implying the
speaker is unsure about where it may end. This uncertainty is reinforced
by the use of maybe, suggesting that the speaker has some vague idea
about where it may end, in the range of set six to set eight. However,
the speaker is not entirely sure: I don’t know and maybe hedge this sug-
gested end. The second I don’t know appears to be literal: at this point
the speaker declares no knowledge of the issue in question.
They are referring to morals as an example of how a large society is
structured.
S1: … um, it’s not immoral, to cut class well maybe it—I mean you
could ma- maybe someone could make an argument like, I don’t
know your parents paid all this money and, you made a promise
to them to go to class and so it’s immoral to cut class but, um, let’s
say uh, let’s say for the sake of argument you know. (L1: 2:68)
S2: The grade. (L1: 2:69)
The I don’t know in (4.7) is somewhat detached from its neighboring

words, and is more of a filler in this case. The first part of S1’s utterance
is a hedged and unclear statement about cutting class and immoral
behavior, and the use of I don’t know becomes a turning point to lead to
a clear assertion that cutting class is immoral, before becoming hedged
again at the end of the speaking turn.
It is around the end of class time and the lecturer is asking one of the
students a question.
S2: Meredith I don’t know who you are, or where you were sitting.
(L1: 3:556)
S24: I’m Meredith. (L1: 3:557)
In (4.8) the role of I don’t know is simple: it literally means ‘I don’t

know’; that the speaker has no idea. This function is not particularly rel-
evant to this study. Like their employment of I guess, the Persian speak-
ers (26 tokens, 9 per cent) used more I don’t know than the other two
groups. The L1s used this phrase 13 times (6 per cent) and the Chinese
a mere five (1 per cent). Most cases of I don’t know clusters, serving as
subjectivizers, occurred either with another vague expression such as
maybe or with a vague discourse marker such as hmm. Again, the Persian
speakers used most of these clusters (12 times); the L1s used them eight
times, and the Chinese only three.
4.1.4 I believe
It is commonly perceived that among the four subjectivizers listed in
Table 4.1, I believe is the most formal and the least used. This could be
because the three data sets in this study are of spoken and face-to-face
interactions, so a formal word would be somewhat less appropriate than
an informal word. Both the L1 and the Persian speakers used I believe
eight times (4 per cent and 3 per cent, respectively), showing some con-
sistency. The Chinese used it three times. This trend is in line with the
common perception.
Further examination reveals that the Persian speakers used I believe
five times in clause-initial position and three times in clause-medial
position; all uses of I believe by the L1s appeared in clause-mid position.

The Chinese group was closer to the Persian pattern, using I believe in
both clause-initial (twice) and clause-medial (once) positions. While
the small overall frequency of I believe in the data makes it difficult to
generate any pattern of usage, it seems clear that the participants in the
data avoided I believe, a relatively formal subjectivizer. When they did
use it, unlike the two L2 groups, the L1s preferred to use I believe in the
middle of a clause rather than at the beginning.
As Figure 4.1 shows, the Chinese group preferred subjectivizers most,
with 741 uses over the Persian speakers’ 282 and the L1’s 205. What
stands out in the Chinese data is that almost all occurrences of sub-
jectivizers involved I think, while the other two groups demonstrated a
more scattered distribution with I think still dominant but I guess, I don’t
know, and I believe also in evidence. I think was used by the Chinese
723 times, a very high concentration in comparison with the Persian
speakers (207) and the L1s (161); but almost none of the other three
categories were used at all. For instance, I guess was used only once by
the Chinese, but 23 times by the L1s and almost twice as often by the
Persian speakers. The same trend occurs with I don’t know: the Chinese
used it once, the L1 group 13 times, and the Persian speakers 26.
L1 Speaker CSLE PSLE

800
732
700
600
500
400
300
161 207
200
100 41
23 13 26 8 8
1 5 3
0
I think I guess I don’t know I believe
Figure 4.1 Frequency distribution of subjectivizers

The Chinese also used I believe less frequently than the other two
groups.
Figure 4.2 illustrates the percentage distribution of the subjectivizers
by the three groups. A striking similarity lies in the percentile rank-
ing of the Persian speakers and L1s (in descending order: I think,
I guess, I don’t know and I believe); the Chinese group demonstrates a
totally different pattern. The percentage values of the Persian
and L1 data are quite close, but as the Chinese did not use I guess
and I believe in their interactions, their entire use of subjectivizers
differed substantially. I think comprises 73 per cent of the Persian
speakers usage and 79 per cent of L1, but the Chinese group’s heavy
use of I think reached 99 per cent of the subjectivizers used in their
interactions. Of the other three subjectivizers, the Persian speakers
used more I guess and I don’t know than the L1s: 15 per cent vs. 11 per
cent and 9 per cent vs. 6 per cent. For I believe, the situation was
reversed, with the L1 group using this more than the Persian speakers
(4 per cent vs. 3 per cent).
What is shown from the summaries in Figure 4.1 (raw frequency)
and Figure 4.2 (percentage frequency) is that there is a discrepancy in
L1 Speaker CSLE
4% 0% 1%
6% 0%
11% I think I think

I guess I guess
I don't know I don't know
I believe I believe
99%
79%
PSLE
3%
9%
I think
15% I guess
I don't know
73% I believe
Figure 4.2 Percentage distribution of subjectivizers

the distribution patterns of subjectivizers, with the Chinese behaving

distinctly differently from the L1 and Persian groups.
4.2 Possibility indicators
Analysis of possibility indicators is an investigation of expressions that

express uncertainty on the part of the speaker. Like subjectivizers, pos-
sibility indicators demonstrate diversity in both frequency and pattern
of use among the three sets of participants, which are statistically mean-
ingful p < 0.01(χ² = 269.453, d.f.8).
The Chinese group again used possibility indicators more than
the other groups, but not as heavily as in the case of subjectivizers
(379 vs. 741). The positions of the Persian speakers and L1s with
regard to the overall frequency of possibility indicators are inverted:
this time L1s used more possibility indicators than the Persian speak-
ers (238 vs. 190).
A pattern shown in Table 4.3 is that the L1s employed the first four
possibility indicators much more evenly than the two L2 groups: maybe
64 times, may 56, might 56, and probably 42. The frequency of these
items in the Chinese and Persian speakers’ data reveals a less even pat-
tern: the Chinese had 312 occurrences of maybe, 50 of may, ten of might,
five of probably and two of possible. In a similar trend, the Persian group
used maybe 156 times, may 15, might 13, and probably once. The L1’s pat-
tern of possibility indicators for the first four items was evenly spread,
while the L2 speakers showed quite scattered distribution. For the last
item in Table 4.3, possible was used mostly by L1s (20 times) and less by
L2 speakers (twice by the Chinese, thrice by the Persian speakers). This
again supports the trend of L1s preferring to use a variety of possibility
indicators to suit different discourses.
Table 4.3 Distribution of possibility indicators
maybe 64 (26%) 312 (82%) 156 (81%)

may 56 (24%) 50 (13%) 15 (8%)
might 56 (24%) 10 (3%) 13 (7%)
probably 42 (18%) 5 (1%) 1 (1%)
possible 20 (8%) 2 (1%) 5 (3%)
Total 238 (100%) 379 (100%) 190 (100%)
4.2.1 Maybe
A similarity among the three groups in using maybe is that all three
preferred it over any other possibility indicator. As Table 4.3 shows,
Chinese used maybe 312 (82 per cent) times, the highest frequency,
about twice as much as the Persian speakers (156 times, 81 per cent) and
five times as much as the L1 group (64 times, 26 per cent). Maybe is the
only item in the category of possibility indicators that the Chinese and
the Persian speakers used more often than the L1s. Except for maybe,
the L1 group used all other categories more than the L2s, including may,
might, probably, and possible.
There were various positions where maybe was placed. In clause-initial
position, it occurred 128 times in the Chinese data, 85 times in Persian
and only 11 times in L1 data. It is obvious that L2 speakers used it about
ten times more than L1s. Turn-initial maybe also was used differently
by the three groups, with the pattern following that of clause-initial
maybe: the Chinese used the most (59), the Persian speakers second-
most (44), and again the L1s least (five); again the L2 speakers’ use was
approximately ten times that of the L1s.
The discrepancy was smaller in the cluster of conj + maybe. The L2
groups used it similarly (24 and 23, respectively), and the L1 group
13 times. This time, the difference between L1 and L2 groups was double,
much smaller than with maybe in clause and turn initial positions. The
pattern emerging is that the Chinese preferred all three maybe positions
and clusters the most, the L1 preferred them the least and the Persian
speakers sat in the middle. The similarity between the two L2 groups
was bigger than between L1 and L2s.
The Persian speakers used 55 per cent of maybe in clause-initial posi-
tion, but only 15 per cent for conj + maybe. The Chinese used 41 per cent
of maybe in clause-initial position, but only 8 per cent for the cluster
conj + maybe. In contrast, the L1 group used 20 per cent conj + maybe
but only 8 per cent maybe in turn-initial position. The gaps for each
group between the most and the least preferred are: 40 per cent for the
Persian speakers, 33 per cent for the Chinese, and only 12 per cent for
the L1s. This again supports the trend of L1 using vague expressions
more evenly than L2 speakers, the same trend encountered in their use
of subjectivizers, as shown in Section 4.1.
The following excerpts show how maybe was used. Both the Chinese
and Persian groups frequently placed maybe at the beginning of a sen-
tence, and in situations when taking over from another interlocutor,
maybe was employed as a turn-taking device.
(4.9): This is a discussion between two Persian participants over two

turns. One has been asked to talk but refuses to do so.
S3: You are forcing. Maybe now she doesn’t have anything to say.
(P: 4:150)
S1: Maybe, we are respecting you ladies. (P: 4:151)
In (4.9), S3 chooses maybe as a clause-initial lead-in, expressing some

kind of guessing. S1 uses maybe to take over from S3 and start his turn.
The second maybe is not exactly the same as the first: while the second
maybe could express S3’s uncertainty whether ‘she’ has something to
say, it may also simply be repeating the maybe used by S1 to show agree-
ment with what S1 says.
In the Persian speakers’ data, there were five occurrences of maybe
after the discourse marker OK, which was not found in either the
Chinese or L1 data.
(4.10): This is a discussion between two Persian participants over
three turns. They are talking about how limitations have been imposed
on society.
S7: But it is a kind of limitation. (P: 6:426)

S5: Ok. It is good for us. (P: 6: 427)
S7: Ok. Maybe the other things are good for you. (P: 6:428)

turns. They are talking about the criteria for choosing a girlfriend or a
boyfriend.
S8: So you don’t, you don’t care about her past? (P: 6:781)
S2: No, no. But I am talking about the effects, Okay? Maybe the effects
will continue. Now we are going to start talking, ok? Because …
(P: 6:782)
In (4.10), S7 uses OK to confirm agreement with the previous speaker,

then maybe to express possibility. The general tone indicated is gentle
and friendly. The situation in (4.11), however, is different, as S2 uses
Okay with a raised question tone. This usage is not an agreement, but
rather sounds like a disagreement or even confrontation. Although
maybe in (4.11) is expressing uncertainty, at the same time, it shows a
firmer stance. That is, while maybe in this context has some degree of
uncertainty, it is less tentative than the maybe in (4.10).
Regardless of the position of maybe in the clause, the three groups

demonstrate different performances in using maybe before a verb. The
Chinese used maybe + verb most (23 times, 7 per cent), the Persian
speakers 14 times (9 per cent) and the L1 group ten times (16 per cent).
The Chinese again used this phrase more than twice as much as the
L1 group. Another noticeable difference occurred in the preference for
types of verb in the cluster maybe + verb, with the Persian speakers pre-
ferring know and think, with six and four occurrences respectively before
maybe. They also used maybe + because five times, a construct absent
from the L1 data.
The pattern maybe + pronoun (subject position) proved to be popular.
The Chinese used it a total of 122 times (39 per cent), the Persian speak-
ers 78 times (50 per cent) and the L1s 18 times (28 per cent). The Chinese
and Persian speakers were keenest on placing maybe before a pronoun.
The Chinese used ten times more maybe + pronoun than the L1 group.
Participants preferred certain types of pronoun over other types and
generally speaking the third person pronoun (for example maybe s/he …)
ranked first, the first person pronoun second, and the second person pro-
noun last. The ranking of pronoun in maybe + pronoun for the Persian
speakers was 47 (third person)/14 (first person)/17 (second person); the
L1 ranking was the same (9/5/4). The Chinese ranking was slightly differ-
ent: 53 (first person)/ 51(third person)/18 (second person). It seems that
in terms of the use of maybe + pronoun, the L1 and the Persian speakers
are the same in ranking order: third top, then first, then second. The
Chinese ranking reverses the top two, putting the first person pronoun at
the top: first top, then third, then second. What seems to be the same in
all the three groups is that second person pronouns are the least favorite.
The data show that maybe tended to cluster with negations. The
Persian speakers used negative statements after maybe 14 times, twice as
often as the L1s (seven times) and a third more than the Chinese
(ten times).
(4.12): This is a turn from an L1 speaker. He is explaining why an item
is not found in a search engine.
S1: Because they have the potential, maybe not necessarily in the
ERIC database. (L1: 3:275)
In (4.12), the speaker uses maybe not to tactically exclude certain things;
the cluster makes the speaker’s tone softer to establish empathy with
the hearer. The use of maybe not is part of hedging taking place in the
utterance.
The data shows that maybe was placed before an ellipsis or similar
structure. This might be linked to the nature of maybe, which expresses
a somewhat weak possibility; thus, omitted utterances seem natural
companions. The ellipsis can enhance the effect of maybe. The cluster
of maybe + ellipsis (or the like) occurred almost equally among Persian
speakers (17 times) and L1s (18 times). The Chinese used it 31 times,
almost as much as the other groups combined.
Other maybe clusters emerged in the data, most from the Chinese
group. I think maybe appeared in the Chinese data 20 times and there
were six uses of maybe I think (swapping the order of maybe and I think);
but none of these was used by Persian or L1 speakers. The Chinese
also employed maybe I/we will 13 times, but the other two groups did
not use any. Another cluster, maybe it is, was used by Persian speakers
20 times and by the Chinese 17 times; none from the L1 group used it.
The trend observed in this group of maybe clusters is that the Chinese
were the leading users while the L1s were not at all interested in using
the clusters. The data exhibits a gap between L1 and L2 speakers’ lin-
guistic behaviors in the use of maybe and its clusters.
Table 4.4 shows that the Chinese used all maybe positions and clusters
the most, except maybe + negation and maybe it is. The L1 group used all
Table 4.4 Maybe positions and clusters
1. maybe (clause-initial) 11 (17%) 128 (41%) 85 (55%)

2. maybe (turn-initial) 5 (8%) 59 (19%) 44 (28%)
3. maybe + conj 13 (20%) 24 (8%) 23 (15%)
4. maybe + verb 10 (16%) 23 (7%) 14 (9%)
5. maybe + pronoun (subject) 18 (28%) 122 (39%) 78 (50%)
6. maybe + third person pronoun* 9 (50%) 51 (42%) 47 (60%)
7. maybe + first person pronoun* 5 (28%) 53 (43%) 14 (18%)
8. maybe + second person 4 (22%) 18 (15%) 17 (22%)
pronoun*
9. maybe + negation 7 (11%) 10 (3%) 14 (9%)
10. maybe + ellipsis 18 (28%) 31 (10%) 17 (11%)
11. I think + maybe/maybe + I think 0 26 (8%) 0
12. maybe I/we will 0 13 (4%) 0
13. maybe it is 0 17 (5%) 20 (13%)
Note: *For items 6, 7, 8, L1 (n=18), Chinese (n=122), Persian (n=78).

maybe positions and clusters the least, except, maybe + ellipsis. In most
cases the Persian speakers used maybe positions and clusters more than
the L1 group but less than the Chinese, holding the middle ground.
There are three exceptional cases, though: Persian speakers were the top
users of maybe + negation and maybe it is, and in the case of maybe +
ellipsis, they were the least users.
As maybe was used by the Chinese the most, it is natural that they
also chose to use the most maybe clusters. The opposite of the Chinese
was the L1 group, who not only used maybe far fewer times in their
interactions but also used the fewest maybe clusters. In the case of
maybe usage, the Persian speakers sat in between the Chinese and L1
groups in terms of the frequency of use of maybe and its clusters. Given
that the Chinese used much more maybe and clusters than the other
two groups, and if we assume that speakers who use more maybe tend to
be more tentative, then the Chinese seem more tentative than the L1
and the Persian speakers, the L1 group looks less tentative than either
L2 group, and the Persian speakers may be less tentative than the
Chinese but more tentative than the L1s.
4.2.2 May
According to Table 4.3, in this study, may appeared as the second most
common vague expression in the category of possibility indicators.
Unlike maybe, used most by the Chinese and least by the L1s, may was
most frequently used by the L1s with the Chinese the second most fre-
quent users. Against the trend shown in the maybe data where the L1s
used the least, this time the L1 group showed a strong tendency to use
may (56 times, 24 per cent), which had the highest frequency among
the three groups. The Chinese were not far behind, using it 50 times
(13 per cent) but much less than their use of maybe (312 times). The
Persian speakers used may only about one third of the other two groups
(15 times, 8 per cent), but used maybe (156 times) reasonably frequently.
The negative form of may was consistently overlooked by the two
L2 speaker groups, who used it once each, meaning that Chinese and
Persian speakers very rarely used may not in their classroom interac-
tions. The L1s showed a frequency of nine for may not. Similar to the
pattern shown in their use of all categories of possibility indicators
(shown in Table 4.3), the L1s used may and may not clusters more
diversely than the other two groups.
Distribution of may co-occurring with verbs in the data was exempli-
fied by two clusters: may be and may have. The L1s were most frequent
users of both, using may be 14 times and may have nine. The Chinese
used may be 11 times and may have six. These two clusters appeared
once each in the Persian data, showing that the Persian speakers had
little tendency to use either cluster. There were some less frequent may
clusters, including may say, may make and may wanna/want: of these the
L1 used four times each of the three clusters, and the other two groups
used them rarely if at all.
Pronoun (subject position) + maybe phrases such as you may, it may,
we may and they may revealed some patterns as well. The L1 group used
you may 18 times, the Chinese 12, and the Persian speakers seven. The
ordering was the same for it may: the L1 group used it nine times, the
Chinese three and the Persian speakers not at all. The pattern was differ-
ent for we may, in that this time the L1s did not use it but the Chinese
used it five times and the Persian speakers three. The Chinese also used
they may five times, and the other groups twice each. The cluster there
may be occurred in the L1 data six times, but was absent from L2 data.
By and large, the L1s used somewhat more may in combination with
pronouns, the Persian speakers had little interest in using this kind of
may cluster, and the Chinese sat in the middle. There were some infre-
quent uses of pronoun + may clusters, such as I may, which the L1 group
used four times, the Chinese two, and the Persian speakers never.
As Table 4.5 shows, L1s used may and all the may clusters the most,
except in two cases. In contrast, the Persian speakers used the least may
and clusters. This time it was the Chinese turn to be in the middle: in
most cases they employed may and its clusters less than the L1s but
more than the Persian group. The cluster that all the three groups used
the most was you may. The pattern shown in Table 4.5 is a part of a big-
ger picture that shows the L1s used possibility indicators more evenly
Table 4.5 May clusters
1. you may 18 (32%) 12 (24%) 7 (47%)

2. may be 14 (25%) 11 (22%) 1 (7%)
3. may have 9 (16%) 6 (12%) 1 (7%)
4. may + not 9 (16%) 1 (2%) 1 (7%)
5. it may 9 (16%) 3 (6%) 0
6. there may be 6 (11%) 0 0
7. they may 2 (4%) 5 (10%) 2 (13%)
8. we may 0 5 (10%) 3 (20%)
than the L2 groups, who concentrated disproportionately on maybe

over any other vague expression in the category (see Table 4.3).
4.2.3 Might
The third commonly used possibility indicator, might occurred with a
frequency of 56 in the L1 data, outweighing its use by Persian (13 times)
and Chinese (ten times) speakers. L1 data for might displays consist-
ency with regard to may and might, in that participants used all three
vague expressions at similar rates (maybe 64, may 56, and might 56).
This consistency is not present in the Chinese and Persian data: for the
Chinese the ordering is maybe 312, may 50, and might 10, with a huge
difference of about 30 times more maybe than might; for the Persian
speakers the discrepancy is smaller but still substantial (maybe 156,
may 15, and might 13), with about ten times more maybe than might.
The L2 groups preferred maybe much more than the other items in the
category of possibility indicators, possibly due to the constraints of their
limited knowledge and proficiency in speaking English.
Table 4.6 shows that the L1s used might clusters more than L2 speak-
ers across the board. In addition to the frequent might clusters dis-
played, there are also infrequent ones, including it might, we might, he
might, and she might.
The investigation of subjects used before might reveals that more
than half of the 31 occurrences of might came before subject pronouns
in the L1 data, and two-thirds (eight) of the Persian and half (five) of
the Chinese. The L1s showed a far greater inclination to use subject
pronouns with might than the Chinese or Persian participants, but as
percentages, the performances of all three groups fall within almost the
same range.
The L1s used six different kinds of subject pronoun in their talk,
although some occurred infrequently. The most common subject
Table 4.6 Might clusters
1. pronoun (subject) + might 31 (55%) 5 (50%) 8 (62%)

2. might be 20 (36%) 6 (60%) 4 (31%)
3. you might 12 (21%) 2 (20%) 1 (8%)
4. it might 6 (11%) 0 0
5. subject + might 5 (9%) 0 5 (39%)
6. might have 5 (9%) 1 (10%) 0
pronoun for this group was you (12 times), followed by it (six), they
(five), I and we each three times, and two occurrences of he. The Persian
speakers’ data demonstrates a preference for combining might with
only four different subject pronouns: you, I, he and she. As they rarely
used might, however, clusters of subject pronouns with might in the
Persian speakers’ data were infrequent. She might (which doesn’t occur
in the L1 data) ranked first in the Persian data with four uses, followed
by two I and one each of you and he. It seems unusual that no Persian
speakers used might with a plural subject pronoun: the one you might
that appeared was singular. By comparison, one third of the subject
pronouns before might in the L1 data were we might and they might. The
Chinese showed reluctance in using might with third person singular
subject pronouns such as he, she, and it, preferring to use this possibil-
ity indicator with a more limited number of subject pronouns than the
other groups.
A pattern observed in the data is the use of subject + might, which
was used five out of 13 times by the Persian speakers. Five out of
56 cases in the L1 data occurred at the beginning of utterances, acting as
turn initiators, but the Chinese never used might with a turn-initiating
subject in their interactions. Besides appearing at the beginning of the
clause, these occurrences turn up at the beginning of the speaking turn.
Despite the same raw frequency of use in the Persian speakers and the
L1s data, the pattern has a substantially different proportional value,
and once it is converted into a percentage, it becomes evident that a
mere 9 per cent of the overall might in the L1 data occurred in subject +
might structure, while in the Persian speakers’ data the proportion is
more than four times as much, 39 per cent.
The most frequent verb combined with might by all three groups is be,
but with different frequency. Might be appears six times (60 per cent) in
the Chinese data, four (31 per cent) in the Persian, but 20 times (36 per
cent) in the L1. Might have appeared five times in the L1 data, only once
in the Chinese and not at all in the Persian. Other less frequent might +
verb combinations include might want and might say.
4.2.4 Probably and possible

As Table 4.3 above shows, the vague expressions probably and possible
are less common than the other three possibility indicators. Probably
was used in substantially different ways by the L1s and the two L2
groups. The L1s used 42 tokens, as opposed to five by the Chinese and
merely one by the Persian speakers. Of the possibility indicators used by
L1s, probably constituted 17 per cent (42/238), but the Chinese and
the Persian speakers used only about 1 per cent of probably each. This
implies that the L1s were more diverse in their use of possibility indica-
tors in the classroom context.
Possible was the least or second least favorite by all three groups in
terms of the category of possibility indicators, as shown in Table 4.3. As
with probably, L1s made the most use of possible: in fact, of the five pos-
sibility indicators apart from maybe, the L1s used all four more than the
other groups. In contrast to the probably trend, the Chinese used possible
twice, less than the Persian speakers’ five.
A diversity of possible clusters appears in the L1 data, such as the last
possible date, the best way possible, and if at all possible. Of the five sen-
tences containing possible used by Persian speakers, two were negative,
two interrogative and one a statement. The L1s produced one negative
and one interrogative, with the rest statements. The Chinese used pos-
sible in one positive and one negative sentence each. Of all five occur-
rences of possible by Persian speakers, three appeared in clause-final
position, a signal that the speaker was going to hand over to another
speaker: this kind of possible acts as both a sentence closer and a turn
closer. In the L1 data, six instances of possible occurred in the final posi-
tion but only one acted as a turn closer, the other five being followed
by another clause by the same speaker. Of the 14 other sentences that
contained the non-final position possible, four were collocated with
that. The Chinese used one out of two possible clusters as a turn closer
in clause-final position, but the other, a mid-clause possible, was not
followed by that.
Figure 4.3 demonstrates that the Chinese used possibility indica-
tors the most, Persian speakers the least, and L1s in between. Ruzaitė
(2007, p. 158) states that ‘maybe or perhaps suggest a lower degree of
the speaker’s commitment to the truth of the claim and make the
claim less categorical’. If we agree, then the total number of tokens
suggests that the Chinese are more tentative and less committed to
their claims than the other two groups, and the L1s are less tenta-
tive and more committed than the Chinese but more tentative and
less committed than the Persian speakers. The other trend shown
in Figure 4.3 is that the Chinese almost exclusively concentrated on
maybe and may; the same pattern exists among the Persian speakers,
but usage was heavily skewed to maybe. The contrastive trend is that
the L1 group spread their uses more evenly across the five expressions
in the category. This phenomenon could be because as native speak-
ers of English, the L1s were better able to use the language resources
available to them than were L2s.

350
312
300
250
200
156
150
100
64
56 50 56 42
50
15 20
10 13 5 1 2 5
0
Maybe May Might Probably Possible
Figure 4.3 Frequency distribution of possibility indicators
The distribution of may in this study demonstrates a pattern in con-

trast with Hyland and Milton’s (1997) finding that L2 users used the
modal auxiliary as a marker of possibility twice as often as L1s. In this
study, L2 speakers did not use may more than L1s, but did use maybe
two and a half times more than the L1s did, which can be taken to sup-
port Hyland and Milton’s finding on may as both may and maybe are
markers of possibility. Another unusual trend revealed in this data is
that the Chinese and the Persian speakers rarely used probably and pos-
sible in their classroom interactions, in contrast with L1s.
The percentages of the five expressions constituting possibility indi-
cators show a substantial difference in terms of the proportion of the
items used by L1s and by the L2 groups. As can be seen in Figure 4.4, in
the L1 data, the first three possibility indicators, maybe, may and might
have an almost even distribution, comprising three quarters of all pos-
sibility indicators. In the L2 data, more than four fifths, 81 and 82 per
cent, of the distribution consists of one item only, maybe. For the L1s,
the fourth quarter consists of probably and possible, the former about
twice as often as the latter. In the Chinese and the Persian data, on the
other hand, the remaining one fifth is made up of four items: may and
might, at 8 per cent and 7 per cent, occur evenly in the Persian data,
L1 Speaker CSLE
1%
3% 1%
8%
Maybe Maybe
26% May 13% May
18%
Might Might
24% Probably Probably
24% Possible Possible
82%
PSLE
1%
3%
7%
8% Maybe
May
Might
Probably
81% Possible
Figure 4.4 Percentage distribution of possibility indicators
but may appears about four times as often as might in the Chinese data.
Probably, at 1 per cent in each L2 speaker group, compares with 18 per
cent in the L1 data. Possible at 3 per cent constitutes the second last
item of possibility indicator for the Persian group, but is the least used
item in the Chinese and L1 data. The overall picture is that the usage of
possibility indicators by the Chinese and Persian speakers is more or less
similar, whereas the L1s demonstrate a very different trend.
4.3 Vague quantifiers
The third category to contribute to an in-depth analysis of vague lan-

guage in this study is what is called ‘vague quantifiers’ in the literature
(Channell 1994; Cutting 2007; Ruzaitė 2007). This category consists
of ten items including some (of), much, many (of), a lot of, most (of),
(a) few of, a little, lots of, a lot, and majority. Their distributions are listed
in Table 4.7.
As shown in Table 4.7, as with subjectivizers and possibility indicators,
the Chinese used vague quantifiers more than the Persian and L1 speakers.
Table 4.7 Distribution of vague quantifiers
some 173 (41%) 264 (36%) 229 (53%)

much 53 (13%) 106 (14%) 40 (9%)
many 46 (11%) 163 (22%) 47 (11%)
a lot of 39 (9%) 85 (11%) 22 (5%)
most 33 (8%) 86 (13%) 38 (9%)
(a) few 21 (5%) 7 (1%) 0
a little 20 (5%) 11 (1%) 9 (2%)
lots of 16 (4%) 11 (1%) 34 (8%)
a lot 16 (4%) 8 (1%) 9 (2%)
majority 6 (1%) 0 7 (2%)
Total 423 (101%)* 741 (100%) 435 (101%)*
Note: *after rounding.
The L1 group again tended to use vague quantifiers evenly (except some),
and the L2 groups did not—although the Chinese had a relatively lower
degree of unevenness. The study of subjectivizers and possibility indicators
found that only the first items, I think and maybe, were heavily used, but the
Chinese demonstrate a different pattern in their use of vague quantifiers,
revealing a preference for using all five terms appearing first in Table 4.7,
some, much, many, a lot of, and most, more than the other groups. The
Chinese used vague quantifiers relatively more evenly than they did with
subjectivizers and possibility indicators.
The Persian speakers used some heavily, but other than that, they
used four other items in a relatively close range: many, much, most, and
lots of. The overall frequency of quantifiers by Persian speakers and
L1s is roughly the same, 435 and 423, but there are 741 instances in
the Chinese data. Statistical analysis reveals significant differences in
the use of vague quantifiers by the three groups, p < 0.05 (χ² = 211.976,
d.f.18).
There is some consistency in the use of the first five items in Table 4.7;
all three groups employed some more than the other four items. The
ranking of the five top items is almost consistent across all groups,
with the exception of a lot of in the Persian data. Of the remaining five
items, consistency is limited: for instance, L1s were the most frequent
users of (a) few, a little, lots of and a lot, whereas the Persian speakers
favored lots of.
4.3.1 Some
The most commonly used vague quantifier for all three groups, some,
occurs in the Chinese interactions 264 times, in the Persian 229, and
in the L1s 173 times. Some makes up 53 per cent of all vague quanti-
fiers in the Persian data, 41 per cent in the L1 and 36 per cent in the
Chinese. Some was used predominantly by all three groups, which sup-
ports Ruzaitė’s (2007) finding that some was the most frequent of vague
quantifier in British and American academic discourse. This study too
finds that some is not only a favorite of L1s of English, but also of L2s.
What is strikingly revealed in this study is that L2 speakers used far
more some than L1s.
Of the 229 occurrences of some in the Persian data, some + noun clus-
ters appear 160 (70 per cent) times; 144 (55 per cent) out of 264 in the
Chinese, and 74 (43 per cent) out of 173 in the L1 data. The Chinese
located about half of their uses of some with nouns, and Persian speakers
more than two thirds; the L1s used less of this some + noun pattern. Of
the two forms some + noun and some + adjective + noun, the latter was
much less preferred by all three groups. The nouns in some + noun were
both countable and uncountable. In the L1 data, most nouns in some +
noun were countable: 50 (68 per cent) countable and 24 (32 per cent)
uncountable. This trend was even more prominent in the L2 groups:
the Chinese used 129 (90 per cent) countable nouns and 15 (11 per
cent) uncountable, the Persian speakers 153 (96 per cent) countable and
a mere seven (4 per cent) uncountable, revealing an almost exclusive
preference for a countable noun plus some. This distribution is in line
with the patterns that have emerged above, that L1s tend to use various
types of vague language more evenly, and the L2 groups concentrate on
a fewer number.
Table 4.8 Some clusters
1. some + noun 74 (43%) 144 (55%) 160 (70%)

2. some + adjective + noun 25 (14%) 25 (9%) 24 (10%)
3. some (clause-initial position) 4 (2%) 10 (4%) 24 (10%)
4. some of 26 (15%) 9 (3%) 16 (7%)
5. some people 3 (2%) 9 (3%) 32 (14%)
6. some other 5 (3%) 16 (6%) 5 (2%)
7. and some 7 (4%) 16 (6%) 2 (1%)
The L2 data contain an overwhelmingly larger number of countable

nouns, indicating that these are consistently preferred by the L2 speaker
groups; L1s prefer to use countable nouns relatively moderately, per-
haps because as native speakers of English, they have more knowledge
of how to use some with both countable and uncountable nouns, and
are more competent in the use of some clusters. The Persian and Chinese
participants show less inclination for using some + uncountable noun,
suggesting that they are less familiar with this pattern.
Table 4.8 shows one similarity in some clusters across three groups
in the use of some + adjective + noun. All three groups use the pattern
at the same rate: the L1 and Chinese speakers used it 25 times and the
Persian speakers 24. However, the consistency disappears when we look
at the different types of adjective used in the cluster: positive, negative
or neutral. The pattern the L1s present seems exponential: positive
adjectives 14 times, twice as many times as neutral adjectives (seven)
and three times as many as negative adjectives (four). This sliding pat-
tern does not appear in the L2 groups. The Chinese group used posi-
tive and neutral adjectives at the same rate, both 11 times, but three
negatives: almost three times less. This suggests that they preferred
non-negative to negative adjectives. Even more differently, the Persian
speakers used all three categories at a similar rate, seven times for nega-
tive, eight for positive, and nine for neutral adjectives, showing little
preference for a particular type.
In Table 4.8, some in clause-initial position refers to occurrences in
clause-initial position by the same speaker as in the previous clause.
Despite the most frequent occurrence of some appearing in the Chinese
data, it does not occur in clause-initial position. The Persian speak-
ers showed an inclination to use some in clause-initial position, with
24 occurrences; the L1s offered only four examples. While one tenth
of all instances of some in the Persian data occurred in clause-initial
position, the Chinese and L1s were sparing of this use. Given its lim-
ited frequency, it is natural that neither group preferred it as a turn-
initiating device: the Chinese used it in this way only three times and
the L1s not at all, but turn-initiating some appeared 16 times in Persian
interactions.
As shown in Table 4.7, some was used most by the Chinese and least
by the L1s. The ranking of some of as a separate item is the reverse. As
illustrated in Table 4.8, the L1s offered 26 instances of some of, roughly
three times as many as the Chinese with nine. As with some, the Persian
speakers remained in the middle, with 16 uses of some of. They used it
in clause-initial position four times, the Chinese just once, and the L1s
not at all.
The examination of words co-occurring before some of reveals that
many of the L1s placed some of after a vague expression (mostly a sub-
jectivizer) or a discourse marker such as I mean, okay:
(4.13): This is a discussion between two L1 participants over two
turns. They are talking about the application of a search engine.
S1: … The, ERIC Clearinghouse on Assessment and Evaluation and

I would say, that this is probably just fine to use. okay? [S4: okay]
little bit out-dated in its design but, <S4: LAUGH> I mean it’s, it’s
okay I mean some of these are, I mean it’s nice that they roll over,
<SS: LAUGH> but, (L1: 3:122)
S4: Cuz I looked_ I found some terms that weren’t in the, in the book
and I ended up not using them just cuz I was [S1: oh really?] nerv-
ous about it. I I wasn’t sure like there’s bullying for violence which
I thought would have been a good term but, (L1: 3:123)
In (4.13), S1 uses the discourse marker I mean three times when recom-
mending something to S4. ‘I mean some of these are’, where I mean
appears to mitigate the force of the recommendation, making it less
imposing. Some here is a vague quantifier providing the right amount
of information. Some of is used by the speakers to reinforce their uncer-
tainty; this function of some of is missing in the Persian data and was
used by the Chinese once.
As Table 4.8 shows, the groups preferred different some clusters. The
Persian speakers made use of some people 32 times, three times more than
the Chinese and ten times more than the L1s. The Chinese used some other
16 times, over three times more than the other two groups separately.
L1s used some kind of (14 times), some of you (nine times), some of the,
some things (eight times, each), some of these (six times), and give me some
(five times); none of these clusters appeared in L2 speakers’ interactions.
The Persian speakers used there are some (ten times), some problems and
some rules (seven times each), some of them and in some tribes (six times
each), some people who and for instance some (five times each); other two
groups use none of these except there are some, used by the Chinese five
times. The Chinese use some students (16 times), to do some (ten times),
I think some (eight times), some, some (seven times), you have some and
some money (six times each), go to some, some other things, and some experi-
ence (five times each); the other two groups use none of these.
Table 4.8 shows that the occurrence of and + some was a favorite of the
Chinese and to a lesser degree of the L1s, with 16 and seven occurrences
respectively. The Persian speakers, on the contrary, use and + some only
twice. Other less frequent clusters of conjunction + some include but
some, used six times by the Persian contingent, once by the Chinese
and never by the L1s). Or some was used by the Chinese five times, three
times by the Persian speakers and twice by the L1s. The total number of
conjunctions used before some by L1 speakers was nine and by Persian
speakers 11: similar results, as opposed to Chinese group’s 22 times.
The Chinese used the pattern the most and L1s the least, just below the
Persian speakers.
The highest frequency of some occurs in the Chinese data (264 times).
The Persian data is not far behind (229 times). Despite using some less
than the Chinese and Persian speakers, the L1s used a more diverse,
wider range of some clusters.
4.3.2 Much
The booster much is often used as an emphasizing marker to enhance a
speaker’s claim. The Chinese contingent used much 106 times, twice as
often as the L1s who used it 53 times and more than twice as often as
the Persian speakers who used it 40 times. This indicates that the
Chinese were more inclined to strengthen their tone of speech than
the other two groups. Compared with the use of much, much of was
used very infrequently: four times by L1s, twice by Chinese, and none
by Persian speakers. Other more frequent much clusters are listed in
Table 4.9.
As shown in Table 4.9, very much is used most frequently by the
Chinese: 48 times, as against two or four times by the other two groups.
Table 4.9 Much clusters
1. so much 9 (17%) 19 (18%) 9 (23%)

2. how much 6 (11%) 3 (3%) 9 (23%)
3. too much 4 (8%) 14 (13%) 4 (10%)
4. very much 2 (4%) 48 (45%) 4 (10%)
5. much much 5 (9%) 39 (37%) 6 (15%)
(clause-final position) so much 2 (4%) 10 (10%) 2 (5%)
very much 1 (2%) 27 (25%) 2 (5%)
The L1 and Persian speakers preferred so much over very much and, to a
lesser degree, over too much. The ranking of booster preferences for the
Chinese is very much, so much, and too much: different from the L1 and
Persian speakers’ so much, too much and very much. The question form of
much is how much; the Persian speakers used how much nine times, the
other groups a half or a third as often. Of its use by the Persian speakers,
five occurred at the beginning of a turn; it occurred as a turn initiator
only once in both the L1 and Chinese interactions.
The Chinese showed an inclination for the heavy use of intensi-
fiers (so, very, and too) before much, a trend that was also apparent
when much appeared in clause-final position. The emergence of much
in clause-final position, modifying a verb, had roughly the same fre-
quency in the Persian and L1 data, but the Chinese showed much heav-
ier use of this: much appears six times in clause-final position in the
Persian data and five times in the L1, but 39 times in the Chinese. Two
other much clusters were found positioned in the end of a clause: so
much and very much. Again, as shown in Table 4.9, the Chinese were
the most frequent users: 27 very much and ten so much; the other
groups rarely used either.
The much clusters discussed above are constructs ending with much;
other types begin with much. For example, much + more occurs fre-
quently in the L1 data but not in L2, which implies that L2 speakers fail
to use much + comparative adjective to intensify their classroom inter-
action. The Chinese differed from other groups in using ‘very much +
conjunction’: nine times for very much and, four times for very much but,
three times for very much because. This shows that the Chinese preferred
more much clusters than the other two groups.
4.3.3 Many
Many is a typical vague quantifier conveying an unspecified but above-
average amount. It can also function as a booster, as does much. Many
occurs 163 times in the Chinese, 47 times in the Persian and 46 times in
the L1 data. There is little difference between L1s and the Persian speak-
ers in their use of many, but the Chinese use is about 3.5 times higher,
showing a strong preference for it in this cohort.
As shown in Table 4.10, the question form of how many was used
17 times by the L1, four by the Chinese and only twice by the Persian
speakers. So many was used 22 times by the Chinese, 14 by the Persian
and six by the L1 group. The Chinese also used many people over four
times more than the other groups. The L1s did not use conjunctions
Table 4.10 Many clusters
1. how many 17 (37%) 4 (2%) 2 (4%)

2. so many 6 (13%) 22 (13%) 14 (30%)
3. many people 3 (7%) 16 (10%) 4 (9%)
4. conjunction + many 0 8 (5%) 4 (9%)
5. many + adj. + noun 5 (11%) 23 (14%) 7 (15%)
6. many or many + noun in 2 (4%) 12 (7%) 19 (40%)
clause-final position
before many in their interaction, but the L2s did, the Chinese twice as
often as the Persian speakers. The three groups had a many construct in
common; ‘many + adjective + noun’, which was used five times by the
L1, 23 by the Chinese and seven by the Persian speakers. Unlike some,
which occurs in clause-initial position (as shown in Table 4.8), what
is noticeable is that many in clause-initial position is a very rare case.
Neither the Persian speakers nor the L1s puts many at the beginning of a
clause, and it occurs this way in the Chinese data only three times. The
analysis of the clause-final position use of many or many + noun reveals
that the Chinese and the Persian speakers, with 12 and 19 examples,
inclined to use many or many + noun this way; the constructs occurred
only twice in L1 data. The Persian and L1 speakers also performed simi-
larly in under-using many of, which occurred only twice in the Persian
data and once in L1, but is relatively more frequent in the Chinese, with
seven uses.
Persian speakers were the most frequent users of many or many +
noun in clause-final position and the L1s the most frequent users of
how many. All remaining items in Table 4.10 occurred mostly in the
Chinese data, indicating that the Chinese group were not only the most
frequent users of many, but also of the majority of the many clusters
discussed here.
There were also some many clusters exclusively used by one group. For
instance, the Chinese used there are many 22 times, have many 13 times,
see many, many years eight times, many places seven times, and many
opportunities and many students five times. As many occurred six times,
all in the L1 data, and so many different with a frequency of seven, and
many problems with five occurrences, appeared in the Persian data only.
Despite the extensive use of many in the Chinese interactions, the
Persian and Chinese speakers showed similarity in terms of total
frequency of many, and there is no consistent pattern in the use of its

clusters, shown as in Table 4.10. As far as the frequency distribution
of many clusters is concerned, in the use of how many and ‘conjunction +
many’, the L2 groups are similar, but in the case of many people and
‘many + adj. + noun’, L1s and the Persian speakers are closer. There is no
case where the Chinese and L1s are close; it appears that they are quite
far apart in the ways in which they utilize many clusters.
4.3.4 A lot of
Like other vague quantifiers discussed so far, the Chinese demonstrated
dominant use of a lot of, using it 85 times, as opposed to 22 times by
the Persian speakers and 39 times by the L1s: four times as often as the
Persian speakers and twice that of the L1s.
The data show that a lot of tends to cluster with countable nouns:
47 times for the Chinese, the most frequent users, 25 times for the L1s
and 14 for the Persian speakers. There is consistency among the groups
in that in all of them, more than half the cases of a lot of are a lot of +
countable noun. For a lot of + uncountable noun, which was much less
in evidence overall, the Chinese again used the pattern most with
19 instances, followed by the L1s (seven) and Persian speakers (six).
Other patterns include a lot of + adjective: the Chinese used this eight
times and the L1s five, but the Persian speakers never.
Some different a lot of clusters emerged among the groups. The
Chinese used there are a lot of eight times and learn a lot of things five, but
neither appeared in the L1 or Persian data. A lot of money was used ten
times and a lot of time four times by the Chinese, but not by L1s or the
Persian speakers. A lot of things occurred 19 times in the Chinese data,
only twice in the L1 and once in the Persian data. The cluster a lot of
people occurred in L1 data six times, three times in the Chinese and only
once in the Persian speakers’ data. This is in conflict with what Drave
(2002) finds in his research on the use of vague language by two cultur-
ally different groups: that the most common phrase of native speakers
of English is a lot of, while a lot of people is used mostly by native speak-
ers of Cantonese conversing in English. The results are in reverse in this
study: a lot of was used mostly by the Chinese, and the L1s used more
a lot of people than the L2 groups. This difference may be attributable to
the different nature of the discourses in the two studies.
4.3.5 Most
The Chinese showed a stronger tendency to use most than the other
groups: 86 times for the Chinese, more than twice as much as the
L1 (33) and Persian speakers (38). The distribution of most of follows a

similar pattern, where the Chinese led again, using it 20 times against
the Persian speakers closely following 15 times. The L1s used most of
five times, considerably less than the L2 groups. A similar trend is that
the Chinese and Persian speakers used most of in clause-initial position
four and five times respectively, whereas the L1s used most of only in
mid position, an indication of the diverse linguistic behaviors of L1 and
L2 speakers.
This study finds that most of tends to follow expressions of emphasiz-
ing, particularly in L1 data:
turns. They are talking about how a construction is distinguished by a
search engine.
S1: What’d we say? On the farm, two intervening words. Two-N farm.
So instead of using, W I use N. In fact most of the constructions
that you see out there, that use, a W, could often be turned around
to be a, two-N. So. (L1: 3:339)
S21: Hairy animals. (L1: 3:340)
In excerpt (4.14), S1 uses in fact before most of to enhance the effect of

the vague quantifier most of. In L2 data, most of tends to follow expres-
sions of uncertainty, as shown in excerpts (4.15) and (4.16):
turns. They are talking about how changes in city life contribute to
changes in children’s characters.
S4: Something like this. But nowadays because I think, huh, most of
the house especially in towns, in cities, such as big cities like Tehran
and the other cities are, the house doesn’t have any. (P: 6:603)
S2: They are like flats. (P: 6:604)
(4.16): This is a discussion between three Chinese participants over

three turns. They are talking about the plot of a movie.
S2: I think, maybe, most of you have seen the film Scrappy. (Ch: 7: 64)
S7: Yeah. (Ch: 7: 65)
S2: In this film a lot of students always make troubles to the teacher
and sometimes they put some glue, on the, the chair and the
teacher sits on it. I think that’s terrible. (Ch: 7: 66)
The expression I think in (4.15) conveys that S4 (a Persian speaker) is

not absolutely sure about the claim made in the following utterance
(‘most of the house …’). The uncertainty is highlighted by huh, indicat-
ing a hesitant moment. The same applies in excerpt (4.16), where the
Chinese speaker (S2) uses both I think and maybe before the statement
‘most of you have seen the film Scrappy’. The two expressions reveal that
S2 is not sure about this statement, and the two vague expressions are
brought in to hedge it.
In some cases the Persian and L1 group seem more in line with
each other rather than with the Chinese. For example, the most occurs
20 times in the Persian data (six times in clause-initial position) and
16 times in L1 data (none in clause-initial position), but is used mini-
mally by the Chinese. Other similar cases are that the Chinese used
most important thing 12 times, think the most six times, and most people
six times, none of which appeared in the other groups’ data. In other
cases, the L2 groups have more in common than they do with the L1
group, using some types of most cluster far more: most important was
used 20 times by the Chinese and 13 times by the Persian speakers,
but only once by the L1s. There are also cases where the Chinese and
L1 group are more alike than they are with the Persian speakers: the
most important was used by the Chinese 17 times and L1s 13 times,
but never by the Persian group. There seems no particular pattern in
terms of using most clusters here, as discrepancies are apparent across
all groups.
4.3.6 (A) few

The L1 group used (a) few 21 times, the Chinese seven, and the Persian
speakers never. The lack of this item in the Persian interactions can be
attributed to the lack of a similar lexical item in the learners’ mother
tongue.
This case of (a) few stands out: it is the first item interrupting the
dominant position of the Chinese as the most frequent users of vague
quantifiers, and it is also the only quantifier that does not occur in the
Persian data. In terms of the total number of vague expressions of all
categories, the Chinese were the most prolific. They also used the high-
est overall number of all vague quantifiers, and indeed of all the vague
quantifiers discussed so far, including some, much, many, a lot of, and
most. Despite the fact that the Chinese used almost twice as many vague
quantifiers as the L1s, they used (a) few exactly three times less often
than the L1 group.
4.3.7 A little
Following (a) few, a little was again more commonly used by the L1s (20
times) than by either L2 group: 11 times by the Chinese and nine by
the Persian speakers. The L1s used this quantifier twice as often as the
other two groups separately.
Around half of the sentences (nine) containing a little in the L1 data
are followed by a comparative adjective, whereas only a third of the
sentences (three) in the Persian data follow the same pattern. There was
only one example from the Chinese, indicating that this pattern may
have occurred accidentally.
4.3.8 Lots of
Lots of was used 34 times by the Persian, 16 by the L1s, and 11 by
the Chinese. This is the first vague quantifier employed more by the
Persian speakers than by the other two groups, and they used lots of
more than twice as often as the L1s and three times more often than
the Chinese.
The occurrence of words before lots of reveals some patterns. For
instance, while there are lots of occurs five times in the Persian data, the
L1s did not use this cluster in their interaction and the Chinese used it
only once. By contrast, the L1s used there is lots of five times, the Persian
speakers once and the Chinese not at all. This difference between L1
and L2 speakers may be attributed to a lack of knowledge of the phrase’s
use on the part of L2 speakers.
There are clusters consisting of elements occurring after lots of. For
example, while the L1 and the Chinese participants did not use lots of
things, this cluster occurred 11 times among the Persian speakers.
Of the total of 34 lots of + noun in the Persian data, 25 involved
countable nouns, as did 12 out of 16 by the L1s and five out of 11
by the Chinese. This finding shows that the Persian and L1 groups
preferred lots of + countable noun more than the Chinese group, who
preferred lots of + uncountable noun instead, although only slightly.
The uncountable nouns found in the data are typically money for the
L1s, and knowledge and time for the Persian speakers.
4.3.9 A lot
A lot occurs 16 times in the L1 data, ranking ahead of the second
favorite, lots of. A lot occurs half as often in the Chinese data, with the
frequency of eight; it appears nine times in the Persian data. A lot is
infrequently used, at less than 5 per cent of all vague quantifiers.
Of the nine occurrences of a lot in the Persian data, three are in clause-
final position. There are four occurrences of a lot in clause-final position
in the Chinese but only one in the L1 data, showing that this group is
much less interested in this construction. There is no consistent pattern
of words used before a lot in any of the data sets, but words appearing
after a lot include comparative adjectives, which the L1s used six times,
and three of which are multi-syllabic adjectives placed after more, two
uses of the irregular comparative adjective better, and two infinitives.
More than half of a lot in the Persian data are followed by conjunctions,
two for reason, two for contrast and one for addition. It seems that the
L1s’ use of a lot is mainly associated with comparisons, but when used
by the Persian speakers, it appears to create the need for a new utterance
through conjunctions. As half of the Chinese use of this phrase occurred
in clause-final position, no meaningful pattern emerged for words
following a lot.
4.3.10 Majority
Majority appears least in all data sets. There were a few majority used by
the Persian speakers and L1s, but the Chinese interactions were devoid
of this vague quantifier. Despite the relative similarity in the use of
majority by the two groups, they present a different pattern from major-
ity. All six occurrences of majority by the L1s are followed by of; none
of the seven occurrences in the Persian data do. Three of the majority in
the Persian data occur in clause-final position.
Figure 4.5 shows the frequency ranking of vague quantifiers as: some
(moderate quantifier); much, many, a lot of, and most (bigger quantifi-
ers); (a) few, and a little (smaller quantifiers); lots of, a lot, and majority
(bigger quantifiers). The pattern is from most frequent to least frequent:
moderate—bigger—smaller—bigger. The result is somewhat similar to
Drave’s (2002) finding that a lot and many are the most common vague
quantifiers.
In terms of overall frequency, the vague quantifiers conveying a
smaller quantity (some, a few, a little, and few) are slightly fewer than
the vague quantifiers conveying a larger quantity (for example many
and lots of). The 1599 quantifiers occurring in this data comprise
734 smaller-amount quantifiers (46 per cent) and 865 bigger-amount
quantifiers (54 per cent). This indicates that the participants use vague
quantifiers to mitigate (for example a little) or emphasize (for example
a lot) at relatively similar rates.
The overall frequency of quantifiers by the three groups suggests that
this vague language category occurred almost evenly in the classroom
300
264
250 229
200 173
163
150
106
100 85 86
53
40 46 47 39 38 34
50 22 33 21 20 20 16 9
7 0 119 11 8 6 7
0
0
f)
ew
tle
of
ity
to
lo
of
an
uc
(o
or
lit
ts
t(
)f
A
lo
M
M
e
Lo
aj
os
A
(A
m
M
M
So
Figure 4.5 Frequency distribution of vague quantifiers
interactions of two groups: the Persian speakers (435) and the L1s
(423). In contrast, the Chinese group used them 74 times: almost twice
as often. Statistical analysis of the occurrences of this category among
the three groups proves significantly different, and close investigation
of some subcategories also reveals discrepancies between the groups at
relatively similar frequencies. Of the ten vague quantifiers studied, the
first five shown in Figure 4.5 were consistently and most heavily used
by the Chinese; this was not the case for the second five, which were
used infrequently, and mainly by L1s (except lots of).
While the discrepancy in the use of the first five items by the three
groups is large, the same does not hold for the second five items, where
the differences are of a smaller scale. Some emerges as the most com-
mon vague quantifier among all three groups, occurring 264 times in
the Chinese interactions, 229 times in the Persian, and 173 times
in the L1. The second most frequent item, much, with 106 instances in
the Chinese data, was used twice as often as by the L1s (53 times), and
even more often than the Persian speakers (40 times). On the whole,
there is consistency in the pattern of small quantity quantifiers in that
(a) few and a little were both used more frequently by the L1s within
the limited frequency range presented by the three groups, while there
were fluctuations in the use of quantifiers expressing large quantities.
Ruzaitė (2007) reports intervarietal differences in the frequency of vague

quantifiers between American and British English, with the speakers
of American English using quantifiers more commonly than British
English speakers. She points out that there are different combinations
of intensifiers occurring before quantifiers, the most common preced-
ing much and many being very, too, and so. This study does not support
Ruzaitė’s finding, demonstrating that despite so being heavily used with
much, many and the like, too and very did not commonly follow vague
quantifiers in the classroom context of this study. This discrepancy may
be attributed to the different discourses under consideration.
Analysis of vague quantifiers from the percentage perspective reveals,
as shown in Figure 4.6, that the L1s used nine vague quantifiers
Figure 4.6 Percentage distribution of vague quantifiers

relatively more evenly than the two L2 groups. The L2 speakers’ use
was skewed toward five of the nine quantifiers, and the others were
used minimally; the unevenness is not so evident in the L1 group. This
difference between L1 and L2 groups echoes and repeats what has been
revealed in their use of subjectivizers and possibility indicators, and the
trend continues in this category as well.
The two L2 speaker groups show a closer ranking distribution, as
shown in Figure 4.6. For instance, the ranking order of the first four
items is the same between the two groups: some, much, many, and a lot
of. More than half of the quantifiers used by the Persian speakers con-
sisted of some; this proportion in the Chinese and the Persian speakers
covered two items (some and much). At the lower end of scale, the total
percentage of the last five items is only 5 per cent for the Chinese group,
while it is 14 per cent and 19 per cent for the Persian speakers and the
L1s, respectively.
4.4 Vague intensifiers
Vague intensifiers are used to enhance the strength of an utterance, and

are exemplified by the items in Table 4.11.
The examination of vague intensifiers is confined to the six items that
appeared most frequently in the data: really, very, actually, so, too, and quite,
as in Table 4.11. Like the previously discussed vague language categories
(subjectivizers, possibility indicators and vague quantifiers), the Chinese
again used vague intensifiers the most, the Persian speakers the least, and
Table 4.11 Distribution of vague intensifiers
Item L1 CSLE PSLE
1. really 174 (44%) 164 (19%) 58 (17%)

2. very 79 (20%) 498 (56%) 108 (32%)
3. actually 67 (17%) 24 (3%) 73 (22%)
4. so 40 (10%) 154 (17%) 75 (23%)
5. too 24 (6%) 34 (4%) 15 (5%)
6. quite 16 (4%) 9 (1%) 4 (1%)
Total 400 (101%)* 883 (100%) 333 (100%)

the L1 group stood in the middle. The total number of vague intensi-
fiers for the Chinese was 2.2 times that of the L1s and 2.7 times that of
the Persian speakers. This indicates that the Chinese participants tended
to emphasize more than the others, and the Persian speakers tended to
emphasize the least. The most preferred item was very for both L2 groups,
but the L1 group preferred really. Only one item was predominantly
used by the Persian speakers: actually. Chi-square test reveals a signifi-
cant difference in the use of vague intensifiers among the three groups,
p < 0.05 (χ² = 321.263, d.f.10).
4.4.1 Really
Not all occurrences of really in the data are vague intensifiers, as some
carry out functions that do not emphasize, as shown in (4.17) below.
(4.17): This is a discussion between three L1 participants over seven
turns. They are talking about the result of a search using a search engine.
S5: Four-oh-two (L1: 3:509)

S8: I have (L1: 3:510)
S1: Huh? (L1: 3:511)
S5: Four-oh-two. (L1: 3:512)
S1: Oh, really? Yeah? (L1: 3:513)
S8: Four thirty-one. (L1: 3:514)
S1: Really? You sure you used D-F? (L1: 3:515)
Really in excerpt (4.17) is used as an exclamation, although it precedes

a question mark. By contrast, really functions as an intensifier mainly
when it occurs with an adjective or a verb.
Really was most frequently used by the L1s, 174 times. This is close to
the frequency of use by the Chinese (164 occurrences), but the Persian
speakers used it significantly less (58 occurrences, about one-third of
those of the other two groups). The percentage value of 44 per cent
highlights the significance of this item in the L1s interaction, while it is
less than one-fifth of the overall vague intensifiers used by the Persian
(17 per cent) and Chinese (19 per cent) speakers respectively.
The examination of really + part of speech (items 1 and 2 in Table 4.12)
displays a consistent pattern: really + verb outnumbers really + adjective
across all three groups, showing that the participants prefer really + verb
more than really + adjective. This consistency is more noticeable between
the Chinese and the L1s, who show roughly the same distribution with
68 (40 per cent) and 69 (41 per cent) occurrences, respectively, similar in
both raw numbers and percentages. The Persian speakers show a lower
Table 4.12 Really clusters
1. really + verb 69 (40%) 68 (41%) 24 (41%)

2. really + adjective 55 (32%) 50 (30%) 9 (16%)
3. subject + really 47 (27%) 51 (31%) 30 (52%)
4. really in negative sentence 39 (22%) 7 (4%) 1 (2%)
5. really + auxiliary 15 (9%) 7 (4%) 3 (5%)
6. really in clause-initial position 2 (1%) 2 (1%) 5 (9%)
7. I really 11 (7%) 20 (12%) 3 (5%)
8. you really (single/plural) 11 (7%) 4 (2%) 4 (7%)
9. it really 8 (5%) 4 (2%) 7 (12%)
10. they really 1 (1%) 6 (4%) 0
11. we really 0 15 (9%) 5 (9%)
12. really want 5 (3%) 9 (5%) 4 (7%)
frequency, with 24 occurrences, but as a percentage (41 per cent) their

use is similar to that of the other groups. In all three groups around
40 per cent of the overall instances contain really + verb.
The pattern really + adjective again shows some consistency between
the L1s and the Chinese: the L1s use it 55 times (32 per cent) and the
Chinese 50 (30 per cent). As with really + verb, the Persian speakers
showed a comparatively lower use (9 occurrences, 16 per cent). The
combination of really + verb and really + adjective amounts to more than
70 per cent of the use of vague intensifiers in the L1 and the Chinese
data, and more than half in the Persian data, showing a strong preference
for these two clusters by the participants. The remaining portion of vague
intensifiers consists of other parts of speech; for instance, the L1 used
really + auxiliary 15 (9 per cent) times, the Chinese seven (4 per cent) and
the Persian speakers three (5 per cent), as shown in item 5 in Table 4.12.
The cluster subject + really made up 27 per cent of the uses of really by
the L1 group, one in three (31 per cent) by the Chinese, and one in two
(52 per cent) by the Persian speakers. The Persian speakers use subject +
really 30 times, 19 of which were subject pronoun + really. L1s used sub-
ject + really 47 times, 31 occurrences with subject pronouns. The Chinese
differed, using subject + really 51 times, only two of which did not have
subject pronouns, showing an extreme preference for this form.
Subject pronoun + really includes five items (7, 8, 9, 10, and 11 in
Table 4.12). The Chinese used a total of 49 such clusters, the L1 group 31
and the Persian speakers 19, less than half the Chinese tally. Of the five
items, only four occur in the Persian and L1 data, but the Chinese used
them all. The four items used by the Persian speakers were distributed
relatively evenly, but the Chinese concentrated heavily on two items
only and the L1 on three. The Chinese preferred the first person sin-
gular and plural, I really and we really, while the other groups showed
no such preference: the L1 group did not use the first person, and
third person was used only once; the Persian speakers did not use
the third person plural.
Really occurs quite differently in negative sentences. The frequency is
quite high in the L1 data, with 39 instances (22 per cent), lower in the
L2 speaker groups (seven and 4 per cent for the Chinese, one and 1 per
cent for the Persian). The L1s prefer really in negation much more than
the L2s, inclining to use really to highlight the force of a negation.
As shown in Table 4.12, really occurs five times in clause-initial posi-
tion in the Persian data, and twice each in the L1 and Chinese. Nine
per cent of the Persian speakers’ intensifiers were positioned at the
beginning of a clause, much more than the 1 per cent of the other two
groups. An example from the L1 group, where really starts a sentence,
appears in (4.18).
(4.18): This is a turn from an L1 participant talking about studying
speech.
S1: I know this wasn’t gonna happ- help work sooner or later. Junior,
high, school. Okay? Makes sense. Really makes sense. –Really
makes sense. Junior high school. I know I’m gonna find that con-
struction. Senior high school. (L1: 3: 346)
In (4.18), both really are used in clause-initial position, where the

dummy subject it is missing from the sentence. The intensifier really
enhances the effect of makes sense. The Chinese also used really when a
dummy subject is needed, but unlike the L1s who used really to modify
a main verb, they used it to intensify the degree of an adjective, as in
excerpt (4.19).
turns. They are talking about the movie Bodyguard.
S1: And the song won a big award and even, 11 very familiar awards.
Really amazing, amazing, and other five songs in this original
song are also very beautiful. (Ch: 6: 71)
S2: What’s the name of this album? (Ch: 6: 72)
In terms of lexical clusters with really, apart from the subject pronoun +
really, other types of cluster are barely present in any of the three groups.
One rare occurrence is the use of really want: the Chinese used this nine
times, the L1s five, and the Persian speakers four. The Chinese used
really like eight times and really need five times, the L1s used them twice
each, but no examples appeared in the Persian data. Really good is used
by the L1s nine times and three times by the Chinese, really important by
the Chinese seven times and four times by the L1s, neither are found in
the Persian data. Some really lexical clusters are used by one group only:
from the L1 group: really start eight times, really interested 11, that’s
really nine, and it doesn’t really five times. From the Chinese: really hard
six times, really I think six, I really want five, and really like to five times.
A trend here seems to be that the L1 and Chinese speakers are similar
in their use of lexical clusters of really, while the Persian speakers do not
use these clusters as often.
The findings demonstrate that the L1 and Chinese cohorts are similar
in many ways and are quite different from the Persian speaking group.
Table 4.12 reveals that the L1 and Chinese groups make more use of
really and its clusters, while the Persian speakers use them less often and
more diversely.
4.4.2 Very
The second most commonly used vague intensifier by the L1s, very, was
the most common vague intensifier used by the Chinese and Persian
speakers. With 498 uses, the Chinese were the leading users of very,
followed by the Persian speakers at 108 occurrences, and trailed by the
L1 group with 79 instances. More than half (56 per cent) of the vague
intensifiers in the Chinese data include very, one third (32 per cent) in
the Persian data, and one fifth (20 per cent) in the L1. It seems that very
is more popular with the L2 groups than the L1s.
An examination of different sentence positions in which very occurs
indicates that while the L1s avoided using this quantifier in clause-
initial position, the Chinese and the Persian speakers used it four and
six times respectively, in this position.
(4.20) This is a discussion between two Persian participants over three
turns. They are talking about the younger generation.
S3: Especially the current generation is very complicated. (P: 6:737)

S6: Very complicated, yes. (P: 6:738)
S3: They are growing up with computer. (P: 6:739)
In excerpt (4.20), the first very by S3 functions as a booster, enhancing

the degree of the complexity. The second very used by S6 confirms and
reinforces what S3 said in the previous turn, differing from the first in
that it appears at the beginning of the sentence, taking a clause/turn
initial position.
(4.21): This is a discussion between three Persian speakers over three
turns. They are talking about why they think Iran is an isolated nation.
S9: I think, I think. I think we are following the wrong path to be

right. (P: 1: 208)
S3: Very good. (P: 1: 209)
S5: Yes, good. (P: 1: 210)
In (4.21), S9 makes a statement first, then S3 uses very good to comment

on S9’s statement in the next turn. Very here increases the strength of
goodness. Note that S5 in the next turn agrees with S3 by saying yes,
adding an evaluation by saying good without very, indicating that S5’s
evaluation is not as high as S3’s, which uses the booster.
The pattern of a very + adjective + noun emerges in the data quite fre-
quently; the nouns are typically singular and countable. There is con-
sistency between the L1 and Persian speakers, who used the cluster at
similar rates: 12 and 14 occurrences, respectively. The Chinese used it
Table 4.13 Very clusters
1. a very + adjective + noun 12 (15%) 27 (5%) 14 (13%)

2. very in negative sentence 2 (3%) 25 (5%) 9 (1%)
3. be + very 4 (6%) 19 (4%) 2 (2%)
4. is + very 3 (4%) 145 (29%) 32 (30%)
5. are + very 1 (1%) 36 (7%) 9 (8%)
6. was + very 0 10 (2%) 6 (6%)
7. very much 2 (3%) 48 (10%) 4 (4%)
8. very important 4 (6%) 28 (6%) 15 (14%)
9. very good 9 (11%) 37 (7%) 14 (13%)
10. very interesting 2 (3%) 16 (3%) 2 (2%)
11. very hard 0 18 (4%) 2 (%)
12. it is very 0 49 (10%) 12 (11%)
about twice as often (27 times); however, proportionally speaking the

Chinese were the least frequent users of this pattern, which is only
5 per cent of the total number of very instances in the Chinese data.
The L1s’ use of the cluster is 15 per cent and the Persian speakers’
13 per cent.
As Table 4.13 shows, the occurrence of very in negative sentences
reveals a trend consistent with the overall frequency of very: the Chinese
were the greatest users, the Persian speakers the second, and L1s the
least users of very in negations. The most frequent patterns are not very
and not a very, exemplified in (4.22) below, where not very functions to
negate importance.
turns. They are talking about how people should obey rules.
S7: I think when he or she knows me, he respects me. (P: 6:111)
S3: Ahuh, the others? (P: 6:112)
S1: I think it is not very important. (P: 6:113)
There are some infrequent occurrences of conjunction + very, consistent

in all three groups, and exemplified by excerpt (4.23) below.
(4.23): This is a discussion between two L1 participants over three
turns. They are talking about how a computer system works.
S1: And so what it does, and the system is doing this not a human
being a computer system. and the computer can do this very
quickly and very efficiently. Computer system starts to number,
each word in the field and preface, the n- word, position informa-
tion or the word position number, by the name of the field so in
this case we have. (L1: 3:268)
SS: Stop word (L1: 3:269)
S1: It’s a, called a stop word and what is a stop word? Anyone know?
(L1: 3:270)
As shown in Table 4.13, the most frequent users of the cluster be + very
were the Chinese (19 instances), far ahead of the other groups (four
times by the L1 and two by the Persian speakers). Despite the differ-
ence in raw numbers, the percentage values are 6, 4, and 2 per cent,
closer than the difference in raw numbers. In the case of is + very,
the Chinese produced 145 instances, the Persian speakers 32, but the
L1s only three. In raw numbers, the Chinese used it five times more
than the Persian speakers and almost 50 times more than the L1s.
The Chinese and Persian percentages are similar (about 30 per cent), but
very different from the L1 group (4 per cent). The situation is similar
to the case of are + very and was + very: where the L2s showed a similar
frequency (by percentage). While the Chinese and Persian speakers
were different in raw frequency, that difference was much smaller than
the difference between them and the L1s. These findings suggest that
there is more diversity between the L1 and L2 groups than within the
L2 groups.
Despite using very the least, L1s revealed some usages that were
uncommon in the L2 data. For instance, the duplication of very was
used in (4:24) below.
turns. They are talking about how culture can affect morals.
S1: And um, that’s actually something there’s a um, a really famous,
philosopher who, who wrote something very, very similar, to
that so, just in case you’re interested, um, Alasdair MacIntyre.
I don’t know jus– … (L1: 2:45)
S11: But definitely I think like something as extreme as a child
killing someone, um, some are more, like are more obvious
than others and haven’t been known to like, work, like certain
actions. (L1: 2:46)
The duplication of very in this instance takes an emphatic function. S1

uses two very to indicate that the similarity is extremely high; otherwise
one very would have been enough. The pile-up highlights the strong
degree of similarity.
Another pattern of very clustering is very + adverb. While the adverbs
used by the L1s and Persian speakers were often regular (adjective + ly),
the Chinese used only two such adverbs and instead presented items
such as very much and very well, modifying a verb. The Persian speak-
ers infrequently used very + adverb, which may be because in informal
contexts in their mother tongue, they could use adjectives instead of
adverbs to modify a verb. The adverb the Persian speakers preferred is
easily; the L1s preferred quickly. The cluster very much was used by the
Chinese 48 times, 10 per cent of the total very used, showing that very
much is one of their favorites. The other two groups did not show much
interest in very much, with four instances in the Persian speakers’ data
and a mere two in the L1s’.
Comparing the use of the lexical clusters constituting items 8 to 12
in Table 4.13, the trend is that the Chinese are always the leading party,
followed by the Persian speakers. The L1 group are the least frequent
users, and do not even use two of the clusters. Of the five clusters, the
Chinese used it is very the most: 49 instances and 10 per cent of the
very instances in their data. The Persian speakers used it is very 12 times,
about five times less than the Chinese. The two items they used a bit
more were very important and very good (15 and 14 times respectively).
The Chinese also used it’s very (a shortened version of it is very) 32 times,
but no instances of this were presented by the other groups.
Some very clusters are absent from one or two groups, suggesting that
they were less popular. For example, very busy and very difficult were used
by the two L2 groups only. Very busy occurred five times in the Chinese
discourse and four times in the Persian; very difficult was used 11 times
by the Chinese and four by the Persian speakers; the cluster a very good
was used eight times by the Chinese and five by the L1s, but not by the
Persian speakers. Some very clusters were used by the Chinese only: there
are very (27 instances), very happy (12), will be very, very beautiful (11), he
is very, very famous (10), this is very (7), very hot, very easy, very convenient,
that’s very (6), very fast, and very young (5). The L1 group used the very
least five times, but it did not appear in any L2 data.
All the adjectives placed after very by the L1s were positive (impor-
tant, good, and interesting), while the Persian speakers used positive and
negative adjectives evenly: three positive adjectives (important, good and
interesting), and three negative (busy, difficult and hard). The Chinese
used more positive than negative adjectives. The trend emerging here
is that very tends to cluster with positive adjectives, and negative adjec-
tives are less often grouped with very.
There is an inconsistency in the patterning of very, and all other
vague intensifiers, among the three groups. The use of vague intensi-
fiers as well as most individual lexical items in the Chinese interactions
outnumbered those of the L1 and Persian speakers, indicating that the
Chinese group tended to enhance their speech more than the others.
4.4.3 Actually
Actually ranks fourth and is one of most frequent intensifiers. Unlike the
trend in very, this time the Persian speakers used it most (73 instances,
22 per cent), the L1 group 67 times (17 per cent), the Chinese least,
with 24 instances (3 per cent), about three times fewer than either other
group. While actually constitutes around one fifth of the overall intensi-
fiers used by the L1 and Persian speakers, the Chinese used it only as
3 per cent of their intensifiers in similar contexts. The Persian speakers
and L1s behaved alike to a large extent, in this case.
The first discrepancy emerges in the analysis of actually serving in

negative sentences. The Chinese presented eight instances and the
Persian speakers six, both more than the L1s (four instances). While the
minute difference among the three groups in the raw number of tokens
is not particularly compelling, the frequency percentage value shows a
stronger distinction: the Chinese at 25 per cent, the Persian speakers at
11 per cent, but the L1 group at only 6 per cent.
The intensifier actually was used as a turn initiator six times by
Persian speakers and four times by Chinese, and served the same pur-
pose three times as often for the L1s (12 instances). The Persian speak-
ers had six instances of actually occurring after conjunctions (because
twice and but four times), but this pattern was absent in the L1 and
the Chinese data.
The L1s used auxiliary verbs + actually more than the L2 groups.
They included verbs such as be, do and modal auxiliaries in both
affirmative and negative forms. Auxiliary verb + actually was used by
the L1s 29 times, whereas in the Persian speakers’ data, despite its
higher overall frequency of actually, it occurred ten times only; the
Chinese used it only twice. There is a cluster of ‘to (infinitive marker) +
actually + verb’ in L1 data (four occasions), exemplified in excerpt
(4.25) below.
turns. They are talking about a possible mistake in searching a term in
a search engine.
S1: I would have to take a look at your search to actually tell you
exactly what went on … my guess is, and I’m, I’m this is just my
hunch, um, that again you may have used an or, instead of an and.
That’s a possibility. Or … (L1: 3:197)
S9: Physical disabilities, and then not-ing out mental retardation (dis-
cussing it back). (L1: 3:198)
S1 in (4.25) is uncertain exactly what the problem is, which is shown

by repetition, hesitation and utterances such as ‘this is just my hunch’.
S1 uses to actually tell to emphasize the conditions needed for the action
of tell to take place.
There is a cluster of discourse marker + actually appearing in the
Persian data three times. An example is given in excerpt (4.26) below.
(4.26): This is a discussion between two Persian speakers over two
turns. They are talking about how a mother would react to hearing that
her daughter has a boyfriend.
S2: For instance, imagine that you understand that she has a boyfriend
what would you do. (P: 6:722)
S1: Well, you know, honestly, what would I do? Just shake but the
most important thing about boyfriends these days in Iran is
I think, actually, I am sorry … you know, actually I talked about
this matter with my daughters. I’ve got three daughters. I had to
talk about these things with them. (P: 6:723)
In (4.26), S2 asks S1 what S1 would do upon hearing the news her

daughter has a boyfriend. S1 starts the answer with discourse markers
well and you know, to buy time to search for the right words to answer
a somewhat difficult question. S1 then uses actually twice, once linked
with the discourse marker you know, to stress the fact that this matter
has already been discussed with the daughters.
The cluster of conjunction + actually occurs five times in the Persian
data (all with and), but appears nowhere else. Another pattern occur-
ring in the Persian data four times, actually + we, does not appear in
the L1 data and only once in the Chinese. Similarly, the cluster I +
actually is found 12 times in the Persian data, five in the Chinese
and three times in the L1. It indicates a clear preference for the term
among the Persian speakers, particularly when they wish to emphasize
something.
4.4.4 So
The Chinese used actually least often, but so most often. Their 154 uses
mean that so was used almost twice as often by the Chinese as by the
Persian speakers (75 instances) and roughly four times as often as by the
L1s (40 instances).
The three groups were relatively consistent in using so in negative sen-
tences. Seven instances each occur in the L1 and Persian data, and nine in
the Chinese. Another cluster is so + adjective, where adjectives are of three
types: positive, neutral, and negative. For the L1 group, the frequency
was 11 for positive, one for neutral, and 11 for negative; for the Chinese
group, 83 for positive, 40 for neutral, and 31 for negative; for the Persian
group, 20 for positive, six for neutral, and 20 for negative. There is some
inconsistency between the Chinese and the other groups, but at the same
time there is a consistent pattern between the Persian and the L1 speakers,
who used negative adjectives as often as positive adjectives, and neutral
adjectives less often. Another consistency lies in the use of neutral adjec-
tives, which happened least often in all groups: one instance by the L1s,
five by the Chinese, and six by the Persian speakers.
There are two frequent lexical clusters in the data: so much and so
many. The L1 and Persian speakers used so much almost evenly, with
eight and nine occurrences, respectively. The Chinese, on the other
hand, used 19 so much, more than twice as often. Like so much, the
Chinese used so many most frequently (17 occurrences), and the L1s
with five occurrences were still the least frequent users of this cluster.
The 15 occurrences of so many by the Persian speakers brought them
closer to the Chinese. In addition to so much and so many, other less
frequent lexical clusters included so difficult, and you are so.
4.4.5 Too
The intensifier too was used most by the Chinese, with 34 instances; the
L1 group presented 24 instances and the Persian speakers 15. Contrary
to so which was used least often by the L1s, the Persian speakers used
too the least commonly, at only half the rate of the Chinese. The three
groups were similar in their use of this term from a percentage perspec-
tive, at about 5 per cent of the total vague intensifiers in the data.
Too occurred in both positive and negative sentences by the L1s and
the Chinese, while the Persian speakers used it in positive sentences
only. There are two instances of too by the Chinese occurring in sen-
tences containing never, but the Persian speakers showed no inclination
to use too in negative sentences.
An example of too used in a negative form is given in (4.27) below. In
turn 128, S1 uses not too soon, where too increases the degree of soon.
(4.27): This is a discussion between two L1 participants over four
turns. They are talking about how state systems will treat religious
beliefs.
SU-m: That’s really interesting we’ll have to see. Probably won’t happen
in this lifetime. (L1: 2:127)
S1: I don’t think so, not too soon anyway, but we’ll see I mean, you
know. (L1: 2:128)
SU-m: No, it’s a slow process. (L1: 2:129)
S1: It is. (L1: 2:130)
The L1s used three times of emphatic just + too, a phrase absent from
the Chinese and Persian data. In excerpt (4.28) below, the word young is
intensified by double boosters, just and too.
turns. They are talking about why young people are more likely to steal
things.
S1: Yeah that’s, yeah, I mean, yeah rebellious, or, or not we, we could
call it socialized which basically just means they’re not accustomed
to society’s norms yet. Yeah. (L1: 2:13)
S2: Or they’re just too young to understand the uh, consequences of,
stealing. (L1: 2:14)
In addition to the discussion of what occurs before too in the data, the
analysis of what occurs after too shows a partial similarity between the
L1 and Persian speakers. The first trend can be seen in the occurrence of
too much: four instances from the L1 and Persian speakers each, 14
from the Chinese. The L1 and Persian groups behave differently, how-
ever, in what too much is paired with: the L1s used too much twice at
the end of a clause, and twice in the middle, once after a verb and the
other before a noun; the Persian speakers placed all four instances of
too much before nouns, one of which is grammatically incorrect. The
Chinese used two tokens of too much at the end of a clause, and the
rest before nouns.
4.4.6 Quite
As with too, quite occurs with patterns that are different across the three
groups. Quite was used 16 times by the L1s, but by the Persian speakers
four times only. The Chinese used it nine times, about half as often as
the L1s. The percentage of quite constitutes a small portion of vague
intensifiers in each group: 1 per cent by the Chinese and Persian speak-
ers, and 4 per cent by the L1s.
As in the case of too, the Persian speakers avoided quite in negative
clauses. The same trend is found in the Chinese sample as well, but half
of the clauses containing quite by the L1s are negative. Quite + adjec-
tive occurs seven times in the L1 data, three in the Persian and six in
the Chinese. In addition, the cluster of quite + adverb occurs four times
in the L1 samples, but the Persian speakers used it only once and the
Chinese cohort never used quite before an adverb.
The L1s used other quite clusters, including quite a bit, quite a while,
quite sure, and quite easy, but the Persian speakers were limited to quite +
adjective and quite + adverb. The Chinese were slightly more diverse,
using quite sure, quite easy, quite + a + adjective + noun, and quite +
preposition in their interactions, making them somewhat similar to the
L1s in the use of quite clusters.
Overall, as Figure 4.7 shows, vague intensifiers form a category with
diversity among the three groups. The discrepancies go beyond simple
frequency of individual items, which are reinforced when patterns of

600
498
500
400
300
174
200
164 154
79 108
75
100 73
58 67
40 34
24 24 15 16
9 4
0
Really Very Actually So Too Quite
Figure 4.7 Frequency distribution of vague intensifiers
use and clustering behavior are taken into consideration. The Chinese
participants, with 883 occurrences, showed a stronger tendency to use
these intensifiers than the L1s with 400 occurrences and the Persian
speakers with 333. The groups had more similar rates of use for too and
quite. Despite this, the difference in the application of vague intensifiers
by the three groups is found to be statistically significant.
In terms of the frequency of individual items, the Chinese used three
categories the most often, very, so, and too, with an extremely high
use of very. The most significant frequency difference among the three
groups lies in this term, which the Chinese used approximately five
times as often as the Persian speakers and six times as often as the L1s.
The L1s used only two items the most often: really and quite (the first
and the last items in Table 4.7); the Persian speakers used only one item
the most often, actually.
Comparison of the frequency of the six vague intensifiers shows
that the L1 and Persian speakers behaved similarly in using four items,
very, actually, so and too. The Chinese and L1 groups were similar
in their use of two items, really and quite. The Chinese and Persian
speakers do not share any similarity in the use of any item. What
can be drawn from this is that as far as the frequency distribution
(raw frequency) is concerned, the L1 and Persian speakers behave

most alike, followed by the L1 and the Chinese. The two L2 groups,
the Chinese and Persian speakers, have little in common in their raw
frequency distributions.
Figure 4.8 shows the same trend as emerged in the raw frequencies:
the L1 and Persian speakers are relatively similar in their even spread
of use of the seven vague intensifiers, while the Chinese concentrate
on very heavily. Comparison of the distribution of intensifiers reveals
significant differences in terms of the ranking of items and of the per-
centage value of each. The top three are really, very, and actually (L1);
very, really, and so (Chinese); very, actually, and so (Persian). Although
in different positions, the top three items in all groups are made up of
four items: very, really, so, and actually. Figure 4.8 shows that the Persian
speakers have the most evenly distributed percentage of vague intensi-
fiers among the three groups.
L1 speaker CSLE
4% 1%
4%
6% Really Really
10% Very 17% 19% Very
Actually 3% Actually
17% 43% So So
Too Too
56%
20% Quite Quite
PSLE
5% 1%
17% Really
23% Very
Actually
32% So
22% Too
Quite
Figure 4.8 Percentage distribution of vague intensifiers

4.5 Placeholders
The term ‘placeholder’ in relation to vague language originates from

Channell (1994). The characteristic of this category of vague items is
that they convey a general meaning with little specific referent.
Among all five categories discussed so far, the placeholder category
is the first and only one where the Persian speakers used vague items
the most; for the other four categories, the Chinese used the most. As
shown in Table 4.14, the Persian speakers used a total of 478 placehold-
ers in their interactions, the L1s 301 and the Chinese 286. Placeholders
consist of seven items: something, thing, things, someone, anything,
somebody, and anybody. Of the seven items, the Persian speakers used
all the most except thing, which the Chinese used slightly more than
the others. The data show a discrepancy in the use of placeholders,
and the difference among the three groups is statistically significant
p. <0.05(χ²=48.906, d.f.12).
There seems some consistency in the usage by the Chinese: for the
top three items in Table 4.14, the frequencies of something, thing and
things are all in the range of 80–85 instances. This similarity continues
to the next three items, someone, anything and somebody, with frequen-
cies in the range of 11–13. Consistency can also be observed in items 4,
5, and 6 for the L1s, where frequencies range from 20 to 28.
4.5.1 Something
Something was used most by the Persian speakers (177 instances), almost
twice as often as the second most frequent users, the L1 group with 94
Table 4.14 Distribution of placeholders
Item L1 CSLE PSLE
1. something 94 (31%) 80 (28%) 177 (37%)

2. thing 72 (24%) 82 (29%) 74 (15%)
3. things 54 (18%) 85 (30%) 108 (23%)
4. someone 28 (9%) 13 (5%) 31 (6%)
5. anything 24 (8%) 12 (4%) 48 (10%)
6. somebody 20 (7%) 11 (4%) 23 (5%)
7. anybody 9 (3%) 3 (1%) 17 (4%)
Total 301 (100%) 286 (101%)* 478 (100%)
Note: *After rounding.

instances; the Chinese were the least frequent users with 80 instances.
The percentage distribution reveals that the difference among the
three groups is not as significant as the frequency suggests. Something
constitutes around one-third of the total placeholders used by each
group, and all three were quite consistent in terms of the percentage of
something in relation to the total number of placeholders, ranging from
28 per cent to 37 per cent.
As shown in Table 4.15, the Persian speakers most frequently put
something in clause-final position. The 47 occurrences in their data
(27 per cent of the total use of something) are more than three times
those of the L1s with 14 occurrences (15 per cent) and more than six
times those of the Chinese with eight tokens (10 per cent).
There is a phenomenon in the data: a discrepancy between something +
conjunction vs. conjunction + something. In terms of something +
conjunction (something precedes the conjunction), the highest
frequency is found in the Persian data where conjunctions after some-
thing appear nine times; this cluster occurs four times in the L1 data
and three in the Chinese. The clusters consist of something + but, and
and or for the L1s and Chinese, while the Persian speakers combined
something with and and or only. This suggests that the Persian speakers
prefer something + conjunction to express addition (and) and choice
(or), but not contrast (but).
Table 4.15 Something clusters
1. clause-final something 14 (15%) 8 (10%) 47 (27%)

2. do something 12 (13%) 20 (25%) 33 (19%)
3. something that 10 (11%) 0 15 (8%)
4. have something 9 (10%) 3 (4%) 5 (3%)
5. say something 7 (7%) 2 (3%) 19 (11%)
6. be something 6 (6%) 0 18 (10%)
7. something + infinitive 8 (9%) 3 (4%) 2 (1%)
8. something like 8 (9%) 0 6 (3%)
9. something else 3 (3%) 7 (9%) 15 (8%)
10. something wrong 2 (2%) 0 6 (3%)
11. something you 0 2 (3%) 16 (9%)
12. to do something 0 8 (10%) 7 (4%)
The use of conjunction + something (something after conjunctions) fol-

lows a different pattern in the L1 data: conjunction + something occurs
15 times, almost twice as often as in the Persian speakers (eight occa-
sions), but something + conjunction appears only half as often. Despite
using something the most, the Persian speakers used conjunction + some-
thing less often than the L1s. The Chinese, with seven occurrences, were
the least frequent users of this cluster. The conjunction most frequently
used by both L1s and the Chinese was or, with 13 and five occurrences
respectively, but was the least frequently used by the Persian speak-
ers (one occurrence). But and and, each with a frequency of one, were
conjunctions least commonly used by the L1s; but the Persian speakers
used and five times, their most frequent conjunction, and but twice,
the second most frequent. The Chinese used three and something; but
something does not appear in their data.
About something occurs four times in the Persian data, three times in
the Chinese and once in the L1; it is not commonly used. However,
when the two words reverse position, something about, it is more fre-
quent: 15 times in the Chinese, six times in the L1 and four times in
the Persian data. There are other clusters of something and preposition:
for example, something for was used eight times by Persian speakers, and
something to and something from each occur twice in the Chinese data.
The most frequent cluster is verb + something: 85 times by the Persians
as the most frequent users, followed by the L1s (34 times) and the
Chinese (33 times). As shown in Table 4.15, do something is the most
frequent verb + something in each group: 33 uses by the Persian speakers,
20 by the Chinese and 12 by the L1s. The groups have two other verb +
something clusters in common: say something and have something. Say
something occurs in the Persian data 19 times, in the L1 seven, and in
the Chinese three times. The frequency rankings for the Persian speak-
ers are do something, say something and have something; for both the L1s
and Chinese do something, have something and say something. The same
item appears first in each group.
Item 11 in Table 4.15 is something you, which the Persian speakers used
16 times, the Chinese twice and the L1 group never. This highlights a
considerable difference among the three groups. The Persian speakers
intended to directly involve the listener in a state of uncertainty when
something you was used in their interaction, exemplified in excerpt (4.29).
(4.29): This is an utterance by a Persian speaker who is defining the
word inevitable.
S5: Inevitable means something you cannot escape from. (P: 4:42)
The L1s do not use something you in their interaction, instead opting for
clusters such as something that, exemplified in excerpt (4.30).
(4.30): This is an utterance by an L1 participant who is referring to a
photo.
S: That might be an example—is this is this something that, needs to

be seen by … (L1: 1:431)
Item 3 in Table 4.15, something that, is one of the clusters most fre-
quently used by the Persian speakers (15); the L1s used it ten times and
the Chinese never. Analysis of the function of that after something shows
that the Persian speakers used that as a conjunction or a subject almost
evenly: seven and eight times, respectively. However, L1 use of that
after something reveals uneven distribution, the speakers preferring that
as a conjunction (seven uses) almost twice as often as that as a subject
(three uses).
Item 7, something + infinitive, appears most frequently in the L1 data
(eight times), while the Chinese and the Persian speakers use it three
and two times respectively, clearly not a favorite for L2s. Item 9 is some-
thing else: it was again most often used by the Persian speakers (15),
while the Chinese used it around half as often (seven) and the L1s least
(three). The Persian speakers used a third of this cluster after the verb
do or does, the Chinese only once; instead, they used it five times after
conjunctions such as and and or.
Some clusters were not used by all three groups. For example, be some-
thing was used 18 times by the Persian speakers, a third of that by the
L1s (six occasions), but not by the Chinese. Nor did the Chinese offer
any examples of something that or something wrong. Two clusters absent
from the L1 data are something you and to do something. Some clusters
were used exclusively by a single group: the Persian speakers provided
seven instances of you do something; six of it is something and something
for eating; five of attention to something, to say something, buy something,
and know something; the Chinese were alone in using learn something
(six instances). This suggests that the three groups differed in their pref-
erences in using something, demonstrated by the absence of particular
something clusters from certain groups.
4.5.2 Things
As with something, things was used the most by the Persian speakers, but
unlike something, where the Chinese were the least users, it was the L1s
that used things least often. The Persian speakers used this placeholder
108 times, exactly twice as often as the L1s (54 instances); the Chinese
used it 85 times, not far from the Persian speakers’ use. Figures in
percentages show that the contribution of things to the formation of
placeholders is around one-fifth in the Persian and L1 data, but one-
third in the Chinese data.
As shown in Table 4.16, things in clause-final position occurs most
in the Persian data (37 times), while it occurs 28 times in the Chinese
and eight times in the L1 data. The cluster makes up about 33 per cent
of things in both L2 groups, and less than half of that (15 per cent) in
the L1 cohort. This seems to be a case of L1 against L2. Around half
the clause-final position things in each group occurred as turn-shifting
devices: 16 by the Persian speakers, 12 by the Chinese, and three by
the L1s. The same trend continues in the use of things in negative and
interrogative sentences, where again the L1s were the least users and the
Persian speakers the greatest. The Persian speakers used things in nega-
tion eight times, the Chinese five, and the L1s never. The L1s used things
in interrogative sentences twice, the Persian speakers six times and the
Chinese four. There is consistency here in that the Chinese and Persian
speakers both used things at a relatively similar rate and ranking when
things was in clause-final position, negative or interrogative sentences,
while the L1s tended to use these less often, coming last in the rankings.
Items 2–6 in Table 4.16 are clusters in the form of lexical item +
things. The frequency of use among the L1s was the opposite of that of
the Persian speakers. A form that was used most by the L1s tended to
be least used by the Persian speakers; an exception is the case of kinds
of things, where the Chinese group was the least user and the Persian
the greatest. Of the five items, four (kinds of things, different things, other
things, and new things) were used most by the Persian speakers, and all
Table 4.16 Things clusters
1. clause-final things 8 (15%) 28 (33%) 37 (34%)

2. some things 8 (15%) 2 (2%) 0
3. kinds of things 3 (6%) 2 (2%) 10 (9%)
4. different things 2 (4%) 3 (4%) 6 (6%)
5. other things 0 10 (12%) 15 (14%)
6. new things 0 4 (5%) 10 (9%)
7. things and 2 (4%) 12 (14%) 7 (6%)
except kinds of things were used the least by the L1s. For the remaining
one, some things, the positions reverse: this time the L1 group was the
greatest user and the Persian cohort, with no uses, the least. This part of
data depicts an opposing tendency between the L1 and Persian speakers.
The Chinese, in the middle, use all five clusters, the L1 group three and
the Persian speakers four.
Some clusters take the form of conjunction + things. Item 7 in Table
4.16 shows the cluster things + and. The most frequent users were the
Chinese, who provided 12 instances; the Persian speakers were the sec-
ond most frequent users with seven instances, and the least frequent
users were the L1s with two instances: a sixth of the frequency of the
Chinese. The contrastive conjunction cluster of things + but was used
by the Persian speakers six times and once by the Chinese, but not by
the L1s. The cluster things + because is used three times by the Persian
speakers and once each by the Chinese and L1s. The cluster things + so
was used exclusively by the Chinese, twice; it did not occur in the other
two groups. The Persian speakers and the Chinese were somewhat simi-
lar in their clustering of things with conjunctions; the L1 group showed
a different behavior.
The Persian speakers frequently used things that and things. OK? The
former was also used frequently by the L1s, but was less frequently
employed by the Chinese. The total occurrence of things that is 23 in the
Persian data and ten in the L1, but only twice in the Chinese. It serves
in the Persian data as either a conjunction or the subject of sentence,
almost equally (12 and 11 times). The L1s preferred that as a conjunc-
tion (ten times), but also used it as a subject (six occurrences). Of the
two occurrences of that after things by the Chinese, one served as a
conjunction and one as a subject.
Some things clusters are used exclusively by one group. The Persian
speakers provided 15 instances of the things that, ten of lots of things; six of
of things but; five each of things that they, things that I, and and other things.
The L1 group offered five instances of of the things. The Chinese gave
19 instances of a lot of things and five of learn a lot of things. These suggest
that despite the closeness between the Persian and the Chinese speakers
in their total use of things, here the Chinese and L1s are closer, in that
they use far fewer things clusters than the Persian speakers.
4.5.3 Thing
The only placeholder that turns up with a close occurrence across
all three groups is thing. The Chinese, with 82 uses, employed it the
most often, but it was only slightly less common in the Persian (74)
and L1 (72) interactions. As a percentage, thing constitutes a quarter

of the placeholders in the Chinese and Persian data, but only 15 per
cent in the L1 data.
The three groups used thing in dissimilar patterns. The first difference
can be observed in clause-final position of thing. The Chinese preferred
this position the most (30 instances, 37 per cent of the total use of
thing). The Persian speakers provided 14 instances (19 per cent), half
those of the Chinese. The least frequent users were the L1 group, with
seven instances (10 per cent), about four times less than the Chinese.
Thing in clause-final position appears as a signal that the speaker is plan-
ning to hand the speaking turn to the next speaker. While this usage is
seen 12 times in the Chinese data and eight times in the Persian data,
the L1s do not use thing as a turn-giving device.
There is no significant difference between the Persian and L1 speak-
ers’ use of thing + conjunction (four and six instances respectively); the
Chinese used thing + conjunction the most often, 11 times, to express
reason, addition and result. The L1s tended to use it for contrast rather
than reason, while the Persian speakers employed it for addition and
reason. The Persian speakers used thing + subject pronoun nine times
with the subject pronouns I and you, and the L1s and the Chinese
slightly more often, 14 times each, with I as the most common pro-
noun; the L1s also used they, but the Chinese did not.
The Persian speakers used thing that the most often, 22 times (30 per
cent of the total number of thing), while it was used ten times (14 per cent)
Table 4.17 Thing clusters
1. clause-final thing 7 (10%) 30 (37%) 14 (19%)

2. thing + conjunction 6 (8%) 11 (13%) 4 (5%)
3. thing + subject pronoun 14 (19%) 14 (17%) 9 (12%)
4. thing that 10 (14%) 3 (4%) 22 (30%)
5. one thing 8 (11%) 4 (5%) 7 (9%)
6. the same thing 7 (10%) 4 (5%) 5 (7%)
7. the only thing 5 (7%) 2 (2%) 3 (4%)
8. another thing 4 (6%) 7 (9%) 4 (5%)
9. the thing 4 (6%) 4 (5%) 17 (23%)
10. good thing 2 (3%) 9 (11%) 2 (3%)
11. most important thing 0 11 (13%) 8 (11%)
by the L1s—about half the Persian use. The Chinese frequency was three
(4 per cent), indicating thing that was not as popular with them. When
that is placed after thing, it serves as the subject of the sentence or a con-
junction: these uses were evenly represented in the L1 interactions (five of
each), but it was used more as a subject than a conjunction by the Persian
speakers (14 instances as a subject, eight as a conjunction). Two of the
three uses of that by the Chinese performed as conjunctions.
In Table 4.17, among seven items (items 5–11), there is relative con-
sistency among the three groups for the first four items (one thing, the
same thing, the only thing, and another thing): the frequencies are all small
and close to each other. This consistency disappears in the last three
items. For item 9, the thing, while the L1 and the Chinese are exactly
the same (four instances each), the Persian speakers stand out by using
the cluster 17 times, about one fourth of the total thing they use. For
item 10, good thing, this time the L1 and Persian speakers are exactly
the same (twice each), but the Chinese use it nine times, four times
more than either group. The last item, 11, most important thing, was
not very different in frequency of use by the L2 groups (11 and eight
instances respectively), but is absent from the L1 data. There are even
more diverse phenomena in the data. Some clusters only occur in one of
the three groups: for instance, the Chinese use interesting thing six times
and most interesting five, but neither group uses them; and the Persian
speakers use thing that I five times, but it does not appear in the data of
the other groups.
4.5.4 Anything
As with all the placeholders examined so far except thing, the Persian
speakers were the most frequent users of anything, employing it twice
as often as the L1s (48 vs. 24) and four times as often as the Chinese
(48 vs. 12). In all groups, the percentage rate was 10 per cent or less of
the total occurrence of anything (Table 4.18).
Table 4.18 Anything clusters
1. anything in negative 10 (42%) 6 (50%) 39 (81%)

sentences
2. clause-final anything 3 (13%) 5 (42%) 15 (31%)
3. do anything 4 (%) 4 (%) 10 (%)
4. anything else 6 (%) 0 2 (%)
The Persian speakers used anything in a negative sentence 39 times:

81 per cent of their total frequency of 48. The Chinese used it six
times (50 per cent) and the L1s ten times (42 per cent). This indi-
cates that the Persian speakers used anything primarily in negative
sentences, twice as often as the L1s. The L1 group used anything more
in non-negative sentences, and the Chinese used anything equally in
negative sentences and in affirmative or interrogative sentences.
Another difference appears in the use of anything in clause-final
position. The L1s use it three times (13 per cent), the Persian speak-
ers 15 times (31 per cent) and the Chinese five times (42 per cent).
As with anything + negative sentence, the L1s were not keen on put-
ting anything at the end of a clause, showing a difference from the
L2 groups. Nearly half of the Chinese uses of anything appear in the
end of a clause, and it appears about one third in the Persian data.
None of the clause-final placeholders presented by the Chinese and
L1s served a turn-shifting role, but the Persian speakers provided six
instances (40 per cent) of clause-ending anything as a turn-shifting
device. That means that when the L1 and the Chinese speakers use
clause-final anything, the same speaker begins the next clause, but
for the Persian speakers, 40 per cent of the clauses ending with any-
thing signal the end of a speaking turn, and the next turn is taken by
another interlocutor.
Compared with placeholders discussed previously, few anything clus-
ters are observed in the data. Do anything is one of the handful clusters
present: four times each from the L1 and Chinese groups, and more
than twice as often (ten examples) from the Persian speakers. Another
noticeable cluster is anything else, but not all groups employ it: the
cluster is missing from the Chinese data, appearing six times in the
L1 and only twice in the Persian data. A number of anything clusters
emerged exclusively in the Persian data: have anything (13 instances),
not have anything (12 instances), not do anything (seven instances),
have anything to (seven instances), and anything + infinitive (seven
instances). This indicates that the Persian speakers have a stronger
preference for anything clusters than the Chinese or L1 speakers.
4.5.5 Someone
The occurrence of someone is rather like that of anything, with the
Persian speakers again the most frequent users and the Chinese the least
frequent. Someone appears up 31 times in the Persian data, slightly less
(28 times) in the L1 and half as often (13 times) in the Chinese. The L1
and Persian speakers are alike in using someone at a similar frequency.
The percentage distributions for all three groups are all less than 10 per
cent: 9 per cent for the L1, 5 per cent for the Chinese and 6 per cent for
the Persian speakers.
Someone is rarely found in clause-initial or clause-final position in
this data. The former is missing in the Chinese data, but occurs once in
the L1 data and twice in the Persian. The Persian group used someone
in clause-final position twice, the other two groups did not use it at all.
It seems that the three groups demonstrate consistency in positioning
someone: the L1 group used conjunction + someone seven times, the
Persian speakers three and the Chinese two. There are four kinds of
conjunctions in the L1s data: if to express condition; and to express
addition; because to express reason; and or to refer to choice. The Persian
speakers use only two, if and and, while the Chinese use the conditional
if. The only conjunction appearing in all three groups is if. The L1s
chose to use a more diverse combination of conjunctions + someone,
while the two L2 groups were less inclined to do so.
Group differences appear in the use of prepositions (about, for, to, and
with) occurring before someone: the Persian speakers used this type of
cluster the most often with 11 occurrences, the Chinese twice and the
L1 speakers once. Somebody + conjunction occurs only once in the L1s
and Chinese interactions respectively, both being someone and; and the
Persian speakers used someone + conjunction three times, with one use
of each and, if and or.
4.5.6 Somebody
As one of the less frequent placeholders, somebody is used slightly less
than someone. The highest frequency of somebody (23) was among the
Persian speakers; the L1s offered 20 examples, and the Chinese, with
11 occurrences, employed someone about half as often as the other two
groups. The percentages reveal a small proportion of the data consisted
of somebody, under 7 per cent for each group.
The participants showed little inclination to use somebody in clause-
initial and final positions. The only interesting trend in this regard was
that this placeholder occurred in clause-initial position and in clause-
final position with the same frequency in each group: that is, somebody
occurs twice in the final position and twice in the initial position in
the Persian data, and once in each position in the Chinese and L1 data.
There are occasional examples of a conjunction occurring before
somebody: the Persian speakers used and once and if three times, and the
Chinese and L1s even less: and was employed once by the Chinese and
or once by the L1 speakers. The pattern of conjunctions occurring after
somebody is more or less similar: the Persian speakers used and twice,
and the L1s used and as the only conjunction after someone; the only
conjunction used by the Chinese was or.
The L1s chose who as one of the two relative pronouns after somebody,
while the Persian speakers preferred that. Relative pronouns were not
used in the Chinese interactions.
4.5.7 Anybody
The least frequently used placeholder in the data is anybody. The Persian
speakers employed anybody 17 times, and it appeared half as often (nine
times) in the L1 interactions and only three times in the Chinese data.
The percentage analysis, however, levels off these differences, and all
occur as less than 4 per cent.
Anybody occurs in clause-initial position as a turn-taking device,
three times each in the Persian and L1 data, while the Chinese used
all three tokens of this placeholder in clause-mid position. The Persian
speakers used anybody in interrogative clauses seven times, in four of
which the question mark appears immediately after anybody. The L1s
presented only three interrogatives including anybody, and none in
which anybody comes directly before the question mark. The Persian
speakers used anybody else four times, two in negative sentences and
two in questions.

200
177
180
160
140
108
120
94
100
85
80
80 72 82 74
54 48
60
40 28 31
13 2412 20 11 23 17
20 9
3
0
ng
gs
ne
ng
y
od
od
in
in
hi
eo
hi
Th
eb
yb
Th
et
yt
m
An
An
m
m
So
So
So
Figure 4.9 Frequency distribution of placeholders

As a category of vague expressions, placeholders were found to be

most popular with the Persian speakers (478 uses), then with the L1s
(301) and least among the Chinese (286). The Persian speakers used
every placeholder item the most with the exception of thing, which was
strongly preferred by the Chinese.
A trend emerges in that in the Chinese data, there are two sets of
similar frequencies. The first consists of the first three items in Figure
4.9: something, thing, and things, which are evenly distributed. The same
trend can be seen in another three items: someone, anything, and some-
body. Even more interesting is that for the second set, the Chinese and
the L1s follow a consistent matched pattern, with the L1s using each
item approximately twice as often as the Chinese.
Analysis of the percentage rate of placeholders reveals some simi-
larities between the Persian and L1 speakers. As Figure 4.10 shows, the
first three placeholders, something, thing, and things, have the same
ranking in the two groups. Together, they constitute three quarters of
L1 speaker CSLE
4%
3% 1%
7% 4%
Something Something
4%
Thing Thing
8% Things
31% Things
Someone 28% Someone
9%
30%
Anything Anything
18% Somebody Somebody
24% 29%
Anybody Anybody
PSLE
4%
5%
Something
Thing
10%
37% Things
6% Someone
Anything
23% 15% Somebody
Anybody
Figure 4.10 Percentage distribution of placeholders

all placeholders: 73 per cent by the L1s and 75 per cent by the Persian
speakers. Eighty-seven per cent of the overall placeholders used by the
Chinese consist of the same three items, which means that the Chinese
used placeholders in a more concentrated way or less uniformly than
the other groups.
The remaining one quarter of placeholders in the L1 and the Persian
data is composed of the other four items, someone, anything, somebody,
and anybody, with similar distribution, the only difference being in the
different ranking of someone and anything. While anything (at 10 per
cent) is more frequent than someone in the Persian data, someone is more
frequent than anything in the L1 data.
Despite the fact that the Chinese performed differently from the
other groups, there is a substantial consistency in its rates of use for the
three least frequent items in the data: someone, anything, and somebody,
each make up 4 per cent of the placeholders.
The lexical realization of vague language was more prominent in L2

classrooms than in L1. The Chinese used twice as much vague lan-
guage as the L1s, and the Persian speakers also used more than the L1s,
although to a lesser extent.
Table 4.19 shows that the L1s used 1,567 vague expressions, the
Chinese 3,030, by far the most, and the Persian speakers 1,718 in the
middle position. Of the five categories of vague language, the Chinese
used four categories the most and the Persian speakers one; the L1s
alternated between second and last positions.
The largest proportion of vague language in the L1 data is represented
by vague quantifiers (27 per cent); in the Chinese by vague intensifiers
Table 4.19 Frequency of lexical categories of vague language
Category L1 CSLE PSLE
Subjectivizers 205 741 282

Possibility indicators 238 379 190
Vague quantifiers 423 741 435
Vague intensifiers 400 883 333
Placeholders 301 286 478
Total 1,567 3,030 1,718
(29 per cent) and in the Persian by placeholders (28 per cent). These per-
centages are similar in quantity but different in kind. It seems that the
Chinese were more emphatic than the other two groups, using substan-
tially more vague intensifiers. In the same way, the smallest proportions
refer to a different category in each group: the smallest in the L1 data
is subjectivizers (13 per cent), in the Chinese placeholders (9 per cent)
and in the Persian possibility indicators (11 per cent). The proportions
within each of these largest and smallest groups show a rough consist-
ency across the three sets of participants.
5
Pragmatic Functions of
Vague Language
The data show that participants aligned their vague language use with
their communicative needs, stretching vague language to smooth the
flow of communication. The growing body of literature on vague lan-
guage acknowledges the pervasiveness of this often underestimated
feature of natural language in communication (Channell 1994; Cutting
2007; Jucker et al. 2003; Ruzaitė 2007; Zhang 2015). The present
chapter will deal with a functional analysis of vague language in the
three data sets.
Channell (1994, p. 194) provides the following categories of vague
language functions: giving the right amount of information, deliber-
ately withholding information, using language persuasively, lexical
gaps, lacking specific information, displacement, self-protection, power
and politeness, informality and atmosphere, and women’s language.
Drave (2002, p. 26) believes ‘The major function of vague language is
to tailor conversational contributions to the perceived informational
needs of the other participant(s) so as to maintain and enhance the
on-going relationship’. The major functions of vague language are
listed as:
• filling lexical gaps (where a speaker cannot recall a word or where

one does not exist in the language)
• filling knowledge gaps (memory lapse)
• emphasizing (and de-emphasizing) certain information
• deliberately withholding specific information
• conveying tentativeness
• conveying an evaluation of, or expectation about, a proposition
• maintaining an atmosphere of friendliness, informality or reference.
(Drave 2002, pp. 26–27)
113
The familiarity of Channell’s (1994) early list of vague language func-

tions and other frequently quoted sources (Cutting 2007; Jucker et al.
2003; Prince et al. 1982) means that this list is not free of dispute. One
source of controversy seems to originate from the different classifica-
tion systems developed to refer to vague language. This can be seen in
the application of the terms ‘softener’ or ‘downtoner’ to refer to the
same phenomena. Another controversial area comes from the incon-
sistent use of functional and lexical terms as interchangeable items. For
instance, Jucker et al.’s (2003) system mixes the functional category of
downtoner with the lexical category of placeholder.
Vague language functions in this study have three broad catego-
ries: mitigation, right amount of information, and structural func-
tion. Each has subcategories: mitigation consists of self-protection,
politeness, and downtoning; right amount of information consists
of approximation and quantification, emphasizing, possibility, and
uncertainty; and structural function consists of repair, hesitation, and
turn management.
As diverse as its lexical realizations, vague language can perform a
wide range of functions, depending on the context where it occurs.
The five lexical categories discussed in Chapter 4 do not necessarily
match up with one, and only one, function. As Ruzaitė (2007) points
out, ‘Quantifiers expressing a big number, e.g. many, much, and loads,
predominantly perform the sub-function of “emphasising”. Quantifiers
referring to small quantities mainly perform the sub-function of
“mitigating”. However, the function of quantifiers can change depend-
ing on the preceding intensifier, for instance quite a few is emphatic,
whereas a few without the intensifier quite is mainly used as a mitigator’
(p. 161). Possibility markers can be used for possibility, politeness, and
right amount of information or discourse management: its diversity is
revealed in the following analysis.
5.1 Mitigation
Martinovski (2006, p. 2) defines mitigation as ‘a pragmatic, cognitive

and linguistic behaviour the main purpose of which is reduction of vul-
nerability’. Focusing on quantifiers and approximators, Ruzaitė (2007)
points out that ‘Mitigating quantifiers can mitigate not only a quantity,
but also the force of request, apologies, advice, instructions and criti-
cism’ (p. 183). Mitigation in this study will fall into three subcategories:
self-protection, politeness, and downtoning.
Pragmatic Functions of Vague Language 115
5.1.1 Self-protection
One of the common uses of vague language is as a tool with which
speakers can protect themselves (self-protection, self-defensive). This
tool ‘attends to the face needs of the speaker’ or serves ‘other-protective
face-work’ (Trappes-Lomax 2007, p. 135). It can also help to protect
the face of the listener or a third party. There seem to be existing areas
of overlap between self-protection and politeness. Discussed under the
term shield by Prince et al. (1982), the uncertainty is found to result
from two different sources: plausibility shield, to express doubt, and
attribution shield, through which the speaker attributes a belief to some-
one else. Below are three examples showing how self-protection works
among the three groups of participants in this study.
(5.1): This is a discussion between two L1 participants over nine turns.
They are talking about online search and database during a tutorial.
S1: Okay, library services to phy- physically handicap. d- yes? (L1:

3:186)
S14: Um, I completely screwed mine up. <SS: LAUGH> and, and, when
I did the search for my first facet I was doing two at a time with
the descriptors, and then when I combined ‘em, to make, at the
end I ha- I think I had four groupings. I used and instead of or,
and then when I finally combined all three facets I got a big fat
zero. (L1: 3:187)
S1: And you know why. (L1: 3:188)
S14: Yes. And I, don’t know why I did it but…<SS: LAUGH> (L1: 3:189)
S1: Okay. Did you have it right on your, script? (L1: 3:190)
S14: Yes. (L1: 3:191)
S1: Or your, okay. Well I’ll take a look at e- you know, it’s good_ did
you realize then when you were online when you did that or after
you signed off and you, stewed about it for a while? (L1: 3:192)
S14: After I stewed a little while. <SS: LAUGH> (L1: 3:193)
S1: Okay, and, this is sometimes typical, when you’re first learn-
ing how to search. that’s why I try to say don’t revise online,
because sometimes it does take you a little while to sort of figure
out, what exactly did I go wrong, where I wound up with, zero
or I wound up with, thirty thousand, you know sometimes,
when you’re online it’s just too hard to assess that. [S14: yeah
unfortunately] yeah I sort of like the old days, I guess, maybe,
I’m just one of those, analog print people, when we had these
long printouts, and you could look through that long printout
while you were online. Now with this awful, Telnet, animal, you
can only really look back a few screens, and I just find it very,
disconcerting. I I wish I could look back, all the way to when
I began, and D S is really the only thing we have, to accom-
modate, that browsing backwards, in your search. <PAUSE:04>
yeah? (L1: 3:194)
In Excerpt (5.1), S1 is discussing the good points of non-online revis-

ing. In supporting her claim, she gives an example of how an analogue
printout can be more advantageous to Telnet (a network protocol) in
turn 194. To reinforce the argument, she gives an example of why she
thinks it can be more helpful. Assuming that there might be some
disagreement, she prefers to use two vague words consecutively to
defend herself against being wrong: I guess, maybe means she might be
wrong and there might be other reasons for preferring these kinds of
old-style revision.
(5.2): This excerpt is a discussion between nine Chinese participants
over fourteen turns. They are talking about what the teaching job
involves and what is required to be a good teacher.
S3: I think, in one word, the most important thing about teaching
is, huh, teaching, teach, teach students how to be a man, yeah.
(Ch: 7: 143)
S9: I have an opinion. Huh, I want, I think if you want to be a good
teacher, you have to be three Ps. < SS: LAUGH> (Ch: 7: 144)
S2: Be patient, profession? (Ch: 7: 145)
S9: Performance. (Ch: 7: 146)
S4: Performance, yeah. (Ch: 7: 147)
S6: Passion. (Ch: 7: 148)
S2: Three Ps. Patience, passion, and performance. (Ch: 7: 149)
S1: What about you? (Ch: 7: 150)
S8: Yes, patience is an important factor. Many students pay more
attention to you, in your class. Therefore your class is efficient
I think. I think this is very important. (Ch: 7:151)
S7: I think, huh, I don’t think passion is a very important thing
because I think passion is a temporary thing which cannot exist
long. I think we should choose what we like, and what we love.
Yes, it is the most important. (Ch: 7: 152)
S5: To be a teacher is maybe very easy, maybe not, maybe very dif-
ficult. It is all up to you and, huh, hope you have a bright future.
(Ch: 7: 153)
S6: What kind of job do you think, the, the teachers, what kind of
job? Very honourable or very boring, simply as, huh, as stable job?
Huh. (Ch: 7: 154)
S5: Different people have different answers. (Ch: 7: 155)
S6: What about you? (Ch: 7: 156)
In Excerpt (5.2) the discussion begins with a moral view of the teaching
job by S3 in turn 143, and proceeds with S9 mentioning three criteria,
the three Ps required for being a good teacher. Following S8’s confir-
mation of one of the required factors in turn 151, S7 through I don’t
think in 152 softly expresses disagreement with the criterion passion as
a required element, and suggests love as a replacement. In turn 153, S5
changes the direction of the conversation to whether or not teaching is
an easy job. She first claims it is very easy to be a teacher, but quickly
uses a vague expression maybe in the negative form, then another maybe
to present an opposing possibility that teaching may be very difficult.
Vague language is used here to protect the speaker against opposing
views. The speaker’s hedged view is reinforced by ‘it is all up to you’ in
the same turn, implying that the speaker is going to attribute the valid-
ity (truthfulness) of her statement to the would-be teacher’s discretion.
S5 uses vague language to protect herself against responsibility for her
utterance.
(5.3): This is a discussion between two Persian speakers over five
turns. They are talking about democracy and freedom.
S3: Don’t say ‘but’ look I know if you want to say something, you get
punished in this place, Ok? The knowledge that you see, the peo-
ple that you see, it is accepted. What you see to happen. But over
there, there is no excuse. When you say to yourself this is supposed
to be the land of freedom, these people are supposed to know
everything. Ok? Then you see abuses there … come and see what
they do with, come and see what the story is over there. Find out
about McDonald’s story, find out about Rockefeller, what he did or
he raised all the crisis in South because there was competition and
in the north there was all his, he raised the High Street Times, ok
guys. Find out about John F Kennedy. Then you can see, hey, this
is not Rafsanjani. Yes. (P: 1: 499)
S2: But. (P: 1: 500)
S3: But. (P: 1: 501)
S2: But, you know this level of intellectuality if I am right, you know
here if we have the maximum, one hundred, ok? I think in Iran
it is twenty. I think I don’t know European countries maybe it is

thirty but it is much more. Ok? Maybe it is fifty. I agree. Maybe it
is the same as here but there are differences. (P: 1: 502)
S3: By the way, at the beginning of what you say, OK? There was
George W Bush’s autobiography (xx). I guess it is. It is a good
movie. You have to see. It is a true story. (P: 1:502)
In Excerpt (5.3), S3 is trying to convince S2 that the USA is no different

from the other countries as long as freedom is concerned, but S2 argues
that it is still ahead of many countries. Self-protective use of vague
language manifests in turn 502, where S2 initially uses ‘if I am right’ to
indicate that he may be wrong and smoothly tries to express contrast.
As he proceeds, he opts for more self-protection devices and uses double
vague language devices, ‘I think, I don’t know’ while insisting on his
position. He finds even this degree of protection insufficient and uses
maybe twice as another self-protective device to emphasize the contrast.
What is distinct in this example is that while emphasizing his stand,
the speaker zigzags between the contrasts with multiple self-protective
vague language devices. On the one hand, he attempts to highlight his
own view, and on the other tries to reserve room for protection in case
the opposite is proven right.
5.1.2 Politeness
When people speak, there are certain linguistic choices they make,
which indicate the social relationship perceived to exist between the
interlocutors. One such phenomenon can arise in expressing polite-
ness. ‘Politeness as it is understood in linguistics involves more than the
common-sense notion of politeness as the conventionalized observance
of certain social norms which spell out the appropriate ways of thank-
ing or greeting’ (Nikula 1996, p. 92). Politeness is mainly associated
with Brown and Levinson’s (1978, 1987, 1994) politeness theory, which
stands on the pillar called ‘face’ raised by Goffman (1967).
Face is defined as ‘the public self-image that every member wants
to claim for himself’ (Brown & Levinson 1978, p. 66). Each individual
needs to look after the face of others in case they need their face to be
maintained: in other words, interactants need to mutually look after
each other’s face. Face in Brown and Levinson’s politeness theory is
divided into two separate but related categories, positive face and negative
face. The former deals with the individual’s desire to be liked and appre-
ciated by others while the latter ‘concerns a person’s want to be unim-
peded and free from imposition’ (Tracy 1990, p. 210). Positive politeness
strategies aim at satisfying participants’ needs for approval, and hence

include things like exaggerating agreement with the interlocutors,
showing interest, and noticing the hearer’s wants and needs. Negative
politeness strategies help satisfy participants’ need for autonomy by
indicating the speaker’s reluctance to impose on others’ territory and
to restrain their freedom of action. Being indirect, using hedges, and
veiling responsibility with the use of impersonal forms are examples of
negative politeness strategies (Nikula 1996, p. 93).
There is also the concept of the face-threatening act, which occurs
when communication ‘runs contrary to the face wants of the addressee
and/or the speaker’ (Brown & Levinson 1978, p. 70). According to
Channell, ‘vagueness is used as one way of adhering to the politeness
rules for a particular culture, and of not threatening face’ (1994, p. 190).
Ruzaitė more specifically refers to quantifiers as ‘a politeness strategy to
minimize face-threat’ (2007, p. 183). Brown and Levinson (1987) and
Aijmer (1997) maintain that one of the manifestations of politeness
strategy occurs in the use of I think to fulfill the function of mitigating
face threat. Below are examples of how vague language performs polite-
ness functions in the three data sets.
(5.4): This is a discussion between five L1 participants over 11 turns.
The speakers are talking about publication rules for newspapers.
S12: Well if you can’t identify whose body it is it’s not really so much
an invasion of privacy. (L1: 1:447)
S1: Mhm. Okay, lots of times, a picture of a body uh, you don’t see
the, the face, [S12: yeah] you know, either it’s covered or just the
angle, s- you just see a, you know, a form. [S12: yeah.] um, so
there’s the question of the distance. What about uh, related to
that is how it’s played, right the play of the photo what, what
does the book say about that? I think that’s in there, isn’t it? On
the checklist? See it on the checklist or in one of the case studies.
is there anything about how the photograph is played questions
they ask about how it’s played? (L1: 1:448)
S13: Didn’t it say something like if it was, pl- like, under the fold it’s
not in the checklist but didn’t it say something like it’s, not as
harmful? (L1: 1:449)
S1: Yeah you’re right um, I think there’s another checklist, after the
first checklist. Um, page two-twelve. (L1: 1:450)
S12: Oh that’s right (L1: 1:451)
S1: This was the checklist, [S13: oh] after the um. [S5: oh, yeah] the
campus tragedy. (L1: 1:452)
S5: Instructional value? (L1: 1:453)

S1: Instructional value, mhm. Is it possible to present the image in
such a way that it reflects, its instructional value without inflict-
ing undue emotional distress? So present the image that could
I- get involved in the play, whether it’s on the front page how big
it is, so forth. Um, and, point four is very important disclosure
what’s what’s that about and why is that so important? Yeah?
(L1: 1:454)
S16: You have to be able to justify why you put the picture in in the
first place, [S1: mhm] to the readers. (L1: 1:455)
S1: So can you remember from, some of the case studies for today
any examples of where, there was an explanation that was, given?
(L1: 1:456)
S16: The wasn’t the, middle finger one, [S1: Yup] in the, the one paper
a couple of them wrote the, right there to the reader, [S1: mhm]
on why they printed it and then there was one that didn’t, that
got like the most complaints and the one asked for feedback
from, all the readers saying, [S1: okay] (that wanted their) opin-
ions. (L1: 1:457)
In Excerpt (5.4), the participants go through each criterion one by one.

S12’s comments on invasion of privacy are not a statement but a request
for further clarification. S1 seems to possess the right knowledge about
this, and in turn 448 tries to politely express disagreement with S12.
When S5 raises the issue of instructional value, S1 in turn 454 makes a
polite request, using ‘is it possible…’, to present the image of the dead
body in such a way that it tends to be viewed for its instructional value
rather than to provoke emotion.
(5.5): This is a discussion between two Chinese students over seven
turns. They are talking about language learning.
S3: I asked some students, teachers and my foreign friends. It is

just, you have, huh; you don’t have, huh, learning a foreign
language atmosphere. You have to create it just every day; look
at the foreign newspaper first and, huh, you’ll speak something.
You should think it in English and then speak it. They just told
us we should build atmosphere for us. And, huh, we just, hu,
listen to something just like BBC, VOA which is familiar for
us. But I think that’s not enough, we need more chance to, to
actually practice it because language is, huh, like our mother
tongue. Why can’t we speak so fluently? Because we speak every
day, every time, every minute, every second. So we are familiar

with most of it, so it is really, it is not really easy for us. Not
just, huh, like English, or French. We even have no chance in
class. We speak English but after class or after school we speak
Chinese or our mother tongue. So our oral English is not very
well and even we can’t, we can’t catch up with the foreign-
ers who speak just five years Chinese. When I am in ‘Expo’
American exhibition, there’s a handsome, handsome boy.
(Ch: 4:146)
S1: <LAUGH> Handsome. (Ch: 4:147)
S3: Yeah, he is really handsome. (Ch: 4:148)
S1: He speaks Chinese very well. When I asked how long have
you, have you, have you studied for Chinese? Just five years.
She, he said to me Chinese is really hard, harder than English.
I said oh really? I think it’s really easy. She said it’s your mother
tongue, so you say it’s really easy, maybe and I said as our
Chinese students, maybe, huh, we have studied English from
our elementary school, right? (Ch: 4:149)
Class: Yes. (Ch: 4:150)
S3: But our English is not very good, maybe that is a question. That
is the focus and our school and even elementary school are,
huh, maybe the department of education should pay attention
to this part. This is really important. Our study of English is
not just, huh, pass the exam. Huh, actually, we should know
how to use it and we should we can be a fluent; we can com-
municate with us, with our friends, our friends and travellers
easily. I think that is our destiny when we choose English as our
majors, right. (Ch: 4:151)
S1: I find, huh, I find a terrific video about how to improve our oral
English. I, I think it can help you. (Ch: 4:152)
In Excerpt (5.5), S3 is explaining what an ideal L2 learning situation

should be like. The first factor she names is the atmosphere, stating
that it can be created by the learner, like reading foreign language
newspapers and listening to BBC or Voice of America; but she finds this
insufficient and talks about few opportunities to speak English as the
main reason for the lack of fluency in English. S1 in turn 149 mentions
a foreign speaker of Chinese who speaks Chinese fluently after five
years. S3 then tries to justify why they do not make much progress in
learning English. Using maybe in turn 51, S3 is politely blaming the
Department of Education and criticizing how English is perceived to be.
She disapproves of the exam-oriented approach in language pedagogy

in the Chinese education system, and believes the ability to communi-
cate should be the goal of learning English.
(5.6): This is a discussion between four Persian speakers over 18 turns.
They are talking about the advantages and disadvantages of modern life.
S6: Okay, now the advantages and disadvantages of modern life.

I can say, there are goals, for instance, disadvantage can be that
people’s life is gonna be (xx) can say. These people are being
somehow like robots. (P: 4:152)
S2: Brainwashed. (P: 4:153)
S6: Yes, that now they sit somewhere, they can do everything. By sit-
ting somewhere and just working with the computer, not being in
variety or somewhere like this. About that question that is it good or
not, about this aspect it is not good, of course, but about the advan-
tages that people are getting more knowledge about. (P: 4:154)
S2: Different cultures. (P: 4:155)
S6: Yes, and huh … the nature they are living in. They are getting
more knowledgeable and they are knowing themselves too, so by
this respect, it can be good for them because they’re finding them-
selves and things like this. For example, Abed says that maybe
in the future we can be sure that there is no difference between
people and animals, just maybe the face, you know, you can see
that people thought that men are somebody and women are
somebody else. They are not like each other, but they are getting
to know that, we are the same in a lot of ways. They are human
and we are the same. You know, it’s an example that we are faced
with, we had in our life. You have proved it. (P: 4:156)
S2: Ok. (P: 4:157)
S1: By this example. You mean? (P: 4:158)
S6: You know. I mean that totally I mean it can be good, it can be
bad. We cannot say that it isn’t good. (P: 4:159)
S10: You know, can I say something? (P: 4:160)
S1: Sure. (P: 4:161)
S10: Ok. About what M said. I agree with M but we know we are going
to know lots of things. We are getting lots of knowledge, but
unfortunately, I think are drowned in lots of knowledge what we
are going to and this is a problem again; knowing lots of knowl-
edge, having lots of knowledge. (P: 4:162)
S2: Which one is wrong and which one is right? (P: 4:163)
S1: Yes. (P: 4:164)
S10: Knowing without having an aim. We have knowing that why we

know this and a-. (P: 4:165)
S2: You know advertising to … that flash colour. Sometime it is easy
to be cheered. What’s good could look good and what is wrong
could look right. (P: 4:166)
S10: Exactly, but it is not correct in all fields. For example, in scientific
a- (P: 4:167)
S2: The basic knowledge. (P: 4:168)
S10: Yeah, we should have a lot of knowledge but some fields just like
religious fields. You know it isn’t good to have a lot of knowledge
because it makes you to be I don’t know. (P: 41:169)
The discussion begins with a two-sided argument, with S6 counting

access to lots of knowledge as a merit. As the conversation goes on,
S10 in turn 160 politely tries to join the conversation, using ‘Can I say
something’. Something here is a vague term, implying that she might
express some disagreement. In turn 162, S10 starts with ‘I agree with M’
(S6) but immediately opposes S6’s argument, stating that what is taken
as a good point (‘getting lots of knowledge’) by S6 is in fact a demerit. It
seems that although the vague term something may generally appear to
be neutral in terms of the speaker’s position, in this particular excerpt
it implies the speaker’s position: disagreement to be expressed in a
polite way.
5.1.3 Downtoning
Downtoners or detensifiers (Hübler 1983) are what Prince et al. (1982)
call ‘adaptors’. Blum-Kulk et al. (1989) define downtowners as ‘sentential
or propositional modifiers which are used by a speaker in order to mod-
ulate the impact his/her request is likely to have on the hearer’ (p. 284).
They include words such as a bit, a little, and a little bit. Jucker et al.
(2003) state the vague terms ‘introduce vagueness into a proposition or
increase the degree of vagueness of an utterance’ (p. 1746). They also
claim that downtowners are used when speakers find that an available
word does not adequately cover the meaning they have in mind.
Pearson (1998, p. 103) states downtowners often consist of adverbi-
als (for instance just), modal words (for example can), and non-factive
predicators (for example one way of defining a … is). Wu et al. (2010)
maintain that mitigators such as probably and maybe may follow I think,
giving a much stronger downtoning function to it. Ruzaitė (2007,
p. 94) adds that ‘The quantifiers (a) little, a bit and a little bit minimize
the force of verbs and downtone the intensity of adjectives’. Below
are three examples of how downtoning is used by the three groups of

speakers.
(5.7): This is a discussion between two L1 participants over three
turns. They are talking about problems occurring in on-line revision.
S1: Or your, okay. Well I’ll take a look at e- you know, it’s good_ did
you realize then when you were online when you did that or after
you signed off and you, stewed about it for a while? (L1: 3:192)
S1: Okay and, this is sometimes typical, when you’re first learning
how to search. That’s why I try to say don’t revise online, because
sometimes it does take you a little while to sort of figure out, what
exactly did I go wrong, where I wound up with, zero or I wound
up with, thirty thousand, you know sometimes, when you’re
online it’s just too hard to assess that. [S14: yeah unfortunately]
yeah I sort of like the old days, I guess maybe I’m just one of
those, analog print people, when we had these long printouts,
and you could look through that long printout while you were
online. Now with this awful, Telnet, animal, you can only really
look back a few screens, and I just find it very, disconcerting. I,
I wish I could look back, all the way to when I began, and DS is
really the only thing we have, to accommodate, that browsing
backwards, in your search. <PAUSE:04> yeah? (L1: 3:194)
In Excerpt (5.7) S1 (teacher/tutor) states that it is a common problem

and she had warned them about it by asking them not to do any online
revision. She also notifies the class of the extra time needed as the
consequence of online revision. Then S1 in turn 194 engages in com-
paring the old system with the new system and gives preference to the
old system as it provides the user with a more convenient service, long
printouts, while she tries to undervalue the new system by downtoning
the significance of the service it provides as just allowing people to go
back ‘a few screens’. This downtoning of the effectiveness of the service
is then turned into an explicit criticism as ‘disconcerting’.
(5.8): This is a discussion between four Chinese participants over six
turns. They are talking about songs and singers and how songs can be
used in English language learning.
S5: Can you show us a song? (Ch: 6: 18)

S1: No, no, no. I maybe listen to those women singers but they really
are, good, very beautiful but it is hard for me to sing. (Ch: 6: 19)
S2: I want to introduce some singers to you just like, huh, Britney. She
has some songs like ‘Everytime’. It is so slow and beautiful. Once
I wanted to train my listening, and it is good to train my listening.
And if you want to improve a high level, you may, you may.
Of course, it is just my suggestion, choose some rap. (Ch: 6: 20)
S1: Rap? (Ch: 6: 21)
S2: It is a little slow, a slow rap, not so quick. I found it just like
Brittany’s Circus, although it is very fast. I like ‘New (xx). It is very
fast. The speed is very fast but (xx) I don’t know how many times
you have, you have heard it. Huh, you will feel it is not slow. Huh,
it is not fast at all and you can hear at your work clearly. I think
it is also a good way. It is up to you what kind of music you like.
(Ch: 6: 22)
S3: Yeah, I think it is, it is a way to enjoy life and some days I told me
that, huh, you, it’s necessary for everybody to learn to sing a song
very well and, huh, only I like ‘Terry Sif’. Yeah, I think her songs
are very beautiful. (Ch: 6: 23)
In Excerpt (5.8), S5 asks S1 to sing them a song, but S1 declines. S2 in

turn 20 tries to introduce some singers to others and recommends rap
as an appropriate style to improve their listening skills. S1in the next
turn asks for confirmation by uttering ‘Rap?’. This can mean either
that in S1’s view rap is not slow enough to be appropriate for language
learning purposes or that S1 is not familiar with rap and needs to
make sure if it really is appropriate to be used for language learning
purposes. S2 in turn 22 tries to adjust the reply to the possible disagree-
ment by S1 and uses the downtoner a little to soften the adjective slow.
This downtoner functions as the point of departure of a continuum,
starting with a little slow, continuing to slow and ending in very slow,
meaning that while rap is not as fast as some other music, it is not
very slow either.
(5.9): This is a discussion between two Persian speakers over three
turns. They are talking about the structure of a government.
S7: Ok, I think there is something about this country. You know for
sure I, I am agree with the previous regime, Shah and the King,
huh, 100 per cent, hey, I think they were wrong and I guess even
now but there is, huh, a little chance that if we changed ourselves
at that time, we changed ourselves, we could improve because they
changed the rules. (P: 1: 440)
S3: Infrastructure. (P: 1: 441)
S7: Yup, but there is some basis for building structures, building, build-
ing a house. You know, you cannot build a two-story building on
some weak basis. (P: 1: 442)
In Excerpt (5.9) S7 expresses his overall agreement with what the previ-
ous regime, the Kingdom, did. He even highlights it with the expression
one hundred per cent. S7 in turn 440 uses the downtoner a little before
chance to underline his view that even changing themselves would not
have been effective, because of the changes in rules that occurred. In
turn 442, S7 gives the example of building a house to show that for
anything to work, some foundation work is needed, just as a building
needs to have a strong foundation. To intensify the point, the speaker
uses two expressions: a two storey building and the weak basis. The latter
is presented in a weaker form, modified by some.
5.2 Right amount of information
This category of vague language functions is mainly associated with

Grice’s (1975) maxim of quantity (Channell 1994). It consists of two
parts: ‘1. Make your contribution as informative as is required (for the
current purposes of exchange). 2. Do not make your contribution more
informative than is required’ (Grice 1975, p. 45). Vague language is
one of the devices speakers can use to tailor their contributions to
interactions (Channell 1994). The category of right amount of informa-
tion is subcategorized into quantification, emphasizing, possibility and
uncertainty.
5.2.1 Quantification
Vague language performs the function of quantification when the
speaker realizes that precision is not necessary. This category is exempli-
fied by expressions such as some, many, much, few, little, several, a lot of,
plenty of, and large amounts of. Channell (1994) believes non-numerical
quantifiers help create implicature and thus avoid breaking the maxim
of quantity. Below are examples of how vague language performs quan-
tification functions for the three groups of participants in this study.
(5.10): This is a discussion between two L1s over eight turns. They are
talking about ERIC (Education Resources Information Centre) and web
searching.
S15: I don’t know I just, my experience in searching in general in

other systems has been that, usually, people don’t wanna wave
through to the end. They’re gonna look at the first ten depending
on, you know what their needs are. (L1: 3:141)
S1: This is actually a re- real important point. Why should it not mat-
ter in this case-searching Dialog? (L1: 3:142)
S15: Cuz everything, should be, as good, as the beginning searches.
(L1: 3:143)
S1: Right. There is no ranking, using Dialog at least, this classic,
Dialog or Dialog Classic that we’re using. A Dialog does have
another system called I believe, Freestyle. Which does do some
ranking, okay? But I think it only gives you the first fifty or, what-
ever. Oh we may get to that or you may wanna do that for one of
you s- uh search reports. Yeah? (L1: 3:144)
S15: So is that okay? (L1: 3:145)
S1: That’s it? Okay, good what facets did you have? (L1: 3:146)
S15: I had three facets, I did pregnancy, um, teenagers and dropping
out. [S1: okay] and then I expanded under each of those and
I didn’t, have as many under pregnancy, um, I limited it to, preg-
nancy slash D F and pregnant students. Um may (L1: 3:147)
S1: Okay, I, I would suggest to you, that there are some more [S15:
okay] like unwed mothers, and early childhood or early parent-
hood or whatever it was, that there some others to use. [S15:
okay] and, since there are so few, I would add more. [S15: okay]
on the other hand you got sixty citations so it didn’t necessarily
hurt but if you wanna be more comprehensive, you wanna add
more descriptors. Any other experiences on the search? Yeah.
(L1: 3:148)
In Excerpt (5.10), S15 starts with commenting on his experience in

using different search systems. The expression other systems that he
uses in turn 141 indicates that the system he is talking about is to
some extent different from them. He then gives an overall view of how
other systems are generally used by others. The way he speaks gives
the impression that he is not taken by this new search system and is in
favor of the others. S1 seems to be trying to underline the benefits of
the new system by asking a question in turn 142 that makes this option
stand out. In supporting the system, S1 in turn 144 tries to imply that
the old system still has drawbacks even if it involves ranking, by stat-
ing that it gives only the first fifty items. She even tries to extend the
privileges of the search system when S15 explains the three facets he
had and could not get sufficient sources for one item. Her expression
of some in turn 148 directly emphasizes the approximating quantity
of facets she could have used. This can be viewed in the examples she
gives through like or even emphasizes by or whatever it was, which indi-
cate that she is reinforcing the quantification by giving examples and
using or whatever.
(5.11): This is a discussion between three Chinese participants
over three turns. They are discussing public transport in Beijing and
Shanghai.
S5: What about transportation in Beijing? (Ch: 4:94)

S1: The transportation is, huh, is convenient but, huh, there are
many, lots of people and every time, every place, you just stand,
it just is just very tight. (Ch: 4:95)
S3: This is similar, is similar to Shanghai. When I, last year I travelled
to Shanghai, it is, it is a holiday maybe I forgot. The subway is
full of people and everybody’s expression is similar. They’re just
not talking, no speaking, just standing or sitting there. And think
them, about themselves and not like ‘Tingwang’ or ‘Tangwan’,
people very friendly. Maybe when we get on the bus, we will talk
with each other. They don’t. I am not. I am not get accustomed
with it. (Ch: 4:96)
In Excerpt (5.11), replying to S5’s question in turn 95, S1 looks at pub-

lic transport in Beijing from two perspectives. One is convenience, the
other, the congestion of passengers on public transport. The overcrowd-
ing is described by vague quantifiers many and lots of, referring to large
numbers or quantities. S1 assumes there is no need to specify a number
but roughly reflects this quantification. S3 confirms S1’s approximate
quantification in turn 96: it is similar in Shanghai where the subways
are overcrowded.
(5.12): This is a discussion between four Persian speakers over ten
turns. They are talking about the disaster that occurred in Hiroshima
during World War II and what happened afterwards in Japan.
S3: American for destroying Hiroshima and killing hundred fifty

thousand people and the nicest deal about Hiroshima history. If
you ever have the chance to go read the life, the biography of the
six people that threw the bombs down, you see what happened to
them. (P: 1: 172)
S4: They all kill themselves. (P: 1: 173)
S3: That is very interesting. Ok? America helped Japan to rebuild itself.
(P: 1: 174)
S2: What happened to them? (P: 1: 175)

S3: America helped Germany to rebuild itself. (P: 1: 176)
S2: What happened? (P: 1: 177)
S4: They killed themselves. (P: 1: 178)
S2: They committed a suicide? (P: 1: 179)
S3: Some of them died. Some of them got killed. The one who was
supposed to throw the bombs didn’t like to do that to happen, so
these guys threw the bombs. (P: 1: 180)
S4: It wasn’t the first time. They didn’t know its war; they are doing it
to kill, maybe. They (xx) one hundred people died. (P: 1: 181)
In Excerpt (5.12), when asked about the destiny of the pilots, S3 uses
some twice to express a vague quantity in turn 180, assuming this quan-
tification, rather than an exact number, can best fit the context. S3 may
or may not know the exact number, but apparently thinks that it is not
important here, and some serves the purpose better; otherwise, in the
next turn, S4 would have asked for clarification of the vague quantity
expressed by some. That did not happen, meaning that the vague infor-
mation was accepted by the listener with no fuss.
5.2.2 Emphasizing
Most intensifiers perform an emphasizing function. Intensifiers have
been referred to as boosters (Holmes 1990; Hyland 2000). Boosters are
mostly studied in comparison with hedges (Bradac, Mulan & Thompson
1995; Holmes 1990; Hyland 2000). Investigating the functions of boost-
ers in writing, Hyland states that ‘Boosters like clearly, obviously and of
course allow writers to express conviction and to mark their involve-
ment and solidarity with an audience’ (2000, p. 179).
Besides the discussion of intensifiers (boosters)/ hedges dichotomy
available in the literature, gender-related examinations of the applica-
tion of these forms make a substantial contribution to the study of
vague language (Bradac, Mulan & Thompson 1995; Holmes 1990).
Holmes’ work shows significant gender differences in the use of dif-
ferent boosters. Bradac, Mulan and Thomson (1995) demonstrate that
women show more consistency in using intensifiers than men. This has
been supported by other studies as well (McMillan et al. 1977; Mulac &
Lundell 1986; Mulac, Lundell & Bradac 1986; Mulac et al. 1988).
Wright and Hosman (1983) claim the overuse of intensifiers by female
speakers brings more interactiveness to their side in communication.
Even the context of communication has been claimed to contribute to
differences in the language used: Bradac et al. (1995) note that women
use more intensifiers when talking to women but more hedges when
talking to men.
Ruzaitė (2007) maintains intensification can also be expressed
through quantifiers. For instance, multal quantifiers can emphasize a
large quantity or long period of time. The other possible ways to add
emphasis, Ruzaitė states, are repeating the same quantifier (lots and lots)
or placing an intensifying premodifier (really) in front of a quantifier.
Below are examples of how intensifiers are used in the data to fulfill
emphasizing functions. The examples are not gender-specific but are
viewed from culturally and linguistically distinct perspectives.
(5.13): This is a discussion between five L1 participants over four
turns. They are talking about the death of a race car driver and the safety
rules applying to race car driving.
S3: Is there any posi- possible cause of death? I mean, if you s- you,
I saw it on T-V. So, you see just, crash. I mean and he was one of
the top, race car drivers right? And so I mean it. (L1: 1:38)
S1: So o- I mean obviously [S3: why do they need them] his death was
caused by the crash so that’s not the question [S7: they’re b-] the
ques- the question is what aspect of the crash specifically caused
his death? (L1: 1:39)
S6: There were a lot of questions about the type of restraints um, what
s- [S2: (like how to change)] like there’re so many different seat
belts and there’s certain ones, yeah. (L1: 1:40)
S1: Like how did it start? What happened to him at the moment of,
impact? Which part of his body, made contact with. (L1: 1:41)
In Excerpt (5.13), the discussion begins with a broad question by S3

regarding the possible cause of death, but S1 in turn 39 narrows the
question by asking about the specific aspect of the crash that caused
the death. S6 in the next turn further narrows it to issues related to
restraints, and as he proceeds, he restricts it further and arrives at a par-
ticular aspect. Using so (many) in turn 40, he emphasizes the diversity
of seat belts available and indicates that the seat belt caused the death.
This highlighting arouses curiosity in S1 as to how it happened and
what happened first; he asks these questions in the next turn (41).
(5.14): This is a discussion between four Chinese participants over
four turns. They are talking about university and university life.
S1: In fact, these prestigious universities provide many, huh, oppor-

tunities to many students. They can do different volunteering
things, but in smaller cities maybe we have, huh, less such activi-
ties. (Ch: 4:120)
S3: My friends in Jason University, in holidays, they are only at home
five days, so after that they will go back to school. They were study
hard, huh. For instance, TOFO and do some experiment, huh, do
many experiences about school. So it is very, they are very busy
and, and. They are, life is very full. (Ch: 4:121)
S4: I think our university life is boring. It is too boring. It is really bor-
ing. Just study and, huh, study, study. We have no to, we have no
chance to. Ok, I want to be a volunteer and to, huh. (Ch: 4:122)
S5: We have no chance. (Ch: 4:123)
In Excerpt (5.14), S1 points to advantages prestigious universities can

offer their students like volunteer jobs to prepare them for their careers.
S3 in turn 121 mainly refers to how hard working students at such uni-
versities need to be, and toward the end of her statement, she uses very
three times to emphasize hard work at such universities. In turn 122, S4
expresses her dissatisfaction with university life by using double inten-
sifiers (too boring and really boring) to overwhelmingly emphasize these
negative aspects. S4 continues, expressing the same idea in different
words: instead of using an intensifier, she repeats the negativity three
times to highlight it: ‘Just study, and, huh, study, study’.
(5.15): This is a discussion between four Persian speakers over seven
turns. The discussion is on the responsibilities of being a parent.
S3: As our parents had, for example, the problems that were between,
the problems that were between my parents with their parents,
huh, were more than our problems, ok? And in the future our
problem will have less I think. (P: 7:322)
S1: This generation is getting more aggressive. (P: 7:323)
S2: Actually, I think your problem will be worse, not less. (P: 7:324)
S5: Exactly. (P: 7:325)
S2: You know nobody can know everything. He is right. Lack of
information can cause problems. I believe personally being a par-
ent comes with a great responsibility. You have to have really,
really. You have to be really talented to be a parent and you have
to have lots of responsibility. For example, if you want to have a
child, you have to think over everything. You have to know a bit
of psychology. You have to know a bit of, I don’t know, whatever.
(P: 7:326)
S1: Sociology. (P: 7:327)
S2: Sociology, yes. You have to be into politics. You have to know lots
of things to be parent. Some people think it really really ideal a
child is coming, is growing up, is as easy as that but it is not this, it
is not this. I mean he is right. If his parents know about computer,
maybe they would encourage him to do it. I mean my daughter
was. Sorry if I keep examples of my children. I am just speaking of
my experience. (P: 7:328)
In Excerpt (5.15), S3 is of the opinion that the gap between the new
generations will be narrower, based on observations of his parents’ and
his generations. He is disagreed with by other interlocutors. S1 uses more
aggressive in turn 323 and is disagreed with by S2, with ‘will be worse
not less’ in turn 324. In turn 326, as an actual parent, S2 gives details
of requirements to be a good parent. She adopts an emphatic approach,
trying to address the problems facing parents and the responsibilities
parents should feel. The first factor she counts, talented, is emphasized
by the intensifier really; then she refers to the broad concept responsibil-
ity again and uses a quantifier, lots of, for emphasis. So she first empha-
sizes by using really and then uses a different phrase, lots of. This shift
between intensifier and quantifier for emphatic purposes is also seen
in turn 328: S2 uses the quantifier lots of first, to refer to what has to
be known to be a parent, then resorts to an intensifier to highlight the
emphatic tone: really really ideal.
5.2.3 Possibility
Prince et al. (1982) examine the possibility function under both approx-
imators and shields. In this study, it serves to express different degrees
of possibility.
(5.16): This is a discussion between six L1s over ten turns. They are
talking about how social control works in the society and why people
do what they do.
S1: Sure. I mean, I I mean, social control is obviously, not perfect,
so um <PAUSE WHILE WRITING ON BOARD> so yo- so young
people. Um what does that tell us about young people, um,
if young people are more likely to say, steal something? You
might wanna talk to her after class just to, find out what she’s
doing and, whether you wanna participate in it. (L1: 2:8)
SU-m: that’s you. (L1: 2:9)
S3: Me? Okay. <SS: LAUGH> (L1: 2:10)
S1: Um, okay so what does that, what does that tell us already, if if
young people, are more likely to steal things than, than o- um,
older people? Yeah. (L1: 2:11)
S4: Either that they’re, more rebellious, or maybe, or just that
they’re not as accustomed to, society’s norms yet. (L1: 2:12)
S1: Yeah that’s, yeah, I mean, yeah rebellious, or, or not we, we
could call it socialized, which basically just means they’re not
accustomed to society’s norms yet. Yeah. (L1: 2:13)
S2: Or they’re just too young to understand the uh, consequences
of, stealing (L1: 2:14)
S1: Oh that’s yeah, and this is important um, the consequences …
somebody else have something they wanted to say? (L1: 2:15)
S5: There also like, isn’t like as many consequences for them.
(L1: 2:16)
S1: Yeah, that’s true, I mean, um why do you think that there are
not as many consequences? I mean why do, why do um, why
does society_ why is our society set up so there won’t be as
many consequences? (L1: 2:17)
In turn 8, S1 chooses the example of a young person stealing something.

S1 continues in later turns to ask whether young people are more likely
to steal than older people. S4’s reply to the question in turn 12 contains
two possibilities which seem to oppose each other, expressed through
either … or … and the possibility indicator maybe. The first possibility
attaches a negative tone to young people, describing them as ‘rebellious’,
whereas the second one associates them with a softer attribute of inability
to adapt themselves to the norms of society. S1 in turn 13 agrees with S4
by reiterating the ideas and possibilities, but the device used to refer to
the possibility is or only. The same device (or) is resorted to by S2 in turn
14 to refer to a possibility, but a new possibility is introduced this time,
which is being too immature to understand the consequences of stealing.
(5.17): This is a discussion between two Chinese participants over eleven
turns. S2 is talking about her plans for a trip to Japan with her pen pal.
S2: My pen pal will come to see me, who is doing a course in Oriental
studies. I will go along with her to Japan. (Ch: 1: 168)
S1: Wow, so it, sounds really interesting. How do you get along?
(Ch: 1: 169)
S2: We’ll probably use one of those very fast and poor trains to get
there and then go by taxi or on foot. (Ch: 1: 170)
S1: Oh, it’s a good choice and how do you go around? (Ch: 1: 171)
S2: On foot or by boat? (Ch: 1: 172)
S1: And, huh, what do you want to wear? (Ch: 1: 173)
S2: Huh, it depends on time of the year. I would want to go there
in spring for the cherry blossom, so probably just jeans and a
sweatshirt. I’d make sure I had a clean pair of, pair of, socks or and
or some slippers shoes because I think you have to take off them
when you visit the temples. (Ch: 1: 174)
S1: Yes, it is a good idea. What would you buy? (Ch: 1: 175)
S2: Nothing. Tourist things. I might buy an electronic gadget like a
calculator. They’re supposed to be cheap in there. (Ch: 1: 176)
S1: Huh, what would you eat and drink? (Ch: 1: 177)
S2: I’d look for ‘Western Food’ and probably end up eating at
McDonalds. She can’t stand raw fish and she doesn’t like rice
much, either. (Ch: 1: 178)
In Excerpt (5.17), S2 uses multiple possibility indicators, as she is not

yet sure if things will work out as planned and prefers to explicitly
highlight this possibility in her talk. Turn 170 is a compound sentence
with a possibility involved in each clause. The first is indicated by
probably, which is applied to the main means of transport in ‘We’ll
probably use one of those very fast and poor trains to get there’. The
second possibility, associated with the other means of transport, is
expressed by a possibility indicator, or, rather than a vague possibility
indicator. What can be inferred from turn 170 is that S2 uses probably
because the number of options to choose from is not specified in the
first part of the sentence; but the two specific options involved in the
second part of the sentence show or as best fit for the sentence. Turn
174 involves possibility as well, expressed through probably, which
demonstrates that choice of clothing by S2 depends on the weather.
This can be confirmed by so in turn 174, which operates as a conjunc-
tion to express result.
In reply to S1’s question on what she is planning to buy, S2 in turn
176 again elevates possibility in her remarks but through a different
word, might, this time. Nothing at the beginning of turn 176 cannot
be interpreted as nothing literally, as S2 immediately continues with
a broad category for shopping: tourist things, which implies that the
speaker has not yet made up her mind what to buy or has no need to
provide a specific list.
What seems to be noticeable in S2’s reply in turn 176 is that she initi-
ates possibility by being broad first and gradually narrowing it down.
Tourist things is too broad to be readily guessable by the listener. It is

narrowed by electronic gadget, but gadget is still a general term. It seems
that the speaker realizes this is an inadequate answer to the question
and attempts to clarify it by giving an example of what she means by
gadget. The last part of the sentence looks at the reason for the possibil-
ity inherent in turn 176, suggesting that the reason S2 prefers to give
more possibility weight to this sentence is that she expects such devices
to be cheap there; otherwise she might not buy any. As with the other
two questions in turn 177, the last question by S1 elicits a possibility-
involving answer: ‘probably end up eating at McDonald’s’ in turn 178
is justified in two regards: S2 prefers Western food, and her pen pal does
not like raw fish, which is common in Japan.
(5.18): This is a discussion between six Persian speakers over 14 turns.
They are talking about non-face-to-face communication.
S7: It’s easier when you write something or speak with someone; talk
with someone on the phone. It’s easier for you to talk some issues
or problems that you cannot tell them face-to-face. (P: 2:75)
S1: Ok. Sometimes writing is much better, huh, when you cannot
speak easily and speak some (P: 2:76)
S8: Problems and a– (P: 2: 77)
S9: But, it isn’t common, you know, the writing. (P: 2:78)
S1: Ok, why in our daily life, sometimes we want to speak with each
other, instead of saying directly, ok? (P: 2:79)
S9: Yes. (P: 2:80)
S1: We say to our partner. Ok. Go home I will call you, yes? Why? (P:
2:81)
S2: Because we are are ashamed of. (P: 2:82)
S8: Because by phone we can talk together easily. (P: 2:83)
S7: Maybe we want plenty of time to speak in a better condition, in.
(P: 2:84)
S1: In a more relaxed situation? (P: 2:85)
S7: Yes. (P: 2:86)
S5: Maybe we need sometimes to prepare ourselves to say that. (P:
2:87)
S1: So these kinds of instruments help us. (P: 2:88)
S7 and S1 both agree that non-face-to-face communication is the most

convenient, but S9 in turn 78 points out that it is not the most com-
mon method. Without expressing agreement or disagreement with S9,
S1 in the next turn raises a question about preferring non-face-to-face
communication. S2 and S8 immediately provide answers using because,

but the following answers by S7 and S5 in turns 84 and 87 are initiated
with the possibility marker maybe, indicating that what they state is not
categorical.
5.2.4 Uncertainty
Lack of information brings about uncertainty, associated with Grice’s
Maxim of Quality which is stated as ‘Do not say that for which you
lack sufficient evidence’ (1975, p. 46). Channell (1994) discusses uncer-
tainty under displacement, which occurs mostly when talking about
past and future events, adding that there are instances that go beyond
tense constraints, as uncertainty exists in the present as well. Channell
points out that vague language is resorted to when the speaker feels
stressed by the uncertainty of the subject, lack of knowledge/vocabu-
lary or an unequal relationship between participants. Examples of how
uncertainty is expressed through vague language across the three data
sets are discussed below.
(5.19): This is a discussion between two L1 participants over nine
turns. They are talking about an editorial on a crime.
S1: Hm. There’re some more examples, from, the case studies yeah?
(L1: 1:466)
S12: Um the, the, the kid that got shot outside the convenience store
[S1: mhm] the editor, um, wrote wrote a, column was it the same
day of the paper? I’m not sure, but he wrote a column explaining
why they, [S1: yeah] why they ran it cuz at first he didn’t wanna
run it. (L1: 1:467)
S1: You can actually see it, y- you can see the column, [S12: yeah]
tu- turn the page. (L1: 1:468)
S12: Yeah, oh. (L1: 1:469)
S1: Well it’s in there somewhere I don’t know. (L1: 1:470)
S12: Yeah yeah it’s right here. Oh wait, (L1: 1:471)
S1: Th- (L1: 1:472)
S12: No no no, that’s not it. (L1: 1:473)
S1: I know that it’s in there somewhere. I just saw it. Anybody r- (L1:
1:474)
In Excerpt (5.19), S1 asks about more examples and S12 refers to one
that was published in the paper, but there is uncertainty as to where
in the newspaper the column appears. In turn 468, S1 tries to help S12
locate it and asks him to turn the page. In turn 470, the vague term
somewhere is indicative of uncertainty of the location of the article,

and the speaker tries to highlight the uncertainty by using I don’t know
in the final position. The double affirmative marker yeah yeah in turn
471 by S12 seems to suggest that he managed to spot it, but the word
wait indicates otherwise. In turn 573, S12 makes it clear that he has
not found the location. S1 in turn 474 then confirms the uncertainty
regarding its exact position, using the vague term somewhere again in
‘it’s in there somewhere’.
turns. They are talking about their future careers.
S5: Shean, what are you going to do when you grow up? (Ch: 7: 9)
S1: Huh, when I was a small child, I always wanted to be a teacher.
Maybe because teaching is the only profession I’ve seen and
I’ve had this dream for about, huh, many years. But and I chose
university teacher when I was in college. But, now, when I, huh,
when I, huh, graduate from this university, I already have doubt
that if I really want to be a teacher in the future, maybe, I will
choose another job for me in a later time. What about you?
(Ch: 7: 10)
In Excerpt (5.20), when asked about her future job, S1 says that there is
a gap between her childhood dream job and the job she might pursue
in the future. She highlights the transition from her childhood job
to a partially different position by pointing to her career at college.
When it comes to her current favourite job she is unsure, saying ‘I
already have doubt if I really want to be a teacher in the future’ in
turn 10. The uncertainty is reinforced by the following vague word
maybe, where she explicitly talks about choosing a different job at a
later stage in life.
(5.21): This is a discussion between five Persian speakers over twelve
turns. They are talking about the reasons for cave paintings.
S1: And go to the past, old time. Why, for instance, in cave now we
found some pictures? (P: 2:103)
S3: Yes. (P: 2:104)
S1: Yes? You know what is the aim of this kind of pictures? In the
cave? (P: 2:105)
S5: It’s I think … (P: 2:106)
S1: By that writing they want to communicate with the next genera-
tion. (P: 2:107)
S7: Yes. (P: 2:108)

S1: Yes? (P: 2:109)
S5: I think some parts of these pictures was some religious reason.
(P: 2:110)
S1: Don’t you think that these persons were alone and didn’t have
anybody to speak with them. Instead of speaking, they write some-
thing. (P: 2: 111)
S8: Because if they were alone, were must be artists. And their paint-
ings in the cave show that … for next generation. (P: 2:112)
S7: Maybe they wanted to transfer some kind of information which
they had. (P: 2:113)
S3: Culture. (P: 2:114)
In Excerpt (5.21) S1, who raises the question of the reasons for cave draw-
ings, tries to answer it in turn 107. The answer is agreed with by S7 in
turn 108, but in turn 109, S1 says Yes?, expecting more detailed replies.
S1’s strategy proves to be effective, as S5 proceeds with a reply (religious
reason) that contains some uncertainty (in the exploratory vague term
I think). In turn 111, S1 tries to elicit more responses by making points
such as ‘feeling alone’ as a reason for cave drawing. As the discussion pro-
ceeds, more reasons come up, and more uncertainty is revealed in replies
expressed through maybe by S7 in turn 113.
5.3 Structural function
Besides fulfilling a lexical function, vague language can also facilitate

the structural flow of information by performing a strategic function
in the communication process. Jucker et al. (2003, p. 1739) state that
‘Vagueness is not only an inherent feature of natural language but
also—and crucially—it is an interactional strategy. Speakers are faced
with a number of communicative tasks, and they are vague for strategic
reasons’.
Focusing mainly on approximators and quantifiers in academic con-
texts, Ruzaitė (2007, p. 187) finds that ‘Discourse management is
especially important in academic discourse since metastatements with
quantifiers help teachers organize discourse and make interrelations
between the future, present and previous discourse’. Her analysis
reveals two important patterns in which vague quantifiers contribute
to discourse management. The first one, called general-specific, refers to
a situation where a specific comment follows a generalization with a
quantifier; the second one, specific-general, denotes a general comment

following a specific comment.
There are different types of pragmatic strategies relating to the
structural functions of vague language that concern the mechanics of
communication; these strategies are mostly employed to deal with dis-
course management. They examine the effect of such factors as lapses
at discourse level and how these are coped with by interactants. In this
section, the structural function of vague language is studied under three
subcategories: repair, hesitation, and turn-management.
5.3.1 Repair
The study of speech repair has been significantly influenced by the
work of Schegloff, Jefferson and Sacks (1977), which introduces repair
sequences as segments. The sequence involves a repair-initiating turn
followed by a coming turn that creates the outcome of the repair,
referred to as execution of repair by Rutter (2008). Either participant
in the conversation can produce these turns. Schegloff et al. (1977)
divide repair sequences into four types, summarized by Rutter (2008,
p. 36) as
(1) Other-initiated other-repair: when the recipient of the trouble

source both initiates and executes its repair.
(2) Other-initiated self-repair: when the recipient of the trouble source
initiates, or calls for, its repair but the architect of the trouble
source executes the repair themselves.
(3) Self-initiated other-repair: when the architect of the trouble source
is responsible for initiating the repair, but its outcome is brought
about by the recipient.
(4) Self-initiated self-repair: when both initiation and execution of
repair are carried out by the trouble source architect.
More specific vague language-related investigation of repairs is associ-

ated with what Prince et al. (1982, p. 94) call non-substantive self-repair.
It involves repetition and self-repairs that include the replacement of
a word or a phrase. Their study reveals frequent use of self-repairs in
physician–physician discourse, and confirms the frequent occurrence
of shields in self-repairs rather than approximators. Ruzaitė (2007)
claims that the act of correcting occurs consciously, and that quanti-
fiers and approximators are deliberately used in such contexts. She
indicates that ‘Self-correction is an important aspect of classroom
communication, where correctness is a principal requirement’

(p. 189).
Repair in this study primarily refers to self-repair, also known as self-
correcting. Below are three examples of how vague language performs
the correction function among the three groups of participants.
(5.22) This is a discussion between three L1 participants over eleven
turns. They are talking about a crime case.
S1: This is anonymity in terms of their names right? (L1: 1:4)

S2: Right (L1: 1:5)
S1: How about their, images, their faces? [S2: um they show a] can can
newspapers take pictures and publish the photograph without the
name? Or is that not in the judge’s ruling? (L1: 1:6)
S2: Um it didn’t say, [S1: mhm] but I mean it shows a picture of them
when they were ten, and so I guess if you’re like, if they look,
I don’t know about you but I look exactly the same as when I did
when I was little, <SS: LAUGH> so it’d be really easy to tell. (L1:
1:7)
S1: Well I have less hair than I, than (I did then.) <SS: LAUGH> (L1:
1:8)
S2: So, um (L1: 1:9)
S3: Was the initial sentence longer than eight years? (L1: 1:10)
S2: Yeah the initial sentence was for fifteen years of [S3: oh okay]
detention, but um, [S1: (xx) a good time] they switched it to eight
years, so just until they were eighteen [S1: mhm] and um, yeah
so it doesn’t I mean I guess they don’t even I don’t even think
the papers know what these boys look like now, [S1: mhm] and
so there’s a chance that they look nothing al- [S1: mhm] alike and
that they’ll, live the rest of their lives in, anonymity or there’s a
chance that maybe people will figure it out, and I don’t it doesn’t
really say what the papers are allowed and not allowed to do it just
says that, (L1: 1:11)
S1: Cuz cuz wherever they’re living, you know let’s say they’re living
in some town somewhere in England, once you publish the pho-
tograph anybody in the town that, sees that person on the street
is gonna know who they are. [S2: right] then you might as well
publish their name because it, it then becomes public knowledge.
(L1: 1:12)
S2: Yeah I don’t think that they’ve been, um press has been, granted
access to them
[S1: mhm] so I don’t know if people know it. (L1: 1:13)

S1: But I just didn’t know whether the judge specified image as well as
name. (L1: 1:14)
In Excerpt (5.22), the participants talk about two ten-year-old boys

who killed a two-year-old in England, and how the court found them
eligible to be released from prison after eight years because they felt
remorse. They are discussing how the boys are to be given new identi-
ties and a chance to begin a new life. S1’s question about the kind of
anonymity the criminals will have is asked in two parts (turn 4 and
turn 6). The first is associated with anonymity in terms of names, but
anonymity of their images seems to be to a certain extent controver-
sial. S2 is trying to address the issue in turn 7 but has two unsuccessful
starts: ‘if they’re like, if they look’ are both abortive. The third start
begins with a vague phrase I don’t know, which reveals that the speaker
is correcting herself.
S3 changes the direction of the conversation in turn 10 by ask-
ing whether the term of their sentence was more than eight years.
Answering S3’s question in turn 11, S2 faces the same false start chal-
lenge as she makes a few corrections: ‘So it doesn’t’ is followed by vague
terms I mean and I guess as correction markers until the speaker comes
up with the appropriate start. In ‘yeah so it doesn’t I mean I guess they
don’t even I don’t even think …’, the speaker finds that ‘they don’t
even’ is still insufficient to express what he wants to say, so he changes
to ‘I don’t even think …’.
(5.23): This is a discussion between five Chinese participants over five
turns. They are talking about what they are planning to be in the future.
S2: Yes, I agree with you. I’ll try my best. (Ch: 2:60)
S5: I am going to be educated further, huh, because during these
years, my second degree is Chinese, so I want to be an editor, so
I want to go to a newspaper office. Maybe, I want to be a teacher
because you know, as a teacher the happiest thing is giving the
knowledge you have learned to the, to your students and you,
and you see them grow happily. So this is my goals. (Ch: 2:61)
S3: I think I will find a job first. Maybe, it’s very hard but just like
Yang Fan [one of the students, noted by the author]I think huh the
four years study really cost my parents so much and I really want
to support myself. So if you, I can, I think, I will find a job. I did,
I don’t know what kind of job I will get, so I have many plans.
I really want to try different kinds of jobs. Since my mother is a

teacher, she warns me to be a teacher, too. Huh, she thinks, huh,
the teacher might be suitable for me, is suitable for me. Maybe she
is right, but I really want to try something new, something differ-
ent. I don’t know, I don’t know what to do but maybe I will try
something different. (Ch: 2:62)
S1: I think if the students can, if the kind, the child can be your stu-
dent, they will be happy. (Ch: 2:63)
S6: I think I also will find a job first. I want to be a tourist guide
because I want to travel very much. I have a dream. I can, can,
travel China in ten years and all my life I can travel all the world.
(Ch: 2:64)
In Excerpt (5.23), S5 provides an assertive answer to the future career

question in comparison with S3. In turn 62, S3 uses I think for the
purpose of correction to compensate for a false start, as in ‘So if you,
I can, I think, I will find a job’. The speaker makes two false starts: ‘so
if you’ and ‘I can’ and makes up for this inadequacy by using I think,
which functions as a self-repair marker. This is in line with what Wu
et al. (2010) claim, that I think can perform self-repair for Chinese EFL
learners.
(5.24): This is a discussion between three Persian speakers over seven
turns. They are talking about social problems they experience in daily life.
S3: That is the price that you have to pay for your own dignity.
(P: 1:360)
S2: You. (P: 1:361)
S12: Alan, it is like that when you go to the nature. Ok? And you see
everyone threw out their garbage in the environment. Ok? Huh,
you said that I want to change myself, I would not do that. Huh,
maybe the others don’t do that, but I actually it is you are saying
it is about yourself, that’s you did it. (P: 1:362)
S3: Said it in an example. (P: 1:363)
S12: And maybe, and maybe. The others when they see you. (P: 1:364)
S2: They learn it. (P: 1:365)
S12: Yeah, they learn it. (P: 1:366)
In Excerpt (5.24), S12 points out, in turn 62, some common problems
faced quite often in the society, and tries to present a simple example
that demonstrates how changing oneself can contribute to the evolu-
tion of a society. The speaker makes a false start where he refers to a
contrast by but, and immediately picks up a vague word, actually, to

make up for the error and makes a new start that flows to the end of
the turn.
5.3.2 Hesitation
Hesitation markers are defined as ‘a set of tools with certain time dura-
tion that are used to solve oral discourse generation and reproduction
problems and that can be both retrospective (for example correction
of a produced discourse piece) and perspective (for example planning
problems of the coming discourse piece)’ (Khurshudyan 2006, p. 1).
Stubbe and Holmes (1995) believe that discourse markers provide
speakers with verbal planning time, but this function has rarely been
attributed to vague expressions. Wiese (1984) claims hesitation may
appear in different forms such as filled pauses (for instance uh, mhm),
repetitions, corrections, and drawls. Focusing on I think by Chinese
learners of English, Wu et al. (2010) find that EFL learners use this
marker to show difficulty in finding the right word to use next. The
following examples present instances of hesitations occurring among
the three groups.
(5.25): This is a discussion between four L1 participants over 13 turns.
They are talking about why a photo of an accident was chosen from
among other photos to be published in a newspaper.
S17: If they had to, sh- show one, of the a- from the accident scene.
(L1: 1:515)
S1: And why would you prefer that? (L1: 1:516)
S17: M- I don’t know. <SU-f: LAUGH>. (L1: 1:517)
S1: Okay. (L1: 1:518)
S17: Cuz it doesn’t, I don’t know I just, think seeing dead bodies in
a newspaper I don’t (L1: 1:519)
S1: I mean it, it bothers you. (L1: 1:520)
S17: Yeah. (L1: 1:521)
S1: Okay. Yeah? (L1: 1:522)
S3: I, I mean I agree that it it’s bothersome and that would be help-
ful but then if you you know after the explanation was made if
you look at, um the letters that they received um, I was kind of,
shocked to think that the reader would I mean that some of these
would be, <CLEARS THROAT> excuse me sent into the newspaper
about, if you think about basically, you know if you think about
how many lives were saved from this or if somebody thinks of
this image, it’s it’s it is upsetting and it is powerful and that’s why
they wanted it in there because, once an image like that is stuck

in your head it’s there, and then whatever, possibly whatever
images were in the photograph or whatever it is, can trigger off
those, images that you have in your mind and it, can prevent
something similar from happening. (L1: 1:523)
S1: Instructional value. Yeah? (L1: 1:524)
S6: I mean th- they didn’t mention it here but also like I think some-
times it hits home more like a picture of the, students, before an
im- like a a normal [S3: mhm] picture of them. [S1: mhm] because
it shows them you know like, a- and then like maybe a picture of
the of the of the actual car or something like that, [S3: mhm] cuz
it it sort of you know it humanizes you I think. (L1: 1:525)
S1: Mhm (L1: 1:526)
S11: Well at that point they didn’t know who the students were or
who the people were. (L1: 1:527)
In Excerpt (5.25), S17 expresses her preference for a particular photo but
once asked the reason for the preference by S1 in turn 516, proceeds
with a quick answer, I don’t know in turn 517. Although S1 implies this
answer is sufficient, S17 decides to elaborate on the reason for her pref-
erence in turn 519, but still finds the reason elusive. She uses the same
answer as before, I don’t know, but this time it functions as a device to
help her cope with her hesitation. In other words, this device allows her
to buy more time to think of an answer. There seem to be other markers
which highlight the state of hesitation by S17, like I just, think right after
I don’t know and even the final position I don’t, when the speaker meant
to say I don’t know again but was interrupted by S1.
S3 in turn 523 expresses agreement that publishing the photo may be
bothersome, but raises the advantages it has brought by making refer-
ence to the letters the newspaper received. Her attempt to specifically
point out one of the advantages raises the need to use some devices
to handle the hesitation brought up by memory lapse. She resorts to
expressions basically and you know to buy time to think and arrive at the
implication of this advantage, which is many lives saved.
(5.26): This is a discussion between five Chinese participants over
eleven turns. They are talking about Disneyland in different countries.
S7: Have you ever been to Disneyland? (Ch: 4: 317)

Class: No. (Ch: 4: 318)
S2: Do you want to be there? (Ch: 4: 319)
S4: Even the Disneyland in Hong Kong, we didn’t. (Ch: 4: 320)
S3: You know, Disneyland is now in a building in Shanghai. Maybe,

maybe one day. Maybe four years later or one year later we can
go there. (Ch: 4: 321)
S1: I hope we can go to Disneyland together. (Ch: 4: 322)
S3: Yeah, that is a good choice. Good idea. [Tapping on S1’s shoul-
der]. (Ch: 4: 323)
S1: We must earn money. First earn money. (Ch: 4: 324)
S1: Money is really important, important, and important. (Ch: 4:
325)
S3: Maybe we are poor people. (Ch: 4: 326)
Class: Yeah. (Ch: 4: 327)
As can be seen in Excerpt (5.26), this extract begins with a question

about Disneyland from speaker 7 in turn 317. As no one has been to
Disneyland, the discussion moves to another aspect following S4 in turn
320, who points to Disneyland in Hong Kong, and S3 in turn 321 who
notifies the class of Disneyland in Shanghai. However, when it comes
to further comments on Disneyland in Shanghai, she appears hesitant,
maybe realizing that there is nothing more to say about it, and imme-
diately decides to express hope that they can visit it in the future. This
hesitance is shown in the repetition of the vague word maybe, maybe.
(5.27): This is a discussion between five Persian speakers over ten
turns. They are talking about why the Iranian nation is alone in the
world and not supported by others.
S3: You think that we are alone? (P: 1: 202)

S10: Yes. (P: 1: 203)
S3: In the world right now. (P: 1: 204)
S10: After this government. (P: 1: 205)
S3: Because we are wrong or because we are right? I ask my question
again. You think that the reason why we are alone in this world
right now is because we say the truth because we are right about
what we say or because we are wrong? (P: 1:206)
S1: Yes. (P: 1:207)
S2: I think, I think. I think we are following the wrong path to be
right. (P: 1: 208)
S3: Very good. (P: 1: 209)
S1: Yes, good. (P: 1: 210)
S9: In some points, we are right. In some points we are wrong but
they are making rules, rules wrong things so big because they’re
powerful, because they have got great advertisements. (P: 1: 211)
Excerpt (5.27) begins with a question by S3 in turn 202. This is con-

firmed by S10 in turn 203. Turn 206 involves a question as to whose
fault it is for creating the current situation: is it because they are telling
the truth or because they are making a big mistake? To answer the
question, S2 in turn 208 goes through a state of hesitation and tries to
buy time by saying I think three times. This triple I think is used as a
device to make up for the delay in the response.
5.3.3 Turn management

Allwood et al. (2007, p. 276) believe turn management ‘is coded by
three general features turn gain, turn end and turn hold’. Section 3.1.2
of this book presents some explanation of turn-management in con-
versation. The following examples show how vague language is used to
manage turns smoothly, across the three data sets.
(5.28): This is a discussion between four L1 participants over twelve
turns. They are talking about whether an autopsy photo should be pub-
lished by a journal.
S1: Have they said we will not publish the photo? Or have they just
said we don’t intend or so … (L1: 1:66)
S12: (xx) We don’t intend to. (L1: 1:67)
S5: Right. (L1: 1:68)
S1: That’s different from saying we won’t, won’t do. (L1: 1:69)
S12: Right. We don’t intend to well what if something else comes up?
(L1: 1:70)
S5: Right the wording was is has no intention of publishing photos.
(L1: 1:71)
S1: No intention of publishing. (L1: 1:72)
S5: It was their attorney who said the photos are important because
they might reveal what caused, [S1: mhm] Earnhardt’s death. [S1:
mhm] and then the other th- as I said before the other thing is that
that, in Florida, granting public access to autopsy photographs is
permissible if it’s not part of a criminal investigation. [S1: mhm] so
the other feather in their cap is the fact that if it was another state
then we we might not be having this argument. (L1: 1:73)
S1: I guess another question would be has this happened in the past?
And it, has there been an autopsy photo that, news media have
obtained access to? You don’t happen to know do you? (L1: 1:74)
S5: Uh <SS: LAUGH> (L1: 1:75)
S1: Not not that you’re the expert on this uh, (L1: 1:76)
S2: Thinking back to who’s died recently no. (L1: 1:77)
In Excerpt (5.28), S12 is probably the one who talked to the people in
charge of the journal and is passing the answer to others. Or he might
have read it in the journal, or heard it from a representative on TV. S5
continues the discussion by pointing out the importance of photos as
stated by the attorney, and also to the requirements for the photo to be
published. S1 in turn 74 tries to take a turn by using the vague expres-
sion I guess to make a new point as to whether autopsy photos have
previously been made available to the news media.
(5.29): This is a discussion between four Chinese participants over
fifteen turns. They are talking about how culture can contribute to dif-
ferences in communication.
S4: Culture difference. Having a difference between China and other

countries, it really exists. (Ch: 5: 391)
S2: We can know what he really means in his words, with his words.
(Ch: 5: 392)
S4: I think as, as soon as that we major in English, we major in
Japanese or major in French, we learn foreign, foreign languages,
and we must first, we must first learn the culture and we must
know something about the culture. Culture is very heritage. For
us, it is very beneficial. (Ch: 5: 393)
S1: Yes. (Ch: 5: 394)
S2: Maybe culture maybe can help us to improve our language.
(Ch: 5: 395)
S3: Yeah, that is right. (Ch: 5: 396)
S4: We can enrich our knowledge. Then we can maybe have, we have
different kinds of information. (Ch: 5: 397)
S1: Do you know any other culture shock between Chinese and for-
eign countries? (Ch: 5: 398)
S4: Let me think a while. (Ch: 5: 399)
S2: I think in our culture. I didn’t know it’s when we nod our head, it
means different things. (Ch: 5: 400)
S4: Yeah, yeah. I remember that we has [nodding head down] this means
yes. [Nodding to side] this means no but in a culture that (Ch: 5: 401)
S3: I know it is an India. (Ch: 5: 402)
S4: [Nodding head down and laughing] Just this is no. It is interesting.
(Ch: 5: 403)
S2: Maybe, it is difficult for Chinese to, to talk with the Indian.
(Ch: 5: 404)
S3: Huh, I think there are some, there are still some things that are
common for which, we have been like each other. (Ch: 5: 405)
Excerpt (5.29) begins with a reference to culture and cultural differ-

ences. To join in the discussion, in turn 393, S4 begins with I think to
highlight the link between language and culture. A noticeable trend
is the consistency S2 shows in taking turns, beginning with a vague
expression in turns 395, 400 and 404. In turn 395, S2 starts the turn
with the turn-taking marker maybe, which leads to the comment that
culture can contribute to better language learning. S1 raises a new ques-
tion in turn 398, which calls for some examples of cultural differences
between China and other countries.
Using another vague expression, I think, as a turn-taking device, S2 in
turn 400 provides a particular example of gesture: nodding or shaking
the head, which can cause cultural confusion. In the next turn, S4 gives
a specific example of this cultural difference and S3 and S4 perform sup-
portive roles and approve S4. S2 in turn 404 agrees with S3 and S4 by
using a vague turn-taking word, maybe.
(5.30): This is a discussion between four Persian speakers over twelve
turns. They are talking about how a mother treated her son and the
consequences of her behaviour.
S6: Unfortunately, his mother because of decreasing the emotional

events and decreasing the sentimental hobbies of her son, he
bought for him everything that he wants. For example, every toys,
every instruments, every books that there was in the CD. In some
buying, I was with him. (P: 7:59)
S1: While he was shopping for the things. (P: 7:60)
S6: Yes, and after all when there is, there was no other thing.(P: 7:61)
S1: Nothing else left <LAUGH>. (P: 7:62)
S6: She decided to buy a computer for him and I saw unfortunately it
became, huh, when her mother asked him Harry do you want any-
thing? Do you want any food? He said that mom, you are wrong.
I am a superman. I am a superman without any eating. (P: 7:63)
S1: I don’t need to eat. I don’t need to be fed. (P: 7:64)
S4: Actually, this kind of thing can influence the personality of these
children. (P: 7:65)
S1: Ok. (P: 7:66)
S4: And about the physical problem. You know, when you spend lots
of time in front of the monitor, your eyes become hard actually
and you. (P: 7:67)
S1: So you will lose your eyesight. (P: 7:68)
S4: Yes, and you need to wear glasses, unfortunately. (P: 7:69)
S5: Ok, you are talking about the psych, you know, actually the physi-
cal problem children will face. (P: 7:70)
In Excerpt (5.30), S6 in turn 59 describes what the story is and how

the mother treated her son, and continues this into turn 63 despite the
two interruptions by S1 in turns 60 and 62. Following S1 in giving an
example of what the child could have said to his mother, S4 tries to
take a turn to comment on this kind of behavior in turn 65. The vague
expression actually indicates that the speaker means to push in and at
the same time show her position. This turn-taking device can also be
interpreted as the beginning of an utterance that will change the direc-
tion of the conversation, as it initiates a discussion on the consequences
of this kind of behavior to children.
It is almost impossible to put vague language into absolute categorical

classifications of pragmatic functions, and consensus on the pragmatic
functions of vague words is hard to reach. However, it is possible and
useful to analyze its pragmatic functions, because it can provide a
tentative picture of how vague language can contribute to enhanced
communication and also how enhanced communication can be taught
to the learners of a language.
This chapter investigates how vague language can be used as a multi-
functional device in communication. Adopting three main categories,
mitigating, right amount of information and structural function, it
examines how one function can be manifested through diversified
lexical categories. Each functional category in this study has been sub-
categorized, such as self-protection, politeness, and downtoning for
mitigation; quantification, emphasizing, possibility and uncertainty for
right amount of information; and finally repair, hesitation and turn-
management for structural function.
Unlike most works on vague language, in addition to considering the
usual pragmatic functions of mitigation, this study also focuses on struc-
tural functions at discourse level. The first part, mitigation, concentrates
on how vague language is used to lessen the strength of an utterance for
different purposes. Right amount of information, the second category
in the functional analysis of vague language, deals with how the effect
of insufficient information can be shown through the use of vague
language. The section on structural function analysis characterizes how
vague language contributes to the dynamics of interactive approaches

in the process of communication. A vague word has the potential to
appear in different contexts and perform different functions. There is
no single match between lexical categories and functional categories of
vague language. For example, as a lexical category, quantifiers can serve
several functions, emphasizing, quantifying, revealing uncertainty, or
self-protecting.
6
General Discussion
This chapter discusses the trends of vague language shown in this

study across the three data sets in relation to linguistic, cultural, and
pedagogic factors involved. The focus here is the combination of overall
and individual occurrences. Differences in the frequency distribution
of items are discussed first, then the causes of the similarities and
discrepancies in the patterns are examined.
6.1 Overall frequency distribution
As can be seen in Table 6.1, vague expressions are used approximately

twice as often by the Chinese (3,030) as they are by the Persian (1,718)
and L1 speakers (1,567). The comparison of the performances of the
three groups highlights meaningful differences from the statistical
perspective, p < 0.05 (χ² = 361, d.f.8).
Based on the frequency distributions of vague items in the data, the
Chinese are found to be substantially vaguer than the Persian speakers
and L1s, and the Persian speakers slightly vaguer than the L1s. The two
L2 groups, particularly the Chinese group, resort to vague language more
often than the L1s: 5.91 per cent of words in the Chinese data (total
51,263 words) and 3.35 per cent of those in the Persian data (total 51,344
words) comprise vague expressions, while as the least vague group, the
L1 data (total 51,403 words) contains 3.05 per cent of vague expressions.
Of the five categories in Table 6.1, the ranking (from most to least) for
the Chinese is intensifiers, subjectivizers/quantifiers, possibility indica-
tors, and placeholders, showing that the Chinese tended to use vague
language most for emphasis and least for generalization. The group
used the first three categories on the ranking list quite heavily. The
ranking for the Persian speakers is placeholders, quantifiers, intensifiers,
151
Table 6.1 Overall distributions of vague expressions
Item L1 CSLE PSLE
Subjectivizers 205 (13%) 741 (24%) 282 (16%)

Possibility idicators 238 (15%) 379 (13%) 190 (11%)
Vague quantifiers 423 (27%) 741 (24%) 435 (25%)
Vague intensifiers 400 (26%) 883 (29%) 333 (19%)
Placeholders 301 (19%) 286 (9%) 478 (28%)
Overall 1,567 (100%) 3,030* (99%) 1,718* (99%)
subjectivizers, and possibility indicators, showing that they used vague

language most to generalize (the opposite of the Chinese) and least to
express possibility. The ranking for the L1 group is quantifiers, intensi-
fiers, placeholders, possibility indicators and subjectivizers, showing
that the L1 group preferred to use vague language most for quantifying
and least for expressing a subjective stance. This result indicates that
the three groups are not similar in their use of the five individual lexi-
cal categories of vague language. The Chinese used subjectivizers and
vague quantifiers evenly; placeholders were employed the most com-
monly by the Persian speakers. The L1s were moderate users of vague
language in this study. Perhaps because, benefiting from the elasticity of
vague language and stretching it to the required degree (Zhang 2011),
the L2 speakers can use it to compensate for inadequacies arising from
insufficient vocabulary and lack of knowledge (Channell 1994; Cheng
& Warren 2001).
The overall uses of vague language by the Persian and L1 speakers
are close, with the smallest difference lying in the frequency of vague
quantifiers (435 occurrences for the Persian speakers and 423 for the
L1s). The largest difference is found in the total number of subjectiviz-
ers: 205 occurrences for L1s and 741 for the Chinese, meaning that the
latter used the item 3.6 times more than the former.
The findings of this study support Metsä-Ketelä’s (2006, 2012) find-
ings that L2 speakers use vague words more heavily than L1s. However,
Drave (2002), who looks at native speakers of English versus native
speakers of Cantonese, concludes that the former turn out to be vaguer
than the latter, and finds ‘the rank order of most frequent items virtually
identical’ between the two groups (p. 29). The discrepancies between his
findings and those of this study may have been caused by a number of
General Discussion 153
factors, chief of which are the different groups of participants: American

English, Mandarin and Persian speakers (this study) vs. Cantonese and
English speakers (Drave’s, but which type of English speaker is unspeci-
fied); and the different scopes of data analysis: this study involves five
vague categories while Drave’s investigation focuses on two categories
of vague language (approximators and placeholders).
The L2 learners in this study found vague language to serve their
communicative needs. To meet these needs and achieve their commu-
nicative goals, they took advantage of the elasticity of vague language
(Zhang 2011) in their interactions and therefore used it more often
than the L1s. L1s did not use any of the five vague language categories
the most among the three groups, indicating that they do not sit at the
maximum occurrence pole of the continuum but shift between middle
and minimum positions. All the maximum occurrence poles are occu-
pied by L2 speakers, the Chinese for all categories except placeholders,
where the Persian speakers are predominant.
As shown in Figure 6.1, L1s move between middle and minimum
positions, which indicates that they used the various vague categories
either the least frequently or moderately. The Chinese use all except
one category the most often. The Persian speakers use one category
the most frequently, and are evenly positioned between middle
and minimum for the other four categories. The L1 group does not
need to make such large use of the elasticity of vague language as
the L2 groups. When using English, the L2 groups, particularly the
Chinese learners, find vague language a versatile tool and a blessing
to communication.
6.2 Clustering patterns
The pattern drawn from the data shows that the expressions used most
freely in multiple positions in the clause (such as subjectivizers) are used
more often by the L2 groups, especially the Chinese. With subjectivizers,
the differences between the three groups are statistically significant
(p < 0.05, χ² = 410.347, d.f.2), with the Chinese using the greatest and L1
using the least. Subjectivizers are employed most often by the Chinese,
and the L1s are found to be the least frequent users of this category.
As with subjectivizers, possibility indicators are significantly different
among the three groups (p < 0.05 χ² + 71.755, d.f.2), with the Chinese
using them most often and the Persian speakers least.
Subjectivizers can appear in clause-initial, clause-mid or clause-
final position. This is where the elasticity of vague language can arm
Maximum occurrence Minimum occurrence

Subjectivizers
CSLE
PSLE
L1S
Possibility indicators
CSLE
L1S
PSLE
Vague quantifiers
CSLE
PSLE
L1S
Vague intensifiers
CSLE
L1S
PSLE
Placeholders
PSLE
L1S
CSLE
Figure 6.1 Positions of the three groups in the use of five lexical categories
the L2 speakers with a powerful tool for communication. Adopting

Zhang’s (2011) slingshot metaphor, placing a vague expression in dif-
ferent positions in a sentence can help interlocutors stretch the rubber
band and arm themselves with a convenient tool for communication.
For instance, most of the possibility indicators in this study such as
may, might and possible, occur in specific positions. Vague intensifiers

primarily occur before adjectives, but subjectivizers are more flexible
applications, serving a wider range of purposes such as turn-taking or
turn-giving. As their flexibility allows them to be used for different
functions, they are highly preferred by L2 speakers, who arm themselves
with this versatile weapon to cope with different communicative needs
in classroom interaction.
The scope of vague language can be examined from the level of
vagueness within an utterance. The use of a vague item can transfer
vagueness to the element immediately preceding or following it (known
as phrasal/local vagueness); some other vague categories can be employed
to extend vagueness beyond the phrasal level and result in global vague-
ness (clausal vagueness), according to Zhang (2014). Vague categories can
vary in scope of vagueness: some categories like intensifiers are narrow
(local vagueness) and apply to adjectives, while others like subjectivizers
extend vagueness to the entire sentence (global vagueness).
Given the dichotomy of local and global vagueness, another possibil-
ity for the popularity of subjectivizers in both L2 speaker groups but
not the L1 is that they can be used at such different levels. The vague
categories popular with the L1s, quantifiers and intensifiers, tend to col-
locate with other immediate components; for instance, quantifiers with
nouns and intensifiers with adjectives; they are to an extent restrictive,
but the subjectivizers popular with the L2 speakers are of a broader
application.
turns. They are talking about limitations being imposed on society and
referring to an example in past Iran.
S2: … everyone wanted to have it, to put it but I think limitation

makes people do the thing that is limited. It is a principle. If you
want people to do something, limit it. (P: 6:482)
S4: Ok. Others? What do you think? What happens to our country in
the next century? (P: 6:483)
(6.2): This is a discussion between two Persian speakers over two turns.
They are talking about how social injustice emerges.
S5: You must hit everybody who violates the others’ rights. You know
I think social injustice is because of some people opinion. Because
they think they are better than the other race, racism. (P: 6:1259)
S4: Ahuh. (P: 6:1260)
In Excerpts (6.1) and (6.2), I think in each sentence covers the entire
sentence. In (6.1) it applies to the whole sentence following but, and in
(6.2) it refers to the entire sentence regarding social injustice. In con-
trast, in Excerpts (6.3) and (6.4) below, which exemplify the emphatic
nature of intensifiers, vagueness is locally linked to the segment which
immediately precedes or follows it rather than to the entire sentence.
They are talking about a photo in a newspaper.
S20: Yeah, I just grabbed this from the Michigan Daily which I thought
was really interesting that this is a normally a colour, uh daily, [S1:
mhm] and they got Ellerbe here in black and white. (L1: 1:735)
S1: Wonder why. (L1: 1:736)

They are talking about social norms.
S1: Well yeah and there’s, <SS: LAUGH> there’s another, there’s another
question. This is a really good, a really good point that you bring
up. Um, I mean we say that, and maybe you would I mean I’m not
saying you wouldn’t, but um there’s lots of people in the world, take
Bill Gates you know for an example, who um, make lots of money,
have lots of money or other people who inherit lots of money. They
don’t have to work, but they do. (L1: 2:82)
S16: It makes ’em happy. (L1: 2:83)
It is evident in these excerpts that really is used by the speakers to

emphasize the adjective immediately following it. In (6.4), even the
speaker finds really insufficient to emphasize the adjective good and pre-
fers to stress the emphasis by repeating a really good twice.
Vague quantifiers tend to hold local vagueness. Items under this
vague category like some, much, many, are of grammatical significance
and therefore ranked as frequent patterns in the syllabi of ELT books.
This provides L2 speakers with adequate practice of these items, which
may explain why in this study the L1 and L2 speakers used vague quan-
tifiers at similar percentage rates: the L1 groups’ use of vague quantifiers
was 27 per cent of their total vague items, the Chinese 24 per cent, and
the Persian 25 per cent.
While individual possibility indicators are evenly distributed in the
L1 data, the L2 speakers employed these items differently, for example
using maybe more than the L1s. The percentage values show close
distribution between four items in the L1 data: maybe (26 per cent),
may (24 per cent), might (24 per cent), and probably (18 per cent). The
values in the Persian and Chinese data reveal huge inconsistencies.
Maybe, with a frequency of 81 and 82 per cent respectively, indicates
that the L2 speaker groups inclined heavily toward this possibility
indicator. Maybe was used by the Chinese the most, while at the other
extreme the L1 group utilized far fewer maybe and the least maybe
clusters. The Persian speakers sat between the Chinese and L1 groups
in their frequency of use of maybe and its clusters. If we assume that
speakers who use maybe tend to be more tentative, then the Chinese
are more tentative than the Persian and L1 speakers, the Persian speak-
ers less tentative than the Chinese but more than the L1s, and the L1s
least tentative of all.
The Persian speakers preferred all placeholder items the most apart
from thing, favoured by the Chinese. While placeholders (with 478
tokens, 28 per cent) were the most frequent vague expressions of the
Persian speakers, the Chinese used this category as the least common
group of vague words (286 tokens, 9 per cent); the L1 group used
301 tokens (19 per cent). Placeholders seem to play a more promi-
nent role in the Persian speakers’ interactions than in the other two
groups’.
In the data, some clustering patterns appear that are uncommon in
English. The unusual patterns appearing in L2 speaker data may have
originated from insufficient exposure to English, which drove the par-
ticipants either to use their own creativity in using clusters or to borrow
relevant knowledge from their own L1. For instance, we may, which
does not appear in the L1 data, is used by the Chinese and the Persian
speakers to show that the speaker is attempting to say something indi-
rectly with a tentative tone. Specifically, we in this context is used by the
Persian speakers to show politeness by creating some kind of intimacy,
avoiding the explicit and direct disagreement and warning.
The Chinese and Persian speakers have a cultural preference
for indirectness in making a request or expressing disagreement
(Gudykunst et al. 1996; Zarei & Mansoori 2007). Additionally, we is
the subject pronoun a Persian speaker usually uses instead of I, to
show politeness. This is confirmed in the data, which shows that the
Persian speakers do not use I may; but the cluster occurs frequently in
the L1s’ discussions.
turns. They are talking about why it is thought that nuclear power is
needed in a country.
S6: Nuclear power is what we are trying to use. (P: 1:37)

S8: That’s our absolute right. Yes? (P: 1: 38)
S6: And, huh, because of some problems, we may not be able to use
it, and we have many problems like our scientists will run away to
another country. (P: 1: 39)
In Excerpt (6.5), S6 does not agree with S8 but is trying to indirectly say
so by using we to involve the interlocutor in the negative reply. This
also indicates a listener-oriented approach by the Persian speakers (see
Section 6.9 for further discussion). S8 starts turn 39 with and and huh,
two hesitation markers signalling the time needed to find the most suit-
able indirect strategy to express disagreement.
The clusters of may not and maybe + not demonstrate discrepancies
between the L1 and L2 speaker groups. While may not and might not are
found to be used more frequently by the L1s, the Persian speakers show
an inclination to use maybe + not. As explained in Section 6.5, which
deals with the influences of a first language on a second, this trend
seems to have emerged from the closeness of maybe to the structure
mainly used to express possibility in the Persian language: that is, maybe
or maybe + negative are structures which the Persian speakers are accus-
tomed to using in their L1 language and the similarity of the structures
between the two languages leads the Persian speakers to prefer maybe
not to may not and might not.
The most different patterns are observed in Table 4:14, the distribu-
tion of placeholders. Two major discrepancies are the large overall
number of uses by the Persian speakers, and the inconsistency in
using clusters of placeholders across the three groups. For instance,
something clusters occur in the Persian data 36 times, in the Chinese
eight times, but never in the L1s data. The clusters with anything show
only one type shared among the three groups. The Persian data pro-
vides the highest number of anything clusters, the Chinese the least.
There are six types of anything clusters in the Persian data, with a total
of 32 occurrences, while the 14 occurrences in the L1 data are of four
types, only two of which are in common with the Persian speakers.
The Chinese data reveal only two types of anything clusters over seven
occurrences.
There are two reasons for using placeholders: they appear to help
speakers cope with no need or inability to speak specifically. They have
the potential to be used almost like a discourse marker to enhance com-
munication. Placeholders are used by L1s when there is no need to speak
specifically, or to compensate for the lack of a word or the inability to
remember the appropriate word; as the Chinese and Persian speakers
were not as proficient in English as the L1s, they encountered such

situations more often and consequently made more use of placeholders.
turns. They are talking about how the younger generation tries to deal
with the older generation.
S5: You are talking about parents who are not young enough to know
about computers. When there is a … there … (P: 6:877)
S2: My parents don’t. My parents don’t can’t work with computer,
too, either. Huh, they don’t know lots of things. They didn’t.
They know a lot of things but about these kinds of things that are
I don’t know common in nowadays. They, most of them time they
don’t agree with that with these, ok? But I try to have a relation-
ship with them, that it solves our problems, ok? For example, I do
my works, I do my, huh, I got my ways, ok? And they just, huh,
watch me send, huh, I don’t know how I can say. (P: 6:878)
S1: This is their way. (P: 6:879)
Excerpt (6.6) illustrates a situation in which the Persian speakers use a

placeholder things to stay general and compensate for a situation where
a specific list of things is not important, or where a limited vocabulary
may not permit S2 to be more specific.
The more frequent use of clusters containing placeholders may occur
if placeholders are more commonly used in the L2s’ first language than
in English. The Persian speakers tend to use placeholders at least in the
academic context of this study, and the data show that the Chinese and
L1 speakers use placeholders less often than other categories of vague
expression.
As discussed in Section 5.1.2, the claim that may and might used by
the Persian speakers and might by the Chinese are to express politeness
is to some extent confirmed when the pronouns cluster with modal
auxiliary verbs. In the Persian data, you may is the most frequent cluster
of subject pronoun and auxiliary verb, followed by we may. Excerpt (6.7)
shows that the Persian speakers use you may for indirectness, in giving
a hedged warning.
turns. They are talking about problems in society.
S5: For example? (P: 6:176)

S2: Ok, your school is a kind of society. For this kind of cheating you
may have problems for your future. (P: 6:177)
S3: Discrimination. (P: 6:178)
Culturally, Persian speakers are accustomed to expressing disagreement

or criticism indirectly through softened statements which include
the party who is being referred to (Behnam & Niroomand 2011).
Disagreement or criticism is implied rather than explicitly stated. In
Excerpt (6.7), S5 seems not to have been convinced by S2’s statement
and in turn 176 asks for a specific example. S2 in turn 177 informs S3
of the potential consequences of a possible situation, using you may to
tentatively and indirectly inform the listener of potential problems and
veiling his disapproval of what is happening. S3 in turn 178 confirms
his understanding of S2’s disapproval of the situation and provides a
tangible problem, ‘Discrimination’.
(6.8): This is a turn by a Chinese participant. She is talking about how
to be a good teacher.
S4: It is very useful for his or her future, for his own development.
I think that is what a teacher should do and we can also make
friends with our students that are quite interesting. And to be a
teacher is not just about teaching, just about giving, just about
giving the same lessons. If we do just this, that might be a little
bit boring. Yeah, that’s what I think. (Ch: 7: 57)
In Excerpt (6.8), the speaker promotes a good way of teaching, at

the same time pointing out a boring way of teaching. S4 does not
want to appear blunt in criticizing the boring approach, so she adds
might, coupled with a little bit, to hedge the statement and appear less
offensive.
6.3 Density of vague language and

concentrated distribution
The occurrence ratio between a vague expression and the total word
count in the data, namely the density of vague language, may present
another perspective as to how these expressions occur differently across
the three groups of participant.
As Table 6.2 shows, on average, there is one subjectivizer per 250
words in the L1 data, whereas this vague expression occurs more fre-
quently, once per 182 words, in the Persian data. The Chinese data
show a more concentrated use at once per 69 words. Possibility indica-
tors occur once per 270 words in the Persian data, fewer than in the L1
once per 216 words. As with subjectivizers, the number of uses in the
Chinese data is the greatest, at one per 135 words.
Table 6.2 Ratio of vague expressions to total word count
Item L1 CSLE PSLE

N = (51,403) N = (51,263) N = (51,344)
Subjectivizers 250 69 182

Possibility indicators 216 135 270
Vague quantifiers 122 69 118
Vague intensifiers 129 58 154
Placeholders 171 179 107
Vague quantifiers occur once per 122 words in the L1 data, slightly
less often than in the Persian at once per 118. The Chinese data again
demonstrate high use, once per 69 words (the same density as subjectiv-
izers). Vague intensifiers appear most often in the Chinese data, once
per 58 words, and least in the Persian at once per 154. L1s used vague
intensifiers once per 129 words. The Chinese used fewest placeholders
at once per 179 words, followed by the L1s at once per 171 and the
Persian speakers at once per 107. This density analysis acts as a supple-
ment to the frequency discussion presented so far.
The Chinese speakers used 741 vague words acting as subjectivizers,
732 of them I think: 99 per cent of their total use of subjectivizers. Both
the Persian and L1 speakers used this subjectivizer item the most as
well, but at slightly lower concentrations (73 per cent and 79 per cent
respectively).
I think may be heavily used by the Chinese because it is being
employed as a discourse marker, appearing where a discourse marker
might seem more appropriate. This trend supports the pattern that
emerged in Wu et al.’s (2010) study, in which Chinese speakers were
found to use I think as a filler. Such usage may meet a speaker’s need for
a filler, achieved by stretching a piece of vague language (Zhang 2011).
Other reasons for the popularity of I think may lie in the need to elasti-
cize vague language, as a turn-taking device, or as a cognitive processing
focus (see sections 6.7 and 6.9 for details). Like subjectivizers, vague
intensifiers have the most concentrated distributions. Very makes up
more than half of the vague words in this category in the Chinese data,
with the remaining use spreading over five other items; the L1 group
concentrates mostly on really.
The most even distribution of possibility indicators is observed
among the L1s, with the first three items in Table 4.3, maybe, may and
might, constituting three quarters of this vague category evenly; the

remaining quarter is composed of the final items, probably and possible.
The Persian and Chinese data predominantly reveal the two first items,
with the other four items all used to a lesser extent. The L2 speakers
used this category more than the L1s.
Vague quantifiers and placeholders present relatively less concen-
trated distribution across all three groups. The data show that the L1s
tend to use vague categories at a lower density, while the Chinese show
concentrated distribution of several categories. This stands out in possi-
bility indicators, where L1s show a relatively even distribution of maybe
(26 per cent), may (24 per cent), might (24 per cent), probably (18 per
cent), and possible (8 per cent). The concentrated distribution shows
82 per cent, 13 per cent, 3 per cent, 1 per cent and 1 per cent for the
Chinese and 88 per cent, 8 per cent, 7 per cent, 1 per cent, and 3 per
cent by the Persian speakers. It seems that the most frequent items in a
category, like I think for subjectivizers and maybe for possibility indica-
tors, are used in more versatile ways by L2 speakers.
6.4 Influence of first language
First language can be a two-edged sword, sometimes impeding L2

learning and sometimes facilitating it. Items which are different in two
languages are claimed to be difficult to learn, while similar items are
more easily learned (Lado 1957).
Individuals tend to transfer the forms and meanings, and the distri-
bution of forms and meanings, of their native language and culture
to a foreign language and culture—both productively when attempt-
ing to speak the language and to act in the culture, and receptively
when attempting to grasp and understand the language and the
culture as practised by natives (Lado, p. 2).
This can be seen in the use of possibility indicators by the two L2 groups
in this study. The Persian speakers seem to have been influenced by
the transfer of maybe from Persian, leading to less frequent use of other
similar items, may and might. These generally appear as expressions
indicating politeness rather than possibility, imitating the function of
the equivalent words in Persian. The Chinese, by contrast, use may at
a frequency close to that of the L1s, because in Chinese yexu (may) can
indicate possibility as well.
May and might are likely to be underemphasized in terms of their dif-

fering functions in L2 instructional materials, and students may have
learned it to recognition level but be unable to use it at production
level. It seems that due to the approximation of the meanings in their
first language, the Persian speakers have opted to express possibility
through maybe. Other modal verbs such as can and must do not have
any other equivalents in Persian and are used more often by the Persian
speakers than by the L1s. May and might are not entirely interchange-
able in English: the former involves stronger possibility than the latter.
In Persian, there is no such difference, the two auxiliary verbs are the
same, and maybe is the term to which almost all possibility roles are
assigned, despite it being a different part of speech.
turns. They are talking about what factors affect the way parents treat
their children.
S4: Different people, different culture, different views. (P: 6:830)

S1: Yes, it influenced you. (P: 6:831)
S6: I don’t know to some other works. Because, due to that, they may,
might not understand us that way. I am not a, that, I don’t know
that shy, or, guy. I did lots of effects and lots of things but when,
the only thing you seen, the garden is beautiful nature, is cows,
sheep, dolls and just taking care of your children. (P: 6: 832)
As Excerpt (6.9) shows, S6 reveals uncertainty as to which modal auxil-

iary to choose to express possibility. He first chooses may but switches
to might, which can be interpreted to have occurred as a result of the
lack of clear distinction between these words in the Persian language, or
uncertainty about the precise meanings of the words in English.
The occurrence of might in the Chinese data shows a similar pattern,
but may appears at a frequency similar to that of the L1 group. Chinese
use might less frequently, perhaps because in Chinese there is an equiva-
lent for may, but not for might: not being familiar with the word in their
first language may cause Chinese speakers to avoid its use, and to per-
ceive a difference between the two auxiliary verbs due to the influence
of their first language.
The findings in the Persian data are in line with Atai and Sadr’s (2008)
findings that the Persian native speaker uses may less frequently than
L1s. The distinction between may and might seems to carry more gram-
matical functions than lexical weighting for learners of English. Römer
(2004) reports that the two words fall in the lower rank in the modal
auxiliary list by the L1s, might sixth and may eighth as the penultimate
item (indicating that even L1 English speakers do not make much use of
these terms or English L1 speakers can’t get the terms right), while can,
should and must occupy higher ranks.
Maybe is the only possibility indicator the L2 groups use much more
than the L1s. It occurs in clause-initial position 85 times, or to 55 per
cent of total possibility indicators for the Persian speakers, 128 times
and 41 per cent for the Chinese, and 11 times and to 17 per cent for
the L1 group. The popularity of this item may lie in its versatility, as it
has the potential to be used in different positions in a clause. Both the
Persian speakers and the Chinese use around half of the overall tokens
of maybe in their clause-initial position: the Persian speakers use possi-
bility indicator items at the beginning of clauses in their L1 and maybe
seems as the most compatible with this pattern. As it is quite common
to use an equivalent of maybe at the beginning of the clause in Persian,
this tendency is also transferred to English. The situation is similar in
Chinese, where maybe can be used in clause-initial position, especially
in spoken exchanges. The tendency to use maybe in English structures
that resemble L1 structures, and the availability of this possibility
marker in both L1 and L2, facilitate the transfer of this item from one
language to another and accounts for the heavy presence of maybe and
the related low appearance of other possibility markers in the L2 data.
(6.10) These are two turns by an L1 describing a photo.
S1: … but they’re they are from the University’s uh Linguistics

Department and this is Janine this is Bonnie and someone will
probably wanna say something about it. (L1: 3:3)
S1: That Photoshop book is bothering me because I’m thinking that
I’m in I’m in six-forty. (L1: 3:4)
As can be seen in Excerpt (6.10), the L1 uses probably in clause-medial

position to express possibility.
(6.11) This is a discussion between three Persian speakers over three
turns. They are talking about modern life in general.
S1: I don’t know. (P: 4:55)

S4: A general idea. Oh, yeah. It is a general idea. But it is true. Maybe
some people have goals and they are searching for it and, huh.
(P: 4: 56)
S7: The attention that they are giving to it. (P: 4: 57)
The Persian speakers in Excerpt (6.11) below, in a similar context, prefer

to use maybe in clause-initial position. L1 linguistic influences on L2 use
emerge in subjectivizers as well. The comparison of I think that among
the three groups indicates that the tendency to use that after I think by
the Persian speakers is different from the other two groups’. The L1s
present 17 occurrences (11 per cent) and the Chinese 16 (1 per cent),
indicating a preference for that acting as the subject (pronoun) of the
sentence. The Persian speakers use the cluster only twice (1 per cent).
The minimal use of subject–serving that in I think that by the Persian
speakers may be because they use this or it interchangeably in the
same position in their first language, while that is mainly used as a
demonstrative pronoun. As there is no such pronoun as it in Persian,
there seems to exist no distinct difference between it and this, and
these two words are used interchangeably. Their data show a relatively
frequent use of I think it is (14 occurrences, versus nil in the L1) over
I think that is. It seems I think that is is replaced by I think it is by the
Persian speakers.
The opposite occurs when that in I think that switches to a conjunc-
tion (complementizer). The pattern is quite common in Persian and
makes its way into their spoken English, shown in Excerpt (6.12)
below.
(6.12) This is a discussion between two Persian speakers over three
turns. They are talking about the problems in society.
S4: People think they should be very brief in everything they, for
example, when they are in a line, they try to go earlier, to for
example buy something or when they are in taxi lines, they try to
go to the taxi sooner than the others. (P: 6:149)
S2: Ok. (P: 6:150)
S4: I think that it is wrong. (P: 6:151)
In Excerpt (6.12), that is a conjunction rather than a subject. The

conjunction usage of that occurs 18 times (9 per cent) in the Persian
speakers data, three times (2 per cent) in L1 data, and seven times
(1 per cent) in the Chinese data. The Persian speakers tend to use this
function more than the Chinese or L1 speakers.
L1 influence is manifested in the use of quantifiers by the Persian
speakers as well. The trend is that, apart from a lot of, the Persian speak-
ers consistently use more nouns after quantifiers than the Chinese or
L1 speakers: 160 occurrences after some versus 74 by the L1s and 144
by the Chinese, and 25 occurrences after lots of versus 12 by the L1s
and five by the Chinese. The preference for countable nouns among
the Persian speakers appears in this study in the form of 11 times of
some + countable noun; the other groups do not have examples of this
pattern.
The low frequency of lots of (11) and high frequency of a lot of (85)
in the Chinese data may be because the Chinese language is insensi-
tive to the distinction between countable and mass nouns. The Persian
speakers use there are lots of five times, while the L1 data does not show
any such combination and the Chinese use it only once. By contrast,
the L1s use there is lots of five times, but this form is non-existent in
the Persian data. The proportion of countable nouns to mass after lots
of reveals that the L1s and the Chinese use them indiscriminately,
whereas the Persian speakers prefer countable nouns to mass nouns by
three to one. This difference may originate from the system of plural
and singular nouns in Persian: a study by Sharifian and Lotfi (2003) on
mass–count distinction finds that Persian speakers’ concept of mass–
count nouns allows them to use some mass nouns as count nouns
(plural) in certain contexts, and that the mass–count distinction across
different languages arises from ‘underlying discrepancies in conceptu-
alizing experience that is being coded in linguistic expression’ (p. 229).
What this implies is that the English, Chinese and Persian mass–count
systems may be different, and that Persian speakers may have devel-
oped a more flexible count nouns system that they prefer to use over
mass nouns. This may explain why the Persian speakers in this study
preferred to use count nouns with quantifiers that otherwise apply to
both mass and count nouns. Due to such discrepancies in mass–count
distinctions, quantifiers are sometimes used quite differently between
L1 and L2 speakers of English.
L1 influence manifests in the use of vague intensifiers as well. The
occurrence of so and too across the three groups of participants displays
two overall trends: the first is that there is consistently heavy use of
these items by the Chinese, but a different pattern is revealed between
the Persian and L1 speakers. The Persian speakers, with 15 uses, were the
least frequent users of too, which the L1s, with 24 tokens, used nearly
twice as often. However, the Persian speakers presented 75 occurrences
of so, while the L1s’ 40 uses made them the least common users. The L1s
seem to resort to other intensifiers to compensate for the low frequency
of so in their data.
turns. They are talking about discrimination in their society.
S4: In important position in government, it is very important. Race,

your religion is so important. (P: 6:188)
S6: I mean for example Aghazadeh.1 This kind of discrimination.
(P: 6:189)
As can be seen in Excerpt (6.13), important clusters with so, a favourite

intensifier among the Persian speakers. The L1s show lexical diversity
by using extremely before important.
(6.14): This is a turn by an L1 discussing a case study in a research
report.
S1: So it’s an exception to the rule. Case study thirty-five photo digi-
tal cover-up very important. This issue is extremely important.
I’m sorry they don’t have more case studies, like this because …
(L1: 1:685)
In Excerpt (6.14), extremely expresses a strong degree. There are other

such adverbs that occur in the L1 data which the Persian speakers either
do not use or use very infrequently: highly, definitely, and extraordinar-
ily. The Persian speakers may demonstrate less lexical density because
of the limited diversity of intensifiers in their L1 language or lack of
knowledge in their L2, giving rise to their heavy use of so compared
with the L1 group.
Both L2 groups employ so five times more often than too, and this
proportion is twice as much as the L1s. One reason can be the lack of
an equivalent for too in the Chinese and Persian languages. Too does
not exist in Persian and the features (attached to it) are not covered by
any other concepts. Its features may be difficult to understand, such as
the negative result which it implies and the particular structures that it
takes (too + adjective + infinitive). For example, ‘The box is too heavy for
him to lift’, which means it is impossible for him to lift the box. These
may make it alien and unwieldy to Persian speakers, who prefer so and
very which are easily transferred from their first language. It appears
that with some modifications to the structure to make it grammati-
cally sound, the Persian speakers used so and very to express what too
expresses in the L1 data.
turns. They are talking about the room they are in.
1
Persian word meaning ‘someone from a privileged family’.
S2: Doesn’t it work in this room? (L1: 3:2)

S1: No it doesn’t it, the room’s just too small. So, that’s, number one,
okay so I like it better like this so if you guys wanna, …. (L1: 3:3)
In Excerpt (6.15), the L1s use too small to emphasize the smallness of the
room. In Excerpt (6.16) below, S5 uses so busy to confirm S2’s statement,
which employs too busy. The frequency of so + adjective in the Persian
data brings up the number of occurrences of so.
turns. They are talking about how stressful study is for a student who
is also a parent.
S2: You are too busy. (P: 6:552)

S5: Yes, I am so busy and have to. In spite all that, my parents were
really worried about me going to the university because they said
you are very busy. You’re always working. You’ve got three kids to
look after. (P: 6:553)
They are talking about discrimination in their society.
S4: In important position in government, it is very important. Race,

your religion is so important. (P: 6:188)
S6: I mean for example Aghazadeh [Persian word: ‘someone from a
privileged family’, noted by the author]. This kind of discrimination.
(P: 6:189)
In Excerpt (6.17), S4 uses so, it is also possible to see very as another

instance where the choice of a Persian speaker is different from that
of where too can be employed by an L1 speaker of English. This can
be confirmed by the proportion of too to very between the two groups:
while the Persian speakers used very over too by seven to one, the L1s
used it only one to three. The same pattern is witnessed in the Chinese
interactions with an even larger proportion: 14 to one, but this seems
to have occurred as a result of linguistic inability to speak at the same
level as L1s of English (see Section 6.6 for details).
turns. They are talking about the hardship of using a search engine.

S1: Where I wound up with, zero or I wound up with, thirty thousand,
you know sometimes, when you’re online it’s just too hard to
assess that. (L1: 3:194)
In Excerpt (6.18), S1 uses too to intensify the hardship of assessing some-

thing online. In (6.19) below, in a similar context, the Persian speaker
opts for very:
turns. They are talking about how culture may change.
S2: Yes? And how can we improve our culture? For instance practic-
ing? What? What should we do? Training? (P: 6:434)
S5: But you know I think our culture is very hard to change because
one of … (P: 6: 435)
S3: No, very easy to change. (P: 6:436)
Excerpts (6.18) and (6.19) demonstrate that the Persian speakers tended
to use very rather than too to intensify some adjectives, as is the case
for the Chinese as well. The Chinese use very heavily as many as 498
occurrences, and the Persian speakers use 108, the least user is the L1
group with 79 occurrences. It is possible that many of very in L2 data
could be replaced by too.
The influence of a first language contributes to discrepancies in the
choice of vague intensifiers when speaking in English. Too does not
exist in Persian but so and very do, and the Persian speakers found
them more comfortable to use where L1s used too. The same applies
to the use of quantifiers, explaining why the Persian speakers used
neither a few nor few. The influence of L1 can sometimes result in
avoiding, or using only rarely, other items: for Persian speakers, these
include such concepts as countable or mass nouns, which are not
distinguished in Persian. It may not be appropriate to conclude that
Persian speakers are unable to use them appropriately, but fairer to
argue that learners of English have not yet reached a level at which
mastery of such cross-language differences in the use of vague items
has been achieved.
6.5 Influence of cultural protocols
Cultural protocols emerge in the way the speaker uses vague language in
interaction and can reveal particular ways in which speakers vary their
use of vague language (Terraschke & Holmes 2007). One such mani-
festation of cultural protocols can be identified in the pattern I think
that with that functioning as a complementizer, a phrase more widely
used by the Persian speakers than the Chinese or L1s. Persian speakers
prefer not to express a proposition directly, especially when there is
uncertainty about it (Abdollahzadeh 2003). That reinforces the speaker’s
doubt, expressed by I think, or indicates that the speaker is going to
express disagreement or contrast and uses that as a somewhat indirect

term that helps to avoid being offensive.
Sharifian and Lotfi (2003) claim that ‘language structure is largely
governed by the ways in which humans conceptualize their experience,
which may be formed or informed by culture’ (p. 241). Differences in
mass–count distinctions in languages may be one result of a cultural
concept influencing the structure of a language. Culturally, Persian
speakers tend to be conservative when expressing disagreement or con-
trast, and especially in academic settings express themselves hesitantly.
This is evidenced when, as in Excerpts (6.20) and (6.21) below, cases of
I think that are preceded by but, huh or you know, discourse markers for
hesitation or a search for words.
turns. They are talking about how people can learn to respect each
other’s privacy.
S2: Also, by fathers and mothers. For example, if in the school or

kindergarten the instructor sees the children that do the opposite
things and you must pay attention and say for him or her fathers
or mothers. (P: 6:324)
S8: But I think that we as we are a traditional country, we can’t change.
Ok. Two or three centuries later, maybe this happens. (P: 6:325)
They are talking about the issue of spirit.
S6: Actually, huh, it is more than here we can say. You don’t, you
say that it is not but I think that it is. My question is that are all
these common with human structure? I mean structure of spirit.
Huh? … (P: 4:110)
S3: Is it good for us? (P: 4:111)
S8 in Excerpt (6.20) and S6 in Excerpt (6.21) oppose what was stated

previously, using I think that introduced by the discourse marker but to
help soften the tone of disagreement. The complementizer that after
I think is quite frequent in the Persian data, while the L1s and Chinese
employ that as a subject after I think more often.
Cultural conceptualizations embedded in a language can be trans-
ferred to second language learning (Sharifian 2003), and in the case
of the Persian speakers, this seems to include vague language use. The
three main salient cultural schemata in the Persian language are aberou,
taarof, and shekaste-nafsi (Sharifian 2007), all closely associated with

politeness. Aberou has something to do with ‘face’:
This schema is manifested in the communicative behaviour of many

Iranian people, partly through repeated attempts to refuse offers and
invitations, hesitation in asking for services and favours, hesitation
in rejecting requests, and so on. Another reflection of tarof is the use
of plenty of hedges.
(Sharifian 2007, p. 39)
O’Shea defines taarof as ‘a formalized politeness that involves verbal

and nonverbal forms and clues [around which] Iranian society revolves’
(2000, p. 122). Koutlaki (2002, p. 1740) defines it as ‘mutual recogni-
tion’, functioning as ‘a tool for negotiating interactants’ relationships’.
This may explain why the Persian speakers in this study use more vague
expressions than the L1s, as their cultural schema requires more such
words. Shekaste-nafsi connotes ‘modesty’, defined as ‘broken-self’ or
‘breaking of the self’ (Sharifian 2005; 2007).
It encourages speakers of Persian to show modesty through the denial
or downplay of any praise or compliment that they receive, reassigning
it to the praiser, to family members, to God or simply to luck. (Sharifian
2007, pp. 41–42)
Shirinbakhsh and Eslami Rasekh (2013) provide extracts which
include several instances of shekaste-nafsi. Extracts (6.22) and (6.23) are
examples from their work, whose subtitle is ‘The case of Shekaste-nafsi
(modesty) in Persian’.
(6.22)
W1: This is a very beautiful dress. Did you sew it yourself?
W2: Yes.
W1: Well done. What an artist.
W2: It’s not as skilful as your sewing.
W1: Thanks, but it’s not true. You are a professional who has surpassed
me. I have become old.
W2: You are welcome. I take my hat off to you. (pp. 100–101)
(6.23)
W1: Your hair is very nice.
W2: It is by chance, this time it became like this.
W1: No, you’re beautiful so anything suits you.
W2: Beauty comes from your eyes. (pp. 102–103)
The manifestation of shekaste-nafsi occurs mainly through intensifiers

(especially very illustrated in Excerpt (6.22/6.23) and actually illustrated
in (6.24)) by which the speaker tries to emphasize a feature, and also
through placeholders that generalize. This might be one reason for the
frequent use of actually and anything in the Persian data:
(6.24): This is a discussion between two Persian speakers over four
turns. They are comparing the older generation with the younger
generation.
S1: She doesn’t know anything how to turn, how to switch the com-
puter on. (P: 7:213)
S2: I am worse than your mom, dear. Because before I bought a com-
puter for my children. First of all, I went to computer classes.
You’ve got that. (P: 7:214)
S1: You know that, you know what you should do but what can my
mother do about it. You went to university, you know that, but my
mother is, has left school. My mother had a child when she was
17. (P: 6:215)
S2: And I had it, actually, when I was 18. (P: 7:216)
In Excerpt (6.24), S1 in turn 213 is pointing out that his mother is

not educated and by anything is emphasizing the fact that she even
does not know the basics of using a computer. To demonstrate that
she is no better, S2 downgrades herself by using ‘worse’ in turn 214,
although it is clear that she has taken courses and knows how to
use a computer. In the next turn, S1 tries to convince S2 that she is
in a better situation than his mother. First he asserts that she knows
enough about computers and then refers to her tertiary education,
which his mother lacks. He then adds that his mother had a child at
a very young age, which S1 counts as a demerit. S2 in turn 216 again
downgrades herself by referring to the fact that she had a child at
a young age as well. This is emphasized by actually. The purpose of
downgrading by S2 is to show respect to S1’s mother and indicate that
she should not be criticized because of her situation. If this were a
conversation between two English speakers, S2 may be playing the
one-upmanship game, pretending to be worse but giving enough
information to prove that she is better: she too had a child, but she
still went to university and computer courses, so she knows these
things and S1’s mother does not.
The overwhelmingly high frequency of vague expressions in the
Chinese interactions suggests that cultural norms are once more influ-
encing vague language use. Zarei and Mansoori (2007) claim that:
while English academic discourse relies on the writer’s responsibil-

ity to provide appropriate transition statements for the reader’s
convenient tracking of the writer’s logic, some other cultures such
as Japanese, Korean and Chinese display an opposite trend, giving
over much of the responsibility to the reader to grasp the writer’s
intention (p. 26).
Their claim is supported by Chan (2013), who argues that Chinese tend
to use indirect and circular styles when they interact, as these allow
them to avoid direct confrontation. They use various strategies such
as contrary-to-face-value (CTFV), or use vague expressions. Ma (1996,
p. 258) defines CTFV as ‘any communication in which what is said is
the opposite of, or different from, what the speaker believes to be true
or what he or she is “logically” expected to say’. This phenomenon can
be realized in ‘yes’ for ‘no’, or vice versa. Another strategy to achieve the
same goal is to remain vague and avoid a direct statement. This norm
in Chinese politeness practice seems to encourage vague language use
in their second-language patterns.
6.6 Impact of lack of language proficiency
A number of vague language patterns in this study may be attributed

to the lack of proficiency on the part of L2 speakers. Overall, the L2
groups used vague language in a more concentrated manner than
the L1s because they did not know as many synonyms as the L1s or
because some terms and concepts did not exist in their first language.
For example, of the four subjectivizers considered in this study, the
Chinese used I think almost exclusively (99 per cent), and of the five
possibility indicators, the Persian speakers used maybe 81 per cent of
the time. Both used very extensively, possibly a function of their lim-
ited English vocabulary. Lack of words, or lack of familiarity with the
ideas and schemata behind some words leads to a tendency to carry
out intensifying tasks with very few vague words.
So is another case in point. The data offer 154 tokens of so by the
Chinese, 75 by the Persian and 40 by the L1 speakers. The low L1 figure
is a consequence of their being able to employ many other intensifiers,
saying, for instance, ‘he was highly pleased to hear the news’ where
Persian and Chinese speakers would most likely use so. The L2 speakers
are limited by the lexical diversity of intensifiers in their L1 language
and also by their ability to use so with all kinds of adjectives, negative
and positive, needing little else. Both groups would benefit from devel-
oping greater lexical diversity in English.
Other discrepancies can be observed: the data provide 15 instances

of the Persian speakers clustering the adjective important with very but
only four of the L1s, who also present four tokens of really important but
the Persian speakers none. There are 17 instances of the Chinese using
very interesting and three of the L1s, who present really interesting nine
times against once by the Chinese. Both L2 speaker groups’ preference
for very over really may be because they are not familiar with the second
word and are unable to integrate it into their English speech patterns.
Insufficient active vocabulary leads to heavy reliance on a few common
items which can serve several purposes, as can be seen in Excerpts (6.25)
and (6.26) below.
(6.25): This is a discussion between two Chinese participants over
three turns. They are talking about Beijing weather in summer.
S4: Have you been there? (Ch: 4:72)

S1: Yes, <LAUGH>. It is very, very, very hot. (Ch: 4:73)
S4: And (xx) (Ch: 4:74)
(6.26): This is a discussion between three L1 participants over four

turns. They are talking about the fact that some people no longer think
about the environment.
S12: Somehow (xx) (L1: 1:319)

S6: When it’s too hot to go to work then you’ll be, <S1: LAUGH>[S12:
yeah] complaining all right? (L1: 1:320)
SU-f: I think it’s important (L1: 1:321)
S12: And you can’t have air conditioning cuz the ozone is totally,
screwed so. (L1: 1:322)
In Excerpt (6.25), the Chinese speaker uses three very in a row to indi-
cate extreme heat; it is possible that the speaker is not familiar with the
alternatives too or really. This is observed of the Persian speakers as well.
Excerpt (6.26) shows that in a context similar to Excerpt (6.25), the L1s
uses too instead.
The L2 speakers in this study made a wide use of different vague
words, but sometimes failed to make appropriate use of vague lan-
guage because of their lack of expertise in English. For example, both
the Persian and Chinese speakers showed that they had not yet acquired
all the applications of the placeholder anything in English, mainly limit-
ing its use to negative sentences. The L1s used it more dynamically, in
negative, interrogative and affirmative phrases. Its elasticity has not yet
been appreciated by the L2 speakers, mainly because of their limited

knowledge of English.
6.7 Impact of cognitive processing focus
As reviewed in Section 2.4.2, one of the main pillars of relevance theory

is the cognitive processing effort needed to achieve maximum effect.
A key concept is ‘cognitive load’, which Sweller (1988) defines as the
total amount of mental activity imposed on working memory at an
instance of time. The effect of cognitive load can emerge in the speaker’s
preference for particular words or phrases in communication.
Vague language can be a tool to meet a speaker’s need to fill gaps
created as a result of concentrating on cognitive processing, as can be
seen in the use of I think + negative across the three groups. As found in
Zhang and Sabet (in press), I think + negative sentence appears predomi-
nantly in the Chinese and Persian data, and I don’t think + affirmative in
the L1 and Chinese; it appears in the Persian data only rarely, and only
in particular contexts. All I don’t think combinations turn up in formu-
laic expressions, as shown in Excerpts (6.27) and (6.28) below.
turns. They are teasing one of the students who claims she once ate
eight apples in a row.
S1: It was one kilo I think. Yes? (P: 5:616)

S5: All of them? No, I don’t think so. (P: 5:617)
S3: One kilo. (P: 5:618)
They are talking about the fact that computer games can play signifi-
cant roles in children’s futures.
S1: … I believe in them, Ok? I don’t think they are. I don’t know they
should be deleted. (P: 6:516)
S5: Yes, of course not. Not deleted. (P: 6:517)
In Excerpts (6.27) and (6.28), I don’t think so and I don’t think they are
are formulaic, explicitly taught in ELT books and used as the result of
explicit instructions.
The data indicate that the Persian speakers prefer I think + negative
statement. I don’t think seems to be a more processed result of the utter-
ance and demands more cognitive load, thus the Persian speakers use
it infrequently. This phenomenon is discussed as negative raising in

which ‘the negation of a complex sentence with a subordinate clause is
read as negation of the subordinate clause’ (Chow 2010, p. 1). It hap-
pens as a result of the application of a transformational rule called ‘Neg-
raising’, which moves the negation from the subordinate clause to the
main clause. The negation raising process requires more cognitive load,
especially in speaking. As I don’t think occurs as a result of the applica-
tion of the transformation, Persian speakers choose to stay clear from it
due to their inadequate linguistic proficiency.
The other impact of cognitive processing is revealed in the occurrence
of I think in clause-final position, where a difference between L2 and L1
speakers is displayed. The Chinese data offer 21 tokens and Persian 16,
but the L1 only two. It seems the Persian speakers use I think at the end
of the sentence as a device to help relieve memory load, giving oppor-
tunity to seek time to think about the next segment.
turns. They are talking about how their country can improve.
S3: Ok. I think, huh, for developing, all the things return to the per-
sonal culture I think. You know if you want to develop, huh, all
the people one by one should start from themselves and (huh)
and to have the culture to be, for being a developed country and,
huh, find the capacity of it. You know I think there are lots of
problems in, in even this class, in this small society that we have.
For instance, he says that <LAUGH>. (P: 1: 270)
S7: Giving ideas. (P: 1:271)
As Excerpt (6.29) shows, S3 in turn 270 uses the subjectivizer I think in

sentence-initial position and uses it again in the end. The clause-final
I think may be a discourse filler to provide space to think and the need
for preparation to make the transition from one sentence to another.
This interpretation is reinforced by the discourse marker you know,
immediately following, because it adds more time.
The other pattern associated with the impact of cognitive processing
effect is shown in the frequency of I think that where that functions as
a complementizer. This pattern occurs more frequently in the Persian
speakers data, showing that the Persian speakers use the complemen-
tizer that, to fill the pause in speaking, thereby getting the chance to
think of the word needed next. What it implies is that the complemen-
tizer that behaves like a discourse marker for the learners of English
when it is used in the right position. That is, to facilitate cognitive
processing in speech, the Persian speakers use that after I think, which
provides them with a momentary pause to unload memory constraints,

as shown in Excerpt (6.30) below.
turns. They are talking about possible problems in the election system
in their country.
S2: Let’s say, let’s say that in the process we cheat, ok? Who is cheating
again? We are cheating us, people again. (P: 1: 112)
S4: But I think that people that are cheating, they should be separated
from people that are living and are, huh. (P: 1: 113)
S2: Maybe that’s the biggest problem in Iran. Maybe that is one of the
biggest problems. That crime does pay in Iran because there is no
way of (xx) people. Ok. (P: 1: 114)
The Chinese go beyond this and use I think twice or even three times, as
can be seen in the examples below, to handle memory load.
turns. They are talking about what they would like to be in the future.
S2: I want to be a teacher. I, I like children. Yeah. They are lovely.

(Ch: 3: 135)
S4: I think, I think the children will like you at the same time. They
are lovely, too. (Ch: 3: 136)

turns. They are talking about what they prefer to be in the future.
S6: … There exist too much difference between Chinese and, huh,
Western culture. For some more, I think, huh, I think we can, we
can, huh, we can enjoy the nature, huh, relax ourselves. And we,
as we know, tourist can earn more money and, huh, strengthen,
strengthen our body. Thank you so much. (Ch: 6: 13)
S1: But many just said that it is just too funny to be real. Yeah, I think,
I think, huh I think it is right to some degree because every career
needs much, much effort and, and, and we have limited energy as
well as limited time. So if I have to choose one as my career, I still
want to be a good learner because linguistics is all, is really, really,
very useful. (Ch: 6: 14)
In Excerpt (6.31), S4 in turn 136 repeats I think twice. In Excerpt (6.32),

S6 uses two I think while thinking how to phrase the next bit. S1 in the
next turn takes up S6’s pattern, going a bit further repeating I think three
times. In both cases the speakers seem to use the repetition of I think to
buy time to think about their next move; the pause marker huh buys
them even more time. There are more such repetitions of I think in the
Chinese data than in the Persian or L1.
Cognitive processing can lead to the rise of the frequency of a vague
word like maybe by the Chinese. Like I think, the Chinese use maybe as
a discourse management device to search for words.
(6.33): This is a discussion between three Chinese participants over
three turns. They are talking about job security in China.
S2: It can be connected to the nation’s interest. Our country, our

government will, huh, have some special benefits to us. That’s
why. (Ch: 7: 137)
S1: Maybe, maybe another reason is, huh, our people, our society
provide, huh, higher principle of our teachers and want people in
various (xx) to teachers and the, huh, yes. This is why it is safer.
(Ch: 7: 138)
S6: I think is is very different to, huh, foreign countries because, huh,
from our English, our teachers think that, huh, teachers in foreign
country their jobs are very. (Ch: 7: 139)
In Excerpt (6.33), S1 in turn 138 uses maybe twice as a turn-taking

device as well as to handle memory load. Like I think, maybe is used
by the Chinese in two ways, the first basically as a turn-taking device
and the second associated with cognitive processing effect. This sug-
gests that the insufficient use of discourse markers is to some extent
compensated for by the heavy use of maybe and similar phrases.
6.8 Different communicative approaches
The pattern I don’t think + positive statement reveals the speaker-

oriented approach by the L1 speakers, demonstrating that the L1 speak-
ers emphasize their own views; on the other hand, by preferring the
pattern I think + negative statement, the PSLE shows more inclination to
a listener-oriented approach, giving more emphasis to the information
in utterance than to their own views (Zhang & Sabet, in press).
turns. They are talking about the different steps needed for a society to
make progress.
S2: I think every person must have strong position to change them-
selves. I must try to change myself and I improve my culture.
I think we can’t improve others’ cultures, can’t change others.

We can just make some rules and, huh, and encourage people to
respect that rule and just this. I, I think by force, we can’t change
people. (P: 6:463)
S7: We must make aware people to know and understand beds of
rules. Just advantages of rules not just thinking about them-
selves. We must be aware to think about others, other persons.
(P: 6:464)
As (Excerpt 6.34) shows, S2 tries to highlight the inability to improve

others’ culture or to change people, but the role of I think is to deem-
phasize ‘I’, the speaker, and place the focus on the negative proposition.
In Excerpt (6.35) below, the L1’s I don’t think gives more emphasis to ‘I’
than to the negative proposition.
turns. They are talking about how morals can shape society.
S10: I don’t think it’s really a difference I think that, in bands tribes
and chiefdoms, they had to do that, to get people to follow ’em.
And, support ’em but if they could’ve, like just kept it all to them-
selves they would’ve. (L1: 2:40)
S1: Maybe so. I mean I’m not saying that’s not true but it was a, it’s
a standard of that society right? Yeah. (L1: 2:41)
Excerpt (6.35) shows that in turn 40, I don’t think + affirmative

statement is more speaker-centred than I think + negative. It makes
a stronger claim, giving more weight to the speakers’ view than
involving the listener directly or concentrating on the proposition.
While I don’t think seems to emphasize the speakers’ view, I think +
negative sentence appears to be more listener-centred. The focus on
the proposition in the second half of the sentence removes it from
the speaker and gives more attention to the listener or to the ideas
being expressed.
Persian language patterns encourage an indirect strategy by which the
speaker adopts a less authoritative position, and this pattern is carried
into their English constructions when they opt to use I think + negative
statement. The L1s, from a culture that expects people to take responsi-
bility for what they say, adopt the more direct mode (see Zhang & Sabet
(in press) for a comprehensive discussion of this issue).
turns. They are talking about whether newspapers should be granted
access to some photos.
S1: Yep (L1: 1:48)

S2: Well, I don’t think the newspapers should be granted access to the
photos because um, like basically we live in a morbid society so if
one newspaper has them even if they don’t publish ’em it’s gonna
get out, be on the net or something like that, [S1: mhm] and so so
people are gonna go see it and I think that is a gross invasion of
privacy, [S8: mhm] to have your pictures of your_ like if I was dead
and I had an autopsy (L1: 1:49)
In Excerpt (6.36), S2 in turn 49 highlights his own stance by using

I don’t think at the beginning of the sentence, which effectively reduces
the focus of the negative load from the proposition and adds it to his
personal view. However, in Excerpt (6.34), the Persian speaker is direct-
ing attention to the negative statement after I think, effectively down-
playing the personal view.
There is a negative correlation: when personal status rises, the nega-
tivity of the proposition falls; and when the negativity of proposition
grows, personal status decreases. In general, the Persian speakers give
more weight to the negative view, while the L1s appear to insist on
the personal view. The Chinese use both structures roughly evenly,
suggesting that they weigh personal views and negative statements as
equally important.
The more listener-oriented approach of the Persian speakers contrast-
ing with the speaker-dominated approach of the L1s is evident when
I think occurs before a discourse marker. The ‘listener involvement
marker’ you know (Romero Trillo 2002), located before I think by the
Persian speakers nine times indicates that the Persian speakers prefer to
orient toward the listener in their talk and create intimacy, but the L1
data lacks this type of cluster. Instead, they use the speaker-dominated
cluster I mean I think seven times: a cluster non-existent in the Persian
interactions. The Chinese again adopt a middle position by avoiding
both clusters.
The claim that the Persian language is more listener-oriented can be
supported by other patterns occurring in the Persian speakers’ interac-
tions. For example, they offer 21 occurrences of something you in the
data, but the L1s offer no examples and the Chinese only two.
turns. They are talking about how feelings can be expressed through
different communication channels.
S2: You can do everything a– (P: 2:72)

S7: How do you feel easily with this? For example, with speaking you
can transfer your felling easily or you’ll be writing or sometimes
you can write or something you can speak with us. (P: 2:73)
S2: It depends. (P: 2:74)
Excerpt (6.37) shows that the Persian speakers use something you to
address the listener. The language the Persian speakers use in English
shows traces of getting the listener involved in the interaction.
A flexible middle position is taken by the Chinese, with one trend
in common with the Persian speakers and one with the L1s. The L2
groups share a preference to include the listener in what they are talk-
ing about, manifested in what occurs after I think: both the Persian and
Chinese speakers use we rather than I. This creates intimacy before any
disagreement is expressed, and so can be used to maintain face of the
other participants. The listener-oriented approach also occurs in sole-
authored papers by the Chinese authors, which use we to involve the
reader in the communication process. The equal proportions of I think
we (6 per cent) in the Persian and Chinese data, against 2 per cent by the
L1s, highlight this difference. At the same time, the Chinese also share
a pattern with the L1s, using I think I at similar proportions to the L1s
(nine and eight respectively), against 2 per cent by the Persian speakers.
I may similarly accounts for 7 per cent of the overall occurrence of
may by L1 speakers and 4 per cent by Chinese, but none by Persian
speakers, who, on the other hand, accounted for half the overall occur-
rences of you may, which involves the listener, as opposed to about one-
third of the Chinese and L1 uses.
Another pattern of the same kind is observed in the subject pronouns
occurring after actually. While the Persian speakers are inclined to use
actually we to involve the listener, the L1s show no interest in using this
cluster, and the Chinese use it only once. The presence of actually I in
interactions of the L1s and Chinese suggests that these groups prefer a
speaker-dominated approach.
6.9 The manifestation of elasticity of vague language
The elasticity of vague language (Zhang 2011) in this study manifests

in the following aspects: linguistic elasticity (turn-taking, turn-shifting
position and clusters), pragmatic elasticity (serving interconnected and
elastic functions), and the versatility between vague language’s linguis-

tic realizations and pragmatic functions.
6.9.1 Linguistic elasticity

The data show that the L2 speakers make a more dominantly versatile
use of vague language than do the L1s. Some vague words provide a
speaker with more opportunities to make a strategic use of an expres-
sion to enrich communication. Elasticity allows vague words to stretch
(Zhang 2011); the first instance of such a use lies in the employment
of maybe. Of all the possibility makers examined in this study, maybe is
the most flexible in use, with the potential to appear in clause-initial,
clause-mid, and clause-final positions. The Chinese preferred maybe in
sentence-initial position as a turn-initiating device 59 times, accounting
for 19 per cent of the total maybe occurrences and the Persian speakers
44 times (28 per cent), while it is sparingly used by the L1s, five times
(8 per cent).
(6.38): This is a discussion between two Chinese participants over
two turns. They are talking about what they would like to be in the
future.
S8: Yes, it is an ideal career. I think I am forced to be a teacher. This is

a good job, and now as I am studying, I feel I will succeed in my
life. I feel, I have the feel to succeed, so … (Ch: 7: 42)
S6: Maybe that comes from our education. (Ch: 7: 43)
As is clear in Excerpt (6.38) above, S6 uses maybe as a turn-taking device

to interrupt S8 and begin his turn. The elasticity of vague language also
links with a multifunctional feature to serve different purposes. For
example, maybe used at the beginning of a clause by the Persian speak-
ers seems to be used to give indirect advice or express disagreement,
while maybe elsewhere in the sentence tends to acknowledge uncer-
tainty or suggest possibility.
turns. They are reacting to a situation in which one has been asked to
talk but refuses to do so.
S2: Right, we’re forcing everyone to speak. (P: 4:149)

S1: You are forcing. Maybe now she doesn’t have anything to say.
(P: 4:150)
S6: Maybe, we are respecting you ladies. (P: 4:151)
In Excerpt (6.39), clause-initial maybe in S1’s utterance expresses disa-

greement or disapproval, meaning that she is not happy with what S2
is doing, whereas S6 shows respect to S1 by using maybe in the same
clause-initial position. Excerpt (6.40) below, on the other hand, presents
a context in which maybe in clause-medial position is used to express
uncertainty or possibility.
(6.40): This is a discussion between three L1 participants over three
turns. They are talking about how to use a search engine.
SU-m: three. (L1: 3:317)

S1: Three-W. I could probably make it five-W or seven-W or ten-W
or eighteen-W, and I’m sure there’s an upper limit I think
the upper limit was twenty-two or something I- I never can
remember. But, you know it’s a sort of rule of thumb. If you want
it adjacent, it’s W. If you want it sort of, nearby three-W and if
(xx) you know you’re willing to sort a maybe make it, maybe in
adjacent sentences I may say something like nine-W, but I rarely
use that. Okay? Is everyone clear, up to this point? (L1: 3:318)
S2: I have one question. (L1: 3:319)
In Excerpt (6.40), S1 in turn 318 uses different vague expressions to

express possibility or uncertainty. The first possibility indicator, prob-
ably, is followed by repeating numbers in a manner that indicates
uncertainty. The next clause also contains a vague item, I think, the
uncertainty in which is reinforced by the approximator twenty-two or
something that follows. The clause-medial position maybe in turn 318
follows another vague word, sorta, to express tentativeness. The second
maybe in the same turn is followed by another clause to indicate uncer-
tainty, ‘I may say something like’. The comparison of Excerpt (6.39) and
Excerpt (6.40) demonstrates that the Persian speakers attached diverse
roles to maybe, finding this possibility indicator versatile, whereas L1s
used a variety of vague expressions. The occurrence of other possibility
indicators in Excerpt (6.40) such as probably, may, maybe, and the sub-
jectivizer I think in S1’s utterances act as confirmation of the elasticity
of vague language in the data.
The data reveal that the overall frequencies and the frequencies of
individual possibility indicators among the three groups are different,
and their patterns reveal discrepancies as well. The tendency to use
a lower frequency of possibility indicators by L2 speakers might be
because their lower proficiency level in comparison to the L1s impeded
them from using hedges (Mauranen 1997). This seems to be true for the
Persian group, who used the fewest possibility indicators, but not for
the Chinese group whose overall frequency of possibility indicators is
actually more than the L1s’.
Elasticity also emerges in the employment of I think in the sense that
both L2 groups use it as a strategy to take turns more often than the
L1s. While the L1s allocated the turn-taking task to this vague expres-
sion only 17 per cent, more than a quarter of the overall occurrences
of I think at the beginning of the Chinese and Persian sentences acted
as a turn-taking device. Both L2 groups used it in clause-final position
as well, whereas the L1s did not show much interest in ending a clause
with I think.
Similarly, elasticity is manifested in the use of I believe in clause-initial
and medial positions by the L2 speakers, while all tokens of I believe in
clause-initial position were in the L1 data. The L2 speakers use I believe
as a versatile tool for communication, as in medial position, it might
take the position of some discourse markers, and function as such. The
fact that the L2 speakers used vague expressions more elastically than
the L1s is because the inherent vagueness and consequent elasticity of
such words allow them to achieve the goal the speakers had set, despite
gaps in their vocabulary that might otherwise have prevented clear
communication.
Elasticity in the use of vague language can be manifested in quanti-
fiers where all instances of some of by the L1s occurred in clause-medial
position, but the L2 speakers preferred it in initial or medial position. As
with the subjectivizers discussed previously, the L2 groups gave the key
role of turn-initiator to some of: in Excerpt (6.41), for example, S2 uses
the turn-taking some of to take over from S6 in the interaction.
turns. They are talking about a nuclear bomb dropped on Hiroshima.
S6: They killed themselves. (P: 1: 178)

S1: They committed a suicide? (P: 1: 179)
S2: Some of them died. Some of them got killed. The one who was
supposed to throw the bombs didn’t like to do that to happen, so
these guys threw the bombs. (P: 1: 180)
The Persian speakers stretched something more than the other groups,
using it highly frequently in clause-final position. This also happened
with their placement of conjunctions after something, which was done
almost twice as often by the Persian speakers as by the other two groups.
While the L1s did not use things in negative sentences, the L2 groups
made strategic uses of this placeholder in such contexts—and not con-
fined to negative sentences: the use of things in interrogative sentences
by the L2 groups is also proportionately more common than by the L1s.
Although the Persian and L1 speakers used thing relatively evenly, 74
and 72 times, the L2 speakers used it twice as often in clause-final posi-
tion. This occurred in the Chinese interactions as well, where clause-
final position thing turned up 30 times, while the Persian speakers used
it only 14 times and L1 speakers only seven times.
The data show that speakers, especially L2 speakers, consistently
attempt to utilize the elasticity of vague language by digging into the
most flexible areas of its use and stretching it to fit their communica-
tive needs. This can be illustrated by the following contrasts between
the overall frequency continuum of a vague language category and the
elastic continuum of a relevant category member (Figures 6.2 and 6.3).
The most frequent
CSLE The least frequent
L1s
PSLE
Figure 6.2 Overall frequency continuum of possibility indicators
The most frequent
CSLE
The least frequent
PSLE
L1s
Figure 6.3 Elastic continuum of maybe

The most frequent
PSLE The least frequent
L1s
CSLE
Figure 6.4 Overall frequency continuum of placeholders
The most frequent
PSLE The least frequent
CSLE
L1s
Figure 6.5 Elastic continuum of things
These contrastive examples in overall frequency and individual fre-

quency are a function of elasticity of the vague items involved. For
example, as Figures 6.4 and 6.5 show, the overall occurrence continuum
of placeholders stretches from Persian, to L1 and to Chinese, but stretches
from Persian, to Chinese and to L1 when the elastic continuum of things
is drawn.
6.9.2 Pragmatic elasticity

The elasticity of vague language is not confined to linguistic elasticity.
The analysis of pragmatic functions of vague language reveals that the
possibility of one vague item serving different functions is a blessing in
language. This can be viewed in the examples discussed in Chapter 5,
where each function is fulfilled by different vague words. For instance,
politeness was expressed through possible by the L1s, maybe by the
Chinese and something by the Persian speakers (see Section 5.1.2). This
elasticity can also be realized in the consecutive occurrence of different
types of vague expression to heighten a single function.
(6.42): This is a turn by a Persian speaker. He is comparing some social
norms in Iran and the USA.
S2: But, you know this level of intellectuality if I am right, you know
here if we have the maximum, one hundred, ok? I think in Iran it
is 20. I think, I don’t know European countries maybe it is 30 but
it is much more. Ok? Maybe it is 50. I agree. Maybe it is the same
as here but there are differences. (P: 1: 502)
In Excerpt (6.42), S2 is trying to make a claim, but to be safe, he uses

multiple self-protection tools in case the opposite is proven right. These
tools, despite serving the same function, do not belong to the same
vague category. The self-protection is initiated by a clause that does
not contain a vague word: ‘if I am right’, but the speaker does not find
it sufficient and adds some vague words. A series of vague expressions
serve the same function of self-protection: the subjectivizer I think and
two possibility indicators maybe.
6.9.3 Versatility between realizations and functions

The analysis of the pragmatic functions of vague language in this study
supports the notion of the elastic nature of vague language (Zhang
2011). It reveals that there is an interconnection between the linguistic
realizations and the pragmatic functions of vague language, in that a
particular vague category tends to serve a particular pragmatic function.
This does not mean that there is a one-to-one correspondence between
a lexical item and a pragmatic function, but that a lexical item may
serve a range of functions depending on the communicative context
as ‘vague language is stretched in varying directions to serve pragmatic
functions and maxims’ (Zhang 2011, p. 592). For instance, placeholders
typically serve to express the right amount of information, mitigation
and downtoning. Quantifiers often serve functions such as the right
amount of information, emphasis, and mitigation.
In communication, a pragmatic function can be served by a vague

expression, but sometimes the speaker might feel a single word does not
express the desired emphasis and makes use of several. This allows the
speaker to stretch the pragmatic elasticity of vague language and align
the expression to the required degree.
The discussion in the present chapter shows that vague language
appears as an appropriate pragmatic tool to enrich communication.
Using this versatile tool, both L1 and L2 speakers try to manage poten-
tial communication pitfalls. As the pitfalls each group encounters are
different, the speakers may employ contrastive vague language in their

communication. Vague language has a substantial manifestation in the
language of EFL speakers, as is indicated by Zhang’s (2011) interpretation
of the elasticity of vague language: namely, that speakers may on many
occasions stretch the slingshot of vague language to meet linguistic and
communicative targets.
7
Conclusions and Implications
Questions such as how often is often or how many is many have been
the concern of the study of vague language. As an integral part of the
language, vague language in this study refers to inexact expressions
that are elastically used to contribute to effective communication. This
study, one of the few looking at vague language in terms of its elasticity,
has investigated how it can meet the communication needs of L2 learn-
ers of English, as compared with L1s, providing insights into its use in
intercultural contexts, including English language teaching.
With the fast-growing body of literature on vague language in recent
years, the inadequacy of studies of this feature of natural language in
ELT has become more conspicuous, and more research in this field
has been inspired. This study is a small step toward a more adequate
account of vague language in the context of academic settings, with
special attention to L1 versus L2 use of English.
The results show greater tendencies for vague language use by the
learner groups, and all three groups showed statistically different perfor-
mances. The Chinese were the most frequent users of vague language, but
with an uneven distribution of items in each vague category. Usability
of a vague expression in multiple positions is found to contribute to its
functionality, which results in the large frequency of vague expressions.
The most intriguing finding of this study is that the elastic feature
of vague language allows speakers to stretch it to satisfy their com-
municative needs. The most versatile vague categories (for example
subjectivizers) and items were most preferred by the L2 groups, as these
met their diverse communicative needs. Elasticity allows vague words
to stretch and provide the speaker with opportunities to make strategic
use of these expressions to enrich communication. For example, vague
language can mitigate, show politeness or solidarity, or maintain face.
189
Not only is vague language convenient for successful communica-

tion, but it can also facilitate the structural management of an interac-
tion. There is an interconnection between the linguistic realizations
of vague items and the particular functions they serve. This is not as
a one-to-one correspondence but as a continuum of particular func-
tions in relation to the linguistic realization of vague items. It is also
revealed that the cultural and linguistic backgrounds of L2 speakers can
influence their vague language use. These can occur as taarof, formal
courtesy, in Persian and indirectness in Chinese.
Learners can be taught to take advantage of the elasticity of vague
language in the process of communication: for instance, by instruction
in the ways vague language can be used to compensate for inadequacies
in their communicative competence. The findings may be applied in
language pedagogy, particularly in curriculum development and teacher
education.
7.1 Conclusions
Based on naturally occurring classroom data among L1, Chinese and

Persian groups, the findings of this study challenge Nikula’s (1996)
and Ringbom’s (1998) claims that vague language is more exten-
sively used in L1s’ interaction: both L2 learner groups in this study
showed a greater tendency to use vague language, although it is found
overwhelmingly to be part of the communicative competence of all
speakers. This study reveals that vague language occurs frequently in
ELT contexts, although each group may reveal trends unique to them.
The total vague expressions used in the data show that L2 speakers
are vaguer than L1s; Chinese are much vaguer than Persians and L1
speakers. One of the reasons could be that in general L2 speakers need
vague language to compensate for their lack of English proficiency, for
example, vocabulary and variety of known phrases. The data show that
the Chinese in particular prefer speaking vaguely more than the other
two groups.
Their use of vague intensifiers indicates that the Chinese were more
emphatic than the other two groups. The frequency of their use of all
five categories indicates that the Chinese tended to use vague language
for emphasis most and for generalization least. The Persian speakers
used vague language to generalize most (the opposite of the Chinese)
and to express possibility least. The L1 group preferred vague language
for quantifying most and for expressing a subjective stance least: clearly,
the groups are dissimilar in their use of the five lexical categories of
Conclusions and Implications 191
vague language. Based on the use of maybe and if it correlates with the
tentativeness, then the Chinese seem to be more tentative than the L1
and the Persian speakers, the L1 group looks less tentative than the two
L2 groups, and the Persian speakers are less tentative than the Chinese
but more tentative than the L1s.
One of the striking findings of this study is that the Persian speak-
ers adopted a listener-oriented approach, against the speaker-oriented
approach of the L1s; the Chinese took a middle position. That is,
Persian speakers were less authoritative than the L1s whose speaker-
dominated approach was evident in the more assertive language used.
The less assertive language of the Persian speakers was manifested in the
frequent application of but I think to softly express disagreement and
indicate contrast, and in the use of I think + negative clause to mitigate
negativity. This was further reinforced by the dominant use of I think
we and you know I think, used to establish intimacy and create a sense of
cooperation. Supporting evidence is the absence of I mean I think in the
Persian data. The L1 group, on the other hand, used the cluster I don’t
think + positive clause, which indicates a self-assertive attitude. The data
reveal a negative correlation: when the speaker’s view is foregrounded,
the focus on the negativity of the proposition is backgrounded; when
the negativity of the proposition is foregrounded, the speaker’s view is
backgrounded.
Elasticity refers to the usability of a category or item in multiple
positions: its diverse functionality. This feature provides speakers with
the opportunity to achieve communication needs appropriately and
effectively. With its elastic nature, vague language can be stretched and
enhance communication. The most versatile category and items were
consistently the two most popular among the L2 speakers in the class-
room interactions under study here. For example, subjectivizers, which
can occur in different positions in a sentence, are found to be the most
popular category with the Chinese and the second most popular with
the Persian speakers.
Elasticity contributes to the frequency of vague language in L2 speak-
ers’ interactions. Versatile items are much preferred by L2 speakers, and
elasticity seems to be processed by them depending on their ease of use
and the potential to meet communicative needs; if an item in one lan-
guage has a recognizably similar function in another, or if it has been
internalized sufficiently to be used effectively in L2 interactions. The
potential to meet communicative needs is judged according to how an
item can convey an intended meaning and enhance communication. It
appears that English learners need to talk about things more than L1s,
and find the elastic nature of vague language particularly helpful in

satisfying these needs.
The overall frequency of vague language reveals significant differ-
ences across the three groups with the L2 learners, in particular the
Chinese, the most frequent users. Except for one, Chinese speakers used
individual categories most heavily. This can be interpreted in terms of
the elasticity of vague language, in that speakers stretch vague language
to the point where their needs are met.
The three groups have nothing in common as far as ranking is con-
cerned. Only a few categories are found to be in common: for instance,
vague quantifiers were the second most frequently used category by
both Chinese and Persian speakers. However, the ranking of items
within categories shows a more consistent pattern: the L1s used each
subcategory more evenly, while the L2 groups were inclined to employ
some items far more frequently than others, so the frequency distribution
of the L2 data is more concentrated than in the L1 data. This mirrors the
effects of such factors as L1 influence, cultural norms, cognitive effects,
and pedagogic contextual influences on vague language use. The lexical
analysis reveals statistically significant differences among the three
groups, and in the frequency distribution of the subcategories and
vague language patterns used, all are attributable to cultural, linguistic,
and pedagogic factors.
The functional investigation acknowledges the diversity of vague
language expressions, whereby speakers have at their disposal diverse
vague language forms, sufficient to deal with the specific functions in
each data set. This verifies the fact that there exists not a rigid con-
nection between one form and one function but an elastic matching
of lexical and functional categories, and a diversity of options avail-
able. What is obvious in the data sets is that not only can vague lan-
guage contribute to more convenient communication, but it can also
facilitate the structural management of interactions in both L1 and L2
communication.
L2 speakers’ cultural and linguistic backgrounds can be influential
in the employment of vague language when they communicate in
English, as in the example of politeness. The Persian speakers’ use of
vague language is a carry-over from the cultural concept of taarof, and
the Chinese preference for indirectness as a cultural norm is expressed
similarly in vague language; L1s in this study preferred more directness
and frankness, as is more common in their cultural background.
Elasticity of vague language can create a versatile continuum of expres-
sions. In this study two sets of continua have been explored, lexical and
versatile continua. None of the lexical items in this study was the most
frequently used by L1s; the L2 speakers used the most vague words.
They also showed a more concentrated use in each category of vague
language, such as I think in subjectivizers, for their versatility. This is
where the maximum potential of vague language elasticity is fulfilled:
to address the speakers’ needs and their goals in communication.
What this implies is that in addition to the overall vague language
continuum which indicates vague language elasticity, there exists an
intra-category continuum (in terms of different rankings) within each
group arranged according to the item’s versatility for that group: the
more versatile a vague word, the more frequently it occurs. The L2 data
reveal that the elasticity of vague language leads to the versatility of a
vague item, and the versatility of the item contributes to the frequency
of its use.
The elasticity of vague language emerges as a result of the uneven
distribution of vague items in each vague category that the L2 speakers
employed in communication. By contrast, the distribution by the L1s
downplayed the continua in the data. One reason why there is no or
little evidence showing L2 speakers having difficulty in communicating
might lie in the fact that they are able to resort to the elasticity of vague
language to compensate for any inadequacies; therefore, the overall
heavier use of vague language by the L2 speaker groups does not mean
they choose to be vague, but that they resort to vague language as a
versatile and reliable tool to secure enhanced communication.
This study confirms the interconnection between the linguistic reali-
zations of a vague lexical item and the particular functions it can serve.
Although there does not seem to be a correspondence between each
vague item and a function, this study reveals a continuum of particular
functions in relation to the linguistic realization of a vague item. The
function of vague language is determined by the context involved.
The findings of this research shed light on the different linguistic
behaviors of L1 and L2 groups, especially in improving the pragmatic
competence of EFL learners. They also promote the need for vague lan-
guage teaching in English. Despite the fast-growing evolution of English
as a global language whereby a uniformity of discourse verities emerges,
some discrepancies are inseparable from discourse communities. That
is, some discrepancies remain, inherited from the cultural and linguistic
norms of different communities. As a result, differences such as vague
language use need to be highlighted among discourse communities
with the aim of fostering better communication among speakers, espe-
cially between different L1 or L2 groups.
7.2 Limitations
A limitation beyond the control of the researchers is the number of

words in each data set. The choice of around 50,000 words, set by the
L1 data, may not be large enough to provide an accurate account of
each group’s vague language use in classroom interaction, but as the
study required three data sets, a total of 150,000 words is sufficient for
a credible analysis.
There is some discrepancy among the data sets: a teacher is clearly
present in the L1 data, minimally in the Persian and not in the Chinese
data. The L1 data is composed of L1 teacher/student or student/student
interactions. In the Persian data, the Persian-speaking teachers have
near native fluency in English, but the main focus in the class is on the
learners’ language, the teacher performing as a facilitator with minimal
speech production to prevent distortion of the learners’ language. The
Chinese data consists of learner language exclusively: because all the
teachers at that level were L1s of English and their participation might
distort the naturalness of the Chinese data, all classes were facilitated
by senior L2 students. While the three data sets were not collected
in exactly the same formats, the validity of the data is believed to be
upheld.
Considering there was no control over the L1 subjects, great effort
was made to keep the same topics of discussion among the three groups.
The new topics were not totally different from the original ones, arising
from and being related to them. Despite the infeasibility of keeping the
topics the same throughout, efforts were made to keep the topics mostly
similar, to a large extent, to minimize the potential distortion of differ-
ent topics on the data.
L1s data came from a ready-made transcript, and the researchers were
unable to make changes to the restricted transcription conventions. For
uniformity, the L1 data’s conventions were applied to the transcribed L2
data sets to facilitate comparability.
7.3 Implications
Despite the growing body of literature in the area of vague language,

and tacit acknowledgment of its crucial role in academic discourse,
little has been written on vague language in language pedagogy. This
research fills a gap in the existing literature. It investigates the mani-
festation of vague language from the perspective of elasticity to find
how its fluid nature enhances communication by speakers from three
different linguistic and cultural backgrounds. The findings of this

research can support the study of language from both linguistic and
pedagogic viewpoints.
This study contributes to a fuller understanding of what comprises
communicative competence. Vague language can be taught to learners
of English, not necessarily adopting L1 language patterns as a model but
focusing on how its elasticity can be used advantageously in the process
of communication. This can be taught in all four components of com-
municative competence: grammatical competence, discourse compe-
tence, sociolinguistic competence, and strategic competence. Students
can be instructed on the appropriate use of vague language elasticity to
compensate for weaknesses in their communicative competence, and
taught discourse management strategies based on vague language use;
for instance, how it can be used to direct spoken discourse like turn-
taking and turn-giving. Vague language is overwhelmingly needed in
academic contexts, and explicit instruction on its appropriate use is
required.
This study may help develop new concepts of the operation of vague
language in communication. Rather than viewing vague language as a
static phenomenon, attention needs to be paid to its dynamic nature.
It is argued in this study that the versatility of a vague word helps to
determine the frequency of its use: the more elastic an item, the more
frequently it occurs. Vague language should therefore be looked at
from the elastic perspective. A continuum-based approach may prove
successful.
An understanding of how vague language can help them meet their
needs and reflect their linguistic and cultural backgrounds would be
helpful to learners of English, and would help prevent miscommunica-
tion or misunderstanding in ESL classes where learners are from diverse
countries. It can promote the co-existence of varieties of English, side-
by-side in a single context, with each variety maintaining its patterns
of vague language use and learning to recognize and accept patterns
different from theirs.
These findings can shed light on classroom interactions, curriculum
development and teacher education. Competency in such areas may
be obtained through formal instructions, a responsibility that lies
partly with curriculum developers and partly with teachers. This study
provides ELT curriculum developers with useful notions for designing
academic materials with due attention to vague language teaching, to
improve language learners’ pragmatic competence. It reveals that stu-
dents also need to gain mastery over the clustering patterns of vague
language, so developers need to incorporate vague language exercises.

The awareness of vague language elasticity can contribute to improve-
ment in both written and spoken modes; the written mode of vague
language, although not addressed in this study, is in particular required
in academic writing where precision is traditionally promoted.
This study suggests that teachers may need to provide learners with
supplementary materials and design classroom exercises to enhance
the appropriate use of vague language. It also indicates that learners
need to learn the use and positions of vague language. It is recom-
mended that explicit instruction is provided on three important
features of vague language; where, when, and how to use it. This will
contribute to the development of pragmatic competence by learners
of English.
This study informs teachers of the kinds of linguistic and cultural
trends in vague language use that can be presented to learners in ESL
settings. These can prevent vague language from raising potential
student/student or student/teacher misunderstandings or miscommu-
nication in ESL contexts by identifying potential cultural or linguistic
conflicts in advance. The elastic nature of vague language confirms the
crucial role it plays in communication, and an important implication
of this study is the need to incorporate vague language instruction into
language pedagogy.
7.4 Further research
Given the significant differences found among the three groups (L1,
Chinese, and Persian speakers), an area for further research is an investi-
gation of vague language within each group, which will neutralize such
factors as linguistic and cultural differences and determine whether
individual psychological factors such as personality-type cause differ-
ences in the frequency of vague language categories or vague language
patterns.
Another area open for investigation is the examination of other vague
categories such as approximators and general extenders among the
same three language groups, to provide a more comprehensive view of
the conceptual dimension of vague language study across the groups.
The data for the present study comprises spoken language in class-
room interactions, but other studies might focus on written discourse
to determine if the mode of discourse is a source of differences in the
frequency and the patterns of vague language use: in other words, a
comparison of spoken and written language can provide a more detailed
account of the use of vague language. Other future research might also
investigate whether different topics can affect the frequency or elasticity
of vague language in both L1 and L2 contexts.
With the significance of vague language for communicative compe-
tence, shown in this study and acknowledged in the literature, there
is no question but that it should be included in the curriculum and
included in language pedagogy. Future research can also address a ques-
tion one step further: whether the explicit teaching of vague language
can lead to enhanced mastery of this important feature of language.
Appendix I: Consent Form for the
Director and Teachers
This consent form will be held for a period of five years.
Title: An investigation of vague language use in academic settings
• I agree to give access to the researcher for recording the classes in my language
centre.
• Students and teachers may participate in the above study if they so wish.
• I acknowledge that the nature of the study and the recording procedure has
been explained to my satisfaction by the researcher and my consent is given
voluntarily. I have had the opportunity to ask questions and have them
answered.
• I understand that the data will be stored in a secure place to safeguard
confidentiality.
• I understand that the data will be stored for any possible future research.
Signature: ---------------------------------------
Name: ------------------------------------
(Please print clearly)
Date: -------------------------------------
Contact number/E-mail: ---------------------------------------
APPROVED BY CURTIN UNIVERSITY HUMAN RESEARCH ETHICS COMMITTEE
FOR……………..YEARS ON………(DATE)……………….,
REFERENCE NUMBER………………………..
198
Appendix II: Consent Form for
Participants
This consent form will be held for a period of five years.
Title: An investigation of vague language use in academic settings
• I agree to take part in this research and to be audio or video taped.

• I acknowledge that the nature of the study and the recording procedure has
been explained to my satisfaction by the researcher and my consent is given
voluntarily. I have had the opportunity to ask questions and have them
answered.
• I am aware that all the information I provide for this research project is con-
fidential and my identity will be protected at all times.
• I give permission to record for about two hours.
• I understand that I can choose to have the recorder turned off at any time and
I am free to delete all or parts of my recordings as I wish. I can withdraw all
the information I give by …………………without giving a reason.
• I understand that the data will be stored in a secure place to safeguard
confidentiality.
• I understand that the data will be stored for any possible future research.
Signature: ----------------------------
Name: -------------------------------
(Please print clearly)
Date: ---------------------------------
Contact Number/E-mail: ------------------------------
APPROVED BY CURTIN UNIVERSITY HUMAN RESEARCH ETHICS COMMITTEE
FOR……………..YEARS ON……… (DATE)……………….,
REFERENCE NUMBER……………………….
199
References
Abdollahzadeh, E. (2003). Interpersonal metadiscourse in ELT papers by Iranian

and Anglo-American academic writers. Paper presented at the International
Conference on Multiculturalism in ELT Practice, Baskent University, Turkey.
Adolphs, S., Atkins, S., & Harvey, K. (2007). Caught between professional require-
ments and interpersonal needs: Vague language in healthcare contexts. In
J. Cutting (ed.), Vague language explored (pp. 62–78). Basingstoke: Palgrave
Macmillan.
Aijmer, K. (1997). I think-an English modal particle. In T. Swan & O. J. Westvik
(eds.), Modality in Germanic languages: Historical and comparative perspectives
(pp. 1–47). Berlin: Mouton de Gruyter.
Alcon, E. (2005). Does instruction work for pragmatic learning in EFL contexts?
System, 33(3), 417–435.
Alcon, E., & Martinez-Flor, A. (2008). Pragmatics in foreign language contexts.
In E. Alcon & A. Martinez-Flor (eds.), Investigating pragmatics in foreign language
learning, teaching and testing (pp. 3–21). Bristol: Multilingual Matters.
Allison, D. (1995). Assertions and alternatives: Helping ESL undergraduates
extend their choices in academic writing. Journal of Second Language Writing,
4, 1–15.
Allwood, J., Cerrato, L., Jokinen, K., Navarretta, C., & Paggio, P. (2007). The
Mumin coding scheme for the annotation of feedback, turn management and
sequencing. Multimodal Corpora for Modelling Human Multimodal Behaviour:
Special Issue of the International Journal of Language Resources and Evaluation,
41(3–4), 273–287.
Aristotle. (1946). The works of Aristotle (Vol. XI.). Oxford: Clarendon Press.
Atai, M. R., & Sadr, L. (2008). A cross-cultural genre study of hedging devices in
discussion section of applied linguistics research articles. Journal of Teaching
English Language and Literature Society of Iran, 2(7), 1–22.
Austin, J. L. (1962). How to do things with words. Cambridge: Clarendon Press.
Bachman, L. F. (1990). Fundamental considerations in language testing. Oxford:
Oxford University Press.
Bardovi-Harlig, K., & Griffin, R. (2005). L2 pragmatic awareness: Evidence from
the ESL classroom. System, 33, 401–415.
Beebe, L. M., & Cummings, M. C. (1996). Natural speech act data versus written
questionnaire data:How data collection method affects speech act performance.
Berlin: Mouton de Gruyter.
Behnam, B., & Niroomand, M. (2011). An investigation of Iranian EFL learners’
use of politeness strategies and power relations in disagreement across differ-
ent proficiency levels. English Language Teaching, 4(4), 204–220. Retrieved from
http://dx.doi.org/10.5539/elt.v4n4p204 doi:10.5539/elt.v4n4p204. (Retrieved
on June 23rd, 2012)
Beighmohammadi, A. (2003). An investigation into the patterns of use of discourse
features of intensity markers in academic research articles of hard science, social sci-
ence and TEFL. (Unpublished Master’s thesis). University of Tehran, Tehran, Iran.
200
References 201
Biber, D., & Barbieri, F. (2007). Lexical bundles in university spoken and written
registers. English for Specific Purposes, 26, 263–286.
Biber, D., Johansson, S., Leech, G., Conrad, S., & Finegan, E. (1999). Longman
grammar of spoken and written English. London: Pearson Education.
Billig, M. (1999). Conversation analysis and the claims of naivety. Discourse and
Society, 10(4), 572–576.
Blum-Kulka, S., House, J., & Kasper, G. (1989). Cross-cultural pragmatics: Requests
and apologies. Norwood: Ablex.
Bradac, J., Mulac, A., & Thompson, S. (1995). Men’s and women’s use of intensi-
fiers and hedges in problem-solving interaction: Molar and molecular analysis.
Research on Language and Social Interaction, 28(2), 93–116.
Brazil, D. (1997). The communicative role of intonation in English. Cambridge:
Cambridge University Press.
Brown, P., & Levinson, S. C. (1978). Universals in language usage: Politeness phe-
nomena. In E. N. Goody (ed.), Questions andpoliteness. Cambridge: Cambridge
University Press.
Brown, P., & Levinson, S. C. (1987). Politeness: Some universals in language usage.
Cambridge: Cambridge University Press.
Brown, P., & Levinson, S. C. (1994). Immanuel Kant among the Tenejapans:
Anthropology as empirical philosophy. Ethos, 22(1), 3–41.
Carter, R., & McCarthy, M. (2006). Cambridge grammar of English: A comprehensive
guide. Spoken and written English grammar and usage. Cambridge: Cambridge
University Press.
Chan, A. Y. (2013). Analysis of Chinese speakers’ indirect and contrary-to-
face value responses to survey interview questions. In Y. Pan & D. Z. Kàdàr
(eds.), Chinese discourse and interaction (pp. 175–198). Sheffield: Equinox
Publishing Ltd.
Channell, J. (1994). Vague language. Oxford: Oxford University Press.
Chen, W. Z., & Wu, S. X. (2002). Fanchou yu Mohu Yuyi Yanjiu (in Chinese) [A
Study on Category and Semantic Fuzziness]. Fujian (China): Fujian Renmin
Press.
Cheng, W. (2007). The use of vague language across spoken genres in an
intercultural Hong Kong Corpus. In J. Cutting (ed.), Vague language explored
(pp. 161–181). Basingstoke: Palgrave Macmillan.
Cheng, W., & Warren, M. (1999). Inexplicitness:What is it and should we be
teaching it? Applied Linguistics, 20(3), 293–315.
Cheng, W., & Warren., M. (2001). The use of vague language in intercultural
conversations in Hong Kong. English World-Wide, 22(1), 81–104.
Cheng, W., & Warren, M. (2003). Indirectness, inexplicitness and vagueness
made clearer. Pragmatics, 13(13), 381–400.
Cheshire, J. (2007). Discourse variation, grammaticalisation, and stuff like that.
Journal of Sociolinguistics, 11(2), 155–193.
Chow, K. F. (2010). A quasi-self-dual interpretation of negative raising academic area:
Formal semantics. Paper presented at the 3rd Postgraduate Research Symposium:
Language and Cultural Studies in the Pearl River Delta, Hong Kong.
Conrad, S., & Biber, D. (2000). Adverbial marking of stance in speech and writ-
ing. In S. Hunston & G. Thompson (eds.), Evaluation in text: Authorial stance
and the construction of discourse (pp. 56–73). New York: Oxford University
Press.
202 References
Connor, U. (1996). Contrastive rhetoric: Cross-cultural aspect of second-language

writing. Cambridge: Cambridge University Press.
Cook, G. (1989). Discourse. Oxford: Oxford University Press.
Cotterill, J. (2007). I think he was kind of shouting or something: Uses and
abuses of vagueness in the British courtroom. In J. Cutting (ed.), Vague language
explored (pp. 97–114). Basingstoke: Palgrave Macmillan.
Creswell, J. W. (1994). Research design: Qualitative and quantitative approaches.
Thousand Oaks: Sage Publications.
Creswell, J. W. (2008). Educational research: Planning, conducting, and evaluating
quantitative and qualitative research (3rd ed.). Upper Saddle River: Pearson/
Merrill Prentice Hall.
Creswell, J. W. (2009). Research design: Qualitative, quantitative, and mixed methods
approaches (3rd ed.). Thousand Oaks, CA: Sage.
Creswell, J. W., & Plano Clark, V. L. (2007). Designing and conducting mixed meth-
ods research. Thousand Oaks, CA: Sage.
Creswell, J. W., Plano Clark, V. L., Gutmann, M. L., & Hanson, W. E. (2003).
Advanced mixed methods research designs. In A. Tashakkori & C. Teddlie
(eds.), Handbook of mixed methods in social and behavioral research (pp. 209–240).
Thousand Oaks: Sage.
Crystal, D., & Davy, D. (1975). Advanced conversational English. London: Longman.
Cutting, J. (2006). Spoken grammar: Vague language and EAP. In R. Hughes (ed.),
Spoken English, TESOL and applied linguistics (pp. 159–181). New York: Palgrave
Macmillan.
Cutting, J. (2007). Introduction to vague language explored. In J. Cutting (ed.),
Vague language explored (pp. 3–20). Basingstoke: Palgrave Macmillan.
Cutting, J. (2008). Pragmatics and discourse (2nd ed.). Oxford: Routledge.
Davies, B. L. (2007). Grice’s cooperative principle: Meaning and rationality.
Journal of Pragmatics, 39, 2308–2331.
De Cock, S., Granger, S., Leech, G., & McEnery, T. (1998). An automated
approach to the phrasicon of EFL learners. In S. Granger (ed.), Learner English
on computer (pp. 67–79). London: Longman.
Deese, J. (1974). Towards a psychological theory of the meaning of sentences.
Hillsdale: Lawrence Ebraham Associates.
Drave, N. (2002). Vaguely speaking: A corpus approach to vague language. In P. Peters,
P. Collins, & A. Smith (eds.), New frontiers of corpus research (pp. 25–40). Amsterdam:
Rodopi.
Dudley-Evans, T. (1991). Socialisation into the academic community: Linguistic
and stylistic expectations of a PhD thesis as revealed by supervisor comments.
In P. Adams, B. Heaton, & P. Howarth (eds.), Socio-cultural issues in English for
academic purposes (pp. 42–51). Basingstoke: Modern English Publications.
Erev, I., Wallsten, T. S., & Neal, M. M. (1991). Vagueness, ambiguity, and the cost
of mutual understanding. Psychological Science, 2, 321–324.
Firestone, W. (1987). Meaning in method: The rhetoric of quantitative and quali-
tative research. Educational Researcher, 16, 16–21.
Flick, U. (2002). An introduction to qualitative research. London: Sage.
Franken, N. (1997). Vagueness and approximation in relevance theory. Journal of
Pragmatics, 28, 135–151.
Goffman, E. (1967). Interaction ritual: Essays on face-to-face behavior. New York:
Anchor.
References 203
Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (eds.), Syntax
and Semantics (Vol. 3: Speech acts, pp. 41–58). New York: Academic Press.
Grice, H. P. (1989). Studies in the way of words. Cambridge: Harvard University
Press.
Gudykunst, W. B., Matsumoto, Y., Ting-Toomey, S., Nishida, T., Kim, K., &
Heyman, S. (1996). The influence of cultural individualism-collectivism, self
construals, and individual values on communication styles across cultures.
Human Communication Research, 22(4), 510–543.
Hamilton, M. A., & Minoe, P. J. (1998). A framework for understanding equivoca-
tion. Journal of Language and Social Psychology, 17, 3–35.
Hanson, W. E., Creswell, J. W., Plano Clark, V. L., Petska, K. P., & Creswell, J. D.
(2005). Mixed methods research designs in counseling psychology. Journal of
Counseling Psychology, 52(2), 224–235.
Heider, E. R. (1973). On the internal structure of perceptual and semantic catego-
ries. In T. E. Moore (ed.), Cognitive development and the acquisition of language
(pp. 111–144). New York: Academic Press.
Holmes, J. (1986). Functions of you know in women’s and men’s speech. Language
in Society, 15(1), 1–22.
Holmes, J. (1988). Doubt and uncertainty in ESL textbooks. Applied Linguistics,
9(1), 21–44.
Holmes, J. (1990). Apologies in New Zealand English. Language in Society, 19,
155–199.
Holmes, J., & Brown, D. (1987). Teachers and students learning about compli-
ments. TESOL Quarterly, 21(3), 523–546.
Hu, Z., Brown, D., & Brown, L. (1982). Some linguistic differences in the
written English of Chinese and Australian students. Language Learning and
Communication, 1, 39–49.
Huang, Y. (2007). Pragmatics. Oxford: Oxford University Press.
Hübler, A. (1983). Understatements and hedges in English. Amsterdam: John
Benjamins.
Hutchby, I., & Wooffitt, R. (1998). Conversation analysis: Principles, practices and
applications. Oxford: Polity Press.
Hutchby, I., & Wooffitt, R. (2008). Conversation analysis (2nd ed.). Cambridge:
Polity Press.
Hyland, K., & Milton, J. (1997). Qualification and certainty in L1 and L2
students’ writing. Journal of Second Language Writing, 6(2), 183–205.
Hyland, K. (1994). Hedging in academic writing and EAP textbooks. English for
Specific Purposes, 13(3), 239–256.
Hyland, K. (1996). Nurturing hedges in the ESP curriculum. System, 24(4),
477–490.
Hyland, K. (1998). Hedging in scientific research articles. Amsterdam: John
Benjamins.
Hyland, K. (2000). Hedges, boosters and lexical invisibility: Noticing modifiers in
academic texts. Language Awareness, 9(4), 179–197.
James, A. R. (1983). Compromisers in English: A cross-disciplinary approach to
their interpersonal significance. Journal of Pragmatics, 7(2), 191–206.
Janicki, K. (2002). A hindrance to communication: The use of difficult and
incomprehensible language. International Journal of Applied Linguistics, 12(2),
194–217.
204 References
Jick, T. D. (1979). Mixing qualitative and quantitative methods: Triangulation in

action. Administrative Science Quarterly, 24, 602–611.
Jucker, A. H., Smith, S. W., & Ludge, T. (2003). Interactive aspects of vagueness in
conversation. Journal of Pragmatics, 35(12), 1737–1769.
Kaltenböck, G. (2010). Pragmatic functions of parenthetical I think. In
G. Kaltenböck, W. Mihatsch & S. Schneider (eds.), New approaches to hedging
(pp. 237–66). Bingley: Emerald.
Kaltenböck, G. (2013). Development of comment clauses. In B. Aarts, J. Close,
G. Leech, & S. Wallis (eds.), The verb phrase in English: Investigating recent change
with corpora (pp. 286–317). Cambridge: Cambridge University Press.
Kärkkäinen, E. (2003). Epistemic stance in English conversation: A description of its
interactional functions, with a focus on I think. Amesterdam: John Benjamins.
Kärkkäinen, E. (2010). Position and scope of epistemic phrases in planned and
unplanned American English. In G. Kaltenböck, W. Mihatsch & S. Schneider
(eds.), New approaches to hedging (pp. 203–236). Bingley: Emerald.
Khurshudyan, V. (2006). Corpus-based study on phonetically-relevant hesitation
markers in oral discourse: On the material of Armenian, Russian, Italian and
English. PhD dissertation abstract. Russian State University for the Humanities,
Moscow.
Koester, A. (2007). About twelve thousand or so: Vaguenss in North American
and UK offices. In J. Cutting (ed.), Vague language explored (pp. 40–61).
Basingstoke: Palgrave Macmillan.
Koutlaki, S. A. (2002). Offers and expressions of thanks as face enhancing acts:
Ta’arof in Persian. Journal of Pragmatics, 34, 1733–1756.
Lado, R. (1957). Linguistics across cultures: Applied linguistics for language teachers.
Ann Arbor: University of Michigan Press.
Lakoff, G. (1972). Hedges: A study in meaning criteria and the logic of fuzzy
concepts. In Papers from the 8th regional meeting of Chicago Linguistic Society,
183–228.
Lakoff, G., & Johnson, M. (1980). Metaphors we live by. Chicago: University of
Chicago Press.
Lakoff, R. T. (1990). Talking power: The politics of language. New York: Basic Books.
Levinson, S. C. (1983). Pragmatics. Cambridge: Cambridge University Press.
Levinson, S. C. (1989). A review of relevance. Journal of Linguistics, 25, 455–472.
Liddicoat, A. J. (2005). Culture for language learning in Australian language-in-
education policy. Australian Review of Applied Linguistics, 28(2), 1–28.
Liddicoat, A. J. (2007). An introduction to conversation analysis. London: Continuum.
Liddicoat, A. J., & Crozet, C. (eds.). (2001). Acquiring French interactional norms
through instruction. In K. Rose & G. Kasper (eds.), Pragmatics in language teach-
ing (pp. 125–144). New York: Cambridge University Press.
Ma, R. (1996). Saying ‘yes’ for ‘no’ and ‘no’ for ‘yes’: A Chinese rule. Journal of
Pragmatics, 25, 257–271.
Martinovski, B. (2006) Framework for analysis of mitigation in courts. Journal of
Pragmatics, 38(12), 2065–2086.
Mauranen, A. (1997). Hedging in language revisers’ hands. In R. Markkanen &
H. Schroder (eds.), Hedging and discourse: Approaches to the analysis of pragmatic
phenomenon in academic texts (pp. 115–133). Berlin: Walter de Gruyter.
McCarthy, M. (1998). Spoken language and applied linguistics. Cambridge:
References 205
McMillan, J. R., Clifton, A. K., McGrath, D., & Gale, W. S. (1977). Women’s lan-
guage: Uncertainty or interpersonal sensitivity and emotionality. Sex Roles, 3,
545–559.
Metsä-Ketelä, M. (2006). Words are more or less superfluous: The case of more
or less in academic Lingua Franca English. Nordic Journal of English Studies,
5(2), 117–143. Retrieved from http://ojs.ub.gu.se/ojs/index.php/njes/article/
view/72/76
Metsä-Ketelä, M. (2012). Frequencies of vague expressions in English as an aca-
demic lingua franca. Journal of English as a Lingua Franca, 1(2), 263–285.
Morse, J. M. (1991). Approaches to qualitative-quantitative methodological trian-
gulation. Nursing Research, 40, 120–123.
Moxey, L. M., & Sanford, A. J. (1993). Communicating quantities: A psychological
perspective. Hove: Lawrence Erlbaum.
Mulac, A., & Lundell, T. L. (1986). Linguistic contributors to the gender-linked
language effect. Journal of Language and Social Psychology, 5, 81–101.
Mulac, A., Lundell, T. L., & Bradac, J. (1986). Male/female language differences
and attributional consequences in a public speaking situation: Toward an
explanation of the gender-linked language effect. Communication Monographs,
53, 115–129.
Mulac, A., Wiemann, J. M., Widenmann, S. J., & Gibson, T. W. (1988). Male/
female language differences and effects in same-sex and mixed-sex dyads: The
gender-linked language effect. Communication Monographs, 55, 315–335.
Myers, G. (1989). The pragmatics of politeness in scientific articles. Applied
Linguistics, 10, 1–35.
Myers, G. (1996). Strategic vagueness in academic writing. In E. Ventola &
A. Mauranen (eds.), Academic writing:Intercultural and textual issues (pp. 1–18).
Amsterdam: John Benjamins.
Neuman, W. L. (1997). Social research methods: Qualitative and quantitative
approaches. Boston: Allyn & Bacon.
Niezgoda, K., & Rover, C. (2001). Pragmatic and grammatical awareness: A func-
tion of the learning environment. In K. R. Rose & G. Kasper (eds.), Pragmatics
in language teaching (pp. 63–79). Cambridge: Cambridge University Press.
Nikula, T. (1996). Pragmatic force modifiers: A study in interlanguage pragmatics.
Jyvaskyla: University of Jyvaskyla.
O’Shea, M. (2000). Cultural shock: Iran. Portland: Graphic Arts Publishing Company.
Ohta, A. S. (2001). Second language acquisition processes in the classroom: Learning
Japanese. Mahwah: Lawrence Erlbaum Associates.
Overstreet, M. (1995). The form and function of general extenders in English interna-
tional discourse. (Unpublished PhD thesis). The University of Hawaii, The USA.
Overstreet, M. (2005). And stuff und so: Investigating pragmatic expressions in
English and German. Journal of Pragmatics, 37, 1845–1864.
Parvaresh, V., Tavangar, M., & Eslami Rasekh, A. (2010). General extenders in
Persian discourse: Frequency and grammatical distribution. Cross-Cultural
Communication, 6(3), 18.
Parvaresh, V., Tavangar, M., Eslami Rasekh, A., & Izadi, D. (2012). About his
friend, how good she is, and this and that: General extenders in native Persian
and non-native English discourse. Journal of Pragmatics, 44, 261–279.
Parvaresh, V., & Tayebi, T. (2014). Vaguely speaking in Persian. Discourse
Processes, 51(7), 565–600.
206 References
Patton, M. Q. (1990). Qualitative evaluation and research methods (2nd ed.).

Newbury Park: Sage Publications, Inc.
Pearson, J. (1998). Terms in context. Amsterdam: John Benjamins.
Plato. (1914). Euthyphro. London: William Heinemann.
Pomerantz, A. (1978). Compliment responses: Notes on co-operation of multiple
constraints. In J. Schenkein (ed.), Studies in organization of conversational interac-
tion (pp. 79–112). New York: Academic Press.
Prince, E. F., Frader, J., & Bosk, C. (1982). On hedging in physician-physician
discourse. In R. Di Pietro (ed.), Linguistics and the professions (pp. 83–97).
Norwood: Ablex.
Reichardt, C. S., & Cook, T. D. (1979). Beyond qualitative versus quantitative
methods. In T. D. Cook & C. S. Reichardt (eds.), Qualitative and quantitative
methods in evaluation research (pp. 7–32). Beverly Hills: Sage.
Ringbom, H. (1998). Vocabulary frequencies in advanced English: A cross-
linguistic approach. In S. Granger (ed.), Learner English on computer (pp. 41–52).
London: Longman.
Römer, U. (2004). A corpus-driven approach to modal auxiliaries and their didac-
tics. In J. Sinclair (ed.), How to use corpora in language teaching. Amsterdam:
Benjamins.
Romero Trillo, J. (2002). The pragmatic fossilization of discourse markers in non-
native speakers of English. Journal of Pragmatics, 34, 769–784.
Rosch, E. H. (1973). Natural categories. Cognitive Psychology, 4, 328–350.
Rossman, G. B., & Wilson, B. L. (1985). Numbers and words: Combining
quantitative and qualitative methods in a single large-scale evaluation study.
Evaluation Review, 9(5), 627–643.
Rowland, T. (2007). Well maybe not exactly, but it’s around fifty basically?:
Vague language in mathematics classrooms. In J. Cutting (ed.), Vague language
Russell, B. (1923). Vagueness. The Australian Journal of Philosophy, 1(2), 84–92.
Rutter, B. (2008). Acoustic properties of repair sequences in dysarthric conversational
speech: An interactional phonetic study. Doctoral Dissertation, University of
Louisiana at Lafayette, Louisiana, USA.
Ruzaitė, J. (2004). Academic precision reconsidered: A corpus-based account. Sky
Journal of Linguistics, 17, 217–247.
Ruzaitė, J. (2007). Vague language in educational settings: Quantifiers and approxima-
tors in British and American English. Frankfurt: Peter Lang.
Sacks, H., Schegloff, E. A., & Jefferson, G. (1974). A simplest systematics for the
organisation of turn-taking for conversation. Language, 50, 696–735.
Schegloff, E. A. (1993). Reflections on quantification in the study of conversa-
tion. Research on Language and Social Interaction, 26, 99–128.
Schegloff, E. A., Jefferson, G., & Sacks, H. (1977). The preference for self-correction
in the organization of repair in conversation. Language, 53(2), 361–382.
Searle, J. R. (1968). Austin on locutionary and illocutionary acts. Philosophical
Review, 77, 405–424.
Sharifian, F. (2003). On cultural conceptualisations. Journal of Cognition and
Culture, 3(3), 187–207.
Sharifian, F. (2005). The Persian cultural schema of shekasteh-nafsi: A study of
compliment responses in Persian and Anglo-Australian speakers. Pragmatics
and Cognition, 13(2), 337–361.
References 207
Sharifian, F. (2007). L1 cultural conceptualisations in L2 learning. In F. Sharifian &

G. B. Palmer (eds.), Applied cultural linguistics: Implications for second language
learning and intercultural communication. Amsterdam: John Benjamins.
Sharifian, F., & Lotfi, A. R. (2003). Rices and waters: The mass/count distinction
in Modern Persian. Anthropological Linguistics, 45(2), 226–244.
Shirinbakhsh, S., & Eslami Rasekh, A. (2013). The effect of age on cultural
schema: The case of Shekaste-nafsi (modesty) in Persian. International Journal of
Research Studies in Language Learning, 2(3), 95–107. Retrieved from http://www.
consortiacademia.org/index.php/ijrsll/article/view/151/159.
Simon-Vandenbergen, A. M. (2000). The functions of I think in political dis-
course. International Journal of Applied Linguistics, 10(1), 41–63.
Soter, A. (1988). The second language learner and cultural transfer in narration.
In A. Purves (ed.), Writing across languages and cultures: Issues in contrastive rheto-
ric (pp. 177–205). Newbury Park: Sage.
Sperber, D., & Wilson, D. (1986a). Loose talk. Proceedings of the Aristotelian Society,
86, 153–171.
Sperber, D., & Wilson, D. (1986b). Relevance: Communication and cognition.
Oxford: Blackwell.
Sperber, D., & Wilson, D. (1995). Relevance: Communication and cognition (2nd ed.).
Oxford: Blackwell.
Stenström, A. B. (1994). An introduction to spoken interaction. London: Longman.
Stenström, A. B. (1995). Some remarks on commentclauses. In B. Aarts &
C. F. Meyer (eds.), The verb in contemporary English (pp. 290–301). Cambridge:
Stubbe, M., & Holmes, J. (1995). You know, Eh and other ‘exasperating expres-
sions’: An analysis of social and stylistic variation in the use of pragmatic
devices in a sample of New Zealand English. Language and Communication,
15(1), 63–88.
Sweller, J. (1988). Cognitive load during problem solving: Effects on learning.
Cognitive Science, 12, 257–285.
Tannen, D. (1989). Talking voices. Cambridge: Cambridge University Press.
Tarnyikova, J. (2009). Vague reference to national categories. Anglica III
Linguistica, 3, 115–132.
Teigen, K. H. (1990). To be convincing or to be right: A question of preciseness.
In K. J. Gilhooly, M. T. G. Keane, R. H. Logie & G. Erdos (eds.), Lines of thinking:
Reflections on the psychology of thought (pp. 299–313). Chichester: Wiley.
Terraschke, A. (2008). The use of pragmatic devices by German non-native
speakers of English. (Unpublished PhD dissertation). Victoria University of
Wellington,Wellington, New Zealand.
Terraschke, A., & Holmes, J. (2007). ‘Und Tralala’: Vagueness and general extend-
ers in German and New Zealand English. In J. Cutting (ed.), Vague language
Thomas, J. (1995). Meaning in interaction: An introduction to pragmatics. London:
Longman.
Tracy, K. (1990). The many faces of facework. In H. Giles & P. Robinson (eds.),
Handbook of language and social psychology (pp. 209–226). Chichester: Wiley.
Tran, G. Q. (2006). The naturalized role-play: An innovative methodology in
cross-cultural and interlanguage pragmatics research. Reflections on English
Language Teaching, 5(2), 1–24.
208 References
Trappes-Lomax, H. (2007). Vague language as a means of self-protective avoid-

ance: Tension management in conference talks. In J. Cutting (ed.), Vague lan-
guage explored (pp. 117–137). Basingstoke: Palgrave Macmillan.
Travers, R. M. W. (1992). History of educational research. In M. C. Alkin
(ed.), Encyclopedia of educational research (6th ed., pp. 384–389). New York:
Macmillan.
Ullman, S. (1962). Semantics. Oxford: Blackwell.
Walsh, S. (2006). Investigating classroom discourse. New York: Routledge.
Walsh, S., O’Keeffe, A., & McCarthy, M. (2008). ‘Post-colonialism, multi-
lingualisam, structuralism, feminism, post-modernism and so on and so
forth’: A comparative analysis of vague category markers in academic dis-
course. In A. Adel & R. Reppen (eds.), Corpora and discourse: The challenges of
different settings ( pp. 9–29). Amsterdam: John Benjamins.
Wiese, R. (1984). Language production in foreign and native languages: Same
or different. In H. W. Dechert, D. Mohle & M. Raupach (eds.), Second language
productions (pp. 11–25). Tubingen: Gunter Narr.
Williamson, T. (1994). Vagueness. In R. Asher & J. M. Y. Sipmson (eds.), The ency-
clopedia of language and linguistics (pp. 4869–4871). Oxford: Pergamon Press.
Wolfson, N. (1981). Compliments in cross-cultural perspectives. TESOL Quarterly,
15(2), 117–124.
Wooffitt, R. (2005). Conversation analysis and discourse analysis: A comparative and
critical introduction. London: Sage.
Wright, H. J., & Hosman, L. A. (1983). Language style and sex bias in the court-
room: The effects of male and female use of hedges and intensifiers on impres-
sion formation. Southern Speech Communication Journal, 48, 137–152.
Wu, T. P. (1979). Mohu yuyan chutan (in Chinese) [Fuzzy Language (I)]. Waiguoyu
[Journal of Foreign Language], 4, 39–44.
Wu, Y., Wang, J., & Cai, Z. (2010). The use of I think by Chinese EFL learners:
A study revisited. Chinese Journal of Applied Linguistics, 33(1), 3–23.
Zadeh, L. A. (1965). Fuzzy sets. Information and Control, 8, 338–353.
Zarei, G. R., & Mansoori, S. (2007). Metadiscourse in academic prose: A contras-
tive analysis of English and Persian research articles. The Asian ESP Journal,
3(2), 24–40.
Zhang, G. (1998/2004, 2nd edition). Mohu Yuyixue (in Chinese) [Fuzzy Semantics].
Beijing: China Social Sciences Press.
Zhang, Q. (1998). Fuzziness-vagueness-generality-ambiguity. Journal of Pragmatics,
29(1), 13–31.
Zhang, Q. (2005). Fuzziness and relevance theory. Foreign Language and Literature
Studies, 2, 73–84.
Zhang, G. (2011). Elasticity of vague language. Intercultural Pragmatics, 8(4),
571–599.
Zhang, G. (2014). The elasticity of I think: stretching its pragmatic functions.
Intercultural Pragmatics, 11(2): 225–257.
Zhang, G., & Sabet, P. G. P. (in press). Elastic ‘I think’: stretching over L1 and
L2. Applied Linguistics, Advance Access published July 1, 2014, http://applij.
oxfordjournals.org
Zhang, G. (2015). Elastic language: How and why we stretch our words? Cambridge:
Subject Index
A go hypothetical 31
aberou 171 go just-right 30
accommodation strategy 12 go subjective 31
adaptors 123
approximation 45 H
attribution shield 115 hesitation markers 143
B I
boosters 129 imprecision 4
incomprehensible language 7
C indirect language 6
classical set theory 5 indirectness 7
cluster 47 inference 27
cognitive effort 27
cognitive load 175 L
common categories 9 Lexical categories 9
contrary-to-face-value (CTFV) 173 lexical density 17
conversation analysis 35, 36 listener-oriented approach 158,
conversational implicature 25 180, 181
cooperative principle 23, 24 local vagueness 155, 156
cultural protocols 169 loose talk 28, 29
D
M
detensifiers 123
managerial mode 15
discourse management 55, 153
material mode 15
discourse markers 143
mitigating 114
downtoning 45
mitigation 45, 114
downtowners 123
mixed methods 38, 39
mode 15
E
modifying devices 18
elasticity 5, 29, 30, 181, 182, 186, 191
multal quantifiers 130
embedded design 39
emphasising 114
N
empirical science 34
negative face 118
epistemic phrases 47
Non-lexical categories 9
equivalent status design 39
F O
face 118 Ostension 27
fuzzy set theory 5
P
G Parallel/simultaneous studies 39
global vagueness 155 placeholder 44, 99, 158
go general 30 plausibility shield 8, 115
209
210 Subject Index
politeness 45, 118 sequential studies 39

positive face 118 shekaste-nafsi 171
possibility 45 slingshot 30
possibility indicators 44, 58 sociolinguistic competence 18
post positivist paradigm 34 speaker-dominated approach 180
pragmatic proficiency 18 subjectivizers 44, 47
pragmatism 38
processing effort 27 T
prototype theory 6 taarof 171, 190
triangulation 39, 40
Q turn management 45
qualitative approach 35
quantification 45, 126 V
quantitative research 33, 34 vague categories 9
vague category identifiers 9
R vague category markers 15
relevance theory 26, 27, 175 vague expressions 23, 29
repair 45, 139 vague intensifiers 44, 84
right amount of information 45 vague language 1, 4, 5, 6,
8, 11
S vague quantifiers 44, 69
scientific method 34 vague tags 19
self-defensive 115 vague work 30
self-protection 45, 115 vagueness 4, 5, 6, 9, 29, 138
Author Index
A Crozet, C. 2
Abdollahzadeh, E. 14, 169 Crystal, D. 8, 9
Adolphs, S. 2, 11 Cummings, M. C. 41
Aijmer, K. 49, 119 Cutting, J. 1, 9, 19, 20, 26, 28,
Alcon, E. 18, 19 69, 113
Allison, D 22,
Alwood, J 146 D
Aristotle 1 Davies, B. L. 25
Atai, M. R. 163 Davy, D. 8, 9
Atkins, S. 2 De Cock, S. 19
Austin, J. L 6 Deese, J. 9
Drave, N. 20, 77, 81, 113, 152
B Dudley-Evans, T. 22
Bachman, L. F. 18
Badovi-Harlig, K. 18 E
Behnam, B. 160 Erev, I. 8
Beighmohammadi, A. 14 Eslami Rasekh, A. 13, 171
Biber, D. 4, 23, 49
Billig, M. 37 F
Blum-Kulka 44 Finegan, E. 5
Bosk, C. 8 Firestone, W. 33
Bradac, J. 129 Flick, U. 34
Brazil, D. 5, 23 Frader, J. 8
Brown, P. 8, 18, 22, 118, 119 Franken, N. 29
C G
Cai, Z. 22 Grice, H. P. 6, 23, 24, 25, 26, 31,
Carter, R. 6 26, 136
Chan, A. Y. 173 Griffin, R. 18
Channell, J. 1, 4, 7, 8, 9, 11, 20, 25, Gudykunst, W. B. 157
26, 44, 45, 69, 99, 113, 119, 126,
136, 152 H
Chen, W. Z. 13 Hanson, H. F. 40
Cheng, W. 3, 4, 6, 7, 11, 12, 17, 18, Harvey, K. 2
19, 20, 27, 45, 152 Heider, E. R. 6
Cheshire, J. 13 Holmes, J. 18, 22, 129,
Chow, K. F. 176 143, 169
Connor, U. 22 Hosman, L. A. 129
Conrad, S. 5, 49 Huang, Y. 24, 25, 26
Cook, G. 5, 23, 33 Hübler, A. 123
Cotterill, J. 2, 11 Hutchby, I. 35, 36, 37
Creswell, J. W. 33, 34, 35, 38, 39, Hyland, K. 5, 16, 17, 21, 22, 23,
40, 41 68, 129
211
212 Author Index
I Niezgoda, K. 2
Izadi, D. 13 Nikula, T. 19, 21, 118, 119, 190
Niroomand, M. 160
J
James, A. R. 18 O
Janicki, K. 7 O’Keeffe, A. 9
Jefferson, G. 36, 139 O’Shea, M. 171
Jick, T. D. 40 Ohta, A. S. 2
Johansson, S. 4 Overstreet, M. 8, 13
Johnson, M. 1
Jucker, A. H. 9, 10, 23, 28, 29, 113, P
114, 123, 138 Parvaresh, V. 13, 14
Patton, M. Q. 38
K Pearson, J. 123
Kaltenböck, G. 49 Plano Clark, V. L. 40, 41
Kärkkäinen, E. 47, 48 Plato 1
Khurshudyan, V . 45, 143 Pomerantz, A. 18
Koester, A. 2 Prince, E. F. 8, 11, 44, 45, 114, 115,
Koutlaki, S. A. 171 123, 132, 139
L R
Lado, R. 162 Reichardt, C. S. 33
Lakoff, G. 1, 5, 11 Remero Trillo, J. 180
Leech, G. 4 Ringbom, H. 20, 189
Levinson, S. C. 6, 8, 18, 26 Roomer, U. 164
Liddicoat, A. J. 2, 36, 37 Rosch, E. H. 6
Lotfi, A. R. 166, 170 Rossman, G. B. 38
Lüdge, T. 9 Rowland, T. 15
Lundell, T. L. 129 Russell, B. 5
Rutter, B. 139
M Ruzaitė, J. 2, 4, 7, 8, 11, 12, 14, 44,
Ma, R. 181 45, 47, 67, 69, 71, 83, 113, 114,
Mansoori, S. 157, 172 119, 123, 130, 138, 139
Martinez-Flor, A. 19 Röver, C. 2
Martinovski, B. 114
Mauranen, A. 184 S
McCarthy, M. 6, 9, 23 Sabet, P. 47, 49, 51, 53, 175,
McMillan, J. R. 129 178, 179
Metsä-Ketelä, M. 8, 10, 21, 152 Sacks, H. 36, 139
Milton, J. 21, 22, 23, 68 Sadr, L. 163
Minoe, P. J. 8 Sanford, A. 10
Morse, J. M. 39 Schegloff, E. A. 36, 139
Moxey, L. M. 10 Searle, J. R. 6
Mulac, A. 129 Sharifian, F. 166, 170, 171
Mulan, A. 129 Simon-Vandenbergen, A. M. 49
Myers, G. 5, 8, 17 Smith, A. 9
Soter, A. 22
N Sperber, D. 23, 25, 26, 27, 28, 31
Neal, m. M. 8 Stenström, A. B. 49
Neuman, W. L. 35 Sweller, J. 175
Author Index 213
T Wang, J. 21
Tannen, D. 10 Warren, M. 3, 4, 6, 7, 12, 17,
Tarnyikova, J. 4 18, 19, 27, 152
Tavangar, M. 13 Wiese, R. 143
Tayebi, T. 14 Williamson, T. 6
Teigen, K. H. 10 Wilson, D. 23, 25, 26, 27, 28,
Terraschke, A. 42, 169 29, 31
Thomas, J. 18 Wolfson, N. 18
Thompson, S. 129 Wooffitt, R. 36, 37
Tracy, K. 118 Wright, H. J. 129
Tran, G. Q. 41 Wu, Y. 13, 22, 123, 142,
Trappes-Lomax, H. 115 143, 161
Travers, R. M. W. 34
Z
U Zadeh, L. A. 5, 6, 12, 14, 28
Ullman, S. 8 Zarei, G. R. 157, 172
Zhang, G. 1, 7, 13, 23, 25,
W 28, 31, 44, 45, 49, 51, 113,
Wallston, T. S. 8 152, 155, 161, 175, 178,
Walsh, S. 9, 15, 16 181, 188

Peyman G. P. Sabet, Grace Q. Zhang (Auth.) - Communicating Through Vague Language - A Comparative Study of L1 and L2 Speakers-Palgrave Macmillan UK (2015)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Peyman G. P. Sabet, Grace Q. Zhang (Auth.) - Communicating Through Vague Language - A Comparative Study of L1 and L2 Speakers-Palgrave Macmillan UK (2015)

Uploaded by

Copyright:

Available Formats

Communicating through Vague Language

Also by Peyman Sabet

Also by Grace Zhang

Peyman G.P. Sabet and Grace Q. Zhang

Typeset by MPS Limited, Chennai, India.

List of Figures and Tables viii

4.1.3 I don’t know 54

5.2 Right amount of information 126

Appendix I: Consent Form for the Director and Teachers 198

Appendix II: Consent Form for Participants 199

Subject Index 209

Author Index 211

4.1 Frequency distribution of subjectivizers 56

3.1 Differences between quantitative and qualitative methods 35

4.6 Might clusters 65

We would like to express our sincere gratitude to Chris Conlan, the

CANCODE Cambridge and Nottingham Corpus of

Vague language (VL) is an integral part of language and has an essential

to give depth to the role of vague language in language learning and

1.1 Purpose of the study

Learners of English tend to use vague language at higher or lower

The objectives of this study are to explore vague language realization

This book consists of seven chapters structured as follows: Chapter 2

Vague language, as an increasingly explored phenomenon in language,

2.1 Vague language: what is it and why do we need it?

Early work on vague language is associated with Bertrand Russell (1923),

supports Zadeh’s theory, as ‘Clearly any attempt to limit truth condi-

giving the right amount of information,

a. memory loss—the speaker forgets the correct word;

Vague language is primarily a strategic phenomenon with important

2.2 Vague language in different settings

Vague language plays different roles in different settings. Focusing

into Chinese linguistics. Wu argues that the concept of membership can

frequency in the Persian corpus. This present study employs a similar

these countries, cultural differences may cause discrepancies in their

2.3 Vague language and education

The vague language works relating to education is reviewed in three

2.3.1 Vague language in the classroom

teaching material or input used, where students answer questions and

tries to keep fact and interpretation balanced. Vague language is an

a. Vagueness in statements of results allows them to be compared to

A good command of vague language is thus a feature of a proficient L2

2.3.2 Vague language in language learning and teaching

vocabulary of the first language, differences in conversational rules

role of vague language in improving learners’ sociolinguistic compe-

2.3.3 Vague language and learner language

in the classroom may be stigmatized); and finally, lack of contact with

Metsä-Ketelä (2006) investigates the use of vague language by non-

cultural backgrounds, language learners display differences in express-

In research on the use of expressions of doubt and uncertainty by L1s

2.4 Theoretical frameworks

Vague language plays an indispensable part in communication.

2.4.1 Vague language and the cooperative principle

be as cooperative as possible. The idea is to ‘Make your conversational

Grice states that conversational implicature is a result of flouting the

When a speaker ostensibly flouts maxims, the hearer may either

utterance creates no barriers to hearers; in fact, neither speakers nor

2.4.2 Vague language and Relevance Theory

truth, whereas others are simply ignored. Thus the truth–conditional

2.4.3 The elasticity of vague language

language as a versatile strategy at interlocutors’ disposal for effective

This is very important. (upward)

This is a bit embarrassing. (downward)