Metalanguage in Interaction

Pragmatics & Beyond New Series (P&BNS)

Pragmatics & Beyond New Series is a continuation of Pragmatics & Beyond and
its Companion Series. The New Series offers a selection of high quality work
covering the full richness of Pragmatics as an interdisciplinary field, within
language sciences.

Editor
Andreas H. Jucker
University of Zurich, English Department
Plattenstrasse 47, CH-8032 Zurich, Switzerland
e-mail: [email protected]

Associate Editors
Jacob L. Mey Herman Parret Jef Verschueren
University of Southern Belgian National Science Belgian National Science
Denmark Foundation, Universities of Foundation,
Louvain and Antwerp University of Antwerp

Editorial Board
Shoshana Blum-Kulka Susan C. Herring Emanuel A. Schegloff
Hebrew University of Indiana University University of California at Los
Jerusalem Angeles
Masako K. Hiraga
Jean Caron St.Paul’s (Rikkyo) University Deborah Schiffrin
Université de Poitiers Georgetown University
David Holdcroft
Robyn Carston University of Leeds Paul Osamu Takahara
University College London Kobe City University of
Sachiko Ide
Foreign Studies
Bruce Fraser Japan Women’s University
Boston University
Catherine Kerbrat- Sandra A. Thompson
University of California at
Thorstein Fretheim Orecchioni
Santa Barbara
University of Trondheim University of Lyon 2
Teun A. van Dijk
John C. Heritage Claudia de Lemos
Pompeu Fabra, Barcelona
University of California at Los University of Campinas, Brazil
Angeles Richard J. Watts
Marina Sbisà
University of Berne
University of Trieste

Volume 181
Metalanguage in Interaction. Hebrew discourse markers
by Yael Maschler
Metalanguage in Interaction
Hebrew discourse markers

Yael Maschler
University of Haifa

John Benjamins Publishing Company

Amsterdamâ•›/â•›Philadelphia
 TM The paper used in this publication meets the minimum requirements of
8

American National Standard for Information Sciences – Permanence of

Paper for Printed Library Materials, ansi z39.48-1984.

Library of Congress Cataloging-in-Publication Data

Maschler, Yael.
€ Metalanguage in interaction : Hebrew discourse markers / Yael Maschler.
€€€€€€ p. cm. (Pragmatics & Beyond New Series, issn 0922-842X ; v. 181)
Includes bibliographical references and index.
1.€ Hebrew language--Discourse analysis. 2.€ Discourse markers. 3.€ Hebrew language--
Spoken Hebrew--Israel.€ I. Title.
PJ4752.M37 â•…â•… 2009
492.4'0141--dc22 2009004733
isbn 978 90 272 5426 9 (hb; alk. paper)
isbn 978 90 272 8950 6 (eb)

© 2009 – John Benjamins B.V.

No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any
other means, without written permission from the publisher.
John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands
John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa
 In memory of my deeply beloved sister
Dorit Bahir Maschler
Jerusalem 8.5.1972 – Singapore 30.8.2000
 Table of contents

Transcription conventions xi
Preface xiii
Acknowledgments xv

chapter 1.╇ Introduction

Metalanguage in interaction: Discourse markers as a system 1
1. Metalanguagingâ•… 1
2. Language alternation at discourse markersâ•… 2
2.1 Relation to previous approaches to discourse markersâ•… 7
3. The dataâ•… 9
4. Distributional patterning: Negotiating frame shifts via discourse
markersâ•… 10
5. Defining discourse markersâ•… 16
5.1 Previous approaches to prosody in defining discourse
markersâ•… 20
6. Functional patterning of discourse markersâ•… 21
6.1 Previous approaches to realms of operation of discourse
markersâ•… 26
7. Structural patterning of discourse markersâ•… 27
8. Grammaticization of discourse markersâ•… 33
8.1 Traugott’s theory of grammaticization and semantic changeâ•… 37
9. Interacting as an Israeliâ•… 39

chapter 2.╇ The interpersonal realm

The discourse marker nu: Israeli Hebrew impatience in interaction 41
1. Introductionâ•… 41
2. Dataâ•… 45
3. Hastening co-participant to get on with some non-verbal actionâ•… 46
4. Nu urging further development within a topicâ•… 52
4.1 Narrational discourseâ•… 52
4.1.1 Urging a move on to the next episode of a narrativeâ•… 52
4.2 Urging further development within a non-narrational topicâ•… 55
4.3 Nu as a continuerâ•… 59
 Metalanguage in Interaction

4.4 Summary of the ‘urging further development of topic’

functionâ•… 64
5. Granting permission to perform an actionâ•… 64
6. Summary of sequential functions of nuâ•… 68
7. Keying functions of nu: joking/provokingâ•… 69
7.1 From joking to provokingâ•… 69
7.2 Nu in combination with other discourse markersâ•… 71
7.2.1 Nu in combination with 'az ma? (‘so what?’)â•… 72
8. Grammaticization of nuâ•… 74

chapter 3.╇ The textual realm

The discourse marker bekitsur: Retroactively constructing digressions 79
1. Introductionâ•… 79
2. Dataâ•… 84
3. Summarizing bekitsurâ•… 85
3.1 Summary following a listâ•… 85
3.2 Summary of episodeâ•… 87
3.3 Drawing conclusions from episodeâ•… 91
4. Resumptive bekitsur: Returning to the main topicâ•… 95
4.1 Collaboration in returning to the main topicâ•… 96
4.2 Returning from interaction-based digressionsâ•… 103
4.3 Returning from content-based digressionsâ•… 104
4.3.1 Following listener-initiated digressionsâ•… 104
4.3.2 Following storyteller-initiated digressionsâ•… 106
4.3.2.1 Following a story asideâ•… 106
4.3.2.2 Following orientationâ•… 108
4.3.2.3 Following evaluationâ•… 110
4.3.2.4 Organizing the hierarchy of episodes in a storyâ•… 112
4.4 Returning from insertion sequencesâ•… 114
5. Foregrounding bekitsur: Introducing a new narrativeâ•… 116
6. Grammaticization of bekitsurâ•… 119
7. Interacting as an Israeli via bekitsurâ•… 124

chapter 4.╇ The cognitive realm

The discourse marker ke'ilu: Realizing the need to rephrase 127
1. Introductionâ•… 127
2. Data and methodologyâ•… 132
3. A quantitative perspective on the different functions of ke'iluâ•… 133
4. Functional distribution of ke'ilu: A qualitative perspectiveâ•… 134
4.1 Ke'ilu as a conjunction in the literal senseâ•… 134
 Table of contents 

4.2 Ke'ilu as hedgeâ•… 137

4.2.1 Post-positioned ke'ilu as hedgeâ•… 137
4.2.2 Pre-positioned ke'ilu as hedgeâ•… 138
4.2.3 Interpersonal constraints on hedging ke'iluâ•… 140
4.3 Ke'ilu as focus markerâ•… 143
4.3.1 Pre-positioned ke'ilu as focus markerâ•… 145
4.3.2 Post-positioned ke'ilu as focus markerâ•… 147
4.3.3 Post-positioned focus-marking ke'ilu in aposiopesisâ•… 148
4.3.4 Interpersonal constraints on focus-marking ke'iluâ•… 150
4.4 Ke'ilu as discourse marker of self-rephrasalâ•… 152
4.4.1 Ke'ilu in self-rephrasals of clarification sequencesâ•… 154
4.4.2 Self-rephrasal ke'ilu tokens as fillersâ•… 157
4.5 Ke'ilu in quotationsâ•… 158
4.5.1 Ke'ilu in double-voiced ironic quotationsâ•… 160
5. The functional itinerary of ke'iluâ•… 162
6. Grammaticization of ke'ilu in cross-linguistic perspectiveâ•… 165
7. Interacting as an Israeli via ke'iluâ•… 166

chapter 5.╇ Between realms

The discourse marker tov: Accepting while shifting 171
1. Introductionâ•… 171
2. Dataâ•… 176
3. Interpersonal tovâ•… 176
3.1 Acceptance of some state of thingsâ•… 177
3.2 Third turn receiptâ•… 180
3.3 Concessionâ•… 181
3.4 Ironic agreement: Disagreementâ•… 185
4. Textual tov: Marking expected transitionâ•… 188
4.1 Transitional tov into the beginning of a narrativeâ•… 188
4.2 Transitional tov between the episodes of narrative discourseâ•… 189
4.3 Transitional tov returning to an interrupted actionâ•… 190
4.4 Transitional tov ending a topic/actionâ•… 191
5. Ambiguous cases: Between interpersonal and textual tovâ•… 193
5.1 Third turn receipt + ending an actionâ•… 193
5.2 Third turn receipt + transition into following episode/actionâ•… 195
5.3 Acceptance of some state of things + transition
into following episode/actionâ•… 196
5.4 Concession + transition into following episode/actionâ•… 198
5.5 Concession and return to main topicâ•… 200
6. Grammaticization of tovâ•… 201
 Metalanguage in Interaction

chapter 6.╇ Concluding remarks

Grammaticization from interaction 207
1. The system of discourse markers permeating interactionâ•… 207
1.1 Functionâ•… 207
1.2 Metalaguaging at frame shiftsâ•… 209
1.3 Structureâ•… 210
2. Grammaticization of discourse markersâ•… 214
2.1 Grammaticization of nu, bekitsur, ke'ilu, and tovâ•… 214
2.2 Common threads and differences
among the grammaticization patterns studiedâ•… 220
2.3 Projectability and grammaticization of constructionsâ•… 225
3. Interacting as an Israeliâ•… 229

Bibliography 233

Author index 251

Subject index 253

 Transcription conventions

Each line denotes an intonation unit (Chafe 1994).

For Hebrew texts, each line is followed by an English gloss. Where this gloss is not
close enough to an English utterance, it is followed by a third line supplying a usu-
ally literal (but sometimes functional) translation.
Transcription basically follows Chafe 1994, with a few additions:
... – half second pause (each extra dot = another 1/2 second)
.. – perceptible pause of less than half a second
(3.22) – measured pause of 3.22 seconds
´ – primary stress
` – secondary stress
´´ – particularly marked primary stress
, – comma at end of line – continuing intonation (‘more to come’)
. – period at end of line – sentence final falling intonation
? – question mark at end of line – sentence final rising intonation, ‘appeal intonation’
(Du Bois et al.€1992).
! – exclamation mark at end of line – sentence final exclamatory intonation
ø – lack of punctuation at end of line – a fragmentary intonation unit, one which
never reached completion.
-- two hyphens – elongation of preceding vowel sound
square bracket to the left of two consecutive lines indicates
overlapping speech, two speakers talking at once
alignment such that the right of the top line
is placed over the left of the bottom line
indicates latching, no interturn pause
/??????/ – transcription impossible
/words within slashes/ indicate uncertain transcription
p – piano (spoken softly)
pp – pianissimo (spoken very softly)
mp – mezzo-piano (spoken fairly softly)
f – forte (spoken loudly)
ff – fortissimo (spoken very loudly)
mf – mezzo-forte (spoken fairly loudly)
 Metalanguage in Interaction

acc. – accelerando (speeding up)

ritard. – ritardando (slowing down)
cresc. – crescendo (progressively louder in volume)
decresc. – decrescendo (progressively softer in volume)
{in curly brackets} – t ranscriber’s comments concerning paralinguistics and pros-
ody which do not have an agreed upon symbol in this tran-
scription system.
[xxxxx] – m  aterial within square brackets in the gloss indicates exuberances of
translation (what is not there in the original (Ortega y Gasset 1959,
Becker 1982)).
' – uninverted quotation mark in the middle of a transliterated word indicates the
glottal stop phoneme.
’ – inverted quotation mark in the middle of a transliterated word indicates an
elided form (e.g., ts’xa instead of tsrixa (‘needs’, fem. sg.)).
Utterances under consideration are given in boldface.
 Preface

I remember making a point to myself as an eight-year-old coming to live in the

U.S. for a year: whenever you need some time to think in the midst of conversa-
tion, don’t say 'e--m, as they do in Israel; rather, say u--h or uhm. This will make
you sound more like the other kids in class.
Many years later, as an undergraduate student of classical Indo-European phi-
lology, certain words, such as Greek gar, men, de; Latin enim, autem, nam; Sanskrit
tadaa, tathaa; or Old English þaa, hwaet caught my attention in the various lan-
guages I was studying. They often appeared at key locations in the text, such as at
beginnings of stories or paragraphs. Each such word had several translations, usu-
ally into high-register Hebrew words. And perhaps most interestingly, my profes-
sors were a bit lost whenever I inquired about the relationships among the differ-
ent translations of a single such word and about its ‘exact meaning’. This is mirrored
by what Longacre has written about what he termed ‘mystery particles’: the “sen-
tential particles continue to defy analysis even at a relatively advanced stage of the
research” (1976:€468).
Later, as a graduate student working on a dissertation on Hebrew-English bi-
lingual discourse, this type of word and the ‘hesitation utterances’ from my child-
hood turned out to have a very interesting property in common. In the Hebrew-
English bilingual corpus I was studying, about 2 out of every 3 such utterances
involved the bilingual discourse strategy of language alternation (what has been
termed ‘code-switching’, a language metaphor I would like to avoid here) (Witt-
genstein 1958, Reddy 1979, Maschler 1988, 1991). Just as my Hebrew-English bi-
lingual informants often alternated from English to Hebrew when about to verbal-
ize such words as look, yeah, but, on the other hand, anyway, or first of all, so they
often switched to Hebrew when they were about to hesitate with an uh or an uhm.
It turned out that this pattern of language alternation showed up again and again
in other language contact situations (e.g., Brody 1987, Salmons 1990, Matras 1998,
and the collection of articles in Maschler 2000a).
The more linguistics I studied, the more phenomena seemed to be encapsu-
lated by these mystery particles. I decided to go back to my first language, Hebrew,
and investigate in depth these words, which by now had a name – ‘discourse mark-
ers’ (Schiffrin 1987).
 Metalanguage in Interaction

Research in the field of discourse markers has flourished in the past two dec-
ades. The goal of the present study is to investigate in depth these elements in
Hebrew discourse, both as individual utterances and as a system. I focus on how
such utterances may come about, on their functions in conversation, on what
processes of grammaticization (Hopper 1987, 1988) they may undergo, and on
what we can learn from them about Israeli culture and society.
Katriel has written several rich ethnographies of communication in Israeli cul-
ture (e.g., 1986, 1991, 1999, 2004). Her research focuses on key words in the contem-
porary Hebrew lexicon as a window onto Israeli culture. Through a close exploration
of the life of certain content-words in Israeli culture (e.g., dugri, lefargen, gibush, ki-
turim), she illuminates the essence of being Israeli. In this book, I am interested in
exploring the essence of interacting as an Israeli. Rather than focus on content-words,
then, I focus on function-words – the grammatical category of discourse markers –
which are employed for frame shifting in Hebrew talk-in-interaction, as will be
shown. I study the system of discourse markers punctuating spoken discourse be-
cause I view these markers as the backbone of interaction. The study thus provides a
bridge between the two disciplines of linguistics and communication.
The opening chapter of this book explores Hebrew discourse markers as a
system. Through this exploration, my own approach to discourse markers is un-
raveled and compared to some previous approaches. In the chapters that follow, I
focus on several particular markers, illuminating through them certain quintes-
sential aspects of Israeli society, identity, and culture. The final chapter ties to-
gether the patterns gleaned from the studies of the individual markers and deline-
ates my approach to the grammaticization of discourse markers.
I do not, of course, claim this to be a comprehensive study of all the discourse
markers of Hebrew – a task far too ambitious for the current project. Rather, it is
intended as an exploration into a variety of such markers, in the quest for un-
raveling the essence of interacting as an Israeli.
This study is informed by a variety of approaches to the study of discourse:
studies of intonation and spoken discourse (e.g., Chafe 1987, 1994, Du Bois et
al.€1992), work in the discourse functional tradition (e.g., Hopper 1979, Chafe 1980,
Becker 1984, Thompson 1984, Du Bois 1985), the methods and findings of conver-
sation analysis (CA) (e.g., Sacks, Schegloff, and Jefferson 1974, Jefferson 1993, Sacks
1992, Schegloff 1993, 1996a,b), and studies in linguistic anthropology (e.g., Becker
1979, Goffman 1981, Gumperz 1982, Katriel 1986, 1991, 1999, 2004, Tannen 1984,
1989). It contributes to a cross-language perspective on the grammaticization of
discourse markers (cf. Auer 1996, Traugott and Dasher 2002, Fleischman and
Yaguello 2004) and asks “how linguistic structures [discourse markers] and pat-
terns of use are shaped by, and themselves shape, interaction” (Selting and Couper-
Kuhlen 2001:€1). This book, then, is about Hebrew interactional linguistics.
 Acknowledgments

Metalanguage in Interaction: Hebrew discourse markers was inspired by the work of

many linguists to whom I feel deeply indebted. It was my great privilege to study
under the mentorship of Alton L. Becker. Although this took place over two dec-
ades ago, his humanistic linguistics, inseparable from his character and personal-
ity, sparkles brightly in my mind. Wallace Chafe taught me how to approach spo-
ken discourse and think about cognitive constraints. From Paul Hopper I learned
about grammaticization and emergent grammar, and also how to read Wittgen-
stein. I have also learned much from John Du Bois, Emanuel Schegloff, Deborah
Schiffrin, Deborah Tannen, Sandra Thompson, and Elizabeth Traugott, all of
whom I had the good fortune to study with at one or another of the Linguistic
Society of America summer institutes in 1985, 1987, and 2001.
I am grateful to the people who have read various parts and versions of this
book and provided many useful comments. They include Mira Ariel, Peter Auer,
Elizabeth Couper-Kuhlen, Cecilia Ford, Susanne Günthner, Paul Hopper, Tamar
Katriel, Deborah Schiffrin, Sandra Thompson, Elizabeth Traugott, and several
anonymous reviewers. Of course, I remain solely responsible for the use I have
made of all their comments. I also owe a debt to Suzanne Fleischman, who sadly
passed away in 2000. It was Suzanne who, back in the summer of 1999, on the ter-
race of the Israel Museum in Jerusalem, encouraged me to begin my first study of
a discourse marker in monolingual Hebrew conversation.
My colleagues and students at the University of Haifa have provided me with
an important context for research. I am especially grateful to Gonen Dori-Hacohen,
Roi Estlein, Tamar Katriel, Irit Meir, Carmit Miller, Bracha Nir, Chaim Noy, Hilla
Polak-Yitzhaki, Rivki Ribak, Michele Rosenthal, and Tamar Zewi for the intellec-
tually stimulating environment they create and for their friendship. Likewise, I am
grateful for the friendship of Jenny Mandelbaum and Susanne Uhmann.
Permission to update and expand two articles is gratefully acknowledged:
Maschler, Yael. 2002. On the grammaticization of ke‘ilu (‘like’, lit. ‘as if ’) in Hebrew
talk-in-interaction. Language in Society 31: 243–276. Cambridge: Cambridge Uni-
versity Press. Maschler, Yael. 2003. The discourse marker nu: Israeli Hebrew impa-
tience in interaction. Text 23: 89–128. Berlin: Mouton de Gruyter.
It has been a pleasure to work with the people at John Benjamins. Andreas
Jucker has been a wise and most efficient series editor. Isja Conen and Martine van
 Metalanguage in Interaction

Marsbergen from the production department have been professional, considerate,

and patient with me.
I am also extremely grateful to the Freiburg Institute for Advanced Studies
(FRIAS) at Albert-Ludwigs-Universität, Freiburg, Germany, for an External Sen-
ior Fellowship I received at the School of Language and Literature. This fellowship
provided me with an ideal setting in which to complete the very final touches on
this book.
In Israel, I have been especially lucky for close friendships, particularly Shari
and David Satran, Lilach and Eran Ishai, Ilan Yaniv and Daniela Yaniv Richter, and
Naomi De-Malach and David Blanc. I would also like to mention my gratitude to
Tamar Kron and David Wieler.
My parents have been a source of much support throughout the years. My
mother Hanna Maschler also proved to be an excellent proof-reader. It is my great
sorrow that my father Michael Maschler did not live long enough to see this book
in print. His genuine interest in my work has always been a source of much happi-
ness, and the example he set of a hard-working, devoted, and honest academic –
an inspiration.
I dedicate this book to the memory of my deeply beloved sister Dorit, whose
sudden untimely death will always darkly color our lives in many more ways than
we would have ever imagined.
My own family deserves the greatest acknowledgment. My children, Shira,
Maya, and Yotam, have not only provided a wealth of excerpts for analysis (some
of which have found their way into the pages of this book); they continue to be my
partners in unearthing Israeli ways of talking and being in the world, while consti-
tuting a source of great joy, pride, and meaning in life. Lastly, and most impor-
tantly, I owe an immense debt of gratitude to my husband and partner Paul Wood-
ward Inbar for listening to me talk about this book endlessly, for his sharp
linguistic insights, criticism and advice, for his sound editing, generosity, and care,
and most importantly, for his love and friendship.
 chapter 1

Introduction
Metalanguage in interaction:
Discourse markers as a system

1. Metalanguaging

I begin by elaborating on what is unique to my own approach to discourse markers.

There have been many studies of discourse markers over the past two decades,
most notably, Schiffrin 1987 (for overviews, see Brinton 1996, Jucker and Ziv 1998,
Schourup 1999, Maschler 2000a, Schiffrin 2001, Fischer 2006, Maschler 2008).
While there exist a few article-length studies of Hebrew discourse markers (e.g.,
Ariel 1998, Even-Zohar 1982, Henkin 1999, Livnat and Yatziv 2003, 2006, Shloush
1998, Ziv 1998, 2001), my own are exceptional in being based on naturally occur-
ring conversation (Maschler 1997a, 1998b, 2001, 2002 a,b, 2003, 2009). However,
the most unique aspect of the present approach is that it focuses on the process of
metalanguaging in relation to employment of discourse markers.
With the term ‘metalanguaging’, I wish to evoke a term coined by A. L. Becker, ‘lan-
guaging’. Becker moved from a notion of ‘language’ to a notion of ‘languaging’ in order
to shift from an idea of language as something accomplished to the idea of languaging as
an ongoing process (1988:€25). A similar shift is found in Hopper’s ideas about emergent
grammar (1987, 1988, see below), which he contrasts with a priori grammar.
Languaging is possible at two levels of discourse. Generally when we use lan-
guage, we look through it at a world we believe to exist beyond language. However,
we can also use language in order to look through it at the process of using language
itself (Maschler 1994b). I investigate here this latter process of metalanguaging –
using language in order to communicate about the process of using language. Met-
alanguaging, I argue, is the semantic-pragmatic process which is at the heart of
both the employment and grammaticization (Hopper 1987) of discourse markers.
Discourse markers are viewed here as linguistic elements employed for meta-
languaging – languaging about the interaction, as opposed to languaging about
the extralingual world. In other words, rather than referring to the world per-
ceived by speakers (through language) to exist beyond language, discourse markers
refer to the text itself, to the interaction among its speakers, or to the cognitive
processes taking place in their minds during verbalization.
 Metalanguage in Interaction

Every act of languaging involves always both levels – the lingual and the meta-
lingual1. On the one hand, the utterances we employ for metalanguaging almost
always (apart from very few utterances such as uh or uhm) refer also to something
in the extralingual world. On the other hand, languaging about the world involves
continually signaling the frame (Goffman 1974) through which our utterance is
intended; in other words, it involves metalanguaging. Languaging, then, always
happens at these two levels, but some utterances are particularly high in their me-
talingual dimension, and it is those utterances which form the focus of this book.
In this chapter, I first show that, from the emic perspective of participants in
an interaction, discourse markers constitute a distinct linguistic category em-
ployed for metalanguaging. I then move on to give a detailed definition of dis-
course markers based on their semantic and structural properties. I show that dis-
course markers constitute a system exhibiting three types of patterning involving:
(1) the moments at which discourse markers are employed in interaction, (2) the
functions fulfilled by discourse markers, and (3) their structural properties. I in-
troduce the topic of grammaticization of discourse markers, which will be ex-
plored for particular markers in the ensuing chapters and portrayed as a whole in
the concluding chapter, and I close with a brief overview of the particular dis-
course markers studied in depth in the remainder of this study.

2. Language alternation at discourse markers

My approach to discourse markers originates in a phenomenon I noticed many

years ago in bilingual conversation – language alternation at utterances which all
have something in common; namely, that they are high in the metalingual dimen-
sion and occur at conversational action (Ford and Thompson 1996) boundaries
(Maschler 1988, 1994b).
Examine, for instance, the following excerpt, typical of much bilingual conver-
sation. Shira and Grace, two bilingual relatives in their early thirties who emi-
grated to Israel from the U.S. 17 and 6 years prior to this conversation, respec-
tively, are discussing two children, a brother and sister (Gad and Nira) in their
family. They are in the midst of a discussion concerning whether boys are treated
differently than girls by their parents2:

1. I prefer the term ‘metalingual’ to ‘metalinguistic’, because we are concerned here with the
level meta to language, not to linguistics.
2. The larger conversation from which excerpts 1 and 2 are taken was analyzed in detail in
Maschler 1988.
 Chapter 1.╇ Introduction 

Excerpt 1:
1 Shira: ... they shòuld not be tréated so differently.
2 I’m sáying
3 Grace: áre they?
4 /really?/
5 .. réally?
6 Shira: .... kén (‘yeah’),
7 ... mà shebarúr (‘what’s for sure’) is that,
8 ... the totsa'á (‘result’) is vèry different.
9 Grace: yés.
10 .. but
11 Shira: 'axshav (‘now’) grá--nted,
12 .... I accé--pt,
13 ... that.. whát’s his name is a lìttle different. {referring to the boy Gad}
14 .. 'aval (‘but’)
15 ... kén (‘yeah’),
16 Grace: Gad?
17 Shira: gám (‘also’),
18 he doesn’t get the sá--me yaxas (‘attentiveness toward the other’),
19 ... as Níra does,
20 .. he gets dífferent
21 .. and sometimes
22 Grace: yeah but that’s also cause she’s ó--lder,
23 .. and hè’s yóu--nger,
24 Shira: bé--seder (‘alright’),
25 Grace: I think there’s something about,
26 ... that you treat a bóy dìfferently
27 .. do yóu treat
28 .. do you think yóu treat Yo'av {Shira’s son}
29 Shira: 'ani betuxá (‘I’m sure’),
30 ... that if hé was a gì--rl,
31 ... I would treat him dífferently.
32 ... 'ani betuxá (‘I’m sure’),
Apart from the nouns yaxas3 (‘attentiveness toward the other’, line 18) and totsa'á
(‘result’, line 8), two relatively highly culture-bound concepts in Israeli culture,

3. Katriel writes that yaxas (‘attentiveness toward the other’) is among the relational terms
which have greatly proliferated in recent years in Israeli society. “These terms denote interper-
sonal behavior that concretizes the notion of interpersonal support. The term yaxas refers to
humanizing gestures of attentiveness and considerateness” (2004:€206–207).
 Metalanguage in Interaction

perhaps even ‘key words’ (Katriel 1999), all the utterances accompanied by the
strategy of language alternation here are high in the metalingual dimension. In
these 32 mostly English intonation units (Chafe 1994), we find 9 such Hebrew ut-
terances. Some of them function metalingually in the realm of the text, signaling
relationships between prior and upcoming discourse: 'axshav (‘now’, line 11), 'aval
(‘but’, line 14), gam (‘also’, line 17). Others function metalingually in the interper-
sonal realm, negotiating relations between speaker and hearer (ken (‘yeah’, lines 6,
15)) or between speaker and text (mà shebarúr (‘what’s for sure’, line 7), 'ani betuxá
(‘I’m sure’, lines 29, 32)). One Hebrew utterance, the concessive béseder (‘alright’,
line 24), functions metalingually both interpersonally – granting the interlocutor
her point – as well as textually, indicating an upcoming utterance which will op-
pose prior discourse (here, that differences in the ways boys and girls are related to
have more to do with their sex than with their relative order in the family).
Yet another type of Hebrew metalingual utterance is illustrated in excerpt 2, in
which Grace compares working women with working men:
Excerpt 2:
1 Grace: ... do you think that they’re any léss ìsolated than mèn are.
2 .. now they have the sàme competítion at work.
3 ... they have the sàme láck of ti--me,
4 .. as a mán who wo--rks,
5 Shira: ze lo4 (‘it’s not’) have the sáme competition,
6 ... sòmetimes they have hárder situations at work,
7 Grace: okáy,
8 okáy,
9 Shira: ... but they ha--ve
10 'e-- (‘uh’).. agaí--n,
11 ... they há--ve a dífferent nétwork of suppórt.
{-------------ritard.---------------------}
Here we find another type of switched utterance, the Hebrew hesitation utterance
'e-- (‘uh’), at the return to a previous point, marked also by English again. Over and

4. Here we find another motivation for language alternation. The strategy is employed here in
an iconic manner (Becker 1982, Maschler 1993), in order to highlight a contrast in the discourse
by presenting the two contrasting segments, each one in a different language. Shira contrasts
her opinion with Grace’s, by repeating Grace’s English utterance (‘have the same competition’)
following its negation in Hebrew (ze lo – ‘it’s not’). The contrast between the two languages is
employed here not to highlight the contrast between the lingual and the metalingual (as in the
case of discourse markers), but rather in order to mirror the contrast between old and new in-
formation, and between Grace’s opinion vs. Shira’s (see Maschler 1997b for a detailed study of
this iconic strategy).
 Chapter 1.╇ Introduction 

over, we find Shira employing the Hebrew hesitation markers 'e-- and 'e--m (with
the front, half-open vowel [ε]) at conversational action boundaries. Only once in
a 40-minute conversation does she employ in this environment the mid-central
vowel, or the schwa [∂]; namely, English uh. This use is particularly prominent
because Shira, having learned Hebrew later in adolescence, generally speaks He-
brew with an accent very much influenced by English. As one of the most difficult
things to adopt when acquiring another language is nativelike usage of fillers, we
would expect to hear an American accent when employing fillers, not vice versa.
Chafe has shown that spoken discourse, as it unfolds, manifests evidence of
the cognitive processes taking place at the time in the mind of the speaker (1987,
1994). One notable phenomenon is that episode boundaries generally involve
longer pauses and hesitations, because more time is necessary at those moments of
discourse for the processing of greater amounts of new information. Shira’s hesita-
tion markers at conversational action boundaries can be viewed as metalingual at
the cognitive realm of talk, referring to (in the sense of attesting) the cognitive
processes taking place in her mind during verbalization. Just as Shira consistently
employs the strategy of language alternation at textual and interpersonal metalin-
gual utterances, so she employs it at cognitive metalingual utterances. Thus, Shira’s
employment of Hebrew fillers can be explained once we view these hesitation
markers as operating on a level meta to discourse, as in the case of the textual and
interpersonal metalingual utterances.
The two excerpts above illustrate three types of metalingual utterance – tex-
tual, interpersonal, and cognitive. We have seen that this classification is not always
clear-cut, as some utterances operate in more than one realm simultaneously. What
is common to all of these metalingual utterances, however, is that each one func-
tions to negotiate a frame shift (Goffman 1981), or a conversational action (Ford
and Thompson 1996) boundary in the interaction. The shift can be a fairly minute
one, such as the shift from a first pair part to a second in an adjacency pair (Sacks
1992, vol. 2: 521–570); for example, the shift from Grace’s assertion in excerpt 1,
lines 22–23 to Shira’s response to it in line 24, negotiated via the discourse marker
béseder (‘alright’, line 24). It could be a somewhat more pronounced frame shift, as
in a return to a previous discourse topic, such as ‘women’s different network of sup-
port’ in excerpt 2 (lines 10–11) negotiated via 'e-- (‘uh’) and English again. Alter-
natively, it could be a very major frame shift, such as the beginning of a completely
new narrative or discourse topic. We shall return to this point in Section 4.
We have seen that what is common to all these switched utterances is that they
are metalingual and occur at conversational action boundaries, or at frame shifts.
Bilingual discourse offers the unique possibility of iconically separating actions of
languaging from those of metalanguaging by performing each of these two types
of activity in a different language. The bilingual employs one of the two available
 Metalanguage in Interaction

languages in order to comment on, or manage, interactions taking place mostly in

the other. The result is a discourse taking place mostly in one language, whose
metalingual frame of discourse markers takes place generally in the other5. In my
data, the phenomenon of language alternation at discourse markers occurred for
66% of the discourse markers throughout the speech of the more balanced bilin-
gual, Shira6. Furthermore, 85% of all her discourse markers occurred in a cluster
of discourse markers including at least one Hebrew marker (Maschler 1994a,b).
The phenomenon of language alternation at discourse markers has since been
documented in many situations of bilingualism (e.g., Brody 1987, McConvell
1988, Salmons 1990, Stroud 1992, de Rooij 1996, Auer 1995, Li Wei and Milroy
1995, Ösch-Serra 1998, Alfonzetti 1998, Matras 1997, Moyer 1998, Maschler 1988,
1998a, 2000a,b,c, Hlavac 2006, and Kyratzis 2007, to mention just a few). It is the
phenomenon responsible for the eventual borrowing of words such as English
okay or Yiddish nu into many of the world’s languages, to the point that such words
have become part of the language now used also by monolingual speakers7.
The study of bilingual discourse strategies thus leads us to a characterization
of discourse markers which is applicable also to the study of discourse markers in
monolingual discourse: they are metalingual utterances occurring at conversa-
tional action boundaries, or frame shifts. We have seen that bilingual discourse
provides evidence for considering fillers and hesitation markers which occur at
conversational action boundaries to be discourse markers. Furthermore, the strat-
egy of language alternation at discourse markers suggests that discourse markers
are perceived as a distinct and unified category – the category of utterances em-
ployed to metalanguage conversational action boundaries – because bilinguals
consistently switch languages in verbalizing discourse markers. Bilingual linguis-
tic behavior thus constitutes evidence that the category ‘Discourse Marker’ is not
merely a linguists’ construct, but rather a substantial and meaningful category for
speakers. Unlike some other linguistic categories, it constitutes a linguistic catego-
ry operating only at the discourse level, with no existence at lower levels of linguis-
tic analysis.

5. Recent studies show that this strategy is in use by bilingual children already at preschool
age (Kyratzis 2007).
6. Furthermore, returning to these same two speakers twelve years later, I found further gram-
maticization of this pattern in their emergent bilingual grammar, such that cognitive and tex-
tual discourse markers occurred in Hebrew significantly more frequently than in the earlier
corpus and interpersonal markers were further grammaticized in a qualitative sense, exhibiting
more fixed, formulaic expressions compared to the earlier corpus (see Maschler 2000c).
7. For the different stages of this process of language contact, see Goss and Salmons, 2000.
 Chapter 1.╇ Introduction 

2.1 Relation to previous approaches to discourse markers

In a recent comparison of some major approaches to discourse markers, Fischer

(2006) posits several dimensions along which to compare the multitude of ap-
proaches to these utterances. One such dimension is that of integratedness – the
degree to which a discourse marker is viewed as part of the utterances it connects,
as opposed to viewing it as an independent element. By showing that discourse
markers constitute the metalingual frame of the discourse, the present approach
views discourse markers as maximally detached from the discourse they frame.
The bilingual discourse strategy of iconically separating out this frame through the
strategy of language alternation underscores the unintegratedness of these ele-
ments of talk.
Another dimension of comparison posited by Fischer is the question of what
type of class discourse markers constitute: semantic, syntactic, functional, or no
class at all (2006:€2). By focusing on the function of discourse markers to negotiate
frame shifts in interaction, the present approach views discourse markers prima-
rily as a functional class. However, we will see that through their repetitive use in
this function, and by nature of the process of grammaticization, discourse markers
also form a syntactic category, one operating at the discourse level.
Many studies approach discourse markers based on relevance theory (Sperber
and Wilson 1986). According to these approaches, discourse markers impose con-
straints on the way an utterance is relevant in its context (e.g., Blakemore 1987,
Watts 1988, Jucker 1993, Andersen 1998). The present approach begins not with a
theory, but rather with a corpus. I ask what the particular discourse markers are
doing in the interactions in which they appear. As Aijmer writes in relation to
relevance theory: “a different approach may be better suited to explain what [dis-
course markers] are doing in authentic texts and why the speaker has chosen one
particle and not another. Moreover, the focus on a single explanation [i.e., the issue
of relevance] means that other reasons for using them […] may be neglected”
(2002:€11).
Many previous approaches to discourse markers relate to their textual func-
tion of relating utterances to other discourse units. For instance:
(1) according to Schiffrin, discourse markers are “sequentially dependent ele-
ments which bracket units of talk. […T]he beginning of one unit is the
end of another” (1987:€31).
(2) according to Fraser, a discourse marker is “an expression which signals the
relationship of the basic message to the foregoing discourse” (1996:€186).
(3) according to Hansen, discourse markers are “linguistic items of variable
scope, and whose primary function is connective” (1997:€160).
 Metalanguage in Interaction

(4) according to Redeker, discourse markers are “linguistic signals of textual

coherence links” (1991:€1139).
Other approaches relate to the interactional functions of discourse markers, for
instance:
(1) Schourup writes that discourse markers “constitute the range of conven-
tionalized reponses in English” (1985:€3).
(2) according to Östman, discourse markers “implicitly anchor the act of
communication to the speaker’s attitudes towards aspects of the on-going
interaction” (1981:€5, 1982:€152).
(3) Stubbs (1983:€193) considers discourse marks as “essentially interactive”.
Many approaches relate the textual and interactive functions of discourse markers
to each other in various theoretical frameworks (e.g., Östman 1982, Schiffrin 1987,
Traugott 1995a, Brinton 1996, Aijmer 2002). However, very few approaches relate
to the cognitive functions of discourse markers. Zwicky, for instance, considers the
English hesitation marker uh a discourse marker (1985), but as Schourup writes,
this item would regularly be excluded from the group of elements considered dis-
course markers by many scholars (1999:€235). Even-Zohar (1982:€191) combines
the textual with the cognitive when he characterizes these elements (‘void prag-
matic connectives’ in his terminology) as indicating “ ‘the speaker’s strong need to
organise his discourse’ or ‘the speaker’s anxiety to draw maximum attention to all
his shifts of mind’ ” (emphasis mine). This last phrase combines the focus on cog-
nitive constraints with the idea of shifting; however, the focus is not on frame shift-
ing in Goffman’s sense, as it is here.
The present approach, as we have seen, differs from most previous approaches
by extending the categorization of discourse markers from elements whose func-
tion is in the textual and/or interpersonal realms, also to those functioning in the
cognitive realm of discourse8. I have presented evidence from bilingual discourse
to justify this categorization. Furthermore, we shall see that the present corpus-
based approach suggests a framework in which these three realms of interaction
– textual, interpersonal, and cognitive – constitute a coherent whole enabling par-
ticular affinities across realms.

8. Two more recent approaches relating discourse markers to the cognitive realm of discourse
are Bazzanella 2006 and Yang 2006.
 Chapter 1.╇ Introduction 

3. The data

We now move to the study of discourse markers in Hebrew monolingual conversa-

tion. The present study is based on the Haifa Corpus of Spoken Israeli Hebrew
(Maschler 2004), comprised of casual Hebrew conversations among college-edu-
cated Israelis in their twenties with their friends and relatives, which I have been
collecting over the past decade or so. The data comprise 50 conversations, ranging
in length from approximately 1 to 9 minutes and taking place between 2 to 5 par-
ticipants. Altogether, the corpus comprises approximately 150 minutes of talk
among 124 different speakers, collected over the years 1994–2002, with the excep-
tion of one conversation recorded by myself in 1986.
The conversations were recorded by my discourse analysis students over the
years, in interactions in which they participated. They were submitted as part of
course projects, along with information concerning the prior-text (Becker 1979) of
the interaction, the nature of interpersonal relations among participants, and, occa-
sionally, relevant non-verbal behavior which participants were able to recall. I include
here only those interactions for which consent was obtained from participants9.
Many of the conversations in the corpus constitute narratives. This is not to
say that discourse markers are employed only in narrative discourse – Maschler
1997a, 1998b show the contrary (and see Section 4 below). This has more to do
with the texts students happened to select for their projects. The instructions given
were to record several long stretches of naturally-occurring casual talk (at least
two audio-cassettes’ worth) and choose a segment one feels enthusiastic about,
one they see themselves spending an entire semester with, a text having a clear
beginning, middle, and end. Interestingly, the majority of conversations selected
turned out to be narratives.
The data have been carefully transcribed and segmented into intonation units
(Chafe 1994), marking intonation contours, pauses, primary and secondary stress,
as well as other paralinguistic features such as tempo and volume which seemed
prominent in the particular excerpts (See the Transcription Conventions Section
which opens this study)10. This transcription method reflects the importance of the
intonation unit as the basic unit of spoken interaction (Chafe 1987, 1994), a unit
serving as the structural basis for my definition of discourse markers (see below).

9. These students attended my courses at three different institutions in Israel: at the Univer-
sity of Haifa, at the Hebrew University of Jerusalem, and at Oranim College near Haifa. They
studied at departments of Communication, Hebrew Language, and English in these three insti-
tutions. I wish to express my gratitude to the students for their generosity with the data.
10. I wish to thank my research assistants Michal Peled-Kaveh, Tamar Rubin, Shlomo Kim,
Tamar Mencher, Tal Dori, Hilla Polak-Yitzhaki, Hadar Netz, Roi Estlein, and Carmit Miller for
help with transcription and coding.
 Metalanguage in Interaction

4. Distributional patterning: Negotiating frame shifts via discourse markers

There is a great deal of systematicity in the ways discourse markers are employed
besides the systematicity of language alternation at discourse markers observed in
bilingual conversation. First, we return to the issue of the employment of discourse
markers as a function of the type of frame shift involved.
In his study of footing, Goffman writes:
A change in footing implies a change in the alignment we take up to ourselves and
the others present as expressed in the way we manage the production or reception
of an utterance. A change in our footing is another way of talking about a change
in our frame for events. [...] participants over the course of their speaking con-
stantly change their footing, these changes being a persistent feature of natural
talk (1981:€128, emphasis mine).

Goffman ends his study with the belief that “linguistics provides us with the cues and
markers through which such footings become manifest, helping us to find our way
to a structural basis for analyzing them” (ibid.: 157). Because of their metalingual
nature, discourse markers play a significant role in framing the discourse. They are
employed in order to create and reflect frame shifts, thus functioning as contextu-
alization cues (Gumperz 1982) signaling those shifts (Maschler 1997a, 1998b).
Discourse markers constitute one of the most prominent linguistic strategies
for negotiating footing, or frame shifts, in interaction. In a study of 11 out of the
50 Hebrew conversations on which the present study is based11, altogether 336
discourse markers were employed. Of the 595 frame shifts found throughout that
database, 247 (42% of all frame shifts) were accompanied by at least one discourse
marker (Maschler 1997a, 1998b).
Furthermore, the corpus shows systematicity in the ways discourse markers
are employed at frame shifts. In general, the ‘higher’ the boundary between con-
versational actions, the more the linguistic material necessary for its construction.
‘Larger packages’ (Sacks 1992), such as the beginning of a new episode in a narra-
tive, open with more discourse markers – and particularly more clusters of dis-
course markers – than ‘smaller packages’, such as the beginning of a ‘second’ in an
adjacency pair. Table 1 illustrates the distribution of discourse markers across the
595 frame shifts found throughout the 11 conversations, separated into segments
of narrative and non-narrative discourse stretches12:

11. These 11 conversations comprise 30 minutes of naturally-occurring monolingual Hebrew

casual conversation among 31 different participants (friends and family relatives), collected
throughout the years 1994–1997.
12. For more detail, see Maschler 1997a, 1998b.
 Chapter 1.╇ Introduction 

Table 1.╇ Distribution of discourse markers and discourse marker clusters at boundaries
(Maschler 1997a, 1998b)

Type of boundary Number of Number Number

boundaries of boundaries of boundaries
+ marker + cluster

I supertopic/story 25 11 (44%) 3 (12%)

II episode (narrative) 61 30 (49%) 9 (15%)

III sub-episode (narrative) 99 38 (38%) 8 (8%)

II subtopic (non-narrative) 99 46 (46%) 10 (10%)

III turn in adjacency pair 311 122 (39%) 20 (6%)
(non-narrative)

Total 595 247 (42%) 50 (8%)

We see that higher-level boundaries between conversational actions (i.e., level-II

boundaries), such as those between episodes in narrative or between sub-topics in
non-narrational discourse, involve the use of a discourse marker in 49% and 46%
of such boundaries, respectively, whereas lower-lever boundaries (i.e., level-III
boundaries), such as between sub-episodes in narrative or between the two parts
of an adjacency pair in non-narrative discourse, involve the use of a discourse
marker in only 38% and 39% of such boundaries, respectively. The pattern is par-
ticularly apparent when examining boundaries involving employment of clusters
of discourse markers (right-most column of Table 1). Clusters are almost twice as
common at level-II boundaries (15%, 10%) as they are at level-III boundaries (8%
and 6%). Level-I boundaries (beginnings of new stories or major discourse topics)
exhibit an employment rate of discourse markers somewhere in between level-II
and level-III boundaries (44% of all level-I boundaries involve discourse markers,
12% involve clusters). This has to do with the fact that at such moments in interac-
tion, we often find longer metalingual utterances, such as rotse lishmoa keta?
(‘wanna hear something weird/funny [lit. ‘a segment’]?’) or 'ani 'asaper lexa mashe-
hu 'al… (‘let me tell you something about…’) – longer utterances (usually sen-
tences) not sufficiently crystallized in order to be considered discourse markers.
An illustration of this point can be seen in excerpt 3, a conversation among 5
family members – a couple and their 3 children Shmulik, Meirav, and Noga. Shmu-
lik and Meirav are in their early twenties. Noga is their little sister. Shmulik is in
the midst of an orientation (Labov 1972) to a story about a little boy, but he is
constantly being interrupted by his family members:
 Metalanguage in Interaction

Excerpt 3 (‘Wounds’):
54 Shmulik: .. 'az hayeled haze--,
so the boy this
so this boy,
55 ... 'eh kaze,
uh [is] like,
56 min profesor,
sort of [a] professor,
57 kmo she'amarti,
as I said,
58 hu sho'el hakol vehu v
he asks everything and he an
59 ve'im 'at lo 'ona lo beduyak
and if you (fem.) don’t answer him in accur
60 be.. bimduyak,
in.. in accuracy
accurately,
61 vehu yode'a be'erex,
and he knows just about
and he knows only the general idea of things,
62 .. 'az 'eh 'i 'efshar lehapil 'oto.
so uh [it’s] impossible to fool him.
63 z’tomeret hu--
in other words he-- [is]
64 Meirav: .. ben kama?
son of how much
how old?
65 Shmulik:→ .. 'eh 'ani yode'a?
uh [do] I know
uh what do I know?
66 'e.. 'ole le.. kita 'aleph sh
u.. going up to grade a
uh.. going up to first grade
 Chapter 1.╇ Introduction 

67 be'od shnatayim shalosh.

in another two years three
in another two-three years.
68 Father: 'eifo hatelefon?
where[’s] the phone?
69 Shmulik:→ ... bekitsu--r,
anyway,
70 Meirav: .. kan kan. {concerning the phone}
here here.
71 Mother: sivanush shalxa 'otam.
Sivanush sent them.
72 Father: 'ah tov. {concerning the phone}
oh good.
73 Noga: shshshshshsh.
74 .. n--u!
75 tnu lo lesape--r!
let him tell!
76 Shmulik: →'az 'eh
so uh
77 → ki haya fadixot,
because there were screw ups,
78 hi 'eh hayta xola sham,
she uh was sick there,
79 ve.. revital ve'ani halaxnu levaker 'otam.
and.. Revital and I went to visit them.
80 →ve'az 'e--h,
and then u--h,
81 hu kaze mis
he like loo
82 ba kaze,
comes like,
Throughout these lines, Shmulik attempts a return to the narrative (a task he fi-
nally accomplishes at line 81) while also dealing with the interruptions. The move
from a first to a second pair part of an interruption (such as from a clarification
question to an answer, or from a side-comment to a response to it) is of a lower
 Metalanguage in Interaction

order than the move from responding to an interruption back to the narrative. Let
us compare Shmulik’s employment of discourse markers throughout his attempts
to return to narration (level-II boundaries) with those he employs in order to deal
with participants’ side comments (level-III boundaries).
One of Shmulik’s sisters, Meirav, latches on at line 64, asking how old the boy
is. After answering her question (lines 65–67), Shmulik attempts a return to the
narrative with the discourse marker bekitsur (‘anyway’, line 69). His father is mean-
while engaged in asking where the phone is (line 68). Meirav answers the father
(line 70), while the mother makes a side comment related to Shmulik’s orientation,
giving the reason why Shmulik and his friend Revital went to the place where they
met the boy whom this story concerns (line 71): sivanush shalxa 'otam. (‘Sivanush
sent them.’). While the father acknowledges finding the phone (72), the little sister,
Noga, quiets everybody down (73) and hastens Shmulik to go on with the narra-
tive (lines 74–75). Shmulik attempts another return, this time with a cluster of two
discourse markers – 'az 'eh (‘so uh’, line 76). He then stops this move at mid-utter-
ance (as indicated by the fragmentary intonation unit at 76) and switches to relate
to the mother’s side comment (lines 77–79). His response begins with one dis-
course marker – ki (‘because’), explaining why Sivanush sent them – she was sick
(line 78). Finally, at line 80, he returns to the story yet again, this time with a clus-
ter of 3 discourse markers – ve'az 'e--h (‘and then u--h’).
We see that at level-II boundaries of attempting to return to his narrative,
Shmulik generally employs more discourse markers, and particularly more clus-
ters: bekitsur (‘anyway’, 69), the cluster 'az 'eh (‘so uh’, 76), and the cluster ve'az 'e--h
(‘and then u--h’, 80)13, whereas at the level-III boundary of relating to a side-com-
ment, he employs only one discourse marker - ki (‘because’, line 77).
However, this pattern is not adhered to in 100% of the cases. For instance, in
responding to Meirav’s question (line 64) concerning the age of the boy (a level-III
boundary), Shmulik actually employs a cluster of 3 discourse markers: 'eh 'ani
yode'a? 'e (‘uh what do I know? uh’, lines 65–66). Furthermore, Meirav’s move at
the level-II boundary of line 64 manifests no discourse markers at all. Perhaps in
order to minimize the disruption, she shapes her question as a co-construction
(Lerner 1991), building on Shmulik’s previous utterance. She grafts her question
onto the subject of the previous utterance – hu (‘he’) – introduced by Shmulik in
the previous intonation unit in the midst of a rephrasal of his own utterance, which
he began with the discourse marker z’tomeret (‘in other words’, line 63).
Table 1 shows a general tendency throughout the corpus – the higher the
boundary in the hierarchy of frame-shifts, the greater the amount of metalingual

13. These 6 discourse markers employed by Shmulik to return to the narrative were counted as
occurring at 3 separate level-II boundaries: lines 69, 76, and 80.
 Chapter 1.╇ Introduction 

material that is employed to construct that boundary. This does not mean, of
course, that no level-III boundaries (for instance) will ever exhibit a discourse
marker cluster. On the whole, however, this happens in only 6% of level-III bound-
aries in non-narrative discourse and in only 8% of level-III boundaries in narrative
discourse. We see, then, that discourse markers constitute a system exhibiting pat-
terning in terms of their quantitative distribution in conversation.
This patterning can be compared with Couper-Kuhlen’s finding (1998) con-
cerning the role of high onsets (the height of the first stressed syllable in a turn
constructional unit) in indicating that a ‘big package’ is underway. This is a non-
verbal strategy for managing the production of a multi-unit turn (1998:€ 19),
whereas the current study shows a metalingual (and therefore verbal) strategy ac-
complishing a similar end.
Studies in the field of historical discourse analysis (e.g., Enkvist and Wårvik
1987) also mention a patterning reminiscent of the one presented here. Studies of
ancient Germanic written texts note that certain formulas, termed ‘introductory
saga formulas’, ‘narrators’ formulas’, ‘transition formulas’, and ‘concluding formu-
las’, although not fixed and quite productive, function to shift scenes and mark
new episodes in the written texts (e.g., Clover 1974, 1982, Lönnroth 1976). “These
formulas are not obligatory and their presence is in large part determined by the
length of time elapsed between scenes. The longer this is, the more likely and the
more elaborate the formula is” (Clover 1982:€88, cited in Brinton 1996:€18). We see
that in this approach as well, the size of the ‘package’ is related to the type and
quantity of metalingual utterances employed at the boundary. The larger the ‘pack-
age’, the greater the amount of metalingual material necessary for the frame shift.
Furthermore, as in the present approach, these metalingual utterances are seen to
have not only a textual function, but often an interactive one as well14. They are
viewed as “a set of narrative directions – formulas in which the narrator addresses
the audience on the mechanics of composition” (Clover 1982:€89). Finally, these
studies provide further evidence for the view taken here on the emergence of some
discourse markers from longer metalingual utterances, as will be elaborated in the
following chapters.

14. Not surprisingly, the cognitive function of such utterances in not discussed in relation to
these written texts.
 Metalanguage in Interaction

5. Defining discourse markers

We are now in a position to define discourse markers more clearly. A variety of

terms have been used to refer to discourse markers15 (e.g., pragmatic markers,
discourse particles, connectives, pragmatic expressions, discourse operators,
markers of pragmatic structure) and they have not always been clearly defined.
Brinton mentions more than twenty such terms (1996:€29) and notes that the many
definitions bear little resemblance to one another. The differences result from the
pragmatic function of these markers that is viewed as primary. Some approaches
stress their function in expressing relevance to prior discourse, others stress their
structural function, or their function as response signals, as a means to achieve
conversational continuity, or their interactive character (ibid.: 30–31).
None of the previous approaches anchors the definition of discourse markers
in the process of metalanguaging involved in their employment. Schiffrin (1980,
1987:€328) mentions that meta-talk expressions such as this is the point, or what I
mean is can function as discourse markers. However, for her, this is a subcategory
of the set she considers discourse markers. In the present approach, on the other
hand, the claim is that all discourse markers are metalingual. This approach stems,
as we have seen, from observations concerning the ways these markers function in
bilingual conversation.
However, there are some metalingual utterances that do not constitute dis-
course markers. The following excerpt, also from the conversation between Shira
and Grace (excerpts 1,2) illustrates this:
Excerpt 4:
1 Grace: ... I don't
2 .. but I meá--n,
3 .. if you say,
4 .. somebody is.. s..eárching for apprecià--tion,
5 ... it seems like they're
6 .. it soú--nds like,
7 .. they're searching to b--e,
8 .. pùt on a pédestal.. n.. to s.. to sóme extent
9 .. I mean /????/
10 Shira: 'a.. lò 'ani lo mitkavénet leze,
oh.. no I don’t mean this
oh.. no that’s not what I mean,

15. For a comprehensive review, see Brinton 1996:€29–31.

 Chapter 1.╇ Introduction 

In intonation unit 10, we find the Hebrew ‘equivalent’ of the English change-of-
state token (Heritage 1984, Schiffrin 1987, Maschler 1994b) 'a (‘oh’) and the dis-
course marker lo (‘no’) followed by the metalingual utterance 'ani lo mitkavénet
leze (‘that’s not what I mean’). This is a longer metalingual utterance not suffi-
ciently grammaticized in order to be considered a discourse marker (cf. Section 4
above). It is, however, reminiscent of the Old Germanic metalingual utterances
cited by Clover (1982) and Lönnroth (1976), a metalingual utterance which might
some day develop into a discourse marker. We shall return to this point below.
Having examined many discourse markers in both bilingual and monolingual
conversation, I noticed that the great majority of them share not only the semantic
property of metalinguality, but also certain structural properties. I thus arrived at
an operational definition of discourse markers in this study.

Definition of discourse markers

In order to qualify as a discourse marker, the utterance in question must fulfill a

semantic requirement:
a. The utterance must have a metalingual interpretation in the context in which it
occurs. In other words, rather than referring to the extralingual world, it must
refer metalingually to the realm of the text (in which case we are concerned
with a ‘textual discourse marker’), to the interaction among its participants (in-
cluding relations between speaker and his/her utterance – ‘interpersonal dis-
course marker’), or to their cognitive processes (‘cognitive discourse marker’).
In order to qualify as a prototypical discourse marker, the utterance must fulfill a
structural requirement as well:
b. The utterance must occur at intonation-unit initial position, either at a point
of speaker change, or, in same-speaker talk, immediately following any into-
nation contour other than continuing intonation. It may occur after continuing
intonation or at non-intonation-unit initial position only if it follows another
marker in a cluster (Maschler 1998b: 31).
We have already seen that all of the boldfaced utterances in excerpts 1 and 2 above
are metalingual in one of the three realms – textual, interpersonal, or cognitive.
This is the case also for the discourse markers discussed in excerpt 3: 'eh is cogni-
tive, attesting the cognitive processes occurring at frame shifts, 'ani yode'a? (‘what
do I know’) is interpersonal, referring to relations of uncertainty of the speaker
towards his utterance, while bekitsur (‘anyway’), ve- (‘and’) 'az (‘so’) and ki (‘be-
cause’) are textual, constructing relations between parts of the text. Similarly, the
'a (‘oh’) of excerpt 4 functions cognitively, realizing new information, and the lo
 Metalanguage in Interaction

(‘no’) following it functions in the interpersonal realm, disaligning the speaker

with the interlocutor’s previous words.
As for the structural requirement, all the discourse markers in the preceding
excerpts occur intonation unit initially, unless they are part of a cluster. The dis-
course markers at excerpt 1, lines 6, 11, 24, and 29, and excerpt 3, lines 65, 76, and
the 'a (‘oh’) of excerpt 4, line 10 occur at points of speaker change. Those at excerpt
1, lines 14, 15, and 32, excerpt 2, line 10, and excerpt 3, lines 69, 77, and 80 occur
in same-speaker talk following non-continuing intonation contours. The remain-
der – those at excerpt 1, lines 6–7 and 14–17 occur in same-speaker talk in dis-
course marker clusters: ken, ma shebarur (‘yes, what’s for sure’) and 'aval[ø16] ken,
gam (‘but[ø] yes, also’), as do those at excerpt 3, lines 65–66 ('eh 'ani yode'a? 'e ‘uh
what do I know? uh’), lines 76 ('az 'eh ‘so uh’) and 80 (ve'az 'e--h ‘and then u--h’)17
and the lo (‘no’) of excerpt 4, line 10. Thus, all discourse markers discussed so far
fulfill both the semantic and the structural requirements for discourse marker-
hood and constitute therefore prototypical discourse markers18.
In Maschler 2002a, I investigated all the discourse markers employed through-
out 16 of the 50 recorded conversations on which the present study is based19.
Altogether, 613 discourse markers fulfilling the semantic requirement above were
employed. Of those, 574 markers (94%) fulfill the structural requirement as well,
i.e. constitute prototypical discourse markers. The remaining 39 discourse mark-
ers (6%) occur at intonation-unit initial position in same-speaker talk, but they
follow continuing intonation and no discourse marker precedes them (i.e., they do
not follow another marker in a cluster). These, then, are the non-prototypical dis-
course markers, to which we shall return in Section 7.
As for the structural requirement concerning intonation-unit initial position,
all but one instance of ke'ilu (Chapter 4, footnote 23) share this property in the
present corpus. I will suggest a reason for this property of Hebrew discourse mark-
ers in the concluding chapter and relate also to the phenomenon of discourse

16. The lack of punctuation at the end of line 14 (highlighted by the ø) indicates a fragmentary
intonation unit (Chafe 1994).
17. In other words, 'ani yode'a? (‘what do I know?’, excerpt 3, line 65), for instance, indeed does
not occur at intonation-unit initial position, but it satisfies the structural requirement because it
follows another marker – 'eh – in a cluster.
18. All other discourse markers appearing in excerpt 3 (but not discussed here) fulfill these two
requirements as well.
19. The database for the 2002a study is comprised of the 11 conversations of the 1997a and
1998b studies (see footnote 11), as well as of five more conversations. Altogether, the 2002a
corpus consists of 40 minutes of naturally-occurring monolingual Hebrew casual conversation
among 43 different speakers, collected throughout the years 1994–1997.
 Chapter 1.╇ Introduction 

markers regularly occurring at intonation-unit final position in languages typo-

logically different from Hebrew (Chapter 6, Section 1.3).
The first study to examine in detail a discourse marker negotiating relations of
speaker to text (‘stance discourse marker’) in spoken Hebrew discourse is Masch-
ler and Estlein 2008. In that study, we followed Biber and Finegan’s definition of
stance as “the overt expression of […] attitudes, feelings, judgments, or commit-
ment concerning the message” (1988:€1) and investigated Hebrew be'emet (‘really,
actually, indeed’, lit. ‘in truth’) in the corpus20. Our study shows that 8 out of 44
(18%) metalingual tokens of be'emet do not fulfill the structural requirement for
prototypical discourse markerhood. The majority of these tokens still occur into-
nation-unit marginally; however, they are found at intonation-unit final posi-
tion21. It seems that in the case of stance discourse markers, a percentage greater
than 6% do not fulfill the structural requirement for prototypical discourse mark-
erhood, but further study is needed here.
With respect to the syntactic category of adverb, based on a study of adverbs
in a variety of languages, Ramat and Rica have suggested a scalar, or even radial
category “with prototypical instantiations and less typical or even marginal ones”
(1994:€289). For example, in one of two approaches they present, prototypical ad-
verbs are those identified as manner predicate adverbs (e.g., English quickly),
whereas ‘setting’ adverbs of space and time, such as today, now, here are less proto-
typical, and focalizers such as only, also, even are even further away from the cent-
er. At the very margin of their classification are what they call ‘text adverbs’ like
firstly, consequently, nevertheless, pertaining to the level above the sentence. These
are the ‘textual discourse markers’ in the present approach.
Based on the study of be'emet, we have suggested a prototypical syntactic cat-
egory of discourse marker, which is constituted by tokens fulfilling both the se-
mantic and the structural requirements in the definition above. Non-prototypical
tokens fulfill only the semantic requirement:
[…] just as the category of adverb is a scalar, or even radial one […], so the cate-
gory of discourse marker can be viewed in a similar fashion: prototypical be'emet
discourse markers (those which are highly metalingual (interactional) and have a
very minor referential dimension) fulfill both the semantic and the structural re-
quirements in the definition of discourse markers, whereas less prototypical dis-
course markers (those operating metalingually (interactionally) as well as

20. By then the corpus was extended to 91 conversations constituting approximately 270 min-
utes of talk among 223 different speakers.
21. Another Hebrew discourse marker which sometimes appears intonation-unit finally is
'aval (‘but’), though no examples were found in the corpus (cf. Mulder and Thompson (in press)
for English but). Interestingly, like some instances of be'emet, this discourse marker expresses
adversativity.
 Metalanguage in Interaction

referentially) fulfill the requirements only partially. Of this latter category, dis-
course markers having a higher interactional dimension […] occur intonation-
unit initially or finally, while those which are both interactional and referential
occur intonation-unit internally (Maschler and Estlein 2008:€312).

I suggest, then, conceptualizing the category of discourse marker as a scalar cate-

gory, with tokens fulfilling both the semantic and the structural requirements con-
stituting the prototypical instantiations. In Section 7 and throughout the continu-
ation of this study, we will see various factors resulting in non-prototypical
discourse markers. We shall also see that in the process of grammaticization, it is
the tendency of discourse markers to gradually ‘gravitate’22 towards the structural
position reserved for discourse markers in the language.

5.1 Previous approaches to prosody in defining discourse markers

No other approach to discourse markers that I am aware of supplies an operational

definition of discourse markers based, among other things, on the concept of the
intonation unit and on prosody in general. As noted by Aijmer (2002:€262), prosody
is a neglected area in the research on discourse markers. Few approaches relate to
prosody at all, and when they do, it is not for the purpose of defining discourse
markers, but rather for separating out different uses of a particular marker23. Tabor
and Traugott, e.g., emphasize the intonation contour of the discourse marker itself
(1998:€ 254). Ferrara 1997, Aijmer 2002, and Yang 2006, e.g., show that different
functions of discourse markers are correlated with particular intonation contours
and durational features across different markers. In the present approach, as we
have seen, prosody constitutes one of the major components in the definition of
prototypical discourse markers, and it is the intonation contour of the intonation
unit preceding the discourse marker that is found to be relevant for the analysis.
Another one of the dimensions along which Fischer compares the different
approaches to discourse markers is that of the medium through which the interac-
tion takes place. According to her, approaches that focus on the unintegrated na-
ture of discourse markers in relation to the discourse they connect (see 2.1 above)
study the medium of conversation, whereas those focusing on the integrated na-
ture of the items often account also for written text (2006:€9). The current defini-
tion of prototypical discourse markers stems, of course, from the spoken nature of
the data on which this study is based. However, Chafe has shown that the concept

22. Cf. Mulder and Thompson (in press) for this metaphor of ‘gravitation’, or “movement along
a pathway of grammaticization” in their terminology.
23. In the following chapters, we shall see that the present approach also correlates different
uses of a particular discourse marker with different prosodic features.
 Chapter 1.╇ Introduction 

of the intonation unit is relevant also to written discourse, where it becomes the
‘idea unit’ (1985). Written idea units are generally separated from each other by
punctuation marks, which partially imitate the intonation contours of spoken dis-
course. Thus, although my approach takes an ‘unintegrated view’ of discourse
markers in Fischer’s terms, the present definition of prototypical discourse mark-
ers could be adapted to written discourse without too much difficulty24.

6. Functional patterning of discourse markers

One way of characterizing frame shifts can be achieved by viewing them through
the prism of Becker’s approach to text as constrained by six different realms of con-
text: the interpersonal, referential, structural, prior-textual, medial, and silential
(1988). In Becker’s view, any act of languaging involves at least these six different
contextual realms, because in using language, we are always performing acts from
these six different realms simultaneously. In languaging, we are always simultane-
ously interacting with people (often with ourselves), referring to an extralingual
world we believe to exist beyond language, shaping linguistic structures, remem-
bering and evoking prior-text (Becker 1979), using a particular medium, and oc-
cupying space which, if not for our utterance, would have consisted of silence.
We have already seen that not all frame shifts are alike. In focusing on changes
in “the alignment we take up to ourselves and the others present” (Goffman
1981:€128), it appears that Goffman views frame shifting as a phenomenon shaped
mostly by interpersonal constraints. However, a closer look reveals that besides a
shift in interpersonal alignment, additional shifts in other contextual realms often
occur at these moments of talk, particularly when we are concerned with a higher-
order frame shift. At this type of frame shift, one also often switches to another
referential world (a new discourse topic or sub-topic), thus making other prior-
textual material relevant. One often switches then to a new linguistic structure
(e.g., different tense or aspect), and one switches into cognitive processes of a dif-
ferent nature in relation to those taking place at other moments of interaction
(Chafe 1994), as we have seen manifested by the fillers and hesitation markers
employed at conversational action boundaries. In his article on footing (1981),
Goffman provides examples of shifts in prosody and non-verbal behavior often
accompanying frame shifting as well. A moment of frame shift (or a conversa-
tional action boundary) can be defined on the basis of Becker’s approach to text as

24. Indeed, in a recent M.A. thesis, this approach has been applied to studying the discourse
markers segmenting Hebrew Quasi-Synchronous Computer-Mediated Communication (‘Chat
discourse’) (Krupik 2007).
 Metalanguage in Interaction

constrained by the various contextual realms. A moment of frame shift is per-

ceived here as a moment at which we find maximal shifts, relative to the sur-
rounding discourse, in constraints from the various contextual realms shaping
discourse (For elaboration and an example, see Maschler 1994b).
I have suggested sorting out discourse markers according to the realm in
which there are maximal shifts in contextual constraints when the particular
marker in question is employed (Maschler 1994b, 1997a). Markers were found to
function in three of the realms identified by Becker (1988) – the interpersonal,
referential, and structural, as well as in the cognitive realm.
Table 2 presents the 574 prototypical discourse markers (fulfilling both the
semantic and the structural requirements) found throughout the 16-conversation
corpus (Maschler 2002a).
Markers constrained mostly by the interpersonal25 realm negotiate issues of
closeness vs. distance among participants of an interaction (Bateson 1972[1956],
Scollon 1982). They include perception verbs, verbs of saying, agreement and dis-
agreement tokens, urging tokens, expressions of amazement, enthusiasm, lack of
enthusiasm, discontent, and maintaining contact, among others.
Markers constrained mostly by the referential realm are comprised of both
deictics and conjunctions. The conjunctions mark relations between conversa-
tional actions in a way that mirrors semantic relations in the extralingual world
marked by those conjunctions (e.g., consequence, coordination, contrast, cause,
disjunction, condition, and concession). The deictics have a discourse function
which mirrors their deictic function in the world narrated by the text (e.g. proxi-
mal vs. distal deixis).
Markers constrained mostly by the structural realm include utterances signal-
ing the ways conversational actions are related to one another in terms of order
and hierarchy (e.g., organizing the order of actions, introducing or ending a con-
versational action, returning to a main action, summarizing an action, or closing
a digression). The referential and structural markers together make up the textual
markers, because they all deal with relationships between different parts of a text.
Markers constrained mostly by the cognitive realm include markers attesting
the cognitive processes taking place at conversational action boundaries (e.g.,
processing information, realizing new information, realizing the need to rephrase
one’s previous utterance).

25. Recall that interpersonal markers negotiating relations between speaker and text were gen-
erally not investigated in these earlier studies.

Table 2.╇ Distribution of the 574 Discourse Markers throughout the Hebrew Database (92 types) (Maschler 2002a)
Interpersonal 35% Referential 40% Structural 10% Cognitive 15%
(57 Types, N=203) (10 Types, N=231) (14 Types, N=56) (11 Types, N=84)

Perception verbs 'at shoma'at? (you hear?) (1) Deictics Organizing order of actions Processing information
tir'e / ti'ri (look) (5) tishma (listen) (1) 'axshav (now) (13) kodem kol (first of all) (3) 'e--h (uh) (39)
'at ro'a! (see!) (1) 'ata yode'a / 'at yoda'at (you know) (7) 'axar kax €(then) (2) hadavar hasheni (second) (1) 'e--m€ (uhm) (4)
ta'amini li (believe me) (2) 'ani yode'a? (what do I know?) (1) rega (one sec) (12) 'e (2)
lo ta'amini (you won’t believe [this]) (1) lo y'dat (I dunno) (2) xake / xaki (wait) (3) hm (1)
'ata mevin? / 'at mevina? (you see?) (2) 'an’lo yoda'at (I don’t know) (1)
tavin (understand) (2)
lo hevanti ((I) didn’t get it) (1)

Verbs of saying 'ani 'asbir lexa€(I’ll explain to you) (1) Consequence Introducing an action Realizing new information
tagidi (say) (3) tsapri li (tell me) (1) 'az (so) (96) tov (okay) (15) 'a (oh) (15)
tagidi li (tell me) (2) kaxa (like this) (1) 'aha (3)
me'anyen ma tagidi (I wonder what 'oy (4)
you’ll say) (1)

Agreement yaxol lihyot (could be) (1) Coordination Ending an action Realizing the need to rephrase
ken (yeah) (28) ze yaxol lihyot €(this could be) (1) ve... (and) (81) zehu €(that’s it) (8) s’tomeret (I mean) (7)
naxon (right) (4) betax (sure) (1) ya (German ‘yes’) (1) ma s’tomeret (meaning) (1)
'okey (okay) (3) mhm€(10) ke'ilu€ (like) (7)
bídyuk (exactly) (1) be'emet (really) (2) klomar (that is to say) (1)
beseder (alright) (2) hm€(7)
metsuyan €(great) (1) 'aha (aha) (1)

Disagreement ma pit'om (no way) (1) Contrast Back to main action

lo (no) (20) lo naxon (wrong) (1) 'aval (but) (28) bekítsur (anyway) (5)
'an’lo betuxa (I’m not sure) (1) higzamt (you’ve gone too far) (1) hakítser €(anyways) (1)
bexáyexa (come on) (1) 'akítser (in short, anyways)(1)
Chapter 1.╇ Introduction 
 
Interpersonal 35% Referential 40% Structural 10% Cognitive 15%
(57 Types, N=203) (10 Types, N=231) (14 Types, N=56) (11 Types, N=84)

Urging speaker to continue ken? (yeah?) (1) Cause Summarizing

nu (go on) (26) yalla (go on) (1) ki (because) (5) besax hakol (all in all) (2)
basof€(in the end) (2)
Amazement yo (wow) (1) Disjunction Closing digression
ken? (yeah?) (4) shyo (wow) (1) 'o € (or) (3) lo xashuv (never mind) (1)
be'emet? (really?) (5) wow (10)
Metalanguage in Interaction

ma€ (what) (10) wai (6) (wow)

Maintaining contact Condition

ken (yes) (1) 'im (if) (1)

Displaying enthusiasm Concession

'axla keta! €(great story!) (1) 'eize tsxok (how funny) (1) lamrot she- (al-
'eize keta! (what a story!) (1) yofi (great) (2) though) (1)
meyle she- (never
mind that) (1)

Lack of enthusiasm
'az ma (so what) (1)
Expressing discontent
lo bseder (very bad) (1) tsk (1)

Other he (stance marker) (1)

bo/bo'i.... (let’s...) (4) ha'emet (truth is) (1)
ha (stance marker) (1) naxon? (right?) (1)
 Chapter 1.╇ Introduction 

Each column consists of utterances operating in a different realm of discourse, but

all markers have something in common – they function metalingually – as we
have seen. Clearly, a discourse marker can operate in more than one of these
realms, which is why we sometimes find the same marker in more than one col-
umn in Table 2. Furthermore, many markers operate in more than one realm si-
multaneously (excerpt 1, line 24, for instance). In such cases, the marker was clas-
sified into the realm which seemed the most dominant in that particular context,
a decision which is not always clear-cut. It turns out that more often than not,
constraints from all contextual realms are at work when a particular discourse
marker is employed, just as constraints from all contextual realms are usually at
work at every moment of languaging (Becker 1988). This is one of the major fac-
tors motivating the grammaticization of discourse markers. We shall return to this
point in Section 8 below and in the following chapters.
Table 2 is, of course, not exhaustive of all prototypical discourse markers em-
ployed in Hebrew, but only of those found in the 16 conversations on which it is
based. The above caveats in mind, one can still draw some general quantitative
conclusions based on Table 2 concerning casual Hebrew talk among friends and
family members of this socioeconomic group. The largest group of discourse
markers (exactly 50%) function in the textual realm, creating and reflecting rela-
tions between different parts of a text. 35% of all discourse markers deal with in-
terpersonal relations between discourse participants. However, the relatively high
number of different types available for this purpose (57 different types) points to
the great complexity of interpersonal relations negotiated in interaction, in con-
trast to the much more limited variety of relations between parts of a text (24
types). Only 15% of all discourse markers attest cognitive processes occurring at
frame shifts26.
As for the most common functions within each realm, we see that in the ref-
erential column, by far the most common relations between conversational ac-
tions are those of consequence and coordination (96 and 81 occurrences, respec-
tively). In the structural column, speakers are engaged in organizing the order of
conversational actions most frequently, with introducing them, and somewhat less
frequently with ending actions. In the interpersonal realm, participants engage
most frequently in agreeing and disagreeing, in displaying their amazement, and
in urging each other to continue. Despite the high tolerance for disagreement in

26. It must be remembered that only hesitation markers occurring at conversational action
boundaries are considered in Table 2. There are far more hesitation markers throughout the
data, such as at instances of self-repair (Schegloff, Jefferson, and Sacks 1977), but as long as they
did not fulfill the structural requirement for discourse markerhood, they were not included in
this table.
 Metalanguage in Interaction

this culture (Katriel 1986, Maschler 1994a, Blum-Kulka 1997), tokens of agree-
ment are far more frequent than those of disagreement (60 as opposed to 25)27.
A glimpse at this table provides a very general quantitative portrait of what it
is that Israelis do most at conversational action boundaries in casual conversation.
Whereas Table 1 is probably less language specific (as suggested by historical dis-
course analysis studies such as Clover 1982, Lönnroth 1976), Table 2 is particular
to the genre of informal discourse among Israeli students, their friends, and fam-
ily relatives studied here.
According to Aijmer, “among the characteristic properties of discourse parti-
cles which have not received enough attention in the literature is their ‘clustering’
(2002:€31). The classification of discourse markers into interpersonal, referential,
structural, and cognitive categories is useful also in explaining their tendency to
cluster at conversational action boundaries. If each discourse marker in a cluster is
associated with a change in a particular type of contextual constraint, and if a con-
versational action boundary (particularly a higher-level one) is characterized by
shifts in constraints from several contextual realms, then indeed we would expect
to find clusterings of discourse markers around conversational action boundaries
(Maschler 1994b). Table 1 gives some initial quantitative findings concerning the
frequency of discourse marker clusters as a function of the level of the boundary.
However, further study is needed in order to reveal the exact patterning of cluster-
ing. Many questions remain open, such as what the restrictions are on the types of
markers that can appear together in a cluster, or what constraints exist on the order
of markers with respect to the contextual realms they come from in a cluster.

6.1 Previous approaches to realms of operation of discourse markers

The present view of discourse markers functioning in the four realms of discourse
– the referential, structural, interpersonal, and cognitive – bears some resemblance
to Schiffrin’s view of discourse markers as “contextual coordinates of talk” (1987:€312).
According to Schiffrin, discourse markers locate an utterance in five planes
(1987:€24–28) or ‘domains’ (2006:317) of talk: the ideational structure (“proposi-
tions with semantic content” (1987:€ 26)), the action structure (the sequence of
speech acts in a discourse), the exchange structure (“the organization of turns at
talk” 2006:317), the participation framework (“the more social side of speaker and
hearer: their identities, alignments, relationships to each other and to what they are
saying” (ibid.), and the information state (“what speakers and hearers know: their

27. This finding must be regarded with caution, however, since many agreement tokens appear
in ‘yes, but’ constructions, as part of a disagreement actually. This word of caution applies to
Table 2 in its entirety: counting forms removed from context can be misleading.
 Chapter 1.╇ Introduction 

organization and management of knowledge and meta-knowledge” (ibid.)). Each

discourse marker is seen to be primarily associated with one of these planes, but all
markers have secondary uses in at least one other plane (1987:€316, Figure 10.2).
The actual classification into realms differs across the two approaches. What is
shared is the idea that a marker may function within more than one realm. This
view is shared by many other previous approaches, as we have seen in 2.1.

7. Structural patterning of discourse markers

So far we have seen systematicity in two aspects regarding employment of dis-

course markers in monolingual conversation – the moments in discourse at which
they are employed and the functions they fulfill. In order to investigate the third
type of systematicity – that involving their structural properties – let us return to
the non-prototypical 6% category of markers fulfilling only the semantic but not
the structural requirement in the definition of prototypical discourse markers28.
For example, in the following excerpt, Sharon recounts a conversation she had
with her father.
Excerpt 5 (‘Obituaries’):
51 Sharon: ... 'amárti lo,
I said to him,
52 .. 'ata yodéa shehu mèt.
you know that he died.
53 ... 'az hu 'omér li,
so he says to me,
54 'òy yoy yó--y!
oh n--o!
55 ... 'az xashávti--,
so I thought,
56 ... she--hu st´ám mitbadeax 'iti 'oy yoy yoy,
{----very high pitch-------} { regular pitch}
that he’s just kidding around with me [saying] oh no,
57 ... ke'ilu ló’xpàt lo,
sort of like [it] doesn’t matter to him,

28. Parts of Section 7 draw on Maschler 2002a. However, whereas that article explores the
construction of multivocality via discourse markers, here I focus on the general structural pat-
terning of discourse markers.
 Metalanguage in Interaction

Sharon constructs her father’s response following her announcement of bad news
(lines 51–52). The father’s constructed dialogue (Tannen 1989) consists of the dis-
course marker 'òy yoy yóy, roughly equivalent to English oh no (line 54). We see
that this discourse marker occurs at intonation-unit initial position in same-
speaker talk (Sharon’s), but it follows a continuing intonation contour (line 53)
with no additional discourse marker preceding it (i.e., not following another
marker in a cluster, as, e.g., in excerpt 1, line 17).
The 'òy yoy yóy of line 54 follows the continuing intonation of an intonation
unit presenting the constructed dialogue into the discourse: 'az hu 'omér li, (‘so he
says to me,’ line 53). In the database, there are no cases in which the intonation
unit introducing constructed dialogue and the discourse marker opening the con-
structed dialogue itself are separated by final intonation. The connection between
these two parts of an utterance is too strong to be separated by final intonation. If
they are at all separated into two intonation units, the first one is verbalized in
continuing intonation. As Chafe writes, the simplest way to signal a link between
one intonation unit and the next is by separating them with continuing intonation
(1988). Thus, the frame shift between the speaker’s voice introducing the con-
structed dialogue and the constructed dialogue itself is accomplished via a com-
promise – employing a discourse marker signaling the frame shift, but one that is
not separated as strongly from the intonation unit introducing it.
This is not the only environment in which discourse markers follow continu-
ing intonation in same-speaker talk. In line 57 we find another discourse marker,
ke'ilu, following continuing intonation (line 56). Sharon tells here how at first she
didn’t think her father was responding seriously to her bad news29: xashávti--, she-
-hu st´ám mitbadeax 'iti 'oy yoy yoy, (‘I thought, that he’s just kidding around with
me [saying] oh no’, lines 55–56), imitating the father’s voice ('oy yoy yoy) in a pitch
contour different than that of the first part of this intonation unit. She then moves
away from imitating the father’s voice and adds a self-rephrasal of this somewhat
unusual construction (see below). The discourse marker ke'ilu precedes the self-
rephrasal: ke'ilu ló’xpàt lo, (‘sort of like [it] doesn’t matter to him’). Ke'ilu will be
studied in depth in Chapter 4. At this point I mention only that a speaker’s realiz-
ing the need to rephrase his or her utterance is a cognitive process attested by
ke'ilu, which is therefore classified here as a cognitive discourse marker.

29. Sharon had good reason to suspect that this might be the case. Her story tells of macabre
humor she tried out on her father: she arbitrarily picked out some name from the obituaries
section of the newspaper he was reading, planned to ask him whether he knew the person (as-
suming he wouldn’t), and upon his negative answer, planned to say that he also never will know
him. This didn’t turn out to work so well, because Sharon happened to pick a name of a person
her father actually did know. For more detail here, see Maschler 1998b.
 Chapter 1.╇ Introduction 

According to Chafe, during verbalization, speakers move from one focus of

consciousness to the next, forming superfoci of consciousness which “bring to-
gether chunks of information too large to be accommodated within a single focus
of consciousness” (1994:€145). These chunks are termed ‘centers of interest’ and
they represent “attempts, with varying degrees of success, to push the mind be-
yond the constraints of active consciousness” (ibid.). Once a speaker judges that a
scanning of a center of interest has been accomplished, he or she expresses that
judgment with a sentence-final intonation contour. The realization that one’s ut-
terance is in need of rephrasal is independent of the judgment concerning comple-
tion of a center of interest. This is why we sometimes find continuing intonation
preceding self-rephrasals30.
In moving from imitating the father’s 'oy yoy yoy to the self-rephrasal at line
57, Sharon also moves back from constructed dialogue to non-constructed dia-
logue. Thus, this move actually blends the two categories illustrated so far: moving
between non-constructed dialogue and constructed dialogue (albeit in the oppo-
site direction here) as well as rephrasing one’s previous utterance. The strategy of
constructed dialogue often involves also moving into a different voice of the speak-
er’s, because speakers often include their own (ridiculing, ironic, joking, etc.) eval-
uation of the speech they are constructing. In Bakhtin’s words, “the speaker’s ex-
pressivity penetrates through the boundaries and spreads to” (1986:€92) the voice
of the person whose voice is being constructed. This is the “layering of voices”
studied by Günthner (1997, 1999a). Thus, the two strategies investigated here of-
ten blend in interaction.
Of the 39 discourse markers following continuing intonation throughout the
data (Maschler 2002a), 59% open constructed dialogue and another 13% begin
self-rephrasals31. What is common to these two environments, besides the con-
tinuing intonation unit preceding the discourse marker, then, is that both involve
the phenomenon of multivocality, or polyphony (Bakhtin 1981, 1986) in conversa-
tion. In both cases, we find a move to a different voice of the speaker – either the
voice of another in some narrated world as constructed by the speaker (construct-
ed dialogue), or another voice of the speaker, rephrasing what he or she just said
in the world of narration. This is a move of a level lower than that between sub-
episodes in narrative or between turns in an adjacency pair (i.e., lower than

30. However, it is still more common to complete a center of interest prior to self-rephrasal.€ In
the study of ke'ilu (Chapter 4), I found 58 tokens of self-rephrasal ke'ilu. Of those, 44 (75.9%)
occurred following non-continuing intonation, whereas only 13 (22.4%) followed continuing
intonation, as seen in excerpt 5. One token occurred intonation-unit finally.
31. In the remaining 28%, we find referential ve- (‘and’, 4 tokens) and 'az (‘then’, 2 tokens), cog-
nitive 'uysh (realizing negative information), and interpersonal 'at yoda'at (‘you (fem.) know’),
ze naxon (‘it’s true’), ma (‘what’), and naxon? (‘right?’). See Maschler 2002a: 28–30.
 Metalanguage in Interaction

Higher-level boundary

complex metalingual utterance

cluster of several consecutive prototypical discourse markers
prototypical discourse marker fulfilling both semantic and structural requirements
non-prototypical discourse marker fulfilling only semantic requirement
Ø discourse marker
Ø intonation unit boundary
blurring of syntactic properties

Lower-level boundary

Figure 1.╇ Strategies for the construction of frame shifts in interaction (Maschler 2002a)

level-III boundaries of Table 1) because it is a move to a voice ‘of second order’ in

relation to the speaker’s voice. In Section 4 we saw that higher-order frame shifts
generally appear with heavier discourse marking compared to lower-order frame
shifts. Here we see that this pattern is adhered to also when lower-level boundaries
(level-IV boundaries) are concerned. The lower the boundary between conversa-
tional actions, the more subtle the strategy constructing it.
Figure 1 sketches the various strategies found throughout the conversations
for constructing frame shifts on a scale from heavier to lighter marking.
As seen earlier, highest-order frame shifts, such as the beginning of a story or
a supertopic, are characterized by the heaviest metalingual material – complex
metalingual utterances which haven’t crystallized into discourse markers because
they are less frozen (i.e., they vary for person, number, gender) and are longer in
nature. With time, some of these will undergo processes of grammaticization and
become full-fledged discourse markers. Next we find clusters of several consecu-
tive prototypical discourse markers, often each from a different contextual realm.
For the majority of boundaries, the frame shift is accomplished via a single proto-
typical discourse marker fulfilling both the semantic and the structural require-
ments in the definition of discourse markers (as do 94% of the markers through-
out the data). For more subtle frame shifts, such as those introducing a voice ‘of
second order’ of the speaker, we find a non-prototypical discourse marker fulfill-
ing only the semantic requirement, as in excerpt 5, lines 54 and 57.
Frame shifts of an order even lower than this exhibit no discourse markers at
all, or no boundary between intonation units, as for instance in the following ex-
cerpt, exhibiting the discourse marker within the same intonation unit introduc-
ing the constructed dialogue:
 Chapter 1.╇ Introduction 

Excerpt 6 (‘Reznik Residence Halls’):

89 Li'at: ...'amarti la tir'i,
I said to her look,
90 tevareri 'od hatsa'ot.
find out [about] other offers.
When listening to intonation unit 89, it is clear that none of the criteria for identi-
fying intonation unit boundaries (Chafe 1994:€57–60) are found between its two
grammatical units. There is no pause between 'amarti la (‘I said to her’) and tir'i
(‘look’), no terminal pitch contour, no acceleration-deceleration pattern, and no
change in overall pitch or voice quality (e.g., creakiness).
The further down we progress along the continuum of Figure 1, strategies for
indicating the ‘upcoming package’ become more subtle, and the ‘packages’ them-
selves become smaller. Higher-level frame shifts are constructed via semantic-
pragmatic strategies such as clusters of prototypical discourse markers and longer
metalingual utterances, whereas lower-level frame shifts are constructed prosodi-
cally through the distinction between final vs. non-final intonation contours or
not marked at all.
Strategies for constructing the most minute frame shifts may even interfere
with the syntactic patterns of the language. We have already mentioned the unu-
sual syntactic construction of excerpt 5, line 56. The discourse marker 'oy yoy yoy
(‘oh no’) imitating32 the father’s constructed dialogue at line 54 occurs at intona-
tion-unit final position, following the verb mitbadeax (‘kidding around, joking’).
This is another case in which the element introducing the (imitated) constructed
dialogue and the (imitation of) the constructed dialogue itself appear with no in-
tonation unit boundary separating them (cf. excerpt 6). However, if the strategy of
constructed dialogue constitutes a voice ‘of second order’ of the speaker, imitating
constructed dialogue constitutes a voice of an even lower order. This is mirrored
by an even more subtle frame shifting strategy at line 56.
The verb mitbadeax functions at intonation unit 56 as a verbum dicendi in an
uncharacteristic way. Mitbadeax (‘joking, kidding around’) is an intransitive verb
and should thus take no complements. In the present context, however, it has be-
come a verb of saying introducing imitated constructed dialogue, with the imita-
tion ('oy yoy yoy, ‘oh no’) having the relation of direct object to the verbum dicendi.
Thus, an intransitive verb has become transitive in this particular context, and the
motivation to mark this extremely low-level boundary preceding imitated

32. That this is a direct quote is supported by the prosodic properties of intonation unit 56: the
very high pitch at the beginning of the intonation unit is contrasted with the regular pitch em-
ployed to imitate the father’s voice in the latter part of the unit.
 Metalanguage in Interaction

constructed dialogue has not only erased intonation unit boundaries but also in-
terfered with the syntax of the language, blurring syntactic properties of verbs.
Thus we see that motivation to mark the very subtle boundary between the
imitation of constructed dialogue and its introduction into the discourse can in-
terfere with the argument structure of a verb. In their study of transitivity and ar-
gument structure based on conversation, Thompson and Hopper come to the con-
clusion that
predicate […] ‘meanings’ are actually generalizations from many repetitions of
hearing predicates used in association with certain types of human events and
situations over the course of a person’s lifetime. What appears to be a fixed ‘struc-
ture’ [such as the argument structure of a particular verb] is actually a set of sche-
mas, some more ‘entrenched’ […] than others, arising out of many repetitions in
daily conversational interactions. […] ‘Argument structure’ needs to be replaced
by a greatly enriched probabilistic theory capturing the entire range of combina-
tions of predicates and participants that people have stored as sorted and organ-
ized memories of what they have heard and repeated over a lifetime of language
use. (2001:€47).

The current use of the verb mitbadeax (‘joking, kidding around’) as a transitive verb
is not a very ‘entrenched’ use at all, one with a rather low probability in Thompson
and Hopper’s proposed probabilistic theory of argument structure. It is the fluidity
of grammatical categories that enables this very ‘unentrenched’ use originating
from the discourse situation – the need to signal an exceptionally minute frame
shift in interaction. Excerpt 5 provides an example of one of the least entrenched
uses of this particular verb, as well as a possible context for its employment.
In a more recent study of the openness of grammatical construction, Hopper
writes:
Because grammar is a result of interactions rather than a prerequisite to them, it
is not a fixed code but is caught up in a continual process of local adaptation
(emergence). These adaptations are microscopic (for instance, the nonce use of
an intransitive verb with a direct object), and either go unnoticed or are dis-
missed as “errors”, but they provide the potential basis for future use and for the
analogical spread of forms. (2004:€153, emphasis mine).

Excerpt 5 is precisely an example of such a “nonce use of an intransitive verb with

a direct object”. This excerpt thus makes the ties between discourse and grammar
apparent both in the structural patterning of discourse markers and in their rela-
tions to the syntactic structures of the language.
This concludes the third patterning involving discourse markers illustrated in
this chapter. Discourse markers were shown to exhibit patterning in terms of (1)
the moments at which they are employed in interaction, (2) the functions they
 Chapter 1.╇ Introduction 

fulfill, and (3) their structural properties. Combining the first and third patterns
discussed here, we see that the system of discourse markers permeating Hebrew
interaction constitutes part of a larger, iconic system of grammatical and prosodic
features (and most likely kinesic ones as well), helping participants distinguish
higher-order frame shifts from those that are more subtle in nature.
Example 5, along with the quote from Hopper 2004 above, lead us next to
consider the grammaticization of discourse markers.

8. Grammaticization of discourse markers

Studies in the Social Sciences often focus on the tension between structure and
event. An event among people, be it a news broadcast, talk show, job interview, or
casual conversation, takes place against the background of a structure, a fixed
scheme functioning as a mold for that event. As this structure is employed in the
various events of that genre, however, a process of negotiation takes place. This
process leads to changes in the initial structure and to the emergence of a new
structure. Hopper pointed out that, parallel to this cultural process, which Gid-
dens (1984) termed “structuration”, there is a linguistic process, grammaticization,
by which new grammatical patterns emerge from interaction.
As an unintended outcome of communicative behavior, grammar is a product of
“structuration” (Giddens, 1984) rather than a bounded object to be thought of as
structure (Hopper 1998:€158).

It is this linguistic process of grammaticization which I am interested in investigat-

ing with respect to discourse markers here.
Hopper and Traugott define grammaticalization as “the change whereby lexi-
cal items and constructions come in certain linguistic contexts to serve grammat-
ical functions and, once grammaticalized, continue to develop new grammatical
functions” (2003: xv). In the earlier edition of their 2003 study, they discuss the
two terms – “grammaticalization” and the newer “grammaticization”:
Some linguists have told us that they avoid the longer word because “grammaticali-
zation” could be understood as “entering the grammar of a language”, i.e., becoming
“grammatical”. “Grammaticization”, by contrast, suggests a process whereby a form
may become fixed and constrained in distribution without committing the linguist
to a view of “grammar” as a fixed, bounded entity. A similar point is sometimes
made in a different way: it is said that “grammaticalization” stresses the historical
perspective on grammatical forms, while “grammaticization” focuses on the impli-
cations of continually changing categories and meanings for a synchronic view of
language, thus placing the entire notion of synchrony into question” (1993: xvi).
 Metalanguage in Interaction

Grammaticalization is the more common term and the one used by linguists prior
to Hopper and Traugott’s 1993 publication (e.g., Meillet 1912, Kuryłowicz 1964,
Heine and Reh 1984, Lehman 1985, Traugott and Heine 1991, Heine, Claudi, and
Hünnemeyer 1991). Hopper and Traugott write that, because it is far from obvious
to them that these distinctions between the two terms actually exist, and because
the work they are writing – an introductory textbook on the subject – doesn’t seem
the right place to use a new term, they opt for the more traditional “grammaticali-
zation” (1993: xvi). I, however, do perceive a distinction between the two terms. I
would like here to move away from the notion of grammaticality and wish to focus
on the continually changing categories and meanings from a discourse-pragmatic,
rather than a historical-diachronic, perspective. I therefore use the newer term, as
it is used by Hopper (1987, 1988, 1998):
The notion of Emergent Grammar is meant to suggest that structure, or regularity,
comes out of discourse and is shaped by discourse as much as it shapes discourse
in an on-going process. Grammar is hence not to be understood as a pre-requisite
for discourse, a prior possession attributable in identical form to both speaker and
hearer. Its forms are not fixed templates, but are negotiable in face-to-face interac-
tion […]

Because grammar is always emergent but never present, it could be said that it
never exists as such, but is always coming into being. There is, in other words, no
“grammar” but only “grammaticization” – movements toward structure which are
often characterizable in typical ways (1987:€142, 148).

Grammaticization has been studied both from a diachronic and a synchronic per-
spective (Lehmann 1985:€303). The majority of studies are historical, investigating
the sources of grammatical forms and their “grammaticalization chains” (Heine,
Claudi, and Hünnemeyer 1991:€171–174). Less frequently, grammaticization is also
seen as a “primarily syntactic, discourse pragmatic phenomenon, to be studied from
the point of view of fluid patterns of language use” (Hopper and Traugott 2003:€2).
It is from this latter synchronic perspective that I approach discourse markers in the
present study. Consistent with Hopper and Traugott’s view that grammaticization
“is motivated by speaker-hearer interactions and communicative strategies” (ibid.:
73), and with Bybee and Hopper’s view that “the notion of emergence [...] relativ-
izes structure to speakers’ actual experience with language” (2001:€ 3), this study
concerns itself with identifying some actual patterns of language use which have
resulted in grammaticization of discourse markers in Hebrew.
Grammaticization has come to be investigated in the context of discourse
markers in a variety of languages: e.g., for English, Romaine and Lange 1991, Finell
1992, Traugott 1995a,b, 2003a,b, Ferrara 1997, Tabor and Traugott 1998; for Old
and Middle English, Brinton 1996; for English, French and other languages,
 Chapter 1.╇ Introduction 

Fleischman and Yaguello 1999, 2004; for Hebrew-English bilingual discourse, Mas-
chler 1994a, 1997b, 1998a; for Hebrew, Maschler 2001, 2002b, 2003, Maschler and
Estlein 2008; for English and Japanese, Traugott and Dasher 2002; for German,
Günthner 2000; for French, Vincent 2005; for Swedish, Lindström and Wide 2005;
for Italian, Visconti 2003, 2005; for Spanish, Pons Bordería and Schwenter 2005.
Let us return now to the metalanguaging property of discourse markers. In
the following chapters, I wish to illuminate some aspects of the emergent grammar
by which the metalanguage employed in interaction crystallizes into the fixed ut-
terances we call ‘discourse markers’. This, then, is a synchronic study of the gram-
maticization of discourse markers, a phenomenon studied so far mostly from the
diachronic perspective.
Book-length studies of the grammaticization of discourse markers are very
few (Brinton 1996, Suzuki 1999, Aijmer 2002, Onodera 2004). Previous article-
length studies of the grammaticization of discourse markers (e.g., Finell 1989,
1992, Brinton 1990, 2001, König 1991, Traugott and König 1991, Abraham 1991,
Onodera 1995, Traugott 1995a,b 2003a,b, Wårvik 1995, Aijmer 1985, 1996, Jucker
1997, Tabor and Traugott 1998, Manoliu 2000, Schwenter and Traugott 2000,
Traugott and Dasher 2002 (Chapter 4), Visconti 2003, 2005, Lindström and Wide
2005, Pons Bordería and Schwenter 2005) trace a particular marker throughout its
history in various written documents from different periods of a language. While
illuminating in many respects, this approach is also problematic. For example, in
the beginning of her book Pragmatic Markers in English, Brinton writes:
I investigate whether medieval narrative might be structured much like contem-
porary oral narrative and whether these linguistic features might be functionally
motivated in ways analogous to pragmatic markers in Modern English discourse,
that is, textually and interpersonally. However, I recognize that the function of
oral features in the written texts of Old and Middle English may be somewhat dif-
ferent from the function of comparable features in the strictly oral discourse of
Modern English, perhaps being used self-consciously as stylized pragmatic mark-
ers (1996:€8).

Almost a decade later, ending a study of the grammaticization of Swedish particles

of the you know type, Lindström and Wide come to a similar conclusion:
There is […] a general difficulty that these oldest texts are not very interactive in
their nature, and thus there is little room for the use of dialogically sensitive ele-
ments. This remains a methodological problem in diachronic studies like the one
carried out here (2005:€232).

The written nature of the data, and the fact that they do not always come from the
same genre, render this approach somewhat problematic.
 Metalanguage in Interaction

The natural habitat of language, its primordial site (Schegloff 1993, 1996a,b),
is casual everyday conversation. Bakhtin (1986) viewed conversation as the most
basic genre from which all genres derive. Since this is the prime locus of the use
and development of language, it is only natural that this is where we should study
grammaticization processes, and in particular, those relating to quintessentially
interactional phenomena such as frame shifting.
The present study is based, as we have seen, on a synchronic corpus of Hebrew
conversations recorded over the years 1994–2002. Ideally, it would have been bet-
ter to have spoken corpora from at least two different periods of the language in
order to investigate grammaticization processes. However, an earlier corpus of
casual Spoken Hebrew does not exist33. As Schourup writes, “historical investiga-
tion of DMs obviously cannot rely on naturally-occurring spoken texts” (1999:€261).
One might wonder how, missing a diachronic perspective, it would still be possi-
ble to investigate grammaticization phenomena. This is possible because, upon a
close synchronic analysis of the tokens of a particular discourse marker in the
spoken corpus, we often find that a certain token functions in more than one way
simultaneously. These ambiguous tokens are particularly illuminating for research
on discourse markers, as close examination of the contexts in which they occur
reveals how a particular marker might come to serve more than one function, thus
suggesting the functional itinerary followed by the particular marker34. Compari-
sons with “equivalent” discourse markers in other languages can also be made in
order to support the grammaticization paths suggested. This is because, as noted
by Hopper and Traugott, grammatical morphemes “tend to be polysemous in sim-
ilar ways across languages, and to undergo similar paths of development as a result
of human discourse and interaction.” (2003:€33). Grammatical morphemes
“reflect the metaphorical processes that are based on human cognitive make-up,
and they reflect the inferences that humans commonly make when they commu-
nicate” (Bybee, Perkins, and Pagliuca 1994:€302).

33. For the chapter on ke'ilu (‘like’) (see below), a discourse marker which has greatly prolifer-
ated in the language in recent years, it was possible to compare conversations collected during
the earlier part of the corpus construction (1994–1997) with those collected later (1998–2002),
and thus arrive at a diachronic perspective based on two spoken corpora.
34. Thus, if we assume that this variation of more than one function per marker is representa-
tive of a diachronic process underway, the study can be viewed as a diachronic analysis of sorts.
I thank Cecilia Ford for pointing this out to me. For other synchronic studies of grammaticiza-
tion, see Thompson and Mulac 1991, Thompson 2002, Kärkkäinen 2003, Imo 2007, Keevallik
2006, 2007 – all on the grammaticization of epistemic parentheticals from the canonical Subject
+ Verb form – and Auer 2005, Günthner 2007, 2008, Pekarek Doehler 2007, Laury and Sep-
pänen 2008, Hopper 2004, Hopper and Thompson 2008, for synchronic studies of grammatici-
zation of other constructions.
 Chapter 1.╇ Introduction 

Other studies of the grammaticization of discourse markers taking this synchron-

ic perspective include Auer 1996, 2005, Günthner 1996, 2000, Fleischman and
Yaguello 1999, 2004.

8.1 Traugott’s theory of grammaticization and semantic change

Recall that in the present approach, the metalinguality of discourse markers was
discovered through a synchronic investigation of the strategy of language alterna-
tion at metalingual utterances (Section 2). This led to the semantic criterion in the
definition of discourse markers – that the utterance have a metalingual interpreta-
tion in the context in which it occurs. What is interesting to remark on from the
very start is that the metalinguality of discourse markers, discovered through syn-
chronic investigation in the current approach, is judged to be the final stage of
grammaticization, a judgement arrived at in diachronic approaches.
Based on a series of diachronic studies of Old, Middle, and Present Day Eng-
lish (e.g., Traugott 1980, 1982, 1986, 1989, 1995a,b, 1999, 2003a,b, Traugott and
Dasher 2002), Traugott proposed a model for the semantic change involved in
grammaticization phenomena. She showed certain pragmatic-semantic tenden-
cies in this change (see Figure 2 below).
As forms become increasingly grammaticized, there is “a tendency toward
metatextual meaning, or more specifically a shift from […] ‘the world being talked
about’ to ‘the speaker’s organization of that world in the act of speaking’” (Traugott
and Dasher 2002:€40). With this shift towards the world of the speaker, as s/he is
involved in the act of speaking, meanings become first more subjective (“increas-
ingly based in the speaker’s subjective belief/state/attitude toward the proposition”
(Traugott 1989:€35)), and then more intersubjective:
[intersubjectivity is best construed] in parallel with subjectivity […as] the explic-
it expression of speaker’s/writer’s attention to the ‘self ’ of addressee/reader in both
an epistemic sense (paying attention to their presumed attitudes to the content of
what is said), and in a more social sense (paying attention to their ‘face’ or ‘image
needs’ associated with social stance and identity). (Traugott 2003a: 128).

a. non/less subjective > subjective > intersubjective

b. content > content/procedural > procedural
c. non-metatextual > metatextual
d. scope within clause > scope over clause > scope over discourse
Figure 2.╇ Pragmatic-semantic tendencies (Traugott 2001)
 Metalanguage in Interaction

There is a move from focusing on the “world out there” to the world of the interac-
tion, with its metatextual structure, procedures for organizing that structure, and
interactional aspects such as intersubjectivity. Correlated with these changes is an
increase in the scope of the grammatical element from an element whose scope is
within the clause, to one whose scope is over the entire clause, and finally to one
whose scope is over a discourse segment. However, “[a]lthough correlated, the
individual horizontal trajectories [in Figure 235] are not necessarily vertically
aligned” (Traugott and Dasher 2002:€40).
Earlier studies by Traugott claimed that early grammaticization involved uni-
directional meaning change which may proceed along the path:
propositional > (textual) > expressive
(Traugott 1982:€257)
where “expressive” was later replaced by “subjective”. (Both terms can be subsumed
under the term “interpersonal”, as it is used in the present study – negotiating rela-
tions between speaker and hearer or between speaker and text). This ordering later
appeared to be too strong (e.g., Heine, Claudi and Hünnemeyer 1991:€ 190–191,
Brinton 1996), and was therefore revised as a set of tendencies which may overlap:
Tendency I: meanings based in the external described situation > meanings based
in the internal (evaluative/perceptual/cognitive) described situation.
Tendency II: Meanings based in the external or internal described situation >
meanings based in the textual and metalinguistic situation.
Tendency III: Meanings tend to become increasingly based in the speaker’s sub-
jective belief state/attitude toward the proposition. (Traugott 1989:€34–35).

Notice that although Traugott’s studies do not address cognitive discourse markers per
se, as does the present study, Tendency I covers this type of discourse marker as well.
In the present, synchronic, approach, we have seen that languaging is per-
ceived of as happening in all realms of discourse simultaneously (Becker 1988),
that is, in the referential, interpersonal, structural, prior-textual, medial, and silen-
tial realms. Languaging is constrained also by the cognitive constraints to which
human interlocutors are subject (Chafe 1987). Correspondingly, metalingual ut-
terances are constrained by these various contextual realms simultaneously,

35. The figure in Traugott and Dasher 2002 is slightly but crucially different than the one of
Traugott 2001, which I have adopted. Instead of the trajectory non-metatextual > metatextual
of Traugott 2001, Traugott and Dasher 2002 have truth-conditional > non-truth-conditional.€
However, below their figure they mention that “it [...] follows that there is a tendency toward
metatextual meaning, or more specifically a shift from de re to de dicto meaning” (ibid.: 40).
Since truth-conditions are completely irrelevant to my analysis, whereas metalanguage is cen-
tral, I adopt the figure in Traugott 2001.
 Chapter 1.╇ Introduction 

although, as we have seen, often a particular realm is more prominent than others
in a particular context. Therefore, there is no fixed cline of unidirectionality in the
present approach, as found in Traugott’s earlier studies (e.g. 1982).
By starting out with the notion of the interaction, with its three aspects of (1)
structure of the text, (2) interaction among participants, and (3) cognitive proc-
esses taking place in a speaker’s mind during verbalization, we saw that metalan-
guage operates at these three realms of discourse – the textual, interpersonal, and
cognitive (Section 2). There are no hierarchical relations between these realms,
and so the present approach avoids getting into the question of whether interper-
sonal meanings stem from textual ones or vice versa. Every act of languaging is
constantly constrained by the various contextual realms shaping discourse, and
discourse markers are no exception. There are, of course, many affinities and inter-
relations among the various contextual realms, but they are not fixed and depend
on the particular utterance in question. In the following chapters, most of all in
Chapter 5, we will see that these affinities are responsible for the particular func-
tional paths followed by the discourse markers.
What is interesting, however, is the fact that Traugott’s final stage of grammati-
cization arrived at through diachronic study – the “metatextual”, relating to the
speaker’s organization of the world being talked about in the act of speaking – cor-
relates precisely with the most essential property of discourse markers – their me-
talingual nature, arrived at synchronically in the present approach. Synchronic
and diachronic approaches, then, are seen to complement, other rather than to
contradict, one another.

9. Interacting as an Israeli

We have seen that Table 2 allows only a bird’s-eye view of all the discourse markers
found throughout the database. In order to achieve a more accurate understand-
ing of the ways discourse markers function in interaction, one must complement
this table with a detailed study of every single discourse marker, a task far too am-
bitious for the present study. I have chosen to focus on four particular discourse
markers. Three of them are representative of the three main realms in which dis-
course markers operate – the interpersonal, textual, and cognitive. In order to
study more closely the ways constraints from these different contextual realms
interact and result in grammaticization of discourse markers, I have chosen an
additional discourse marker – one that functions ‘in between’ realms.
For the interpersonal realm, I focus on the urging token nu ( ‘yeah, go on’), for
the textual realm – on the resumptive token bekitsur (‘anyway’, lit. ‘in short’), and
for the cognitive realm – on the token of self-rephrasal ke'ilu (‘like’, lit. ‘as if ’). As a
 Metalanguage in Interaction

marker functioning ‘in between’ realms, we shall examine tov (‘okay, fine’, lit.
‘good’), which operates both interpersonally and textually throughout the interac-
tions. Table 3 supplies quantitative information concerning the extent to which
each one of these markers is employed throughout the database – the 50 conversa-
tions on which the present study is based.
We see that nu is the most widespread of all four markers. In fact, it is the sec-
ond most common interpersonal discourse marker throughout the database (fol-
lowing ken (‘yeah’)). Of the markers investigated in depth in this study, nu is fol-
lowed by tov in terms of frequency, with 62 tokens in this corpus. Ke'ilu is the most
widespread utterance of the four (120 tokens), but only 58 function as discourse
markers. Finally, the least frequently employed discourse marker among the four
is bekitsur, with 49 tokens.
These four discourse markers have not been chosen for further study arbitrar-
ily. Not only is each one relatively prevalent throughout the conversations in rela-
tion to other discourse markers from the same realm36, but also each marker illu-
minates certain quintessential aspects of Israeli identity, society, and culture. I
therefore believe these four discourse markers are particularly telling for under-
standing the essence of interacting as an Israeli. In order to explain this statement,
however, we must turn to the actual studies of the particular markers.

Table 3.╇ Distribution of the four discourse markers focused on in this study

Interpersonal Textual Cognitive Interpersonal and Textual

nu bekitsur ke'ilu tov

115 49 120 (58) 62

36. Although based on a smaller corpus, Table 2 provides quantitative support for this claim.
 chapter 2

The interpersonal realm

The discourse marker nu:
Israeli Hebrew impatience in interaction

1. Introduction1

Growing up in a Jewish-Israeli household to a mother of Austrian-German back-

ground, there were several things I was not allowed to say to my mother. One of
them was nu. If I used nu in conversation, I would get scolded. My mother per-
ceived this word as disrespectful and impolite2, though she couldn’t help employ-
ing nu sometimes in her own talk to me.
There is no satisfying English equivalent for this word, but some tokens may
be translated as ‘well?’, ‘go on’, ‘so’, or ‘so what’. Nu often sounds impolite also to
Americans living in Israel. My American husband once said to me: ‘Don’t nu me!’
following my use of this word in our English discussion.
In another genre, classroom discourse, nu is considered highly inappropriate,
particularly in the talk of student to teacher, at least at the university level. I dis-
tinctly recall only two such incidents in the course of my teaching over the past ten
years, both of which struck me at the time as rather impolite.
Hebrew casual conversation, however, is abundant with nus. Although the
word is often generally perceived by Israelis as having an impolite aura, most of its
tokens do not seem to be interpreted as impolite in actual ordinary talk-in-inter-
action because in fact they function to encourage the flow of discourse rather than
obstruct it.
Examine, for instance, the following conversation between a man and two
women in their early twenties. The women are co-telling the man about their ex-
periences in a diving course down in Eilat, humorously describing how most of
the women in the group dropped out at various stages of the course:

1. This chapter was originally published in 2003 in Text 23: 89–128. It has been extended and
updated here to examine additional nu tokens from 20 more conversations so as to fit in with the
larger study.
2. The term ‘impolite’ is used here in its everyday sense, rather than as a term in a theory of
politeness.
 Metalanguage in Interaction

Excerpt 1 (‘Diving Course’):

29 Orna: harofé hipìl 'ota,
the doctor dropped her [i.e., caused her to drop out],
30 .. ya'ani.
sorta.
31 .. 'al.. 'e--h.. refu'í hi naflà.
on uh medical [reason] she fell.
for uh medical reasons she dropped out.
32 All: {laughter}
33 Ran: /'atem 'al?????????/
you [dropped out] for????????
34 Orna: ló /naxon/.
{--laughing--}
not true.
35 Anat: ... 'anàxnu nafálnu,
we dropped out,
36 'al sotsyométri.
{---laughing---}
on sociometric [grounds]
for sociometric reasons. [i.e., for not having the right social atti-
tude]
37 All: {laughter}
38 Orna: ... nish'àrnu 'árba banot.
{----------laughing---------}
we were left four girls.
39 ... to
al
40 ... tóv!
alright!
41 Ran: ... nu,
42 'ani maksh
I’m listn
43 .. 'ani mexake lá--
{-----laughing---}
I’m waiting for the
 Chapter 2.╇ The interpersonal realm 

44 .. lapo'énta.
{--laughing--}
for the point.
45 Anat: la-- pántsh layn.
{-------laughing-----}
for the punch line
46 Orna: ... tóv!
alright!
47 .. magía hayom hashenì.
arrives the day the second
the second day arrives.
The story (which had actually begun several minutes before line 29) proceeds at
length, with many long digressions and much laughter. Ran employs nu at line 41
immediately preceding his metalingual comment 'ani maksh.. 'ani mexake la--..
lapo'enta. (‘I’m listn.. I’m waiting for the.. for the point’, lines 42–44) in the midst
of Orna’s narrative (lines 38–40), urging her on to get to this story’s point. This is
flavored with humor indicated by Ran’s laughter as well as with perhaps some de-
gree of patronizing. However, no impoliteness is oriented to by the other partici-
pants here: Anat picks up the humor and laughingly co-constructs (Lerner 1991)
with Ran, completing his 'ani mexake la-- (‘I’m waiting for the’, line 43) with the
English borrowing la-- pantsh layn, (‘for the punch line’, line 45); and Orna pro-
ceeds enthusiastically, as her tov3! (‘alright!’, line 46) indicates, to the next episode
of the narrative, concerning the second day of the course (line 47).
Nu is a non-lexical form in Hebrew. Only two of five Hebrew dictionaries I
checked list nu as an entry. Even-Shoshan’s dictionary (2003) classifies nu as an
interjection borrowed from European languages4 meaning 'efo (‘therefore’) hava
(‘let us’) uvxen (‘well then’)–all words of high register. However, this classification
says little about the uses of nu in casual Hebrew conversation, such as the one
above. Avneyon’s dictionary (1997) comes closer: milat zeruz (‘hastening word’).
Although nu was borrowed from European languages, particularly Yiddish and
Russian, this word has gained new meanings in Israeli Hebrew, different from
those it had in the languages of origin5. For instance, as I am often told by Israelis

3. For the study of tov, see Chapter 5.

4. See Goss and Salmons 2000, Maschler 2000 a, b, and Chapter 1 of the present study, for
studies of the role of language alternation in the grammaticization of discourse markers.
5. Nu is often heard also in varieties of Jewish American English whose speakers originate in
Eastern Europe. In Weinreich’s Yiddish-English dictionary (1977), we find Yiddish nu trans-
lated as English ‘go on! well? on!’. According to Rosten (1968: 271–273), aside from oy and the
 Metalanguage in Interaction

upon hearing that I am studying nu, one characterization of new immigrants from
the former Soviet Union to ‘Israeli ears’ is that when these people speak Hebrew,
they employ nu much more frequently in comparison to native-born Hebrew
speakers, and often in contexts which do not seem appropriate. In a pilot study,
Mazo and Voloshin (1999) found that Russian nu has some softening functions
which are not found for Israeli nu.
In previous studies of the system of discourse markers segmenting Hebrew
talk-in-interaction (Maschler 1997a, 1998b, 2002a. See Table 2, Chapter 1), nu was
classified as a discourse marker functioning in the interpersonal realm of discourse
(Becker 1979) ‘urging speaker to continue’.
Recall that in the present study, a prototypical discourse marker is defined as
an utterance fulfilling two conditions; one semantic, the other structural:
a. Semantically, the utterance must have a metalingual interpretation in the con-
text in which it occurs. In other words, rather than referring to the extralin-
gual world, it must refer metalingually to the realm of the text, to the interac-
tion between its participants, or to their cognitive processes.
b. Structurally, the utterance must occur at intonation-unit initial position, ei-
ther at a point of speaker change, or, in same-speaker talk, immediately fol-
lowing any intonation contour other than continuing intonation. It may occur
after continuing intonation or at non-intonation-unit initial position only if it
follows another marker in a cluster (Maschler 1998b: 31)6.

articles, nu is the most frequently used word in Yiddish. Rosten adds: ‘From Russian nu: ‘well’,
‘well now’, etc.; cognates are common in Indo-European languages.’ (Rosten 1968: 271). I wish
to express my gratitude to Sandra Thompson for showing me this reference.
In a recent article, Sorjonen studies two Finnish particles, no, which she characterizes as a
go-ahead response (Schegloff 1990), and nii, another Finnish continuer (Schegloff 1982), both
manifesting some of the functions of Hebrew nu (see below). She mentions that the Finnish
etymological dictionary compares Finnish no to German nun and na and suggests that ‘it is
partially an original Finnish word, also found in other Baltic-Finnish and Sámi languages, [...]
which has also been influenced by foreign languages, most notably by Germanic languages’
(2002: 166). See also Hilmisdóttir’s study of Icelandic nú and núna (2007). It seems, then, that
other languages besides Hebrew and Jewish varieties of English have borrowed forms related to
nu from Germanic languages.
6. In Chapter 1 we saw that the two criteria in this definition coincide for 94% of the discourse
markers found throughout the database. We saw that the remaining 6% satisfy the semantic, but
not the structural requirement, and that the majority of these 6% category discourse markers are
employed to construct two types of multivocality in discourse–constructed dialogue (Tannen
1989) and self-rephrasal.€ In this chapter, I include nu tokens following continuing intonation in
same-speaker talk as well. All but two such instances occur in the move from non-constructed
to constructed dialogue.
 Chapter 2.╇ The interpersonal realm 

The nu of excerpt 1, for example, satisfies both requirements for prototypical dis-
course markerhood. It functions metalingually in the interpersonal realm as Ran
is urging Orna to continue her story and finally get to its point (lines 41–44), and
it fulfills the structural requirement as it appears at intonation unit initial position
at a point of speaker change. This nu, then, like the majority of tokens throughout
the database (68.7% of all instances of nu, see Table 1 below), fits the characteriza-
tion ‘interpersonal discourse marker urging speaker to continue’.
However, there are many subtleties in the functions of nu which the charac-
terization above does not capture, and we will also see some tokens of nu for which
this characterization is less relevant. The present chapter attempts a finer charac-
terization of the functions and flavor of nu in a corpus of casual Hebrew talk-in-
interaction. At the same time, it relates to the possible grammaticization path fol-
lowed by nu in Hebrew and to the role of impatience in creating interpersonal
involvement in Israeli discourse.

2. Data

Nu was found to be the second most common interpersonal discourse marker

throughout the 50 conversations of the Haifa Corpus of Spoken Israeli Hebrew
(Maschler 2004) on which the present study is based (see Chapter 1, Section 3). Fol-
lowing ken (‘yeah’) with over 240 occurrences, nu manifests 115 tokens, which cor-
responds to an average of one nu token every 1.3 minutes7. A few additional excerpts
manifesting tokens of nu were broadly transcribed immediately after having heard
them employed in interactions in which I participated. They, too, will be discussed
in the present chapter, although they are not counted in the following table.
Table 1 sketches an overall picture of the functions of nu in the recorded
database:

Table 1.╇ Distribution of nu tokens according to function throughout the database

Hastening Urging further Granting Providing Total

non-verbal development permission to joking/provoking key
action within main topic perform action

6 (5.2%) 79 (68.7%) 3 (2.6%) 27 (23.5%) 115 (100%)

7. The next most common interpersonal marker in the corpus after nu, lo (‘no’), was em-
ployed significantly less frequently.
 Metalanguage in Interaction

The following sections elaborate on the headings of this table and illustrate the
various functions of nu in this corpus of casual Hebrew talk-in-interaction.

3. Hastening co-participant to get on with some non-verbal action

Let us begin with the least metalingual use of nu. This word can be employed in
order to hasten a co-participant to get on with some non-verbal action. For exam-
ple, in the following interaction, David is cooking some soup in the kitchen while
telling his close friend Dalia a story. At the end of the story, he wants her to taste
his soup, an action she repeatedly refuses to perform:
Excerpt 2 (‘Soup’):
124 David: .. pashút,
simply,
125 .. ló rotse lehaxzir ta’ xovòt.
doesn’t want to return the debts.
126 'axsháv nit'am mehamara/ve/k8.
now we’ll taste from the soup
now we’ll taste the soup.
127 {David tastes the soup}
128 Dalia: (4.05) nu 'éx yatsa?
nu how came out
nu how did it come out?
129 David: .. 'aní xoshev sheyatsa tòv,
I think that came out good
I think it came out good,
130 .. bó'i tit'amì.
come taste [it].
131 Dalia: .. ló rotsa.
don’t want [to].
132 David: bó'i tit'amì!
{-----ff-----}
come taste [it]!

8. The Hebrew word for soup is marak and we know that this is what is cooking from other
parts of the text. However, something approximating the syllable -ve- is heard inserted within
the final syllable of this word here: mara/ve/k.
 Chapter 2.╇ The interpersonal realm 

133 Dalia: ló rotsa.

don’t want [to].
134 David: 'ani 'eshpóx 'et ze 'alàyix.
I’ll spill this on you
I’m gonna spill this on you.
135 Dalia: .. ló rotsa.
don’t want [to].
136 David: (3.92) 'amrù lax pá'am
they told you once
did they ever tell you
137 'amárti lax pà'am,
I told you once
did I ever tell you,
138 she'at domà egárfild?
that you similar to Garfield
that you look like Garfield?
139 Dalia: .... kén,
yes,
140 'amárta li,
you’ve told me,
141 shloshím 'elef pà'am,
thirty thousand times,
142 David: nu!
get on with it already!
{moving soup spoon closer to Dalia’s mouth}
143 Dalia: she'ani domá le
that I similar to
that I look like
144 .. ló rotsà!
{-------ff-------}
don’t want [to].
145 David: tit'amí--!
{----ff----}
taste [it]!
At line 127 David tastes the soup and Dalia asks him nu 'ex yatsa? (‘nu how did it
come out?’, 128). This use of nu will be dealt with in Section 7. David then invites
 Metalanguage in Interaction

her to taste the soup herself in a series of two requests (lines 130, 132), both of
which Dalia refuses (lo rotsa (‘don’t want [to]’), lines 131, 133). It takes several ad-
ditional increasingly direct requests to get Dalia to taste the soup. At line 142,
David moves the soup spoon closer to her mouth9, uttering nu! in sentence final
exclamatory intonation. Dalia refuses yet again (line 144), this time more loudly,
at which point David resorts to a direct order: tit'ami--! ‘taste [it]!’, line 145. The nu
at line 142, then, urges Dalia to perform a non-verbal action. Here it is accompa-
nied by impatience mixed with humor, as indicated by the prosody of lines 142
and 145, particularly the fact that the intonation contour of 145 imitates that of
144, with a marked high-low-high intonation pattern on the last syllable of
tit'amí--! (line 145) uttered in loud volume.
In excerpt 3, Yosi tells his wife about a conversation he engaged in with two
cashiers at the local grocery store. Yosi retells how one of the cashiers was telling
the other about some extraordinary place which her husband had taken her to
while they were abroad. The other cashier responded with some embarrassment
that she, too, had been there, at which point Yosi joins in, expressing his curiosity
about the place. At a crucial point in the story, right before the identity of the place
is revealed, Yosi describes the cashier dealing with the customer ahead of him in
line. Yosi employs nu (line 68) in his own constructed dialogue (Tannen 1989) ad-
dressed to the cashier, dramatizing his hastening of the cashier to be done with
that customer so that she could tell him the end of the story:
Excerpt 3 (‘Cashiers’):
63 Yosi: .. 'ani kvar ló yax
I already not ca
I can’t anymo
64 .. 'ani / muxráx/ lishmòa,
I / must/ hear,
65 .. 'ani mét mimètax 'axshav.
I’m dying of suspense now.
66 Tali: {laughter}
67 Yosi: ... 'ani ló yaxol lehaxzík ma'amàd,
{---------laughing------------}
I not can hold on
I can’t wait any longer,

9. The interaction was audio-, rather than video-taped, but the person who first transcribed it was
Dalia, who remembered the non-verbal action accompanying this utterance. The warning 'ani
'eshpox 'et ze 'alayix (‘I’ll spill this on you’) at 134, and the ensuing discussion concerning her resem-
blance to Garfield (136–142), further support this location of the spoon relative to Dalia’s face.
 Chapter 2.╇ The interpersonal realm 

68 .. nú,
{laughing, change of tone}
come on,
69 .... tigmerí maher 'im haben'adám haze!
{-------------laughing-----------------}
finish quickly with the human being this
be done with this person already!
70 Tali: ..... nú?
71 Yosi: {laughter}
72 Tali: {laughter}
73 Yosi: ... 'az hi 'oméret,
so she says,
74 .. tishmá,
{����---pp---}
listen,
75 .... lakxú 'otanu,
{-------pp----------}
they took us,
76 lexanùt le'avizarèy mín.
{--------pp---------} {----laughing-----}
to a store for accessories of sex
to a sex shop.
The key is quite humorous in this narrative, indicated by the laughter, as Yosi con-
structs his talk to the cashier (lines 68–69) and her response (lines 74–76). In fact,
Yosi’s laughter while uttering the cashier’s response (line 76) reveals his evaluation
of her constructed speech: as he later says explicitly (excerpt 3a below), he views
these cashiers’ bashful attitude towards sex shops ridiculous. The laughter accom-
panying line 76 manifests the layering of voices (Bakhtin 1986, Günthner 1999a)
in discourse–specifically, the layering of Yosi’s ridiculing voice over the cashier’s
constructed dialogue. The action Yosi would like the cashier to hasten in the midst
of this humorous segment is described by the verb of the intonation unit following
nu: tigmeri maher (‘finish quickly’) 'im haben'adam haze (‘with this human be-
ing’), or ‘be done with this person already’, line 69. Unlike the previous example,
then, here nu occurs in continuing intonation and is followed by an utterance
spelling out the action being urged. Note that while Yosi’s nu of line 68 hastens an
action in the extralingual world, Tali’s nu of line 70, to which we shall return later,
relates to the world of the text and is therefore considered metalingual here.
 Metalanguage in Interaction

Discourse markers are often employed in the little dialogues we carry on with
ourselves. A speaker can also urge him/herself to perform some action. No such
examples were found in the corpus, but an interaction around our dinnertable
supplies an example. In the following excerpt, which I transcribed immediately
after the utterance took place10, five-year-old Yotam begins to sing to us his ver-
sion of a song they learned that day at school:
Excerpt 4 (‘Song’):
1 Yotam: hashu'al,
{--singing--}
the fox,
2 'avar miderex,
{-----singing---}
passed from a path,
3 'el 'el 'el,
{--singing--}
to to to,
4 ... nu.
5 .... 'ex kor'im le'eifo sheha'anavim?
how one calls to where that the grapes
what do you call the place where the grapes are?
6 Mother: kerem.
vinyard.
7 Yotam: 'el 'el 'el,
{--singing--}
to to to,
8 hakerem.
{--singing--}
the vinyard.
At line 4 Yotam stops singing because he had forgotten the continuation of the
song. Employing nu in sentence-final falling intonation, he hastens himself to re-
member the missing words. Following a longer than average pause, in the follow-
ing intonation unit, he requests my help in this task. Once reminded, he continues
singing (line 7 and on).

10. These excerpts are transcribed using broader transcription conventions because they were
not recorded.
 Chapter 2.╇ The interpersonal realm 

Children are aware of this use of nu quite early. Following Yotam’s employment
of nu at a similar context, I asked him why he said nu there. His answer was quick
and simple: nu she'ani 'ezkor kvar! (‘nu [so] that I remember already!’). In this case,
then, nu may be better characterized as a cognitive discourse marker–one provid-
ing information about cognitive processes occurring at frame shifts which are often
revealed in the medium of spoken discourse, rather than as an interpersonal dis-
course marker. The context of recalling a word also places this use of nu closer to
being metalingual, rather than functioning in the extralingual realm.
The following excerpt supplies another nu token urging an action which is on
the boundary between the lingual and metalingual. It comes from an interaction I
heard between my two daughters. Nine-year-old Maya was trying again and again
to get her sister, twelve-year-old Shira, to play the Harry Potter trivia game with
her, which Shira repeatedly refused. Following a series of attempts on Maya’s part,
Shira finally said:
Excerpt 5 (‘Trivia Games’):
1 Shira: nu,
2 'aval 'ani sonét lesaxek,
but I hate to play
but I hate playing,
3 misxakèy trívya!
games of trivia
trivia games!
When I immediately asked why she said nu there, the answer was as quick as her
brother’s was in the previous example: nu sheday lenadned! (‘nu [so] that enough
already with the nagging!’). In other words, the action Maya was urged to do in
this context was ‘stop nagging her sister’. The action of stopping to nag somebody
here clearly involves language, although it involves other aspects of behavior as
well, a fact making this instance of nu another token functioning on the fuzzy
boundary between the lingual and the metalingual.
Nu as a hastener of non-verbal actions is not that common in the corpus–only
about 5.2% of the tokens–because most of the conversations do not take place
while participants are also performing accompanying non-verbal actions. Yet this
function seems to me to be the basic function of nu, the non-metalingual function
from which other functions originate. It is difficult to confirm this, as it would
involve examining the functions of nu in Yiddish, Russian, and Hebrew spoken
corpora from the time in which nu was first borrowed into Hebrew. However, it is
in agreement with characterizations of the semantic process involved in
grammaticization as involving a metaphorical shift from the concrete to the
 Metalanguage in Interaction

abstract (Traugott 1982, Sweetser 1988), or from the propositional to the ‘metalin-
guistic’ (Traugott 1988, Traugott and König 1991).

4. Nu urging further development within a topic

Just as one can be urged to perform a non-verbal action, so can a speaker urge an
interlocutor to perform a verbal action. The majority of nu tokens in this database
(more than 2 out of every 3 tokens, 68.7%) function to urge the speaker to move
on in the development of a discourse topic (Chafe 1994). This happens in both
narrational and non-narrational discourse, and the flavor of nu changes as a func-
tion of the genre in which it is employed. We will see that this type of nu can be
characterized as a ‘continuer’ (Schegloff 1982) with some modification.

4.1 Narrational discourse

Nu can function to urge a participant to move on in the development of a narra-

tive. This includes urging the speaker (1) to begin a narrative the hearer knows is
about to be told, (2) to move on to a current narrative’s next episode (Chafe 1987)
thus advancing towards its point, or (3) to return to the narrative following a di-
gression. I illustrate only the second category here.

4.1.1 Urging a move on to the next episode of a narrative

In the following excerpt, Tamir describes to his male friend Erez a fight he
witnessed between two men:
Excerpt 6 (‘Violence’):
35 Tamir: .. 'axsháv,
now,
36 .. hahù kvar haya-- 'al haritspá,
the other one already was on the floor,
the other one [i.e., the victim] was already on the floor,
37 .. ke'ilu,
like,
38 .. hu baràx bizxilá,
he had escaped crawling,
39 .. hu ló-- barax beritsà.
he hadn’t escaped running.
40 ... 'ata mevín?
 Chapter 2.╇ The interpersonal realm 

you understand
see?
41 .. ke'ilu--,
like,
42 .. kaze,
like
43 {demonstrating victim’s crawling}.
44 Erez: .... wá wá!
wow wow
45 ... nú?
46 Tamir: .. vedofék bo makòt,
and slamming at him blows
and [he] is hitting him hard,
47 .. vehu ló,
and he doesn’t,
48 Erez: /xxxxxxxx/
49 Tamir: ... lò meraxém 'alav.
doesn’t have [any] pity for him.
At line 36 Tamir describes the state the victim of the fight was in at that point in
the story: hahù kvar haya-- 'al haritspá (‘the other one was already on the floor’).
He backtracks to point out that the victim had been hurt in the fight so badly, that
he had escaped crawling, not running (lines 37–39). This is done via two rephras-
als employing kaze and ke'ilu (lines 37–43) (two ‘equivalents’ of English like11),
both verbally (lines 38–39) and non-verbally by demonstrating the victim’s crawl-
ing (lines 42–43). At line 44 Erez manifests his involvement in this description/
dramatization with the interjection wa wa! uttered in sentence-final intonation.
He then employs nu? (line 45) in appeal intonation12. This is followed by Tamir’s
move on to the next episode of the narrative, describing the next complicating ac-
tion: vedofek bo makot, (‘and [he] is hitting him hard’, line 46). The shift to the new
episode from the background of the narrative to its foreground is indicated by
Tamir’s use of the discourse marker ve- (‘and’, line 46) as well as by his switch in
tense and aspect (Hopper 1979) from barax (‘had escaped’, past perfective, lines

11. For the functions of Hebrew kaze (lit. ‘like this’) and ke'ilu (lit. ‘as if ’), see Maschler 2001
and Chapter 4, respectively.
12. “The question mark (?) indicates a class of intonation contours whose transitional continu-
ity is regularly understood as an appeal [...] ‘Appeal’ here refers to when a speaker, in producing
an utterance, overtly seeks a validating response from a listener” (Du Bois et al.€1992: 30).
 Metalanguage in Interaction

38, 39) to dofek (‘is hitting’, present imperfective, line 46). Thus, this nu was indeed
interpreted as a request to move on to the following episode of the narrative.
Another such move can be seen in excerpt 3, line 70, when Tali’s nu? occurs at
the boundary between Yosi’s complication (Chafe 1994) describing his hastening
of the cashier (lines 68–69) and the climax (Chafe 1994) of the story (lines 73–76)
revealing the identity of the place–a sex shop. In fact, nu here can be seen as a di-
rect response to the well known evaluative strategy of delaying the point of a nar-
rative (Labov 1972), thereby increasing audience involvement.
This use of nu is reminiscent of the use described by Sorjonen for Finnish nii:
[T]he utterance to which nii(n) responds forms a place of maximum incomplete-
ness. That is, the speaker has reached a place in her talk after which the delivery of
a point of the activity is imminent. Before proceeding to the point she momentar-
ily stops her talk, thereby yielding a place for a possible response by the recipient
(202: 183).

However, Sorjonen does not discuss the interpersonal dimension of this momen-
tary stop preceding the use of nii. The nu of excerpt 3, line 70 shows a particularly
high degree of interpersonal involvement partly due to the iconicity (Becker 1982)
it reveals: Tali’s impatience in the interaction, concerning getting to the story’s
point, mirrors Yosi’s impatience in the storyworld, concerning discovering the
identity of the place.
One more example is, of course, the one opening this chapter. Ran’s nu (ex-
cerpt 1, line 41) is uttered precisely at the end of Orna’s episode describing how the
first of five women dropped out of the diving course on the first day: nish'àrnu
'árba banot... to... tóv! (‘we were left four girls... al... alright!’, lines 38–40). Follow-
ing Ran’s metalingual comment (lines 41–45) which begins with nu, Orna pro-
ceeds to the next episode, describing the second day of the course (lines 46–7).
It is in this sense that nu can reflect a high degree of interpersonal involvement
between speaker and hearer. By exhibiting their impatience in moving towards the
climax of a story–to the point of taking the liberty to control the flow of another’s
discourse–hearers can show maximal involvement in the narrative. This is another
aspect of the Israeli dugri (‘direct’) way of speaking (Katriel 1986), an aspect which
may be paraphrased in the following manner: ‘we are so close, that not only can I
tell you things without worrying too much about your positive face wants (Brown
and Levinson [1978] 1987, Katriel 1986, Blum-Kulka 1992), I also don’t have to
waste too much time on your negative face wants and can thus take the liberty to
speed up the flow of your talk’. Within this frame (Goffman 1981), impatience
contributes positively to the interaction, rather than being interpreted as impolite,
because it is viewed as indicative of the audience’s high involvement in the talk.
 Chapter 2.╇ The interpersonal realm 

This is in agreement with previous studies of the high degree of interpersonal

involvement in Israeli discourse (Blum-Kulka and Katriel 1991, Maschler 1994a,
Blum-Kulka 1997), and it is similar to the positive functioning of overlaps (Tan-
nen 1984) and argumentativeness (Schiffrin 1984) in the talk of Jewish Americans
of East European background. The parallelism is particularly apparent in the case
of overlap, because overlap and employment of nu can be viewed as two diametri-
cally opposed strategies: in the strategy of overlap, the hearer takes the liberty to
begin his or her talk without waiting for a transition relevance place which is to
follow the turn constructional unit being produced (Sacks, Schegloff, and Jeffer-
son 1974). In the case of nu, the hearer not only gives up the floor, but also takes
the liberty to hasten the speaker upon reaching a transition relevance place13, in-
dicating that the speaker should go on. In both cases there is an attempt to control
the flow of discourse. In the former case–in order to grab the floor, in the latter–in
order to hasten the other’s occupying of it.

4.2 Urging further development within a non-narrational topic

As an example of nu in non-narrational discourse, examine excerpt 7 from the

argumentative genre–a family dinner conversation between two parents, their
son, Gabi, and daughter, Shani, both in their early twenties. The context here is a
political argument between father and son, following the peace agreement signed
between Israel and Jordan in 1994.
Excerpt 7 (‘Political Argument’):
20 Gabi: .... 'atá lò maskím 'iti,
you don’t agree with me,
21 .. shehaheském hateritoryali,
that the agreement the territorial,
that the territorial agreement,
22 'im yardén,
with Jordan,
23 hu mutsláx?
is successful
is good?

13. This transition relevance place, however, is not necessarily a complex transition relevance
place (a ‘CTRP’, Ford and Thompson 1996), because intonation can be non-final at this point
(see below).
 Metalanguage in Interaction

24 Father: ... 'ìm 'anì lo maskim 'itxá,

if I don’t agree with you,
25 Shani: todá. {to Mother, concerning the food being served}
thank you.
26 Father: .... 'ìm 'anì lo maskim 'itxá,
if I don’t agree with you,
27 ... 'ata titén li lehagìd 'et da'ati?
you will let me say my opinion
will you let me express my opinion?
28 Gabi: ... /rak / tipa ki 'ani lo yaxol /????????/.
/just/ a bit because I can’t /????????/.
29 .. 'avál,
but,
30 .... 'ata maskím 'iti,
you agree with me
do you agree with me,
31 .. shehaheském hateritoryali,
that the agreement the territorial,
that the territorial agreement,
32 'im yardén,
with Jordan,
33 hu--
is
34 Father: ló.
no.
35 'eh.. /tir'e/,
uh.. look,
36 .. 'ani lò maskím 'itxa!
I don’t agree with you!
37 .. 'amárti lexa,
I’ve told you,
38 Gabi: láma lo.
why not.
39 Father: 'ata kól pa'am sho'èl 'oti,
you every time ask me,
40 'et 'otá sh
 Chapter 2.╇ The interpersonal realm 

the same qu
41 .. ló maskìm.
don’t agree.
42 Mother: shshsh--
43 Gabi: láma hu lo mutslàx?
why it’s not successful
why isn’t it good?
44 Mother: 'ani ló sovèlet,
I can’t stand [it],
45 ..she/'anaxnu tso'akím bamitbax/.
that /we scream in the kitchen/.
46 Father: 'ani 'asbír lexa,
I’ll explain to you,
47 làma ló.
why not.
48 Gabi: ... nú.
49 Father: ... xaké!
wait!
50 Gabi: ... nu,
51 tasbír!
explain!
52 Father: ... 'áleph,
a {first letter of Hebrew alphabet},
53 ...... 'im
if
54 Gabi: 'ata 'omèr leshem havikúax?
[are] you saying [this] for the sake of the argument?
55 Father: ... ló.
no.
56 Gabi: ..... nu 'az láma hu lo mutslàx.
nu so why it not successful
nu so why isn’t it good.
57 Father: .... mipney she,
because,
 Metalanguage in Interaction

Already at the beginning of this excerpt we learn that this is not these speakers’
first political argument and that the father feels his son doesn’t allow him to ex-
press himself (lines 26–27). At line 38 and then again at line 43, the son, Gabi, asks
his father why he views the peace agreement with Jordan as lo mutslax (‘not suc-
cessful’) or ‘not good’. The father answers announcing in a rather patronizing way
that he will explain to him why the agreement is not a good one: 'ani 'asbír lexa,
làma ló. (‘I’ll explain to you, why not’, lines 46–47). Perhaps as a response to this
patronizing tone, at line 48, employing nu, Gabi urges his father to get on with the
argument. But unlike the cases above, this particular urging to move on in the
development of the topic does not serve to move the discourse forward. On the
contrary; the father pauses even further in his argument and orders his son: xake!
(‘wait!’, line 49).
In the previous Section (4.1.1), we have seen that in casual Hebrew conversa-
tion, it is acceptable for one to attempt to control the flow of another’s discourse by
hastening the speaker, as long as it is clear that this is done in the name of genuine
enthusiasm concerning his or her talk. When a participant takes the liberty of
controlling another’s flow of discourse, a momentary inequality between partici-
pants is created in the interaction. In the argumentative context, such a move is
more likely to be perceived as ‘be done with your argument already, so that we can
get to my turn finally’. In such contexts, an attempt to control the other speaker’s
talk may well be perceived negatively, and the nu may be destructive to the flow of
talk. The speaker might decide to pause purposely and make the urging partici-
pant wait even longer, in protest of this hastening. This is precisely what happens
in the present interaction. Note the minor dueling over who will get to control the
flow of discourse in the exchange at lines 48–51. Gabi does not accept his father’s
order to wait, and once more attempts to control the situation with an elaborated
rephrasal of line 48: nu, tasbir! (‘nu, explain!’, lines 50–51), explicitly mentioning,
via the metalingual verb tasbir (‘explain’), the action which nu hastens.
We are now in a position to understand the infrequent employment of nu in
the talk of student to teacher in classroom discourse. Nu is highly inappropriate in
this context not only because this is a relatively formal situation, but particularly
because the student is perceived as attempting to control the discourse flow of a
participant quite higher than him/her in status. On the other hand, the frequency
of nu in the discourse of parents to their young children can be explained by the
parents’ higher status relative to the child and the informality of the situation.
Thus we see that the nature of the interaction can mitigate those aspects of nu
related to impoliteness. While nu usually does carry an impolite aura, at least in
speakers’ metalinguistic awareness, this aura is often neutralized by the nature of
the interaction and the relationship between its participants. Whether or not nu is
perceived as impolite and functions to obstruct the flow of talk depends on the
 Chapter 2.╇ The interpersonal realm 

genre and the general context of the interaction. Interpretation of nu, then, is an
emergent phenomenon, shaped by a variety of contextual constraints.

4.3 Nu as a continuer

In discussing English uh huh, yeah, mm hm and the like, Schegloff writes that a
continuer is a minimal vocalization
exhibit[ing...] an understanding that an extended unit of talk is underway by an-
other, and that it is not yet [...] complete. It takes the stance that the speaker of that
extended unit should continue talking, and in that continued talking should con-
tinue that extended unit. [It] exhibit[s] this understanding, and take[s] this stance,
precisely by passing an opportunity to produce a full turn at talk. (1982:€81).

This characterization of a ‘continuer’ fits nu with some modification: nu involves

not only taking the stance that the speaker should continue talking, but actively
urging the speaker to do so. The extended unit of talk underway can vary in size
from a single prosodic sentence14 to a whole narrative.
One Hebrew continuer is ken (‘yeah’), the most common interpersonal dis-
course marker throughout the database (over 240 tokens). Indeed, the functions of
ken and nu show some overlap, as can be seen from the following non-narrational
excerpt. This is a conversation between David, Avner, and myself. A certain degree
of tension exists between Avner and David throughout this interaction, resulting
from their different attitudes towards patriotism, about which I have written
elsewhere (Maschler 1997a). About half an hour into the beginning of this din-
nertable conversation, David summarizes why he wanted to go into an army pilot
training course after high school:
Excerpt 8 (‘Pilot Training’):
1 David: ... bekìtsur 'anáxnu hayinu patriyotim,
in short we were patriots,
2 Yael: .... kén?
yes?
3 Avner: .... nu--?
4 David: ... vekulà--nu ratsinu lihyot kraviyí--m,
and we all wanted to be battle unit [soldiers],

14. A prosodic sentence is defined by Chafe (1994:€137–145) as an intonation unit, or a group

of successive intonation units separated by continuing intonation, ending in sentence-final into-
nation. Chafe observes that once a speaker judges that the scanning of a center of interest (such
as a sub-episode in a story-episode) has been completed, s/he expresses that judgment with a
sentence-final intonation contour.
 Metalanguage in Interaction

5 .. ve--'éx ze holex hayom,

and how it goes today
and how does it go these days,
6 .. be--xèl 'avir kulám harey holxim,
in force air everybody PART. goes
to the air force, as you know, everybody goes,
7 .. ma,
.. what [do you mean],
8 ze--.. ze.. kavòd gadó--l.
it it honor great
it’s.. it’s.. a great honor.
David ends intonation unit 1 with continuing intonation. He pauses for a period
which both Avner and I judge to be too long, and we both employ continuers at
precisely the same moment (manifesting what Erickson and Shultz (1982) call
rhythmic synchrony). I employ the continuer ken? (line 2), and Avner employs nu?
(line 3), both in appeal intonation. David then continues his utterance in the fol-
lowing lines, ending his prosodic sentence at line 8. In this instance, then, nu func-
tions as a continuer, overlapping one of the uses of ken. However, the nature of the
interaction between Avner and David results in the flavor of Avner’s continuer be-
ing not only ‘I’m listening, go on’, as in the case of my ken?, but also ‘go on, let’s hear
what you’ve now got to say’.
Nu, however, can be employed as a continuer even when no particular tension
in the discourse is involved, as in the following narrative in which a young man
begins to tell two women friends about a revelation he had concerning the type of
woman he liked:
Excerpt 9 (‘Women’):
134 Eyal: .. li,
to me,
135 .. ze karà pa'am 'axát,
it happened one time,
136 .. kshehayiti be'ey--ze kontsért,
when I was at some concert,
137 Yafit: .... nu.
138 Eyal: be'avonotáy.
in [all] my sins
to my shame.
 Chapter 2.╇ The interpersonal realm 

Eyal begins to tell a story that took place when he went to some concert. He ends
intonation unit 136, the orientation to his narrative, in continuing intonation and
pauses for longer than average (indicated by 4 dots at line 137). A point of incom-
pletion is thus achieved. Yafit judges this pause to be ‘too long’, at which point she
employs nu as a continuer: she displays her understanding that a larger discourse
unit (a prosodic sentence in this case) is underway, and urges Eyal to continue its
production. Eyal indeed continues and adds be'avonotáy (lit., ‘in [all] my sins’), a
formulaic phrase roughly equivalent to English ‘to my shame’, only then ending in
sentence-final intonation (line 138). Going to a concert is apparently not a ‘cool’
thing to do among members of this social group, and Eyal jokingly admits to hav-
ing performed this ‘shameful’ deed. His narrative continues in a light key, with no
tension between participants throughout the interaction. This use of nu as a con-
tinuer, then, carries a neutral flavor, unlike the one of excerpt 8.
Another typical place for the continuer nu is right after an initial conditional
clause. The following excerpt comes from a conversation between three male students
in their early twenties arguing over the quality of their departmental newspaper. Here
they compare this newspaper (called perspektivi,‘Perspectival’) to the newspaper ed-
ited by the university’s student union (called pesek zman ‘Time Out’). The latter news-
paper’s readers are compared to the readers of Bazuka gum wrappers:
Excerpt 10 (‘Departmental Newspaper’):
122 Yoni: ... 'ím 'ata ben 'adà--m,
if you[’re a] person,
123 shel 'e--h,
of u--h,
124 ... sheyésh’xa koax lashèvet,
that there is to you energy to sit,
that you have the energy to sit,
125 velikró,
and to read,
126 velehit'amék,
and to go deeply [into things],
127 ... 'az 'ata tikrà perspektívi.
then you’ll read Perspectival.
128 Ido: ha.. ha'idiot hax
the the idiot the most
the bi[ggest] idiot
 Metalanguage in Interaction

129 Yoni: 'aval 'ím 'ata ben 'adàm,

but if you[’re a] person,
130 shekore bazúka,
who reads Bazuka,
131 Avner: nú.
132 Yoni: .. 'ata tikrà pesek zmán.
you’ll read Time Out.
Avner employs his continuer immediately following Yoni’s second conditional
clause 'aval 'ím 'ata ben 'adàm, shekore bazúka, (‘but if you’re a person, who reads
Bazuka’, lines 129–130). Yoni pauses just slightly past this conditional clause–long
enough for Avner’s continuer, which in fact overlaps the beginning of Yoni’s ‘then’-
clause 'ata tikrà pesek zmán. (‘you’ll read pesek zman’, line 132). Note that Avner
does not employ nu between the two clauses of the first conditional sentence in
this excerpt (lines 122–127), even though there is actually a longer pause between
these two clauses (line 127) in comparison to the barely audible pause in line 132.
We have here a parallel construction of two conditional sentences, contrasted by
'aval (‘but’), whose ‘if ’-clauses both open with the formula 'ím 'ata ben 'adàm, she
(‘if you’re a person, who’) and whose ‘then’-clauses both open with ['az]'ata tikrà
((‘then) you’ll read’). This parallelism makes a strong projection that the end of
this rather long figure is approaching, as well as a projection concerning what its
ending might be. Avner and Yoni are on the same side in this argument–they both
think the departmental newspaper is of high quality, while the student union’s
paper is not (Ido is the one who thinks less highly of it). Thus, Avner’s nu urging
Yoni at line 131 is neutral; it is not prompted by any particular tension between
himself and the person he is hastening, as was the case in excerpt 8, but rather by
a projection of the end of a rather long figure15.
Previous studies (e.g., Haiman 1978, Schiffrin 1992, Ford 1993) document the
function of conditionals in presenting given information. In her study of adver-
bial clauses in American English conversation, Ford writes that initial adverbial

15. However, the 7 tokens of nu throughout the three-minute male interaction from which
excerpt 13 is taken are one of many factors contributing to its somewhat more rough nature,
placing this conversation closer to a ‘masculine’ end of a continuum of discourse styles as a func-
tion of gender (Coates 1997). This is particularly so because employment of this token of nu
between two clauses of a conditional sentence–a likely place for the strategy of co-construction
(Lerner 1991) or co-production (Ferrara 1992), and particularly within the parallel construction
we have seen above–comes at the expense of employing this latter collaborative strategy. This is
one of the features placing this conversation closer to a ‘masculine’ end of a continuum of dis-
course styles, in which participants assume less collaborative footings (Goffman 1981) towards
each other (Maltz and Borker 1982, Tannen 1990, 1999).
 Chapter 2.╇ The interpersonal realm 

clauses, and initial conditional clauses in particular, “form pivotal points in the
development of talk and present explicit background for material that follows”
(1993:€62). This would motivate the function of nu to urge further development of
the (‘new’) topic in excerpt 10, particularly in this double contrastive conditional,
where the second initial conditional is ‘given’ not only because of its initial posi-
tion, but also because of its contrast with the previous conditional clause. Ford also
writes that “ordering an adverbial clause before its main clause insures that no
point of possible completion will be reached before the entire complex is deliv-
ered” (ibid.: 52). Nu between the adverbial and main clauses can be viewed as the
interlocutor’s response to this strategy: nu hastens precisely that part of the utter-
ance that presents the ‘given’ background and is there to delay reaching a point of
possible completion. Nu, then, can be a response to two types of delay on the part
of the speaker–delaying the point of a narrative (Section 4.1.1) or delay in reaching
a point of possible completion, further supporting its characterization as a marker
of impatience.
Clancy et al. define a reactive token as “a short utterance produced by an inter-
locutor who is playing a listener’s role during the other interlocutor’s speakership.
That is, reactive tokens will normally not disrupt the primary speaker’s speaker-
ship, and do not in themselves claim the floor” (1996:€356). They further subcate-
gorize reactive tokens into four groups, one of which is the backchannels–a non-
lexical vocalic form, serving as “a ‘continuer’ (Schegloff 1982), display of interest,
or claim of understanding” (Clancy et al. 1996:€359). This is relevant to those in-
stances of neutral nu functioning as continuers, as seen in the last two excerpts.
Clancy et al. compared reactive tokens in English, Japanese, and Mandarin, show-
ing that these languages differ in the reactive tokens they favor and in their
frequency and distribution in conversation. The goal of their study was “to exam-
ine the communicative strategies in each language with respect to culture-specific
expectations about the degree of interaction that the non-primary speaker will
engage in”. None of the languages investigated in their study manifest a reactive
token similar to nu. We have seen that in Hebrew discourse, there are culture-
specific expectations about a relatively high degree of interaction that the non-
primary speaker will engage in, to the point of allowing this non-primary speaker
to attempt to control the flow of the primary speaker’s talk.

4.4 Summary of the ‘urging further development of topic’ function

The majority of nu tokens in this database, then, are employed in order to urge a
speaker to move on in the development of a topic. This can be done in both nar-
rational and non-narrational topics, at discourse unit boundaries of various hier-
archical levels, from a prosodic sentence to a whole narrative. We saw that the
 Metalanguage in Interaction

general impolite aura accompanying nu in speakers’ metalinguistic awareness can

be neutralized by the nature of the interaction and the relationship between its
participants. In fact, the impatience exhibited by nu is often a sign of high involve-
ment and enthusiasm. Finally, we saw that this type of nu can be characterized as
a continuer, except that rather than just “taking the stance that the speaker should
continue talking” (Schegloff 1982), nu actively urges it.

5. Granting permission to perform an action

Another function of nu, not nearly as common as the function described in Section
4, is that of granting permission to perform some action. This ‘allowing’ of some
action, usually verbal, is often done somewhat reluctantly. Only 2.6% of all tokens
throughout the database function in this role. The following excerpt supplies two
examples. This is a conversation between a young woman in her early twenties,
Orna, and her parents. The mother is telling a story which is interrupted at line 193
by the father suddenly making a comment concerning the tape recorder:
Excerpt 11 (‘Family Gossip’):
192 Mother: .. mà 'ód hu 'amar,
what else he said
what else did he say,
193 Father: tagídi,
tell (fem. sg.) me,
194 .. 'at hexláft po batarìya?
you changed here battery
did you change batteries here?
195 Orna: .. ló.
no.
196 Father: ... 'át yoda'at shehadavàr haze 'ovéd
you know that the thing the this works
you know that this thing works
197 .... 'ani ganávti 'et ze--,
I stole this,
198 Mother: 'ani be'émtsa mishpàt.
I[‘m] in [the] middle of [a] sentence.
199 Orna: 'ima be'émtsa
Mom[‘s] in [the] middle of
 Chapter 2.╇ The interpersonal realm 

200 {laughter}
201 Mother: .. 'ani mishtagá'at mimxa!
I’m going crazy from you
you’re driving me crazy!
202 Father: tòv 'aval ze norá xashuv,
okay but it’s terribly important,
203 ladá'at 'et ze.
to know this.
204 Orna: {laughter}
205 Father: {laughter}
206 ... beshiv'ím veshmòne,
in seventy eight,
207 ganàvti 'et ze mimé'ir 'adív,
I stole this from Meir Adiv,
208 .. meshumásh.
used.
209 Orna: ... tizahér hakol maklitím po,
be careful everything is being recorded here,
210 .. 'ába.
Dad.
211 {laughter}
212 Father: ... ga
I st
213 .. ganavti bemerxa'ót.
I stole in quotation marks.
214 Orna: {laughter}
215 Mother: ... 'at shomá'at?
you (fem. sg.) hear
do you hear what I’m saying?
216 Orna: nu.
217 Father: hish'árti lo 'et hamaxshev sheli 'al hashulxan.
I left him the computer mine on the table
I left him my computer on the table.
218 Orna: 'ába,
Dad,
 Metalanguage in Interaction

219 .. 'ima hayta be'èmtsa mishpát!

Mom was in [the] middle of [a] sentence!
220 Father: ... nú.
221 Mother: ... 'az hu 'amàr la shehi 'itít,
so he told her that she is slow
so he told her she was slow,
In all of the excerpts of Section 4, nu was employed as a first pair part of an adja-
cency pair (Schegloff and Sacks 1973). In Excerpt 11, nu is used as a second pair
part. Rather than initiate a new move, a return to the main topic, each of the two
nu tokens in this excerpt responds to an utterance which itself initiated a new
move. In line 215 Orna’s mother addresses her daughter with 'at shoma'at? (lit.,
‘you (fem. sg.) hear?’ or ‘do you hear what I’m saying?’). This is a pre-pre (Schegloff
1980), as the mother here requests her daughter’s cooperation in getting back to
the story despite her husband’s interruption. Orna responds to this pre-pre with a
nu (line 216) granting the mother permission to go on with the story, i.e., as a go-
ahead token (Schegloff 1990). However, the father overlaps this nu with a continu-
ation of his justification for having ‘stolen’ the tape recorder back in 1978–in re-
turn, he left his own computer for the person whose tape recorder he ‘stole’
(line 217). This further attempt to make himself look better is interrupted by Or-
na’s objection, a partial repetition of lines 198–199 earlier: 'aba, 'ima hayta be'emtsa
mishpat! (‘Dad, Mom was in the middle of a sentence!’, lines 218–219). This objec-
tion functions also as the daughter’s request (and thus, first pair part) that her fa-
ther stop interrupting and let her mother go on with the telling. The father agrees
to this somewhat reluctantly, as the prosody of line 220 suggests, with another nu
in second position, responding to Orna’s first pair part of lines 218–219, and al-
lowing the mother to continue her story, as she indeed does at line 221.
Although relatively few tokens of nu function in this role in the corpus, family
discourse in which young children are present (a context abundant with pre-pre’s,
as Sacks (1972) has shown) often manifests such uses of nu, as, for example, in the
following interaction between my then nine-year-old daughter and myself:
Excerpt 12 (‘Ballet’):
1 Maya: ... 'ima,
Mom,
2 ... yesh li she'ela.
there is to me question
I have a question.
3 Mother: ...nu.{impatiently, in the midst of preparations for leaving the
house}
 Chapter 2.╇ The interpersonal realm 

4 Maya: ... 'ani lo yoda'at ma la'asot,

I don’t know what to do,
5 ... ki,
because,
6 .. 'im 'ani 'eshan 'etsel hadasi,
if I sleep over at Hadasi’s,
7 .. 'ani 'etstarex lehafsid,
I’ll have to miss,
8 shi'ur balet.
lesson ballet
[my] ballet lesson.
Again we find a pre-pre (Maya’s announcing her upcoming question, lines 1–2).
The tone of my nu (line 3) is rather impatient, and allows her to proceed with the
requested act somewhat reluctantly, thus functioning as a go-ahead token. Unlike
the impatience in most of the previous excerpts, here it is definitely not genuine
interest in the topic that motivates the tone. In this instance, we might tie the im-
patience to the rush in the midst of which this utterance occurred, but many such
nu tokens are found in our family’s discourse when no special rush is involved.
This use of nu is very similar to the use of Finnish no described by Sorjonen:
[A]s a turn of its own no typically provides a response to coparticipant’s prior ut-
terance that acts as a preliminary (pre) to something else yet to come [...] The
particle no offers a second pair part to the preliminary: it provides a ‘go-ahead’
response (Schegloff 1990:€61), that is, it invites the coparticipant to get on with the
production of the main action. (2002:€166–167).

The adjustment I would suggest for Israeli discourse is that rather than inviting the
coparticipant to get on with the talk, nu allows it, and often rather reluctantly.
Thus, while Finnish employs two different particles, one–nii(n)–urging fur-
ther development within the topic, the other–no–functioning as a ‘go-ahead’ to-
ken, in Hebrew, a single utterance–nu (originating probably from the same Ger-
manic source as the two Finnish particles)–performs both functions.
Lines 43–53 from excerpt 7 are telling for understanding how it might come
about that a single form will come to be used in these two functions. The nu of line
48 is a borderline case between urging further development within a main topic
and granting permission to perform an action. If lines 46–47 'ani 'asbír lexa, làma
ló. (‘I’ll explain to you, why not.’) are viewed as a response (and, therefore, a second
pair part) to line 43 láma hu lo mutslàx? (‘why isn’t it [the peace agreement with
Jordan] good?’), then the nu of line 48 can be viewed as a first pair part of the fol-
lowing sequence, in which Gabi is urging his father to go on with the explanation.
 Metalanguage in Interaction

But if we view lines 46–47 as beginning something new, rather than as responding
to line 43 (and the mother’s comment at lines 44–45 can certainly be said to break
the continuity of the previous sequence), then 'ani 'asbír lexa, làma ló. (‘I’ll explain
to you, why not.’) can be viewed as a pre-pre. We have seen that nu granting per-
mission to perform an action generally occurs in response to a pre-pre, and often
reluctantly so. Viewed this way, the nu at line 48 can be interpreted as a very
reluctant granting of permission by the son to his father to proceed with his expla-
nation. It is these ambiguous cases that constitute transitional examples between
the function of nu to urge further development of a topic and the function of
granting permission to perform an action.

6. Summary of sequential functions of nu

All of the functions of nu seen so far in Sections 3, 4, and 5, have to do with mov-
ing the (verbal) action along, and thus have sequential implications. I would like
to suggest a continuum of ‘degree of encouragement to proceed with action’ along
which this discourse marker could be described. The continuum includes (from
‘most encouraging’ to ‘least encouraging’): hastening, urging, granting permis-
sion, allowing, and allowing reluctantly, and it involves both metalingual and non-
metalingual actions. All of the tokens of nu examined thus far fall at various points
along this continuum.

7. Keying functions of nu: joking/provoking

In sharp contrast to the previous three categories, which all have to do with se-
quential functions, the fourth category, consisting of about 23.5% of all nu tokens,
has to do with the notion of key. Structurally, what is common to these tokens is
that they are always followed by additional talk of the same speaker, very often
within the same intonation unit. This additional talk does not specify an action
being urged (as in excerpt 3, for instance). Keying nu adds a joking or provoking
tone to the following utterance and it often involves some resultative meaning.
Excerpt 2, line 128 constitutes such an example. At 126 David declares some-
what dramatically: 'axsháv nit'am mehamara/ve/k. (‘now we’ll taste the soup’). At
line 127 he proceeds to taste it, after which Dalia asks nu 'éx yatsa? (‘nu how did it
come out?’, line 128). Dalia’s question is a result of David’s declaration of the action
he was about to engage in and of his actual tasting of the soup. Nu adds a joking
tone to Dalia’s utterance, perhaps in response to David’s dramatic declaration of
this rather mundane act.
 Chapter 2.╇ The interpersonal realm 

7.1 From joking to provoking

A clearer example of the joking key is found in excerpt 13. Shani constructs here
an utterance which was said upon the descent of her boyfriend, Shaxar, her father,
and herself from an approximately $80,000 new jeep of a friend:
Excerpt 13 (‘Jeep’):
122 Shani: ... basof sháxar /??? /
in the end Shaxar /??????/
123 'e--h yarádnu,
uh we came down [from the jeep],
124 .. 'az 'ába sheli 'omèr leshaxar,
so father my says to Shaxar
so my father says to Shaxar,
125 .. nu mat'ím lexa jìp kazè?
nu suits to you jeep like this
nu does a jeep like this suit you?
126 .. 'az /ab/ shàxar 'osé lo,
so /my f/ Shaxar does to him
so /my f/ Shaxar goes,
127 ló!
no!
128 Dalia: {laughter}
129 Shani: {laughter}
130 ... ka'éle xamudìm,
like these cute
so cute,
131 .. ya 'ála!
boy!
The joking key originates from the fact that it is evident to all that such a jeep is
way beyond Shaxar’s means. Without this nu, it would have been possible to ask
Shaxar seriously whether such a jeep suited him. Following this utterance, Shani
tells (line 126) that Shaxar responded to this question in the same joking key lo!
(‘no!’, line 127), i.e., that such a jeep does not suit him, both women laugh at the
telling of this short interchange, and Shani concludes with an evaluation of her
boyfriend and her father: ka'éle xamudìm, ya 'ála! (‘so cute, boy!’)–all further sup-
porting the joking reading of this excerpt. The resultative meaning comes from the
 Metalanguage in Interaction

fact that her father’s question is relevant as a result of the fact verbalized in the
preceding utterance (line 123)–that they had just descended from this jeep.
The boundary between joking and provoking is, of course, not always clear. In
the following conversation, which took place the morning after a party, Amir is
telling his girlfriend Tami about the breakup of another couple:
Excerpt 14 (‘Breaking Up’):
19 Tami: .. hu nifràd midalít?
he broke up from Dalit?
20 Amir: ken.
yeah.
21 Tami: .... 'ani lò ma'aminá.
I don’t believe
I can’t believe it.
22 Amir: ... nu ma xashávt,
nu what did you think,
23 shehu bá 'etmòl e--h,
that he came yesterday u--h,
24 .... levád,
on his own,
25 .. ki bá lo?
because came to him
because he felt like it?
Amir is half mocking, half ridiculing Tami for not having understood why the
man came to the party unaccompanied by his girlfriend Dalit. He does so by em-
ploying the utterance nu ma xashavt (‘nu what did you think’, line 22) followed by
the ridiculous interpretation of the event, that the man showed up at the party on
his own because he happened to feel like it (lines 23–25). Nu allows Amir to con-
struct this partly mocking, partly ridiculing tone.
The doubt concerning whether the key is mocking or ridiculing/provoking, is
no longer apparent in the case of the political argument of excerpt 7, when Gabi
asks his father nu 'az láma hu lo mutslàx (‘nu so why isn’t it good’, line 56 ), about
the peace agreement with Jordan, following the dueling over who will get to con-
trol the flow of discourse (lines 48–51). The general tone of the conversation at that
point indicates that we are no longer dealing with mocking or even with ridiculing.
A provocation, quite disrespectful, is at issue. The resultative meaning is construct-
ed here also by the discourse marker 'az (‘so’), but the disrespect is provided not by
'az, but rather by the nu, as well as by the prosodic elements of the utterance.
 Chapter 2.╇ The interpersonal realm 

We have seen that the nature of the interaction and the relationship between
its participants determines whether or not the impatience exhibited by nu is per-
ceived as impolite, and whether or not the continuer nu functions neutrally. Simi-
larly, whether keying nu functions jokingly or provocatively is also a matter of the
context in which it is employed. In the present corpus, out of the 27 tokens of this
type of nu, 12 are employed in a joking key whereas 15 are used provocatively. This
suggests a characterization of Israeli humor; namely, that the boundary between
humor and provocativeness in this culture is elusive and blurry. However, more
study is needed in order to support this claim.

7.2 Nu in combination with other discourse markers

Keying nu often occurs in conjunction with other discourse markers. In this data-
base, it is found most frequently (in 8 out of 27 keying tokens) with 'az (‘so’) (sup-
porting the resultative meaning frequently involved in its use), as well as with
betax (‘sure, of course’), bexayex (‘come on’, lit. ‘in your life’), and ve- (‘and’). An-
other common cluster is nu be'emet (‘nu really’), though it is not found in the
present database16. Nu is almost always the first member of these discourse marker
clusters. I elaborate here on only one combination, because it is particularly rele-
vant to the interpersonal function of nu.

7.2.1 Nu in combination with 'az ma? (‘so what?’)

The idiom 'az ma? is the Israeli version of the much-discussed ‘so what?’ question
(Labov 1972:€366). As Labov has noted in relation to narratives (and as Tannen
(1989) has observed in relation to non-narrational discourse), this question is the
worst response to be encountered by a storyteller upon ending their story. Much
of what conversationalists do is ward off this question by continually employing
involvement creating strategies throughout their talk.
The ‘so what?’ question is accompanied by an additional threat for the Israeli
storyteller. The Israeli hearer might not only wonder about the worthiness of the
telling. This hearer might also take the opportunity to mock/ridicule/provoke/act
derogatorily towards the speaker for not having considered this matter before-
hand. This additional tone is what nu provides when added to the 'az ma? ques-
tion, as, for example, in the continuation of excerpt 3:

16. Surprisingly, Maschler and Estlein 2008 found only one instance of this cluster in the ex-
tended corpus, comprised of 91 conversations (270 minutes of talk among 223 different speak-
ers), adding 41 conversations to the present corpus.
 Metalanguage in Interaction

Excerpt 3a (continued) (‘Cashiers’):

73 Yosi: ... 'az hi 'oméret,
so she says,
74 .. tishmá,
{����---pp---}
listen,
75 .... lakxú 'otanu,
{-------pp----------}
they took us,
76 lexanùt le'avizarèy mín.
{--------pp---------} {----laughing-----}
to a store for accessories of sex
to a sex shop.
77 {laughter}
78 Tali: {laughter}
79 ... habe'alím shelahem?
the husbands theirs
their husbands?
80 Yosi: .... habe'a
the husb
81 .. habe'alím,
the husbands,
82 .. 'o ha'áx shel habà'al,
or the brother of the husband,
83 .. /???????/.
84 {laughter}
85 Tali: {laughter}
86 Yosi: ... 'ani 'omér,
{laughing}
I say,
87 .. nú,
{laughing}
88 .. 'az má--,
{laughing}
so what,
89 .. yófi--!
 Chapter 2.╇ The interpersonal realm 

{laughing}
great!
90 ... hi 'oméret li,
she says to me,
91 .. 'aval hadór shelí,
{imitating cashier’s talk (Russian accent?)}
but the generation mine
but my generation,
92 .. shonè legámre--y!
{imitating cashier’s talk (Russian accent?)}
different completely
is completely different!
93 ... /???? holex lesham/.
{imitating cashier’s talk (Russian accent?)}
/???? goes there/.
94 .. 'az nú,
so nu,
95 .. 'az hirgàsht tse'irá,
so you felt young,
96 .. 'az má yesh.
{laughing}
so what is there
so what’s the big deal.
97 {laughter}
98 Tali: {laughter}
We have already seen that Yosi views these cashiers’ bashful attitude towards sex
shops as ridiculous. After recounting the cashier’s revealing of the identity of the
place to him, and after much laughter on the part of both Yosi and his wife, Yosi
proceeds to recount his response to the cashier (lines 86–89): 'ani 'omér, nu, 'az
ma, yofi! (‘I say, nu, so what, great!’). This is an explicit instance of the dreaded ‘so
what?’ question, ornamented by the extra ridiculing tone provided by nu. After
reporting the cashier’s objection to this unimpressed and ridiculing response, say-
ing (in what sounds like a poor imitation of perhaps a Russian accent) that in her
generation things are different (lines 90–93), Yosi recounts (back in his own voice)
his further belittlement of the point of the cashier’s story: 'az nú, 'az hirgàsht tse'irá,
'az má yesh. (‘so nu, so you felt young, so what’s the big deal’).
 Metalanguage in Interaction

The occurrence of nu in conjunction with the ‘so what?’ question is related, of

course, to the fact that nu is often employed at points of maximal incompleteness,
such as right before the climax of a narrative in response to the storyteller’s delay
of the delivery of the point, as illustrated in excerpt 3, line 70. Nu preceding the ‘so
what?’ question is, in fact, the mirror image of nu immediately preceding the cli-
max of a narrative, because it is employed immediately following the climax, and
rather than exhibiting enthusiasm, it communicates a lack of enthusiasm concern-
ing the point.

8. Grammaticization of nu

We have seen that nu is the second most common interpersonal discourse marker
throughout the database, and that speakers in this corpus employ nu most fre-
quently (in 68.7% of the cases) in order to urge further development of an ongoing
topic. This suggests that part of the Israeli discourse experience is to allow a fairly
high degree of impatience in casual interaction. Impatience is manifested in hear-
ers’ attempts to speed up the flow of their interlocutors’ discourse.
However, despite the impolite aura nu carries in Israelis’ metalinguistic aware-
ness, the impatience constructed by this discourse marker in Hebrew casual con-
versation is not necessarily perceived negatively. Quite the contrary. We have seen
that impatience is most often interpreted in this corpus as indicative of the
audience’s high involvement in the speaker’s talk. By exhibiting their impatience in
moving forward in a topic–to the point of attempting to control the flow of an-
other’s discourse via nu–hearers show maximal involvement in the talk. Thus, the
nature of the interaction and the relationship among its participants can neutralize
the impolite aura often associated with nu. In this way, its meaning is emergent
from the discourse.
We have also seen that of the four main functions of the discourse marker nu,
three have to do with moving the (verbal) action along, and thus involve the se-
quential aspect of talk: hastening non-verbal actions, urging further development
within a topic, and granting permission to perform an action. The fourth function
has to do with key; namely, providing a joking/provoking tone.
I have already related to the functional itinerary (Traugott 1995a) of nu with
respect to the first three functions. Of course, without a corpus providing dia-
chronic perspective (as in Traugott 1989 or Traugott and Dasher 2002, e.g.), the
path of development of nu can only be hypothesized. I suggested the primacy of nu
as a hastener of non-verbal actions. The one example provided in the dictionary
(Avneyon 1997) which is relevant to nu in casual conversation–nu kvar (‘nu al-
ready’), hastening mostly non-verbal actions–provides partial support for this, as-
 Chapter 2.╇ The interpersonal realm 

suming examples given in dictionaries have something to do with providing rep-

resentative, basic examples. Certainly, this is the first function learned by children
acquiring nu. More support comes from studies of grammaticization, claiming
that such processes generally happen in the less metalingual to more metalingual
direction (e.g., Sweetser 1988, Traugott 1988), but only external evidence of nu
from Yiddish, Russian and earlier periods of Hebrew will clarify this point.
We have seen that just as one can be urged to perform a non-verbal action, so
can a speaker urge an interlocutor to perform a verbal action. Sequential nu was
described along a continuum of ‘degree of encouragement to proceed with action’,
moving from ‘most encouraging’ to ‘least encouraging’: hastening, urging, grant-
ing permission, allowing, allowing reluctantly. Nu could be described along an-
other continuum–that of key, with joking and acting derogatorily towards address-
ee on either end, and the additional possibilities (mocking/ridiculing/provoking/
provoking disrespectfully) in between. (These two continua are illustrated sche-
matically by the diagonal lines in Figure 1 below). Thus, nu is ambiguous both with
respect to its sequential nature as well as with respect to key. Again, the meaning of
a particular token emerges from the context in which it is employed.
The question is, how is it that a word involved in encouragement to proceed
with an action comes also to have a keying function in discourse? I would like to
suggest that the keying function of nu originates in the impolite aura nu may carry
in its sequential functions. Attempting to control the actions of another is inher-
ently impolite; indeed in some cases we have seen that nu is accompanied by a
derogatory tone (excerpt 7, e.g.). However, when the nature of the interaction neu-
tralizes its inherent impoliteness, the speaker employing nu will often do it hu-
morously (excerpt 1, e.g.). The humor can compensate for the impoliteness some-
times associated with this type of move, and it can also attribute the utterance with
a somewhat self-mocking tone, as if saying ‘I know I’m doing something which
could, in some contexts, be considered impolite’.
To further explain how a token functioning in the sequential realm of dis-
course comes to function also in its interpersonal realm, we need to posit yet a
third continuum along which nu could be described–that of metalanguage (illus-
trated by the horizontal line in Figure 1). On one of its ends would be nu urging
non-metalingual actions (such as in excerpt 2, line 142); on the other would be nu
urging metalingual actions via explicit metalingual utterances. The majority of cas-
es fall in between these two ends–urging metalingual actions (mostly, further de-
velopment of a topic) unaccompanied by a longer metalingual utterance. In other
words, sometimes the metalingual utterance is spelled out, as in excerpt 7, lines
50–51: nu, tasbir! (‘nu, explain!’). But more frequently, the metalingual utterance is
only implied and is to be gathered from context, as in excerpt 6, line 45: nu?.
 Metalanguage in Interaction

This is the case particularly with the tokens in the fourth category; namely, key-
ing nu. The metalingual utterance is seldom spelled out in these instances. For exam-
ple, the nu of excerpt 13, line 125 nu mat'ím lexa jìp kazè? (‘nu does a jeep like this
suit you?) can be interpreted as urging an implied metalingual utterance such as nu
tagid li, or nu ma 'ata 'omer, mat'ím lexa jìp kazè? (‘nu tell me’ or ‘nu what do you
say, does a jeep like this suit you?’), or the nu of excerpt 3a, lines 94–95 'az nu, 'az
hirgasht tse'ira (‘so nu, so you felt young’) can be interpreted as implying a longer
metalingual utterance such as 'az nu, ma 'at mitlonenet, 'az hirgasht tse'ira (‘so nu,
what are you complaining about, so you felt young’). Other metalingual utterances
implied by keying nu tokens found in the corpus are: nu, 'ata lo mevin she... (‘nu,
don’t you see that...’), nu, tasik kvar 'et hamaskana (‘nu, draw the conclusion al-
ready!’), or even nu 'az tasbir li 'im 'ata yaxol, ve'ani lo ma'amin she'ata yaxol (‘nu
so explain to me if you are able to, and I don’t believe you are’) in the political argu-
ment of excerpt 7, line 56: nu 'az láma hu lo mutslàx (‘nu so why isn’t it good’).
An interaction around our dinnertable supports this suggestion. Yotam, 7
years old at the time, was telling us how he and his friends were cracking unshelled
pine nuts with hammers. His 11-year-old sister Maya said that she preferred crack-
ing them with a big stone, because this doesn’t crack the meat inside. To this, Yo-
tam responded in a slightly belittling tone:
Excerpt 15 (‘Pine Nuts’):
Yotam: nu gam 'im patish ze lo shover.
nu also with [a] hammer it doesn’t break [them].
When I asked him why he said nu there, Yotam couldn’t answer the question.
Maya, however, said that it was because she didn’t understand that when he had
told us about cracking the nuts with a hammer, hu kvar hitkaven shegam 'im patish
hem lo nishbarim (‘he had already meant that also with a hammer they don’t
break’). In other words, this nu stands for a longer metalingual utterance such as
nu 'at lo ro'a she'hitkavanti shegam 'im patish ze lo shover (nu don’t you see that I
meant that also with a hammer they don’t break’). Thus, a longer metalingual ut-
terance is implied in her interpretation of this belittling nu token.
Figure 1 summarizes the three continua along which nu could be described.
Had the keying nu tokens occurred accompanying such explicit metalingual
utterances, they might have carried a derogatory, impolite aura, as in excerpt 7,
lines 50–51: nu, tasbir! (‘nu, explain!’). Alternatively, their impolite aura could
have been neutralized by the nature of the interaction, and a humorous tone could
have replaced it, as in many of the nu tokens throughout the data (e.g., excerpt 1,
lines 41–44). Accordingly, what is left of these longer metalingual utterances–the
nu–embodies only the key, from derogatory to joking; it verbalizes only the tone
encompassed by the longer utterance vaguely in the background. Like the smile of
 Chapter 2.╇ The interpersonal realm 

the Cheshire Cat, then, the keying nu token is all that remains of the longer meta-
lingual utterance implied. In this way, a word functioning in the sequential realm
of discourse comes also to have an interpersonal keying function.

joking reluctantly
allowing action
key nu
urging non- urging metalingual action
metalanguage
metalingual action via metalingual utterance
sequentiality
hastening action derogatory

Figure 1.╇ Nu along the dimensions of sequentiality, key, and metalanguage

 chapter 3

The textual realm

The discourse marker bekitsur:
Retroactively constructing digressions

1. Introduction

The Hebrew word bekitsur (‘succinctly, in short’) is a Modern Hebrew manner ad-
verb associated with the adjective katsar (‘short’), found already in Biblical texts.
Bekitsur is composed of the preposition b(e)- (‘in’, ‘with’)1 attached to the verbal
noun2 kitsur (‘doing something in short, shortening’) of the verb lekatser which
came to be used only later on, starting in the Talmudic period. Bekitsur is not found
as a separate entry in Even-Shoshan’s 2003 dictionary, but in a sub-entry of the ad-
jective katsar (‘short’) it is listed as an equivalent of another manner adverb from
the same root, biktsara, a higher-register equivalent of bekitsur (also from the Mod-
ern period): lo ba'arixut, bekitsur, bemilim mu'atot (‘not at length, succinctly, in few
words’). An example in fairly high register is provided: saper-na biktsara 'ex kara
hadavar. A translation might be: ‘pray-tell succinctly how the matter happened’.

1. The adverb bekitsur belongs to a group of adverbs, all composed of the preposition b(e)- at-
tached to a nominal form associated with an adjective, such as: ba'arixut (‘at length’, lit. ‘with
length’), bimhirut (‘quickly’, lit. ‘with quickness’), bizrizut (lit. ‘with agility’), be'itiyut (‘slowly’, lit.
‘with slowness’), bixvedut (‘heavily’, lit. ‘with heaviness’), bekalut (‘easily’, lit. ‘with ease’), etc. In
fact, prefixing adjectival nouns with the b(e)- morpheme is a productive way of forming adverbs
in Hebrew.
2. This verbal noun belongs to the verbal group (binyan) pi'el, which in 3rd person sg. past
tense consists of the three verbal root consonants, √k.ts.r., the first two separated by the vowel
i, the second two by the vowel e: kitser (masc. sg. past, ‘he lessened, reduced’ (Biblical period),
‘he did something in short’ (Talmudic period), or ‘he shortened’ (Modern period (Enlighten-
ment period and on)).
The symbol ts in the middle root consonant (sometimes transcribed as c, as e.g. in Shloush
1998), denotes here the affricate composed of the stop t and the fricative s, which constitutes a
single phoneme in Hebrew.
Note that the adjective katsar is found already in Biblical Hebrew (though only in the form
ktsar-, as the first member of the smixut construction (‘construct state’), e.g., ktsar-yad (lit.
‘short-handed’, ‘weak, incapable of doing things’)), whereas the verb lekatser meaning ‘to short-
en’ is not found before the Modern period (Even-Shoshan 2003, Avneyon 1998).
 Metalanguage in Interaction

However, in the corpus of 50 casual conversations on which the present study

is based, bekitsur is never found to function as an adverb. A typical usage is given
in excerpt 1, a conversation between two young women in their early twenties, in
which Hanna tells the story of how she was wounded during her army service at
the Lebanese border. In the orientation (Labov 1972) to the narrative, she describes
what she and her women soldier companions were wearing – heavy equipment
which caused her to stumble and “be wounded”:
Excerpt 1 (‘Wounded in Lebanon’):
91 Hanna: .. hem 'e--h sàmu 'aleynu shaxpáts.
they u--h put on us [a] shaxpats {a type of bullet-proof vest}.
92 ... 'aval kavéd kazè,
but heavy like3,
93 .. mehaye--shaním.
of the old [ones].
of the old kind.
94 Galia: .. max.. matól neged tànkim?
{------------laughing-----------}
max.. matol {a type of weapon} against tanks?
95 Hanna: né--shek,
[a] gun,
96 Galia: {laughter}
97 Hanna: {laughter}
98 .. né--shek,
{---laughing--}
[a] gun,
99 ... shéva maxsaniyòt mele'òt,
seven magazines full
seven full magazines,
100 Galia: má 'at 'omèret,
what are you saying,
101 .. she'áf 'axat ló yoda'at lirò--t betax.
that no one not knows to shoot probably
when no one knows how to shoot probably.
102 Hanna: ... máshehu kazè.
something like that.

3. For a study of Hebrew kaze, one ‘equivalent’ of English like, see Maschler 2001.
 Chapter 3.╇ The textual realm 

103 .... xagór,

[an] ammunition belt,
104 ... sh--téy meymiyòt,
two canteens,
105 .. shehayu tsrixót lihiyot,
that had to be,
106 .. mele'ót kòl 'axat,
full each one
full each,
107 Galia: mhm.
108 Hanna: .. ve--kaplá--d!
and [a] kaplad!
109 ... kaplá--d!
[a] kaplad!
110 Galia: ze,
it’s,
111 ... má ze kaplád?
what[’s a] kaplad?
112 : kaplád,
113 ... ke'ílu,
like4,
114 .. kasdá,
[a] helmet,
115 .. 'avàl mipladá,
but [made out] of steel,
116 Galia: 'áh!
oh!
117 Hanna: ... ló ha'ele hakalòt.
not the those the light
not those light ones.
118 Galia: mhm.
119 ... s.. slixá 'al haburùt.
s.. sorry for the ignorance.

4. For the study of Hebrew ke'ilu, another ‘equivalent’ of English like, see Chapter 5.
 Metalanguage in Interaction

120 Hanna: ... ló meshane.

doesn’t matter
never mind.
121 ..... bekitsú--r,
122 az 'e
so uh
123 ... 'e--h,
u--h,
124 ... baleyló--t,
during the nights,
125 ... 'axarey shehayinu mexapsòt lalevanòniyot bayamím,
after that we would search the Lebanese (fem.pl.) during the
days
after we would search the Lebanese women during the days,
126 .... baleylòt hayinu yoshvót 'e--h,
during the nights we would sit u--h,
127 ba.. xamál.
in the.. xamal {acronym5 for war room}.
Bekitsur of line 121 appears as a separate intonation unit, is not part of any gram-
matical sentence, and cannot be said to function as an adverb of any verb. It occurs
after the rather long orientation describing the equipment Hanna and the women
were carrying, which includes a clarification sequence concerning the blending
kaplad (‘steel helmet’) (lines 110–120), which Galia was not familiar with. At the
end of this clarification sequence, Galia apologizes half-jokingly for not knowing
what a kaplad was: s.. slixa 'al haburut. (‘s..sorry for the ignorance’, line 119), Han-
na accepts with lo meshane (‘never mind’, line 120), pauses a rather long pause, and
then moves on to describe the location of the incident (lines 122–127). This is still
part of the orientation to her narrative, but a different sub-episode (Chafe 1994) of
it – one describing the location rather than the equipment. The move to this new
sub-episode, and out of the preceding clarification sequence, apology, and accept-
ance, is signaled by bekitsur (line 121).
In previous studies of the system of discourse markers segmenting Hebrew
talk-in-interaction (Maschler 1997a, 2002a. See Table 2, Chapter 1), bekitsur was
classified, along with its slang variants hakitser and 'akitser, as a textual discourse

5. Israeli army parlance is noted for its pervasive invention of acronyms and blendings. The
focus of the clarification sequence of this segment, kaplad, is another case in point: kasda (‘hel-
met’) miplada (‘made out of steel’).
 Chapter 3.╇ The textual realm 

marker denoting the function ‘back to main action’, somewhat equivalent to that
of the English discourse marker anyway (e.g., Kussmaul 1978, Owen 1985, Alten-
berg 1986, Schourup and Waida 1987, Chafe 1987, Schiffrin 1987, Bublitz 1988,
Lenk 1995, 1998, Ferrara 1997, Takahara 1998, Tabor and Traugott 1998).
Recall that in the present study, a prototypical discourse marker is defined as
an utterance fulfilling two conditions; one semantic, the other structural:
(a) Semantically, the utterance must have a metalingual interpretation in the con-
text in which it occurs. In other words, rather than referring to the extralin-
gual world, it must refer metalingually to the realm of the text, to the interac-
tion between its participants, or to their cognitive processes.
(b) Structurally, the utterance must occur at intonation-unit initial position, ei-
ther at a point of speaker change, or, in same-speaker talk, immediately fol-
lowing any intonation contour other than continuing intonation. It may occur
after continuing intonation or at non-intonation-unit initial position only if it
follows another marker in a cluster (Maschler 1998b: 31)6.
For example, bekitsur of excerpt 1 satisfies both conditions. Structurally, it occurs
at intonation-unit initial position following final intonation (line 120) (here in
same-speaker talk). Semantically, bekitsur cannot be interpreted to function liter-
ally here – the ensuing discourse is neither ‘short’ nor ‘succinct’ and, as we have
seen, there is no verb in the surrounding intonation units to which bekitsur could
function as adverb. This word refers to the realm of the text, rather than to the
extralingual world. In other words, it functions metalingually, rather than lingual-
ly, and signals a move from the aside clarifying the word kaplad back to the main
topic of (the orientation to) the narrative.
Bekitsur is reminiscent of the English discourse marker anyway, marking “prior
discourse as tangential to the main point” (Schiffrin 1987:€165). Another argument
for the discourse marker status of bekitsur relies on the tendency of discourse mark-
ers to cluster at conversational action boundaries7: in excerpt 1, bekitsur opens a
cluster of four discourse markers: bekitsur,..'az 'e...'e--h (‘bekitsur,..so uh... u--h’).
When linguistic elements come to function as discourse markers, they often
undergo prosodic and phonological changes (Schiffrin 1987:€328, Ferrara 1997).
This is apparent in a series of prosodic and phonological changes undergone by

6. In Chapter 1 we saw that the two criteria in this definition coincide for 94% of the discourse
markers. The remaining 6% satisfy the semantic, but not the structural requirement. These
statistics are based on a sub-corpus consisting of 16 conversations (40 minutes of discourse
among 43 participants, see Maschler 2002a). In the present extended corpus of 50 conversa-
tions, only 2 out of 49 instances of bekitsur (4.1%) did not satisfy the structural requirement. We
will examine one of these cases below.
7. See Chapter 1, Section 6.0.
 Metalanguage in Interaction

bekitsur and its variants, beginning with the change in the position of the stress of
this word from bekitsúr to bekítsur8, following with a vowel change to bekítser, the
deletion of the prepositional component b(e)- in hakítser, 'akítser, and kitsúr, and
on to the more complicated sound changes (modeled after Arabic word patterns)
of be'uktsur, all of which are found in the present corpus. Other, more playful forms
which I have heard on occasion are bekítskets, hakítskits, and even 'akíbitser9.
Based on made-up examples, and paying no attention to prosody, Shloush
studied bekitsur with reference to relevance theory. Her work will be discussed
below. The goal of the present chapter is to deepen our understanding of the func-
tions of bekitsur and of the conversational actions of digressing and returning from
a digression, based on the study of naturally occurring Hebrew conversation. First,
I examine the employment of bekitsur in my corpus and compare it to that of Eng-
lish anyway. Second, I suggest a path of grammaticization (Hopper 1987, 1988,
1991, Hopper and Traugott 1993, Bybee and Hopper 2001) for bekitsur. Finally, I
consider this discourse marker in the wider context of Israeli culture.

2. Data

In the 50 conversations of the Haifa Corpus of Spoken Israeli Hebrew (Maschler

2004) on which the present study is based (see Chapter 1, Section 3), altogether 49
tokens of bekitsur are employed. Thus, bekitsur is employed at an average frequen-
cy of one token every 3.06 minutes in this corpus. The tokens are distributed
among the following functions:
Table 1 shows that, far and away, the main function of bekitsur in this corpus
(approximately 77.6% of all tokens) is to signal a move by the speaker back to the
main topic, or, in other words, to return to the main topic following what is marked
retrospectively as a digression. Let us examine these functions in more detail.

Table 1.╇ Functions of bekitsur throughout the database

Summarizing Signaling resump- Requesting Foregrounding Total

tion of main topic resumption of new narrative
main topic

5 (10.2%) 38 (77.6%) 1 (2%) 5 (10.2%) 49 (100%)

8. This stress displacement does not take place in all instances of the discourse marker bekit-
sur. The token in excerpt 1 is such an instance where it does not. More on this below.
9. This last token involves fusion with the Yiddish word for ‘beggar’, kibitser.
 Chapter 3.╇ The textual realm 

3. Summarizing bekitsur

I begin with the least metalingual use of bekitsur, that of summarizing some point
in the discourse. We will see two variants here: summarizing a list (3.1) and sum-
marizing an episode (3.2). A related function, drawing conclusions from an epi-
sode, will be discussed in Section 3.3.

3.1 Summary following a list

In the following excerpt, Nili, a woman in her forties, tells another couple and her
own husband about her first experience using Chat on the internet:
Excerpt 2 (‘Chat’):
40 Nili: .... 'az,
so,
41 .. 'az hayù kàma hitkatvuyó--t,
so there were [a] few correspondences
so there was some writing back and forth,
42 .. s’tomeret,
I mean,
43 .. 'eh
uh
44 .. nehenèti mehamisxá--k,
I enjoyed the game,
45 ... shel 'e--h la'anó--t,
of u--h answering,
46 .. velehakshí--v,
and listening,
47 .. vetikra pratí--m,
and you read details
and reading details,
48 .. ve,
and,
49 .... bekitsúr,
in short,
 Metalanguage in Interaction

50 .. letargél 'et ha'inyàn.

drilling the issue
getting some practice with the whole thing.
51 Gil'ad: ken,
yeah,
52 .. ze ktsàt nexmád kaze bahatxala.
it a bit nice like at first
it’s sort of like nice at first.
Shloush mentions an ‘appositional’ or ‘equative’ use of bekitsur, whose occurrence
“is motivated by an urge to conclude, or more specifically [...] to summarize the
preceding list under a ‘super category’, once the information is taken to be self-
evident” (1998:€ 63). Shloush’s characterization fits here. Nili lists the things she
enjoyed about that first experience of using chat: neheneti mehamisxa--k,... shel
'e--h la'ano--t,.. velehakshi--v,.. vetikra10 prati--m, (‘I enjoyed the game, of uh an-
swering, and listening, and reading details’, lines 44–47). The final two members of
this three-item list begin with the coordinative conjunction ve- (‘and’, lines 46, 47),
and ve- introduces also a fourth member of the list which is never completed, end-
ing in continuing intonation (line 48). After a pause and a token of bekitsur (line
49), Nili summarizes the list of things she enjoyed: letargel 'et ha'inyan (‘getting
some practice with the whole thing’, line 50). The actions of answering, listening,
and reading on Chat are summarized by the ‘super-category’ of ‘getting some prac-
tice with the whole thing’. Gil'ad’s response at lines 51–52, overlapping the sum-
mary, ken,.. ze ktsat nexmad kaze bahatxala. (‘yeah, it’s sort of like nice at first’),
shows that the information Nili was conveying was indeed ‘taken to be self-evi-
dent’ by at least some of the audience immediately before it was summarized. Pro-
sodically, bekitsur occurs following a series of continuing intonation contours
(lines 45–48) which it summarizes under a ‘super-category’.
Here, the literal English translation of bekitsur, ‘in short’, is adequate. This is
paralleled by the fact that this is one of only two tokens of bekitsur following con-
tinuing intonation throughout the database (see footnote 6). In this instance, it is
the continuing intonation of the unfinished list ending at line 48, which contrib-
utes to the impression that the list actually contains many more items that just did
not get mentioned. As we have seen (Chapter 1, Section 7), iconically, higher-level

10. Notice, however, the lack of equivalence between the syntactic elements conjoined in this
list. Whereas the first two consist of infinitives – la'anot, lehakshiv (‘to answer’, ‘to listen’) – the
third consists of a finite verb + complement: tikra pratim (‘read details’). Here we find the sec-
ond person masculine singular future form of the verb, tikra (‘you will read’), used as an im-
perative in the impersonal sense. Summarizing the list, the speaker returns to the infinitive –
letargel (‘to drill/get some exercise’).
 Chapter 3.╇ The textual realm 

boundaries are marked by a ‘heavier’ strategy, namely, final intonation, whereas

lower-level boundaries are marked by a ‘lighter’ strategy – continuing intonation.
The boundary between listing items of some category and summarizing them by a
‘super-category’ is certainly of a lower level than the boundary between a digres-
sion and a return to the main topic, as in excerpt 1. Therefore, it is not surprising
to find continuing, rather than final intonation at line 48.
In this instance, then, the structural requirement for discourse-markerhood is
not fulfilled, and the semantic requirement is only partially fulfilled: the literal
translation ‘in short’ is adequate here, but this bekitsur does not function as adverb
to any verb in the surrounding intonation units. The bekitsur of line 49 is inter-
preted as standing for (in a way that is no longer apparent to the speaker) some
longer metalingual utterance, such as 'im lomar bekitsur (lit., ‘if to say in short’, ‘to
say succinctly’ or ‘put succinctly’), and is therefore considered to function meta-
lingually here.
One of Hopper’s principles of grammaticization is persistence, which he for-
mulates in the following manner:
When a form undergoes grammaticalization from a lexical to a grammatical func-
tion, so long as it is grammatically viable some traces of its original lexical mean-
ings tend to adhere to it, and details of its lexical history may be reflected in con-
straints on its grammatical distribution (1991:€22).

This principle is a refinement of a phenomenon first observed by Bybee and Pagliu-

ca and termed ‘retention’, according to which older meanings of a form “glimmer
through” as they are retained from their original lexical sources (1987:€109). “The
phenomenon of retention is usually associated with single tokens of a form observ-
able at different stages of development and in different linguistic contexts” (Ziegeler
2000:€31). We see the retention, or persistence, of lexical features at this stage of the
grammaticization of bekitsur from adverb to discourse marker.

3.2 Summary of episode

In the following excerpt, which takes place during a breakfast conversation about
a couple who recently broke up, Amir explains to his girlfriend Tami why the man,
Shmulik, who had been married before, left his girlfriend, Dalit:
 Metalanguage in Interaction

Excerpt 3 (‘Breaking Up’):

39 Tami: .. haya lahem kol kax tóv beyàxad harey.
it was to them so good together PART11.
they were so happy together after all.
40 Amir: ... gam ló haya tov,
also to him it was good
he too was happy,
41 .. 'aval 'e--h,
but u--h,
42 .. pashut hi--,
simply she
it’s just that she,
43 .... Dalít kvar ló tse'ira.
Dalit already not young
Dalit isn’t young anymore.
44 .... ratsta--,
[she] wanted,
45 .. hi bikshá mimeno,
she asked him,
46 ke'ílu--,
like,
47 .. sheya'avrù-- lagur beyáxad,
that they move in to live together
to move in together,
48 ve--
and
49 .... bekitsúr,
in short,
50 lehitmaséd shuv.
to become established again.
51 (2.5) hahu nivhál.
that one became scared
that one [i.e., Shmulik] got cold feet.

11. In a study of this particle, Ariel notes: “A speaker who uses Hebrew harey indicates to her
addressee that the proposition under its scope imparts information which is already available to
him.” (Ariel 1998: 224).
 Chapter 3.╇ The textual realm 

52 (4.5) 'amár la,

told her,
53 shetsàrix la'asot 'eyze hafsaká.
that it’s necessary to do some break
that they had to take a break.
54 (3.00) bekítsur,
in short,
55 .. hem ló medabrim kvar shalósh12 shavu'ot.
they not talking already three weeks.
they haven’t been talking for three weeks now.
56 Tami: .. lò medabrím kvar shloshà shvu'ot?
not talking already three weeks
haven’t been talking for three weeks now?
57 Amir: lo medabrím.
not talking
haven’t been talking.
58 .... lo nifgashím,
not meeting
haven’t been meeting,
59 .. lo klúm.
not anything
nothing.
In this elicitational discourse segment (Chafe 1994, see footnote 16) discussing the
couple Dalit and Shmulik, Amir begins, in ‘listing intonation’, to recount the things
Dalit asked Shmulik prior to his leaving her: hi bikshá mimeno, ke'ílu--,.. sheya'avrù--
lagur beyáxad, ve-- (‘she asked him, like, to move in together, and’, lines 45–48).
This list is never completed and ends with a fragmented intonation unit (line 48).
Instead, after a longer than average pause at line 49, Amir summarizes this never-
to-be list with bekitsúr, lehitmaséd shuv. (‘in short, to become established again.’).
This case is similar to the one we saw in excerpt 2, except that the list here is much
shorter. After a 2.5 second pause, at line 51, Amir moves on to describe Shmulik’s

12. Agreement between numerals and the nouns they modify is one of the first morphological
patterns to become neutralized in Israeli Hebrew (see Ravid 1995). Notice, however, Tami’s correc-
tion of this grammatical ‘mistake’ in the following intonation unit: Amir’s shalosh shavu'ot (‘three
(fem.) weeks (masc. pl.)’) gets corrected to shlosha shvu'ot (‘three (masc.) weeks (masc. pl.)’).
 Metalanguage in Interaction

reaction to Dalit’s request. This is done via two prosodic sentences13 separated by a
long pause: hahu nivhál. (‘he got cold feet.’) (4.5) 'amár la, shetsàrix la'asot 'eyze
hafsaká. (‘told her, that they had to take a break.’). Following another long pause at
line 54, he summarizes the result of these actions: (3.00) bekítsur,.. hem ló medab-
rim kvar shalósh shavu'ot. (‘in short, they haven’t been talking for three weeks
now’). Here bekitsur occurs following a series of final intonation contours. Line 55
both summarizes the result of the actions described in the preceding intonation
units (lines 51–53) and returns the participants to the discussion of the state Shmu-
lik and Dalit are at now, the topic they were discussing prior to this segment.
Excerpt 3, then, manifests a transitional case between summarizing an epi-
sode and returning from a digression. Indeed, a likely conversational action prior
to returning to a main action is summarizing the previous action. Summarizing in
this way can also help construct the previous action as ‘tangential to the main
point’, as noted by Schiffrin (1987) for anyway. Thus, even if throughout lines 42–
53 the participants did not perceive the discourse taking place as a digression from
some main topic, employment of bekitsur following this stretch of talk helps to
construct it as such in retrospect.
Note the position of the stress in these two instances of bekitsur. We see that
stress placement here is not a matter of personal style, as both tokens are uttered
by the same speaker. At line 49, summarizing a list, Amir verbalizes bekitsúr with
the stress in standard position. However, in the transitional case between summa-
rizing an episode and returning to a main action, he verbalizes bekítsur, displacing
the stress in the more colloquial fashion. In this instance, then, we find a prosodic
feature distinguishing the two functions of bekitsur, such that the function further
away from the literal meaning of the word (summarizing an episode and returning
to the main action) is the one accompanied by the prosodic change. This is remi-
niscent of a different prosodic shift described for anyway: Ferrara found that this
resumptive discourse marker is accompanied by an intonation contour that differs
markedly from the intonation contours of the adverbial uses of anyway as an ad-
ditive (‘besides’) and a dismissive (‘nonetheless’), which are similar to one another
(1997:€354–359).

13. A prosodic sentence is defined by Chafe (1994:€137–145) as an intonation unit, or a group

of successive intonation units separated by continuing intonation, ending in sentence-final into-
nation. Chafe observes that once a speaker judges that the scanning of a center of interest (such
as a sub-episode in a story-episode) has been completed, s/he expresses that judgment with a
sentence-final intonation contour.
 Chapter 3.╇ The textual realm 

3.3 Drawing conclusions from episode

There is one case throughout the database in which bekitsur is employed not pre-
ceding a summary of the episode, but rather preceding a conclusion that is drawn
from it. This comes from a story that Ada, a woman who accompanied the Israeli
delegation to Oslo during the peace talks with the Palestinians in the mid-1990’s,
tells her friend Micha. She tells about an incident in which she and a member of
the Palestinian delegation, Abdullah, found themselves lost on a tram in Oslo be-
cause the group they were with had gotten off the tram while they weren’t paying
attention. After realizing their group was no longer there, they got off the tram and
started walking back:
Excerpt 4 (‘Ada and Abdullah in Oslo’):
60 Ada: .... hitxalnu lexapés ta’kvutsà,
we started looking for the group,
61 ... /'aval/ lo matsánu otam.
/but/ not found them
/but/ we didn’t find them.
62 .... velo yadànu gam le'éyfo hem holxim,
and not know also to where they are going
and we also didn’t know where they were going,
63 .. ki lo hikshávnu,
because we didn’t listen
because we didn’t pay attention,
64 Micha: (2.45) /nu/ 'éyfo ze haya,
/go on/ where it was
/go on/ where was it,
65 be'óslo?
in Oslo?
66 Ada: ken.
yeah.
67 Micha: (2.47) nu?
go on?
68 Ada: ..... bekítsur,
69 .. hitxàlnu laxzór laxzór laxzór laxzór,
we started to go back to go back to go back to go back,
we started going back back back back,
 Metalanguage in Interaction

70 ve'az 'amarnu tóv 'èn 'e--h.. derex,

and then we said okay there isn’t u--h a way [to find them],
71 'az nitkashér la--madrixìm shelanu.
so we’ll call to the guides ours
so we’ll call our guides.
72 ...... 'az 'e--h,
so u--h,
73 Micha: hamadrixìm yad'u--.. norvégi
the guides knew.. Norweg
74 'e--h... 'anglít?
u--h.. English?
75 /dibártem 'itam/?
/you spoke to them/?
76 Ada: ken.
yes.
77 .... 'az 'amarti,
so I said,
78 yá--,
{interjection}
79 yésh li t’a--
there is to me the
I’ve got the
80 yesh li t’a télefon shel ha--,
there is to me the phone of the
I’ve got the phone [number] of the,
81 .. madrixim batík,
guides in [my] bag,
82 ... ve'ani matxila lexapès batík,
and I start looking in the bag,
83 ... ló motsèt.
don’t find
can’t find anything.
84 Micha: {laughter}
85 Ada: matxila lexapes ba --xagoràt béten shehayta li,
[I] start looking in the--beltstomach that there was to me
I start looking in the money belt I had,
 Chapter 3.╇ The textual realm 

86 ... ló motsèt.
don’t find
can’t find anything.
87 .... ve'az hu 'omér,
and then he says,
88 'á--h,
89 yesh lax t’atélefon shel hamadrìx haze vehaze?
there is to you the phone of the guide this and this
do you have the phone [number] of some particular guide?
90 ... kén ken bétax! {Ada’s voice to Abdullah}
yes yes sure!
91 .. 'ani matxila lexapés beze,
{--laughing--}
I start looking in this
I start looking here,
92 .. 'an’ló motsèt.
can’t find [anything].
93 .. mexapéset,
[I] look [there],
94 .. ló motsèt.
can’t find [anything].
95 ... bekítsur hayinu mevohalím laxalutín.
bekitsur we were alarmed completely
bekitsur we were completely alarmed.
96 .. 'amarti tóv,
I said okay,
97 nitkasher venish'al beshagrirùt yisra'él.
we’ll call and ask at embassy Israel
we’ll call and ask at the Israeli embassy.
98 .... hitkasharti lamerkaziyà shel.. 'e--h.. 'óslo,
I called the operator of.. u--h.. Oslo,
99 ... bikáshti /ta’ telefon shel/ hashagrirùt,
I asked for /the telephone [number] of/ the embassy,
100 ... hitkashárti lashagrirùt,
I called the embassy,
 Metalanguage in Interaction

In lines 60–72 Ada describes in past tense the actions she and Abdullah took in order
to try to locate their group (there is a short clarification sequence concerning where all
this happened in the midst of this description (lines 64–67), with bekitsur employed to
return from it at line 68). After dealing with another clarification request from Micha
concerning the language they spoke with the guides (lines 73–76), she moves on to the
present tense, to the episode of looking to no avail for the guide’s phone number in
various places: in her bag (lines 78–83), in her money belt (lines 85–86), and in other
places (lines 87–94). In conclusion, she returns to the past tense and describes the state
they were in as a result of these various unsuccessful searches: hayinu mevohalím lax-
alutín. (‘we were completely alarmed’, line 95). This conclusion is preceded by bekitsur.
However, with this word, Ada also returns to the past tense description of the actions
they took in order to try to locate their group which she began earlier (lines 60–72).
At line 96 she begins to describe the next action; namely, calling the Israeli embassy.
Thus, the ‘looking to no avail’ episode (lines 77–94) is retroactively constructed as a
digression from the main storyline of describing the actions taken in order to locate
the group. In this sense, the bekitsur of line 95, too, is a transitional case – here, be-
tween drawing a conclusion from an episode and returning from a digression. Indeed,
a likely way to end an episode is to describe its conclusions or results14. As in excerpt
2, this token of bekitsur likewise follows a series of prosodic sentences ending in final
intonation contours: lines 77–83, 85–86, 87–89, 90, 91–92, 93–9415.
Throughout the database, there is only one such case of drawing conclusions
from an episode preceded by bekitsur. Because drawing conclusions and summariz-
ing are closely related activities, I consider them both to be in the same category of
‘summarizing’. Table 1 shows that only 5 instances (10.2%) of this less metalingual
use of bekitsur are found in the corpus, two of which are already borderline between
this function and that of returning to the main topic following a digression, as we
have seen in excerpts 2 and 3. Let us move now to this latter function, the main func-
tion of bekitsur, which is never found following continuing intonation.

14. This point is underscored by one of the meanings of the English verb ‘to conclude’; namely,
‘to end’.
15. Note that this is different from Shloush’s ‘inferential use’ of bekitsur, for which she con-
structs the following example: david mevsSel, menake, Sotef ricpa, megahec ... bekicur, keday lax
leha’asik oto (‘David, cooks, cleans, does the floors, irons ... in short, you should hire him’).
(1998: 63). Shloush characterizes this case in the following manner: “once sufficient informa-
tion has been provided, what follows bekicur is an inference rather than a mere equation [as in
the case of summarizing a list]” (ibid.: 64) and “the cutting of the list in question by bekicur
signals a conclusion related to the discourse topic” (ibid.: 67). No such lists were found in the
corpus. The only time bekitsur was employed to draw conclusions from previous discourse oc-
curred following an episode which was made up of a series of prosodic sentences (i.e., units
ending in final intonation contours), unlike Shloush’s constructed example. This is one of the
problems of constructing examples – one cannot look at prosody.
 Chapter 3.╇ The textual realm 

4. Resumptive bekitsur: Returning to the main topic

I use the term ‘topic’ or ‘discourse topic’ here in Chafe’s (1994) sense. In Chafe’s
model of information flow in discourse, as people speak, ideas of events, states,
and the referents that participate in them are constantly changing activation states
in their minds. Information can be in any one of three activation states: active,
semiactive, or inactive. As new concepts become activated, previously used con-
cepts gradually become inactive. “Intonation units are hypothesized to be the lin-
guistic expression of information that is, at first, active in the consciousness of the
speaker, and then, by the utterance of the intonation unit, in the consciousness of
the listener, or at least that is the speaker’s intent” (1994:€69). In chapters 10 and 11
of his book, Chafe considers units larger than the intonation unit:
We need now a name for the larger amount of information that can be semiactive
[in the speaker’s mind]. I will use the term topic in this way, qualifying it when nec-
essary as discourse topic to distinguish it from the other phenomena to which the
term has been applied. The usage here is thus in accord with such expressions as the
topic of a paragraph, changing the topic, and the like (cf. Keenan and Schieffelin 1976,
Brown and Yule 1983, pp. 71–106). [...] We can think of each such topic as an ag-
gregate of coherently related events, states, and referents that are held together in
some form in the speaker’s semiactive consciousness (ibid.: 120–121).

In Chafe’s approach, then, topics are on the one hand aggregates of semiactive in-
formation that segment discourse into chunks larger than intonation units. On the
other hand, a topic is interpreted in the lay sense of what speakers understand they
are talking about at a given time. The question of when one topic ends and an-
other begins, or of when one topic is briefly interrupted for a digression (in itself a
new topic) and then returned to, is open to negotiation by the participants of an
interaction, as is the question of whether or not a topic is of secondary importance
and therefore a digression.
The notion of topic, then, is a flexible one, and whether or not a digression has
occurred is a relative issue. In what follows, we will continue to see that one of the
main functions of bekitsur is to retroactively construct the topic that has just ended
as a digression relative to the continuation of the discourse.

Table 2.╇ Distribution of bekitsur according to discourse type

Narrational discourse Elicitational discourse

45 (91.8%) 4 (8.2%)
 Metalanguage in Interaction

Table 2 shows the distribution of bekitsur tokens in the present database according
to the type of discourse16 in which they appear.
We see that the great majority of bekitsur tokens (45 tokens or 91.8%) occur in
narrational discourse, although this word is not restricted to this type of talk. In-
terestingly, 3 out of the 4 tokens of bekitsur employed in elicitational discourse are
of the summarizing variety (Section 2). We will see bekitsur employed in particu-
lar in the moves from elicitational back to narrational discourse.

4.1 Collaboration in returning to the main topic

Ferrara found that the majority of anyway discourse marker tokens in her data,
which consisted of narrative discourse17, were teller-triggered. In other words,
74% of anyway tokens followed a digression initiated by the teller of the story,
while only 26% followed a digression initiated by the listener. She concludes that
the discourse marker anyway is mainly a self-digression management marker
(1997:€ 359). By contrast, in the present corpus, bekitsur is more often listener-
triggered than it is teller-triggered:

Table 3.╇ Distribution of resumptive bekitsur according to trigger of digression

Listener- Teller- Listener+ Situational Indetermin- Total

triggered triggered teller digression able19
triggered18

22 (57.9%) 11 (28.9%) 1 (2.6%) 1 (2.6%) 3 (7.9%) 38 (100%)

16. Chafe distinguishes two types of discourse topics: narrational and elicitational.€ While
movement through a narrational topic develops according to the narrational scheme (cf. Labov’s
narrative structure (1972)), movement through an elicitational topic is driven by the interaction
between participants. One participant functions as an eliciter and the other as a responder. The
eliciter introduces the topic, but it is the responder “who possesses the bulk of the interesting
information. The topic is, in other words, one about which the eliciter finds it interesting to gain
more knowledge, and he or she accomplishes that goal partly by asking questions, partly in
other ways” (Chafe 1994: 123).
17. There is a difference in employment of the term ‘narrative discourse’ here in comparison to
Ferrara (1997). Whereas Ferrara takes the elicitational segments interrupting the stories to be
part of her corpus of narratives, I count bekitsur tokens appearing in these elicitational segments
as occurring in elicitational discourse.
18. This is a case in which bekitsur is employed after a long stretch of laughter by both teller and
listener, following an orientation of a narrative. Thus, the digression was triggered both by the
teller and the listener.
19. No recording of the earlier segment is available.
 Chapter 3.╇ The textual realm 

In other words, bekitsur occurs following digressions initiated by the listener

roughly twice as often as it occurs following digressions initiated by the teller.
However, listener-triggered digressions in this corpus often develop into long elic-
itational segments, and the question of who triggered the digression is no longer
relevant because speakers are probably no longer aware of who initiated the di-
gression when they attempt to resume the main topic. What is more important is
their collaboration in getting back to the main topic.
A case in point is the following excerpt from a conversation between two
women in their early forties. Nurit is recounting to Sarah the long story of the
dealings her husband Ofer had with various medical doctors:
Excerpt 5 (‘Doctors’):
39 Nurit: .. 'axsháv doc
now Doc
40 .. haláxnu lif
we went bef
41 .. lifnéy shehalaxnu la'ortopèd,
before that we went to the orthopedic surgeon
before we went to the orthopedic surgeon,
42 hayìnu 'etsel doctor 'ábu,
we were at Doctor Abu[’s],
43 .. kvar ló zoxeret bishvil mà.
no longer remember for what
[I] can’t remember anymore what for.
44 .. nidmé li,
seems to me
I think,
45 bishvil hapétek shel haxofesh maxalà.
for the note of the vacation sickness
for the sick leave note.
46 .. ló meshane.
doesn’t matter
never mind.
47 .. ve.. doctor 'ábu 'amàr,
and.. Doctor Abu said,
48 Sarah: xòfesh maxala lemí?
vacation sickness for whom
sick leave for whom?
 Metalanguage in Interaction

49 Nurit: .. le--'ófer,
for Ofer,
50 lir'ót,
to see,
51 'uláy 'e--h.. bitùax te'unot 'ishiyo--t,
maybe u--h.. insurance accidents personal
maybe u--h.. personal accident insurance [will pay for his sick leave],
52 Sarah: 'ah 'ah 'ahá,
oh oh aha,
53 Nurit: 'ulày bituax le'umí,
maybe insurance national
maybe social security,
54 Sara: hevánti.
I understood
got it.
55 Nurit: míshehu.
somebody.
56 tishme'i,
listen,
57 xodsháyim hu kvar lò 'ovéd!
two months he already not working
it’s been two months now that he hasn’t been working!
58 Sarah: naxón,
right,
59 naxon.
right.
60 .. yésh bitùax leze,
there is [some] insurance for this,
61 .. 'ad káma she'ani yodà'at.
as much that I know
as far as I know.
62 Nurit: ló yoda'at.
don’t know.
63 Sarah: yésh mashehu.
there is something.
 Chapter 3.╇ The textual realm 

64 Nurit: ló yoda'at.
don’t know.
65 Sarah: nú?
go on?
66 kén?
yeah?
67 Nurit: .. bekitsú--r,
68 .. doctor 'àbu 'amár,
Doctor Abu said,
The listener, Sarah, is the one who interrupted the flow of this story, asking for
clarificatory information (sick leave for whom?, line 48), but she is also the one who
initiates the return to the main topic with nu? (‘go on?’) ken? (‘yeah?’) in lines
65–66. Only after this urging does Nurit return to the main storyline, opening
with bekitsur (line 67). This digression was listener-triggered, but the storyteller
collaborated with the listener in following the digression for a relatively long
stretch of talk (lines 49–64), so that at line 65 it was the listener, not the storyteller,
who initiated the move back to the story. We see that this type of bekitsur, like the
episode-summarizing variety, follows a series of prosodic sentences ending in fi-
nal intonation contours (lines 49–55, 56–57, 58–59, 60–61, 62, 63, 65, 66). Most
importantly, the intonation unit immediately preceding it (ken?, ‘yeah?’) ends in
sentence-final intonation. This is the case with all but one bekitsur token in the
resumptive category as well as with all narrative-introducing bekitsur tokens.
Collaboration with the sequence
listener: nu? (‘go on’)
storyteller: bekitsur,
can also be found in much shorter digressions, as in excerpt 4, lines 64–68. In fact,
about a fifth of resumptive bekitsur tokens and bekitsur tokens introducing new
narratives (9 out of 43 cases) in this corpus follow a token of nu (‘go on’) uttered by
the listener. The mere fact that the second most prevalent interpersonal discourse
marker in Hebrew, nu (‘yeah, go on’), is one whose main function is to urge further
development within a main topic20 is proof of the highly collaborative nature of
returning from digressions in this culture.
In a study of English anyway and some of its corresponding Japanese dis-
course markers, Takahara writes: “a close of digression or return to the main topic
is almost always marked by this characteristic linguistic unit anyway” (1998:€335).

20. See Chapter 2.

 Metalanguage in Interaction

Bekitsur is indeed the discourse marker most frequently employed to return from
digressions in my Hebrew data. There seems to be an important difference, how-
ever, between bekitsur and English anyway with respect to collaboration. The Is-
raeli language games of digressing and returning from a digression are more col-
laborative. First, as we have seen, in contrast to Ferrara’s study of anyway (1997),
digressions preceding bekitsur are more often initiated by the listener than by the
teller. Second, the listener is also more involved in returning from them. This is in
line, of course, with other features of this relatively high-involvement (Tannen
1984, Schiffrin 1984) speaking style (Katriel 1986, Blum-Kulka and Katriel 1991,
Maschler 1994a, Blum-Kulka 1997).
To explain the relatively high frequency of digressions in discourse, Ferrara
quotes Basgoz, who suggests a possible rationale for this frequency – that of bridg-
ing the cultural gap between speakers and hearers (Ferrara 1997:€366): “A gap –
small or big, historical, linguistic, social or ecological in nature – develops between
the past and present cultures. [....] Digression bridges this gap, making the un-
known known, irrational rational, obscure clear, incredible credible...” (Basgoz
1986:€3). In the conversations collected here, the cultural gap between participants
is minimal because they are usually close friends or family members. Yet their
conversations abound with digressions. The reason for this is interactional: digres-
sions tie discourse participants to each other, enabling them to make their inter-
locutors’ discourse more relevant to their own experience. When digressing-lis-
teners then also initiate a move back to the main topic, they are able to reassure the
speaker of their continuing involvement in their talk.
Even the listener herself can employ bekitsur in order to return the storyteller
from a digression. This happens in only one instance throughout the corpus, in a
conversation between two women in their early twenties about a couple who had
broken up. Beforehand, the couple had bought sofas the woman did not like:
Excerpt 6 (‘Sofas’):
95 Gila: ..... hayu shlosha tsva'ím shel 'e--h,
there were three colors of u--h,
96 ... praxìm 'al hasapá.
flowers on the sofa.
97 .. 'az hi hexlìta shehi titsbá.
so she decided that she will color [in].
98 ... 'exad mehatsva'ím shehi lo 'ahavà.
one of the colors that she didn’t like.
99 {laughter}
 Chapter 3.╇ The textual realm 

100 .... tikne tséva,

[she]’ll buy [some] color,
101 vetitsbá 'otam.
and color them [in].
102 ... 'ani nisiti lehoríd 'ota mehara'ayòn,
I tried to get off her from the idea
I tried to get her off the idea,
103 .. shetetsapé 'et zè!
that she put upholstery [on] it!
get her to upholster them!
104 .... 'o máshehu.
or something.
105 .. 'an’ló yoda'at,
I dunno,
106 .. 'aval she
but that
107 Yonat: lamròt sheze 'ód kesef,
{------laughing-----}
although that that more money
although that’s more money,
108 .. ken?
yeah?
109 Gila: .... ze be'ayà shel mi shekone sapót,
this [is the] problem of [one] who buys sofas,
110 velo bodék,
and doesn’t check,
111 .. 'im hu 'ohév 'otam,
if he likes them,
112 'o ló 'ohèv 'otam,
or doesn’t like them,
113 .. 'o 'im hem mat'imòt lo labáyit.
or if they match for him to the house
or if they match his house.
114 Yonat: .. tòv 'akítser,
okay 'akitser,
 Metalanguage in Interaction

115 sof hasipùr hu shehem nifredú?

[the] end of the story is that they broke up?
116 Gila: sóf hasipùr
end of the story
shehem nifredu,
that they broke up
in the end they broke up,
117 vehadavar harishón sheli'at 'amrà,
and the thing the first that Li'at said
and the first thing Li'at said was,
118 'òy yófi,
oh great,
119 'ani 'etèn lo 'et hasapót,
{--------laughing------}
I’ll give him the sofas,
120 ve'ani 'eknè li xadashót.
{-------laughing--------}
and I’ll buy myself new [ones].
121 ..... zéhu.
that’s it.
At lines 102–104, Gila says she tried to dissuade her friend from coloring in the
flowers on the upholstery and to get her to re-upholster the sofas instead. At line
107 Yonat objects to this idea with the concessive discourse marker lamrot she-
(‘although’) (cf. Günthner 2000), saying that that would have cost even more mon-
ey. Gila answers her at lines 109–113, beginning with the formula ze be'aya shel mi
she... (‘this is [the] problem of one who...’), justifying her solution with the idea
that if one buys unwisely, one should ‘pay for it’ (i.e., it’s the unwise woman’s prob-
lem, not hers). Yonat does not pursue this line, and at line 114 returns Gila to her
story, requesting its ending: tov 'akitser, sof hasipur hu shehem nifredu? (‘okay 'ak-
itser, the end of the story is that they broke up?’), which Gila then provides (lines
116–121). We see that Yonat brings Gila back from the digression (which Yonat
herself started at line 107) with the colloquial 'akitser in the discourse marker clus-
ter tov21 'akitser (‘okay 'akitser’). Using this more jocular form of bekitsur may also
be functioning to compensate for the disagreement between them. Thus, while
initiating return from a digression allows participants to manifest their involvement

21. On the dual nature of tov (‘okay’) here, as both an interpersonal and a textual discourse
marker, see Chapter 5, Section 5.
 Chapter 3.╇ The textual realm 

in each others’ talk, when the digression is a disagreement, it can also help avoid
unwelcome topics.
Let us move now to a closer look at the various types of digression from which
bekitsur returns. Dascal and Katriel classify digressions into three types: utterance-
based, interaction-based, and insertion sequences:
Utterance-based digressions are characterized by the fact that some kind of ‘con-
tent’ relation holds between the mainstream utterance and the digressional ones.
Interaction-based digressions exhibit no such relations. Insertion sequences, in
our use of the term, refer to a large variety of corrective and clarificatory speech-
acts. They have an intermediary standing between utterance- and interaction-
based digressions (1982:€82).

While defining a ‘content’ relation is not a simple matter, and Dascal and Katriel
acknowledge this, I use their classification as a preliminary way of describing the
employment of resumptive bekitsur in the corpus.

4.2 Returning from interaction-based digressions

Only one instance of bekitsur throughout the data functions to return participants
from an interaction-based digression to the main topic. However, this paucity of to-
kens is due mainly to the nature of the data; namely, that participants in this corpus
are not generally having a conversation while also doing something else. I have often
heard this discourse marker employed to return from interaction-based digressions,
such as following a phone call or an interruption of a conversation by a child.
In the following excerpt two women, Sharon and Dalia, close friends in their
early twenties, are sharing a night-shift in one of the offices of Israel’s Electric
Company. As Sharon later explained to me, they are sitting in front of a map in
which locations of callers who call in to report problems are flagged. Their job is to
dispatch engineers in order to deal with the problems. In line 1, Dalia asks Sharon
whether she has dealt with two calls that had come in:
Excerpt 7 (‘Prince’):
1 Dalia: ... 'asìt 'et shneyhém?
{-----------------pp-----------------}
you did both of them
did you do both of them?
2 Sharon: ... 'asiti 'exád.
{ ---pp---}
I did one.
 Metalanguage in Interaction

3 Dalia: .. betuxá?
{ ---pp---}
[are you] sure?
4 Sharon: ken.
{ -pp-}
yeah.
5 Dalia: ... hakítse--r,
{------f------}
6 .. hi 'oméret li,
{------f------------}
she says to me,
7 .... yashávnu dibárnu veze--,
{-------------f---------------}
we sat [and] talked and so on,
Following this short interchange concerning the calls they are supposed to deal
with (lines 1–4), Dalia returns to a story she heard from another friend (line 6).
The return to this story is preceded by hakitser (5) and accompanied by a marked
prosodic change in volume, as well as by employment of the pronoun hi (‘she’, line 6),
which would not have been employed at the beginning of a narrative. No ‘content’
relation holds between this digression and the narrative, and we have here an in-
teraction-based (or ‘situational’, Lenk 1998:€60) digression.

4.3 Returning from content-based digressions

By far, the great majority of bekitsur tokens in this corpus function to return par-
ticipants to the main topic following a digression which does bear some ‘content’
relation to the main topic. I present here a finer classification of these instances,
based on the nature of the digression.

4.3.1 Following listener-initiated digressions

Oftentimes during narration, the storyteller pauses for a question or an elicita-
tional segment related to the story which is initiated by some member of the audi-
ence. After dealing with this new topic, the storyteller moves back to the narrative
with bekitsur or one of its variants.
In the following excerpt, David describes to his close friend, Dalia, a chase
after a tractor that he was supposed to repossess as part of his work that day:
 Chapter 3.╇ The textual realm 

Excerpt 8 (‘Tractor Chase’):

11 David: haxokér hapratí 'amar,
the investigator private said
the private investigator said,
12 shehu nimtsa be'eyze makóm,
that he is located in some place
that he [the defaulting purchaser] can be found in some place,
13 ... géshem zal'afót baxuts,
rain raging outside
it’s raining cats and dogs outside,
14 ke'ílu,
I mean,
15 ... má ze gèshem,
what is rain
what [do I mean by] rain,
16 ... shexavál 'al hazmàn.
that it’s a waste of time.
unimaginable rain.
17 .. ràk metumtamím,
only idiots,
18 yots'ím mehabàyit.
go out of the house.
19 Dalia: kamóxa.
like you.
20 David: naxón.
right.
21 .. nu 'az mí yefarnès,
nu so who will earn the living,
22 .. 'át?
you?
23 Dalia: nu--?
go on.
{smiling}
24 David: .... be'uktsú--r,
{in American accent}
 Metalanguage in Interaction

25 .... 'anaxnu mezahìm 'et ha.. klí,

we identify the.. vehicle,
To illustrate how strong the rain was during this tractor chase, David adds: rak me-
tumtamim, yots'im mehabayit. (‘only idiots, go out of the house’., lines 17–18). Dalia
quickly joins in, mocking her close friend: kamoxa (‘like you’, line 19), to which
David responds with humorous sarcasm: nu 'az mí yefarnès,.. 'át? (‘nu22 so who will
earn the living, you?’, lines 21–22). At line 23, smiling, Dalia urges him to go on
with the story with the discourse marker nu--?, which David does at line 25 follow-
ing the most playful variation on bekitsur in this corpus – be'uktsur. This form is
modeled after an Arabic word pattern and is uttered in an American accent con-
tinuing the mockery and playfulness of this digression. We see, then, a return to the
main topic following an audience-initiated, topically-relevant digression. In this
case, the digression is only 5 intonation units long, but in other instances through-
out the data (such as excerpt 4), the digression often becomes a long elicitational
segment and bekitsur functions to move from elicitation back to narration.
The mocking tone of be'uktsur in the preceding example is influenced by the
general mocking tone of the digression. However, in considering the variation in
form of the discourse marker anyway (anyways, anyhow, at any rate), Ferrara men-
tions the jocular form anyhoo as one of the variants of anyhow (1997:€370). It is not
surprising to find a jocular tone when a speaker returns from a digression, par-
ticularly one initiated by the listener. In doing so, the speaker signals to the lis-
tener that what they have just said is of secondary importance, a move potentially
carrying interpersonal ramifications. Joking is a natural compensation for an in-
terpersonally threatening move of this type.

4.3.2 Following storyteller-initiated digressions

4.3.2.1 Following a story aside
Bekitsur can be employed by the storyteller to return from a digression she herself
initiated, as in the following case in which Li'at tells her friend Anat a story about
trying to switch residence halls with some woman:
Excerpt 9 (‘Reznik Residence Halls’):
34 Li'at: .. kódem kol,
first of all,

22. While nu at line 23 functions to urge the speaker to go on with the story, at line 21 it pre-
cedes the discourse marker 'az (‘so’) and functions to add a mocking tone to this statement. See
Chapter 2.
 Chapter 3.╇ The textual realm 

35 hi kol hazman norà paxadá,

she all the time terribly was afraid
she was constantly terribly scared,
36 .. biglal zé she,
because of the fact that,
37 ... yitfesú 'otanu--,
they’ll catch us,
38 .. veya'alú 'al ze,
and they’ll find out about it,
39 .. vema yikré--,
and what[’s gonna] happen,
40 ve..ma haprotsedúra.
and what[’s] the procedure.
41 hi hayta betuxa shehì teshalem lí ta’,
she was sure that she would pay me the,
42 .. 'et ma she'ani be.. berám?
what I [pay] in.. in Ram?
43 ... begiv'at rám?
in Ram Hill [residence hall]?
44 vehi
and she
45 'aní 'ashalèm la ta’reznik,
I’ll pay her the Reznik [residence hall],
46 ve--bixlál,
and in general,
47 .. hi haytà me'ód lo ba'inyanim.
she was very much not in the issues
she was very much out of it.
48 ... vehitbarèr bixlal shehi lomedet 'itánu!
and it turned out at all that she studies with us
and it turned out she even studies with us!
49 Anat: /????/
50 /????/
51 Li'at: ken,
yeah,
 Metalanguage in Interaction

52 shaná rishonà.
year first
first year.
53 .... hakítser,
54 'amárti la,
I said to her,
At line 48 the storyteller digresses to give an external fact about the woman she
was trying to switch residence halls with; namely, that it turned out she was a first
year student in their department. After a very short discussion of this (lines 49–52),
she returns to the narrative with hakitser (line 53), clearly marking the preceding
digression as tangential to the main point.
The relativity of digressions is underscored by the employment of bekitsur by
the storyteller in the midst of a narrative, between its various episodes (Labov 1972,
Chafe 1994). Such episodes are clearly not digressions from the narrative, but in
the progression towards the climax of the story (Chafe 1994), they can be retroac-
tively constructed as secondary, thus aiding the listener in following the main sto-
ryline. Sections 4.3.2.2 - 4.3.2.4 illustrate this retroactive construction of back-
grounding (Hopper 1979) in narrative via bekitsur.

4.3.2.2 Following orientation

In the following excerpt, the storyteller employs 'akitser to move from the orienta-
tional episode of her story to the beginning of the complication. First, she de-
scribes a computerized form that workers in her office are supposed to fill out
when interviewing women for some job:
Excerpt 10 (‘Office Gossip’):
154 Galit: .. 'axshav yésh bamaxshe--v,
now there is in the computer
155 .. 'e--h.. kmo-- rúbrika,
u--h like [a] rubric,
156 .. rúbrika kazot,
[a] rubric like,
157 ... shel 'e--h,
of u--h,
158 .... hofa'á.
appearance.
159 ... 'az yésh 'e--m,
so there is uh--m,
 Chapter 3.╇ The textual realm 

160 ... gru'á,

awful (fem. sg.),
161 ... regilá,
ordinary (fem. sg.),
162 ... 'e--h tová,
u--h good (fem. sg.),
163 .. yitsugít,
presentational (fem. sg.),
164 .. vetovà me'ód.
and good (fem. sg.) very
and very good.
165 .... beséder?
okay?
166 ... ve'ef
and wh
167 ... vebedérex klal,
and usually
168 .. mà she'ani 'amura lasím,
what (that) I[’m] supposed to put [down],
169 ... ze regilá.
is ordinary.
170 .. ke'ilu,
like,
171 .. 'ata yodé'a,
you know,
172 .. lò--.. yitsugí bi..myuxad,
not presentational especially
not particularly presentational,
173 .. velò.. garú'a bimyuxad.
and not awful especially
and not particularly awful.
174 Nir: {cough}
175 Galit: ..... 'akítser,
176 .. hayom 'ani-- korét xavat dà'at,
today I[’m] read[ing a] report,
 Metalanguage in Interaction

177 ... 'az 'eh 'ari'éla kotèvet,

so uh Ariela writes,
178 ... 'e--h.. baxurà mamash mexo'éret.
{-----laughing----}
uh young woman really ugly
u--h.. a really ugly woman.
Following the orientation explaining the form workers in this office are supposed
to fill out, at line 175 Galit moves on to describe a particular form she encountered
that day that was filled out rather tactlessly by one of her co-workers. This is the
beginning of the complication of her narrative, and it is preceded by 'akitser
(line 175). Certainly, lines 154–170 are not a digression from her story. However,
in progressing towards the climax of the story, they are of secondary importance
compared to the complicating action, and 'akitser helps construct them as such.
This is similar to Ferrara’s characterization of English anyway: in 40% of this
discourse marker’s tokens throughout her data, anyway occurs “after preceding
orientational detail given by the teller” (1997:€359). This use is much less prevalent
in the Hebrew data – only 2 such cases (out of 37 resumptive bekitsur tokens, 5.4%)
were found in the present database. However, as the following sections show, bekit-
sur occurs not only between orientation and complication in narratives, but also at
other narrational episode boundaries.

4.3.2.3 Following evaluation

A storyteller can retroactively construct an evaluative section of the narrative as
being of secondary importance. One example comes from the Tractor Chase con-
versation, in the midst of the complicating action, as David describes his chase
after the owner of the tractor that he is supposed to repossess:
Excerpt 11 (‘Tractor Chase’):
46 David: .... ve'ani kmo metumtám,
and I like an idiot,
47 ... rodéf 'axaràv.
am running after him.
48 ... ke'ilu
like
49 .. barégel,
by foot,
50 'axarèy tráktor,
after a tractor,
 Chapter 3.╇ The textual realm 

51 betòx kol habóts haze?

in all the mud this
in all this mud?
52 ... veke'ilu haragláyim sh’xa,
and like the feet your (masc. sg.)
and like your feet,
53 nitka'ot bifním kaze,
get stuck inside like
get stuck inside like,
54 kx kx kx kx. {sounds imitating the sound of feet in the mud}
55 ... bekitsúr vela'inyá--n,
in short and to the point,
56 ... 'axaré--y,
after,
57 .. mirdáf she--,
a chase that,
58 ne'eràx kemispàr dakó--t,
lasted roughly a few minutes,
59 .... hitslàxti lehagia 'ad latráktor,
I managed to get to the tractor,
60 .. lisgór lo ’ta
to close for him the
to close his
61 ke'ìlu lexabót ta’manòa,
like to turn off the engine
like to turn off his engine,
62 lakáxti ta’maftèax.
I took the key.
In lines 46–47, David describes the ridiculousness of the situation. Lines 48–51 are
a self-rephrasal of this description, preceded by the discourse marker of self-re-
phrasal ke'ilu (‘like’, line 48), explicating the phrase kmo metumtam (‘like an idiot’)
of line 46. From these lines we learn why David felt like an idiot running after the
guy whose tractor he was supposed to repossess: he was running after a moving
tractor by foot, in the mud. To further enhance his story and involve the audience,
he adds a short evaluative episode depicting the feet getting stuck in the mud (lines
 Metalanguage in Interaction

52–53) along with the sound they make in this state (line 54)23. At line 55, how-
ever, he returns to the complication describing the actions he took – reaching the
tractor, switching off its engine, and taking the key (lines 59–62). This return to the
complicating actions is preceded by the discourse marker bekitsur embellished by
the fixed formula (Pawley and Syder 1983) bekitsúr vela'inyán (‘in short and to the
point’), implying that the previous utterance was perhaps not precisely ‘to the
point’. Again, the evaluative episode, contribute as it might to audience involve-
ment in the story, is constructed as being of secondary importance as far as pro-
gressing towards the climax of the story is concerned.
The characterization of evaluative material as being of secondary importance
is consistent with Labov’s description:
A complete narrative begins with an orientation, proceeds to the complicating
action, is suspended at the focus of evaluation before the resolution, concludes with
the resolution, and returns the listener to the present time with the coda. The
evaluation of the narrative forms a secondary structure which is concentrated in
the evaluation section but may be found in various forms throughout the narra-
tive (Labov 1972:€369, emphasis mine).

This characterization of the secondary nature of evaluation is also in agreement

with other studies of foregrounding and backgrounding in discourse, e.g., Rein-
hart 1984:€783, Brinton 1996:€45.

4.3.2.4 Organizing the hierarchy of episodes in a story

In long and involved stories which often include embedded stories as episodes,
bekitsur functions to organize the hierarchy of episodes, thus aiding the listener in
following the main storyline. In the continuation of the ‘Doctors’ narrative, Nurit
tells that Doctor Abu had recommended some antibiotic, but the orthopedic sur-
geon they then went to on a Thursday decided to take a culture and give another
antibiotic instead. Nurit then proceeds to an episode constituting an embedded
story lasting 37 intonation units (only the end of which is given here) about how
the nurse sent this culture to the lab only on Sunday, and how they found that out
from Doctor Abu. In the following segment, Nurit summarizes this embedded
story and moves on to the continuation of the first story:
Excerpt 12 (‘Doctors’):
172 Nurit: .. ve... vedoctor 'ábu berèr,
and and Doctor Abu checked,

23. For a detailed study of this segment, particularly with respect to the tokens of kaze (‘like’)
and ke'ilu (‘like’) found in it, see Maschler 2001 and Chapter 4.
 Chapter 3.╇ The textual realm 

173 .. pashút,
simply,
174 .. hitkashér lama'abadà--,
called (to) the lab,
175 .. venatàn ta’mispar te'udat zehú--t,
and gave the number of certificate identification
and gave the number of the ID card,
176 Sarah: ke--n.
yeah.
177 Nurit: .. veta’
and ev
178 Sarah: ken.
yeah.
179 Nurit: .. ta’kó--l,
everything,
180 .. vehistabé--r,
and it turned out,
181 sheràk beyom rishón,
that only on day first
that only on Sunday,
182 hi masrá 'et ze.
she sent it.
183 .... kitsú--r,
184 .. hu-- 'amár,
he said,
185 shexayavìm lehaxlif ta’'antibyótika.
that they must change the antibiotic.
At line 183 Nurit employs kitsur to return from this embedded story about the
nurse, and proceeds to the following complicating action of the main story de-
scribing Doctor Abu’s instructions to change the antibiotic given by the ortho-
pedic surgeon. This token of bekitsur, then, helps organize the various embedded
stories in this 262 intonation unit long story.
Several linguists suggest that grounding is not a dichotomous matter of some
binary foreground vs. background distinction, but rather, that it is a scalar matter
(e.g., Hopper and Thompson 1980:€ 252, Fleischman 1985:€ 854, 862, 1990:€ 169,
184–185, Enkvist and Wårvik 1987:€ 224, Wårvik 1987:€ 380, 385–386, Brinton
 Metalanguage in Interaction

1996:€269). The majority of these studies focus on the formal aspects of the verb
constructing grounding in discourse. The present study examines employment of
discourse markers for this purpose. The existence of a marker retroactively con-
structing background episodes supports the view of grounding as scalar, because
it shows that participants themselves treat grounding as a flexible matter that they
can manipulate.

4.4 Returning from insertion sequences

The final category in Dascal and Katriel’s classification of digression is that of in-
sertion sequences; namely, corrective and clarificatory conversational actions
which “have an intermediary standing between utterance- and interaction-based
digressions” (1982:€82). One such case was seen in excerpt 1, line 121, following
the clarification sequence initiated by the addressee concerning the term kaplad.
Another case, less content-related, is found in the following excerpt, in a conversa-
tion between three young women about Yonit’s shopping with her friend Yo'av for
a pair of gym shoes which turned out to be quite a bargain:
Excerpt 13 (‘New Shoes’):
61 Yonit: histovávnu,
we turned around,
62 ra'ìnu 'et 'éle.
we saw these. {i.e., the pair of gym shoes she is wearing}
63 Michal: míshehu rotsè? {offering some food}
anybody wants
does anybody want some?
64 Yonit: pìt'om kám
suddenly how
65 .. káma ze 'olè? {Yonit’s voice while shopping}
.. how much [does] this cost?
66 Morit: lo todá. {concerning the food offer}
no thanks.
67 Yonit: lo-- todá.
no thanks.
68 ... tish'ìm vetésha shekel.
ninety (and) nine(fem.) shekels.
69 Michal: ... tish'á.
nine (masc.)
 Chapter 3.╇ The textual realm 

70 All: {laughs}
71 Michal: .... tish'im vetish'á shkalìm.
{---------laughing-----------------}
ninety (and) nine (masc.) shekels.
72 Yonit: .. shkalí--m,
shekels (pl.),
73 'aval 'ani 'amàrti shékel!
{--------slightly laughing----}
but I said shekel (sg.)!
74 Michal: .. nu 'az má?
nu so what?
75 : ... bekitsú--r,
76 Michal: shèkel ze gám zaxar.
shekel is also masculine.
77 Yonit: ... yo'àv mistakel 'al hana'aláyim,
Yo'av looks at the shoes,
78 ... hu 'omér li,
he says to me,
79 ló yaxol lihiyòt,
not can be
can’t be,
80 hem betax moxrìm na'al 'axát.
they probably selling shoe one
they must be selling one shoe.
In giving the price of the shoes, 99 shekels (roughly 25 dollars at the time), Yonit
employs the feminine form of the numeral, tesha (‘nine, fem.’) although the mon-
etary unit shekel is a masculine noun (line 68). This is an instance of the common
neutralizing of agreement patterns with numerals in casual spoken Hebrew (Ravid
1995). Michal and Yonit are both students at a Hebrew Language Department.
Perhaps this (and the fact they are being recorded24) is why Michal corrects Yonit
at line 69 with tish'a (‘nine’, masc.). She then rephrases Yonit’s utterance: tish'im
vetish'a shkalim (‘ninety nine shekels’), using the plural form for shekel, shkalim25.

24. Their laughter throughout lines 70–73 is perhaps evidence of their awareness of the record-
ing of this ‘grammatical mistake’.
25. In certain contexts of enumerating in Hebrew, it is possible to use the singular instead of the
plural form of the noun.
 Metalanguage in Interaction

At lines 72–73, Yonit objects, saying she used the singular shekel, not shkalim (pre-
sumably justifying the feminine tesha, which she had used back in line 68). At lines
74 and 76, we find Michal’s nu so what (the Israeli ridiculing/mocking version of
Labov’s famous ‘so what?’ question (1972:€366)26) ridiculing Yonit’s justification
(line 73) because shekel ze gam zaxar (‘shekel is also masculine’, line 76). Overlap-
ping this, Yonit returns to her story at line 75 with bekitsur. Thus we find this word
employed to return from an insertion sequence that is relatively unrelated to the
content of the discourse.

5. Foregrounding bekitsur: Introducing a new narrative

The final function of bekitsur found in this corpus underscores, more than any
other function, the retroactive construction that is accomplished by this word.
This type of bekitsur is found opening narratives, with no digression in the imme-
diately preceding discourse. As a result of employment of bekitsur, whatever hap-
pened beforehand (and this could be another foregrounded narrative) ‘moves to
the background’, and the new narrative is foregrounded even more brightly.
Such is, for instance, the beginning of theTractor Chase narrative. This narra-
tive is told by David while cooking some soup in his kitchen, while Dalia, a close
friend of his, with whom he is not romantically involved, is visiting him. In the
immediately preceding moments of the interaction, David is answering Dalia’s
question concerning how to prepare a béchamel sauce. Following his sketchy in-
structions, Dalia requests an accurate recipe, and David jokingly says, ze gám 'ole
kèsef! (‘this, too, costs money!’). This is a reference back to the immediately pre-
ceding topic, in which David tells Dalia about a woman who calls him up and
comes to his house solely for sexual purposes, and how disgusted he is with the
situation. Dalia mocks him, saying that he should charge her for this, something
he jokingly says he has considered doing. He then recounts to Dalia a conversation
he had recently had with this woman on the phone, telling her he was fed up with
the meaninglessness of their meetings. In the moments following this reference
back to the previous topic (transcribed below and immediately preceding
excerpt 8), David and Dalia return to this previous topic of the phone conversa-
tion, and David constructs his imaginary dialogue to the woman, in which he
charges her money. This is the only time throughout the 60 minute cassette I re-
ceived from Dalia in which they speak any English:

26. See Chapter 2, Section 7.2.1 for the collocation nu 'az má? (‘nu so what?’).
 Chapter 3.╇ The textual realm 

Excerpt 14 (‘Tractor Chase’):

–15 David: ... ze gám 'ole kèsef!
this too costs money!
–14 Dalia: ... 'ó--y,
–13 hakó--l kè--sef.
everything [is] money.
–12 David: ... mhm.
–11 Dalia: mhm.
–10 David: ... óne nìght, {David’s imaginary dialogue to the woman}
–9 ... a míllion dòllar,
–8 Dalia: òne-night stánd,
–7 David: for óne night,
–6 Dalia: ... with your wífe,
–5 David: wi
–4 .... I’m gonna téll you,
{-------pp-------}
–3 Dalia: /??????/
–2 David: ... gò to hé--ll!
{-------pp-------}
–1 .... 'ex hi nexneká mize!
{------- mf, slightly laughing---}
how she choked from it!
{i.e., from his telling her he was fed up with the situation}
0 Dalia: ..... /????/
1 David: ..... bekítsu--r,
{-----ff-----}
2 .... shavùa she'avá--r,
week last
last week,
3 .... haya lanu 'eyze kéta--27,
there was to us some segment
we had something happen to us,
4 ... ba'avodá--,

27. For a discussion of the word keta, see Maschler 1998b as well as Chapter 6, Section 1.2.
 Metalanguage in Interaction

at work,
5 .. má ze kèta madlìk.
what is segment cool
such a cool thing.
6 ... halàxnu lexapes 'eyze tráktor,
we went to look for some tractor,
7 .. le--'akél,
to repossess,
8 .. be.. támra,
in.. Tamra,
9 Dalia: .. nu.
go on.
{laughing}
10 David: .. betox hasadòt shel támra.
in the fields of Tamra.
At lines -4 and -2, David constructs his imaginary dialogue telling the woman to
go to hell (line -2). The co-constructed move to English and mention of a million
dollars for one night with your wife (lines (-10) - (-6)) go back to the film Indecent
Proposal, which this imaginary situation most likely evoked for the speakers at this
point in their conversation (the film came out in 1993, about 4 years before this
conversation took place). The imaginary dialogue is constructed pianissimo, in
English, and in future tense. At line -1, back in average loudness and in Hebrew
past tense, he repeats the woman’s response to the words he did have with her;
namely, that he was fed up with the situation. Thus, the contrasts English vs. He-
brew, future vs. past tense, pianissimo vs. mezzo forte all mirror the contrast be-
tween the imaginary and real parts of the previous narrative28. In other words,
they mirror the contrast between its background (what could have happened) and
foreground (what did actually happen). Line -1, then, constitutes the foreground
of the previous narrative. Following a long pause, at line 1, David begins a com-
pletely unrelated story about a tractor he had to repossess that day at work. In
order to highlight this new (much less personal, and therefore perhaps inherently
less foregrounded) narrative, David employs bekitsur in very loud volume and
continues the new story, never again mentioning the previous topic. This func-
tions to move the previous story into the background, making room for the new

28. For a study of the bilingual iconic strategy of language alternation for mirroring discourse
contrasts, see Maschler 1994a and 1997b.
 Chapter 3.╇ The textual realm 

story and foregrounding it more vividly. Dalia’s laughing nu at line 9 is perhaps

evidence of her perception of the abruptness of this move.
Note that the same speaker, David, who in excerpt 8 used be'uktsur, is using
bekítsur in excerpt 14, and bekitsúr in excerpt 11, line 55. Thus, it is not only a mat-
ter of personal style, as in Ferrara’s (1997) study of the sociolinguistic variation
between anyway, anyways, anyhow, and at any rate and the patterns of their distri-
bution in the population. David employs the particular variant of bekitsur as a
function of the particular context and not only as a function of the social group he
belongs to.
This use of bekitsur to introduce new narratives is found in at least29 5 in-
stances (10.4%) throughout the corpus. We see that no digression or background
material is involved here – only the highlighting of a new topic.

6. Grammaticization of bekitsur

Recall that grammaticization is defined as the change “whereby lexical items and
constructions come in certain linguistic contexts to serve grammatical functions,
and, once grammaticalized, continue to develop new grammatical functions”
(Hopper and Traugott 2003: xv). Hopper and Traugott note the principle of unidi-
rectionality in grammaticization, according to which a typical path of grammati-
cization follows a cline of de-categorialization (2003:€106–109). Thus, typically, the
starting point of the process is a full category, such as noun or verb, and over time,
the lexical item becomes more grammatical and “tends to lose the morphological
and syntactic properties that would identify it as a full member of a major gram-
matical category [...] In its most extreme form such a change is manifested as a
cline of categoriality, statable as:
major category (> adjective/adverb) > minor category” (ibid.: 107).
Adjective and adverb thus constitute an intermediate category, whereas the term
‘minor category’ includes grammatical categories such as preposition, conjunc-
tion, auxiliary, pronoun, and demonstrative. In her study of anyway, Ferrara
(1997:€371) adds that the grammatical category of discourse marker should also be
considered a minor category.
We saw that bekitsur came to function as a manner adverb (‘succinctly’) ac-
cording to a productive process of deriving adverbs from adjectival nominals via
prefixation with the preposition b(e)-:

29. There are two possible additional cases of bekitsur beginning new narratives, but the seg-
ment preceding them was not provided in the recordings I received.
 Metalanguage in Interaction

Preposition (b(e)-) + Verbal Noun (kitsur) Adverb (bekitsur).

Once it became an adverb (i.e., entered the ‘intermediate category’), bekitsur fol-
lowed the second half of Hopper and Traugott’s schema and came to function in
the minor grammatical category of discourse marker30. We have also seen that this
process was accompanied by prosodic and phonological changes of bekitsúr into
bekítsur, bekítser, hakítser, 'akítser, kitsúr, be'úktsur, bekítskets, hakítskits and 'akí-
bitser. In its later development, bekitsur followed a path similar to that of other
adverbs which have evolved into discourse markers, such as actually (Aijmer 1986,
Watts 1988, Tabor and Traugott 1998, Traugott and Dasher 2002), well (Jucker
1997, Traugott and Dasher 2002), in fact, indeed (Traugott 1995a, 2003b, Traugott
and Dasher 2002), really, basically (Watts 1988), frankly (Traugott 2001), anyway
(Ferrara 1997, Traugott 2003b), and Hebrew be'emet (‘really, actually, indeed’, lit.
‘in truth’) (Maschler and Estlein 2008). Consistent with the view that “grammati-
calization [...] is motivated by speaker-hearer interactions and communicative
strategies” (Hopper and Traugott 2003:€73), and with the view that “the notion of
emergence [...] relativizes structure to speakers’ actual experience with language”
(Bybee and Hopper 2001:€3), this chapter concerned itself with identifying the pat-
terns of language use which have resulted in this grammaticization in the case of
Hebrew bekitsur.
Grammaticization of bekitsur did not stop once it became a discourse marker.
In this minor grammatical category, it continued to develop. The present corpus
suggests that it came to have three functions. The question arises as to the connec-
tion between these three functions; in other words, what was the functional itiner-
ary followed by bekitsur once it became a discourse marker? I will argue that in its
continuing semantic development, bekitsur came to function more and more me-
talingually. This is in agreement with characterizations of the semantic process
involved in grammaticization as involving a metaphorical shift from the concrete
to the abstract (Traugott 1982, Sweetser 1988), or from the propositional to the
‘metalinguistic’ (Traugott 1988, Traugott and König 1991). In the absence of a

30. The verbal noun kitsur itself followed a more complex path. The root √k.ts.r. is found in
Biblical Hebrew in the verb (‘major category’) of the verbal pattern (binyan) pa'al; i.e., katsar,
with the meaning ‘to reap, to harvest’ in the context of cutting wheat and other grains. The pas-
sive meaning of this verb was then extended (Even-Shoshan 2003) and the verb katsar came to
mean ‘to be short’. From this, according to the cline of de-categorialization, the adjective (‘inter-
mediate category’) katsar (‘short’) was developed. Thus, the first half of Hopper and Traugott’s
cline of de-categorialization (major category > intermediate category) was also followed. From
the adjective katsar, found already in Biblical Hebrew, a verb in the verbal pattern (binyan) pi'el
was later formed, i.e. kitser ‘to lessen, reduce’ (Biblical period), ‘to do something in short’ (Tal-
mudic), ‘to shorten’ (Modern). Kitsur is the verbal noun of this verb.
 Chapter 3.╇ The textual realm 

diachronic spoken corpus, of course, the following grammaticization path of bekit-

sur can only be hypothesized.
As for the first function illustrated in this study, the list summary function, the
discussion of excerpt 2 suggested that a metalingual utterance consisting of a
clause involving a metalingual verb + manner adverb, such as 'im lomar bekitsur
(‘if to say in short’, ‘said in short’, or ‘put succinctly’) developed into a discourse
marker of summarizing, which can be translated as English in short. In this sum-
marizing use, bekitsur manifests Hopper’s principle of persistence in grammatici-
zation (1991:€ 22) in its original sense of ‘retention’ (Bybee and Pagliuca 1987)
rather transparently. We see here a metaphorical shift from the realm of the extra-
lingual world (the concept of ‘shortness’ in the extralingual world) to that of the
text (making a list ‘short’) with respect to the concept of space (cf. Heine, Claudi,
and Hünnemeyer’s discussion of the space-discourse metaphor (1991:€179–186)).
Not only lists, but episodes, too, can be summarized. Following the summary
or conclusion of an episode, a previous topic is often resumed. Excerpts 3 and 4
exhibit two transitional cases in which bekitsur could be said to function both to
summarize / draw a conclusion from a previous episode as well as to return to a
previous topic following a digression.
The summarizing function of bekitsur is not as prevalent as that of returning
from digressions in this corpus. It was employed by only 10.4% of the tokens
throughout the database. I suggest that bekitsur gradually came to be used more
frequently as a resumptive discourse marker, the seeds of which were found in
transitional examples such as excerpts 3 and 4. Approximately 80% of all bekitsur
tokens in this database function as resumptive discourse markers (77.6% signaling
resumption and 2% requesting resumption by listener). In this later use, one is no
longer required to follow bekitsur with an actual summary of the previous episode,
and bekitsur comes to signal a return to the main topic following the digression
that has just ended (excerpt 5, e.g.). Hopper’s principle of persistence (1991:€22) is
still manifested, though less transparently than for summarizing bekitsur: by end-
ing digressions, the discourse as a whole is made shorter. Because of this lesser trans-
parency, resumptive bekitsur is considered higher in the metalingual dimension
compared to summarizing bekitsur.
This stronger metalingual dimension is paralleled by a prosodic fact: when
bekitsur functions as a resumptive discourse marker, it is preceded by sentence-
final intonation, whereas in the case of list-summarizing bekitsur, the immediately
preceding unit may end in continuing intonation. Thus, while Tabor and Traugott
emphasize the intonation contour of the English discourse marker anyway itself
(“Typically […] disjunctive, requiring comma intonation” (1998:€254)), here it is
the intonation contour of the preceding intonation unit that is focused on. We
have seen that episode-summarizing bekitsur, too, follows sentence-final
 Metalanguage in Interaction

intonation, being a transitional case between the two categories. These prosodic
patterns are consistent with the iconic strategy of marking the boundary more
clearly (i.e. with sentence-final and not continuing intonation) when a higher-
level boundary is concerned (see Chapter 1, Section 7, Maschler 2002a).
On the basis of the current corpus, it is not possible to prove that a metalingual
utterance such as one involving the phrase lomar bekitsur ‘put succinctly’ or ‘said in
short’ is the origin of the resumptive discourse marker bekitsur. However, in support
of this claim I found two pieces of evidence. The first is a quote from an excerpt of
written Hebrew published in the Hebrew daily Ha'aretz by Amia Lieblich, a leading
professor of clinical psychology and author of several novels. In this rather formal
genre, we would not expect to find the colloquial resumptive discourse marker bekit-
sur which is so frequent in casual conversation. Instead, we find something else. In
reviewing a recent collection of psychotherapeutic case-studies published by the
psychotherapist Yoram Yovel, Lieblich begins the following paragraph:
ve'axen, giborey hasefer haze 'eynam haponim le'ezra 'ela dmut hametapel-hames-
aper ha'ola mimenu. yihiyu mi sheya'asku beshushelet hayexasim shel yovel (hu
nexdo shel yesha'ayahu leybovitch), kfi she'ose hamol, ve'af hamexaber, bitsni'ut
'amitit, po vesham beyn dapey hasefer. 'ani 'omar bekitsur: ze 'adam, shekol mi
shnizkak 'o yizdakek 'ey pa'am letipul nafshi hayiti rotsa le'axel lo sheyifgosh bo.
'emdato hatipulit hi kazot shehayiti rotsa lehaknot letalmiday. hu metapel ragish
ve'empati (kfi she'omrim 'etslenu) haboxen lelo heref 'et tguvotav shelo umesik me-
hen maskanot lemahalax hatipul. veyaxad 'im zot hu maskil me'od....

Indeed, the protagonists of this book are not those seeking [therapeutic] help, but
rather the image of the therapist-narrator that emerges from it. There are those
who will concern themselves with Yovel’s family tree (he is the grandson of
Yesha'ayahu Leibovitch [a prominent philosopher]), as does the publisher, and
even the author, in genuine modesty, here and there between the pages of the book.
I will say bekitsur: This is a person I wish for anyone, who needs or ever will need
therapy, to have the good fortune of meeting. His therapeutic stance is one which I
would like to impart to my students. He is a sensitive and empathetic therapist (as
we say in our jargon) who constantly examines his own responses and draws con-
clusions from them for the continuation of the therapy. And in addition to all this,
he is highly knowledgeable... (Lieblich 2001, translation and emphasis mine).

With the sentence 'ani 'omar bekitsur (‘I will say in short / succinctly’), the writer
is not summarizing previous discourse. Rather, she is moving from what she con-
siders of secondary importance (the fact that Yovel is the grandson of a very prom-
inent philosopher) with which others will concern themselves, to what she consid-
ers most prominent – Yovel’s virtues as therapist. Furthermore, what comes next is
a long elaboration of Yovel’s virtues (only the beginning of which is reproduced
here); in other words, nothing ‘short’. Because of the similarities between this
 Chapter 3.╇ The textual realm 

context and the colloquial ones we have seen throughout this study, I surmise that
metalingual utterances of this type are the source of the colloquial resumptive
discourse marker bekitsur.
The way to prove this would have been to collect instances of the adverb bekit-
sur from an earlier corpus of spoken discourse and see how frequently they oc-
curred with a verb of saying, such as lomar, lehagid (‘to say’) or lesaper (‘to tell’) in
such contexts31. The present corpus, as will be recalled, contains no bekitsur tokens
functioning as adverbs. Given that the discourse marker function was not pro-
vided in the dictionary at all (and considering that dictionaries are less likely to list
newer, more colloquial meanings), this in itself is support for the grammaticiza-
tion pattern suggested here32.
In the absence of an earlier corpus, I offer a second piece of evidence support-
ing the origin of bekitsur in metalingual utterances. The explanation given in the
dictionary for the higher register equivalent of bekitsur – biktsara – involves a
highly metalingual context: ‘lo ba'arixut, bekitsur, bemilim mu'atot’ (‘not at length,
succinctly, in few words’). Furthermore, the example provided: saper-na bikstara
'ex kara hadavar (‘pray-tell succinctly how the matter happened’) involves a verb
of saying, saper (‘tell’), also high in the metalingual dimension. Thus, when pro-
viding an example for the adverbial form, the example Even-Shoshan comes up
with is highly metalingual. Examples given in dictionaries presumably have some-
thing to do with frequency and representativeness. One might conclude, then, that
the frequency of co-occurrence of bekitsur with a metalingual verb such as lomar
(‘to say’) or lesaper (‘to tell’) was high when this dictionary was written, further
supporting the grammaticization path suggested here.
In resuming a previous topic, the immediately preceding discourse is retroac-
tively constructed as a digression and backgrounded, while future discourse is fore-
grounded. This backgrounding function of bekitsur was extended to contexts in which
no real digression is apparent, such as between the various episodes of a narrative
(excerpts 10, 11, 12) in order to guide the listener in following the main storyline.

31. Cf. this methodology with that of Tao 2001, 2003, the volume entitled ‘Frequency and the
Emergence of Linguistic Structure’ (Bybee and Hopper 2001), Thompson 2002, and Hopper and
Thompson 2008.
32. Cf. Ferrara’s study of English anyway: “A further indication of grammaticalization is the
noted increased frequency of the discourse marker (89% discourse marker vs. 11% adverbs in
the corpus)” (1997:€372).
 Metalanguage in Interaction

Adverb Summarizing D.M.

within clause, List-summarizing: following continuing intonation
within intonation unit Episode-summarizing: following final intonation
Resumptive D.M. Foregrounding D.M.
following final intonation following final intonation, accompanied by addi-
tional prosodic means
Figure 1.╇ Functional itinerary of bekitsur and corresponding prosodic properties

In the most grammaticized and least transparent function of bekitsur, its foreground-
ing function was extended to the foregrounding of new narratives. This strategy can
move a previous foregrounded narrative into the background, thus foregrounding
the new narrative more brightly (excerpt 14). Prosodically, this type of bekitsur is
not only preceded by sentence-final intonation, but also accompanied by other pro-
sodic means, such as a longer pause or a drastic change in volume.
The functional itinerary of bekitsur and the corresponding prosodic properties
are summarized in Figure 1.
We have seen that all but one function of discourse marker bekitsur (‘list-sum-
marizing: following continuing intonation’) fulfill both the semantic and struc-
tural requirements for prototypical discourse markerhood and thus constitute
prototypical discourse markers.

7. Interacting as an Israeli via bekitsur

To conclude this chapter, I would like to return to the most common function of
bekitsur in this corpus. It is not self-evident that in order to return to a main topic,
a word meaning also succinctly or in short would come to be used. The most wide-
ly used English resumptive discourse marker, anyway, for example, does not tie
resumption of main topic with the concept of ‘shortness’. Furthermore, other high-
er register Hebrew resumptive discourse markers which are more literal transla-
tions of anyway exist, but they are much less frequent: bexol 'ofen (‘in any way/
manner’, 3 tokens in this database), bexol mikre (‘in any event’, ‘at any rate’, 0 tokens
in this database).
Upon hearing that I was studying bekitsur, I have often been told by Israeli
friends that using this word implies some reproach that one is not speaking tersely
and to the point. This hue of reproach does not accompany a resumptive discourse
marker with no associations of terseness, such as anyway. The grammaticization of
the resumptive discourse marker bekitsur from an adverb meaning succinctly or in
short is consistent with a particular aspect of Israeli culture and identity.
 Chapter 3.╇ The textual realm 

According to Katriel, the term ‘Sabra’,

which literally denotes a local cactus bush and its thorn-covered fruit, was applied
metaphorically [‘thorny outside but sweet inside’, cf. Oring 1981] to Jewish chil-
dren born in Palestine during the British mandate (roughly, between the early
1920s and late 1940s). These youngsters’ formative years were spent in the preÂ�
statehood educational system and in other socializing contexts such as the Zionist
youth movements, where the mythic image of the native-born New Jew, the Sabra,
was cultivated. (2004:€139).

The construction of the identity of this New Jew, an identity now in decline (Kim-
merling 2001, Katriel 2004, and see Chapter 4), involved first and foremost con-
structing new social practices and in particular, new ways of speaking and inter-
acting. The main staple of these new ways of interacting is the Sabra’s dugri (‘direct’)
speaking style (Katriel 1986, 2004), which emerged in opposition to the speaking
style of the Diaspora Jew.
While forthrightness is not at issue in this chapter (but see Chapter 4), this
characteristic of the Sabra speaking style goes along with another, more relevant
characteristic:
According to the “New Jew” ideology, the Israeli Jew was to be everything the
Diaspora Jew was not. In communicative terms, this cultural gesture of refusal
implied the rejection of ways of speaking associated with Jewish diaspora life,
[and] with European bourgeois culture. [...] As depicted in Zionist ideology, di-
aspora Jews’ ways of responding to life’s exigencies, and especially their ways of
interacting with the non-Jewish world, were marked by a sense of restrictiveness,
defensiveness, and passivity as an adaptive mechanism. Traditionally, Jews recog-
nized the value of using speech adroitly as a way of coping with a potentially
hostile environment. Furthermore, Jewish talmudic tradition colored Jews’ dispo-
sition toward pilpul, an elaborate form of discussion that involved a recognition of
the complexity and many-sidedness of issues, the inherent ambiguity of human
affairs, and the role of confrontation and debate in clarifying issues.
[…The New Jew, by contrast, was to prefer] the modernist, pragmatic ethic un-
derlying the Sabra’s straight talk, which privileged nonmanipulative openness,
simplicity of expression, and explicitness of purpose. (2004:€140–141).

In this new way of speaking, then, not only forthrightness, but also terseness and
speaking ‘to the point’ are valued. The New Jew is no longer to be debilitated by a
sensitivity to the complexity of issues and to external social pressures leading to
obfuscation and indirectness. When “simplicity of expression” and “explicitness of
purpose” are the preferred ways of speaking, it is not surprising that digressions,
while very common, will be perceived as an object of reproach.
 chapter 4

The cognitive realm

The discourse marker ke'ilu:
Realizing the need to rephrase

1. Introduction1

In a comparative study of the English discourse marker like and its French equiva-
lent genre, Fleischman and Yaguello (1999, 2004) address the question posed by
Traugott of “whether there are cross-language generalizations to be made about the
development of discourse particles in terms of both their likely semantic sources
and their semantic-pragmatic paths” (Traugott 1995a: 4). In particular, Fleischman
and Yaguello address this question in relation to the phenomenon of different lan-
guages independently grammaticizing markers with the same range of functions,
and having similar lexical sources in their corresponding languages. In the case of
English like and French genre, this lexical source involves some comparative mean-
ing “whereby an item is considered in relation to a norm or paradigm” (2004:€139).
In the case of Hebrew talk-in-interaction, we find the use of two more ele-
ments involving comparative meaning, kaze (‘like’, lit. ‘like this’) and ke'ilu (‘like’,
lit. ‘as if ’). Examine, for instance, the following excerpt recorded in Haifa in 1998,
from a conversation between two close friends in their early twenties, David and
Dalia. David, while cooking some soup in his student-apartment kitchen, de-
scribes to Dalia how, as part of his day at work, he ran after a guy whose tractor it
was his job to repossess:
Excerpt 1 (‘Tractor Chase’):
46 David: .... ve'ani kmo metumtám,
and I like an idiot,
47 ... rodéf 'axaràv.
am running after him.

1. A similar study, based on only 28 conversations comprising the present corpus, appeared as
Maschler 2002b ‘On the Grammaticization of ke'ilu (‘like’, lit. ‘as if ’) in Hebrew Talk-in-Interac-
tion’ in Language in Society 31: 243–276. It has been extended and revised here to fit in with the
present study. The article was dedicated to the memory of Suzanne Fleischman. Section 7 is new
but contains a few paragraphs published in Maschler 2001.
 Metalanguage in Interaction

48 ... ke'ilu
like
49 .. barégel,
by foot,
50 'axarèy tráktor,
after [a] tractor,
51 betòx kol habóts haze?
in all the mud this
in all this mud?
52 ... veke'ilu haragláyim sh'xa,
and like the feet your (masc. sg.)
and like your feet,
53 nitka'ot bifním kaze,
get stuck inside like
get stuck inside like,
54 kx kx kx kx. {sounds imitating the sound of feet in the mud}
55 ... bekitsúr vela'inyá--n,
in short and to the point
in short,
56 ... 'axaré--y,
after,
57 .. mirdáf she--,
a chase that,
58 ne'eràx kemispàr dakó--t,
lasted roughly a few minutes,
59 .... hitslàxti lehagia 'ad latráktor,
I managed to get to the tractor,
There are two tokens of the utterance ke'ilu in this excerpt and one token of kaze,
both translated into English here as ‘like’. Both words begin with the Hebrew prepo-
sition of comparison k(e)- (‘as’, ‘like’)2. In the case of ke'ilu, this preposition is at-
tached to the conjunction of hypothetical conditionals, or counterfactual conjunc-
tion 'ilu (‘if ’, irrealis). In the case of kaze, it is attached to the demonstrative ze (‘this’,
masc. sg.). Literally, then, ke'ilu is to be translated as ‘as if, as though’ and kaze as
‘like this’. However, new uses have emerged for these two utterances in Hebrew

2. The element of comparison k(e)- is apparent also in line 46, in the preposition kmo ‘like’.
 Chapter 4.╇ The cognitive realm 

talk-in-interaction, different from their literal meanings. This chapter is concerned

with the emergence of these new meanings, focusing on the case of ke'ilu.
Non-literal uses of kaze and ke'ilu have become extremely widespread over the
past decade or so, particularly in the language of younger Israelis, to the point that
they are often stigmatized as ‘the kaze ke'ilu generation’, who employ these expres-
sions ‘too frequently’ in prescriptivists’ eyes. The Slang Lexicon (1993) defines
both of these words as “void fillers which find their way into the sentence for no
reason” (translation mine). In September 1990, an issue of the Israeli magazine
Politika, dealing with Israeli society, policy, and culture (Samet 1990), was devoted
to the study of the then 16–21 year-olds. The subtitle of that issue was dor hakaze
ke'ilu, ‘the kaze ke'ilu generation’. In that magazine, we find an article by Ariel Hir-
schfeld entitled 'al kaze veke'ilu (‘On kaze and ke'ilu’) in which Hirschfeld, a liter-
ary critic at the Hebrew University who also teaches at a prestigious high school in
Jerusalem, discusses these expressions from a cultural-philosophical perspective,
based, among other things, on his conversations with his students. Hirschfeld’s
study is often linguistically naïve, but his students provide an angle which comple-
ments the present study from a cultural perspective, adding what might be found
if one were to conduct playback interviews (Gumperz 1982) with participants. I
will therefore return to Hirschfeld’s findings throughout this chapter.
Hirschfeld ends his essay with the statement that when these high school stu-
dents “grow up and graduate from the army, the kaze and the ke'ilu disappear from
their language” (1990:€9, translation mine). Over almost two decades later, now
that these teenagers have long graduated from the army to become adults, kaze
and ke'ilu have far from disappeared from the language.
About a decade following Hirschfeld’s article, kaze and ke'ilu had become as-
sociated with a style of speaking called tsfóni (‘Northern’) in Israel. The stress on
the first syllable of tsfóni indicates that the implication is not Northern Israel, but
rather Northern Tel Aviv, i.e. the most urban and yuppie region in the country. In
an interview with Shula Modan, a woman in her early fifties, a successful pub-
lisher, psychotherapist, and author of novels, children’s books, and a ‘down to
earth’ cookbook, we find two relevant utterances. She is cited in a weekend issue of
the Ha'aretz daily newspaper, criticizing the pretentious yuppie culinary tradition
that has been sweeping Israeli society in recent years:
hakol kaze-ke'ilu, lo shayax latarbut shelanu, ba mibaxuts.
‘Everything is kaze-ke'ilu, doesn’t belong in our culture, comes from out-
side’.
 Metalanguage in Interaction

Modan is described by the interviewer as

soledet mitarbut hahay vehabay umityaxeset bevuz lanefixut shenotsra sviv nose
habishul, bimdorey ha'oxel ba'itonim uvesifrey bishul 'axerim. “hakaze vehake'ilu”,
hi mexana 'et hatofa'a, kmo bat kfar shebemikre legamrey nikle'a la'ir hagdola.
‘revolted by the culture of the ‘hi’ and the ‘bye’ [i.e., by the use of English words
often punctuating Hebrew speech] and treating with contempt the pompousness
created around the topic of cooking, in the food columns of the papers and in
other cookbooks. “The kaze and the ke'ilu”, she calls these phenomena, like a vil-
lager who happened upon the big city’ (Lori 2000:€87–88, translation mine).

There was also a comedy show in the 1990s called kaze ke'ilu. And in a best-seller
(Zeltzer 2000:€68–69) about charming corners to visit in Israel, a section describ-
ing trendy cafés on Shenkin Street – perhaps Tel Aviv’s most yuppie street – in
which to enjoy a Friday3 morning breakfast, is entitled ‘Tel Avivi kaze ke'ilu’. One
of these cafés is called kafe kaze. Finally, my daughter attending elementary school
during the late 1990s was instructed by her teacher not to use kaze ke'ilu when
speaking up in class.
I cite all this to emphasize that the frequent employment of kaze and ke'ilu has
gained attention not only among linguists, but among lay-people as well. Further-
more, they ‘go together’ in people’s perception, suggesting that they share some
properties in common. Kaze, however, is not a discourse marker since it is not
very high in the metalingual dimension and hardly ever occurs intonation-unit
initially (see Chapter 1, Section 5). I focus here only on ke'ilu, while the functions
of kaze are investigated in Maschler 2001.
New uses of ke'ilu have emerged, of course, not only in the language of young-
sters, and not only among yuppies from Northern Tel Aviv. For example, in a study
of Hebrew-English bilingual conversation, a close comparison was made between
the speech of the same two speakers at two time periods, twelve years apart (Mas-
chler 2000c). The first set of data was collected in Jerusalem in 1986 and 1987, and
the second set was collected in 1998, when the speakers were in their mid-forties.
From no cases of non-literal ke'ilu in the 40-minute 1986 corpus or in the 20 hours
of follow-up interviews conducted with these speakers in 1987, to 9 cases of non-
literal ke'ilu in the 33-minute interaction twelve years later, we can clearly see that
ke'ilu has entered the bilingual grammar emerging in these speakers’ talk as well
(Maschler 1997b).

3. The weekend spans Friday and Saturday for most Israelis. Friday morning, when the chil-
dren are in daycare or at school but the shops are open, is a time many young adults enjoy leisure
activities not involving children, such as relaxing with the weekend paper in a café.
 Chapter 4.╇ The cognitive realm 

Fleischman and Yaguello 2004 mention ke'ilu as one ‘equivalent’ of English

like4.Ke'ilu was investigated as part of the general study of the system of discourse
markers segmenting Hebrew talk-in-interaction (Maschler 1997a, 1998b, 2002a,
see Chapter 1, Table 2). In those studies, which examined only tokens collected in
the period 1994–1997 (at most 16 of the 50 conversations of the present database),
only those tokens of ke'ilu (1) functioning metalingually in the contexts in which
they occurred, and (2) occurring at intonation-unit initial position (or in a dis-
course marker cluster at intonation-unit initial position) were investigated. Ke'ilu
satisfying these two requirements was characterized as a cognitive discourse
marker of ‘realizing the need to rephrase’.
For example, the ke'ilu of excerpt 1, line 48, manifests this function. In the two
preceding lines (46–47), David describes himself running after the man driving
the tractor: ve'ani kmo metumtám, rodéf 'axaràv. (‘and I like an idiot, am running
after him.’). Following this statement comes a rephrasal, further explicating the
phrase ‘like an idiot’ and elaborating on the ridiculousness of the situation: by foot,
after a tractor, in all this mud? (lines 49–51). This further elaboration along with
the rising question intonation at the end of the utterance suggest that the motiva-
tion for rephrasal here is the need to involve the audience. Preceding this self-re-
phrasal is the discourse marker ke'ilu, fulfilling both requirements for prototypical
discourse markerhood: (1) It functions metalingually, referring to the cognitive
processes of the speaker (realizing the need to rephrase) as opposed to referring to
something in the extralingual world. In other words, the expression ke'ilu itself
refers to what is happening in the speaker’s mind during verbalization, a process
that may be paraphrased as ‘what I just said should be rephrased / elaborated on....’.
(2) It appears at intonation-unit initial position (line 48), following final intona-
tion in same-speaker talk (line 47)5.
Henkin 1999 studied the hedging functions of kaze and ke'ilu, among other
expressions, based on modern Hebrew informal written discourse6. However, her

4. Given the fact that the Israelis of this corpus are quite fluent in English (it is the most
prominent second language among the Jewish population of the country and is learned at
schools starting from around fourth grade and often earlier), we must consider the possibility
that the grammaticization of Hebrew ke'ilu was influenced by that of English like. This possibil-
ity is highly unlikely, first, because there are some significant differences in the grammaticiza-
tion of the two (see below); and second, because, as Jane Hill has pointed out to me (p.c.), Israe-
lis speaking English do not use like in the American way.
5. However, we will see below that ke'ilu differs from the other discourse markers in the data-
base in that it often follows non-final intonation in same speaker talk as well.
6. A few tokens of ke'ilu in Henkin’s study consist of isolated utterances collected on a Kibbutz
(1999:120, footnote 21).
 Metalanguage in Interaction

study focuses mostly on kaze. Henkin is aware of the “distortion of the spoken lan-
guage in its representation in literature” (1999:€118, footnote 2, translation mine).
Ke'ilu has other functions in the discourse besides hedging and self-rephrasal.
Maschler 2002b investigates these functions in 28 out of the 50 conversations
comprising the corpus of the present study, i.e. those collected during the years
1994–1999. In the present chapter, I investigate the various uses of ke'ilu on the
basis of its distribution in the entire 50 conversation corpus, i.e. including the
2000–2002 data. I examine all tokens of this word found throughout the database
– those considered prototypical discourse markers in previous studies as well as
those that were not, and those that occur both intonation-unit-initially and non-
initially – in hope of further contributing to the elucidation of the grammaticiza-
tion and functional itinerary of this word and to the question posed in Traugott
1995a concerning implications for cross-language pragmatics.

2. Data and methodology

The 50 conversations of the Haifa Corpus of Spoken Israeli Hebrew (Maschler

2004) on which the present study is based (see Chapter 1, Section 3) are divided
into two groups according to time of data collection: (1) 25 conversations com-
prising approximately 79 minutes of talk collected during the years 1994–1997, (2)
25 conversations comprising approximately 72 minutes of talk collected during
the years 1998–20027.
All tokens of ke'ilu employed throughout the database were considered for this
study. Altogether, there were 120 tokens of ke'ilu. Table 1 shows only the non-liter-
al tokens, as they were distributed throughout the two time periods of data collec-
tion. Altogether, there were 113 tokens of non-literal ke'ilu.

Table 1.╇ Distribution of non-literal tokens of ke'ilu

1994–1997 data 1998–2002 data Total (1994–2002)

79 minutes 72 minutes 151 minutes

31 82 113

7. The cut-off point was decided to be 1997/98 because this is when it seemed to me that kaze
and ke'ilu had begun to become extremely widespread. Coincidentally, it turned out that this
resulted in two sub-corpora of similar size. Since the conversations were collected in a similar
fashion in the case of both sub-corpora, the average age of speakers is roughly the same in the
two databases (mostly students in their early twenties, with family members occassionally par-
ticipating). None of the speakers contributed to both sub-corpora.
 Chapter 4.╇ The cognitive realm 

The 1998–2002 data, then, show significantly more employment of ke'ilu in its
non-literal uses compared to the 1994–1997 database. Ke'ilu is employed at an
average frequency of one token approximately every 2.55 minutes in the earlier
data, as opposed to one token approximately every 0.88 minutes in the later cor-
pus. Compared with the figure of one token approximately every 0.4 minutes for
the 1998–1999 database (Maschler 2002b), this data supports the general impres-
sion that this word has continued to greatly proliferate in the language during the
later years of corpus construction.

3. A quantitative perspective on the different functions of ke'ilu

Surveying several studies of English like and examining their own data on like and
its French equivalent genre, Fleischman and Yaguello (1999, 2004) summarize the
following pragmatic functions of English like:
HEDGE (Schourup 1985, Jucker and Smith 1998, Andersen 1998)
FOCUS MARKER (Underhill 1988, Miller and Weinert 1998)
QUOTATIVE (Blythe et al. 1990, Butters 1982, Schourup 1985, Tannen
1986, 1989, Romaine and Lange 1991, Ferrara and Bell 1995, Fleischman
1995, Haiman 1993, Lucy 1993, Tagliamonte and Hudson 1999).
These functions of English like were found relevant for Hebrew ke'ilu to differing
extents. Table 2 shows that another function – self-rephrasal (not considered as a
separate category by Fleischman and Yaguello) is crucial when we consider this
expression.
Table 2 shows that about half the ke'ilu tokens (48.3%) are employed in self-
rephrasals. The second most common use of ke'ilu is that of focus marker (29.2%);
next comes its hedging function (13.3%). Seven ke'ilu tokens (5.8%) are employed
in the literal sense, and there are only 4 ke'ilu tokens (3.3%) preceding quotes.
Fleischman and Yaguello suggest the following functional itineraries for Eng-
lish like and French genre.

Table 2.╇ Functional distribution of ke'ilu tokens

LITERAL HEDGE FOCUS SELF- QUOTE TOTAL

(‘as if ’) REPHRASAL

7 16 35 58 4 120
(5.8%) (13.3%) (29.2%) (48.3%) (3.3%) (100%)
 Metalanguage in Interaction

LEXICAL MEANING: ‘similar to’, ‘approximately’

FOCUS ? QUOTATIVE ? HEDGE

Figure 1.╇ Development of the pragmatic functions of like/genre (Fleischman and Yaguello 1999)

As illustrated in this figure, Fleischman and Yaguello see the three functions of
focus, quotative, and hedge as stemming from the lexical meanings of these words
in English and French, and they hypothesize that there are also paths of develop-
ment between these three functions in the two languages. The analysis presented
below will support this hypothesis for Hebrew as well. We turn now to a qualitative
analysis of the functions of ke'ilu in order to understand the employment and
functional itinerary of this utterance in Hebrew.

4. Functional distribution of ke'ilu: A qualitative perspective

4.1 Ke'ilu as a conjunction in the literal sense

Ke'ilu has an additional element of lexical meaning in comparison to English like,

because it is composed not only of the comparative k(e)-, but also of the irrealis
conditional 'ilu. We will see how this affects its functional itinerary. 'ilu itself is com-
posed of the conditional conjunction 'im (‘if ’) plus the hypothetical conditional, or
counterfactual conjunction lu. The definitions Even-Shoshan’s dictionary (1986)
gives for ke'ilu are: kmo 'ilu (‘as hypothetical if, as though’) (i.e., a conjunction) and
dome le- (‘similar to’). I relate to the latter meaning of ke'ilu in the next section. As
a conjunction in the literal sense, ke'ilu is found already in Talmudic Hebrew:
bexol dor vador xayav 'adam lir'ot 'et 'atsmo ke'ilu hu yatsa mimitsrayim
‘In every generation one is obliged to see oneself as though he/she him/herself
had actually gone forth from Egypt’ (Psaxim 10:5, translation mine).

There are only 7 tokens of ke'ilu throughout the database in this literal conjunction
sense (‘as if ’, ‘as though’). For instance, in the following excerpt, Roni, a secular
Israeli, ridicules the minute details of keeping the Jewish custom of letting the soil
rest every seventh year (called ‘the shmita year’) by not planting anything in it dur-
ing that whole year. In the following utterance he is talking about a plant in a
flower pot with a hole in its bottom (i.e., having contact with the soil):
 Chapter 4.╇ The cognitive realm 

Excerpt 2 (‘Shmita Year’):

144 Roni: 'asúr lehaziz 'oto mimkomò bixlal.
it’s forbidden to move it from its place at all
it’s forbidden to move it at all from its place.
145 .. ze ke'ilu she'ata shotél 'oto bemakom 'axèr.
it’s as though that you plant it in place another
it’s as though you plant it in another place.
Roni says that a flower pot that has contact with the soil may not be moved from
its place all shmita year (line 144). If one moved it, that would count ke'ilu (as
though) one had planted it in another place (line 145), an action that is against the
shmita custom. We see that ke'ilu is used here literally at intonation-unit non-ini-
tial position, immediately preceding the subordinator she- (‘that’) and a full ac-
companying clause: 'ata shotel 'oto bemakom 'axer (‘you plant it in another place’).
Literal ke'ilu is also found at intonation-unit initial position. The following
conversation concerning an opera singer takes place between a man, Eyal, and two
women, Yonat and Hila:
Excerpt 3 (‘Women’):
145 Eyal: .. hi hayta kazót magnivà,
she was like this cool
she was so cool,
146 ... vehi gam shára,
and she also sang,
147 .. kol káx magnìv,
so cool,
148 .. shepit'óm,
that suddenly,
149 Yonat: baxura tse'irá?
woman young
a young woman?
150 Eyal: ... kén.
yeah.
151 Yonat: ... 'ani yodá'at mì zot.
I know who this
I know who she is.
152 Eyal: .. pit'om kaláteti,
suddenly I got it,
 Metalanguage in Interaction

153 ve'amárti,
and I said,
154 ... waí,
wow
155 Hila: má 'at yodà'at mì zot!
what [do you mean] you know who this
what [do you mean] you know who she is!
156 .. ke'ìlu sheyésh 'eh,
as if that there is uh
as if there’s uh,
157 .. zamèret 'opera 'axát.
singer opera one
one opera singer.
Eyal begins to tell about an opera singer he found very attractive (lines 145–147).
Yonat overlaps him, asking whether she is young (line 149). When Eyal answers
affirmatively (line 150), Yonat says she knows who the singer is (line 151). Hila
then objects at line 155: what [do you mean] you know who she is! ke'ilu (‘as if ’)
there's uh one opera singer (lines 155–157). Here, again, the literal meaning of ke'ilu
is used, and it is followed by the subordinator she- (‘that’) and a full clause (span-
ning two intonation units). Note that both here and in the preceding example, the
hypothetical condition is ridiculed by the speaker.
In this database, ke'ilu as a conjunction in the literal sense almost always8 pre-
cedes a full clause which often begins with the subordinator she- (‘that’). We have
seen that this use of ke'ilu is extremely limited – only 5.8% of all tokens of ke'ilu
throughout the database (7 out of 120 tokens).

8. In one of the 7 cases, there is no full clause following literal ke'ilu. The speaker describes
two drunken women who had had a serious fight upon coming home at night, but in the morn-
ing woke up ke'ilu klum (‘as if nothing’):
‘Drunk on Campus’:
169 Yifat: .. lemoxorat,
the next day,
170 kamot baboker,
[they] wake up in the morning,
171 ke'ilu klum.
as if nothing.
172 .. lo zaxru klum.
not remember nothing
didn’t remember anything.
 Chapter 4.╇ The cognitive realm 

4.2 Ke'ilu as hedge

Henkin 1999 examined ke'ilu as a hedge in informal written Hebrew discourse. In

the present corpus of spoken discourse, we find only 13.3% of ke'ilu tokens func-
tioning as hedges. The hedging function is related to one of the uses of ke'ilu listed
in Even-Shoshan’s 1986 dictionary: dome le- (‘similar to’), and is derived directly
from the approximative meaning of the preposition k(e)-. When ke'ilu functions as
a hedge, the speaker acknowledges a mismatch between what is said and “what he/
she has in mind or feel ideally might or should be said” (Fleischman and Yaguello
2004:€131). There is a ‘loose fit’ (Schourup 1985) between the expression and the
intended meaning.
Interestingly, since the publication of the initial version of my study of ke'ilu
(Maschler 2002b), in the recently published 2003 edition of Even-Shoshan’s dic-
tionary, ke'ilu has changed its status. In the 1986 version of this dictionary, it does
not appear as an entry of its own but rather as a sub-entry of 'ilu (the irrealis con-
ditional conjunction) with the two meanings of kmo 'ilu (‘as hypothetical if, as
though’) (i.e., a conjunction) and dome le- (‘similar to’). It is listed there, as we have
seen, as originating already in Talmudic Hebrew. In the 2003 version of the dic-
tionary, however, ke'ilu appears in addition to this as an entry of its own. There it
is listed as an adverb, and two additional meanings from the modern period of
Hebrew are given: 1) kivyaxol (‘supposedly’) and 2) [colloquial] haba'a hamexu-
venet lehaxlish 'et mashma'ut hadvarim (‘an expression meant to weaken the mean-
ing of the utterance’), in other words, a hedge. Neither of these new definitions,
however, covers the additional functions of ke'ilu found in the present study.
Henkin characterizes hedging ke'ilu as common mostly in conversation be-
tween people of differing status (1999:â•›116). The spoken data upon which the
present study is based consist mostly of interactions among friends, often very
close ones. Status distinctions play no role in their relationships. Nevertheless,
their speech shows a considerable degree of hedging ke'ilu.
The position of hedging ke'ilu relative to the element it hedges is flexible. In
this corpus, we find it preceding the utterance it modifies, following it, and also in
the middle of the modified utterance.

4.2.1 Post-positioned ke'ilu as hedge

In the following excerpt, the mismatch between the words chosen and the mean-
ing intended is commented on in the continuation of the text:

The following intonation unit (line 172), comprised of a repetition of klum (‘nothing’) of line
171 in a clause construction (as object of the clause lo zaxru klum (lit. ‘didn’t remember noth-
ing’)), supports the interpretation of this occurrence of ke'ilu as the conjunction ‘as if ’ in the
literal sense.
 Metalanguage in Interaction

Excerpt 4 (‘Family Gossip’):

183 Sara: ... 'axshav hú--,
now he
184 .... tsoxék 'aleha ke'ìlu,
laughs at her sort of,
185 .. káxa ze hate
this is the desc
186 .. te'ùr shel ronít.
description of Ronit
this is Ronit’s description.
187 .. má zot 'omeret hu tsoxèk 'aleha,
what this says he laughs at her
what does it mean ‘he laughs at her’,
188 .. hu 'omér la,
he says to her,
189 .. 'an’ló mevin mà lakaxt 'arixà--,
I don’t understand what you took editing
I don’t understand why you took editing,
190 ... 'at me'òd 'ití--t,
you’re very slow,
Sara is discussing a couple, Ronit and her boyfriend. The man is highly critical of
Ronit’s choice of taking an editing course. Sara uses the verb tsoxek 'aleha (‘laughs
at her/ridicules her’, line 184) to describe this lack of respect for Ronit’s decision,
but she immediately realizes it is not adequate for what she intends. Sara hedges it
with ke'ilu and continues to relate to the inadequacy of her choice of words: they
are Ronit’s words kaxa ze hate'ur shel ronit (‘this is Ronit’s description’, lines 185–
186), not Sara’s. She further explicates this choice of words – ma zot 'omeret9 hu
tsoxek 'aleha (‘what does it mean “he laughs at her”’, line 187) – by providing an
instantiation of his lack of respect: he says to her ‘I don’t understand why you took
editing, you’re very slow’ (lines 188–190). These further explanations indicate the
speaker’s lack of satisfaction with her initial word choice.

4.2.2 Pre-positioned ke'ilu as hedge

In the following excerpt, Eyal describes the type of woman he likes:

9. Note the metalingual utterance ma zot 'omeret (‘what does it mean’) here. This, I think, is
the origin of one equivalent of the discourse marker ke'ilu, s'tomeret (‘I mean’, lit. ‘this means’),
see Section 4.4.
 Chapter 4.╇ The cognitive realm 

Excerpt 5 (‘Women’):
99 Eyal: lì gám,
to me also
I too,
100 haytà 'et hahitgalút hazot.
there was the revelation this
had this revelation.
101 .. tamíd 'ani 'ahàvti,
always I liked
I always used to like
102 baxurót ke'ilu,
young women like
young women like,
103 shenir'òt mamash fíks.
who look really perfect.
104 .. ke'ilu,
like,
105 .. 'at yoda'at,
you know,
106 .. razót ka'elu,
thin like these
thin sort of,
107 .. 'im... xazè gadól,
with... bust large
with a large bust,
108 .. ve.. yi.. sheyihiyù norà yafót,
and..the..they should be terribly pretty,
109 .. ve-- 'eh bapaním,
and uh in [their] face,
110 .. vekol miney shtuyót ka'èlu.
and all sorts of nonsense like these.
To describe the type of woman he used to like, Eyal uses the expression nir'ot ma-
mash fiks (‘look really fiks’, line 103). Fiks is slang for something that is perfect, just
the way it should be. The inadequacy of this expression for the present context in
Eyal’s mind is indicated by the hedge ke'ilu immediately preceding the expression
 Metalanguage in Interaction

(line 102), as well as by his continuing to elaborate10 on what this woman should
look like (lines 106–110). The feeling of inadequacy may also be related to a cer-
tain discomfort Eyal may feel about this sexual and not very politically correct
topic. His point, or the ‘revelation’ he refers to (lines 99–100), is that he discovered
that he is not really attracted to this type of woman any more.

4.2.3 Interpersonal constraints on hedging ke'ilu

Hedging ke'ilu is often employed at ‘uncomfortable’ moments, such as moments of
a certain embarrassment about the topic discussed, or about imposing on the ad-
dressee. In excerpt 6, a conversation between one unmarried couple about another
couple, the woman is described as having asked the man, previously married, to
move in with her:
Excerpt 6 (‘Breaking up’):
45 Amir: .. hi bikshá mimeno,
she asked him,
46 ke'ílu--,
like,
47 .. sheya'avrù.. lagur beyáxad,
that they move in to live together
to move in together,
48 ve--
and
49 .... bekitsúr,
in short,
50 lehitmaséd shuv.
to become established again.
The delicate matter of asking a partner to move in together is hedged by ke'ilu (line
46). Although these words are in the speaker’s voice11, rather than being the con-
structed dialogue (Tannen 1989) of the woman who had made the request, this is

10. This elaboration is also preceded by ke'ilu (line 104), a case of self-rephrasal dealt with in
Section 4.4.
11. The distinction between ‘direct’ and ‘indirect’ speech in Hebrew is quite clear, as the latter
always involves some ‘subordinating’ element such as she-. At the beginning of line 47 we find
this element of ‘subordination’ she-. In addition, the verb of line 47 is in third person: ya'avru
(‘they will move in’).
 Chapter 4.╇ The cognitive realm 

a delicate matter for the speaker, since he is telling this story to his girlfriend, and
the issue of getting married is in the background for this couple as well12.
In the following part of ‘Family Gossip’, we find self-mockery about the speak-
er’s own use of hedging ke'ilu:
Excerpt 7 (‘Family Gossip’):
264 Sara: 'az hi 'oméret,
so she says,
265 .. ma--
what
266 .. bikóret notnì--m,
cricitism one gives,
267 ... kshemáshehu,
when something,
268 ... boné.
is constructive.
269 ... 'aval 'im ben 'adá--m 'eh,
but if a person uh,
270 ... ma'avìr bikóret,
passes criticism
criticizes,
271 .. bebèn zúg,
[one’s] partner,
272 ... she--,
that,
273 ... xayím ke'ilu beyàxad,
live like together
one like lives with,
274 ... veke'ìlu 'ohavím,
and like love
and like loves,

12. The employment of ke'ilu here may also be related to the fact that we are dealing with a case
of reported speech. The speaker is reporting on the speech of someone else, and this is always
only an approximation of what was actually said (see Section 4.5). Furthermore, ke'ilu may be
hedging not only the request itself, but also the verb biksha (‘asked’) introducing it, as the re-
quest was perhaps not made explicitly. However, the interpersonal constraints seem to me more
acute here.
 Metalanguage in Interaction

275 .. ve ke'ìlu ke'ílu,

and ke'ilu ke'ilu,
276 .... 'az,
then,
277 ... má habikoret hazot 'osá,
what the criticism this does
what does this criticism do,
278 .. rak mesaréset,
only castrates,
279 .. ve'od moridá 'et habitaxón ha'atsmì.
and also lowers the confidence the self
and also lowers one’s self-confidence.
The speaker, Sara, is a parent discussing a couple of the younger generation living
together out of wedlock. She is speaking in the voice of the woman, but this voice
coincides with her own thoughts on the matter, as we know from the remainder of
the conversation. The first two ke'ilu tokens (lines 273–274) are hedges of sensitive
or emotional topics: xayim beyaxad (‘living together’, line 273) and 'ohavim
(‘love’, line 274). The second two ke'ilu tokens (line 275), however, are a self-mock-
ery about her own abundant use of ke'ilu’s. Line 273 also provides an example of
hedging ke'ilu occurring at mid-utterance position: xayim ke'ilu beyaxad (‘live
ke'ilu together’).
Hedging ke'ilu is found not only in self-mockery, but also in actual contexts of
parody. I could not find any such instances in the database, but a newspaper re-
view of the televised Miss Israel contest, entitled ke'ilu malkot yofi (‘ke'ilu Miss
Universes’, lit. ‘ke'ilu beauty queens’), supplies a wealth of examples. The journalist
sarcastically describes the 20 participants in the contest – floating in gondolas in a
hotel in Las Vegas, of which one of the floors was reconstructed as Venice – in a
show hosted by Nir Xaxlili, who is known, according to the journalist, as “the Is-
raeli Richard Gere”:
lo be'emet venetsia, ke'ilu. vexaxlili hu ke'ilu richard gir. ve“malkat hayofi” hi ke'ilu
malkat yofi. veha'ish baxalifat hadonald dak bedisnilend hu ke'ilu donald dak. kax
shelo tsarix lehizda'em mitaxarut malkat hayofi. ze misxak beke'ilu. harov yod'im
sheha'ish baxalifat hadonald dak 'eyneno donald dak, veshehabaxura hanirgeshet
'im haketer 'al harosh 'eynena malkat hayofi.
'ela she-20 hamo'amadot hasofiyot lo behexrax yod'ot shehakol ke'ilu. lo yod'ot she-
hen ke'ilu 20 hamo'amadot hasofiyot betaxarut ke'ilu malkat hayofi. ze 'ikron
hake'ilu: kedey lihiyot “ke'ilu” yesh lehityaxes 'elav ke'el “be'emet”.
mila ktana, ke'ilu. yesh lishkol lehosifa lato'ar: “ke'ilu malkat hayofi”.
 Chapter 4.╇ The cognitive realm 

Not really Venice, ke'ilu. And Xaxlili is ke'ilu Richard Gere. And “the beauty
queen” is a ke'ilu beauty queen. And the man in the Donald Duck suit in Disney-
land is ke'ilu Donald Duck. So there’s no need to become enraged about the beau-
ty queen contest. It’s a game of ke'ilu. Most people know that the man in the Don-
ald Duck suit is not Donald Duck, and that the ecstatic young woman with the
crown on her head is not the beauty queen.

But the final 20 candidates don’t necessarily know it’s all ke'ilu. Don’t know that
they are ke'ilu the final 20 candidates in the ke'ilu beauty queen contest. This is the
ke'ilu principle: in order to be a ke'ilu one must relate to it as real.

A tiny word, ke'ilu. One should consider adding it to the title: “the ke'ilu beauty
queen” (Alper 2000, Ha'aretz daily newspaper, translation mine13).

There is an allusion here to a very common activity among young children, pre-
tend play, which in Israel is known as lesaxek beke'ilu (‘playing ke'ilu’)14. This ex-
pression is, of course, directly related to the literal meanings of ke'ilu – the con-
junction ‘as if ’, as well as the hedging meaning ‘close (to reality), but not quite it’.
All hedging ke'ilu tokens in the database hedge the predication of the utter-
ance15 – either the predicate along with some of its complements (e.g., as in ex-
cerpt 4), or the entire clause (e.g., as in excerpt 6). We have seen that the position
of hedging ke'ilu is flexible in relation to the utterance it modifies.

4.3 Ke'ilu as focus marker

In this database, 29.2% of all tokens of ke'ilu function as focus markers. These to-
kens of ke'ilu do not carry any of the meanings specified in the dictionaries for this
utterance (i.e., ‘as though’ or ‘similar to’, ‘supposedly’, or ‘used as a hedge’). The con-
cept of focus marking employed here builds on that developed by Lambrecht: “The
focus is that portion of a proposition which cannot be taken for granted at the time
of speech. It is the UNPREDICTABLE or pragmatically NON-RECOVERABLE
element in an utterance” (1994:€207). Lambrecht’s full definition is as follows:

13. I thank Rivki Ribak for showing me this passage.

14. Henkin (1991) argues that in order to mark their play as ‘pretend play’, Israeli children
employ the past tense in utterances referring to the present. This, she argues, has to do with the
fact that the counterfactual conditional ('i)lu (‘if, irrealis’) must combine with a verb in the past
tense (in adult speech); as, e.g., in Fiddler on the Roof ’s famous song lu hayiti rotshild (‘if I were
a rich man’, lit. ‘if I were Rothschild’). The verb hayiti (‘I were’) is in simple past tense (lit. ‘I
was’). The past tense here represents a modal distance from reality.
15. Most tokens of ke'ilu in the newspaper passage above (Alper 2000) are different from the
tokens appearing in the spoken database, in that they hedge nominal nuclei as opposed to pred-
ications, e.g., ke'ilu malkat hayofi (‘the ke'ilu beauty queen’).
 Metalanguage in Interaction

Focus: The semantic component of a pragmatically structured proposition where-

by the assertion differs from the presupposition. (ibid.: 213).

The terms ‘assertion’ and ‘presupposition’ are defined as follows:

Pragmatic Presupposition: The set of propositions lexicogrammatically evoked in
a sentence which the speaker assumes the hearer already knows or is ready to take
for granted at the time the sentence is uttered.
Pragmatic Assertion: The proposition expressed by a sentence which the hearer is
expected to know or take for granted as a result of hearing the sentence uttered.
(ibid.: 52).

Although focused information is frequently the accented constituent, Lambrecht

argues (ibid.: 214–257) that at the theoretical level, the two concepts of focus and
emphasis must be kept distinct. The means for focus marking may be prosodic (as
in accentuation), morphological, syntactic, or some combination of these.
However, Lambrecht’s criteria of ‘unpredictable’, ‘non-recoverable’, and ‘cannot
be taken for granted’ are somewhat too weak, since they can also apply to utter-
ances which are not in focus. I would add the adjective ‘surprising’ to Lambrecht’s
list of attributes. The focused part, then, is that portion of the utterance which the
speaker finds unpredictable, non-recoverable, and surprising. It is the portion
which cannot be taken for granted that the speaker draws the hearer’s attention to
most.
Fleischman and Yaguello, citing Underhill 1988, mention that like “marks as fo-
cal information whatever occurs to the right16 of it […], which can be a word […], a
phrase […], a sentence constituent […], or an entire utterance” (2004:€131). Focus-
marking ke'ilu, too, is flexible in terms of its position relative to the focused utter-
ance. In 23 out of 35 cases (65.7%), ke'ilu marks as focal information immediately
following it; in the remaining 12 cases (34.3%) it marks as focal information imme-
diately preceding it. In this corpus, it does not occur at mid-utterance position.

16. The written- (and formal syntax-) biased expressions ‘to the right of ’ and ‘to the left of ’ are
of course irrelevant for Hebrew, which is written right-to-left. In any event, I agree with Auer
2000, who concludes that “the supposed parallel between ‘left’ and ‘right’ in syntax [. . . ] is fun-
damentally mistaken when applied to spoken syntax; in speaking, there is no ‘left’ and ‘right’,
but only ‘earlier’ and ‘later’. At least for an approach to syntax which takes the in-time (‘on-line’)
emergence of (particularly) oral language units seriously, what is dealt with first and what is
taken care of later cannot be seen as a decision between two logical equivalents (as between
‘right’ and ‘left’). Rather, it involves one of the most basic and far-reaching decisions a speaker
can make, with all kinds of cognitive, interactional and structural repercussions” (199).
 Chapter 4.╇ The cognitive realm 

4.3.1 Pre-positioned ke'ilu as focus marker

In the present database, initial ke'ilu focuses phrases, verbs along with their com-
plements, and full clauses. In the following excerpt, ke'ilu marks a phrase as focused
information. This comes from a conversation between four religious Israelis in
their early twenties, in which one of them tells an unusual Midrash from the Ge-
mara17 about Jesus:
Excerpt 8 (‘Lions’):
28 Yo'av: .... nixnás yeshu.. lekodesh hakodashìm,
entered Jesus to the holy of holy places
Jesus entered the holy of holies [in the Temple],
{19 intervening intonation units}
47 ... lakáx 'et hashem hameforàsh,
took the name explicit
took the Name of God,
48 .. ra'á 'otò,
saw it,
49 .. he'etìk 'al xtáv.
copied [it] out in writing.
50 ... he'etík.
copied [it] out.
51 ... xatàx 'et ha.. yaréx shelo,
cut the thigh his
cut his thigh,
52 ..... kará,
tore,
53 ... ke'ílu,
like,
54 ... 'im sakín.
with [a] knife.
55 Eran: mhm.
56 Yo'av: hixnis 'et hapètek bifnóxo,
put in the note inside
put the note inside,

17. The Talmud is comprised of the Mishnah, the first major redaction into written form of
Jewish oral traditions, and the Gemara, which contains rabbinical commentaries and analysis of
the Mishnah. Midrash is one method of textual interpretation employed in the Gemara.
 Metalanguage in Interaction

Yo'av relates Jesus’ act of cutting his thigh in order to put the note with the Name
of God inside the thigh (so as not to forget it when he came out of the holy of ho-
lies in the Temple). He refers to this act with two verbs – xatax (‘cut’, line 51) and
kara (‘tore’, line 52). Whereas one can cut a thigh or tear things without any instru-
ment, in order to further focus this unpredictable and surprising information, he
adds the method by which he cut it – with a knife (line 54). This information is
prefaced with ke'ilu. Thus, ke'ilu marks here a phrase as focused information. In-
deed, the audience responds to this further focus with the minimal response ‘mhm’
(line 55). We see that while hedging ke'ilu always hedges the predicate of the utter-
ance (either with or without its complements), focus marking ke'ilu can be found
modifying the verbal complement alone.
In the following excerpt, ke'ilu marks a verb along with its complement as fo-
cused information. Excerpt 9 consists of the orientation of a narrative (Labov
1972) in which Eynat describes to her friend a burglary of an apartment of a mu-
tual acquaintance:
Excerpt 9 (‘Burglary’):
74 Eynat: .... 'az 'e--h,
so uh,
75 ... hig'ìa habáyta,
[she] came home,
76 ... 'axarèy hatsohoráyim,
after noon
in the afternoon,
77 .... 'olá bamadregòt veze,
[she]climbs the stairs and so on (lit. ‘and this’),
78 ... ra'atà-- zug garbáyim sh’la,
[she] saw [a] pair of socks of hers,
79 .. ke'ilu zrukót 'al ha--madregòt.
like thrown on the staircase.
80 Meytal: .. wow!
The focused information here is that the pair of socks the friend saw was thrown
on the staircase (line 79) leading to her apartment. Not only is this not the place
one normally expects to find socks, this was also the friend’s first clue to the fact
that someone had broken into her apartment. Thus, we find a clause constituent
(V+ verbal complement) in focus, preceded by ke'ilu. Again, the audience responds
to this focus (‘wow’, line 80).
 Chapter 4.╇ The cognitive realm 

In excerpt 1 above, ke'ilu focuses a full clause. Earlier, we saw how David de-
scribes himself running in the mud after the tractor. To further enhance his story,
he adds the unpredictable information of his feet getting stuck in the mud (lines
52–53), along with the sound they make in this state (line 54). This is focused in-
formation, an entire clause preceded by ke'ilu (lines 52–53).
Fleischman and Yaguello mention that the scope of like is often ambiguous,
being marked in most cases only on the left of a focused segment (2004:€132). One
solution to the problem of ambiguity that they mention is repeating like at the end
of the focused segment. Interestingly, in excerpt 1, we find the other element, kaze,
at the end of the focused segment (line 53). However, kaze has additional func-
tions in this context (see Maschler 2001).

4.3.2 Post-positioned ke'ilu as focus marker

There are 12 cases of ke'ilu following the constituent it focuses. In the following
excerpt, ke'ilu focuses a phrase immediately preceding it. This comes from the
‘Burglary’ conversation, towards the end of the narrative. Eynat says that the cur-
rent mode of operation of thieves is that they don’t even bother to bring their own
gloves when breaking into an apartment. They simply use socks they find in the
apartment, and the only thing they touch with their bare hands – the lock of the
front door – they take with them:
Excerpt 10 (‘Burglary’):
182 Eynat: mistabèr shehayó--m,
it turns out that today,
183 .. hastíl18 hu,
the deal is,
184 .... hem lokxìm garbáyim,
they take socks,
185 .. shel ha--nignáv ke'ìlu,
belonging to the [person who was] robbed like,
186 .. hem /'afilu/ lo megi'ìm 'im ze mehabáyit.
they /even/ don’t arrive with it from home
they don’t even bother to bring it from home.
The focused information here is that the socks used by the thieves belong to the
person whose apartment is robbed (and not, as one may predict, to the thief). This
information is focused via ke'ilu following the focused phrase shel hanignav (lit. ‘of

18. The slang expression stil was explained to me by the participant who recorded the interac-
tion as a blending of the Englishisms ‘style’ and ‘deal’.
 Metalanguage in Interaction

the stolee’, ‘of the robbed’) (line 185). However, this is a borderline case, since here
ke'ilu can also be functioning as a hedge. The passive participle nignav (‘the stolee’)
is not commonly used as an animate noun, to refer to the person who was robbed19.
The speaker may be hedging it with the thought that she perceives some inadequa-
cy in this word (see discussion of hedges, Section 4.2 above).
It is these ambiguous cases that constitute transitional examples for the shift
from the hedging meaning to the focus meaning of ke'ilu. The functional itinerary
here is fairly straightforward, since hedged material is also what is in focus. In
other words, the fact that the expression is not a precise ‘fit’ for the present context,
too, is judged by the speaker ‘unpredictable’ or ‘non-recoverable’ for the hearer at
the time of utterance.

4.3.3 Post-positioned focus-marking ke'ilu in aposiopesis

In the following excerpt, Anat describes to Ran the first dive she went on. She was
taken on this dive alone by the diving instructor of a course that she, Orna (also
present in the interaction), and a few other girlfriends took, in order to convince
her not to leave the course on the first day:
Excerpt 11 (‘Diving Course’):
91 Anat: ... sha'á hayinu mitàxat lamàyim.
an hour we were under (the) water.
92 Ran: .. sha'á?
an hour?
93 Anat: ... vehaya.. madhí--m!
{----------pp---------}
and it was amazing!
94 Orna: sha'à ze harbé--.
an hour is a lot.
95 .. vetaxshóv,
and think (masc. sg.) [about it],
96 Anat: .. haya madhím.
it was amazing.
97 Orna: .. leyòm rishón.
for a day first
for a first day.
98 Ran: mhm.
99 Anat: .. ve.. le--míshehi,
and for someone,

19. The participle nignav (from the root √g.n.v. in the nif 'al (passive) pattern) would generally
refer to the property stolen.
 Chapter 4.╇ The cognitive realm 

100 .. she'áf pa'am lò haytà,

who never was
who had never been,
101 mitàxat lamáyim,
under the water
under water,
102 .. ke'ilu--,
I mean,
103 ..... ve--,
and,
104 .... vehu 'asa
and he did
105 hu 'amár li,
he said to me
106 .. lifnéy ze,
beforehand,
107 ta'así ma she'áni 'ose bamàyim.
do what [ever] I do in the water.
Anat describes their being under water for a whole hour (line 91) and evaluates
how amazing it was (madhim, line 93). Orna, who also participated in this diving
course, further points out to the main addressee of this story, Ran, that an hour is
a lot, particularly for a first day [of diving] (lines 94–95, 97). Overlapping Orna
(and therefore also probably not completely aware of Orna’s utterance), Anat, to
further involve her audience, repeats the predicate madhim (‘amazing’) at line 96,
and adds the information and for someone who had never been under water (lines
99–101) to focus on how unusual this experience is. This information can be
viewed as ‘non-recoverable’ at the time of utterance in a very local sense. Anat’s
story up to this point has been very long and involved. It began over ten minutes
beforehand and was punctuated by many digressions. This may be contributing to
her assumption (as well as to Orna’s) that at this point in the discourse, Ran may
not be thinking about this being Anat’s first dive. In other words, the speaker can
assume the hearer is not ready to take this information for granted at the time the
sentence is uttered. Following this focused information we find ke'ilu20.
Here ke'ilu focuses a sentence constituent, the noun mishehi, (‘someone’, fem.
sg.) along with a relative clause modifying it she'af pa'am lo hayta mitaxat lamayim
(‘who had never been under water’). Beginning with the preposition le- (‘for’), we
see that this entire sentence constituent functions as a complement of a predicate

20. I translate this token with English I mean, which seems to better capture the function of
ke'ilu here.
 Metalanguage in Interaction

in a sentence that was never completed. Furthermore, ke'ilu is uttered in continu-

ing intonation, but the speaker, following a longer than average pause (line 102),
never completes it and moves on to a different matter – what the guide had told
her before the dive (lines 103–107). This is a case of the trope, or figure of speech,
rhetoricians term ‘aposiopesis’ (Quinn 1982); its function is “suddenly discontinu-
ing speech as if one is unable or unwilling to continue (for example, rendered
speechless by emotion)” (Tannen 1989:€25). Post-positioned ke'ilu is often found
in such contexts in the interactions.
This particular context fits well with Hirschfeld’s description concerning the
place of ke'ilu in the sentence:
[... ] the main thing is the gaps in the sentence: ke'ilu requires a continuation. ke'ilu
what? And here the ke'ilu appears by itself, as if it were something in and of itself
and not a component, a scaffolding for another floor in the sentence. (1990:€6,
translation mine).

4.3.4 Interpersonal constraints on focus-marking ke'ilu

The question arises as to why a speaker would point out to the addressee something
which is unpredictable or non-recoverable at the time of utterance. In his study of
narrative discourse, Chafe writes that “a narrative that fails to conflict with expecta-
tions is no narrative at all” (1994:€122). In order to ward off the threatening ‘so
what?’ question (Labov 1972), people usually tell narratives about topics they judge
their interlocutors will find interesting in some way. “The topics worth verbalizing
are those that have a point, which is usually to say that they conflict with expecta-
tions” (Chafe 1994:€122). Thus, one reason to point out unpredictable information
is precisely in order to involve the audience. Indeed, in all the focus marking exam-
ples involving ke'ilu, the motivation to put things in focus in the discourse has to
do with interpersonal involvement. It stems either from the speaker’s need further
to involve the audience (as in all the examples of Section 4.3 above), or, less fre-
quently, from the speaker’s own involvement with the text. Indeed, in the former
case, the audience very often responds to this involvement creating strategy with
an utterance showing involvement, such as mhm (excerpt 8, line 55), wow (excerpt
9, line 80), or a longer utterance such as madhim (‘amazing’).
The following example from a conversation between two young women, best
friends discussing their intimate relationships, illustrates the speaker’s own in-
volvement with the text. Merav tells Tali that there are times when she clearly feels
that she loves the man she is involved with, but that she has difficulty telling him
so explicitly:
 Chapter 4.╇ The cognitive realm 

Excerpt 12 (‘Best Friends’):

85 Merav: ... vekén yesh pe'amìm,
and yes there are times,
and there are times,
86 .. she'ani--,
that I,
87 ... sheze--
that this
88 ... ke'ílu,
like,
89 'ani kol kbx margishà 'et ze,
I so much feel it,
I feel it so deeply,
90 .. shehu 'e
that he uh
91 .. hu yodéa,
he knows,
92 ... vehu shoméa,
and he hears,
This token of ke'ilu focuses information that the speaker clearly feels very emotional
about: 'ani kol kbx margisha 'et ze (‘I feel it so deeply’, line 89). The fumbling at line
87 and the expressive prosody of these lines further contribute to the interpretation
that she was deeply involved in her utterance, to the point of losing her train of
thought. Thus, the motivation for focus-marking here seems related more to the
speaker’s involvement with the text than to her need to involve the audience.
We have seen, then, that both the focus-marking and the hedging functions of
ke'ilu are often associated with the evaluative aspect of discourse (Labov 1972,
Tannen 1989). They are frequently employed around topics that are sensitive or
emotional for the speaker and/or addressee. These empirical findings fit well with
Hirschfeld’s cultural-philosophical analysis:
How would a real boy tell a real girl something that really goes from the heart and
to the heart despite the fact he is not a poet? He will lead her to the edge of lan-
guage, to the place from which one sees what one doesn’t see from the center of
language. The kaze and the ke'ilu point to whatever is beyond what was said, to the
non-language. [... ]
kaze and ke'ilu have to do with emotion and they proliferate with the increase of
feelings. There are moments in which there is perhaps only kaze and ke'ilu – and
this is where language goes completely bankrupt. [... ]
 Metalanguage in Interaction

If one listens carefully, one hears that the kaze and the ke'ilu are in fact the expres-
sive parts of the sentence, and not just a verbal connecting tissue. (1990:€ 8–9,
translation mine).

However, excerpt 12 involves also self-rephrasal, as Merav switches to a new direc-

tion and to a new grammatical construction, not continuing the ‘subordinate’
clause she began at line 86 or 87 with the complementizer she- (‘that’). Being am-
biguous, this token of ke'ilu therefore constitutes another transitional example,
this time for the shift between the focus-marking use and the self-rephrasal one.
Again, the functional itinerary is rather straightforward, since what is self-re-
phrased in the discourse is also what is in focus. In other words, it is precisely the
reformulation of the utterance that the speaker judges ‘unpredictable’ or ‘non-re-
coverable’ at the time of verbalization. This leads us to the next function of ke'ilu.

4.4 Ke'ilu as discourse marker of self-rephrasal

The majority of ke'ilu tokens throughout the data (48.3%) function as discourse
markers opening self-rephrasals21. Fleischman and Yaguello, too, mention that
English like often connects two utterances the second of which “provides an exam-
ple, explanation, justification, or elaboration of the first” (2004:€134). For them,
this is a subcategory of the hedging function of like. I consider these two cases
separately because of the different levels of discourse at which they operate – the
lingual (hedging) and the metalingual (self-rephrasing).
As mentioned earlier, previous studies of discourse markers in Hebrew talk-
in-interaction (Maschler 1998b, 2002a) have shown that intonation-unit-initial
ke'ilu often functions metalingually as a discourse marker of ‘realizing the need to
rephrase’. We have seen one example in excerpt 1, line 48.
This use of ke'ilu is different from previous uses we have seen so far because it
is metalingual. Rather than referring to something in the extralingual world, such
as some hypothetical condition, some unpredictable and surprising information,
or the inadequacy of the words chosen to describe some state of affairs, metalin-
gual uses of utterances refer to the text, to the interaction between its participants,
and/or to the cognitive processes taking place in their minds during verbalization.
Ke'ilu of excerpt 1, line 48 is a cognitive discourse marker because it refers to the
speaker’s realization that his previous utterance should for some reason be re-
phrased. This reason, too, often has something to do with the need to involve the
addressee.

21. For another look at this discourse marker, relating to the ways it fits in the system of dis-
course markers negotiating frame shifts in Hebrew talk-in-interaction, see Chapter 1, Section 7
and Maschler 2002a.
 Chapter 4.╇ The cognitive realm 

There are also structural arguments for considering this use of ke'ilu a dis-
course marker. In excerpt 1, line 48, ke'ilu appears following sentence final intona-
tion in the immediately preceding intonation unit (line 47), as do 94% of the dis-
course markers in the database22. However, the discourse marker ke'ilu is the most
frequently employed discourse marker in the remaining 6% category of discourse
markers which follow continuing intonation in the immediately preceding intona-
tion unit. Of the 58 tokens of the discourse marker ke'ilu employed in the present
database (50 interactions), 44 (75.9%) occur following non-continuing intonation
in the immediately preceding intonation unit (as in excerpt 1), whereas 13 (22.4%)
occur following continuing intonation23.
For example, earlier on in the orientation to this story, David tells about the
place and participating characters:
Excerpt 13 (‘Tractor Chase’):
11 David: haxokér hapratí 'amar,
the investigator private said
the private investigator said,
12 shehu nimtsa be'eyze makóm,
that he is located in some place
that he [the defaulting purchaser] can be found in some place,
13 ... géshem zal'afót baxuts,
rain raging outside
it’s raining cats and dogs outside,
14 ke'ílu,
I mean,
15 ... má ze gèshem,
what is rain
what [do I mean by] rain,
16 ... shexavál 'al hazmàn.
that it’s a waste of time.
unimaginable rain.

22. This is the database on which Maschler 2002a is based, comprising 16 out of the 50 conver-
sations on which the present study is based, in other words, 40 minutes of naturally-occurring
casual conversation among 43 different speakers, collected throughout the years 1994–1997 (see
Chapter 1, sections 5, 7).
23. One case of ke'ilu throughout the database appears at intonation unit final position. Since
it functions metalingually, I counted it as a discourse marker, even though it does not appear at
intonation-unit initial position.
 Metalanguage in Interaction

David mentions that it was raining very heavily outside (line 13). He elaborates: ma
ze geshem (‘what [do I mean by] rain’), shexaval 'al hazman (‘unimaginable rain24’).
Preceding this self-rephrasal is the discourse marker ke'ilu (line 14), following con-
tinuing intonation in the immediately preceding intonation unit (line 13).
Chafe (1994) ties prosodic phenomena such as types of intonation contour to
cognitive processes taking place during verbalization. For instance, he notes that
once speakers judge that the scanning of a center of interest (such as a sub-episode
in a story-episode) has been completed, they express that judgment with a sen-
tence-final intonation contour. The realization that one’s utterance is in need of
rephrasal is, of course, independent of one’s decision concerning whether or not a
center of interest has been completed. Therefore, we find both continuing and
non-continuing intonation preceding self-rephrasals that begin with ke'ilu25.

4.4.1 Ke'ilu in self-rephrasals of clarification sequences

Sometimes the need for rephrasal is not first realized by the speaker but instead
originates from the addressee. A common case of the discourse marker ke'ilu fol-
lowing continuing intonation occurs in sequences of request for clarification,
schematized as follows:
1. A: utterance
2. B: request for clarification
3. A: repetition of problematic utterance (in continuing intonation contour)
4. ke'ilu
5. rephrasal of problematic utterance
For example, in the same conversation, David tells how he pulled ‘it26’ out of the
mud, and the hearer requests clarification concerning the identity of ‘it’:

24. The expression xaval 'al hazman literally means ‘it’s a pity on the time’, or ‘it’s a waste of
time’. Originally it was used in the more literal meaning of ‘not worth wasting time on / paying
attention to’. Over the last fifteen years or so, the expression has gained a slang meaning, and is
used also to intensify the property attributed (to take the present context, ‘it was raining so hard
that it’s a waste of time to waste words describing it’, whence the translation ‘unimaginable rain’).
In this new use, and by influence of the many acronyms originating in Israeli army slang, the
expression has also become an acronym in casual spoken Hebrew: xavlaz (acronym of sorts of
xaval 'al hazman, taking the first two consonants, /x/ and /v/, from the first word xaval (‘pity’),
the middle consonant /l/ either from the preposition 'al or from the third consonant of xaval,
and the final consonant /z/ from the word zman (‘time’)).
25. For an elaboration of this, see Maschler 2002a: 5–8.
26. Since there is no neuter in Hebrew, there is no way of knowing whether this ‘it’ ('oto, ex-
cerpt 14, line 92) refers to an animate or inanimate entity (whence the ambiguity in the transla-
tion: ‘him/it’). In other words, David may be referring here to pulling out the tractor, or the
 Chapter 4.╇ The cognitive realm 

Excerpt 14 (‘Tractor Chase’):

91 David: ... higía hagràr.
arrived the towing truck
the towing truck arrived.
92 ... shaláf 'oto,
pulled out him/it
pulled him/it out,
93 'axarey 'eyze.. sha'á shel 'avoda sham,
after about [an] hour of work there
94 betox habóts,
in the mud,
95 .. 'et ha'emét?
the truth?
96 Dalia: má,
what,
97 shalàf má?
pulled out what?
98 David: .. shaláf ’to,
pulled out him/it
pulled him/it out,
99 ke'ílu,
like,
100 ... garár 'oto haxútsa.
towed him/it out [of the mud].
Following Dalia’s request for clarification (lines 96–97), David repeats the prob-
lematic utterance shalaf ’to27 (‘pulled him/it out’, line 98) in continuing intonation,
and then rephrases it: garar 'oto haxutsa (‘towed him/it out’, line 100). This self-
rephrasal is preceded by ke'ilu. David still does not specify in his clarification (lines
98–100) what 'oto (‘him/it’) refers to, but the verb garar (‘towed’) seems sufficient
to disambiguate the reference for Dalia, since it is usually tractors that are towed,
not people.

owner of the tractor (who was perhaps sitting inside the tractor). This is the source of the re-
quest for clarification in line 96.
27. In fast speech, the initial glottal stop + vowel 'o of 'oto are often dropped.
 Metalanguage in Interaction

In a related use of ke'ilu, A’s response to the request for clarification proceeds
directly to ke'ilu + self-rephrasal, skipping stage 3 in the scheme above, the repeti-
tion of the problematic utterance, as in the following excerpt:
Excerpt 15 (‘Sex Shop’):
50 Yosi: ... 'az hi 'oméret lah--ì,
so she says to the other one,
51 .. hù mishelánu,
he’s of ours
he’s one of our people,
52 .. ze beséder,
it’s okay,
53 .. 'efshár lesapèr lo.
{-------laughing---------}
it’s okay to tell him.
54 {laughter}
55 Tali: ... má ze shelànu?
what is ‘ours’?
56 Yosi: .... ke'ìlu 'ani mishelahém,
like I’m one of theirs,
like I’m one of their people,
57 .. she'efshár le--
that it’s okay to
58 Tali: /nu/.
{laughing}
/go on/.
59 Yosi: ... 'at ló makira 'et habitùy ba'àrets,
you don’t know the expression in the country,
don’t you know the Israeli expression,
60 .. hu mishelánu?
‘he’s one of ours’?
Tali requests clarification of the utterance hù mishelánu (‘he’s one of ours’, line 51).
Yosi’s self-rephrasal here occurs immediately following the request for clarification
ma ze shelanu? (‘what is “ours”?’, line 55), without repetition of the problematic ut-
terance (i.e., skipping stage 3 of the scheme above). It is preceded by ke'ilu (line 56).
The function of ‘realizing the need to rephrase’ is performed not only by the
discourse marker ke'ilu, but also by several other overlapping discourse markers
 Chapter 4.╇ The cognitive realm 

throughout the database, such as s’tomeret (lit., ‘this means’), klomar (lit. ‘that is to
say’), ya'ani and ya'anu (these last two forms from Arabic28 ‘it means’). However,
ke'ilu is by far the most common discourse marker employed for this purpose.

4.4.2 Self-rephrasal ke'ilu tokens as fillers

As one realizes the need to rephrase, one is often also groping for the right word.
It is in this sense that ke'ilu can also function as a ‘filler’ (as we find it in The Slang
Lexicon (1993): milat miluy reyka: mila hameshurbevet lelo siba lamishpat (‘a void
filler which finds its way into the sentence for no reason’), see also Ben Shachar
2000). This ‘filler’ function of ke'ilu is illustrated in the following example, which
will lead us to a very recent development in the employment of ke'ilu in Hebrew
talk-in-interaction. In the continuation of excerpt 1, David says:
Excerpt 16 (‘Tractor Chase’):
55 David: ... bekitsúr vela'inyá--n,
in short and to the point,
in short,
56 ... 'axaré--y,
after,
57 .. mirdáf she--,
a chase that,
58 ne'eràx kemispàr dakó--t,
lasted roughly a few minutes,
59 .... hitslàxti lehagia 'ad latráktor,
I managed to get all the way to the tractor,
60 .. lisgór lo ’ta
to close for him the
to close his
61 ke'ìlu lexabót ta’manòa,
like to turn off the engine
like to turn off his engine,
62 lakáxti ta’maftèax.
I took the key.

28. For the phenomenon of language alternation at discourse markers leading to borrowing
phenomena, see Goss and Salmons 2000, Maschler 2000a,c, and Chapter 1, Section 2.
 Metalanguage in Interaction

63 .. 'ani omér lo,

I say to him,
64 Dalia: 'í 'efshar leta'èr.
it’s not possible to imagine
one cannot imagine.
65 David: ke'ílu,
like,
66 .. 'ani omér lo,
{----laughing----}
I say to him,
67 .. léx le'azazèl,
go to hell,
68 .. tamút.
die
I wish you dead.
In line 60, David begins with the verb lisgor (‘to close’), but then self-corrects to
lexabot (‘to turn off ’), a more appropriate verb in the context of switching off the
engine of a tractor. This is a straightforward case of the discourse marker ke'ilu
(line 61) beginning self-rephrasal. The more interesting case for our purpose here,
however, starts at line 63, when David begins to tell what, upon finally getting hold
of the tractor, he said to the owner who had tried to escape: 'ani omer lo (‘I say to
him’). He is overlapped by Dalia (line 64) still showing involvement in the previ-
ous episode: 'i 'efshar leta'er (‘one cannot imagine’, or ‘incredible’). David then re-
peats his previous utterance – 'ani omer lo, ‘I say to him’ (line 66) – but not before
he precedes his repetition with ke'ilu (line 65). Presumably, Dalia’s overlap dis-
tracted him for a moment, and ke'ilu is employed here as a filler, while he takes a
split second to return to his line of thought. This brings us to the final use of ke'ilu
found in the interactions – ke'ilu in quotations.

4.5 Ke'ilu in quotations

I found only four instances of ke'ilu in quotations (3.3% of all tokens in the data-
base). This is a relatively new use of ke'ilu, as indicated also by the fact that all four
tokens come from two interactions recorded in 199929, in spite of the fact that the
database abounds in instances of constructed dialogue (Maschler 2002a).

29. Conversations collected later on in the process of corpus construction and not included in
the present study indeed show more cases of quotative ke'ilu.
 Chapter 4.╇ The cognitive realm 

In the following excerpt, Uriel tells three women about an incident in which
he had told a friend that he was feeling somewhat depressed. A few hours later, this
friend’s friend, Orit (whom he barely knows), called him up to ask him how he was
feeling. Uriel was amazed that such an incident could happen in today’s alienated
age. We find a token of ke'ilu introducing constructed dialogue, as well as the over-
lapping marker ya'anu (from Arabic ‘I mean’):
Excerpt 17 (‘Intruding One’s Privacy’):
172 Uriel: ... 'aval 'amárti ke'ilu,
but I said like,
173 ... ba'idán,
in the age,
174 ... ba'idàn hamodérni ya'anu.
in the age the modern like.
in the modern age like.
175 ... ha.. ha.. ha'adam haxàd meymadí shel merkúza,
the.. the.. the individual the one dimensional of Marcuse
the.. the.. the One Dimensional Man of Marcuse,
176 Dalit: ze ki orìt xamudá.
that’s cause Orit is nice.
177 Uriel: ... 'az 'e--h,
so uh,
178 ... ze kéta30,
it’s a segment
it’s a thing,
179 ... pashút 'e--h,
simply uh,
180 ... madlík.
cool
[it’s just a] cool [thing].
The statement of lines 173–175 lacks an ending. The implication is that in the al-
ienated age described in Marcuse’s 1968 book One Dimensional Man (line 175),
one wouldn’t expect a phone call from a person one barely knows, inquiring about
how one is feeling. In line 172, we find ke'ilu following the verbum dicendi

30. For the use of the word keta in Hebrew talk-in-interaction, see Maschler 1998b and Chap-
ter 6, Section 1.2.
 Metalanguage in Interaction

'amarti (‘I said’), immediately preceding constructed dialogue31 of Uriel to himself

in the story world. This is similar to the quotative function found for English like.
The connection between ke'ilu as a hedge and ke'ilu in quotes is straightforward:
in the strategy of constructed dialogue, the actual dialogue is seldom repeated
verbatim; it is always an approximation of what was actually said, always a hedge.
Notice the ‘equivalent’ ya'anu (‘like’) in Uriel’s constructed dialogue to himself
(line 174). Here ya'anu hedges the expression ba'idan hamoderni (‘in the modern
age’), which the speaker seems to judge not completely fit for the present context,
as his further elaboration of the expression indicates: ha'adam haxad meymadi
shel merkuza (‘Marcuse’s One Dimensional Man’, line 175); in contrast, ke'ilu (line
172) is employed to hedge the entire constructed dialogue32. Another possible rea-
son for Uriel’s judgment concerning the unsuitability of this expression for the
present context has to do with the ‘academic’ register to which the words ba'idan
hamoderni (‘in the modern age’) belong. This is Uriel’s ironic comment on his
employing such an academic register in this highly casual conversation among
friends33.
Nevertheless, ke'ilu as a quotative is different from English like in this func-
tion, in that, at least in the present database, it does not occur unaccompanied by
a verbum dicendi, as often found for the English ‘equivalent’34:
(30) And I’M LIKE: “What the hell’s going on here?”
(Fleischman and Yaguello 2004:€135).

4.5.1 Ke'ilu in double-voiced ironic quotations

Fleischman and Yaguello mention a related use of English like and French genre,
the double voiced ironic quotation (1999, 2004): “like and genre enable dual-voiced
utterances in which narrators can use the direct quote form to report thoughts/
attitudes they attribute to participants in their narratives while at the same time
superimposing onto those internal quotations their own evaluative judgments”
(2004:€137–138). Note Haiman’s example:
Yeah, sure, LIKE I haven’t heard that one before (Haiman 1998:€53).

31. The lack of a ‘subordinator’ such as she- (‘that’) is evidence that this is ‘direct speech’.
32. However, in other parts of this text, ya'anu functions as an equivalent of ke'ilu not only in
the hedging function, but also in the self-rephrasal and quotative functions.
33. I thank Tamar Katriel for suggesting this point to me.
34. However, the second utterance involving comparative meaning, kaze (‘like’, lit. ‘like this’),
does appear in the present database as a quotative which is often unaccompanied by a verbum
dicendi. See Maschler 2001.
 Chapter 4.╇ The cognitive realm 

This use was not found in the present database, but in our household, an interac-
tion between an eleven-year-old and her parents provides a lively example. Our
eldest daughter was asked to do her part in helping with the preparations for din-
ner and the bedtime routine. When she objected, she was told by her father that
there would be no dessert unless she kept her part of the deal. Storming out of the
room on her way to perform the task, she shouted bitingly35:
Excerpt 18 (‘Sweets’):
1 Shira: hu me'ayém 'alay,
he’s thereatening me,
2 she'ani ló 'akabel dessèrt.
that I won’t get [any] dessert.
3 ke'ilu 'eize mamtakím,
like what sweets,
4 yesh babáyit haze!
there are in the house this
there are in this house!
This young teenager is ridiculing her father for the admittedly poor strategy of
threatening to withhold sweets (lines 1–2), and she is attributing to him the
thought that there is a wide array of tasty sweets around the house (lines 3–4).
Prefacing this attribution with ke'ilu and employing very marked prosody (cf.
Günthner 1999a), at the same time she expresses her own evaluative judgment
about the quality of these sweets.
A special case of the double-voiced ironic quotative ke'ilu can be found in a
relatively new idiom, ke'ilu da?. Livneh (2002) cites this idiom in a slang dictionary
she published in a special Independence Day supplement of the Ha'aretz daily
newspaper. One of the then latest Hebrew slang expressions was da?, very much
related to the American slang expression duh?, and interpreted by Livneh as ‘bez-
ilzul, ma 'ata 'omer? (‘in contempt, you don’t say!’). Livneh supplies the following
example:
shokolad ze mashmin hu 'omer li. ke'ilu da? (‘chocolate is fattening, he tells
me. Like duh?’).
In this constructed example, the speaker verbalizes self-evident constructed dia-
logue (Tannen 1989) concerning chocolate: shokolad ze mashmin (‘chocolate is
fattening’). As we know from studies of polyphony in discourse (Bakhtin 1981,

35. I transcribed this excerpt immediately after it happened. However, as I do not have a re-
cording of it, I was not able to obtain the minute details of transcription, such as information
about pauses between intonation units.
 Metalanguage in Interaction

Günthner 1999a), in the strategy of constructed dialogue the speaker often super-
imposes his or her own stance towards the speech constructed. This is generally
done via paralinguistics or kinesics (as seen in the preceding excerpt). In the
present case, the speaker superimposes a contemptuous stance on the constructed
dialogue (i.e., ‘what is he telling me the obvious for?’). Just in case the recipient did
not pick up the contemptuous key, the speaker adds da?, lexically elaborating the
stance expressed earlier via paralinguistics and kinesics. Quotative ke'ilu is thus
employed here to introduce a speaker’s self-quote concerning the previous utter-
ance. It connects the previous utterance about chocolate (including the prosodi-
cally superimposed stance of the speaker toward this utterance) with a self-quote
(da?) which elaborates the speaker’s stance towards this utterance (‘you don’t say!’)
and spells it out more explicitly.

5. The functional itinerary of ke'ilu

We have seen, then, five different uses of ke'ilu. Three of these uses are more di-
rectly related to its literal meaning: the conjunction ‘as though’, hedging, and quo-
tation; two are more distant: focus-marking and self-rephrasal. I suggest two func-
tional itineraries: one in the lingual realm, originates more from the comparative
element in ke'ilu; the other, in the metalingual realm, originates more from its ir-
realis conditional component. Of course, without a corpus providing diachronic
perspective, the following comments concerning the path of development of ke'ilu
can only be hypothesized.
In the lingual realm, we start out with the lexical element of comparison and
approximation k(e)-. From the lexical meaning of this preposition to the hedging
use of ke'ilu, the itinerary is rather direct: what is intended is ‘like but not quite
what was said’. This fact often requires elaboration (‘how is what was said different
from what was intended?’), whence the self-rephrasal function of ke'ilu and the
common juxtaposition of ke'ilu in these two functions, as seen, for example, in
excerpt 5, lines 102–104, and in excerpt 19, a continuation of excerpt 9:
Excerpt 19 (‘Burglary’):
74 Eynat: .... 'az 'e--h,
so uh,
75 ... higì'a habáyta,
[she] came home,
76 ... 'axarèy hatsohoráyim,
after noon
in the afternoon,
 Chapter 4.╇ The cognitive realm 

77 .... 'olá bamadregòt veze,

[she]climbs the stairs and so on (lit. ‘and this’),
78 ... ra'atà-- zug garbáyim sh’la,
[she] saw [a] pair of socks of hers,
79 .. ke'ilu zrukòt 'al ha--madregót.
sort of like thrown on the staircase.
80 Meytal: .. wów!
81 Eynat: .. yeshanót ke'ilu,
old like,
82 ... ke'ilu,
like,
83 .. shehi mizmàn kvar lo halxá 'itam,
that she for a long time already didn’t go with them
that she hadn’t worn for a while,
84 .. spórt ka'ele.
sport like these
sporty sort of.
In line 79 we see the focus-marking ke'ilu with which Eynat describes the socks
her friend found thrown on the staircase. Eynat further describes them as yesh-
anot ke'ilu (‘old sort of ’, line 81). This is hedging ke'ilu. Immediately following this
hedge, Eynat adds a further elaboration on the socks: shehi mizman kvar lo halxa
'itam36 (‘that she hadn’t walked with for a while’, line 83). In other words, the socks
are ‘old’ not in the usual sense, but in the sense that they hadn’t been used for a
while. This elaboration is preceded by the discourse marker of self-rephrasal ke'ilu,
line 82. Thus we see how a hedge is naturally followed by a case of elaboration and
self-rephrasal. Furthermore, excerpt 19 provides us with examples of three of the
functions of ke'ilu – focus-marking (line 79), hedge (line 81), and self-rephrasal
(line 82) – in close proximity to each other.
When a speaker is constructing someone else’s speech, the report is never ver-
batim but always an approximation, whence ke'ilu in quotations: ‘like what was
actually said, but not quite’. Thus we see the contribution of the lexical element of
similarity and approximation k(e)- to the discourse functions of ke'ilu. Of course,

36. Here, as in many other spoken contexts, agreement of fem. pl. is no longer maintained in
the discourse and we find the masculine singular resumptive pronoun 'itam (‘with them’, masc.),
rather than feminine 'itan, referring to what the speaker perceives of as a feminine plural entity
garbayim yeshanot (‘old (fem. pl.) socks’). For studies relating to resumptive pronouns in spoken
Hebrew discourse, see Ariel 1999 and Maschler and Shaer (forthcoming).
 Metalanguage in Interaction

the hypothetical conditional component 'ilu also adds to the hedging-based func-
tions: a condition, and a hypothetical condition in particular, is at a modal dis-
tance from the reality described (Fleischman 1989:€2).
We have seen that a self-rephrased utterance (excerpt 12) and a hedged utter-
ance (excerpt 10) are generally also what is in focus. However, a more direct route
to focus marking was studied by Haiman, who noted a universal link between
comparison and focus: “[a] comparative construction is one which contrasts, and
hence, focuses the elements which are compared;... the element compared... [be-
ing] more highlighted” (Haiman 1988:€ 310, cited in Fleischman and Yaguello
2004:€ 140). From the lexical component of comparison k(e)-, then, the focus-
marking function can also be derived, though less directly. Fleischman and
Yaguello note also that a direct quote is always necessarily the focus of the utter-
ance in which it occurs. “The QUOTATIVE function [...] thus represents a natural
extension of an item already operating as a FOCUS marker” (2004:€141).
In the metalingual realm, we relate particularly to the lexical element of the
irrealis conditional 'ilu (‘as if ’, ‘as though’) with which ke'ilu is composed. Em-
ployed in the literal sense, this conjunction would require a full clause following it.
A paraphrase of the self-rephrasal function of ke'ilu would be something like ke'ilu
'amarti ‘as if I were saying’ or ke'ilu lomar ‘as if to say’. In other words, when a
speaker rephrases an utterance, he or she prefaces it with ke'ilu which may ‘stand
for’ (in a way that is no longer apparent to the speaker) a longer metalingual utter-
ance such as ‘as if I were saying’ or ‘as if (I were) to say’. This functions metalin-
gually, in the realm of the text, the interaction, and the cognitive processes in the
mind of the speaker, as opposed to referring to the extralingual world.
The most common function of ke'ilu is that of self-rephrasal (48.3% of all to-
kens). Fleischman and Yaguello’s study is not quantitative, but it seems to me that
English like is much less common in self-rephrasals than ke'ilu. We have seen that
the function of self-rephrasal is arrived at by three paths (paths 1, 4, and 6 schema-
tized in Figure 2), both in the metalingual and lingual realms. We can hypothesize
the development of ke'ilu’s pragmatic functions as follows:

lexical meaning lingual realm metalingual realm

comparison and 1) hedge self-rephrasal

approximation 2) hedge quote
3) focus hedge
4) focus self-rephrasal
5) focus quote
6) irrealis self-rephrasal
Figure 2.╇ Hypothesized development of the pragmatic functions of ke'ilu
 Chapter 4.╇ The cognitive realm 

Comparing this figure with Fleischman and Yaguello’s itineraries for like and genre
(1999) reproduced earlier in this chapter as Figure 1 (Section 3), we see that al-
though the number of realms involved differs, there are many similarities between
the two figures. The present study provides additional evidence for the functional
itinerary paths, suggested with question marks in Figure 1, between the focus
marking, quotative, and hedging functions.

6. Grammaticization of ke'ilu in cross-linguistic perspective

We return, finally, to Traugott’s question concerning cross-language generalizations

to be made about the development of discourse particles in terms of both their like-
ly semantic sources and their semantic-pragmatic paths. In the conclusion of their
study of English like and French genre, Fleischman and Yaguello write:
The interest of our data for cross-language pragmatics lies in presenting a case-
study in which similarly functioning discourse markers in two languages have
evolved, independently, from similar lexical sources and have followed relatively
parallel paths of development, particularly once they came to function as prag-
matic operators. If additional cases of the same parallelisms can be found, these
might be used to identify lexical sources and pathways of grammaticalization for
discourse markers in the same way as we have identified lexical sources and path-
ways of grammaticalization for many grammatical categories. We might then be
able to predict that, e.g., expressions of comparison/approximation may give rise
to focus markers, hedges, and/or quotatives. (2004:€142).

The data presented in this study, from a language unrelated to English or French,
provides further evidence for this prediction. Close analysis of Hebrew ke'ilu to-
kens in context further supports the affinity between the discourse functions of
hedging, focus marking, and quotation across languages as well as the claim con-
cerning the development of these particular discourse functions from expressions
of comparison/approximation. The case of Hebrew constitutes an example of a
system in which one of the functions (quotation) has only recently begun to de-
velop. This provides an opportunity to observe a discourse function in emergence
and suggests the secondary nature of the function of quotation as compared to
those of hedging and focus marking in this particular functional itinerary.
More generally, this chapter supports the claim that there exist parallel path-
ways of grammaticization for discourse markers in various unrelated languages.
Moreover, we find further support here that the nature of the lexical sources influ-
ences the uses of the markers that derive from them. According to Hopper’s prin-
ciple of ‘persistence’ in grammaticization, “when a form undergoes grammaticali-
zation from a lexical to a grammatical function, so long as it is grammatically
 Metalanguage in Interaction

viable some traces of its original lexical meanings tend to adhere to it, and details
of its lexical history may be reflected in constraints on its grammatical distribu-
tion” (1991:€22). Since the lexical source of ke'ilu involves the additional semantic
element of the hypothetical condition 'ilu, and not only the comparative element
k(e)-, we should not be surprised to find a widespread additional function of ke'ilu
in comparison to the functions of English like and French genre – the function of
self-rephrasal.

7. Interacting as an Israeli via ke'ilu

We started out noting that ke'ilu and kaze have greatly proliferated in Hebrew casual
conversation in recent years. Whereas the most widespread function of ke'ilu is self-
rephrasal, that of kaze is hedging (around 70% of all kaze tokens in the database in-
volve hedging, as opposed to only 13.3% of all ke'ilu tokens, Maschler 2001).
In Fleischman and Yaguello (1999), we find a list of a variety of languages ex-
hibiting similar functional itineraries of words originating in the concept of com-
parison which have developed at least one of the functions of hedge (including, in
Fleischman and Yaguello’s categorization, self-rephrasal), focus, or quotative. Be-
sides English like and French genre, as we have seen, they list German so, Finnish
niinku, Bislama (New Guinea) olsem, Swedish likson, Italian tipo, Swedish typ,
Lahu (Tibeto-Burman) qhe, Japanese nanka, Tok Pisin olsem, and Buang (New
Guinea) (na)be.
This tendency across many languages to independently develop similar func-
tional itineraries for elements having similar lexical sources can shed light not
only on linguistic processes, but also on cultural ones (cf. Tagliamonte and Hud-
son 1999:€ 147). The recent proliferation of ke'ilu and kaze in Hebrew reflects a
cultural change witnessed in many societies recently. However, in the case of kaze
and ke'ilu, there is a specific Israeli bent to this cultural change.
Recall Katriel’s characterization of the Sabra (1986, 2004), the mythic image of
the New Jew born in Palestine during the British mandate (see Chapter 3, Section 7).
Katriel has shown that constructing the identity of this New Jew involved acquir-
ing new social practices, and, in particular, new ways of speaking and interacting,
first and foremost of which is the dugri speaking style – direct, straightforward,
often blunt, consciously “suspend[ing] face-concerns so as to allow for the free
expression of the speaker’s thoughts, opinions, or preferences in cases in which
they might pose a threat to the addressee’s face” (2004:€152). As one of Katriel’s
informants put it, “to speak dugri is to act like a Sabra” (2004:€143).
In a chapter entitled “Confrontational Dialogues: The Rise and Fall of Dugri
Speech”, Katriel (ibid.) argues that during the past quarter of a century, this
 Chapter 4.╇ The cognitive realm 

speaking style, along with the Sabra myth, have been in decline (cf. Kimmerling
2001). She traces the beginnings of the decline of the dugri mode of speaking to
the early 1980s, by which time Israelis “refused to accept the Sabra myth at face
value, yet were similarly reluctant to give it up completely, at least as a reminder of
past longings and dreams” (ibid.: 195). By the late 1990s, the change has become
more definite:
For a couple of decades, the small group of Sabras of European extraction retained
its status as a cultural elite. This position of social advantage became translated in
terms of military, political, and civilian careers, keeping the Sabra ethos and dugri
style alive in some influential domains of the Israeli social scene, especially in
military and political circles. However, even in these contexts it became increas-
ingly contested in the 1990s. This was indicated by the scandal triggered by [Knes-
set member and former General] Ori Or’s remarks that touched on interethnic
relations in Israel, which were described by some as dugri and by others as racist.
The scandal that came to be known as the Ori Or affair [in the fall of 1997] hints
at what has become of the dugri idiom in a society where cultural diversity could
no longer be ignored, and social exclusion buttressed by the demands for a mono-
lithic interactional style no longer held their sway. (ibid.: 197).

At the core of a dugri interactional style rests a decided, self-assured, and undoubt-
ing voice. One cannot constantly hedge and rephrase oneself and speak dugri at
the same time. I have argued (Maschler 2001) that the proliferation of kaze and
ke'ilu beginning in the 1990s is a linguistic manifestation of the decline of the du-
gri interactional style. According to Katriel, this “newfound tolerance for tenta-
tiveness, lack of resoluteness, and self-questioning” is a manifestation of what she
calls the softening of the dugri mode, “the affective display of tentativeness with
respect to one’s own speech through a variety of mitigating devices”. One of the
reasons for this softening is a growing recognition of the interpersonal costs of the
bluntness implicated by the dugri style in a society that has become increasingly
heterogeneous and hierarchical (Katriel 2004:€206–208).
Another reason for the softening of the dugri mode is “a shift towards inter-
personal focus and introspective stance”, which has been “reinforced by the in-
creasing influence in Israeli culture of the Western therapeutic ethos” (ibid.: 207).
To Katriel’s surprise, some of her teenage informants now interpret the term dugri
“with reference to self-disclosure and intimacy” (ibid.). I have found an additional
change in the way this word is used among teenagers, one quite relevant to our
interest in discourse markers.
For example, a conversation in which my 15-year-old daughter participated
(in June 2006) illustrates the word dugri employed as a discourse marker of em-
phatic agreement, somewhat reminiscent of some functions of the discourse
marker be'emet (‘really, actually, indeed’, lit. ‘in truth’) (Maschler and Estlein 2008),
 Metalanguage in Interaction

another word involving the notion of truthfulness. Complaining about the heavy
workload at their high school, these teenagers lament the huge amount of mate-
rial to be studied for an upcoming exam:
Excerpt 10 (‘Unfair Exam’)37:
1 Maya: 'ex hem metsapim,
how [do] they expect,
2 shenilmad kazot kamut shel xomer,
that we study such a quantity of material,
3 lamivxan?
for the exam?
4 Noy: veze lo ke'ilu38 she,
and it’s not as if SUBORD,
5 natnu lanu 'et haxomer,
they gave us the material,
6 harbe zman merosh.
lots of time in advance.
7 Maya: dugri!
exactly/just so!
In this excerpt, Maya presents an argument in support of her claim about the un-
fairness of the upcoming exam (namely, the large quantity of material, lines 1–3).
Her friend Noy adds a further argument in support of this claim (namely, the short
notice, lines 4–6). Upon presentation of a further argument in support of her own
claim, Maya cannot but agree emphatically. This she does with the utterance dugri,
unattached to any other word, occurring at intonation unit initial position (as a
separate intonation unit, in fact) at a point of speaker change, thus fulfilling the
structural requirement for prototypical discourse markerhood. This unattached
dugri does not modify any noun or verb (as in Katriel’s examples, see below) refer-
ring to the extralingual world. Rather, it functions metalingually in the world of
the interaction among speakers, expressing emphatic agreement, thus fulfilling
the semantic requirement for discourse markerhood as well.
Of course, the use of dugri in this sense is not unrelated to its original sense of
‘directness’. What I say in a ‘direct’ and self-assured voice is also exactly what I
strongly believe in and completely agree with. However, this use is far removed

37. This example has been reconstructed, as I do not have a recording of it. Its transcription is
therefore broader.
38. Note the literal use of ke'ilu here, followed by a subordinate clause.
 Chapter 4.╇ The cognitive realm 

from the adjectival and adverbial uses found in Katriel’s data (e.g., hu dugri (‘He is
dugri’) or tedaber dugri (‘Speak dugri’, i.e. ‘Speak straight’), (2004:€152)).
Note that this shift from an adjective or adverb to a discourse marker follows
a common grammaticization path (Traugott 1995a, Traugott and Dasher 2002,
and also Chapter 3, Section 6 and Chapter 6). Furthermore, the Hebrew word
dugri comes from Arabic dugri (which itself derives from Turkish dogru). How-
ever, whereas in Arabic (as well as in Turkish), dugri utterances are those in which
the speaker claims to be true to the facts “out there”, in Hebrew the speaker claims
to be true to him/herself, reporting subjective experiences, thoughts, and opinions
(and to some extent feelings) (Katriel 2004:€152). Of course, this being true to one’s
self has ramifications for the addressee, particularly pertaining to his/her face
wants, as illustrated by Katriel (1986). Thus we see that in the shift from Arabic to
present day teenage Hebrew, dugri has not only changed grammatical categories
but also undergone both subjectification and intersubjectification, in accordance
with the tendencies noted by Traugott for semantic change (Traugott 1989, 2003a,
and also Chapter 1, Section 8.1). We see that the path followed by dugri is a typical
grammaticization path of discourse markers.
Returning to kaze and ke'ilu, the words of an eighteen-year-old high school
student, interviewed by Hirschfeld (1990) in his cultural-philosophical study of
kaze and ke'ilu, manifest an interesting reflection on the decline of the decided,
undoubting dugri voice:
“‘halaxti kaze lakolnoa ke'ilu (‘I went kaze to the movies ke'ilu’)’ [... ] is said with the
feeling that there is something pretentious, scornful, or pathetic [... ] about really
wanting, really going, really doing. In other words, I really did go to the movies, but
this going didn’t have in it what was supposed to be in it. I went kaze. It’s only
similar to going. The movies, too, are just similar to movies. Things are much less
what they are, what we’ve heard they were, what we know they could be. I went –
indeed. But what happened was ‘kaze’”. (Hirschfeld, 1990:€9, translation mine).

However, this change in Israeli culture did not occur in vacuum. The widespread
functional path followed by equivalents of kaze and ke'ilu in so many of the world’s
languages (Fleischman and Yaguello 1999) suggests a wider phenomenon. The
words of Hirschfeld’s young man echo the larger postmodern context Israelis are
exposed to.
Gergen (1991) writes about the saturation of society by multiple voices. Accord-
ing to him, our exposure to new technologies in the postmodern era has resulted in
our being able to sustain relationships with an ever-expanding range of other per-
sons, and, therefore, in our increased awareness of the multiplicity of voices:
“[This] social saturation brings with it a general loss in our assumption of true and
knowable selves. As we absorb multiple voices, we find that each ‘truth’ is relativized
 Metalanguage in Interaction

by our simultaneous consciousness of compelling alternatives. We come to be aware

that each truth about ourselves is a construction of the moment, true only for a
given time and within certain relationships.” (ibid.: 16, emphasis mine).

In this postmodern world, in which “persons exist in a state of continuous con-

struction and reconstruction, [...] where anything goes that can be negotiated”
(ibid.: 7), it is no surprise that so much is hedged and rephrased. As we have seen
throughout this chapter, kaze and ke'ilu have a major role in this. “The limits of
language... mean the limits of my world”, wrote Wittgenstein in his Tractatus Log-
ico-Philosophicus. The motivation to hedge and rephrase utterances originates in
the same place, in recognizing the limits of one’s language (and of one’s world). It
is an attempt to ke'ilu overcome these limits.
In the Israeli context, however, a special bent is added to this postmodernity.
Besides the dialogue with the now-fading Sabra ethos, there are other cultural fac-
tors shaping this new way of interacting as an Israeli. Hirschfeld writes that in the
silence that was created in the classroom following the words of the youth quoted
above, another student raised her hand and said plainly:
“We go to the movies ke'ilu, because I feel that the important things in the world
are not us and what we do or want to do. The important things are (and she didn’t
think for long) Memorial Day, Intifada, The Holocaust. We’re just ke'ilu kaze”.
(1990:€9, translation mine).

There is an awareness here not only of the multiplicity of voices, but also of a particu-
lar history and political situation leading to their multiplicity. As in Du Bois’ model of
the internal and external forces to which grammars adapt in the process of grammati-
cization (1985:€361), cultural change, too, is subject to both types of influence.
Following the words of the young woman in Hirschfeld’s classroom, the “com-
plete sadness that descended suddenly” (ibid.) upon this room full of eighteen-
year-olds is instructive of what is so Israeli about this way of being in the world.
 chapter 5

Between realms
The discourse marker tov:
Accepting while shifting

1. Introduction

The Hebrew word tov is listed in Even-Shoshan’s dictionary (2003) in two separate
entries. The first, with which we shall not concern ourselves in this chapter, is as a
verb (a use found only in Biblical Hebrew). The second, with which we will be
concerned, is as an adjective meaning ‘good’ (ba'al 'erex xiyuvi (‘of positive val-
ue’)), a meaning it has had since Biblical Hebrew times, with cognates in the other
Semitic languages of Ugaritic, Aramaic, and Arabic. Even-Shoshan mentions that
tov may function as an adverb as well and, particularly in Biblical Hebrew, also as
a masculine noun. The final, least frequent meaning he lists, found already in Bib-
lical texts, is the following: milat haskama vexiyuv (‘a word of agreement and af-
firmation’) [meaning] ken (‘yes’), nixa (‘fine, let it be’, [translated by Sivan and
Levenston’s Hebrew-English dictionary (1967) as ‘good!’, ‘all right!’]). The Biblical
context provided is the first Book of Kings, Chapter 2, verse 18. There, King Solo-
mon’s older brother, Adoniyahu, asks Bath Sheba (Solomon’s mother) to request
her son, King Solomon, to allow Adoniyahu to marry Avishag Hashunamit. The
Biblical author continues: vatomer bat sheva tov 'anoxi 'adaber 'alexa 'el hamelex
(‘And Bath Sheba said: tov, I shall speak to the King on your behalf ’).
The common understanding of tov both as an adjective as well as ‘a word of
agreement and affirmation’ is perceived early by young Israelis, as the words of my
son, seven years old at the time, reveal. Upon hearing that I was studying tov, he
said: yesh shney tov (‘there are two tovs’), 'exad ze kshemashehu hu tov (‘one is when
something is good’), vehasheni ze kshemaskimim la'asot mashehu (‘and the second
is when you agree to do something’).
The present chapter is concerned with this last meaning of tov, as well as with
another meaning common in spoken Hebrew but not listed by any of the diction-
aries with which I am familiar.
For the first meaning, in which tov is employed by the recipient, examine the
following excerpt from a conversation between two women in their early twenties.
 Metalanguage in Interaction

Yafit is attempting to begin a story about a pair of shoes she found on one of her
shopping trips, while Meirav wants to know first how much they cost:
Excerpt 1 (‘New Shoes’):
25 Meirav: káma 'alu?
how much [did they] cost?
26 Yafit: /xip/.. réga,
/[I] loo/.. [just a] sec,
27 .. 'ani 'agía leze.
I’ll get to it.
28 Meirav: 'im 'at lò 'omeret káma 'alu,
if you don’t tell how much [they] cost,
29 'í 'efshar lehitrakèz.
[it’s] impossible to concentrate.
30 Yafit: 'ani 'agía.
I’ll get [to it].
31 takshívi!
listen!
32 Meirav: tóv.
okay.
33 Yafit: haláxti,
I went,
34 xipàsti na'aláyim.
looked [for] shoes.
Meirav humorously tells Yafit that unless she tells her first how much the shoes
cost, she won’t be able to focus on the story (lines 28–29). Yafit promises her again
that she’ll get to it (line 30), and asks her to listen to the story from its beginning
with takshivi! (‘listen!’, line 31). Meirav agrees to this request with tov (line 32), and
Yafit begins the story (lines 33–34).
This use of tov is equivalent to the English use of okay to affirmatively respond
to a request (Merritt 1978, Condon 1986, Beach 1993), as in the following excerpt
provided by Beach:
(4) UTCL: Family Phone: 2 (Beach 1993:€330).
Subscriber: Hang on I got a call on the other line.
Partner: ’Kay.
 Chapter 5.╇ Between realms 

In both the Hebrew and the English examples, a speaker requests the recipient
to do something, and the latter agrees. I have, therefore, translated the tov of
line 32 as ‘okay’.
The second function of tov, not listed by any of the dictionaries, is employed
not by the recipient, but rather by the speaker him/herself. In the following ex-
cerpt, two archaeology students at the University of Haifa are discussing a paper
they were assigned concerning an ancient city of their choice in the region. They
are making fun of the strange names some of these cities have (Constantinople,
Serageyopolis). Ami then jokingly tells Liraz to let their professor know that he’ll
be writing his paper on the ‘ancient’ city of Haifa. This should suffice, he adds,
because Haifa, too, has a ‘strange’ name:
Excerpt 2 (‘Archaeology’):
114 : .. tagídi lo,
tell him,
115 .. she'aní 'ose 'avodà,
{---laughing---}
that I am doing a paper
that I’ll be writing a paper,
116 'al xeyfá,
on Haifa,
117 Liraz: (laughs)
118 Ami: .. she/??/
that /??/
119 (laughs)
120 ... gam káxa,
also this way
this way too,
121 .. ze shèm meshuné.
{------laughing------}
this strange name
it’s a strange name.
122 {laugh} 'ani nish'àr bexeyfá.
I’m staying with [the topic of] Haifa.
123 Liraz: (laughs)
124 Amir: ..... tóv,
okay,
 Metalanguage in Interaction

125 'az legabèy hanose hashení--,

so concerning the topic the second
so about the second topic,
126 ... 'eh.. ze-- ha--'inyán haze,
uh.. it’s the issue this
uh.. it’s this issue,
127 shel ha--,
of the,
128 .... yom hastudént.
day of the student
Students’ Day.
1 29 Liraz: ... má 'ito?
what with it
what about it?
After Ami’s joking comment, Liraz laughs (line 123), and there is a long silence
(line 124). Ami then begins a new topic, one concerning Students’ Day celebra-
tions on campus. The transition to this topic is prefaced by tov, by the discourse
marker 'az (‘so’), and by the longer metalingual utterance legabey hanose hasheni
(‘about the second topic’). I do not have a recording of the beginning of the con-
versation, but, because hanose hasheni (‘the second topic’) is accompanied here by
the definite article (ha-), it can be inferred that a second topic had been on the
agenda for a while, and the transition to it at line 124 was expected at some point
following the topic of the archeology paper.
This is equivalent to the English use of ‘discourse ok’ at significant transitions
in the discourse, as described by Condon: ‘discourse ok participates in a default
organization of language behavior that distinguishes unmarked, routine sequences
and marked, nonroutine departures from expected events’ (2001:€491). While okay
has previously been shown to have a transitional function in English discourse
(Schegloff 1968, 1979, 1986, Schegloff and Sacks 1973, Merritt 1978, Hopper 1989,
Beach 1990, 1993, 1995), Condon shows that ‘discourse ok’ marks expected transi-
tions. In other words, okay ‘marks the transition across [the] boundary as a default
or expected one’ (ibid.: 496). For example, Condon observes that okay is often the
first linguistic form in highly contextualized interactions such as the beginning of
a class (ibid.: 508). Tov, too, is very often the first linguistic form employed in a
class taught in Hebrew. I have thus translated tov with ‘okay’ in excerpt 2 as well.
In both its agreement and ‘transition to an expected course of action’ mean-
ings, tov functions as a discourse marker. Recall that in this study an utterance is
 Chapter 5.╇ Between realms 

considered a prototypical discourse marker if it fulfills two conditions; one seman-

tic, the other structural:
a. Semantically, the utterance must have a metalingual interpretation in the con-
text in which it occurs. In other words, rather than referring to the extralin-
gual world, it must refer metalingually to the realm of the text, to the interac-
tion between its participants, or to their cognitive processes.
b. Structurally, the utterance must occur at intonation-unit initial position, ei-
ther at a point of speaker change, or, in same-speaker talk, immediately fol-
lowing any intonation contour other than continuing intonation. It may occur
after continuing intonation or at non intonation-unit initial position only if it
follows another marker in a cluster (Maschler 1998b: 31).
As for the structural requirement, in both excerpts tov occurs at intonation-unit
initial position. The tov of excerpt 1 fulfills the structural requirement because it
occurs at a point of speaker change. The tov of excerpt 2 fulfills the structural re-
quirement because it occurs in same-speaker talk following final intonation. With
respect to the semantic requirement, rather than referring to the extralingual
world (as, for instance, when tov functions as modifier of some noun referring to
the world “out there”), in both of the excerpts above tov carries a metalingual func-
tion: in excerpt 1 tov functions in the realm of interaction between speaker and
addressee; in excerpt 2 it functions in the realm of the text, marking transitions
within it. Thus each one of the tokens fulfills both requirements for prototypical
discourse markerhood1.
Hebrew tov and English okay, then, are equivalents of each other in at least the
two functions of agreement and transition to an expected course of action. To be
sure, English okay has entered the Hebrew discourse marking system just as it has
in many other languages2. Hebrew 'okey is sometimes employed throughout this
corpus in these two functions. However, in the present corpus, tov is much more

1. In Chapter 1 we saw that the two criteria in this definition coincide for 94% of the discourse
markers. The remaining 6% satisfy the semantic, but not the structural requirement. These
statistics are based on a sub-corpus consisting of 16 conversations (40 minutes of discourse
among 43 participants, see Maschler 2002a). The majority of the 6% that satisfy the semantic
but not the structural requirement are employed to construct two types of multivocality in dis-
course – constructed dialogue (Tannen 1989) and self-rephrasal (see Chapter 1, Section 7). In
the present extended corpus of 50 conversations, a larger percentage – 10 out of 62 instances of
tov (16%) – do not satisfy the structural requirement for prototypical discourse markerhood.
All of these tokens are employed at the beginning of constructed dialogue, following a verbum
dicendi, as will be illustrated below.
2. See Goss and Salmons 2000 and Maschler 2000a,b and Chapter 1, Section 2 for the role of
the bilingual strategy of language alternation in the borrowing and grammaticization of dis-
course markers.
 Metalanguage in Interaction

common than 'okey, and employment of 'okey seems to be a personal variation of

some individuals. From a linguistic point of view, tov has the advantage of being
easily traced back to a particular lexeme, which okay does not, despite the various
studies attempting to pin down the origins of English okay (e.g., Read 1963a,b,
1964a,b,c, 1988; Cassidy 1981).
The present chapter further explores the functions of discourse marker tov in
Hebrew casual conversation and tackles the question of how a particular discourse
marker might come to have both interpersonal as well as textual functions. I ad-
dress this question while keeping in mind the literal meaning of tov, ‘good’, in or-
der to shed light on the grammaticization path of this word in Hebrew.

2. Data

In the 50 conversations of the Haifa Corpus of Spoken Israeli Hebrew (Maschler

2004) on which the present study is based (Chapter 1, Section 3), altogether 62
tokens of the discourse marker tov are employed. Thus tov is employed at an aver-
age frequency of one token every 2.4 minutes in this corpus. The tokens are dis-
tributed among the following functions:

Table 1.╇ Functions of tov throughout the database

Interpersonal tov Textual tov - Textual tov Ambiguous Total

Transition into -Closing current between interper-
following action action sonal/textual
tov
29 (46.8%) 22 (35.5%) 2 (3.2%) 9 (14.5%) 62 (100%)

Table 1 shows that tov in its interpersonal uses is somewhat more common (46.8%)
than tov in its textual uses (35.5% + 3.2% = 38.7%) throughout the database. An-
other 14.5% of the tokens perform both functions simultaneously, and have thus
been classified as ‘ambiguous’. It is these ambiguous cases that are particularly help-
ful in tracing the connection between the interpersonal and textual functions of tov.
I now examine the functions of tov in these various categories in more detail.

3. Interpersonal tov

Interpersonal tov is employed in this corpus to express:

1. agreement to an action
 Chapter 5.╇ Between realms 

2. acceptance of some state of things

3. third-turn receipt
4. concession, often preceding a ‘but’ response
5. ironic agreement, i.e., disagreement.
The first category, ‘agreement to an action’ was illustrated in excerpt 1. Let us now
examine the other four categories.

3.1 Acceptance of some state of things

Just as a recipient can agree to some course of action, they can also agree to or accept
a description of some state of things. In the following excerpt, Gila and Yonat, two
women in their early twenties, are having a chat, in the course of which Gila offers
Yonat some citrus fruit which she hopes is a pomelit (a cross between a pomelo and
a grapefruit, common in Israel in the winter months) and not a grapefruit:
Excerpt 3 (‘Sofas’):
51 Gila: (9.00) 'ani mekavà sheze pomeliyót,
I hope that it’s pomelits,
52 .. velò 'eshkoliyót.
and not grapefruit.
53 Yonat: 'o?
or?
54 (laughter) /xamúts/ {tastes fruit}
/sour/
55 Gila: (laughter)
56 Yonat: ... 'o--y,
57 (slight laughter)
58 ... tipa már,
drop bitter
a bit bitter,
59 .. hayiti /mexaná 'et ze/.
I would /call it/
I’d say.
60 Gila: (laughs)
61 tóv,
okay,
 Metalanguage in Interaction

62 'az bó’i,
so come
so let’s,
63 .. 'ani 'etén lahem 'od 'eize,
I’ll give them another,
64 Yonat: .... yom yomáyim,
day two days
day or two,
65 Gila: la.. la'asòf ktsat pazám3 mashehu.
to to gather some time something
to become more ripe
After tasting some of the fruit, Yonat announces that it’s kind of bitter (lines 56–59).
To this Gila responds with some laughter and a tov (line 61), accepting Yonat’s
description of the state of the fruit. She then proceeds, following the discourse
marker 'az (‘so’) and the directive (functioning as discourse marker here4) bo'i
(‘let’s’, lit. ‘come’), which is directed towards herself as well, to say that she’ll let the
fruit sit around a bit longer until it becomes more ripe (lines 62–65). In fact, this
suggestion is co-constructed (Lerner 1991) by both of the participants, as Yonat
verbalizes the length of time – a day or two (line 64) – which they should wait.
Tov of acceptance is often employed in this corpus as part of the strategy of
constructed dialogue (Tannen 1989). This happens when a speaker describes
someone’s acceptance of the state of things in the storyworld. In the following ex-
cerpt from the mid 1990s, Ada, who accompanied the Israeli delegation to Oslo
during the peace talks with the Palestinians, tells about getting lost on a tram in
Oslo with Abdullah, a member of the Palestinian delegation. First, she tells of their
attempts to locate the phone number of their guide:
Excerpt 4 (‘Ada and Abdullah in Oslo’):
87 Ada: .... ve'az hu 'omér,
and then he says,
88 'á--h,

3. pazam is an acronym originating in army slang: pesek zman minimali (‘minimal period of
time’). It refers to the length of time which a soldier has accumulated in some rank.
4. Hebrew bo/'i/'u (lit. ‘come’ masc. sg./ fem. sg./ masc. or fem. pl.) has undergone grammati-
cization from an imperative to a hortative and then a discourse marker, somewhat like English
let us > let’s (Traugott 1995b: 36–37).
 Chapter 5.╇ Between realms 

89 yesh lax ta’télefon shel hamadrìx haze vehaze?

there is to you the phone of the guide this and this
do you have the phone [number] of some particular guide?
90 ... kén ken bétax! {Ada’s voice to Abdullah}
yes yes sure!
91 .. 'ani matxila lexapés beze,
{--laughing--}
I start looking in this
I start looking here,
92 .. 'an’ló motsèt.
can’t find [anything].
93 .. mexapéset,
[I] look [there],
94 .. ló motsèt.
can’t find [anything].
95 ... bekítsur hayinu mevohalím laxalutín.
in short we were alarmed completely
in short we were completely alarmed.
96 .. 'amarti tóv,
I said okay,
97 nitkasher venish'al beshagrirùt yisra'él.
we’ll call and ask at embassy Israel
we’ll call and ask at the Israeli embassy.
98 .... hitkasharti lamerkaziyà shel.. 'e--h.. 'óslo,
I called the operator of.. u--h.. Oslo,
99 ... bikáshti /ta’ telefon shel/ hashagrirùt,
I asked for /the telephone [number] of/ the embassy,
100 ... hitkashárti lashagrirùt,
I called the embassy,
Following the description of a series of failed attempts to locate the guide’s phone
number (lines 91–94), Ada sums up, following the discourse marker bekitsur (‘in
short’, ‘anyway’), the state of mind they were in as a result of their failed attempts:
hayinu mevohalím laxalutín. (‘we were completely alarmed’, line 95). She then moves
on to describe her response to this situation (lines 96–97): 'amarti tóv, (‘I said okay’),
nitkasher venish'al beshagrirùt yisra'él. (‘we’ll call and ask at the Israeli embassy’). Tov
 Metalanguage in Interaction

here verbalizes Ada’s acceptance of the state of things at the time, and it precedes the
new strategy she devised as a result; namely, calling the Israeli embassy.
Note that tov of excerpt 4 does not appear at intonation-unit initial position,
and thus, like 16% of the tov tokens throughout the corpus (see footnote 1, cf. Chap-
ter 1, excerpt 6), does not satisfy the structural requirement for prototypical dis-
course markerhood. However, in accordance with what I have claimed earlier
(Maschler 2002a and Chapter 1, Section 7), tov here follows the verbum dicendi
'amarti (‘I said’) and constitutes Ada’s constructed dialogue to herself in the story-
world. In Chapter 1, Section 4, we saw that the system of discourse markers is an
iconic one: ‘big packages’ (Sacks 1992), such as the beginnings of new stories, are
announced by ‘heavier’ strategies such as clusters of several consecutive prototypi-
cal discourse markers and longer metalingual utterances, while ‘small packages’
such as sub-episodes in a narrative are announced by ‘lighter’ strategies. The bound-
ary between the unit introducing constructed dialogue and the constructed dia-
logue itself is a lower-level boundary in the discourse. Such boundaries often do not
involve the use of any discourse markers at all. When they do, the markers they
employ are ‘lighter’ in character, in that they satisfy only the semantic, but not the
structural requirement for prototypical discourse markerhood. They have different
prosodic qualities such as following continuing intonation in same-speaker talk or
not appearing at intonation-unit initial position, as is the case in excerpt 4.

3.2 Third turn receipt

Frequently, the acceptance of the state of things is given in a particular sequential

position, known in the conversation analytic literature as ‘third turn receipt’. For
example, examine the following segment from the continuation of the ‘new shoes’
conversation:
Excerpt 5 (‘New Shoes’):
35 Yafit: .... ve--
and
36 no.. 'ani--,
/??/.. I[’m],
37 ... tipùs shel.. nyu bálans.
a type of.. New Balance
a New Balance type.
38 ... ke'ilu,
like,
 Chapter 5.╇ Between realms 

39 'ani xolá 'aleyhem.

I sick about them
I’m crazy about them.
40 Meirav: ... má ze nyu bàlans.
what is New Balance.
41 xevrá?
[a] company?
42 Yafit: xevrá.
[a] company.
43 kmo hay te
like High Te
44 'e--h.. hay tek lemashal.
u--h.. High Tech for instance.
45 .. 'az 'e--h,
so u--h,
46 Meirav: tóv.
{laughing}
okay.
47 Yafit: ra'íti ka'ele shel nyu bàlans,
I saw like these of New Balance
I saw such shoes by New Balance,
Following Meirav’s clarification request concerning the term ‘New Balance’ (lines
40–41), Yafit explains that it is a brand name of gym shoes (lines 42–44) and moves
on to continue her narrative (lines 45 and 47). Overlapping her, at line 46, Meirav
shows her acceptance of this answer with a tov, much in the same way that okay
can be used in this ‘third’ position: “Free-standing ‘Okays’ are also employed by
current speakers who initiate such activities as questions, and having received an
affirmative, acceptable, and/or clarifying answer from recipient, move next to
mark recognition and/or approval in third slot via ‘Okay’” (Beach 1993:€331).

3.3 Concession

Tov can be employed by a recipient to pay lip service towards accepting some state of
things. In these instances, it is often immediately followed by a ‘but’ response. The
result is a fleeting concession of the state of things as described by the interlocutor.
 Metalanguage in Interaction

In the following interaction between a young woman and her parents, the
mother is in the middle of a story when the father suddenly interrupts with a ques-
tion concerning the tape recorder:
Excerpt 6 (‘Family Gossip’):
192 Mother: .. mà 'ód hu 'amar,
what else he said
what else did he say,
193 Father: tagídi,
tell (fem. sg.) me,
194 .. 'at hexláft po batarìya?
you changed here battery
did you change batteries here?
195 Orna: .. ló.
no.
196 Father: ... 'át yoda'at shehadavàr haze 'ovéd
you know that the thing the this works
you know that this thing works
197 .... 'ani ganávti 'et ze--,
I stole this,
198 Mother: 'ani be'émtsa mishpàt.
I[’m] in [the] middle of [a] sentence.
199 Orna: 'ima be'émtsa
Mom[’s] in [the] middle of
200 {laughter}
201 Mother: .. 'ani mishtagá'at mimxa!
I’m going crazy from you
you’re driving me crazy!
202 Father: tòv 'aval ze norá xashuv,
tov but it’s terribly important,
203 ladá'at 'et ze.
to know this.
204 Orna: {laughter}
205 Father: {laughter}
206 ... beshiv'ím veshmòne,
in seventy eight,
 Chapter 5.╇ Between realms 

207 ganàvti 'et ze mimé'ir 'adív,

I stole this from Meir Adiv,
208 .. meshumásh.
used.
When the father continues the tape recorder topic (lines 196–197), both mother and
daughter overlap, objecting that the mother was in the middle of a sentence (lines
198–199). The mother then adds an expression of her annoyance at the situation 'ani
mishtagá'at mimxa! (‘you’re driving me crazy!’, line 201), which the father overlaps
with tòv 'aval ze norá xashuv, ladá'at 'et ze. (‘tov but it’s terribly important, to know
this’). With this tov, the father accepts the state of things as described by mother and
daughter, and pays lip service towards acknowledging the inadequacy of his inter-
ruption. His acceptance is quite momentary though – it is immediately followed by
'aval (‘but’), a humorous justification of his behavior (lines 202–203), and a continu-
ation of the topic of the tape recorder (lines 206–208). I have left this token of tov
untranslated, because the concessive character of tov seems stronger to me than that
of okay; a fact related, I think, to the clearer semantic origin of tov (‘good’).
Concessive tov is not always followed by an explicit ‘but’. In the following ex-
cerpt, Orna describes some old fashioned medicine in the form of mustard pow-
der mixed with water, which her mother recommended she use:
Excerpt 7 (‘Old Fashioned Medicine’):
33 Orna: ... la'asót mize--,
[you need] to make out of it,
34 'at yoda'at,
you know,
35 .. 'is
dou
36 ktsat,
a bit,
37 'isá kazoti?
a dough sort of?
38 Xava: ... nu?
go on?
39 Orna: .. limróax 'et ze,
[you need] to spread it,
40 'al 'e--h,
on u--h,
 Metalanguage in Interaction

41 .. xatixot neyá--r?
pieces of paper?
42 .... velasím 'et ze,
and to put it,
43 'al ha--,
on the--,
44 re'ót.
lungs.
45 ... me'axóra?
from the back?
46 Xava: nu,
yeah,
47 vemá ze 'ose?
and what it does
and what does it do?
48 Orna: ... hi 'omeret,
she says,
49 z
i
50 ze,
it,
51 ze soféax xòm,
it gathers heat,
52 ze lokeax 'et kól hahitkarerùt.
it takes all the cold
it makes the entire cold go away.
53 ... /mahér/,
/quickly/,
54 /mahér/,
/quickly/,
55 'amárti le'ima sheli,
I said to mother my
I said to my mother,
{------smiling-------}
56 tóv,
fine,
 Chapter 5.╇ Between realms 

57 'ani ló ya'ase 'et ze.

I won't do it.
Following the mother’s constructed dialogue5 concerning the benefits of this med-
ication (lines 49–54), Orna reports, again in constructed dialogue, her response to
her mother in a smiling tone: 'amárti le'ima sheli, (‘I said to my mother’), tóv
(‘fine’), 'ani ló ya'ase6 'et ze. (‘I won’t do it’). Tov here does not accept the mother’s
advice even momentarily, as it is accompanied by a smile and a declaration of its
unacceptability (line 57). I have translated the tov of excerpt 7 as English fine since
this word seems to capture its concessive flavor better than okay.
A similar phenomenon was described by Auer for certain German particles.
Of particular relevance to tov is Bavarian fei (< fein, lit. ‘fine’, now roughly mean-
ing: ‘contrary to what you may assume’) (1996:€317) and German gut (Auer p.c.).
One could easily interpret the tov of lines 56–57 in this way: ‘contrary to what you
may assume, I won’t do it’. We see that although there is no explicit ‘but’ here, the
contrastive stance is clear.

3.4 Ironic agreement: Disagreement

For some readers, the tov of excerpt 7, line 56 will be interpreted with a certain
degree of irony – the irony of a daughter ridiculing her mother’s unacceptable sug-
gestions. I submit that, for all readers, the irony expressed by the tov of the follow-
ing excerpt is unquestionable. This is a political argument that took place in 1994
at a family meal between the father, a right-leaning supporter of the Likud party,
and his left-leaning son, who is in his early twenties and a supporter of the Ma'arax
(Labor) party. The argument concerns the peace agreement with Jordan that had
just been signed by the Labor government:
Excerpt 8 (‘Political Argument’):
4 Gabi: bentáyim,
meanwhile,
5 mi shedafák 'et tahalìx hashalòm,
he who screwed up the process peace
those who screwed up the peace process,
6 xamésh 'esre shanà,
[for] fifteen years,

5. In Hebrew, the absence of the subordinator she- (‘that’) is evidence that an utterance is in
‘direct speech’, i.e., constructed dialogue.
6. Note the use of the third person masc. sg. future prefix ya- (‘he will’) instead of the first
person sg. (both masc. and fem.) prefix 'a in the verb ya'ase, as noted in Bolozky 1984, 1999.
 Metalanguage in Interaction

7 ... ze halikúd,
is the Likud,
8 lo ha--
not the--
9 Shani: ta'avír li ta’pirè?
pass me the puree?
10 Gabi: .. ló hama'aràx.
not the Ma'arax.
11 Father: ze naxón,
it’s true,
12 .. halikúd,
the Likud,
13 'asú 'et ze.
did it.
14 ... 'avál,
but,
15 (3.5) ze ló 'omer,
this doesn’t mean,
16 shegàm hama'aráx,
that also the Ma'arax,
17 lo 'osím 'et ze.
aren’t doing it.
18 Gabi: bentáyim,
so far,
19 hem 'osìm yafe me'ód,
they’re doing nice very
they’re doing very well,
20 .. 'im 'efshàr letsayén.
if it’s possible to remark
if I may say so.
21 Father: ... 'aval /????/
but /??????/
22 Gabi: / bentáyim hem menahalim 'et/ hamasà umatan bexoxmá.
/so far they’re carrying on the / negotiations with wisdom
/so far they’re carrying on the/ negotiations skillfully.
 Chapter 5.╇ Between realms 

23 Father: tóv.
{ironic tone}
fine.
24 ....... 'ókey.
{ironic tone}
okay.
25 Gabi: .... 'atá lò maskím 'iti,
you don’t agree with me,
26 .. shehaheském hateritoryali,
that the agreement the territorial,
that the territorial agreement,
27 'im yardén,
with Jordan,
28 hu mutsláx?
is successful
is good?
29 Father: ... 'ìm 'anì lo maskim 'itxá,
if I don’t agree with you,
30 Shani: todá. {to Mother, concerning the food being passed}
thank you.
31 Father: .... 'ìm 'anì lo maskim 'itxá,
if I don’t agree with you,
32 ... 'ata titén li lehagìd 'et da'ati?
you will let me say my opinion
will you let me express my opinion?
At lines 18–22, Gabi claims in a rather agonistic tone that the Labor government has
been carrying on the negotiations quite skillfully. To this the father responds with
an ironic (as well as sarcastic) tov (line 23) followed by a long silence and an ironic
'okey (‘okay’, line 24). With these two tokens, the father clearly shows that he does
not agree with his son that the Labor government is carrying on the negotiations
with Jordan skilfully. Gabi’s continuing talk 'atá lò maskím 'iti, (‘you don’t agree with
me...’, lines 25–28) is evidence that this was Gabi’s interpretation at the time of the
argument. The father’s response 'ìm 'anì lo maskim 'itxá (‘if I don’t agree with you...,
lines 29–32) is evidence that this was the father’s interpretation at the time as well.
Thus the tov (as well as the 'okey) here is a token of ironic agreement, which is to say,
disagreement. An English translation of fine seems to capture this irony.
 Metalanguage in Interaction

All of the excerpts of Section 3 illustrate interpersonal tov employed by the

recipient. Let us now examine more carefully textual tov (columns 2 & 3 of Table
1), employed by the speaker.

4. Textual tov: Marking expected transition

Textual tov is employed by the speaker to mark transitions:

1. at beginnings of major topics, whether narrational or elicitational
2. between the episodes and sub-episodes of a narrative
3. to return to an action which has been interrupted
4. at the end of a topic/action.
I now illustrate each of these categories.

4.1 Transitional tov into the beginning of a narrative

We have already seen an example of tov beginning an elicitational topic (excerpt

2). As for narrational topics, the following example comes from the beginning of a
story about the army days of one of the women:
Excerpt 9 (‘Wounded in Lebanon’):
1 Hanna: .... tóv,
okay,
2 .. 'az 'ani 'asapèr 'et hasipúr,
so I will tell the story,
3 .. 'al 'éx niftsáti belevanòn?
about how I was wounded in Lebanon?
4 Galia: 'ókey.
okay.
5 Hanna: ... tóv.
okay.
6 .. 'az 'e--h,
so u--h,
7 ... kshehitgayásti latsavà,
when I was drafted to the army,
Although I do not have a recording of the beginning of this interaction, it is rela-
tively safe to assume that the two women met here in order to record ‘a good nar-
rative’ for the course assignment for which this conversation was collected. This
 Chapter 5.╇ Between realms 

tov, then, like the one in excerpt 2, is equivalent to ‘discourse okay’ as described by
Condon, which marks ‘significant, yet expected, transitions in the discourse’
(2001:€495). Telling a story was clearly on the agenda, as we learn from the definite
article preceding the Hebrew word for story – hasipur (‘the story’, line 2). The re-
cipient (who was also the student recording the conversation) responds with the
agreement token 'okey (‘okay’, line 4), and Hanna continues the story, beginning
with another tov (line 5), to which we shall return in the discussion of the ambigu-
ous cases (Section 5).

4.2 Transitional tov between the episodes of narrative discourse

Examine the following narrative, told by a young man, Adi, to another, Gil, and to
a woman, Hadas, about a suitcase left unattended at the airport:
Excerpt 10 (‘Said’s suitcase’):
10 Adi: ..... ze hitxíl baze she,
it started with the [fact] that,
11 ... kshetásnu,
when we flew,
12 betèrminal shtáyim,
in terminal two,
13 .. 'anaxnu omdím,
we[’re] standing [there],
14 .. doxafím,
[people are] pushing,
15 bardák,
[it’s] pandamonium,
16 'áf 'exad lo záz.
no one is moving.
17 tóv,
18 /má ze?/
/what’s this?/
19 ..... pít'om,
suddenly,
20 ... pinúy!
evacuation!
 Metalanguage in Interaction

21 .. la'azóv hakól,
leave everything,
22 borxìm haxútsa.
[people] are running outside.
In this excerpt, tov is found at a significant boundary in the narrative, as the nar-
rator moves from the orientation to the complicating action (Labov 1972). It is
employed along with pit'om (‘suddenly’) as the speaker begins to recount the ac-
tions leading to the climax of the story. The transition from the orientation of a
narrative to its complicating action is a relatively expected one, originating from
the narrative scheme familiar to both participants. Thus, like ‘discourse okay’, tov
between narrative episodes constitutes a transition to an expected course of action
(Condon 2001).

4.3 Transitional tov returning to an interrupted action

Returning to an action following an interruption is also an expected course of ac-

tion. The following excerpt comes from a story told by Yif 'at to a married couple,
Avi and Inbar. In the middle of the story, the couple’s toddler shows up and they
deal with her and with opening something:
Excerpt 11 (‘Drunk on Campus’):
99 Avi: ... 'at lo yoda'at /liftoax 'et ze/?
{---------------pp-------------------}
you don’t know /how to open it/?
100 Inbar: toda.
{pp}
thank you.
101 ... bàt lashevet pó? {to toddler}
you came to sit here?
102 Yif 'at: ... tóv,
okay,
103 bekitsú--r,
anyway,
104 Inbar: ma, {to toddler}
what,
105 lehoríd lax? {to toddler}
[shall I] bring [this] down for you?
 Chapter 5.╇ Between realms 

106 Yif 'at: yom 'exád,

one day,
107 hem halxú,
they went,
108 le'eyze mesibá,
to some party,
Following the interruption dealing with the toddler, Yif 'at returns to her story
(line 106) following a cluster of two discourse markers – tov, bekitsur (‘okay, any-
way’, lines 102–3). While bekitsur has the main function of resuming a main topic
following a digression, as we saw in Chapter 3, tov here marks the close of the in-
terruption and the transition to an expected course of action – a return to the
previous topic.

4.4 Transitional tov ending a topic/action

Occasionally (at only 3.2% of the cases), the transition to the following conversa-
tional action is marked at the end of the preceding action rather than at the begin-
ning of the next one. One example comes from a narrative co-told by Orna and
Anat to Ran concerning a diving course they attended with three other women
friends. The narrative goes on for a long time, humorously describing how each
one of the women dropped out as the course went on:
Excerpt 12 (‘Diving Course’):
29 Orna: harofé hipìl 'ota,
the doctor dropped her [i.e., caused her to drop out],
30 .. ya'ani.
sorta.
31 .. 'al.. 'e--h.. refu'í hi naflà.
on uh medical [reason] she fell.
for uh medical reasons she dropped out.
32 All: {laughter}
33 Ran: /'atem 'al?????????/
you [dropped out] for????????
34 Orna: ló /naxon/.
{--laughing--}
not true.
35 Anat: ... 'anàxnu nafálnu,
we dropped out,
 Metalanguage in Interaction

36 'al sotsyométri.
{---laughing---}
on sociometric [grounds]
for sociometric reasons. [i.e., for not having the right social attitude]
37 All: {laughter}
38 Orna: ... nish'àrnu 'árba banot.
{----------laughing---------}
we were left four girls.
39 ... to
al
40 ... tóv!
alright!
41 Ran: ... nu,
42 'ani maksh
I’m listen
43 .. 'ani mexake lá--
{-----laughing---}
I’m waiting for the
44 .. lapo'énta.
{--laughing--}
for the point.
45 Anat: la-- pántsh layn.
{-------laughing-----}
for the punch line.
46 Orna: ... tóv!
alright!
47 .. magía hayom hashenì.
arrives the day the second
the second day arrives.
In lines 29–31 Orna ends the description of the first woman dropping out on the
first day of the course. This is interrupted by Ran’s humorous comment and a hu-
morous objection to it by Orna and Anat (lines 33–37). Following some laughter,
Anat summarizes the first day of the course: nish'àrnu 'árba banot. (‘we were left
four girls’, line 38). This is ended with tov! (which I have translated here as ‘alright!’,
line 40). Next comes Ran’s humorous prodding to get to this story’s point finally
(lines 41–44). Orna continues her story with the next episode – the second day of
 Chapter 5.╇ Between realms 

the course: magía hayom hashenì. (‘the second day arrives.’, line 47) – preceded by
another token of tov (line 46), this time opening a new episode in the narrative (as
in excerpt 11). Employment of the same discourse marker both at the ending of
the previous episode and at the beginning of the next one is further evidence of its
role in marking significant transitions in the discourse.

5. Ambiguous cases: Between interpersonal and textual tov

In the two previous sections, we have seen that interpersonal tov is employed by
the recipient whereas textual tov is used by the speaker. However, all of our utter-
ances are simultaneously constrained by the various contextual realms shaping
discourse (Becker 1988), not just by the interpersonal and textual realms. It should
not surprise us, therefore, that in 9 cases throughout the database (14.5%) tov
functions both interpersonally and textually. These tokens are generally employed
by the recipient, and it is the recipient, not the speaker, who also either initiates a
move on to the next episode/action or closes the current one. These ambiguous
instances are instructive for understanding the connections between the functions
of tov and for positing a possible grammaticization path between them.
In a study first published in 1983, Jefferson noted the configuration of Minimal
Acknowledgment Recycle, in which speakers “can manage both exhibiting
attention to the overlapping utterance [...] and immediately getting back to the over-
lapped talk [...]”. She characterized it as “an attention on the way to something else”
(Jefferson 1993 [1983]: 3, emphasis in original). Jefferson noticed that a similar phe-
nomenon can occur in the case of topic shift: “a recipient will at some point produce
a minimal acknowledgment of a prior utterance and follow that with a shift in topic”
(ibid.: 4). This is another case of “exhibiting attention while shifting” (ibid.: 29).
This observation is critical for understanding the connection between the in-
terpersonal and textual uses of tov. Some of the tov tokens in the present database
also exhibit this dual nature. A recipient can simultaneously provide some ac-
knowledgment towards accepting the state of things, while at the same time initi-
ating a move to begin the next action (or to end the current one). This duality has
also been found for English okay: “participants rely on ‘Okay’ as a means of simul-
taneously attending to prior turn while also setting-up next-positioned matters
(topics/activities)” (Beach 1993:€329).

5.1 Third turn receipt + ending an action

In the case of third turn receipt, it is often the case that the recipient’s tov also ends
the current action, because recognition of having received a response also generally
 Metalanguage in Interaction

ends the sequence. This is particularly apparent in the following conversation be-
tween a young man, Uriel, and three young women concerning the high school
Uriel attended:
Excerpt 13 (‘Intruding on One’s Privacy’):
116 Dalit: ... 'ata lamàdeta bebet sefer 'im ma'alót?
you studied at school with Ma'alot [kids]?
117 Uriel: ... kén.
yeah.
118 Galit: hù mikfar vradím!
he[’s] from Kfar Vradim!
119 Dalit: ... ze karóv?
{whispering, laughing}
that[’s] close?
120 Uriel: ..... bet sèfer 'ezorí!
school regional
it’s a regional school!
121 Liora: .. kòl ha'ezór,
all the region
the entire region,
122 .. zehu.
that’s it.
123 Dalit: .. tóv.
okay.
124 (slight laughter)
125 Uriel: ... 'az zéhu,
so that’s it,
126 ... 'az bekitsúr,
so anyway,
127 ... bekitsúr,
anyway,
128 ... 'az 'e--h,
so u--h,
129 ... 'az 'ani 'omer,
so I say,
130 ... kén,
yeah,
 Chapter 5.╇ Between realms 

131 veze--,
and so on,
132 .. 'az hi 'oméret,
so she says,
133 ... dibàrti 'im danít.
I talked to Danit.
At line 116, Dalit is surprised to learn that Uriel went to school with children from
the town of Ma'alot in the Northern Israeli region of the Galilee. This town is
known as a ‘development town’, which in the Israeli context generally means a
post-statehood peripheral town a significant percentage of whose population be-
longs to a lower socio-economic strata than the general Israeli population. Uriel
answers positively (line 117), and Galit overlaps him reminding Dalit that Uriel is
from Kfar Vradim (line 118). Kfar Vradim, which neighbors Ma'alot, is known as
a village whose population generally enjoys a rather high socio-economic status.
At line 119 Dalit whispers shyly, asking whether Kfar Vradim is near Ma'alot. Uri-
el answers that the school is a regional school (line 120), implying that therefore
children from various settlements in the Galilee (and perhaps regardless of their
socio-economic status) study there. Liora reinforces this (lines 121–122), and Dalit
accepts the answer at line 123 with a tov, accompanied by slight laughter, perhaps
indicating embarrassment at her ignorance and at the lack of political correctness
implied by her surprise. Dalit’s tov constitutes a case of third-turn receipt, but we
see that it also ends the current clarification sequence concerning the school: at
line 125, with a cluster of 8 consecutive discourse markers (lines 125–129), Uriel
moves on to continue his story. Thus we see that a token of tov indicates recogni-
tion at having received a clarifying answer from the recipient at the same time that
it serves to end the current action.

5.2 Third turn receipt + transition into following episode/action

A less obvious case occurs when a recipient acknowledges some state of things and
at the same time moves on to the next action. We have seen this in excerpt 9, line 5:
Excerpt 9 (‘Wounded in Lebanon’):
1 Hanna: .... tóv,
okay,
2 .. 'az 'ani 'asapèr 'et hasipúr,
so I will tell the story,
3 .. 'al 'éx niftsáti belevanòn?
about how I was wounded in Lebanon?
 Metalanguage in Interaction

4 Galia: 'ókey.
okay.
5 Hanna: ... tóv.
okay.
6 .. 'az 'e--h,
so u--h,
7 ... kshehitgayásti latsavà,
when I was drafted to the army,
In lines 1–3 Hanna requests permission to tell the story. Her utterances in these
lines also function as the abstract (Labov 1972) to her narrative. Her request is
responded to positively (line 4) with 'okey, and in line 5 Hanna acknowledges this
response with the third-turn receipt tov. However, this tov also serves to begin the
orientation episode of the narrative; in other words, it constitutes a transition into
the next expected action.

5.3 Acceptance of some state of things + transition

into following episode/action

A token of tov can perform both an interpersonal and a textual function even
when it does not occur in third position. In the following interaction between two
women in their early forties, Nurit tells Sara about the dealings her husband had
with various medical doctors while they were on vacation. In this section of the
story, they are on the phone with their family practitioner, who is wondering why
the results of the lab test the husband took still aren’t in:
Excerpt 14 (‘Doctors’):
129 Nurit: .. hu 'omér,
he says,
130 má ze,
what[’s] this,
131 'én totsa'òt!
no results!
132 .. ma.. masàrtem beyom xamishí,
y..you handed [the culture] in on Thursday,
133 .. ze tsarix lihiyót,
it has to be [in],
134 hatotsa'ót!
the results!
 Chapter 5.╇ Between realms 

135 ... yom shishí,

Friday,
136 yom Shabát,
Saturday,
137 yom rishón,
Sunday,
138 .. ts’xot lihiyót kvar hatotsa'òt.
must be already the results
the results should be in already.
139 Sara: naxón.
right.
140 Nurit: .. tóv.
okay.
141 tén li ta’tèlefon shelxa,
give me the phone yours
give me your phone number,
142 .. 'ani 'avarér.
I’ll find out.
In lines 130–138, which comprise the constructed dialogue of the family practi-
tioner, the doctor calculates the days that have elapsed since the culture was taken
and concludes that the results should have already been in. At line 139 Sara, the
recipient of this story, agrees with the doctor’s conclusion, and Nurit continues the
story at line 140 with the doctor’s constructed dialogue: tov. ten li ta’telefon shelxa,
'ani 'avarer. (‘okay. give me your phone number, I’ll find out’, lines 140–142). This
tov signals the doctor’s acceptance of the situation (i.e., that there are still no results
although they should have been in already). At the same time, it begins a new epi-
sode in the story, the episode describing a new action of the doctor’s: he requests
their phone number, promising to notify them about what he finds out at the lab.
Thus we see that the same token of tov functions both as acceptance of some state
of things as well as the beginning of a new sub-episode7 in the narrative.
The difference between this and the two preceding cases is that here tov is ac-
tually verbalized not by the recipient in the interaction but rather by the story-
teller. However, the storyteller is constructing the utterances of the doctor, who is

7. This is similar to excerpt 4, except that the boundaries between the units are more clear
here (partly since this tov satisfies all the requirements for prototypical discourse markerhood),
and tov clearly precedes the beginning of a new sub-episode (while in excerpt 4, lines 96–97, it
does not).
 Metalanguage in Interaction

a recipient in the storyworld, in the sense that he is accepting a state of things

which Nurit and her husband have presented to him.

5.4 Concession + transition into following episode/action

In a study of the position preceding the front-field of the finite German verb (‘the
pre-front field’) as a preferred locus of grammaticization, Auer writes:
Semantically, these pre-front field agreement adverbials and tokens often preface
a possible counter-argument, which the speaker presumes to be relevant in some
kind of imagined dialogue with a partner who is not necessarily identical with the
one co-present. This counter-argument, although it has not been made by the
other party, at least not explicitly and not in the prior turn, is taken up and ‘agreed
with’ in a yes-but strategy (1996:€316–317).

In excerpt 7, we saw that a real, not an imagined, dialogue (with the mother) was
reported, and the counter-argument was made via tov: tov (‘fine’), 'ani lo 'a'ase 'et
ze. (‘I won’t do it’), (or ‘contrary to what you may assume, I won’t do it’, see Section
3.3). The idea of an imagined dialogue with some partner is enlightening for un-
derstanding tov in the following excerpt from a story a mother tells her husband
and twenty-something year old daughter about her own mother:
Excerpt 15 (‘Grandma Can’t Remember’):
10 Mother: ... xamishím shana,
fifty years,
11 hi xáya sham be--,
she lived there on,
12 .. behamèlex jórj.
on King George [street].
13 ... ze ló kàl,
it’s not easy,
14 le'ishà mevugéret,
for woman old
for an older woman,
15 pit'òm la'avór dira.
to suddenly move apartments.
16 Father: naxón.
right.
17 Mother: .. tóv 'anaxnu 'azàrnu,
of course we helped,
 Chapter 5.╇ Between realms 

18 veze ve
and so on and
19 .. veyóm
and [a] day
20 .... yóm lifney haha'avarà,
[a] day before the move,
21 .... 'ani hayíti sham 'ita--,
I was there with her,
22 ... ve.. sax hakol haya mesudár,
and all in all was arranged
and all in all things were arranged,
23 ... vehe'evárnu,
and we moved [everything],
24 ... venixnásnu habàyta,
and we entered [her new] home,
25 .... ladirá,
into the apartment,
26 .... pit'òm betsura me'òd meshunà--,
suddenly in fasion very strange
suddenly in a very strange way,
27 ... hi sho'élet 'oti,
she asks me,
28 .... má 'anaxnu 'osìm po?
29 what we doing here
what are we doing here?
30 ... láma 'anaxnu nimtsa'ìm po?
why we here
why are we here?
31 ... lama 'ani nimtset po,
why I here
why am I here,
32 ... hi sho'elet 'oti.
she asks me.
This tov (like the one in excerpt 14) is employed by the narrator of the story, not by
the recipient, and there is no prior speaker (neither in the storyworld nor in the
world of the interaction) to whom this token responds. However, the tov of line 17
 Metalanguage in Interaction

‘introduces a syntactic gestalt’, to use Auer’s words again (ibid.: 316), and prefaces
a possible argument, which the speaker presumes to be relevant in some kind of
imagined dialogue, such as ‘it is not easy for an older woman to suddenly move
apartments, particularly with no help’. This imaginary argument ‘is taken up and
“agreed with” [or rather countered] with a yes-but strategy’: tov 'anaxnu 'azarnu
veze (‘tov we helped and so on’), in other words, ‘yes, you may assume we did not
help, but we did’, or ‘contrary to what you may assume, we helped’. The translation
of tov as ‘of course’ seems more adequate here (cf. Auer’s discussion of German
natürlich ‘of course’, ibid.: 315–316).
With this example of imaginary dialogue, we begin to see how a metaphorical
extension of the primary interactional usage of the agreement token tov “opens the
way to grammaticalization into a connective with quasi-conjunctional function”
(ibid.: 317). We shall continue to explore this last excerpt in the concluding chap-
ter of this book in order to further develop a theory about the grammaticization of
discourse markers (Chapter 6, sections 2.1 and 2.3).
However, the tov of excerpt 15, line 17, also has a textual function in the dis-
course, because it constitutes the beginning of a new sub-episode in the narrative.
Lines 10–12 are part of the orientation to the story, describing the Grandmother’s
living situation over the past 50 years. The narrator then interrupts (lines 13–15)
to make a general assertion concerning older people and moving, to which the
recipient responds with naxon (‘right’, line 16). At line 17 the narrator returns to
the narrative, to a new sub-episode of the orientation which describes the specific
background to the story she is telling – the happenings of the day preceding the
move. Thus we see that a single token of tov functions both as a concessive and as
a transition into the following sub-episode of the narrative.

5.5 Concession and return to main topic

Tov is employed in this corpus twice in order to return to a main action following
an interruption, both times followed by the discourse marker bekitsur (‘anyway’).
We have seen one instance in excerpt 11. In the other instance, tov is employed by
the recipient. Here Gila is telling Yonat a story. Yonat had interrupted with some
comment concerning the story, which Gila responds to in lines 109–113:
Excerpt 16 (‘Sofas’):
109 Gila: .... ze be'ayà shel mi shekone sapót,
this [is the] problem of [one] who buys sofas,
110 velo bodék,
and doesn’t check,
 Chapter 5.╇ Between realms 

111 .. 'im hu 'ohév 'otam,

if he likes them,
112 'o ló 'ohèv 'otam,
or doesn’t like them,
113 .. 'o 'im hem mat'imòt lo labáyit.
or if they match for him to the house
or if they match his house.
114 Yonat: .. tòv 'akítser,
okay anyway,
115 sof hasipùr hu shehem nifredú?
[the] end of the story is that they broke up?
116 Gila: sóf hasipùr
end of the story
shehem nifredu,
that they broke up
the end of the story is that they broke up,
From the surrounding context8 we learn that Yonat does not entirely agree with
Gila’s assertion in lines 109–113. However, rather than pursue the disagreement, at
lines 114–115 she urges Gila to get on with the story. This action is begun with the
discourse marker cluster tòv 'akítser (‘okay anyway9’), this time employed by the
recipient of the story. The tov of this excerpt is borderline between textual and in-
terpersonal, because it also constitutes the recipient’s response of paying lip service
towards accepting the speaker’s previous statement (as in excerpt 6 of Section 3.3).
This tov, then, both carries a concessive flavor and marks the close of a digression
and the transition into an expected course of action – the return to the main topic.

6. Grammaticization of tov

In Hopper and Traugott’s definition, grammaticization is “the change whereby

lexical items and constructions come in certain linguistic contexts to serve gram-
matical functions, and, once grammaticalized, continue to develop new grammat-
ical functions” (2003: xv). As we have seen in Chapter 3, according to their princi-
ple of unidirectionality in grammaticization, a typical path of grammaticization

8. For a detailed study of this excerpt, see Chapter 3 Section 4.1.

9. As we saw in Chapter 3, the discourse marker bekitsur has several variants, one of which is
'akítser, much like the English variant of anyway, anyways.
 Metalanguage in Interaction

follows a cline of de-categorialization (ibid.: 106–109). Thus, typically, the starting

point of the process is some lexical item that fully belongs to a major category such
as noun or verb; over time, the lexical item becomes more grammatical and “tends
to lose the morphological and syntactic properties that would identify it as a full
member of a major grammatical category [...] In its most extreme form such a
change is manifested as a cline of categoriality, statable as:
major category (> adjective/adverb) > minor category” (ibid.: 107),
where ‘minor category’ includes categories such as preposition, conjunction, aux-
iliary, pronoun, and demonstrative (Hopper and Traugott, ibid), as well as the cat-
egory of discourse marker (Ferrara 1997:€371), as we have seen in the discussion of
bekitsur (Chapter 3).
The Hebrew word tov has followed this path of grammaticization. In the ab-
sence of a diachronic corpus, I first rely on information found in dictionaries.
Even-Shoshan (2003) lists the period of Hebrew at which the meanings were first
used (Biblical, Talmudic, Medieval, Modern). In the introduction to his dictionary
(ibid.: ix), he writes that the various meanings are listed in the order of their fre-
quency in Modern Hebrew, from most frequent to least so. He lists the following
meanings for tov, in the following order:
1. (Adjective) of positive value
2. (Adjective) nice, of good deeds
3. (Adverb) worthwhile (to...)
4. (Noun) a good thing, a good deed, goodness
5. a word of agreement and affirmation.
6. (Modern Hebrew only) The mark [school grade] 8 (between 75 and 84).
7. (Modern Hebrew only) ‘tov be-’ ‘good in’ as in ‘She’s good in math’.
The first five meanings are found already in Biblical Hebrew. The nominal use of
tov (‘goodness’) is highly infrequent in Modern Hebrew and restricted only to very
high registers such as poetry and archaic language. All citations given for this use
are Biblical. From this I conclude that tov as a major category (noun) has become
less frequently employed over the years. The most common use according to Even-
Shoshan is that of adjective (‘good’), but tov is also found, though much less fre-
quently, as an adverb. This information suggests, then, that the word tov has moved
from the major category of noun to the intermediate categories of adjective, and,
much less frequently, adverb. From there, it has moved on to the minor category
of (interpersonal) discourse marker. This happened already in Biblical Hebrew (‘a
word of agreement and affirmation’), as we saw in the example from the Book of
Kings which opened this chapter. This meaning ranks fifth on Even-Shoshan’s list,
implying that it is one of the least frequently employed meanings of tov in the
 Chapter 5.╇ Between realms 

modern language. I suspect, however, that had Even-Shoshan based his lexicogra-
phy on a corpus of spoken language, this meaning would have been much more
frequent, perhaps the most frequent.
Grammaticization of tov did not stop once it became a discourse marker. In
this minor grammatical category it continued to develop. We have seen another
use of tov as a discourse marker: the textual use of ‘transition to an expected course
of action’. The fact that it is not found in any of the dictionaries, which survey He-
brew from Biblical times to the present, supports the claim that this is the most
recent development of this word.10
Note that the development from interpersonal to textual discourse marker in
the case of tov contradicts the cline posited in Traugott 1982:
propositional > (textual) > expressive (later termed ‘subjective’).
This cline suggests that when meanings develop from propositional to interper-
sonal (‘expressive’ or ‘subjective’ in Traugott’s terminology), they must pass through
a textual stage first. I have suggested that tov first developed interpersonal func-
tions and only later – textual ones. Indeed, this cline later came to be considered
too strong (e.g., Heine, Claudi and Hünnemeyer 1991:€ 190–191, Brinton 1996)
and was restated as a set of three tendencies (Traugott 1989:€34–35, see Chapter 1,
Section 8.1).
How does a word meaning ‘good’ develop first into an interpersonal discourse
marker and then into a textual one? I would like to suggest that from the adjectival
meaning of tov (‘good’), the interpersonal discourse marker emerged through use
in metalingual utterances such as ze tov, velaxen 'ani maskim laze/lehamshix (‘this
(referring to the state of things) [is] good, and therefore I agree to it / for you to
continue’). In other words, before moving on with the rest of the discourse, an
agreement must be reached between its participants to the effect that they both
find the state of things satisfactory. This is what I referred to as ‘accepting while
shifting’ in the title of this chapter, improvising on Jefferson’s ‘exhibiting attention
while shifting’ (1993 [1983]: 29). This ‘finding the state of things satisfactory’ was
perhaps first verbalized as the nominal sentence ze tov (‘this [is] good’), after which
it became just tov (‘good’). Of course, in the absence of a spoken corpus from an
earlier period of the language, this can only be hypothesized.
The ambiguous cases presented in Section 5 show that a recipient may exhibit
acceptance of the state of things and at the same time initiate a transition to the

10. Grammaticization of tov as a discourse marker has recently been accompanied also by a
structural change. No examples are found in the present corpus, but recent talk over the past two
years or so manifests the rather marked form to--'ov (with a lengthened vowel and inserted glot-
tal stop).
 Metalanguage in Interaction

next action (or to the ending of the current one). Excerpts 14 and 15 show that this
can even be done by the narrator of a story, not only by the recipient. In Section 5.4
we saw Auer’s suggestion (1996) concerning imaginary dialogue with some re-
cipient to which the speaker responds, beginning with a concessive such as one of
the German ‘equivalents’ of tov. I would like to extend this idea of imaginary
dialogue to the dialogue a narrator may hold with him/herself concerning ‘the
state of things as they have so far been presented’. In the absence of a comment
addressed to the speaker before continuing, a speaker may carry on a metalingual
dialogue with him/herself and examine the state of his/her discourse thus far. If
everything is satisfactory to the speaker’s mind, he or she may indicate that this is
so via tov, and move on to the next, expected, action or episode. Over time, of
course, this inner metalingual dialogue ceases to be carried out at every single
shift, and tov comes to be used unambiguously, as in the excerpts of Section 4, to
mark expected transitions. In this way a word that had originally served interper-
sonal purposes comes also to serve textual functions in the discourse.
Again, in the absence of a spoken corpus from an earlier period there is no
way to ascertain this suggestion. However, the fact that actions and episodes some-
times end with formulaic statements involving the word tov lends further support
to the idea of speakers scanning their discourse thus far and indicating their satis-
faction at it. Such formulaic utterances include: 'ad kan hakol tov (‘so far every-
thing [is] good’) (cf. English so far - so good), or, as the fixed phrase goes, kol ze tov
veyafe (‘all this [is] good and nice’) (cf. English that’s all very nice or that’s all well
and good) or even just yafe (‘nice’) (cf. English very well then). And for the end of
stories, we find sof tov - hakol tov (‘end [is] good - everything [is] good’ (cf. English
all is well that ends well). These proverbs and formulaic utterances found at the
ends of actions/episodes suggest that speakers are indeed involved in some kind of
(inner) metalingual dialogue taking stock of the state of their discourse thus far.
Another source of support for the grammaticization path suggested here
comes from studies of ‘equivalents’ of tov in other languages. We have already seen
that Bavarian German employs fei (< fein, lit. ‘fine’, now roughly meaning: ‘con-
trary to what you may assume’, Auer 1996) in a concessive way similar to that of
tov. In Spanish, de Fina’s study of bien in the classroom talk of teachers shows that
it has both organizational functions (‘move to another activity’) and evaluative
functions (in response to student’s answer to teacher’s question) (de Fina 1997).
Schiffrin adds that “like okay, the positive connotation (i.e. ‘I accept this’) of bien
has been semantically bleached [...] in transitional (but not evaluative) environ-
ments” (2001:€64). In Chinese, hao, an adjective meaning ‘good’, is also employed
as a discourse marker with two functions, one indicating approval or agreement,
the other marking completion of social actions. It ‘appears to function in Chinese
in a manner similar to that of OK in English’ (Miracle 1989:€226). Finally, Sherzer
 Chapter 5.╇ Between realms 

analyses the Brazilian thumbs-up gesture, showing that it has a paradigmatic

meaning of ‘“good”, “positive”, or “OK”, and [an] interactional or syntagmatic
meaning “social obligation met”’ (1991:€193), and that it also functions as “an in-
teractional link between moves, units, or moments within an interaction”
(ibid.: 192). In this last function, Sherzer mentions that it is similar to the use of
OK in service encounters in the United States, as found by Merritt (1978).
One of Hopper’s principles of grammaticization is persistence, which he for-
mulates in the following manner:
When a form undergoes grammaticalization from a lexical to a grammatical func-
tion, so long as it is grammatically viable some traces of its original lexical mean-
ings tend to adhere to it, and details of its lexical history may be reflected in con-
straints on its grammatical distribution (1991:€22).

If in several mostly unrelated systems – for example German, Spanish, Chinese,

and Hebrew, as well as Brazilian gestures – an adjective meaning ‘good’ has inde-
pendently developed the same discourse functions as an interpersonal and textual
discourse marker, this is further support of the persistence of its lexical source in
its functional itinerary.
Finally, following his micro-level discourse and social interaction analysis of
the Brazilian thumbs-up gesture (TUG), Sherzer relates it to the larger context of
Brazilian social and cultural life: “the TUG is public, friendly, and playful, and cov-
ers up brutal conflicts in the society in which it occurs” (1991:€195–196). Quoting
his personal communication with Da Matta, a scholar of Brazilian culture (e.g., Da
Matta 1991), Sherzer continues: “[t]he gesture, thus, has the power of disarming
disagreement. [...] Approaching the Brazilian TUG in this way is to explore its
place in a Brazilian ethnography of communication” (ibid.) Another culture in
which disagreement is highly pervasive, of course, is the Israeli one (Katriel 1986,
2004). That tov is one of the most frequent discourse markers employed through-
out the present database could be viewed as grammatical compensation for this
particular facet of Israeli culture.
 chapter 6

Concluding remarks
Grammaticization from interaction

1. The system of discourse markers permeating interaction

This study set out tracing a framework for conceptualizing discourse markers ar-
rived at through close synchronic analysis of the metalingual function of discourse
markers in bilingual and monolingual interaction. On the basis of this analysis, I
arrived at a definition of prototypical discourse markers having both a semantic
and a structural component, in which the concept of metalanguaging plays a cru-
cial role. Discourse markers were shown to be employed for the purpose of meta-
languaging: in contrast to utterances referring to the extralingual world, discourse
markers refer to the text itself, to the interaction among its participants, and/or to
the cognitive processes taking place in their minds during verbalization. I showed
that from the emic perspective of participants in an interaction, discourse markers
constitute a distinct linguistic category with particular patternings and regulari-
ties, employed for the purpose of accomplishing frame shifts in interaction.
Discourse markers were first investigated as a system. I showed that this sys-
tem exhibits three types of patterning involving (1) the functions fulfilled by dis-
course markers, (2) the moments at which they are employed in interaction, and
(3) the structural properties of discourse markers.

1.1 Function

The functions fulfilled by discourse markers were conceptualized through Becker’s

approach to text as constrained by contextual constraints (1988), and also with ref-
erence to Chafe’s theory of cognitive constraints on information flow (1987, 1994).
A moment of frame shift was defined as a moment at which maximal shifts occur,
relative to the surrounding discourse, in constraints from the various contextual
realms shaping discourse (Maschler 1994b). I suggested sorting out each discourse
marker token according to the realm in which there are maximal shifts in contex-
tual constraints when the particular marker in question is employed (Maschler
1994b, 1997a). This resulted in a preliminary classification of the markers into in-
terpersonal, referential, structural, and cognitive realms, and in some preliminary
 Metalanguage in Interaction

quantitative findings concerning the contextual realms involved in frame shifting

in Hebrew casual conversation. We saw that in the particular corpus in which this
was investigated, 50% of all discourse markers function to create and reflect rela-
tions in the textual realm of discourse (including both the referential and struc-
tural realms), 35% negotiate interpersonal relations among discourse participants,
and 15% of all makers attest cognitive processes involved in frame shifting (Chapter
1, Table 2). We also saw that in contrast to the relatively limited variety of relations
between parts of a text, created and reflected by 24 textual discourse marker types
in that particular corpus, there is a very wide array of interpersonal relations to be
negotiated via interpersonal discourse markers (57 types).
The conceptualization of discourse markers according to the contextual realms
at which they operate was tied to the phenomenon of clustering of discourse mark-
ers (Maschler 1994b, Aijmer 2002) often found at frame shifts: frame shifts are
characterized as moments in discourse at which contextual constraints change
maximally. Thus, the verbalization of a cluster of discourse markers – a marker
from each contextual realm – at such moments is to be expected.
This necessarily limited bird’s-eye view of the functions accomplished by dis-
course markers as sketched by the above-mentioned table was complemented by
detailed investigations into the functions and nature of four particular markers
representing each of the contextual realms: interpersonal nu, textual bekitsur, cog-
nitive ke'ilu, and tov, which was found to be both interpersonal and textual.
Nu was shown to function in (1) hastening a non-verbal action, (2) urging
further development within a topic, (3) granting permission to perform an action,
or (4) as a keying discourse marker providing a joking/provoking tone.
For bekitsur we saw the following functions: (1) summarizing, (2) resuming
main topic following a digression, and (3) foregrounding a new narrative.
The investigation of ke'ilu revealed four functions apart from its literal func-
tion as the conjunction ‘as if ’: (1) hedging, (2) self-rephrasal, (3) focus-marking,
and (4) quotation. Only in its self-rephrasal role was ke'ilu shown to function as a
discourse marker (Chapter 4, Section 4.4).
Finally, we saw that tov tokens employed by the addressee function in the in-
terpersonal realm as (1) agreement, (2) acceptance, (3) third turn receipt, (4) con-
cession, or (5) ironic agreement, i.e. disagreement; whereas those employed by the
speaker function in the textual realm to mark transitions (6) at beginnings of ma-
jor narrational or elicitational topics, (7) between episodes or sub-episodes of a
narrative, (8) returning to an interrupted action, or (9) at the end of a topic/action.
We also saw that 15% of all tov tokens perform both interpersonal and textual
functions simultaneously.
 Chapter 6.╇ Grammaticization from interaction 

1.2 Metalaguaging at frame shifts

As for the second patterning exhibited by discourse markers, that concerning the
moments at which they are employed in interaction, we saw that the higher the
boundary in the hierarchy of frame shifts permeating interaction, the more meta-
lingual material employed to construct that boundary.
Metalingual material at the highest level in the hierarchy is sometimes longer
and not yet sufficiently crystallized to be considered a discourse marker. However,
such longer metalingual utterances are precisely the material that often develops
into discourse markers. As an example, compare the longer metalingual utterance
rotse lishmoa keta? (‘wanna hear something weird/funny [keta, lit. ‘a segment’]?’)
generally found at beginnings of narratives, which inflects for person, number,
and gender, and allows noun modifiers, such as demonstratives, determiners, ad-
jectives, etc., with a related expression, crystallized and sufficiently frozen to be
considered a discourse marker, the enthusiasm-displaying fixed phrase, 'eize keta!
(lit. ‘what [a] segment!’, ‘what a story/incident/coincidence!’, perhaps best ren-
dered in English by ‘wow!’) following highly involving discourse strategies em-
ployed by the interlocutor1. Longer metalingual utterances were treated here as
potentially ‘on their way’ to becoming discourse markers. This motif was returned
to throughout each of the four chapters studying the particular markers, and evi-
dence was presented in support of the various metalingual utterances from which
the discourse markers are hypothesized to have originated.
For nu, metalingual utterances such as nu tagid li (‘nu tell me’), nu ma 'ata 'omer
(‘nu what do you say’), nu, ma 'at mitlonenet (‘nu, what are you complaining for’),
nu, 'ata lo mevin she... (‘nu, don’t you see that...’), nu, tasik kvar 'et hamaskana (‘nu,
draw the conclusion already!’), or nu 'az tasbir li 'im 'ata yaxol, ve'ani lo ma'amin
she'ata yaxol (‘nu so explain to me if you are able to, and I don’t believe you are’) were
posited in order to explain how a non-lexical item involved in encouraging one to
proceed with an action might come to have also keying functions in discourse.
For the discourse marker bekitsur, a metalingual utterance such as 'im lomar
bekitsur ‘put succinctly’ (lit. ‘if to say in short’) or 'ani 'omar bekitsur (‘I will say
succinctly’) was suggested.
For self-rephrasal ke'ilu, ke'ilu 'amarti (‘as if I were saying’) or ke'ilu lomar (‘as
if to say’) were hypothesized.
Finally, for the discourse marker tov, we saw that positing a metalingual utter-
ance such as ze tov, velaxen 'ani maskim laze/lehamshix (‘this (referring to the state
of things) [is] good, and therefore I agree to it / for you to continue’) and an inner
metalingual dialogue of similar sorts carried out by the speaker allows us to tie the

1. See Maschler 1998b for a detailed analysis of this particular example.

 Metalanguage in Interaction

interpersonal functions of this discourse marker, which operates between realms,

to its textual functions.
In the absence of a diachronic corpus, the type of evidence presented in sup-
port of the claim that such metalingual utterances might be the origin of these
discourse markers included a close analysis of the polysemies of specific discourse
marker tokens in interaction, secondary source information (dictionaries), utter-
ances exhibiting the actual use of such collocations with metalingual elements,
interpretations provided by interaction participants, and evidence from other lan-
guages, in which elements of similar lexical nature have developed into discourse
markers of similar pragmatic function.
The assumption underlying this last source of evidence is that grammatical
morphemes – here discourse markers – show similar polysemies and similar de-
velopment paths across languages, resulting from similar human discourse and
interactional features (Hopper and Traugott 2003:€33) because these morphemes
“reflect the metaphorical processes that are based on human cognitive make-up,
and they reflect the inferences that humans commonly make when they commu-
nicate” (Bybee, Perkins, and Pagliuca 1994:€302, see also Section 2 below).

1.3 Structure

The third type of patterning exhibited by discourse markers involves their struc-
tural properties – linguistic as well as prosodic. First, the structural properties
with which a discourse marker is verbalized were shown to often be associated
with different functions for a particular marker. For instance, in the case of nu, we
saw that stand-alone nu? tends to function in the sequential realm, whereas keying
nu is always followed by more talk, usually within the same intonation unit.
In the case of bekitsur, we saw that a function which is further removed from
the literal meaning of the word (summarizing an episode and returning to the
main action, as opposed to summarizing a list) is accompanied by a prosodic
change (position of the accented syllable) in the verbalization of this word, such
that bekitsúr becomes bekítsur. In its foregrounding function, bekítsur is accompa-
nied by additional prosodic features such as a longer preceding pause and a drastic
change in volume. Furthermore, the resumptive and foregrounding functions are
accompanied by phonological reduction phenomena such that any of the follow-
ing variants of bekitsur may occur: bekítser, hakítser, 'akítser, kitsúr. The resump-
tive and foregrounding functions may also be articulated by playful variants, such
as be'úktsur, bekítskets, hakítskits and 'akítbitser.
In the case of ke'ilu, the structural feature separating discourse marker (self-
rephrasal) tokens from hedging, focus-marking, quotative, and literal ke'ilu tokens
is that only the self-rephrasal tokens are consistently found at intonation-unit
 Chapter 6.╇ Grammaticization from interaction 

initial position2. Thus, the tokens employed intonation-unit initially were also the
highest in the metalingual dimension.
In the case of tov, no completely consistent correlation was found between its
structural properties and various uses, except that tov tokens exhibiting the textual
function almost always occupy the entire intonation unit, whereas tov tokens fol-
lowed by an additional utterance within the same intonation unit tend to exhibit the
interpersonal function. The lack of complete consistency here mirrors the frequent
blending of the textual and interpersonal functions in employment of tov.
Whereas others have also pointed to the role of prosody in separating out the
various uses of a particular discourse marker (e.g., Ferrara 1997, Tabor and Traugott
1998), the present approach is unique in that structural properties of discourse
markers were found to be of primary importance for defining prototypical discourse
markers. From a study based on a sub-corpus of the present study’s database
(Maschler 2002a), we see that 94% of the discourse markers fulfill not only the se-
mantic requirement of metalinguality, but also a structural requirement: they occur
at intonation-unit initial position, either at a point of speaker change, or, in same-
speaker talk, immediately following any intonation contour other than continuing
intonation (unless they follow another discourse marker in a cluster). For the defini-
tion of prototypical discourse markers, then, it is not the prosody with which the
discourse marker is articulated which plays a crucial role, but rather the prosody of
the intonation unit immediately preceding the discourse marker.
The finding that almost all discourse markers in this corpus are employed at
intonation-unit initial position3 corroborates a structural feature of Traugott’s the-
ory of semantic change relating to discourse markers (1995a). Based on studies of
English indeed, in fact, and besides, Traugott argues that as discourse markers orig-
inating from clause-internal adverbials become increasingly grammaticized, their
scope increases from operating within the clause to operating over the clause, to
operating over a stretch of discourse. This process is paralleled by the fact that
many discourse markers occur at syntactically marginal positions in the clause,
depending on the typology of word order in the language. In English they tend to
appear preceding the clause, whereas in Japanese they tend to occur following it.
In other languages, such as Greek, they often occur as the second element of the
clause (“Wackernagel’s position”) (Traugott and Dasher 2002:€156). Since Modern

2. The exceptions to this are ke'ilu tokens in constructed dialogue, in which there is no into-
nation-unit boundary between the utterance introducing the constructed dialogue and the dia-
logue itself (e.g., Chapter 1, excerpt 6).
3. Again, the exceptions consist mainly of discourse markers opening constructed dialogue
following an utterance introducing the constructed dialogue with no intonation unit boundary
between the two as in Chapter 1, excerpt 6. There is also one instance of ke'ilu intonation-unit
finally (Chapter 4, footnote 23).
 Metalanguage in Interaction

Hebrew is generally an SVO language (Ravid 1977), like English, we would expect
Hebrew discourse markers to occur at initial position.
I would like to suggest that the position occupied by discourse markers is interac-
tionally motivated. Auer has written about projection in interaction and projection in
grammar (2005). Projection, according to Auer, is the phenomenon “that an indi-
vidual action or part of it foreshadows another”. It is based both on interactional as
well as on grammatical (particularly syntactic) knowledge. The term ‘trajectory’ refers
to “the time course over which [a projection] develops and comes to closure or reso-
lution” (ibid.: 8). Projection establishes a cognitive rhythm in interaction, in that par-
ticipants go through a phase of maximal effort in the early part of a trajectory (both
planning the emerging structure and understanding it), while these efforts decrease
the closer a trajectory comes to a close, as the final parts are more easily predicted.
Projections also vary in strength. For example, in a language like Hebrew, in
which prepositions precede noun phrases, prepositions make a strong projection
about the nature of the following element – it must be either a determiner (often
already fused with the preposition) or a noun4. “Some words, such as conjunctions
like [German] und ‘and’ or oder ‘or’, produced after a syntactic closure, leave all
options open apart from not continuing, i.e., they project syntactic continuation in
an extremely vague manner” (ibid.: 16). I would add here that discourse markers
carry an even weaker syntactic projection concerning the syntactic nature of the
utterance to follow. At the same time, however, they carry a strong interactional
projection of an immediately pending frame shift, as well as a projection concern-
ing the type of frame shift that is about to occur. Since, as I have argued, frame
shifts tend to occur at clause boundaries (a point of maximal shift in constraints of
linguistic structure, among other constraints), the strong interactional projection
of an imminent frame shift carried by discourse markers constitutes motivation
for verbalizing them in syntactically marginal positions relative to the clause – ei-
ther initially (as in English or Hebrew) or finally (as in Japanese).
Auer writes that the question of whether a particular instance of projection is a
grammatical or an interactional one is not always unambiguous (ibid.: 27). Although
he does not call it a discourse marker, as an example for this blurring he discusses
German nur (‘only’), which can be used as a scalar adverbial when part of a noun
phrase, as part of a predicate, or, when in the pre-front field, as a ‘framing device’ for
the following utterance (ibid.: 17). In some contexts, “it may be asked whether nur
projects on the basis of its syntactic status or on the basis of being a (subsidiary)
verbal activity in its own right that foreshadows disagreement” (ibid.: 27). The

4. For the effect this word order typology (i.e., function words precede the content words they
serve as satellites to) has on self-repair strategies in English, Hebrew, and German, see Fox,
Maschler, and Uhmann (forthcoming).
 Chapter 6.╇ Grammaticization from interaction 

framing Auer mentions in the context of nur is the frame shifting which, as I have
argued, all discourse markers are involved in constructing. The example of German
nur is of course reminiscent of the English discourse markers which have developed
from sentence internal adverbs discussed by Traugott (1995a).
According to Auer, “in some cases, activity type and syntactic type project at
the same time” (2005:€28). I have suggested that, because of their role in construct-
ing frame shifts in interaction, all discourse markers manifest double projection.
On the one hand, because just about any syntactic construction is possible follow-
ing a discourse marker, these utterances carry weak grammatical projection. On
the other hand, because discourse markers always entail a frame shift, they carry
strong actional projection. Furthermore, their projection includes clues as to the
type of frame shift imminent. The projection entailed by lexemes which have de-
veloped into discourse markers, then, is in a sense ‘floating’ between grammatical
and interactional projection, depending on the particular context in which they
are employed. This accounts for the varying scopes and positions which these lex-
emes may have in discourse (Traugott 1995a). We shall return to the double pro-
jection of discourse markers in Section 2.3.
We have seen, then, that 94% of the discourse markers in the sub-corpus
(Maschler 2002a) fulfill the semantic requirement of metalinguality in my definition
of prototypical discourse markers as well as the structural requirement. Returning
now to the remaining 6%, the non-prototypical discourse markers fulfilling only the
semantic requirement, the majority of these markers were shown to involve a move
of a lower order compared to the moves accomplished by prototypical discourse
markers: moving to a different voice of the speaker’s – either the voice of some other
speaker in the narrated world (constructed dialogue) or the self-rephrasal voice of
the speaker (following a token of ke'ilu preceded by continuing intonation).
Employing a discourse marker which does not fulfill the structural require-
ment for discourse markerhood is a more subtle strategy for constructing a bound-
ary between conversational actions. We have seen that the structural patterning of
discourse markers fits in with the patterning concerning moments at which dis-
course markers are employed in interaction (Section 1.2): the lower the boundary
between conversational actions, the more subtle the strategy constructing it.
We have seen an array of structural strategies employed for constructing frame
shifts in interaction (Chapter 1, Figure 1) ranging from longer metalingual utter-
ances at the most prominent frame shifts, to clusters of prototypical discourse
markers, to a single prototypical discourse marker (as in the majority of frame shifts
throughout the data), to non-prototypical discourse markers fulfilling only the se-
mantic requirement, to frame shifts involving no discourse markers at all, to frame
shifts involving no intonation unit boundaries, and finally to frame shifts involving
the blurring of syntactic properties of constructions (Chapter 1, Section 7).
 Metalanguage in Interaction

These strategies progress from most prominent to most subtle, and they per-
meate interaction in an iconic manner: higher-level frame shifts are constructed
via the heavier semantic-pragmatic strategies (e.g., longer metalingual utterances,
prototypical discourse marker clusters), whereas lower-level frame shifts are con-
structed via the lighter prosodic and structural strategies (non-final intonation
contour preceding the discourse marker, a lack of an intonation unit boundary or
of a discourse marker altogether). This iconicity helps participants distinguish
higher-order frame shifts from the more subtle ones, thus contributing to a clearer
specification of the projected action.
We have seen (Chapter 1, Section 7) that the most minute frame shifts may even
interfere with the syntactic patterns of the language and result in highly ‘unentrenched’
uses of constructions (Thompson and Hopper 2001, Hopper 2004). This brings us to
reconsider the phenomenon of grammaticization of discourse markers.

2. Grammaticization of discourse markers

I have suggested four different grammaticization paths for the four discourse
markers nu, bekitsur, ke'ilu, and tov. Below I first review them and then consider
the common threads as well as the differences among them. This is then followed
by a discussion of projectability and grammaticization of discourse markers. I will
make use of the grammaticization path of tov in order to elaborate on this last
topic of projectability.

2.1 Grammaticization of nu, bekitsur, ke'ilu, and tov

I have argued that the non-lexical item nu entered Hebrew from Yiddish and Rus-
sian through the bilingual discourse strategy of language alternation (‘code-
switching’) at discourse markers (Brody 1987, Maschler 1988). Upon becoming a
Hebrew lexeme, however, it gained new meanings, different from those it had in
the languages of origin. From a discourse marker hastening non-verbal actions, I
suggested that it developed into a discourse marker urging further development
within a topic. From there it came also to function as granting permission to per-
form an action. Finally, through frequent employment in particular metalingual
utterances attempting to control the actions of the interlocutor, and therefore car-
rying an impolite aura in speakers’ metalinguistic awareness, it also came to have
keying functions in the discourse. Over time, these longer metalingual utterances
were no longer verbalized, and all that remained was the element carrying the key
– nu – from derogatory to joking. Figure 1 summarizes this process:
 Chapter 6.╇ Grammaticization from interaction 

hastening non-verbal actions

urging further development within a topic

granting permission to perform an action keying

Figure 1.╇ Hypothesized functional itinerary of nu

The grammaticization path of bekitsur is the most similar to the one described in
Traugott 1995a for adverbs developing into discourse markers. First, according to a
productive process in Hebrew grammar of deriving adverbs from nouns via prefixa-
tion with the preposition b(e)- (‘in’), bekitsur became a manner adverb (‘succinctly’):
Preposition (b(e)- ‘in’) + Verbal Noun (kitsur ‘shortening’)
Adverb (bekitsur).
In this process, it moved from a major category (verbal noun5) to an intermediate
category (adverb), in accord with Hopper and Traugott’s cline of de-categorializa-
tion (2003:€106–109):
major category (> adjective/adverb) > minor category.
Once an adverb, it followed the second half of this cline to become a discourse
marker (‘minor category’, Ferrara 1997). I have suggested that by association with
a hypothesized metalingual utterance such as 'im lomar bekitsur (‘if to say in short’,
‘said in short’, or ‘put succinctly’), it became a discourse marker summarizing lists
or episodes. Following the summary or conclusion of an episode, a previous topic
is often resumed. Over time, bekitsur came to be used as a resumptive discourse
marker. When resuming a previous topic, the immediately preceding discourse is
retroactively constructed as a digression and backgrounded, while future dis-
course is foregrounded. The function of backgrounding immediately-preceding
discourse by bekitsur was extended to contexts in which no real digression is ap-
parent, such as between narrative episodes. In the most grammaticized function of
bekitsur, its foregrounding function was extended to the foregrounding of new
narratives. Figure 2 summarizes this process schematically:

Adverb summarizing D. M.
resumptive D.M. foregrounding D.M.
Figure 2.╇ Hypothesized functional itinerary of bekitsur

5. For the path followed by the verbal noun kitsur, see Chapter 3, footnote 30.
 Metalanguage in Interaction

lexical meaning lingual realm metalingual realm

comparison and 1) hedge self-rephrasal

approximation 2) hedge quote
3) focus hedge
4) focus self-rephrasal
5) focus quote
6) irrealis self-rephrasal
Figure 3.╇ Hypothesized development of the pragmatic functions of ke'ilu

The self-rephrasal function of the discourse marker ke'ilu is arrived at via three
paths – two in the lingual realm, which originate mostly in the preposition of com-
parison and approximation k(e)- (‘as, like’), and a third in the metalingual realm,
which originates mostly from its irrealis component, the conditional 'ilu (‘if ’, irrea-
lis). For the lingual realm, we have the lexical meaning of comparison and ap-
proximation developing into the hedging use. A hedge often requires elaboration,
whence the self-rephrasal function (path (1) of Figure 3). Quotations are always an
approximation, thus the quotative use of ke'ilu also developed from the hedging
use (path (2)). From the comparative lexical element, the focus-marking function
can also be derived (Haiman 1988). A hedged utterance and a self-rephrased ut-
terance are generally also what is in focus (paths (3), (4)). A quote, too, is always
necessarily the focus of the utterance in which it occurs (Fleischman and Yaguello
2004). The quotative function is thus also a natural extension of the focus-marking
function (path (5)).
In the metalingual realm, the irrealis conditional 'ilu plays a significant role. Em-
ployed in the literal sense, this conjunction would generally require a full clause follow-
ing it. I have suggested that by association with a hypothesized metalingual utterance
such as ke'ilu 'amarti (‘as if I were saying’) or ke'ilu lomar (‘as if to say’), this conjunc-
tion developed as a discourse marker with a self-rephrasal function (path (6)).
Thus the self-rephrasal function of the discourse marker ke'ilu is arrived at by
three separate paths ((1), (4), and (6)), both in the lingual as well as in the meta-
lingual realms of discourse.
The most dialogical of all grammaticization paths in this study is exhibited by
tov. Like bekitsur, tov has probably developed from a major category (noun) to an
intermediate category (adjective (and, less frequently, adverb)), in accord with the
first half of Hopper and Traugott’s cline of de-categorialization (2003:€106–109):
major category (> adjective/adverb) > minor category.
From there, the adjective tov further developed into the minor category of interper-
sonal discourse marker – agreeing to some action – a usage attested already in
 Chapter 6.╇ Grammaticization from interaction 

Biblical Hebrew. I suggest that this has happened by association with a hypothe-
sized metalingual utterance such as ze tov velaxen 'ani maskim la'asot 'et ze (‘this
(the action you suggest) [is] good, and therefore I agree to do it’). Just as a recipient
can agree to some course of action, they can also agree to, or accept, a description
of some state of things. Hence, interpersonal tov functions also as acceptance of
some state of things. Frequently, the acceptance of the state of things is given in a
particular sequential position, known in the conversation analytic literature as
‘third turn receipt’. Thus, interpersonal tov functions also as acceptance of an action
requested by the speaker, such as an answer. Interpersonal tov can also be employed
by a recipient to pay lip service towards accepting some state of things. In this case
it is often followed by a ‘but’ response. This results in a fleeting concession of the
state of things as described by the interlocutor. I showed that this concessive tov
could be interpreted as ‘contrary to what you may assume’. From the concessive use,
tov of ironic agreement, or rather disagreement, is only a step away: only the first
half of the ‘yes, but’ response is verbalized, and it is done in a manner that makes it
clear that the ‘but’ to follow is so obvious that there is no need to verbalize it.
In all the above functions, discourse marker tov is employed by the recipient
and functions in the interpersonal realm of discourse. Before moving on to the
textual uses of tov, a word about the dialogical nature characterizing the gram-
maticization illustrated by the development into concessive tov. Here we see a clear
case of an interactional structure crystallizing into a grammatical one. I have ar-
gued (Chapter 5, excerpts 6 and 7) that this grammaticization path constitutes a
condensation of the following discourse structure:
A: assertion
B: agreement
stepping back from the agreement
In this discourse structure, B makes two moves – one agreeing with A’s assertion,
the other retracting his/her own previous agreement. We have seen that concessive
tov condenses these two moves into one. By employing a discourse marker essen-
tially meaning ‘contrary to what you may assume’ in this context, the speaker is
able to perform both the agreement and the retraction of it simultaneously. As
Auer writes,
[i]n some cases, the same linguistic element can either constitute an independent
action to be dealt with and responded to or be a grammatical element of a syntac-
tic construction. There is reason to believe that the second is a grammaticalized
version of the first (2005:€28).
 Metalanguage in Interaction

I have argued that concessive tov is indeed a grammaticization of this interac-

tional structure6. Auer (ibid.) discusses two additional cases of interactional pat-
terns having sedimented into grammatical ones: vocatives originating in sum-
mons-answer exchanges and conditionals originating in yes/no question-answer
exchanges in German.
A recipient can simultaneously provide some acknowledgment of his/her ac-
ceptance of the state of things, while at the same time initiating a move on to the
next action (Jefferson 1993 [1983]). This acknowledgement, however, does not
necessarily have to be directed towards some other interlocutor. It could take place
in some inner dialogue within the speaker concerning the state of things thus far
in the discourse. I have suggested that this is the phenomenon which is at the heart
of the further evolution of discourse marker tov into a textual discourse marker.
Thus, an inner metalingual utterance such as ze tov, velaxen 'ani maskim lehamshix
(‘this (referring to the state of things in the discourse) [is] good, and therefore I
agree to continue’) allows us to link the interpersonal and textual functions of this
discourse marker which operates in between realms. Textual tov tokens are gener-
ally employed by the speaker, not the recipient. As a textual discourse marker, tov
can thus be employed in order to mark expected transitions into beginnings of
major topics, episodes of a narrative, back to an interrupted action, or in order to
end a topic or an action.
A more elaborate dialogical grammaticization path is illustrated in the case of
concessive tov which at the same time functions also as transition into a new epi-
sode. I reproduce part of the example we have seen in Chapter 5 here:
Chapter 5, Excerpt 15 (‘Grandma Can’t Remember’) (in part):
13 Mother: ... ze ló kàl,
it’s not easy,
14 le'ishà mevugéret,
for woman old
for an older woman,
15 pit'òm la'avór dira.
to suddenly move apartments.
16 Father: naxón.
right.
17 Mother: .. tóv 'anaxnu 'azàrnu,
of course we helped,

6. Couper-Kuhlen and Thompson 2000 discuss a similar interactional concessive pattern for
English (2000: 382).
 Chapter 6.╇ Grammaticization from interaction 

18 veze ve
and so on and
This time, another interactional structure has crystallized into a grammatical one.
I have argued that this grammaticization is a condensation of the following imag-
ined interactional structure in the speaker’s consciousness:
A: assertion
(Lines 13–15: ‘it’s not easy for an older woman to suddenly move apart-
ments’)
A: imagining a counter-argument of some voice to this assertion
(‘it is not easy for an older woman to suddenly move apartments, particularly
with no help’. More specifically: ‘you didn’t help’)
A: addressing the imagined counter-argument with a ‘yes, but’ strategy, i.e., disa-
greeing with it
(Line 17: ‘yes, you may assume we did not help, but we did’, or ‘contrary to
what you may assume, we helped’)
A: transition into new episode
(Line 18)
Speaker A makes an assertion. He or she then imagines a possible counter-argu-
ment to this assertion and addresses it in a ‘yes, but’ strategy. By employing con-
cessive tov (i.e., ‘contrary to what you may assume’) preceding this ‘yes, but’ disa-
greement, the speaker is able to both allude to the imaginary counter-argument
and disagree with it simultaneously. Once this has been resolved, the speaker con-
tinues on with the expected next course of action – the continuation of the narra-
tive. Now we see concessive tov condensing two other moves in an interactional
structure (imaginary counter-argument + disagreeing with it) into a single one.
However, if we take into account the fact that the second move, the disagreement,
is in itself a case of the previously illustrated condensation into concessive tov of
the interactional structure of (1) agreement and (2) stepping back from it, the
present case actually consists of a condensation of a three-step interactional struc-
ture: (1) alluding to an imaginary counter-argument, (2) agreeing with it, and (3)
stepping back from the agreement. Another interactional structure – one involv-
ing three steps – has thus crystallized into a grammatical one.
Hopper writes that grammatical constructions are open in two ways. The most
obvious one is that “they contain open slots into which lexical items can be in-
serted with a greater or lesser degree of freedom” (2004:€172). The more important
way in which constructions are open stems from the fact that “their structure is
emergent, that is to say, their structure never reaches a point of closure and com-
pletion as a construction” (ibid.). The examples concerning concessive tov above
illustrate the discourse motivation behind this second type of openness: as a
 Metalanguage in Interaction

Adjective agreement to some course of action

acceptance of description of some state of things

concessive acceptance/concession + transition into

next expected course of action

ironic agreement (disagreement) transition into next expected course of

action
Figure 4.╇ Hypothesized functional itinerary of tov

construction is used in various interactional exchanges among speakers, new ex-

changes may crystallize and become condensed into that construction. In this
manner, a construction that initially condensed two actions into one may come to
condense an additional third action (as illustrated above), or it may come to con-
dense a new interactional structure altogether. Thus, the nature of constructions is
open and continually changing as a function of the interactions in which those
constructions participate. In the case of the discourse marker construction ‘con-
cessive tov’, the change in the interactional exchange condensed by the construc-
tion has not resulted in a structural change in the construction itself (as is some-
times the case with the English take NP and construction explored in Hopper
2004). This, however, is not always the case with discourse markers. We shall re-
turn to this point in Section 2.3.
Figure 4 summarizes the grammaticization path of tov.

2.2 Common threads and differences

among the grammaticization patterns studied

Onodera concludes her recent book about grammaticization of Japanese discourse

markers noting two clines of grammaticization of English discourse markers sug-
gested in earlier research:
1. Clause internal Adv > Sentence Adv > Discourse Marker (Traugott 1995a)
2. Matrix clause [subject + verb] > Parenthetical adjunct (Brinton 2001)
She hypothesizes that “it is plausible that there are plural clines of grammaticaliza-
tion to be proposed in addition to the traditional view of its unidirectional tenden-
cies” (2004:€203). The paths suggested above for nu, bekitsur, ke'ilu, and tov suggest
that this indeed is the case. While bekitsur exhibits linearity (Figure 2) somewhat
similar to the first cline above, nu, ke'ilu, and tov exhibit grammaticization pat-
terns which involve branching (figures 1, 3, 4) and represent the multiple direc-
tions the utterance has taken as a function of the recurrent contexts in which it was
 Chapter 6.╇ Grammaticization from interaction 

employed. For example, sequential nu preceding particular metalingual utterances

developed into keying nu, while in other environments it developed additional
sequential functions along the continuum of ‘degree of encouragement to proceed
with an action’ (Figure 1).
Another difference is that while Traugott 1995a and Brinton 2001 deal with
the ways a non discourse-marker becomes a discourse marker, the patterns dis-
cussed in the current study focus in particular on the ways an utterance which has
already become a discourse marker continues to develop additional discourse
marking functions.
The discourse markers studied here follow different functional itineraries, but
their development is characterized by common threads that exhibit many of the
features associated with grammaticization:
Semantic loss
Apart from the non-lexical item nu, which had little semantic substance to begin
with, the remaining three utterances, bekitsur, ke'ilu, and tov, are all semantically
bleached when used as discourse markers.
Phonological reduction
Phonological reduction was shown particularly for bekitsur, but some tokens of
ke'ilu also manifest this feature. Not surprisingly, no tokens manifesting phono-
logical reduction were found for tov and nu, both single syllable items. However,
the playful forms of bekitsur (bekítskets, hakítskets, 'akibitser) and of tov (to--'ov,
see Chapter 5, footnote 10) actually show phonological expansion.
De-categorialization
A grammatical category – adverb in the case of bekitsur, adjective in the case of tov,
and conjunction in the case of ke'ilu – “tends to lose the morphological and syntactic
properties that would identify it as a full member” (Hopper and Traugott 2003:€107)
of that grammatical category (Hopper and Thompson 1984, Hopper 1991). The cline
from major to minor category (Hopper and Traugott 2003) is followed for the two
discourse markers originating in the intermediate categories: bekitsur (adverb) and
tov (adjective). Discourse marker ke'ilu too shows fewer of the syntactic properties
that characterize the conjunction ke'ilu (e.g., the lack of a following subordinate
clause). For the non-lexical item nu, which could perhaps be considered an interjec-
tion, there is not much sense in discussing de-categorialization.
Bonding within the phrase
In the case of the two utterances which consist of more than one lexical element
– bekitsur (preposition + verbal noun) and ke'ilu (preposition + conjunction) –
there is a fixation accompanying their de-categorialization, such that the preposi-
tion opening each is no longer perceived as a separate entity. Furthermore, if we
 Metalanguage in Interaction

consider the metalingual utterances posited as the utterances from which the dis-
course markers bekitsur, ke'ilu, tov and keying nu have originated (Section 1.2),
bonding within them has become so strong that it is no longer necessary to verbal-
ize the entire utterance.
In a study of grammaticization of constructions, Bybee cites Haiman (1994) who
views the process of grammaticization as ritualization resulting from repetition:
[R]epetition leads to the automatization of a sequence of units, and the reanalysis
of the sequence as a single processing chunk […R]epetition also leads to the re-
duction of form through the weakening of the individual gestures comprising the
act, and through the reorganization of a series of formerly separate gestures into
one automated unit; and emancipation occurs as the original, more instrumental
function of the practice gives way to a more symbolic function inferred from the
context in which it occurs (Bybee 2003:€603).

I suggest that, through repetition, each of the longer metalingual utterances pos-
ited as the constructions from which the discourse markers have originated has
reduced in form and reorganized from a series of separate gestures into one auto-
mated unit. In the process of this reorganization into one automated unit, it was no
longer necessary to verbalize each of the individual gestures comprising the meta-
lingual utterance. As Couper-Kuhlen and Thompson write, “under certain condi-
tions, [… a move] can be so strongly projected […] that its explicit expression is
made redundant” (2000:€397). Not only moves can be strongly projected, but also
parts of constructions. The result is a situation in which verbalizing the key com-
ponent of the construction – i.e., nu, bekitsur, ke'ilu, or tov – is sufficient in order
to activate the entire metalingual utterance.
Note the different type of projection considered here. This is not the interac-
tional projection of a certain type of frame shift, but rather the grammatical pro-
jection of the remainder of the metalingual utterance announcing it. As the title of
Auer’s 2005 study (“Projection in interaction and projection in grammar”) indi-
cates, these are two different types of projection, but there is a very tight connec-
tion between them.
There is, however, a difference between this type of syntactic projection and the
syntactic projection discussed in Auer 2005. Discourse markers crystallized from
such metalingual utterances are sometimes the final element of the metalingual ut-
terance posited as their origin. This is not the case for nu and ke'ilu, but it is the case
for bekitsur and tov: 'im lomar bekitsur (‘if to say in short’), ze tov (‘this [is] good’). In
such cases, we are concerned with ‘backwards’, or retrospective projection.
Generalization of meaning
For all four utterances, meaning generalizes in the sense that “more and more
domains (polysemies) become available” (Traugott 1995a: 14). We have seen that
 Chapter 6.╇ Grammaticization from interaction 

this is clearly the case for the adverb bekitsur, for the adjective tov, and for the
conjunction ke'ilu, which no longer function in these semantic-syntactic roles as
they become discourse markers, but it is also the case for non-lexical nu, origi-
nally associated with hastening and impoliteness, which has acquired keying func-
tions in the discourse.
Pragmatic strengthening
All four utterances increase in their pragmatic functions. They all move from be-
ing more to less referential and they illustrate a “movement away from [their]
original specific and concrete reference […] toward increasingly general and ab-
stract reference” (Pagliuca 1994: ix, cited in Traugott 1995a). In the case of non-
lexical nu, it is of course not possible to speak of an original concrete reference;
however, even nu exhibits development away from hastening actions in the extra-
lingual realm to hastening metalingual utterances. Nu’s acquisition of a keying
function also represents a movement away from the referential, because it con-
structs the tenor of the relationship among participants in the world of the interac-
tion rather than referentially.
Most crucially, all four discourse markers show an increase in their metalin-
gual function. As forms become increasingly grammaticized, there is “a tendency
toward metatextual meaning, or more specifically a shift from […] ‘the world be-
ing talked about’ to ‘the speaker’s organization of that world in the act of speak-
ing’” (Traugott and Dasher 2002:€40). This involves a metaphorical shift from the
propositional to the ‘metalinguistic’ (Traugott 1988, Traugott and König 1991) fol-
lowing the non-metatextual > metatextual tendency (Traugott 1995a, Traugott
2001, Traugott and Dasher 2002:€ 40), or, as Dasher put it, speakers move from
encoding “elements of the referenced event independently of the speech event” to
encoding “features of the speech event independently of the referenced event”
(1995:€266–271).
That the metalingual requirement is satisfied by all the discourse markers in-
vestigated here is no surprise. This is so by nature of my definition of discourse
markers. Only utterances employed for the purpose of referring to the text itself,
to the interaction among its participants, and/or to the cognitive processes taking
place in their minds during verbalization were considered in the first place. Even
so, we have seen that for each of the four clines, the endpoints of the paths fol-
lowed by each marker are all more metalingual than the initial stages. For nu, this
was discussed in the paragraph before last. For bekitsur, foregrounding a new nar-
rative when no real digression is apparent is the most metalingual (and least trans-
parent) of all its functions. For ke'ilu, its metalingual nature is most prominent in
its self-rephrasal function. Finally, tov marking expected transitions in a text is
 Metalanguage in Interaction

more metalingual than tov signaling agreement and acceptance, because these lat-
ter two involve acceptance of some state of things in the extralingual world.
In general one might say that those functions originating in some posited me-
talingual utterance are all more metalingual than those not originating in such
utterances.
Subjectification and intersubjectification
All four discourse markers have become “increasingly based in the speaker’s sub-
jective belief/state/attitude toward the proposition” (Traugott 1989:€35), and then
more intersubjective, i.e., more concerned with the ‘self ’ of the addressee (Traugott
2003a). Of the four discourse markers investigated here, subjectification and inter-
subjectification are perhaps most prominent in the path followed by tov. It was
argued that its grammaticization path to becoming a textual discourse marker of
expected transition is a residue of an inner dialogue concerning the state of things
in the discourse thus far. Similarly, concessive tov exhibits the speaker’s concern
with the ‘self ’ of the addressee to the point of constructing his or her voice and
addresing it (Chapter 5, excerpt 15). Such dialogicity in grammaticization shows
maximal attention to the ‘self ’ of the (imagined or actual) addressee.
However, the discourse markers nu, bekitsur, and ke'ilu also exhibit increasing
subjectification and intersubjectification in their paths of development. The end-
points of the paths followed by each marker are all more subjective and intersub-
jective than the initial stages: keying nu – in that it introduces the speaker’s joking
or provoking stance towards the addressee; foregrounding bekitsur – in that it sig-
nals what the speaker wishes the addressee to put in the foreground at that par-
ticular moment in the discourse; and self-rephrasal ke'ilu – in that it spells out
what the speaker him/herself meant in the preceding utterance, under his or her
assumption that it was not sufficiently clear for the addressee.
Retention / Persistence
Finally, another common thread running through the studies of the four discourse
markers is the role of the lexical item to be grammaticized. Bybee and Pagliuca
talked about retention here, according to which older meanings of a form “glim-
mer through” as they are retained from their original lexical sources (1987:€109).
Hopper later refined this into the persistence principle of grammaticization:
When a form undergoes grammaticalization from a lexical to a grammatical func-
tion, so long as it is grammatically viable some traces of its original lexical mean-
ings tend to adhere to it, and details of its lexical history may be reflected in con-
straints on its grammatical distribution (1991:€22).

We have seen that, apart from the non-lexical item nu, for which it is irrelevant to
discuss the original lexical source here because it came from another language, the
 Chapter 6.╇ Grammaticization from interaction 

remaining three functional itineraries are strongly influenced by lexical history:

the concept of shortness is crucial for the functional itinerary of bekitsur, as are the
comparison/approximation + irrealis conditional components for the course ke'ilu
has taken. Finally, for the path followed by tov, the lexical meaning of the adjective
‘good’ is pivotal.
The metalingual utterances posited as the origins of these discourse markers
offer a clue as to why older meanings of a form might “glimmer through”. This is a
result of the fact that in the posited metalingual utterances, these items actually do
function in their original lexical meaning: in 'im lomar bekitsur (‘if to say suc-
cinctly’), bekitsur is in fact the adverb bekitsur (‘succinctly’); in ke'ilu lomar (‘as if
to say’), ke'ilu does function as the preposition of comparison + irrealis conjunc-
tion; and in ze tov (‘this [is] good’), tov is actually an adjective meaning ‘good’. It is
no surprise, therefore, that details of their lexical history are reflected when these
lexemes function as discourse markers.

2.3 Projectability and grammaticization of constructions

Bybee argues for taking the word ‘constructions’ in the definition of grammaticiza-
tion seriously: “In fact, it may be more accurate to say that a construction with
particular lexical items in it becomes grammaticized, instead of saying that a lexi-
cal item becomes grammaticized” (2003:€602). She ends her study of mechanisms
of change in grammaticization with the following statement:
Repetition alone, however, cannot account for the universals of grammaticization.
[…] It is not just the fact of repetition that is important, but in addition what is
repeated that determines the universal paths. The explanation for the content of
what is repeated requires reference to the kinds of things human beings talk about
and the way they choose to structure their communications (ibid.: 622).

One of the things human beings talk about are the frame shifts in which they are
about to engage in the course of interaction. And they do so with constructions
that over time are repeated and eventually grammaticized. The present study con-
cerned the grammaticization of such constructions into discourse markers.
Grammaticization of constructions has been examined from both diachronic
and synchronic perspectives. Underlying Traugott’s diachronic approach to the
study of grammaticization (Chapter 1, Section 8.1) is the Invited Inferencing The-
ory of Semantic Change (IITSC) (Traugott 1999), accounting for the ways conver-
sational implicatures (Grice 1975), or “invited inferences” (Geis and Zwicky 1971),
become conventionalized based on processes of metaphor and metonymy in com-
munication. This process was described by Dahl:
 Metalanguage in Interaction

if some condition happens to be fulfilled frequently when a certain category is

used, a stronger association may develop between the condition and the category
in such a way that the condition comes to be understood as an integral part of the
meaning of the category (1985:€11).

If inferences occur frequently, they come to play a significant role in grammatici-

zation. “In early stages of grammaticalization conversational implicatures fre-
quently become ‘semanticized’ […], that is, become part of the semantic polysem-
ies of a form” (Hopper and Traugott 2003:€82). For example, Hopper and Traugott
show that inferences of causality from temporal sequence (“the well known logical
fallacy […] characterized as […] ‘after this, therefore because of this’” (ibid.)) are
the basic process underlying grammaticization of Present Day English since (which
could be either temporal or causal) from the Old English connective siþþan, ‘from
the time that, after’ (which could only be temporal) (ibid.). Thus, a conventional-
izing of a conversational implicature of cause has occurred.
Another example is Aijmer’s (1985:€13) of the development of first person voli-
tional ‘will’ to pure future: “If the speaker is willing to do something, it follows con-
versationally that he intends to do it and that the future action will take place”.
Traugott explains that conversational implicatures may come to be used stra-
tegically over time; that is, speakers may invite addressees to “let implicatures go
through” (2003b: 634).
Invited inferencing is a kind of conceptual metonymy within the speech chain […]
since it is primarily associative in character, being derived from the uses to which
interlocutors put linear sequences of utterances and associations in context (ibid.).

Another, more synchronic way of looking at these linear sequences concerns the
concept of projectability. Frequently occurring linear sequences give rise to the
phenomenon of projectability. Once the first component of a frequent linear se-
quence has been verbalized, it makes a projection concerning the nature of the
following component.
In a recent study of projectability and clause combining in interaction, Hopper
and Thompson 2008 examine three constructions from a synchronic perspective
– English wh-cleft (pseudocleft)7, English extraposition, and German Wenn-claus-
es (Günthner 1999b), which in English translate into either when- or if-clauses.
Much like the analysis suggested here of the grammaticization of concessive tov
from two- and three-step interactional exchange structures (Section 2.1), and like
Auer’s analysis of the grammaticization of vocatives from summons-answer ex-
changes or of conditionals from yes/no question-answer exchanges (2005), Hopper

7. The pseudocleft construction has been studied from this perspective already in Hopper
2001.
 Chapter 6.╇ Grammaticization from interaction 

and Thompson argue that each of the three syntactic constructions they are ana-
lyzing “can be understood as a syntacticized version of an interactional pattern”
(2008: 13). As such, they carry a strong projection “whereby the first part [i.e., the
pseudocleft, extraposition, or Wenn-clause] projects not another clause, as has be-
come the norm in more formal varieties of linguistic communication, but a com-
plex of one or more social actions which is typically manifested as a span of talk
of indeterminate length” (ibid., emphasis mine).
Hopper and Thompson are in fact engaged with the two types of projection –
grammatical and interactional – suggested by Auer (2005). Whereas traditional
grammatical approaches focus on grammar and on the fact that the first part
(pseudocleft, extraposition, Wenn-clause) grammatically projects another clause,
Hopper and Thompson suggest that a better way to understand the phenomenon
is to view the first part as interactionally projecting one or more social actions.
In the present study we have been concerned with a particular type of social
action – frame shifting. From the shifting into a new narrative (e.g., foregrounding
bekitsur) to the shifting into a disagreement with a previous assertion (e.g., tov of
ironic agreement), we have seen that the social action of frame shifting is quite
variable and has a much wider range than the social actions projected by Hopper
and Thompson’s three projecting constructions above. Nevertheless, frame shift-
ing is a particular social action which, as Goffman (1981) has taught us, speakers
continually engage in, and which, as illustrated here, speakers often announce in
some way, in other words – project. There is a particular type of ‘first parts’ em-
ployed to project frame shifts – the longer metalingual constructions we have been
considering in the present study. I would like to suggest that, like Hopper and
Thompson’s syntacticized constructions, discourse markers, too, have grammati-
cized out of such ‘first parts’ which strongly project other social actions to come.
Discourse markers are involved in two parallel crystallizations, then; one con-
cerned with actions, the other with structure. On the one hand, they comprise a
crystallization of an interactional structure – announcing the upcoming frame
shift (in a necessarily metalingual utterance) + continuing the discourse in some
particular way – into a single move. (Recall the discussion of concessive tov (Sec-
tion 2.1) condensing two or three moves in an interactional structure into a single
one). On the other hand, they comprise a crystallization of a frequently employed
metalingual utterance (announcing a particular frame shift) into a fixed utterance
(e.g., 'im lomar bekitsur ‘if to say succinctly’ > bekitsur).
There are three differences between the constructions discussed in Hopper
and Thompson and the discourse markers discussed here. To begin with, the ‘first
parts’ projecting social actions to come discussed here are always metalingual.
Second, the projection discussed here is far more variable because it encompasses
a whole range of frame shifting phenomena. Third, the projection a particular
 Metalanguage in Interaction

discourse marker carries is usually also more specific than the projections carried
by the constructions considered by Hopper and Thompson. The nature of the
projected frame shift is commented on in the metalingual utterance, be it a full-
fledged one or a crystallized version.
However, there is evidence to support the suggestion that the phenomenon
discussed in Hopper and Thompson and the one studied here are essentially the
same. A recent study of the N-be-that (die Sache ist ‘the thing is’) construction in
German (Günthner 2007) shows that this matrix clause is positioned in the pre-
front field, a position whose function is, according to Auer 1996, “to frame (often
metapragmatically) the subsequent utterance” (Günthner 2007:€10):
This bi-part division between a framing element (positioned in the pre-front field)
[…] and a following, syntactically somewhat independent, clause […] shows for-
mal and functional parallels with other elements in spoken German which are
regularly positioned in the pre-front field, such as ‘topic formulas’ […] and dis-
course markers. (ibid.:11, emphasis mine).

Thus, the German N-be-that construction is reanalysed as a framing device, rather

than as the main clause of a bi-clausal construction, and it is shown to project the
upcoming stretch of discourse. Likewise, we have seen throughout this study that
discourse markers are framing devices projecting the upcoming discourse and that
the ensuing discourse is syntactically independent of them. Günthner continues,
these phrases become routinized and formulaic and are highly skewed toward
certain communicative functions: they project and defer an upcoming segment of
discourse, focusing the recipients’ attention to the following segment of discourse.
(ibid.: 15).

In spite of this routinization, German die Sache ist (‘the thing is’) does not show
any reduction in form (Günthner, p.c.). In informal spoken English, however, we
do find reduction into a shorter version of the ‘equivalent’ construction (the form
thing is8, see Aijmer 2007). This is similar to the reduction in form we have been
concerned with in the case of discourse markers. Unlike the construction studied
by Günthner, we have seen that discourse markers have additional, more specific
roles besides deferring an upcoming segment of discourse and, with the exception
of ke'ilu, they do not generally engage in focusing phenomena.
Günthner shows that the further grammaticized cases of the German N-be-that
construction are followed not by a subordinate clause (as traditional approaches,
which regard such constructions as [matrix clause + complement clause] structures,

8. This is particularly common among younger speakers. I recently heard this reduced con-
struction even in scripted discourse -- the animated film ‘Ratatouille’, in which the animals are
presented as fast moving, quick talking, and young.
 Chapter 6.╇ Grammaticization from interaction 

would lead us to expect), but rather by a complex stretch of discourse (e.g., a nar-
rative, as in the example she provides). Similarly, no subordination is involved in
the case of our discourse markers. The stretch of discourse following a discourse
marker is not subordinate to the discourse marker. Neither is it subordinate to the
metalingual utterance from which it is said to have developed. However, we often
find related constructions in the discourse which do show subordination. In the
case of tov, for example, we often find the construction tov she-… (‘[it’s] good
that…’), where the part following the complementizer she- is traditionally ana-
lyzed as the complement subject clause of the adjectival predicate tov. One exam-
ple from our corpus is the following: tov she'asita halixa hayom (‘[it’s] good that
you did a walk today’). It is not difficult to imagine how in some, more dialogical
contexts, a simple ze tov (‘this [is] good’) would be employed, with the demonstra-
tive ze (‘this’) anaphorically referring to an earlier assertion such as 'asiti halixa
hayom (‘I did a walk today’). And from here, the path to a simple tov, as we have
seen, is direct. Thus, although discourse markers involve parataxis, there are re-
lated constructions which involve hypotaxis, as Günthner has shown for the
N-be-that construction.
Hopper and Thompson write that these projecting constructions are best seen
as “relatively open-slot ‘prefabs’ […] which strongly project certain types of social
actions” (ibid.: 13). A discourse marker is a projecting construction that has devel-
oped in a way that locates it on the rather closed end of these relatively open-slot
‘prefabs’. We have seen that in the process of grammaticization, it has arrived at a
relatively high degree of crystallization and reduction in form. At the same time,
however, its structure may continue to evolve as new interactional exchanges are
crystallized and condensed into it.
In this way, phenomena of grammatical, and particularly interactional projec-
tion shed light on the crystallization of metalingual utterances into the discourse
markers permeating Hebrew talk and constructing the frame shifts into the new
actions taking place throughout interaction.

3. Interacting as an Israeli

I have written in the preface that one of my interests in writing this book has been
to explore the essence of interacting as an Israeli. By studying four particular ut-
terances employed to shift frames in interaction, a particular profile of Israeliness
is revealed.
The most frequently employed discourse marker explored here (Chapter 1,
Table 3) is nu. The study of nu has taught us that Israelis converse rather impa-
tiently, but that they do not necessarily view this impatience negatively. On the
 Metalanguage in Interaction

contrary, the impatience manifested by interlocutors in the course of their hasten-

ing a narrator to move towards the climax of a story, for instance, is indicative of
their high degree of involvement in the story, which is seen as a particularly posi-
tive feature. Addressees take the liberty of controlling the speed of another’s flow
of discourse in the name of exhibiting curiosity, enthusiasm, and involvement in
the other’s talk. We saw also that employment of nu involves culture-specific ex-
pectations that the non-primary speaker will engage in a relatively high degree of
interaction (Clancy et al. 1996) in Hebrew discourse. These findings are in line
with previous characterizations of the Israeli conversational style as high in degree
of interpersonal involvement (Katriel 1986, Blum-Kulka and Katriel 1991, Masch-
ler 1994a, Blum-Kulka 1997).
The closeness among participants often serves to neutralize the impolite aura
associated with nu (at least in participants’ metalinguistic awareness) and gives
rise to a humorous, self-mocking interactional key in which the hastener relates to
his or her supposedly impolite behavior jokingly. In this culture, the boundary
between joking and provoking, as illustrated by the use of nu, has been shown to
be somewhat fuzzy.
The second most frequently employed discourse marker explored here is tov.
Katriel has written much about the high tolerance for disagreement in Israeli cul-
ture (1986, 2004). I have suggested that, like the Brazilian thumbs-up gesture
meaning ‘good’ but also functioning as “an interactional link between moves,
units, or moments within an interaction” (Sherzer 1991:€192), which “has the pow-
er of disarming disagreement” in a society profuse with brutal conflicts (ibid.:
195–196), the pervasiveness of tov in casual Hebrew conversation may be viewed
as compensation for the high degree of disagreement imbuing the culture. Thus, to
interact Israeli-style is also to show a high tolerance for disagreement, while at the
same time maintaining an amiable tone when possible. Concessive tov with its
‘yes, but’ structure is instructive here.
The grammaticization of the resumptive discourse marker bekitsur from an
adverb meaning succinctly or in short is quite telling with regard to the Israeli in-
teractional mode. We have seen that it is not self-evident that a resumptive dis-
course marker would be associated with the concept of shortness (English anyway,
for example, is not), particularly when other markers such as bexol 'ofen (‘in any
way/manner’) and bexol mikre (‘in any event’) are possible. These latter two are not
employed very frequently in this corpus. Israelis strongly prefer bekitsur, an ex-
pression which often also implies some degree of reproach that one is not speaking
tersely and to the point. As Katriel has taught us (1986, 2004), the new way of
speaking endorsed for the “New Jew”, according to the Sabra ethos, is not only
dugri (direct and straightforward), but also concise and ‘to the point’ (katsar
 Chapter 6.╇ Grammaticization from interaction 

vela'inyan). In this conversational style, a digression will be perceived as an object

of reproach.
The study of bekitsur has also highlighted the collaborative nature of Hebrew
discourse. Many more digressions preceding bekitsur are listener-initiated in com-
parison to the case of the English ‘equivalent’ anyway (Ferrara 1997). However,
because of the high degree of involvement on both sides, the question of who initi-
ated the digression quickly becomes irrelevant. Participants also collaborate high-
ly in returning to the main line of discourse. We have seen that about a fifth of
resumptive bekitsur tokens and bekitsur tokens introducing new narratives follow
a token of nu uttered by the listener. Thus digressions, as well as the return from
them, were shown to be a highly cooperative venture that further ties participants
to each other in this high-involvement style (Tannen 1984) of discourse.
Another motivation for joking, besides the one discussed in relation to nu, is
exhibited by some uses of bekitsur. In returning from a digression, particularly one
initiated by the listener, the speaker signals that the preceding discourse was of
secondary importance, a potentially face-threatening act for the listener. As in the
case of nu, a potentially face-threatening strategy is compensated for by a jocular
key that often characterizes the Israeli mode of casual interaction. Considering
that tov, too, can be employed as compensation for a pervasive face-threatening
act (disagreement), we see that three of the discourse markers studied here involve
an inherent contradiction in the Israeli casual interactional mode. This contradic-
tion is reminiscent of the mythic Sabra image ‘thorny outside but sweet inside’
(Oring 1981, cited in Katriel 2004:€139). One might say that thorny behavior is
compensated for by sweet, or at least jocular modes of interaction: joking nu, tov
of acceptance and agreement, and jocular bekitsur. Note, however, that if we take
the substantive component (the ‘inside’) to be the thorny behavior, and the joking
to be its external embellishment, as I would be inclined to do here, then applying
the Sabra metaphor involves a reversal of interior and exterior.
By the beginning of the 1980s, along with the decline of the dugri mode of speak-
ing, we find a refusal to accept the Sabra myth at face value, but also a reluctance to
give it up entirely (Katriel 2004). This change has become more definite by the late
1990s, by which time we see it grammaticized into the language. I have argued that
the great proliferation of hedging and rephrasing via ke'ilu and kaze, particularly
towards the beginning of the 21st century, is a linguistic manifestation of the decline
of the dugri interactional style. It mirrors the softening of the dugri mode of speaking
in a society which has become much more heterogeneous and cognizant of the in-
terpersonal costs of the dugri speaking style (Katriel 2004). I have suggested that this
change was influenced not only by internal change in the culture, but also by the
saturation of society by multiple voices (Gergen 1991) which characterizes the post-
modern era, a phenomenon responsible for the great proliferation recently of
 Metalanguage in Interaction

‘equivalents’ of ke'ilu and kaze in many of the world’s languages (Fleischman and
Yaguello 1999, 2004). Thus, just as grammar is sensitive to competing motivations
deriving from both internal and external forces (Du Bois 1985), so culture too
changes as a function of influences from both within and without.
Interacting as an Israeli, according to the study of Hebrew discourse markers,
implies on the one hand retaining many of the characteristics of the dugri speak-
ing style, along with a certain amount of jocular linguistic behavior compensating
for them. On the other hand, interacting Israeli-style has recently come to imply
also being aware of the multiplicity of voices and truths, and adopting therefore a
softer, less decided and self-assured mode of conduct. This results in allowing for
hedgings and rephrasals and in a new tolerance for tentativeness. Grammatical
and cultural processes are indeed seen to occur in parallel, as grammaticization
patterns of a language mirror the processes of structuration (Giddens 1984, Hop-
per 1998) undergone by the culture which binds its speakers together.
 Bibliography

Abraham, Werner. 1991. The grammaticization of German modal particles. In: Elizabeth Closs
Traugott and Bernd Heine (eds.), Approaches to Grammaticalization, vol. II. Amsterdam/
Philadelphia: John Benjamins. 331–380.
Aijmer, Karin. 1985. The semantic development of ‘will’. In: Jacek Fisiak (ed.), Historical Seman-
tics, Historical Word-Formation. New York: Mouton. 11–21.
Aijmer, Karin. 1986. Why is actually so popular in spoken English? In: Tottie Gunnel and Ingeg-
erd Bäcklund (eds.), English in Speech and Writing. A Symposium. Uppsala: Almquist &
Wiskell. 119–129.
Aijmer, Karin. 1996. I think – an English modal particle. In: Toril Swan and Olaf Jansen Westvik
(eds.), Modality in Germanic Languages. Berlin: Mouton de Gruyter. 1–47.
Aijmer, Karin. 2002. English Discourse Particles: Evidence From a Corpus. Amsterdam/Philadel-
phia: John Benjamins.
Aijmer, Karin. 2007. The interface between discourse and grammar: The fact is that. In: Agnès
Celle and Ruth Huart (eds.), Connectives as Discourse Landmarks. Amsterdam/Philadel-
phia: John Benjamins. 31–46.
Alfonzetti, Giovanna, 1998. The conversational dimension in Italian-dialect code-switching. In:
Peter Auer (ed.), Code-Switching in Conversation. London: Routledge. 180–211.
Altenberg, Bengt. 1986. Contrastive linking in spoken and written English. In: Tottie Gunnel
and Ingegerd Bäcklund (eds.), English in Speech and Writing. A Symposium. Uppsala:
Almquist & Wiskell. 13–40.
Andersen, Gisle. 1998. The pragmatic marker like from a Relevance-Theoretic perspective. In:
Andreas H. Jucker and Yael Ziv (eds.), Discourse Markers: Descriptions and Theory. Amster-
dam/Philadelphia: John Benjamins. 147–170.
Ariel, Mira. 1998. Discourse markers and form-function correlations. In: Andreas H. Jucker and
Yael Ziv (eds.), Discourse Markers: Descriptions and Theory. Amsterdam/Philadelphia: John
Benjamins. 223–259.
Ariel, Mira. 1999. Cognitive universals and linguistic conventions: The case of resumptive pro-
nouns. Studies in Language 23: 217–269.
Auer, Peter. 1995. The pragmatics of code-switching: A sequential approach. In: Lesley Milroy
and Peter Muysken (eds.), One Speaker, Two Languages: Cross-Disciplinary Perspectives on
Codeswitching. Cambridge: Cambridge University Press. 115–135.
Auer, Peter. 1996. The pre-front field in spoken German and its relevance as a grammaticaliza-
tion position. Pragmatics 6: 295–322.
Auer, Peter. 2000. Pre- and post-positioning of wenn-clauses in spoken and written German. In:
Elizabeth Couper-Kuhlen and Bernd Kortmann (eds.), Cause, Condition, Concession, and
Contrast: Cognitive and Discourse Perspectives. Berlin: Mouton de Gruyter. 173–204.
Auer, Peter. 2005. Projection in interaction and projection in grammar. Text 25: 7–36.
Avneyon, Eitan. 1998. Milon Sapir Entsiklopedi (‘The Encyclopedic Sapphire Dictionary’). Or
Yehuda: Hed 'Artsi Publishing Ltd./ Eitav Publishing House Ltd.
 Metalanguage in Interaction

Bakhtin, Mikhail M. 1981. The Dialogic Imagination. Trans. by Carl Emerson and Michael
Holquist, Michael Holquist (ed.). Austin: University of Texas Press.
Bakhtin, Mikhail M. 1986. Speech Genres and Other Late Essays. Trans. by Vern W. McGee, Carl
Emerson, and Michael Holquist (eds.). Austin: University of Texas Press.
Basgoz, Ilhan. 1986. Digressions in oral narrative: A case study of individual remarks by Turkish
romance tellers. Journal of American Folklore 99: 5–23.
Bateson, Gregory. 1972 [1956]. A theory of play and fantasy. In: Steps to an Ecology of Mind. New
York: Ballantine. 177–193.
Bazzanella, Carla. 2006. Discourse markers in Italian: “Compositional” meaning. In: Kristine
Fischer (ed.), Approaches to Discourse Particles. North Holland: Elsevier. 449–464.
Beach, Wayne A. 1990. Language as and in technology: Facilitating topic organization in a vide-
otex focus group meeting. In: Martin J. Medhurst, Alberto Gonzales, and Tarla Rai Peter-
son (eds.), Communication and the Culture of Technology. Pullman, Wash.: Washington
State University Press. 197–219.
Beach, Wayne A. 1993. Transitional regularities for ‘casual’ “Okay” usages. Journal of Pragmatics
19: 325–352.
Beach, Wayne A. 1995. Preserving and constraining options: “Okays” and ‘official’ priorities in
medical interviews. In: G. H. Morris and Ronald J. Chenail (eds.), The Talk of the Clinic:
Explorations in the Analysis of Medical and Therapeutic Discourse. Hillsdale, N.J.: Lawrence
Erlbaum. 259–289.
Becker, Alton L. 1979. Text-building, epistemology, and esthetics in Javanese shadow theater. In:
Alton L. Becker and Aram Yengoyan (eds.), The Imagination of Reality. Norwood, N.J.:
Ablex. 211–43. Reprinted in Becker 1995.
Becker, Alton L. 1982. Beyond translation: Esthetics and language description. In: Heidi Byrnes
(ed.), Georgetown Univeristy Round Table on Languages and Linguistics (GURT) 1982: Con-
temporary Perceptions of Language: Interdisciplinary Dimensions. Washington, D.C.: Geor-
getown University Press. 124–38. Reprinted in Becker 1995.
Becker, Alton L. 1984. The linguistics of particularity: Interpreting superordination in a Javanese
text. Proceedings of the Tenth Annual Meeting of the Berkeley Linguistics Society 10: 425–36.
Berkeley: Berkeley Linguistics Society. Reprinted in Becker 1995.
Becker, Alton L. 1988. Language in particular: A lecture. In: Deborah Tannen (ed.), Linguistics
in Context: Connecting Observation and Understanding. (Lectures from the 1985 LSA/TESOL
and NEH Institutes). Norwood, N.J.: Ablex. 17–35. Reprinted in Becker 1995.
Becker, Alton L. 1995. Beyond Translation: Essays Toward a Modern Philology. Ann Arbor: Uni-
versity of Michigan Press.
Biber, Douglas and Finegan, Edward. 1988. Adverbial stance types in English. Discourse Proc-
esses 11: 1–34.
Blakemore, Dianne. 1987. Semantic Constraints on Relevance. Oxford: Blackwell.
Blum-Kulka, Shoshana. 1992. The metapragmatics of politeness in Israeli society. In: Richard J.
Watts, Sachiko Ide, and Konrad Ehlich (eds.), Politeness in Language: Studies in its History,
Theory and Practice. Berlin: Mouton de Gruyter. 255–281.
Blum-Kulka, Shoshana. 1997. Dinner Talk. Mahwah, N.J.: Lawrence Erlbaum.
Blum-Kulka, Shoshana and Katriel, Tamar. 1991. Nicknaming practices in families: A cross-
cultural perspective. In: Stella Ting-Toomey and Felipe Korzenny (eds.), Cross-Cultural
Interpersonal Communication. International and Intercultural Communication Annual, vol.
15. Newbury Park: Sage. 59–78.
 Bibliography 

Blythe, Carl Jr., Recktenwald, Sigrid, and Wang, Jenny. 1990. I’m like, ‘say what?!’: A new quota-
tive. American Speech 65: 215–227.
Bolozky, Shmuel. 1984. Subject pronouns in colloquial Hebrew. Hebrew Studies 25: 126–130.
Bolozky, Shmuel. 1999. On the special status of the vowels a and e in Israeli Hebrew. Hebrew
Studies 40: 233–250.
Brinton, Laurel J. 1990. The development of discourse markers in English. In: Jacek Fisiak (ed.),
Historical Linguistics and Philology. Berlin: Mouton de Gruyter. 45–71.
Brinton, Laurel J. 1996. Pragmatic Markers in English: Grammaticalization and Discourse Func-
tions. Berlin: Mouton de Gruyter.
Brinton, Laurel J. 2001. From matrix clause to pragmatic marker: The history of look-forms.
Journal of Historical Pragmatics 2: 177–199.
Brody, Jill. 1987. Particles borrowed from Spanish as discourse markers in Mayan languages.
Anthropological Linguistics 29: 507–521.
Brown, Gillian and Yule, George. 1983. Discourse Analysis. Cambridge: Cambridge University
Press.
Brown, Penelope and Levinson, Stephen 1987[1978]. Politeness: Some Universals in Language
Usage. Cambridge: Cambridge University Press.
Bublitz, Wolfram. 1988. Supportive Fellow-Speakers and Cooperative Conversations. Amsterdam/
Philadelphia: John Benjamins.
Butters, Ronald. 1982. Editor’s note. American Speech 57: 149.
Bybee, Joan L. 2003. Mechanisms of change in grammaticization: The role of frequency. In:
Brian D. Joseph and Richard D. Janda (eds.), The Handbook of Historical Linguistics. Ox-
ford: Blackwell. 602–623.
Bybee, Joan L. and Hopper, Paul J. (eds.). 2001. Frequency and the Emergence of Linguistic Struc-
ture. Amsterdam/Philadelphia: John Benjamins.
Bybee, Joan L. and Pagliuca, William. 1987. The evolution of future meaning. In: Anna Gia-
calone Ramat, Onofrio Carruba, and Giuliano Bernini (eds.), Papers from the 7th Interna-
tional Conference on Historical Linguistics. Amsterdam/Philadelphia: John Benjamins.
109–122.
Bybee, Joan L., Perkins, Revere, and Pagliuca, William. 1994. The Evolution of Grammar: Tense,
Aspect, and Modality in the Languages of the World. Chicago: University of Chicago Press.
Cassidy, Frederick. 1981. OK – is it African? American Speech 56: 269–273.
Chafe, Wallace L. 1980. The deployment of consciousness in the production of a narrative. In:
Wallace L. Chafe (ed.), The Pear Stories: Cognitive, Cultural, and Linguistic Aspects of Nar-
rative Production. Norwood, N. J.: Ablex. 9–50.
Chafe, Wallace L. 1985. Linguistic differences produced by differences between speaking and
writing. In: David R. Olson, Nancy Torrance, and Angela Hildyard (eds.), Literacy, Lan-
guage, and Learning. Cambridge: Cambridge University Press. 105–123.
Chafe, Wallace L. 1987. Cognitive constraints on information flow. In: Russell S. Tomlin (ed.),
Coherence and Grounding in Discourse. Amsterdam/Philadelphia: John Benjamins. 21–51.
Chafe, Wallace L. 1988. Linking intonation units in spoken English. In: John Haiman and San-
dra A. Thompson (eds.) Clause Combining in Grammar and Discourse. Amsterdam/Phila-
delphia: John Benjamins. 1–27.
Chafe, Wallace L. 1994. Discourse, Consciousness, and Time: The Flow and Displacement of Con-
scious Experience in Speaking and Writing. Chicago: University of Chicago Press.
 Metalanguage in Interaction

Clancy, Patricia M., Thompson, Sandra A., Suzuki, Ryoko, and Tao, Hongyin. 1996. The conver-
sational use of reactive tokens in English, Japanese, and Mandarin. Journal of Pragmatics
26: 355–387.
Clover, Carol J. 1974. Scene in saga composition. Arkiv för Nordisk Filologi 89: 57–83.
Clover, Carol J. 1982. The Medieval Saga. Ithaca/London: Cornell University Press.
Coates, Jeniffer. 1997. Competing discourses of femininity. In: Helga Kotthoff and Ruth Wodak (eds.),
Communicating Gender in Context. Amsterdam/Philadelphia: John Benjamins. 285–314.
Condon, Sherri. 1986. The discourse functions of OK. Semiotica 60: 73–101.
Condon, Sherri. 2001. Discourse ok revisited: Default organization in verbal interaction. Journal
of Pragmatics 33: 491–513.
Couper-Kuhlen, Elizabeth. 1998. On high onsets and their absence in conversational interac-
tion. Interaction and Linguistic Structures (InLiSt) 8: 1–26.
Couper-Kuhlen, Elizabeth and Thompson, Sandra A. 2000. Concessive patterns in conversa-
tion. In: Elizabeth Couper-Kuhlen and Bernd Kortmann (eds.), Cause, Condition, Conces-
sion, Contrast: Cognitive and Discourse Perspectives. Berlin: Mouton de Gruyter. 381–410.
Da Matta, Roberto. 1991. Carnivals, Rogues and Heroes: An Interpretation of the Brazilian Di-
lemma. Notre Dame, Indiana: University of Notre Dame Press.
Dahl, Östen. 1985. Tense and Aspect Systems. Oxford: Blackwell.
Dascal, Marcelo and Katriel, Tamar. 1982. Digressions: A study in conversational coherence. In:
Janos S. Petöfi (ed.), Text vs. Sentence (Papers in Textlinguistics 29). Hamburg: Helmut
Buske. 76–95.
Dasher, Richard B. 1995. Grammaticalization in the System of Japanese Predicate Honorifics.
Ph.D. Dissertation, Stanford University.
de Fina, Anna. 1997. An analysis of Spanish bien as a marker of classroom management in
teacher-student interaction. Journal of Pragmatics 28: 337–354.
de Rooji, Vincent. 1996. Cohesion Through Contrast: Discourse Structure in Shaba Swahili/French
Conversations. Amsterdam: IFOTT.
Du Bois, John W. 1985. Competing motivations. In: John Haiman (ed.), Iconicity in Syntax.
Amsterdam/Philadelphia: John Benjamins. 343–365.
Du Bois, John W., Cumming, Suanna, Schuetze-Coburn, Stephan, and Paolino, Danae. 1992.
Discourse Transcription: Santa Barbara Papers in Linguistics, vol. 4. Santa Barbara: Depart-
ment of Linguistics, University of California, Santa Barbara.
Enkvist, Nils Erik and Wårvik, Brita. 1987. Old English tha, temporal chains, and narrative
structure. In: Anna Giacalone Ramat, Onofrio Carrua, and Guiliano Bernini (eds.), Papers
from the 7th International Conference on Historical Linguistics. Amsterdam/Philadelphia:
John Benjamins. 221–237.
Erickson, Frederick and Shultz, Jeffrey 1982. The Councelor as Gatekeeper: Social Interaction in
Interviews. New York: Academic Press.
Even-Shoshan, Avraham. 1986. hamilon hexadash (‘The New Dictionary’). Jerusalem: Kiryat
Sefer.
Even-Shoshan, Avraham. 2003. milon even shoshan (‘The Even Shoshan Dictionary’). Jerusalem:
Hamilon Hexadash Inc.
Even-Zohar, Itamar. 1982. The emergence of speech organisers in a renovated language: The
case of Hebrew void pragmatic connectives. In: Nils E. Enkvist (ed.), Impromptu Speech: A
Symposium (Meddelanden från Stiftelsens för Åbo Akademi Forskiningsinstitut 78). Åbo:
Åbo Akademi. 179–193.
 Bibliography 

Ferrara, Kathleen 1992. The interactive achievement of a sentence: Joint productions in thera-
peutic discourse. Discourse Processes 15: 207–228.
Ferrara, Kathleen. 1997. Form and function of the discourse marker anyway: Implications for
discourse analysis. Linguistics 35: 343–378.
Ferrara, Kathleen and Bell, Barbara. 1995. Sociolinguistic variation and discourse function of
constructed dialogue introducers: The case of Be – Like. American Speech 70: 265–290.
Finell, Anne. 1989. Well now and then. Journal of Pragmatics 13: 653–6.
Finell, Anne. 1992. The repertoire of topic changers in personal, intimate letters: A diachronic
study of Osborne and Woolf. In: Matti Rinassen, Ossi Ihalainen, Terttu Nevalainen, and
Irma Taavitsainen (eds.), History of Englishes: New Methods and Interpretations in Historical
Linguistics. Berlin/New York: Mouton de Gruyter. 720–735.
Fischer, Kerstin. 2006. Towards an understanding of the spectrum of approaches to discourse
particles: Introduction to the volume Approaches to Discourse Particles. North Holland:
Elsevier. 1–20.
Fleischman, Suzanne. 1985. Discourse functions of tense-aspect oppositions in narrative: To-
ward a theory of grounding. Linguistics 23: 851–882.
Fleischman, Suzanne. 1989. Temporal distance: A basic linguistic metaphor. Studies in Language
13: 1–50.
Fleischman, Suzanne. 1990. Tense and Narrativity: From Medieval Performance to Modern Fic-
tion. Austin: University of Texas Press.
Fleischman, Suzanne. 1995. Imprefective and irrealis. In: Joan L. Bybee and Suzanne Fleischman
(eds.), Modality in Grammar and Discourse. Amsterdam/Philadelphia: John Benjamins.
519–551.
Fleischman, Suzanne and Yaguello, Marina. 1999. Discourse markers in comparative perspec-
tive: A contribution to cross-language pragmatics. Paper presented at the PRAGMA con-
ference, Tel Aviv University, Israel, June 13–16, 1999.
Fleischman, Suzanne and Yaguello, Marina. 2004. Discourse markers in comparative perspec-
tive: A contribution to cross-language pragmatics. In: Carol Lynn Moder and Aida Marti-
novic-Zic (eds.), Discourse Across Languages and Cultures. Amsterdam/Philadelphia: John
Benjamins. 129–147.
Ford, Cecilia E. 1993. Grammar in Interaction. Cambridge: Cambridge University Press.
Ford, Cecilia E. and Thompson, Sandra A. 1996. Interactional units in conversation: Syntactic,
intonational, and pragmatic resources for the management of turns. In: Elinor Ochs,
Emanuel A. Schegloff, and Sandra A. Thompson (eds.), Interaction and Grammar. Cam-
bridge: Cambridge University Press. 134–184.
Fox, Barbara A., Maschler, Yael, and Uhmann, Susanne. (forthcoming). A cross-linguistic study
of self-repair: Evidence from English, German, and Hebrew.
Fraser, Bruce. 1996. Pragmatic markers. Pragmatics 6: 167–190.
Geis, Michael L. and Zwicky, Arnold M. 1971. On invited inferences. Linguistic Inquiry 2:
561–566.
Gergen, Kenneth J. 1991. The Saturated Self: Dilemmas of Identity in Contemporary Life. Basic
Books.
Giddens, Anthony. 1984. The Constitution of Society: Outline of a Theory of Structuration. Cam-
bridge: Polity Press.
Goffman, Erving. 1974. Frame Analysis. New York: Harper and Row.
Goffman, Erving. 1981. Forms of Talk. Philadelphia: University of Pennsylvania Press.
 Metalanguage in Interaction

Goss, Emily L. and Salmons, Joseph C. 2000. The evolution of a bilingual discourse marking
system: Modal particles and English markers in German-American dialects. International
Journal of Bilingualism 4: 469–484.
Grice, H. P. 1975. Logic in conversation. In: Peter Cole and J. L. Morgan (eds.), Syntax and Se-
mantics, vol. 3: Speech Acts. New York: Academic Press. 41–58.
Gumperz, John. 1982. Discourse Strategies. Cambridge: Cambridge University Press.
Günthner, Susanne. 1996. From subordination to coordination? Verb-second position in Ger-
man causal and concessive constructions. Pragmatics 6: 323–356.
Günthner, Susanne. 1997. Complaint stories: Constructing emotional reciprocity among wom-
en. In: Helga Kotthoff and Ruth Wodak (eds.), Communicating Gender in Context. Amster-
dam/Philadelphia: John Benjamins. 179–218.
Günthner, Susanne. 1999a. Polyphony and the “layering of voices” in reported dialogues: An
analysis of the use of prosodic devices in everyday reported speech. Journal of Pragmatics
31: 685–708.
Günthner, Susanne. 1999b. Wenn-Sätze im Vor-Vorfeld: Ihre Formen und Funktionen in der
gesprochenen Sprache. InLiSt. (<http://uni-potsdam.de/u/inlist/issues/11index.htm>).
Günthner, Susanne. 2000. From concessive connector to discourse marker. The use of obwohl in
everyday German interaction. In: Elizabeth Couper-Kuhlen and Bernd Kortmann (eds.),
Cause, Condition, Concession, and Contrast: Cognitive and Discourse Perspectives. Berlin:
Mouton de Gruyter. 439–468.
Günthner, Susanne. 2007. N-be-that-constructions in everyday German conversation. Paper
presented at the 10th International Pragmatics Conference, July 13, Göteborg, Sweden.
Günthner, Susanne. 2008. Between emergence and sedimentation: Communicative patterns in
interaction. Workshop on Emergent Constructions, Freiburg Institute of Advanced Studies
(FRIAS), May 8–9, Freiburg, Germany.
Haiman, John 1978. Conditionals are topics. Language 54: 564–589.
Haiman, John. 1988. Incorporation, parallelism, and focus. In: Michael Hammond, Edith
Moravcsik and Jessica R. Wirth (eds.), Studies in Syntactic Typology. Amsterdam/Philadel-
phia: John Benjamins. 303–320.
Haiman, John. 1993. Life, the universe, and human language (A brief synopsis). Language Sci-
ences 15: 293–322.
Haiman, John. 1994. Ritualization and the development of language. In: William Pagliuca (ed.),
Perspectives on Grammaticalization. Amsterdam/Philadelphia: John Benjamins. 3–28.
Haiman, John. 1998. Talk is Cheap: Sarcasm, Alienation, and the Evolution of Language. Oxford/
New York: Oxford University Press.
Hansen, Mosegaard Maj-Britt. 1997. Alors and donc in spoken French: A reanalysis. Journal of
Pragmatics 28: 153–187.
Heine, Bernd and Reh, Mechthild. 1984. Grammaticalization and Reanalysis in African Lan-
guages. Hamburg: Helmut Buske.
Heine, Bernd, Claudi, Ulrike, and Hünnemeyer, Friederike. 1991. Grammaticalization: A Con-
ceptual Frameword. Chicago: University of Chicago Press.
Henkin, Roni. 1991. ktanim 'im harbe 'avar: shimushim yixudiyim shel zman 'avar bilshon
yeladim (‘Little with lots of past: Special uses of past tense in children’s language’). Lesho-
nenu 55: 333–362. (In Hebrew).
Henkin, Roni. 1999. ma beyn ‘hashamayim kxulim ka'ele’, ‘hashamayim kxulim kaze’
ve‘hashamayim kxulim ke'ilu’: 'al hashimush bexinuyey remez mashvim visodot 'axerim lehi-
staygut. (‘The difference between ‘the sky is blue like this’, ‘the sky is blue kaze’, and ‘the sky
 Bibliography 

is blue ke'ilu’: On the use of comparative demonstratives and other elements for hedging’).
In: Rina Ben Shachar and Gideon Touri (eds.), Ha'vrit safa xaya II (Hebrew – A Living
Language II). Tel Aviv: Porter Institute for Poetics and Semoitics, Tel Aviv University. 103–
122. (In Hebrew).
Heritage, John. 1984. A change-of-state token and aspects of its sequential placement. In: J.
Maxwell Atkinson and John Heritage (eds.), Structures of Social Action: Studies in Conversa-
tion Analysis. Cambridge: Cambridge University Press. 299–345.
Hilmisdóttir, Helga. 2007. A Sequential Analysis of nú and núna in Icelandic Conversation.
Ph.D. dissertation, Department of Scandinavian Languages and Literature, University of
Helsinki.
Hirschfeld, Ariel. 1990. 'Al kaze veke'ilu (‘On kaze and ke'ilu’). Politika 34: 6–9. (In Hebrew).
Hlavac, Jim. 2006. Bilingual discourse markers: Evidence from Croatian-English code-switch-
ing. Journal of Pragmatics 38: 1870–1900.
Hopper, Paul. J. 1979. Aspect and foregrounding in discourse. In: Talmi Givón, (ed.), Discourse
and Semantics (Syntax and Semantics, vol. 12). New York.: Academic Press. 213–241.
Hopper, Paul J. 1987. Emergent grammar. In: Jon Aske, Natasha Beery, Laura Michaelis, and
Hana Filip (eds.), Proceedings of the Thirteenth Annual Meeting of the Berkeley Linguistics
Society 13: 139–157. Berkeley: Berkeley Linguistics Society.
Hopper, Paul J. 1988. Emergent grammar and the a priori grammar postulate. In: Deborah Tan-
nen (ed.), Linguistics in Context: Connecting Observation and Understanding (Lectures from
the 1985 LSA/TESOL and NEH Institutes). Norwood, N.J.: Ablex. 117–134.
Hopper, Paul J. 1991. On some principles of grammaticalization. In: Elizbeth C. Traugott and
Bernd Heine (eds.), Approaches to Grammaticalization. Amsterdam/Philadelphia: John
Benjamins. 17–35.
Hopper, Paul J. 1998. Emergent grammar. In: Michael Tomasello (ed.), The New Psychology of
Language: Cognitive and Functional Approaches to Language Structure. Mahwah, N.J.: Law-
rence Erlbaum. 155–175.
Hopper, Paul J. 2001. Grammatical constructions and their discourse origins: Prototype or fam-
ily resemblance? In: Martin Pütz, Susanne Neimeier, and René Dirven (eds.), Applied Cog-
nitive Linguistics I: Theory and Language Acquisition. Berlin/New York: Mouton de Gruyter.
109–129.
Hopper, Paul J. 2004. The openness of grammatical constructions. Papers from the 40th Regional
Meeting of the Chicago Linguistic Society, 153–175.
Hopper, Paul J. and Thompson, Sandra A. 1980. Transitivity in grammar and discourse. Lan-
guage 56: 251–299.
Hopper, Paul J. and Thompson, Sandra A. 1984. The discourse basis for lexical categories in
universal grammar. Language 60: 703–783.
Hopper, Paul J. and Thompson, Sandra A. 2008. Projectability and clause combining in interac-
tion. In: Ritva Laury (ed.), Crosslinguistic Studies of Clause Combining: The Multifunctional-
ity of Conjunctions. Amsterdam/Philadelphia: John Benjamins. 99–123.
Hopper, Paul J. and Traugott, Elizabeth Closs. 1993. Grammaticalization. Cambridge: Cam-
bridge University Press.
Hopper, Paul J. and Traugott, Elizabeth Closs. 2003. Grammaticalization. Second edition. Cam-
bridge: Cambridge University Press.
Hopper, Robert. 1989. Speech in telephone openings: Emergent interaction vs. routines. West-
ern Journal of Speech Communication 53: 178–184.
 Metalanguage in Interaction

Imo, Wolfgang. 2007. Clines of subordination – constructions with the German “complement-
taking-predicate”. Paper presented at the 10th International Pragmatics Conference, July
13, Göteborg, Sweden.
Jefferson, Gail. 1993. Caveat speaker: Preliminary notes on recipient topic-shift implicature. Re-
search on Language and Social Interaction 26: 1–30. (First circulated as Tillburg Papers in
language and Literature, 1983).
Jucker, Andreas H. 1993. The discourse marker well: A relevance-theoretical account. Journal of
Pragmatics 19: 435–452.
Jucker, Andreas H. 1997. The discourse marker well in the history of English. English Language
and Linguistics 1: 91–110.
Jucker, Andreas H. and Smith, Sara W. 1998. “And people just you know like ‘wow’ ”: Discourse
markers as negotiating strategies. In: Andreas H. Jucker and Yael Ziv (eds.), Discourse
Markers: Descriptions and Theory. Amsterdam/Philadelphia: John Benjamins. 172–201.
Jucker, Andreas H. and Ziv, Yael (eds.). 1998. Discourse Markers: Descriptions and Theory. Am-
sterdam/Philadelphia: John Benjamins.
Kärkkäinen, Elise. 2003. Epistemic Stance in English Conversation: A Description of its Interactional
Functions, with a Focus on I think. Amsterdam/Philadelphia: John Benjamins.
Katriel, Tamar. 1986. Talking Straight: ‘Dugri’ Speech in Israeli Sabra Culture. Cambridge: Cam-
bridge University Press.
Katriel, Tamar. 1991. Communal Webs. Communication and Culture in Contemporary Israel. Al-
bany: State University of New York Press.
Katriel, Tamar. 1999. milot mafteax: dfusey tarbut vetikshoret beyisrael (‘Key Words: Patterns of
Culture and Communication in Israel’). Tel Aviv: Haifa University Press and Zmora Bitan.
(In Hebrew).
Katriel, Tamar. 2004. Dialogic Moments: From Soul Talks to Talk Radio in Israeli Culture. Detroit:
Wayne State University Press.
Keenan, Elinor Ochs and Schieffelin, Bambi. 1976. Topic as a discourse notion: A study of topic
in the conversations of children and adults. In: Charles N. Li (ed.), Subject and Topic. New
York: Academic Press. 335–384.
Keevallik, Leelo. 2006. From discourse pattern to epistemic marker: Estonian (ei) tea ‘don’t
know’. Nordic Journal of Linguistics 29: 173–200.
Keevallik, Leelo. 2007. Interrogative complements as independent clauses. Paper presented at
the 10th International Pragmatics Conference, July 13, Göteborg, Sweden.
Kimmerling, Baruch. 2001. The Invention and Decline of Israeliness: State, Society, and the Mili-
tary. Berkeley: University of California Press.
König, Ekkehard. 1991. The Meaning of Focus Particles: A Comparative Perspective. London:
Routledge.
Krupik, Shani. 2007. Samaney siax batikshoret hakwazi-sinxronit metuvexet-hamaxshev (siax
hachet). (‘Discourse Markers in Israeli Hebrew Quasi-Synchronous Computer-Mediated
Communication (Chat Discourse)’). M.A. Thesis, Department of Hebrew Language, Uni-
versity of Haifa. (In Hebrew).
Kuryłowicz, Jerzy. 1964. The Inflectional Categories of Indo-European. Heidelberg: Winter.
Kussmaul, Paul. 1978. In fact, actually, anyway Indikatoren illokutiver Sprechakte. Die neueren
Schprachen 27: 357–369.
Kyratzis Amy. 2007. Mexican-immigrant children’s code-switching and alignment in peer play
interactions in a bilingual U.S. preschool. Paper presented at the 10th International Prag-
matics Conference, July 8–13, Göteborg, Sweden.
 Bibliography 

Labov, William. 1972. Language in the Inner City. Philadelphia: University of Pennsylvania
Press.
Laury, Ritva and Seppänen, Eeva-Leena. 2008. Clause combining, interaction, evidentiality, par-
ticipation structure, and the conjunction-particle continuum: The Finnish että. In: Ritva
Laury (ed.), Crosslinguistic Studies of Clause Combining: The Multifunctionality of Conjunc-
tions. Amsterdam/Philadelphia: John Benjamins. 153 -178.
Lambrecht, Knud. 1994. Information Structure and Sentence Form: Topic, Focus, and the Mental
Representations of Discourse Referents. Cambridge/N.Y.: Cambridge University Press.
Lehman, Christian. 1985. Grammaticalization: Synchronic variation and diachronic change.
Lingua e Stile 20: 303–318.
Lenk, Uta. 1995. Discourse markers and conversational coherence. In: Brita Wårvik, Sanna-
Kaisa Tanskanen, and Risto Hiltungen (eds.), Organization in Discourse: Proceedings from
the Turku Conference (Anglicana Tukuensia 14). Turku: University of Turku. 341–352.
Lenk, Uta. 1998. Marking Discourse Coherence: Functions of Discourse Markers in Spoken Eng-
lish. Tübingen: Gunter Narr Verlag.
Lerner, Gene. 1991. On the syntax of sentences-in-progress. Language in Society 20: 441–458.
lexicon hasleng ha'ivri vehatsva'i. 1993. (‘The Hebrew and Army Slang Lexicon’). Tel Aviv: Prolog
Publishers.
Li Wei and Milroy, Lesley. 1995. Preference marking and repair strategies in bilingual conversation:
Evidence from a Chinese-English bilingual community. Journal of Pragmatics 23: 281–99.
Lindström, Jan and Wide, Camilla. 2005. Tracing the origins of a set of discourse particles:
Swedish particles of the type you know. Journal of Historical Pragmatics 6: 211–236.
Livnat, Zohar and Yatsiv, Ilil. 2003. Causality and justification: The causal marker ki in spoken
Hebrew. Revue de Semantique et Pragmatique 13: 99–119.
Livnat, Zohar and Yatsiv, Ilil. 2006. koherentiyut besiax davur: 'iyun betafkidav shel saman hasiax
'az (‘Coherence in spoken discourse: A study in the functions of the discourse marker 'az
(‘then/so’). In: Rina Ben Shachar (ed.), Ha'vrit safa xaya IV (Hebrew – A Living Language
IV). Tel Aviv: Porter Institute for Poetics and Semoitics, Tel Aviv University. 175–189. (In
Hebrew).
Longacre, Robert. 1976. Mystery particles and affixes. In: Salikoko S. Mufwene, Carol A. Walker,
and Sanford B. Steever (eds.), Papers from the Twelfth Regional Meeting of the Chicago Lin-
guistic Society. 468–475.
Lönnroth, Lars. 1976. Njáls Saga: A Critical Introduction. Berkeley/Los Angeles/London: Uni-
versity of California Press.
Lucy, John. 1993. Metapragmatic presentationals: Reporting speech with quotatives in Yucatec
Maya. In: John Lucy (ed.), Reflexive Language. Cambridge: Cambridge University Press.
91–126.
Maltz, Daniel N. and Borker, Ruth A. 1982. A cultural approach to male-female miscommunica-
tion. In: John J. Gumperz (ed.), Language and Social Identity. Cambridge: Cambridge Uni-
versity Press. 195–216.
Manoliu, Maria. 2000. From deixis ad oculos to discourse markers via deixis ad phantasma. In:
John C. Smith and Delia Bentley (eds.), Historical Linguistics 1995, vol. I: General Issues and
Non-Germanic Languages. Amsterdam/Philadelphia: John Benjamins. 243–260.
Marcuse, Herbert. 1968. One Dimensional Man. London: Routledge & K. Paul.
Maschler, Yael. 1988. The Games Bilinguals Play: A Discourse Analysis of Hebrew-English Bi-
lingual Conversation. Ph.D. dissertation, University of Michigan, Ann Arbor.
 Metalanguage in Interaction

Maschler, Yael. 1991. The language games bilinguals play: Language alternation at language
game boundaries. Language and Communication 11: 263–89.
Maschler, Yael. 1993. Iconicity in discourse: The story of Echo. Poetics Today 14: 653–690.
Maschler, Yael. 1994a. Appreciation ha'araxa 'o ha'aratsa (‘valuing or admiration’)?’: Negotiat-
ing contrast in bilingual disagreement talk. Text 14: 207–238.
Maschler, Yael. 1994b. Metalanguaging and discourse markers in bilingual conversation. Lan-
guage in Society 23: 325–366.
Maschler, Yael. 1997a. Discourse markers at frame shifts in Israeli Hebrew talk-in-interaction.
Pragmatics 72: 183–211.
Maschler, Yael. 1997b. Emergent bilingual grammar: The case of contrast. Journal of Pragmatics
28: 279–313.
Maschler, Yael. 1998a. On the transition from code-switching to a mixed code. In: Peter Auer
(ed.), Code-Switching in Conversation. London: Routledge. 125–149.
Maschler, Yael. 1998b. Rotse lishmoa keta? ‘wanna hear something weird/funny [lit. ‘a seg-
ment’]?’: The discourse markers segmenting Israeli Hebrew talk-in-interaction. In: Andreas
H. Jucker and Yael Ziv (eds.), Discourse Markers: Descriptions and Theory. Amsterdam/
Philadelphia: John Benjamins. 13–59.
Maschler, Yael (ed.). 2000a. Discourse Markers in Bilingual Conversation. Special issue of Inter-
national Journal of Bilingualism 4: 437–561.
Maschler, Yael. 2000b. What can bilingual conversation tell us about discourse markers?: Intro-
duction to the special issue on Discourse Markers in Bilingual Conversation. International
Journal of Bilingualism 4: 437–445.
Maschler, Yael. 2000c. Toward fused lects: Discourse markers in Hebrew-English bilingual con-
versation twelve years later. International Journal of Bilingualism 4: 529–561.
Maschler, Yael. 2001. veke'ilu haragláyim sh’xa nitka'ot bifním kaze (‘and like your feet get stuck
inside like’): Hebrew kaze (‘like’), ke'ilu (‘like’), and the decline of Israeli dugri (‘direct’)
speech. Discourse Studies 3: 295–326.
Maschler, Yael. 2002a. The role of discourse markers in the construction of multivocality in Is-
raeli Hebrew talk-in-interaction. Research on Language and Social Interaction 35: 1–38.
Maschler, Yael. 2002b. On the grammaticization of ke'ilu (‘like’, lit. ‘as if ’) in Hebrew talk-in-in-
teraction. Language in Society 31: 243–276.
Maschler, Yael. 2003. The discourse marker nu: Israeli Hebrew impatience in interaction. Text
23: 89–128.
Maschler, Yael. 2004. The Haifa Corpus of Spoken Israeli Hebrew. http://hevra.haifa.ac.il/com/
maschler/.
Maschler, Yael. 2008. Discourse Markers. The International Encyclopedia of Communication (ed.
Donsbach), Volume IV. Wiley-Blackwell (Oxford, UK and Malden, MA), 2008.
1364–1366.
Maschler, Yael. 2009. ma'arexet samaney hasiax shel ha'ivrit hayomyomit hadvura (‘The dis-
course marker system of casual spoken Hebrew’). Balshanut 'Ivrit (Hebrew Linguistics).
62–63: 99–129. (In Hebrew)
Maschler, Yael and Estlein, Roi. 2008. Stance-taking in Israeli Hebrew casual conversation via
be'emet (‘really, actually, indeed’, lit. ‘in truth’). Discourse Studies 10(3): 283–316.
Maschler, Yael and Shaer, Susan. (forthcoming). On the emergence of adverbial connectives
from Hebrew relative clause constructions. In: Peter Auer and Stephan Pfänder (eds.),
Emergent Constructions. Berlin: Mouton de Gruyter.
 Bibliography 

Matras, Yaron. 1998. Utterance modifiers and universals of grammatical borrowing. Linguistics
36: 281–331.
Mazo, Natalie and Voloshin, Oleg 1999. Forms and functions of the discourse marker nu in Rus-
sian and Hebrew conversations. A Seminar Paper. Department of English, University of
Haifa at Oranim.
McConvell, Patrick, 1988. Mix-im-up: Aboriginal code-switching, old and new. In Monica Hel-
ler (ed.), Codeswitching: Anthropological and Sociolinguistic Perspectives. Berlin: Mouton de
Gruyter. 97–169.
Meillet, Antoine. 1912. L’évolution des formes grammaticales. Scientia (Rivista di Scienze) 12,
no. 26, 6. Reprinted in Meillet 1958:€130–148.
Meillet, Antoine. 1958. Linguistique historique et linguistique générale. Paris: Champion.
Merritt, Marylin. 1978. On the use of ‘O.K.’ in service encounters. Sociolinguistic Working Paper
42, University of Texas at Austin, April. [Also in: James Baugh and Joel Sherzer (eds.), Lan-
guage in Use. Englewood Cliffs, N.J.: Prentice-Hall. 139–147.]
Miller, Jim and Weinert, Regina. 1998. Spontaneous Spoken Language: Syntax and Discourse.
Oxford: Clarendon Press.
Miracle, W. Charles. 1989. Hao: A Chinese discourse marker. Papers from the 25th Regional
Meeting of the Chicago Linguistic Society, 213–227.
Moyer, Melissa. 1998. Bilingual conversation strategies in Gibraltar. In: Peter Auer (ed.), Code-
Switching in Conversation: Linguistic Perspectives on Bilingualism. London: Routledge. 215–
234.
Mulder, Jean and Thompson, Sandra A. 2008. The grammaticization of English but as a final
particle in English conversation. In: Ritva Laury (ed.), Crosslinguistic Studies of Clause
Combining: The Multifunctionality of Conjunctions. Amsterdam/Philadelphia: John Ben-
jamins. 179–204.
Onodera, Noriko O. 1995. Diachronic analysis of Japanese discourse markers. In: Andreas H.
Jucker (ed.), Historical Pragmatics. Amsterdam/Philadelphia: John Benjamins. 393–437.
Onodera, Noriko O. 2004. Japanese Discourse Markers: Synchronic and Diachronic Discourse
Analysis. Amsterdam/Philadelphia: John Benjamins.
Oring, Elliot. 1981. Israeli Humor: The Content and Structure of the “Chizbat” of the Palmach.
Albany: State University of New York Press.
Ortega y Gasset, José. 1959. The difficulty of reading. Diogenes 28: 1–17.
Ösch-Serra, Cecilia. 1998. Discourse connectives in bilingual conversation: The case of an
emerging Italian-French mixed code. In: Peter Auer (ed.), Code-Switching in Conversation:
Linguistic Perspectives on Bilingualism. London: Routledge. 101–122.
Östman, Jan-Ola. 1981. “You know”: A Discourse Functional View. Amsterdam/Philadelphia:
John Benjamins.
Östman, Jan-Ola. 1982. The symbiotic relationship between pragmatic particles and impromptu
speech. In: Nils E. Enkvist (ed.), Impromptu Speech: A Symposium (Meddelanden från Stif-
telsens för Åbo Akademi Forskiningsinstitut 78). Åbo: Åbo Akademi. 147–177.
Owen, Marion. 1985. The conversational functions of anyway. Nottingham Linguistic Circular
14: 72–90.
Pagliuca, William (ed.). 1994. Perspectives on Grammaticalization. Amsterdam/Philadelphia:
John Benjamins.
Pawley, Andrew and Syder, Frances. 1983. Two puzzles for linguistic theory: Nativelike selection
and nativelike fluency. In: Jack C. Richards and Richard W. Schmidt (eds.), Language and
Communication. London: Longman. 191–267.
 Metalanguage in Interaction

Pekarek Doehler, Simona. 2007. Subordination in grammar – subordination in action: Projec-

tion sequences in French conversation. Paper presented at the 10th International Pragmat-
ics Conference, July 13, Göteborg, Sweden.
Pomerantz, Anita. 1984. Agreeing and disagreeing with assessments: Some features of preferred/
dispreferred turn-shapes. In: J. Maxwell Atkinson and John Heritage (eds.), Structures of
Social Action: Studies in Conversation Analysis. Cambridge: Cambridge University Press.
79–112.
Pons Bordería, Salvador and Schwenter, Scott A. 2005. Polar meaning and “expletive” negation
in approximative adverbs: Spanish por poco (no). Journal of Historical Pragmatics 6:
262–282.
Quinn, Arthur. (1982). Figures of Speech. Salt Lake City: Gibbs M. Smith.
Ramat, Paolo and Ricca, Davide. 1994. Prototypical adverbs: On the scalarity/radiality of the
notion of ADVERB. Rivista di Linguistica 6: 289–326.
Ravid, Dorit. 1977. mispar heybetim shel be'ayat seder hamarkivim be'ivrit yisra'elit modernit (‘Sev-
eral aspects of the problem of element order in general Israeli Hebrew’). balshanut 'ivrit xofÂ�
shit (‘Hebrew Computational, Formal, and Applied Linguistics’) 11: 1–45. (In Hebrew).
Ravid, Dorit. 1995. Neutralization of gender distinctions in Modern Hebrew numerals. Lan-
guage Variation and Change 7: 79–100.
Read, Allen Walker. 1963a. The first stage in the history of ‘O.K.’ American Speech 38: 5–27.
Read, Allen Walker. 1963b. The second stage in the history of ‘O.K.’ American Speech 38:
83–102.
Read, Allen Walker. 1964a. The folklore of ‘O.K.’ American Speech 39: 7–25.
Read, Allen Walker. 1964b. Later stages in the history of ‘O.K.’ American Speech 39: 83–101.
Read, Allen Walker. 1964c. Successive revisions in the explanation of ‘O.K.’ American Speech 39:
243–268.
Read, Allen Walker. 1988. Exuberance, a motivation for language. Word Ways 21: 71–74.
Reddy, Michael J. 1979. The conduit metaphor: A case of frame conflict in our language about
language. In: Andrew Ortnoy (ed.), Metaphor and Thought. Cambridge: Cambridge Uni-
versity Press. 284–324.
Redeker, Gisela. 1991. Linguistic markers of discourse structure. Linguistics 29: 1139–1172.
Reinhart, Tanya. 1984. Principles of Gestalt perception in the temporal organization of narrative
texts. Linguistics 29: 1139–1172.
Romaine, Suzanne and Lange, Deborah. 1991. The use of like as a marker of reported speech and
thought: A case of grammaticalization in progress. American Speech 66: 227–279.
Rosten, Leo. 1968. The Joys of Yiddish. New York: McGraw-Hill.
Sacks, Harvey. 1972. On the analyzability of stories by children. In: John J. Gumperz and Dell
Hymes (eds.), Directions in Sociolinguistics: The Ethnography of Communication. New York:
Holt, Rinehart & Winston. 325–345.
Sacks, Harvey. 1992. Lectures on Conversation, edited by Gail Jefferson. 2 vols. Oxford: Blackwell.
Sacks, Harvey, Schegloff, Emanuel A. and Jefferson, Gail. 1974. A simplest systematics for the
organization of turn-taking in conversation. Language 50: 696–735.
Salmons, Joseph C. 1990. Bilingual discourse marking: Code switching, borrowing, and conver-
gence in some German-American dialects. Linguistics 28: 453–480.
Samet, Gideon (ed.). 1990. Dor hakaze ke'ilu. (‘The kaze ke'ilu generation’). Special issue of Poli-
tika 34. (In Hebrew).
Schegloff, Emanuel A. 1968. Sequencing in conversational openings. American Anthropologist
70: 1075–1095.
 Bibliography 

Schegloff, Emanuel A. 1979. Identification and recognition in telephone conversational open-

ings. In: George Psathas (ed.), Everyday Language: Studies in Ethnomethodology. New York:
Irvington. 23–78.
Schegloff, Emanuel A. 1980. Preliminaries to preliminaries: “can I ask you a question”. In: Don
Zimmerman and Candace West (eds.), Language and Social Interaction, special edition of
Sociological Inquiry 50: 104–152.
Schegloff, Emanuel A. 1982. Discourse as an interactional achievement: Some uses of ‘uh huh’
and other things that come between sentences. In: Deborah Tannen (ed.), Georgetown Uni-
versity Roundtable on Languages and Linguistics (GURT) 1981: Analyzing Discourse: Text
and Talk. Washington, D.C.: Georgetown University Press. 71–93.
Schegloff, Emanuel A. 1986. The routine as achievement. Human Studies 9: 111–151.
Schegloff, Emanuel A. 1990. On the organization of sequences as a source of “coherence” in talk-
in-interaction. In: Bruce Dorval (ed.), Conversational Organization and its Development
(Advances in Discourse Processes XXIX). Norwood, N.J.: Ablex. 51–77.
Schegloff, Emanuel A. 1993. Reflections on quantification in the study of conversation. Research
on Language and Social Interaction 26: 99–128.
Schegloff, Emanuel A. 1996a. Turn organization: One intersection of grammar and interaction.
In: Elinor Ochs, Emanuel A. Schegloff, and Sandra A. Thompson (eds.), Interaction and
Grammar. Cambridge: Cambridge University Press. 52–133.
Schegloff, Emanuel A. 1996b. Issues of relevance for discourse analysis: Contingency in action,
interaction and co-participant context. In: Eduard H. Hovy and Donia R. Scott (eds.), Com-
putational and Conversational Discourse: Burning Issues – An Interdisciplinary Account.
Heidelberg: Springer Verlag. 3–35.
Schegloff, Emanuel A., Jefferson, Gail, and Sacks, Harvey. 1977. The preference for self-correc-
tion in the organization of repair in conversation. Language 53: 361–382.
Schegloff, Emanuel A. and Sacks, Harvey. 1973. Opening up closings. Semiotica 8: 289–327.
Schiffrin, Deborah. 1980. Meta-talk: Organizational and evaluative brackets in discourse. In:
Don H. Zimmerman and Candace West (eds.), Language and social interaction, special edi-
tion of Sociological Inquiry 50: 199–236.
Schiffrin, Deborah 1984. Jewish argument as sociability. Language in Society 13: 311–335.
Schiffrin, Deborah. 1987. Discourse Markers. Cambridge: Cambridge University Press.
Schiffrin, Deborah 1992. Conditionals as topics in discourse. Linguistics 30: 165–197.
Schiffrin, Deborah. 2001. Discourse markers: Language, meaning, and context. In: Deborah
Schiffrin, Deborah Tannen, and Heidi E. Hamilton (eds.), The Handbook of Discourse Anal-
ysis. Oxford: Blackwell. 54–75.
Schiffrin, Deborah. 2006. Discourse marker research and theory: Revisiting and. In: Kerstin
Fischer (ed.), Approaches to Discourse Particles. North Holland: Elsevier. 315–338.
Schourup, Lawrence C. 1985. Common Discourse Particles in English Conversation. New York:
Garland (dissertation, Columbus: Ohio State University, 1983).
Schourup, Lawrence. 1999. Discourse markers: Tutorial overview. Lingua 107: 227–265.
Schourup, Lawrence and Waida, Yukiko. 1987. English Connectives. Tokyo: Kurishio Publishing
Company.
Schwenter, Scott A. and Traugott, Elizabeth Closs. 2000. Invoking scalarity: The development of
in fact. Journal of Historical Pragmatics 1: 7–25.
Scollon, Ron. 1982. The rhythmic integration of ordinary talk. In: Deborah Tannen (ed.), Geor-
getown Univeristy Round Table on Languages and Linguistics (GURT) 1981: Analyzing Dis-
course: Text and Talk. Washington, D.C.: Georgetown University Press. 335–349
 Metalanguage in Interaction

Selting, Margret and Couper-Kuhlen, Elizabeth (eds.). 2001. Studies in Interactional Linguistics.
Amsterdam/Philadelphia: John Benjamin.
Sherzer, Joel. 1991. The Brazilian thumbs-up gesture. Journal of Linguistic Anthropology 1: 189–
197.
Shloush, Shelley. 1998. A unified account of Hebrew bekicur ‘in short’: Relevance theory and
discourse structure considerations. In: Andreas H. Jucker and Yael Ziv (eds.), Discourse
Markers: Descriptions and Theory. Amsterdam/Philadelphia: John Benjamins. 61–82.
Sivan, Reuben and Levenston, Edward A. 1967. The Megiddo Modern Dictionary: English-
Hebrew/Hebrew-English. Tel Aviv: Megiddo Publishing Co. Ltd.
Sorjonen, Marja-Leena 2002. Recipient activities: The particle no as a go-ahead response in
Finnish conversations. In: Cecilia E. Ford, Barbara A. Fox and Sandra A. Thompson (eds.),
The Language of Turn and Sequence. New York: Oxford University Press. 165–195.
Sperber, Dan and Wilson, Deirdre. 1986. Relevance: Communication and Cognition. Cambridge:
Cambridge University Press.
Stroud, Christopher. 1992. The problem of intention and meaning in code-switching. Text 12:
127–155.
Stubbs, Michael. 1983. Discourse Analysis: The Sociolinguistic Analysis of Natural Language. Ox-
ford: Basil Blackwell.
Suzuki, Ryoko. 1999. Grammaticization in Japanese: A Study of Pragmatic Particle-ization.
Ph.D. dissertation, University of California, Santa Barbara.
Sweetser, Eve. 1988. Grammaticalization and semantic bleaching. In: Shelley Axmaker, Annie
Jaisser, and Helen Singmaster (eds.), Proceedings of the Fourteenth Annual Meeting of the
Berkeley Linguistics Society 14: 389–405. Berkeley: Berkeley Linguistics Society.
Tabor, Whitney and Traugott, Elizabeth Closs. 1998. Structural scope expansion and grammati-
calization. In: Anna Giacalone Ramat and Paul J. Hopper (eds.), The Limits of Grammati-
calization. Amsterdam/Philadelphia: John Benjamins. 229–272.
Tagliamonte, Sali and Hudson, Rachel. 1999. Be like et al.€beyond America: The quotative system
in British and American youth. Journal of Sociolinguistics 3: 147–172.
Takahara, Paul, O. 1998. Pragmatic functions of the English discourse marker anyway and its
corresponding contrastive Japanese discourse markers. In: Andreas H. Jucker and Yael Ziv
(eds.), Discourse Markers: Descriptions and Theory. Amsterdam/Philadelphia: John Ben-
jamins. 327–351.
Tannen, Deborah. 1984. Conversational Style: Analyzing Talk Among Friends. Norwood, N.J.:
Ablex.
Tannen, Deborah. 1986. Introducing constructed dialogue in Greek and American conversa-
tional and literary narrative. In: Florian Coulmas (ed.), Direct and Indirect Speech. Berlin:
Mouton de Gruyter. 311–332.
Tannen, Deborah. 1989. Talking Voices: Repetition, Dialogue, and Imagery in Conversational Dis-
course. Cambridge: Cambridge University Press.
Tannen, Deborah. 1990. You Just Don’t Understand: Women and Men in Conversation. New
York: Ballantine.
Tannen, Deborah. 1999. The display of (gendered) identities in talk at work. In: Mary Bucholtz,
A. C. Liang, and Laurel A. Sutton (eds.), Reinventing Identities: The Gendered Self in Dis-
course. Oxford & New York: Oxford University Press. 221–240.
Tao, Hongyin. 2001. Discovering the usual with corpora: the case of remember. In: Rita C. Simp-
son and John Swales (eds.), Selected Papers from the North American Symposium on Cor-
pora and Applied Linguistics. Ann Arbor: University of Michigan Press. 116–144.
 Bibliography 

Tao, Hongyin. 2003. A usage-based approach to argument structure: ‘Remember’ and ‘Forget’ in
spoken English. International Journal of Corpus Linguistics 8: 75–95.
Thompson, Sandra A. 1984. ‘Subordination’ in formal and informal discourse. In: Deborah
Schiffrin (ed.), Georgetown University Round Table on Languages and Linguistics (GURT)
1984: Meaning, Form, and Use in Context: Linguistic Applications. Washington, D.C.: Geor-
getwon University Press. 85–94.
Thompson, Sandra A. 2002. “Object complements” and conversation: Towards a realistic ac-
count. Studies in Language 26: 125–163.
Thompson, Sandra A. and Hopper, Paul J. 2001. Transitivity, clause structure, and argument
structure: Evidence from conversation. In: Joan Bybee and Paul J. Hopper (eds.), Frequency
and the Emergence of Linguistic Structure. Amsterdam/Philadelphia: John Benjamins. 27–
60.
Thompson, Sandra A. and Mulac, Anthony. 1991. A quantitative perspective on the grammati-
cization of epistemic parentheticals in English. In: Elizabeth Closs Traugott and Bernd
Heine (eds.), Aproaches to Grammaticalization, vol. II. Amsterdam/Philadelphia: John Ben-
jamins. 313–329.
Traugott, Elizabeth Closs. 1980. Meaning-change in the development of grammatical markers.
Language Sciences 2: 44–61.
Traugott, Elizabeth Closs. 1982. From propositional to textual and expressive meanings: Some
semantic-pragmatic aspects of grammaticalization. In: Winfred P. Lehman and Yakov Mal-
kiel (eds.), Perspectives on Historical Linguistics. Amsterdam/Philadelphia: John Benjamins.
245–271.
Traugott, Elizabeth Closs. 1986. From polysemy to internal semantic reconstruction. Proceed-
ings of the Twelfth Annual Meeting of the Berkeley Linguistics Society 12: 539–550. Berkeley:
Berkeley Linguistics Society.
Traugott, Elizabeth Closs. 1988. Pragmatic strengthening and grammaticalization. In: Shelley Ax-
maker, Annie Jaisser, and Helen Singmaster (eds.), Proceedings of the Fourteenth Annual Meet-
ing of the Berkeley Linguistics Society 14: 406–416. Berkeley: Berkeley Linguistics Society.
Traugott, Elizabeth Closs. 1989. On the rise of epistemic meanings in English: An example of
subjectification in semantic change. Language 65: 31–55.
Traugott, Elizabeth Closs. 1995a. The role of the development of discourse markers in a theory
of grammaticalization. Twelfth International Conference on Historical Linguistics. Man-
chester, August 1995 (version of 11/97). http://www.stanford.edu/~traugott/traugott.html
Traugott, Elizabeth Closs. 1995b. Subjectification in grammaticalisation. In: Dieter Stein and
Susan Wright (eds.), Subjectivity and Subjectivisation: Linguistic Perspectives. Cambridge:
Cambridge University Press. 31–54.
Traugott, Elizabeth Closs. 1999. The role of pragmatics in semantic change. In: Jef Verschueren
(ed.), Pragmatics in 1998: Selected Papers from the 6th International Pragmatics Conference,
vol. II. Antwerp: International Pragmatics Association. 93–102.
Traugott, Elizabeth Closs. 2001. Class notes from LSA Summer Institute course “The role of se-
mantics in pragmatic change”. University of California at Santa Barbara, June 25 – August 3.
Traugott, Elizabeth Closs. 2003a. From subjectification to intersubjectification. In: Raymond
Hickey (ed.), Motives for Language Change. Cambridge: Cambridge University Press.
124–139.
Traugott, Elizabeth Closs. 2003b. Constructions in grammaticalization. In: Brian Joseph and
Richard D. Janda (eds.), The Handbook of Historical Linguistics. Oxford: Blackwell.
624–647.
 Metalanguage in Interaction

Traugott, Elizabeth Closs and Dasher, Richard B. 2002. Regularity in Semantic Change. Cam-
bridge: Cambridge University Press.
Traugott, Elizabeth Closs and Heine, Bernd (eds.). 1991. Approaches to Grammaticalization, 2
vols. Amsterdam/Philadelphia: John Benjamins.
Traugott, Elizabeth Closs and König, Ekkehard. 1991. The semantics-pragmatics of grammati-
calization revisited. In: Elizabeth Closs Traugott and Bernd Heine (eds.), Approaches to
grammaticalization, vol. II. Amsterdam/Philadelphia: John Benjamins. 189–218.
Underhill, Robert. 1988. Like is, like, focus. American Speech 63: 234–236.
Vincent, Diane. 2005. The journey of non-standard discourse markers in Quebec French: Net-
works based on exemplification. Journal of Historical Pragmatics 6: 188–210.
Visconti, Jacqueline. 2003. Scalar meanings in contrast: A diachronic approach. Paper presented
at the 8th International Pragmatics Conference, July 14, Toronto, Canada.
Visconti, Jacqueline. 2005. On the origins of scalar particles in Italian. Journal of Historical Prag-
matics 6: 237–261.
Wårvik, Brita. 1987. On grounding in narratives. In: Ishrat Lindblad and Magnus Ljung (eds.),
Proceedings from the Third Nordic Conference for English Studies, Vol. 1. Stockholm: Almqvist
and Wiskell. 379–393.
Wårvik, Brita. 1995. The ambiguous adverbial/conjunctions þa and þonne in Middle English. In:
Andreas H. Jucker (ed.), Historical Pragmatics. Amsterdam/Philadelphia: John Benjamins.
345–357.
Watts, Richard J. 1988. A relevance-theoretic approach to commentary pragmatic markers: The
case of actually, really, and basically. Acta Lintuistica Hungarica 38: 235–260.
Weinrich, Uriel 1977. Modern English-Yiddish Yiddish-English Dictionary. New York: Shocken
Books.
Wittgenstein, Ludwig. 1922. Tractatus Logico-Philosophicus. Trans. by D. F. Pears and B. F.
McGuinness. London: Routledge and Kegan Paul.
Wittgenstein, Ludwig. 1958. Philosophical Investigations. Trans. by G.E.M. Anscombe. New
York: Macmillan.
Yang, Li-Chiung. 2006. Integrating prosodic and contextual cues in the interpretation of dis-
course markers. In: Kerstin Fischer (ed.), Approaches to Discourse Particles. North Holland:
Elsevier. 265–298.
Ziegeler, Debra. 2000. The role of quantity implicatures in the grammaticalization of would.
Language Sciences 22: 27–61.
Ziv, Yael. 1998. Hebrew kaze as discourse marker and lexical hedge: Conceptual and procedural
properties. In: Andreas H. Jucker and Yael Ziv (eds.), Discourse Markers: Descriptions and
Theory. Amsterdam/Philadelphia: John Benjamins. 203–221.
Ziv, Yael. 2001. ze bixlal lo “pashut”: samaney hasiax ba'ivrit hameduberet (‘It’s not at all pashut
(‘simple’): the discourse markers of colloquial Hebrew). balshanut 'ivrit (Hebrew Linguis-
tics) 48: 17–29. (In Hebrew).
Zwicky, Arnold M. 1985. Clitics and particles. Language 61: 283–305.

Texts

Alper, Rogel. 2000. ke'ilu malkot yofi (‘ke'ilu Miss Universes’, lit. ‘ke'ilu beauty queens’). Ha'aretz
Daily Newspaper, 31.3.2000, p. A16.
 Bibliography 

Ben Shachar, Rina. 2000. tov 'az 'eh... kaxa... ma 'ani 'agid laxem?: 'al kasharim rekim balashon
hameduberet (‘Okay so uh... like so... what shall I say to you?: On empty connectives in the
spoken language’). Ha'aretz Daily Newspaper, 22.12.00.
Kuzar, Ron. 2000. hasafa ma-ze xaya, 'aval hamexkar kaze 'ayef ke'ilu (‘The language sure is
alive, but the research is kaze tired ke'ilu’). Ha'aretz Daily Newspaper, 12.4.00, book review
section, pp. 19, 21.
Lieblich, Amia. 2001. Gvulot ha'ishiyut (‘The boundaries of personality’). Ha'aretz Daily News-
paper, 5.9.2001, book review section.
Livneh, Neri. 2002. 'en matsav: 'idkuney sleng le'aviv 2002 (‘No way: Slang updates for Spring
2002’). matsav ha'ivrit. musaf yom ha'atsma'ut, Ha'aretz, (‘The situation of the Hebrew Lan-
guage: Independence Day supplement’), Ha'aretz Daily Newspaper, 16.4.2002.
Lori, Aviva. 2000. Shula hamevashelet (‘Shula the cook’). Ha'aretz Weekend Supplement,
24.3.2000, pp. 87–88.
Talmud, Masexet Psaxim.
Zeltzer, Elisheva. 2000. Pinat Xemed. (‘A Charming Corner’). Tel Aviv: Yedi'ot Axaronot.
 Author index

A Chafe, Wallace L.╇ 4–5, 9, 18, G

Aijmer, Karin╇ 7–8, 20, 26, 35, 20–1, 28–9, 31, 38, 52, 54, 59, Geis, Michael L.╇ 225
120, 208, 226, 228 82–3, 89–90, 95–6, 108, 150, Gergen, Kenneth J. ╇ 169, 231
Alfonzetti, Giovanna╇ 6 154, 207 Giddens, Anthony╇ 33, 232
Alper, Rogel╇ 143 Clancy, Patricia M.╇ 63, 230 Goffman, Erving╇ 2, 5, 8, 10, 21,
Altenberg, Bengt╇ 83 Claudi, Ulrike╇ 34, 38, 121, 203 54, 62, 227
Andersen, Gisle╇ 7, 133 Clover, Carol J.╇ 15, 17, 26 Goss, Emily L.╇ 6, 43, 157, 175
Ariel, Mira╇ 1, 88, 163 Coates, Jeniffer╇ 62 Grice, H. P. ╇ 225
Auer, Peter╇ 6, 36–7, 144, 185, 198, Condon, Sherri╇ 172, 174, 189–90 Gumperz, John╇ 10, 129
200, 204, 212–3, 217–8, 222, Couper-Kuhlen, Elizabeth╇ 15, Günthner, Susanne╇ 29, 35–7, 49,
226–8 218, 222 102, 161–2, 226, 228–9
Avneyon, Eitan╇ 43, 74, 79
D H
B Dahl, Östen╇ 225 Haiman, John╇ 62, 133, 160, 164,
Bakhtin, Mikhail M.╇ 29, 36, Da Matta, Roberto╇ 205 216, 222
49, 161 Dascal, Marcelo╇ 103, 114 Hansen, Mosegaard Maj-Britt╇ 7
Basgoz, Ilhan╇ 100 Dasher, Richard B.╇ 35, 37–8, 74, Heine, Bernd╇ 34, 38, 121, 203
Bateson, Gregory╇ 22 120, 169, 211, 223 Henkin, Roni╇ 1, 131–2, 137, 143
Bazzanella, Carla╇ 8 de Fina, Anna╇ 204 Heritage, John╇ 17
Beach, Wayne A.╇ 172, 174, 181, Du Bois, John W.╇ 53, 170, 232 Hilmisdóttir, Helga╇ 44
193 Hirschfeld, Ariel╇ 129, 150–1,
Becker, Alton L.╇ 1, 4, 9, 21–2, 25, E 169–70
38, 44, 54, 193, 207 Enkvist, Nils Erik╇ 15, 113 Hlavac, Jim╇ 6
Bell, Barbara╇ 133 Erickson, Frederick╇ 60 Hopper, Paul J. ╇ 1, 32–4, 36, 53,
Ben Shachar, Rina╇ 157 Estlein, Roi.╇ 9, 19–20, 35, 71, 84, 87, 108, 113, 119–21, 123, 165,
Biber, Douglas╇ 19 120, 167 201–2, 205, 210, 214–6, 219–21,
Blakemore, Dianne╇ 7 Even-Shoshan, Avraham╇ 43, 79, 224, 226–9, 232
Blum-Kulka, Shoshana╇ 26, 120, 123, 134, 137, 171, 202–3 Hopper, Robert╇ 174
54–5, 100, 230 Even-Zohar, Itamar╇ 1, 8 Hudson, Rachel╇ 133, 166
Blythe, Carl Jr.╇ 133 F Hünnemeyer, Friederike╇ 34, 38,
Bolozky, Shmuel╇ 185 Ferrara, Kathleen╇ 20, 34, 62, 83, 121, 203
Borker, Ruth A.╇ 62 90, 96, 100, 106, 110, 119–20,
Brinton, Laurel J.╇ 1, 8, 15–6, I
123, 133, 202, 211, 215, 231 Imo, Wolfgang╇ 36
34–5, 38, 112–3, 203, 220–1 Finegan, Edward╇ 19
Brody, Jill╇ 6, 214 Finell, Anne╇ 34–5 J
Brown, Gillian╇ 95 Fischer, Kerstin╇ 1, 7, 20–1 Jefferson, Gail╇ 25, 55, 193, 203,
Brown, Penelope╇ 54 Fleischman, Suzanne╇ 35, 37, 113, 218
Bublitz, Wolfram╇ 83 127, 131, 133–4, 137, 144, 147, Jucker, Andreas H.╇ 1, 7, 35, 120,
Butters, Ronald╇ 133 152, 160, 164–6, 169, 216, 232 133
Bybee, Joan L. ╇ 34, 36, 84, 87, Ford, Cecilia E.╇ 2, 5, 36, 55, 62–3
120–1, 123, 210, 222, 224–5 Fox, Barbara A.╇ 50, 212 K
Fraser, Bruce╇ 7 Kärkkäinen, Elise╇ 36
C Katriel, Tamar╇ 3–4, 26, 54–5,
Cassidy, Frederick╇ 176 100, 103, 114, 125, 160, 166–9,
205, 230–1
 Metalanguage in Interaction

Keenan, Elinor Ochs╇ 95 Oring, Elliot╇ 125, 231 Stubbs, Michael╇ 8

Keevallik, Leelo╇ 36 Ösch-Serra, Cecilia╇ 6 Suzuki, Ryoko╇ 35
Kimmerling, Baruch╇ 125, 167 Östman, Jan-Ola╇ 8 Sweetser, Eve╇ 52, 75, 120
König, Ekkehard╇ 35, 52, 120, 223 Owen, Marion╇ 83 Syder, Frances╇ 112
Krupik, Shani╇ 21
Kuryłowicz, Jerzy.╇ 34 P T
Kussmaul, Paul╇ 83 Pagliuca, William╇ 36, 87, 121, Tabor, Whitney╇ 20, 34–5, 83,
210, 223–4 120–1, 211
L Pawley, Andrew╇ 112 Tagliamonte, Sali╇ 133, 166
Labov, William╇ 11, 54, 71, 80, Pekarek Doehler, Simona╇ 36 Takahara, Paul, O.╇ 83, 99
96, 108, 112, 116, 146, 150–1, Perkins, Revere╇ 36, 210 Talmud, Masexet Psaxim.╇ 145
190, 196 Pons Bordería╇ 35 Tannen, Deborah╇ 28, 44, 48, 55,
Lambrecht, Knud╇ 143, 144 62, 71, 100, 133, 140, 150–1, 161,
Lange, Deborah╇ 34, 133 Q 175, 178, 231
Laury, Ritva╇ 36 Quinn, Arthur╇ 150 Tao, Hongyin╇ 123
Lehman, Christian╇ 34 R Thompson, Sandra A. ╇ 2, 5,
Lenk, Uta╇ 83, 104 Ramat, Paolo╇ 19 19–20, 32, 36, 44, 55, 113, 123,
Lerner, Gene╇ 14, 43, 62, 178 Ravid, Dorit╇ 89, 115, 212 214, 218, 221–2, 226–9
Levenston, Edward A.╇ 171 Read, Allen Walker╇ 61–2, 85–6, Traugott, Elizabeth Closs.╇ 8, 20,
Levinson, Stephen╇ 54 109, 176 33–9, 52, 74–5, 83–4, 119–21,
Lieblich, Amia 122 Redeker, Gisela╇ 8 127, 132, 165, 169, 178, 201–3,
Lindström, Jan╇ 35 Reh, Mechthild╇ 34 210–1, 213, 215–6, 220–6
Livnat, Zohar╇ 1 Reinhart, Tanya╇ 112
Livneh, Neri╇ 161 U
Romaine, Suzanne╇ 34, 133 Uhmann, Susanne╇ 212
Li Wei╇ 6 Rosten, Leo╇ 43–4
Lönnroth, Lars╇ 15, 17, 26 Underhill, Robert╇ 133, 144
Lori, Aviva╇ 130 S V
Lucy, John╇ 133 Sacks, Harvey╇ 5, 10, 25, 55, 66, Vincent, Diane╇ 35
174, 180 Visconti, Jacqueline╇ 35
M Salmons, Joseph C.╇ 6, 43, 157,
Maltz, Daniel N.╇ 62 Voloshin, Oleg╇ 44
175
Manoliu, Maria╇ 35 Samet, Gideon╇ 129 W
Marcuse, Herbert╇ 159, 160 Schegloff, Emanuel A.╇ 25, 36, Waida, Yukiko╇ 83
Maschler, Yael╇ 1–2, 4, 6, 9–11, 44, 52, 55, 59, 63–4, 66–7, 174 Wårvik, Brita╇ 15, 35, 113
17–20, 22–3, 26–30, 35, 43–5, Schieffelin, Bambi╇ 95 Watts, Richard J.╇ 7, 120
53, 55, 59, 71, 80, 82–4, 100, 112, Schiffrin, Deborah╇ 1, 7–8, 16–7, Weinert, Regina╇ 133
117–8, 120, 122, 127, 130–3, 137, 26, 55, 62, 83, 90, 100, 204 Wide, Camilla╇ 35
147, 152–4, 157–60, 163, 166–7, Schourup, Lawrence C.╇ 1, 8, 36, Wilson, Deirdre╇ 7
175–6, 180, 207–9, 211–4, 230 83, 133, 137 Wittgenstein, Ludwig╇ 170
Matras, Yaron╇ 6 Schwenter, Scott A.╇ 35
Mazo, Natalie╇ 44 Scollon, Ron╇ 22 Y
McConvell, Patrick╇ 6 Seppänen, Eeva-Leena╇ 36 Yaguello, Marina╇ 35, 37, 127, 131,
Meillet, Antoine╇ 34 Shaer, Susan╇ 163 133–4, 137, 144, 147, 152, 160,
Merritt, Marylin╇ 172, 174, 205 Sherzer, Joel╇ 204–5, 230 164–6, 169, 216, 232
Miller, Jim╇ 9, 133 Shloush, Shelley╇ 1, 79, 84, 86, 94 Yang, Li-Chiung╇ 8, 20
Milroy, Lesley╇ 6 Shultz, Jeffrey╇ 60 Yule, George╇ 95
Miracle, W. Charles╇ 204 Sivan, Reuben╇ 171 Z
Moyer, Melissa╇ 6 Smith, Sara W.╇ 133 Zeltzer, Elisheva╇ 130
Mulac, Anthony╇ 36 Sorjonen, Marja-Leena╇ 44, Ziegeler, Debra╇ 87
Mulder, Jean╇ 19–20 54, 67 Ziv, Yael╇ 1
O Sperber, Dan╇ 7 Zwicky, Arnold M.╇ 8, 225
Onodera, Noriko O.╇ 35, 220 Stroud, Christopher╇ 6
 Subject index

A appeal intonation╇ 53, 60 Biblical Hebrew language╇ 171,

a priori grammar╇ 1, 239 appositional use of bekitsur╇ 86 202, 217
acceptance of state of approximative meaning╇ 137 bien (French)╇ 204, 236
things╇ 177, 197, 217, 224 Arabic language╇ 84, 106, 157, biktsara╇ 79, 123
action: order of actions╇ 22 159, 169, 171 bilingualism╇ 6, 238, 241–3
returning to main╇ 22, 83, 90, Aramaic language╇ 171 bilingual conversation╇ 2, 10,
94, 96 argument genre╇ 55–8, 62, 70, 16, 241–3
structure╇ 26; 185–7, 245 bilingual discourse╇ 6–8, 35,
actually╇ 120, 234, 240, 248 argument structure╇ 32, 247 214, 238
additive╇ 90 as if╇ 39, 127, 128, 134, 136, 143, bonding╇ 221, 222
adjacency pair╇ 5, 10–1, 29, 66 164, 168, 208, 209, 216, 225, borrowing╇ 6, 43, 243, 244
first pair part╇ 5, 66–7, 227 242 boundary╇ 2, 5–6, 10–1, 14–5, 26,
second pair part╇ 13, 66–7 aside╇ 83, 106 29–32, 51, 54, 63, 70–1, 87, 122,
adverbial clause╇ 62, 63 at any rate╇ 106, 119, 124 174, 180, 190, 209, 212–4, 230,
in American English conver- attentiveness╇ 3 242, 250
sation╇ 62 audience involvement╇ 54, 112 level-I╇ 11
adverbs: in text╇ 19 see also involvement in level-II╇ 11, 14
of space and time╇ 19 discourse level-III╇ 11, 14, 15, 30
manner predicate adverbs╇ 19 automatization╇ 222 lower-level╇ 30, 87
prototypical╇ 19 average frequency╇ 84, 133, 176 higher-level╇ 11
syntactic category of╇ 19 Brazilian thumbs-up ges-
agreement (discourse)╇ 22, 112, B ture╇ 205, 230, 247
115, 120, 167, 171, 174–7, 185, backgrounding╇ 108, 112, 123, 215
187, 189, 198, 200, 202–4, 208, basically╇ 120 C
217, 219–20, 224, 231 bekitsur╇ 79–125, 191, 200–2, causality╇ 226, 241
emphatic agreement╇ 167–8 208–10, 214–6, 220–5, 227, cause╇ 22, 226, 233, 236, 238
ironic╇ 177, 187, 208, 217, 220, 230–1 center of interest╇ 29, 154
227 appositional use of╇ 86 change in volume╇ 104, 124, 210
tokens of╇ 26 backgrounding function change-of-state token╇ 17, 239
agreement (grammatical)╇ 89, of╇ 123 chat (internet)╇ 85–6, 240
115, 163 episode-summarzing func- Chinese language╇ 63, 204–5,
allowing╇ 64, 66, 68, 75 tion of╇ 121 231, 241, 243
allowing reluctantly╇ 68, 75 equative use of╇ 86 claim of understanding╇ 63
amazement╇ 22 foregrounding function clarification╇ 13, 82, 94, 114,
although╇ 101–2 of╇ 224, 227 154–6, 181, 195
American accent in Hebrew╇ 5, grammaticization of╇ 87, clarification question╇ 13
105–6 119–124, 214–29 clarification sequence╇ 82, 94,
anyhoo╇ 106 resumptive╇ 96, 99, 103, 110, 114, 154, 195
anyhow╇ 106, 119 121, 231 classroom discourse╇ 41, 58
anyway╇ 13–4, 17, 39, 83, 84, 90, summarizing╇ 121 clause combining╇ 226, 235, 239,
96, 99, 100, 106, 110, 119, 120, besides╇ 211 241, 243
121, 124, 179, 190, 191, 194, 200, betax╇ 71, 80, 115 cline of de-categorializa-
201, 230, 231, 237, 240, 243, 247 bexayex╇ 71 tion╇ 119–20, 202, 215–6, 221
anyways╇ 106, 119 bexol mikre╇ 124, 230 closeness vs. distance╇ 22
aposiopesis╇ 148, 150 bexol 'ofen╇ 124, 230
 Metalanguage in Interaction

cluster╇ 6, 10–1, 14–5, 17–8, 26, contrast╇ 4, 22, 31, 62–3, 118, storyteller-initiated╇ 106
28, 30–1, 44, 71, 83, 102, 131, 164, 185, 207–8, 233, 236, 238, disagreement╇ 22, 26, 102–3, 177,
175, 180, 191, 195, 201, 208, 211, 242, 249 185, 187, 201, 205, 208, 212, 217,
213–4 controlling discourse flow╇ 54–5, 219–20, 227, 230–1, 242
of discourse markers╇ 6, 10–1, 58, 63, 70, 74 high tolerance for╇ 25, 230
26, 208 conversational action╇ 2, 5–6, tokens of╇ 22
co-construction╇ 14, 62 11, 21–2, 25–6, 83–4, 90, 114, discourse level╇ 6–7
code-switching╇ 214, 233, 239, 191, 213 discourse markers╇ 1–2, 6–11,
240, 242–3, 247 conversational action bounda- 14–23, 25–30, 32–7, 39–40, 44,
see also language alternation ries╇ 5–6, 10, 21–2, 26, 30, 71, 82–3, 120–1, 124, 131, 152–3,
cognitive constraints╇ 8, 38, 207 83, 213 156, 165, 167, 169, 180, 191, 195,
cognitive discourse marker╇ 8, conversational implica- 205, 207–15, 220–5, 227–9,
17, 28, 38, 51, 131, 152–8 ture╇ 225–6 231–3, 235, 239–42, 246–9
cognitive processes╇ 1, 5, 17, 21–2, coordination╇ 22, 25, 238 definition of╇ 2, 9, 16, 19–20,
25, 39, 44, 51, 83, 131, 152, 154, corpus╇ 7–10, 14, 18–9, 22, 36, 40, 30, 37, 223
164, 175, 207–8, 223 45–6, 50–1, 66, 71, 74, 76, 80, clusters of╇ 11, 18, 26, 71
cognitive realm╇ 5, 8, 22, 39, 84, 94, 96–7, 99–100, 103–4, distribution of╇ 10
127, 207 106, 116, 119, 120–4, 130, 132–3, employment of╇ 1, 10, 14,
come on╇ 49, 71 137, 144, 162, 175–6, 178, 180, 27, 114
communicative strategies╇ 34, 200, 202–4, 208, 210–1, 213, grammatical category
63, 120 229–30, 233, 248 of╇ 119–20
comparison, element of╇ 128, 162 counterfactual conjunction╇ 128, interactive functions of╇ 8
complementizer╇ 152, 229 134 intonation contour of╇ 20
complicating action╇ 53, 110, cross-language generaliza- intonation unit preceding╇ 20,
112–3, 190 tions╇ 127, 165 29
complication╇ 54, 108, 110, 112 cross-language pragmatics╇ 132, metalanguaging property
concession╇ 22, 177, 181, 198, 200, 165, 237 of╇ 35
208, 217, 220, 233, 236, 238 metalinguality of╇ 37
concessive discourse mark- D operational definition of╇ 17,
er╇ 102 da?╇ 161–2 20
condition╇ 22, 136, 152, 164, 166, deictics╇ 22 previous approaches to╇ 7–8,
226, 233, 236, 238 delaying narrative point╇ 63, 74 16, 27
conditional clause╇ 62–3 evaluative strategy of╇ 54 prototypical╇ 17–22, 25, 27,
conditional conjunction╇ 134, 137 dialogue: constructed╇ 28–32, 30–1, 44–5, 83, 124, 131–2,
conjunction of hypothetical 48–9, 140, 158–62, 178, 180, 168, 175, 180, 207, 211, 213–4
conditionals╇ 128 185, 197, 213, 237, 24; 7 non-prototypical╇ 18, 20, 213
conjunctions╇ 22, 212, 239, 241, imaginary╇ 116–7, 198, 200, realms of operation of╇ 26
243, 249 204 resumptive╇ 90, 121–4, 215, 230
connectives╇ 8, 16, 233, 236, dialogical grammaticization semantic requirement for dis-
242–3, 246, 250 path╇ 218, 224 course markerhood╇ 17–9,
consequence╇ 22, 25 digression╇ 43, 52, 79, 84, 87, 90, 30, 87, 168, 175, 211, 213
constraints╇ 7, 22, 25–6, 29, 94–7, 99–100, 102–4, 106, 108, structural patterning
39, 59, 87, 140, 150, 166, 205, 110, 114, 116, 119, 121, 123, 125, of╇ 27–33, 213
207–8, 212, 224, 234–5 149, 191, 201, 208, 215, 223, 231, structural properties of╇ 2, 17,
contexts of parody╇ 142 234, 236 27, 33, 207, 210–1
contextual realms╇ 21–2, 25–6, closing a╇ 22 structural requirement for
38–9, 193, 207, 208 collaborative nature of discourse markerhood╇ 17–
contextualization cues╇ 10 returning from╇ 99 20, 22, 27, 30, 45, 87, 124,
continuer╇ 52, 59, 60–4, 71 content-based╇ 104 168, 175, 180, 211, 213
continuing intonation, xi╇ 17–8, language games of╇ 100 of summarizing╇ 121
28–9, 44, 49, 60–1, 83, 86–7, listener-initiated╇ 104 as syntactic category╇ 7
94, 121–2, 124, 150, 153–5, 175, listener-triggered╇ 97 system of╇ 207
180, 211, 213 returning from╇ 84, 90, 94, textual╇ 17, 19, 22, 203, 205,
100, 231 208, 218, 224
 Subject index 

discourse operators╇ 16 figure of speech╇ 150 go-ahead token╇ 66–7

discourse package╇ 15, 31, 180 filler╇ 5, 6, 21, 129, 157–8 grammatical category
discourse particles╇ 16, 26, 127, nativelike usage of╇ 5 intermediate╇ 119–20, 215–6
165, 233–4, 237, 241, 246, 249 fine╇ 40, 171, 184–5, 187, 198, 204 major╇ 119, 202, 215–6
discourse topic╇ 5, 11, 21, 52, 94–6 Finnish particles╇ 67 minor╇ 119, 202, 215–6, 221
highlighting new╇ 119 first pair part╇ 5, 66–7, 227 grammaticalization╇ 33–4, 87,
main╇ 66–7, 83–4, 87, 90, 95, floor╇ 52–3, 55, 63, 150 120, 165, 200, 205, 220, 224,
97, 99–100, 103–4, 106, 121, focus╇ 143–5, 216 226, 233, 235–6, 238–9, 241,
124, 191, 200–1, 208 focus marker╇ 133, 143, 145, 147, 243–4, 247–9
returning to main╇ 22, 90 165 chains of╇ 34
supertopic╇ 30 focus of consciousness╇ 29 discourse-pragmatic perspec-
disjunction╇ 22 footing╇ 10, 21, 62 tive on╇ 34
dismissive╇ 90 foregrounding╇ 112, 116, 119, 124, historical-diachronic per-
domains╇ 26, 167, 222 208, 210, 215, 223–4, 227, 239 spective on╇ 34
dome le-╇ 134, 137 formulaic utterances╇ 15, 112, grammaticization╇ 1–2, 7, 20, 25,
dugri╇ 54, 125, 166–9, 230–2, 204, 228 30, 33–9, 45, 51, 74–5, 84, 87,
240, 242 concluding formulas╇ 15 119–21, 123–4, 132, 165, 169–70,
introductory saga formu- 176, 193, 200–5, 207, 214–22,
E las╇ 15 224–6, 229–30, 232–3, 235,
element of comparison╇ 128, 162 narrative direction formu- 242–3, 247–8
elicitational discourse╇ 89, 96 las╇ 15 of discourse markers╇ 2, 25,
elicitational topic╇ 188, 208 forthrightness╇ 125 33–5, 37, 39, 200, 214
embedded stories╇ 112–3 frame╇ 2, 5–8, 10, 14–5, 17, 21, discourse-pragmatic perspec-
emergent grammar╇ 1, 34–5, 120 25, 28, 30–3, 36, 51, 54, 207–9, tive on╇ 34
emic perspective of partici- 212–4, 222, 225, 227–9, 237, historical-diachronic per-
pants╇ 2, 207 242, 244 spective on╇ 34
ending an action╇ 22, 193 frame shifts╇ 5–7, 10, 17, 21, 25, locus of╇ 198
English language╇ 4–6, 8, 17, 19, 30–1, 33, 51, 207–9, 212–4, 225, unidirectionality in╇ 119, 201
28, 34–5, 37, 41, 43, 53, 59, 61, 227, 229, 242 of Swedish particles╇ 35
63, 83–4, 86, 92, 99–100, 110, in interaction╇ 7, 30, 207, 213 grammaticization path╇ 36, 45,
116, 118, 121, 124, 127–8, 130–1, moment of╇ 21–2, 207 121, 123, 169, 176, 193, 204, 210,
133–4, 152, 160, 164–6, 171–6, framing device╇ 212, 228 214–7, 220, 224
185, 187, 193, 204, 209, 211–3, frankly╇ 120 granting permission╇ 64, 67–8,
220, 226, 228, 230–1, 233–41, French language╇ 34–5, 127, 74–5, 208, 214–5
243, 246–9 133–4, 160, 165–6, 236, 238, grounding╇ 113–4, 235, 237, 248
enthusiasm╇ 22, 58, 64, 74, 209, 243–4, 249 gut (German)╇ 185
230 frequency╇ 26, 40, 58, 63, 100,
lack of╇ 22, 74 123, 202, 235, 248 H
episode╇ 5, 10, 53–4, 82, 87, 90–1, functional itinerary╇ 36, 74, 120, Haifa Corpus of Spoken Israeli
94, 99, 108, 110–2, 121, 124, 124, 132–4, 148, 152, 162, 165–6, Hebrew╇ 9, 45, 84, 132, 176,
154, 158, 193, 196–7, 200, 204, 205, 215, 220–1, 225 242
215, 218–9 hakitser╇ 82, 104, 108
episode boundaries╇ 5, 110 G hao (Chinese)╇ 204, 243
sub-episodes╇ 11, 29 Gemara╇ 145 harey╇ 60, 88
evaluation╇ 29, 49, 69, 110, 112 genre╇ 26, 33, 35–6, 41, 52, 59, 122 hastening╇ 43, 46, 48, 54, 58, 62,
exchange structure╇ 26, 226 genre (French)╇ 127, 133–4, 160, 68, 74–5, 208, 214–15, 223, 230
extralingual world╇ 1–2, 17, 21–2, 165–6 hastening non-verbal action╇ 74,
44, 49, 83, 121, 131, 152, 164, German language╇ 35, 41, 44, 214–5
168, 175, 207, 224 166, 185, 198, 200, 204–5, hastening word╇ 43
extraposition╇ 226–7 212–3, 218, 226, 228, 233, hava╇ 43
237–8, 240, 244 Hebrew preposition of compari-
F Germanic languages╇ 233, 241 son╇ 128
face wants╇ 54, 169 given information╇ 62 Hebrew-English bilingual con-
fei (Bavarian)╇ 185 go-ahead response╇ 44, 67, 246 versation╇ 130, 241–2
 Metalanguage in Interaction

hedging╇ 131–3, 137, 140–3, 146, interpersonal realm╇ 4, 8, 18, 25, Japanese discourse markers╇ 99,
148, 151–2, 162–6, 208, 210, 39, 41, 44–5, 75, 208, 217 220, 243, 247
216, 231, 239 interpersonal relations╇ 9, 25, jocular form╇ 102, 106
hesitation╇ 4–6, 8, 21 208 joking╇ 29, 31–2, 68–71, 74–6,
markers of╇ 5, 6, 21 interrupted action╇ 190, 208, 218 106, 174, 208, 214, 224, 230–1
utterance╇ 4 interruption╇ 13–4, 66, 103, 183,
hierarchy: of episodes╇ 112 190–1, 200 K
of frame shifts╇ 209 intersubjectification╇ 169, 224, k(e)-╇ 128, 134, 137, 162–4, 166,
high involvement in dis- 248 216
course╇ 53–4, 64, 71, 74, 100, intersubjectivity╇ 37–8 katsar╇ 79, 230
102, 150–1, 158, 230–1 intonation contour╇ 9, 18, 20–1, kaze╇ 12–3, 53, 86, 111, 127–32,
high onsets╇ 15, 236 28–9, 31, 44, 48, 53, 59, 83, 86, 147, 151–2, 166–7, 169–0, 231–2,
humor: Israeli╇ 71, 243 90, 94, 99, 121, 154, 175, 211 238–9, 242, 244, 249–50
mixed with impatience╇ 48 sentence-final╇ 29, 154 kaze ke'ilu generation╇ 129, 244
humorous sarcasm╇ 106 intonation unit╇ 4, 9, 14, 17–8, ke'ilu╇ 127–170, 216, 220–5
hypotaxis╇ 229 20–1, 28–32, 45, 59, 68, 82–3, as a focus maker╇ 146
hypothetical conditional╇ 134, 90, 95, 99, 121, 124, 153–4, 168, functional itinerary of╇ 162–6
164 210–1, 213–4, 235 grammaticization of╇ 162–6,
intonation-unit final posi- 214–29
I tion╇ 19–20, 29 in hedging╇ 137, 140–3, 146,
Icelandic language╇ 239 intonation-unit initial posi- 163
iconic strategy╇ 4, 118, 122 tion╇ 17–8, 20, 28, 44–5, 83, quotative function of╇ 161
iconicity╇ 54, 214, 236, 242 130–1, 132, 135, 152–3, 168, 175, ken╇ 4, 18, 40, 45, 59–60, 70, 86,
idea unit╇ 21 180, 210–1 91–3, 99, 101, 104, 107, 113,
ideational structure╇ 26 intonation-unit non-initial 171, 179
imaginary dialogue╇ 116–8, 200, position╇ 17, 44, 83, 132 key╇ 45, 49, 68, 69–71, 74–7,
204 intonation-unit marginal posi- 111–2, 157, 162, 209, 214, 222–3,
impatience╇ 41, 45, 48, 54, 63–4, tion╇ 19 230–1, 240
67, 71, 74, 229–30, 242; invited inferences╇ 225, 237 key words╇ 4, 240
impoliteness╇ 43, 58, 75, 223 Invited Inferencing Theory of kitsur╇ 79, 113, 120, 215
in fact╇ 120, 211, 240 Semantic Change╇ 225 kivyaxol╇ 137
in short╇ 39, 59, 79, 85–90, 111–2, involvement in discourse╇ 45, klomar╇ 157
121–2, 124, 128, 140, 157, 179, 54–5, 71, 74,100–2, 150, 230 kmo 'ilu╇ 134, 137
209, 215, 222, 230, 247 ironic agreement╇ 177, 187, 208,
indeed╇ 19, 120, 167, 211, 242 L
217, 220, 227 lamrot she-╇ 102
inferences╇ 36, 210, 226 irony╇ 185, 187
information flow╇ 95 language alternation╇ 2, 5–6, 118,
double-voiced ironic quota- 214, 242
information state╇ 26 tions╇ 160–2
insertion sequences╇ 103, 114 theory of╇ 4–7, 37
irrealis╇ 128, 134, 137, 162, 164, at discourse markers╇ 2–6, 10,
interacting as an Israeli╇ 40, 216, 225, 237
124–5, 166–70, 229–32 43, 157, 175, 214,
Israeli culture and society╇ 2–3, languaging╇ 1–2, 21, 25, 38, 39
interactional projection╇ 212–3, 26, 39–41, 43–5, 54–5, 67,
222, 229 layering of voices╇ 29, 49, 238
71, 74, 84, 91, 93–4, 100, 116, lekatser╇ 79
interactional structure╇ 217–20, 124–5, 129, 134, 142, 156, 166–7,
227 lesaxek beke'ilu╇ 143
169–70, 178–80, 195, 205, let us╇ 27, 43
interjection╇ 43, 53, 92, 221 229–32, 234–5, 240, 242–4
interpersonal constraints╇ 21, levels of discourse╇ 1, 152
Israeli discourse╇ 45, 55, 67, 74 lexical history╇ 87, 127, 166, 205,
140–1, 150–1 Israeli humor╇ 71, 243
interpersonal discourse 224–5
Italian language╇ 35, 166, 233–4, like╇ 12–3, 16, 27–8, 36, 39, 52–3,
marker╇ 17, 40, 45, 51, 59, 74, 243, 249
99, 203, 208, 216 80–1, 86, 88–9, 111, 128, 130–1,
interpersonal involvement╇ 45, J 133–5, 139–41, 144–7, 151–2,
54–5, 71, 74,100–2, 150, 230 Japanese language╇ 35, 63, 99, 155–66, 180–1, 189–90, 199,
166, 211–2, 220, 236, 243, 247 204
 Subject index 

like this╇ 69, 76, 127–8, 135, 238 N perception verbs╇ 22

lip service╇ 181, 183, 201, 217 narration╇ 14, 29, 104, 106 permission╇ 64, 66, 68, 196
lo╇ 4, 12–3, 16–8, 27–8, 48, 56–8, narrational discourse╇ 96, 188 persistence╇ 87, 121, 165, 205, 224
65, 67, 69, 70, 76, 79, 82, 89, 91, non-narrational dis- phonological changes╇ 83, 120
100–2, 107, 111, 114, 122–3, 129, course╇ 52, 55, 63, 71 phonological expansion╇ 221
142, 147, 149, 156–8, 163, 168, narrative╇ 5, 9–11, 13–5, 35, 43, phonological reduction╇ 210, 221
173, 186–7, 189–90, 198, 201, 49, 52–4, 59–61, 63, 74, 80, playful forms╇ 84, 221
209, 249 82–3, 96, 99, 104, 108, 110, playful variation╇ 106
112, 116, 118, 123–4, 146–7, 150, playfulness╇ 106
M 180–1, 188–91, 193, 196–7, 200, point of incompletion╇ 61
madhim╇ 149–50 208, 215, 218–9, 223, 227, 229, point of speaker change╇ 17,
maintaining contact╇ 22 234–7, 244, 247 44–5, 83, 168, 175, 211
manner predicate adverbs╇ 19 N-be-that╇ 228–9, 238 polyphony╇ 29, 161, 238
markers of pragmatic struc- need to rephrase╇ 127, 157 see also multivocality, multi-
ture╇ 16 negotiating: footing╇ 10 plicity of voices
meta-knowledge╇ 27 relations between speaker polysemies╇ 210, 222, 226
metalanguaging╇ 1–2, 5, 16, 35, and hearer╇ 4, 38 postmodernity╇ 169–170, 231
207, 242 new information╇ 5, 17–8 pragmatic markers╇ 16, 35, 235,
metalingual dimension╇ 2, 4, 121, nii (Swedish)╇ 54, 67 249
123, 130, 211 no (Swedish)╇ 67 pragmatic presupposition╇ 144
boundary with lingual nonetheless╇ 90 pre-front field╇ 198, 212, 228, 233
dimension╇ 5, 51 non-verbal action╇ 46, 48, 51–2, pre-pre╇ 66–8
metalingual frame╇ 6–7 74–5, 208 pretend play╇ 143
of discourse markers╇ 6, 37 Northern Tel Aviv╇ 129–30 processing information╇ 22
metalingual interpretation╇ 17, nu╇ 41–77, 91, 99, 105–6, 115–6, projectability╇ 214, 225–6, 239
37, 44, 83, 175 118–9, 156, 183–4, 192, 208–10, projecting construction╇ 227, 229
metalingual utterance╇ 4–6, 214–5, 220–4, 229–31 projection╇ 62, 212–3, 222, 226–7,
11, 15–7, 30–1, 37, 38, 75–7, as a continuer╇ 61 233, 244
87, 121–3, 164, 174, 180, 203, in keying╇ 71, 76–7, 210, 221–2, double╇ 213
209–10, 213–8, 221–5, 227–9 224 retrospective╇ 222
metalinguality╇ 37, 211, 213 sequential functions syntactic╇ 212, 222
semantic property of╇ 17 of╇ 52–68, 75, 221 propositional meaning╇ 38, 52,
metaphor╇ 225, 231, 237, 244 grammaticization of╇ 74–77, 120, 203, 223, 248
space-discourse╇ 121 214–29 prosodic sentence╇ 59–61, 63,
metaphorical extension╇ 200 nu be'emet╇ 71 90, 94, 99
metaphorical shift╇ 51, 120–1, 223 nur (German)╇ 212–3 prosody╇ 20–1, 48, 66, 84, 151,
meta-talk expressions╇ 16 161, 211
metonymy╇ 225–6 O provoking╇ 68–70, 74–5, 208,
Midrash╇ 145 OK: as a discourse marker╇ 174, 224, 230
milat zeruz╇ 43 189–90 pseudocleft╇ 226–7
minimal acknowledgment╇ 193 orientation╇ 11, 14, 61, 80, 82–3,
mm hm╇ 59 108, 110, 112, 146, 153, 190, Q
mockery╇ 106 196, 200 quotations╇ 158, 160, 163, 216
mocking╇ 70, 75, 106, 116, 230 original lexical sources╇ 87, 166, double-voiced ironic╇ 160–2
monolingual discourse╇ 6 205, 224 quotative╇ 134, 160–2, 165–6, 210,
monolingual speakers╇ 6 overlap╇ 38, 55, 59, 62, 136, 158, 216, 235, 247
multiplicity of voices╇ 169–70, 183, 195
231–232 R
see also multivocality, po- P reactive token╇ 63, 236
lyphony paralinguistics╇ 9, 162 realizing new information╇ 17, 22
multi-unit turn╇ 15 parallelism╇ 55, 62, 238 really╇ 3, 19, 110, 120, 139, 167,
multivocality╇ 27, 29, 44, 175, 242 parataxis╇ 229 242, 248
see also multiplicity of voices, participation framework╇ 26 referential realm╇ 22
polyphony patronizing╇ 43, 58 register╇ 79, 123–4, 160
 Metalanguage in Interaction

high╇ 43, 79, 202 so what╇ 41, 71–4, 115–6, 150 interpersonal╇ 176, 188, 193,
relations of speaker to text╇ 19 space-discourse metaphor╇ 121 217
relevance╇ 7, 16, 84, 233–4, 240, Spanish language╇ 35, 204–5, quasi-conjunctional function
245–6, 249 235–6, 244 of╇ 200
theory of╇ 7, 84, 246 speaker change╇ 17–8, 44–5, 83 textual╇ 188, 193
repetition╇ 7, 66, 154, 156, 158, 168, 175, 211 transitional╇ 188–91
222, 225, 247 stance╇ 19, 37, 59, 64, 162, 224, trajectories╇ 38, 212
rephrasal╇ 14, 28–9, 58, 111, 131, 234, 240 transcription╇ 9, 236
152, 154, 157, 163–4, 167, 170, of uncertainty╇ 17 conventions╇ 9
208, 210, 216 story aside╇ 106 transition╇ 15, 55, 174–6, 188,
realizing the need for╇ 22, 28, stress╇ 9, 90, 129 190–1, 195–6, 198, 200–1, 203,
131, 152, 156 position of╇ 84, 90 218–20, 242
response signals╇ 16 structural realm╇ 22, 208 into following episode/ac-
resultative meaning╇ 68–71 structuration╇ 33, 232, 237 tion╇ 195–6, 198
retention╇ 87, 121, 224 subjectification╇ 169, 224, 247 expected╇ 174, 188, 204, 218,
retroactive construction╇ 108, subordinator╇ 135–6, 160, 185 223–4
116 succinctly╇ 79, 87, 119, 121–4, 209, transition relevance place╇ 55
returning from digression╇ 84, 215, 225, 227, 230 trope╇ 150
90, 94, 100, 231 summarizing╇ 22, 85, 87, 90, 94, turn╇ 11, 15, 29, 55, 58–9, 67, 177,
returning to main action╇ 22, 96, 99, 121–2, 124, 208, 210, 215 180, 193, 195–6, 198, 208, 217,
90, 94, 96 summarizing a list╇ 85, 90, 210 237, 244, 246–7
rhythmic synchrony╇ 60 summarizing an action╇ 22 turn constructional unit╇ 15, 55
ridiculing╇ 29, 49, 70, 73, 75, 116, summarizing an episode╇ 85,
161, 185 90, 210 U
Russian language╇ 43–4, 51, 73, superfoci of consciousness╇ 29 Ugaritic language╇ 171
75, 214, 243 sure, of course╇ 71 uh huh╇ 59, 246
speakers in Israel╇ 44 SVO language╇ 212 urging╇ 22, 25, 39, 43–5, 50–2, 55,
Swedish language╇ 35, 166, 241 58–9, 62–3, 67–8, 74–6, 99,
S syntactic gestalt╇ 200 208, 214–5
Sabra╇ 125, 166–7, 170, 230–1, 240 syntactic projection╇ 212, 222 uvxen╇ 43
same-speaker talk╇ 17–8, 28, 44,
83, 131, 175, 180, 211 T V
scalar category╇ 20 Talmudic Hebrew language╇ 78, ve-╇ 17, 29, 53, 60, 71, 81, 86, 88–9,
scope╇ 7, 37–8, 147, 211, 247 134, 137 107, 139, 140, 149, 180
second pair part╇ 13, 66–7 tense and aspect: switch in╇ 53 verbal noun╇ 79, 120, 215, 221
self-mockery╇ 141–2 textual realm╇ 4, 17, 25, 44, 39, 79, verbs of saying╇ 22
self-rephrasal╇ 28–9, 39, 111, 83, 164, 175, 193, 208 verbum dicendi╇ 31, 159–60, 180
131–3, 152, 154–6, 158, 162–4, the thing is╇ 228 W
166, 208–10, 213, 216, 223–4 theory of grammaticization and well╇ 43–44, 120, 204, 237, 240
semantic change╇ 37, 169, 211, semantic change╇ 37–9, 225–6 well?╇ 41
248–9 therefore╇ 43, 203, 209, 217–8, 226 well then╇ 43, 204
semantic-pragmatic path third turn receipt╇ 180, 193, 208, Wenn-clauses (German)╇ 226,
see grammaticization path, 217 233
functional itinerary tolerance for tentativeness╇ 167, wh-cleft╇ 226
sequences╇ 103, 154, 174, 226, 232
244, 246 topic Y
shifting╇ 8, 21, 31, 36, 171, 193, see discourse topic yeah╇ 3–4, 39–40, 45, 59, 70, 86,
203, 208, 213, 227 tov╇ 13, 40, 43, 88, 102, 171–205, 91, 99, 101, 104, 107, 113, 135,
side comments╇ 13–4 208–9, 211, 214, 216–27, 160, 184, 194
similar to╇ 47, 134, 137, 143 229–31, 250 Yiddish language╇ 6, 43, 51, 75,
slang╇ 82, 129, 139, 147, 154, 157, grammaticization of╇ 201–5, 84, 214, 244, 248
161, 178, 241, 249 214–29
Pragmatics & Beyond New Series
A complete list of titles in this series can be found on www.benjamins.com

193 Suomela-Salmi, Eija and Fred Dervin (eds.): Cross-Linguistic and Cross-Cultural Perspectives on
Academic Discourse. vi, 297 pp. + index. Expected December 2009
192 Filipi, Anna: Toddler and Parent Interaction. The organisation of gaze, pointing and vocalisation.
xiii, 265 pp. + index. Expected December 2009
191 Ogiermann, Eva: On Apologising in Negative and Positive Politeness Cultures. x, 291 pp. + index.
Expected November 2009
190 Finch, Jason, Martin Gill, Anthony Johnson, Iris Lindahl-Raittila, Inna Lindgren,
Tuija Virtanen and Brita Wårvik (eds.): Humane Readings. Essays on literary mediation and
communication in honour of Roger D. Sell. xi, 156 pp. + index. Expected November 2009
189 Peikola, Matti, Janne Skaffari and Sanna-Kaisa Tanskanen (eds.): Instructional Writing in
English. Studies in honour of Risto Hiltunen. 2009. xiii, 240 pp.
188 Giltrow, Janet and Dieter Stein (eds.): Genres in the Internet. Issues in the theory of genre.
ix, 290 pp. + index. Expected November 2009
187 Jucker, Andreas H. (ed.): Early Modern English News Discourse. Newspapers, pamphlets and scientific
news discourse. 2009. vii, 227 pp.
186 Callies, Marcus: Information Highlighting in Advanced Learner English. The syntax–pragmatics
interface in second language acquisition. 2009. xviii, 293 pp.
185 Mazzon, Gabriella: Interactive Dialogue Sequences in Middle English Drama. 2009. ix, 228 pp.
184 Stenström, Anna-Brita and Annette Myre Jørgensen (eds.): Youngspeak in a Multilingual
Perspective. 2009. vi, 206 pp.
183 Nurmi, Arja, Minna Nevala and Minna Palander-Collin (eds.): The Language of Daily Life in
England (1400–1800). 2009. vii, 312 pp.
182 Norrick, Neal R. and Delia Chiaro (eds.): Humor in Interaction. 2009. xvii, 238 pp.
181 Maschler, Yael: Metalanguage in Interaction. Hebrew discourse markers. 2009. xvi, 258 pp.
180 Jones, Kimberly and Tsuyoshi Ono (eds.): Style Shifting in Japanese. 2008. vii, 335 pp.
179 Simões Lucas Freitas, Elsa: Taboo in Advertising. 2008. xix, 214 pp.
178 Schneider, Klaus P. and Anne Barron (eds.): Variational Pragmatics. A focus on regional varieties in
pluricentric languages. 2008. vii, 371 pp.
177 Rue, Yong-Ju and Grace Zhang: Request Strategies. A comparative study in Mandarin Chinese and
Korean. 2008. xv, 320 pp.
176 Jucker, Andreas H. and Irma Taavitsainen (eds.): Speech Acts in the History of English. 2008.
viii, 318 pp.
175 Gómez González, María de los Ængeles, J. Lachlan Mackenzie and Elsa M. González
Ælvarez (eds.): Languages and Cultures in Contrast and Comparison. 2008. xxii, 364 pp.
174 Heyd, Theresa: Email Hoaxes. Form, function, genre ecology. 2008. vii, 239 pp.
173 Zanotto, Mara Sophia, Lynne Cameron and Marilda C. Cavalcanti (eds.): Confronting
Metaphor in Use. An applied linguistic approach. 2008. vii, 315 pp.
172 Benz, Anton and Peter Kühnlein (eds.): Constraints in Discourse. 2008. vii, 292 pp.
171 Félix-Brasdefer, J. César: Politeness in Mexico and the United States. A contrastive study of the
realization and perception of refusals. 2008. xiv, 195 pp.
170 Oakley, Todd and Anders Hougaard (eds.): Mental Spaces in Discourse and Interaction. 2008.
vi, 262 pp.
169 Connor, Ulla, Ed Nagelhout and William Rozycki (eds.): Contrastive Rhetoric. Reaching to
intercultural rhetoric. 2008. viii, 324 pp.
168 Proost, Kristel: Conceptual Structure in Lexical Items. The lexicalisation of communication concepts in
English, German and Dutch. 2007. xii, 304 pp.
167 Bousfield, Derek: Impoliteness in Interaction. 2008. xiii, 281 pp.
166 Nakane, Ikuko: Silence in Intercultural Communication. Perceptions and performance. 2007. xii, 240 pp.
165 Bublitz, Wolfram and Axel Hübler (eds.): Metapragmatics in Use. 2007. viii, 301 pp.
164 Englebretson, Robert (ed.): Stancetaking in Discourse. Subjectivity, evaluation, interaction. 2007.
viii, 323 pp.
163 Lytra, Vally: Play Frames and Social Identities. Contact encounters in a Greek primary school. 2007.
xii, 300 pp.
162 Fetzer, Anita (ed.): Context and Appropriateness. Micro meets macro. 2007. vi, 265 pp.
161 Celle, Agnès and Ruth Huart (eds.): Connectives as Discourse Landmarks. 2007. viii, 212 pp.
160 Fetzer, Anita and Gerda Eva Lauerbach (eds.): Political Discourse in the Media. Cross-cultural
perspectives. 2007. viii, 379 pp.
159 Maynard, Senko K.: Linguistic Creativity in Japanese Discourse. Exploring the multiplicity of self,
perspective, and voice. 2007. xvi, 356 pp.
158 Walker, Terry: Thou and You in Early Modern English Dialogues. Trials, Depositions, and Drama
Comedy. 2007. xx, 339 pp.
157 Crawford Camiciottoli, Belinda: The Language of Business Studies Lectures. A corpus-assisted
analysis. 2007. xvi, 236 pp.
156 Vega Moreno, Rosa E.: Creativity and Convention. The pragmatics of everyday figurative speech. 2007.
xii, 249 pp.
155 Hedberg, Nancy and Ron Zacharski (eds.): The Grammar–Pragmatics Interface. Essays in honor of
Jeanette K. Gundel. 2007. viii, 345 pp.
154 Hübler, Axel: The Nonverbal Shift in Early Modern English Conversation. 2007. x, 281 pp.
153 Arnovick, Leslie K.: Written Reliquaries. The resonance of orality in medieval English texts. 2006.
xii, 292 pp.
152 Warren, Martin: Features of Naturalness in Conversation. 2006. x, 272 pp.
151 Suzuki, Satoko (ed.): Emotive Communication in Japanese. 2006. x, 234 pp.
150 Busse, Beatrix: Vocative Constructions in the Language of Shakespeare. 2006. xviii, 525 pp.
149 Locher, Miriam A.: Advice Online. Advice-giving in an American Internet health column. 2006.
xvi, 277 pp.
148 Fløttum, Kjersti, Trine Dahl and Torodd Kinn: Academic Voices. Across languages and disciplines.
2006. x, 309 pp.
147 Hinrichs, Lars: Codeswitching on the Web. English and Jamaican Creole in e-mail communication.
2006. x, 302 pp.
146 Tanskanen, Sanna-Kaisa: Collaborating towards Coherence. Lexical cohesion in English discourse.
2006. ix, 192 pp.
145 Kurhila, Salla: Second Language Interaction. 2006. vii, 257 pp.
144 Bührig, Kristin and Jan D. ten Thije (eds.): Beyond Misunderstanding. Linguistic analyses of
intercultural communication. 2006. vi, 339 pp.
143 Baker, Carolyn, Michael Emmison and Alan Firth (eds.): Calling for Help. Language and social
interaction in telephone helplines. 2005. xviii, 352 pp.
142 Sidnell, Jack: Talk and Practical Epistemology. The social life of knowledge in a Caribbean community.
2005. xvi, 255 pp.
141 Zhu, Yunxia: Written Communication across Cultures. A sociocognitive perspective on business genres.
2005. xviii, 216 pp.
140 Butler, Christopher S., María de los Ængeles Gómez González and Susana M. Doval-Suárez
(eds.): The Dynamics of Language Use. Functional and contrastive perspectives. 2005. xvi, 413 pp.
139 Lakoff, Robin T. and Sachiko Ide (eds.): Broadening the Horizon of Linguistic Politeness. 2005.
xii, 342 pp.
138 Müller, Simone: Discourse Markers in Native and Non-native English Discourse. 2005. xviii, 290 pp.
137 Morita, Emi: Negotiation of Contingent Talk. The Japanese interactional particles ne and sa. 2005.
xvi, 240 pp.
136 Sassen, Claudia: Linguistic Dimensions of Crisis Talk. Formalising structures in a controlled language.
2005. ix, 230 pp.
135 Archer, Dawn: Questions and Answers in the English Courtroom (1640–1760). A sociopragmatic
analysis. 2005. xiv, 374 pp.
134 Skaffari, Janne, Matti Peikola, Ruth Carroll, Risto Hiltunen and Brita Wårvik (eds.):
Opening Windows on Texts and Discourses of the Past. 2005. x, 418 pp.
133 Marnette, Sophie: Speech and Thought Presentation in French. Concepts and strategies. 2005.
xiv, 379 pp.
132 Onodera, Noriko O.: Japanese Discourse Markers. Synchronic and diachronic discourse analysis. 2004.
xiv, 253 pp.
131 Janoschka, Anja: Web Advertising. New forms of communication on the Internet. 2004. xiv, 230 pp.
130 Halmari, Helena and Tuija Virtanen (eds.): Persuasion Across Genres. A linguistic approach. 2005.
x, 257 pp.
129 Taboada, María Teresa: Building Coherence and Cohesion. Task-oriented dialogue in English and
Spanish. 2004. xvii, 264 pp.
128 Cordella, Marisa: The Dynamic Consultation. A discourse analytical study of doctor–patient
communication. 2004. xvi, 254 pp.
127 Brisard, Frank, Michael Meeuwis and Bart Vandenabeele (eds.): Seduction, Community,
Speech. A Festschrift for Herman Parret. 2004. vi, 202 pp.
126 Wu, Yi’an: Spatial Demonstratives in English and Chinese. Text and Cognition. 2004. xviii, 236 pp.
125 Lerner, Gene H. (ed.): Conversation Analysis. Studies from the first generation. 2004. x, 302 pp.
124 Vine, Bernadette: Getting Things Done at Work. The discourse of power in workplace interaction. 2004.
x, 278 pp.
123 Márquez Reiter, Rosina and María Elena Placencia (eds.): Current Trends in the Pragmatics of
Spanish. 2004. xvi, 383 pp.
122 González, Montserrat: Pragmatic Markers in Oral Narrative. The case of English and Catalan. 2004.
xvi, 410 pp.
121 Fetzer, Anita: Recontextualizing Context. Grammaticality meets appropriateness. 2004. x, 272 pp.
120 Aijmer, Karin and Anna-Brita Stenström (eds.): Discourse Patterns in Spoken and Written Corpora.
2004. viii, 279 pp.
119 Hiltunen, Risto and Janne Skaffari (eds.): Discourse Perspectives on English. Medieval to modern.
2003. viii, 243 pp.
118 Cheng, Winnie: Intercultural Conversation. 2003. xii, 279 pp.
117 Wu, Ruey-Jiuan Regina: Stance in Talk. A conversation analysis of Mandarin final particles. 2004.
xvi, 260 pp.
116 Grant, Colin B. (ed.): Rethinking Communicative Interaction. New interdisciplinary horizons. 2003.
viii, 330 pp.
115 Kärkkäinen, Elise: Epistemic Stance in English Conversation. A description of its interactional
functions, with a focus on I think. 2003. xii, 213 pp.
114 Kühnlein, Peter, Hannes Rieser and Henk Zeevat (eds.): Perspectives on Dialogue in the New
Millennium. 2003. xii, 400 pp.
113 Panther, Klaus-Uwe and Linda L. Thornburg (eds.): Metonymy and Pragmatic Inferencing. 2003.
xii, 285 pp.
112 Lenz, Friedrich (ed.): Deictic Conceptualisation of Space, Time and Person. 2003. xiv, 279 pp.
111 Ensink, Titus and Christoph Sauer (eds.): Framing and Perspectivising in Discourse. 2003. viii, 227 pp.
110 Androutsopoulos, Jannis K. and Alexandra Georgakopoulou (eds.): Discourse
Constructions of Youth Identities. 2003. viii, 343 pp.
109 Mayes, Patricia: Language, Social Structure, and Culture. A genre analysis of cooking classes in Japan and
America. 2003. xiv, 228 pp.
108 Barron, Anne: Acquisition in Interlanguage Pragmatics. Learning how to do things with words in a study
abroad context. 2003. xviii, 403 pp.
107 Taavitsainen, Irma and Andreas H. Jucker (eds.): Diachronic Perspectives on Address Term
Systems. 2003. viii, 446 pp.
106 Busse, Ulrich: Linguistic Variation in the Shakespeare Corpus. Morpho-syntactic variability of second
person pronouns. 2002. xiv, 344 pp.
105 Blackwell, Sarah E.: Implicatures in Discourse. The case of Spanish NP anaphora. 2003. xvi, 303 pp.
104 Beeching, Kate: Gender, Politeness and Pragmatic Particles in French. 2002. x, 251 pp.
103 Fetzer, Anita and Christiane Meierkord (eds.): Rethinking Sequentiality. Linguistics meets
conversational interaction. 2002. vi, 300 pp.
102 Leafgren, John: Degrees of Explicitness. Information structure and the packaging of Bulgarian subjects
and objects. 2002. xii, 252 pp.
101 Luke, K. K. and Theodossia-Soula Pavlidou (eds.): Telephone Calls. Unity and diversity in
conversational structure across languages and cultures. 2002. x, 295 pp.
100 Jaszczolt, Katarzyna M. and Ken Turner (eds.): Meaning Through Language Contrast. Volume 2. 2003.
viii, 496 pp.
99 Jaszczolt, Katarzyna M. and Ken Turner (eds.): Meaning Through Language Contrast. Volume 1. 2003.
xii, 388 pp.
98 Duszak, Anna (ed.): Us and Others. Social identities across languages, discourses and cultures. 2002.
viii, 522 pp.
97 Maynard, Senko K.: Linguistic Emotivity. Centrality of place, the topic-comment dynamic, and an ideology
of pathos in Japanese discourse. 2002. xiv, 481 pp.
96 Haverkate, Henk: The Syntax, Semantics and Pragmatics of Spanish Mood. 2002. vi, 241 pp.
95 Fitzmaurice, Susan M.: The Familiar Letter in Early Modern English. A pragmatic approach. 2002.
viii, 263 pp.
94 McIlvenny, Paul (ed.): Talking Gender and Sexuality. 2002. x, 332 pp.
93 Baron, Bettina and Helga Kotthoff (eds.): Gender in Interaction. Perspectives on femininity and
masculinity in ethnography and discourse. 2002. xxiv, 357 pp.
92 Gardner, Rod: When Listeners Talk. Response tokens and listener stance. 2001. xxii, 281 pp.
91 Gross, Joan: Speaking in Other Voices. An ethnography of Walloon puppet theaters. 2001. xxviii, 341 pp.
90 Kenesei, István and Robert M. Harnish (eds.): Perspectives on Semantics, Pragmatics, and Discourse. A
Festschrift for Ferenc Kiefer. 2001. xxii, 352 pp.
89 Itakura, Hiroko: Conversational Dominance and Gender. A study of Japanese speakers in first and second
language contexts. 2001. xviii, 231 pp.
88 Bayraktaroğlu, Arın and Maria Sifianou (eds.): Linguistic Politeness Across Boundaries. The case of
Greek and Turkish. 2001. xiv, 439 pp.
87 Mushin, Ilana: Evidentiality and Epistemological Stance. Narrative Retelling. 2001. xviii, 244 pp.
86 Ifantidou, Elly: Evidentials and Relevance. 2001. xii, 225 pp.
85 Collins, Daniel E.: Reanimated Voices. Speech reporting in a historical-pragmatic perspective. 2001.
xx, 384 pp.
84 Andersen, Gisle: Pragmatic Markers and Sociolinguistic Variation. A relevance-theoretic approach to the
language of adolescents. 2001. ix, 352 pp.
83 Márquez Reiter, Rosina: Linguistic Politeness in Britain and Uruguay. A contrastive study of requests and
apologies. 2000. xviii, 225 pp.
82 Khalil, Esam N.: Grounding in English and Arabic News Discourse. 2000. x, 274 pp.
81 Di Luzio, Aldo, Susanne Günthner and Franca Orletti (eds.): Culture in Communication. Analyses of
intercultural situations. 2001. xvi, 341 pp.
80 Ungerer, Friedrich (ed.): English Media Texts – Past and Present. Language and textual structure. 2000.
xiv, 286 pp.
79 Andersen, Gisle and Thorstein Fretheim (eds.): Pragmatic Markers and Propositional Attitude. 2000.
viii, 273 pp.
78 Sell, Roger D.: Literature as Communication. The foundations of mediating criticism. 2000. xiv, 348 pp.
77 Vanderveken, Daniel and Susumu Kubo (eds.): Essays in Speech Act Theory. 2002. vi, 328 pp.
76 Matsui, Tomoko: Bridging and Relevance. 2000. xii, 251 pp.
75 Pilkington, Adrian: Poetic Effects. A relevance theory perspective. 2000. xiv, 214 pp.
74 Trosborg, Anna (ed.): Analysing Professional Genres. 2000. xvi, 256 pp.
73 Hester, Stephen K. and David Francis (eds.): Local Educational Order. Ethnomethodological studies of
knowledge in action. 2000. viii, 326 pp.
72 Marmaridou, Sophia S.A.: Pragmatic Meaning and Cognition. 2000. xii, 322 pp.
71 Gómez González, María de los Ængeles: The Theme–Topic Interface. Evidence from English. 2001.
xxiv, 438 pp.
70 Sorjonen, Marja-Leena: Responding in Conversation. A study of response particles in Finnish. 2001.
x, 330 pp.
69 Noh, Eun-Ju: Metarepresentation. A relevance-theory approach. 2000. xii, 242 pp.
68 Arnovick, Leslie K.: Diachronic Pragmatics. Seven case studies in English illocutionary development. 2000.
xii, 196 pp.
67 Taavitsainen, Irma, Gunnel Melchers and Päivi Pahta (eds.): Writing in Nonstandard English.
2000. viii, 404 pp.
66 Jucker, Andreas H., Gerd Fritz and Franz Lebsanft (eds.): Historical Dialogue Analysis. 1999.
viii, 478 pp.
65 Cooren, François: The Organizing Property of Communication. 2000. xvi, 272 pp.
64 Svennevig, Jan: Getting Acquainted in Conversation. A study of initial interactions. 2000. x, 384 pp.
63 Bublitz, Wolfram, Uta Lenk and Eija Ventola (eds.): Coherence in Spoken and Written Discourse. How
to create it and how to describe it. Selected papers from the International Workshop on Coherence, Augsburg,
24-27 April 1997. 1999. xiv, 300 pp.
62 Tzanne, Angeliki: Talking at Cross-Purposes. The dynamics of miscommunication. 2000. xiv, 263 pp.
61 Mills, Margaret H. (ed.): Slavic Gender Linguistics. 1999. xviii, 251 pp.
60 Jacobs, Geert: Preformulating the News. An analysis of the metapragmatics of press releases. 1999.
xviii, 428 pp.
59 Kamio, Akio and Ken-ichi Takami (eds.): Function and Structure. In honor of Susumu Kuno. 1999.
x, 398 pp.
58 Rouchota, Villy and Andreas H. Jucker (eds.): Current Issues in Relevance Theory. 1998. xii, 368 pp.
57 Jucker, Andreas H. and Yael Ziv (eds.): Discourse Markers. Descriptions and theory. 1998. x, 363 pp.
56 Tanaka, Hiroko: Turn-Taking in Japanese Conversation. A Study in Grammar and Interaction. 2000.
xiv, 242 pp.
55 Allwood, Jens and Peter Gärdenfors (eds.): Cognitive Semantics. Meaning and cognition. 1999.
x, 201 pp.
54 Hyland, Ken: Hedging in Scientific Research Articles. 1998. x, 308 pp.
53 Mosegaard Hansen, Maj-Britt: The Function of Discourse Particles. A study with special reference to
spoken standard French. 1998. xii, 418 pp.
52 Gillis, Steven and Annick De Houwer (eds.): The Acquisition of Dutch. With a Preface by Catherine E.
Snow. 1998. xvi, 444 pp.
51 Boulima, Jamila: Negotiated Interaction in Target Language Classroom Discourse. 1999. xiv, 338 pp.
50 Grenoble, Lenore A.: Deixis and Information Packaging in Russian Discourse. 1998. xviii, 338 pp.
49 Kurzon, Dennis: Discourse of Silence. 1998. vi, 162 pp.
48 Kamio, Akio: Territory of Information. 1997. xiv, 227 pp.
47 Chesterman, Andrew: Contrastive Functional Analysis. 1998. viii, 230 pp.
46 Georgakopoulou, Alexandra: Narrative Performances. A study of Modern Greek storytelling. 1997.
xvii, 282 pp.
45 Paltridge, Brian: Genre, Frames and Writing in Research Settings. 1997. x, 192 pp.
44 Bargiela-Chiappini, Francesca and Sandra J. Harris: Managing Language. The discourse of corporate
meetings. 1997. ix, 295 pp.
43 Janssen, Theo and Wim van der Wurff (eds.): Reported Speech. Forms and functions of the verb. 1996.
x, 312 pp.
42 Kotthoff, Helga and Ruth Wodak (eds.): Communicating Gender in Context. 1997. xxvi, 424 pp.
41 Ventola, Eija and Anna Mauranen (eds.): Academic Writing. Intercultural and textual issues. 1996.
xiv, 258 pp.
40 Diamond, Julie: Status and Power in Verbal Interaction. A study of discourse in a close-knit social network.
1996. viii, 184 pp.
39 Herring, Susan C. (ed.): Computer-Mediated Communication. Linguistic, social, and cross-cultural
perspectives. 1996. viii, 326 pp.
38 Fretheim, Thorstein and Jeanette K. Gundel (eds.): Reference and Referent Accessibility. 1996. xii, 312 pp.
37 Carston, Robyn and Seiji Uchida (eds.): Relevance Theory. Applications and implications. 1998. x, 300 pp.
36 Chilton, Paul, Mikhail V. Ilyin and Jacob L. Mey (eds.): Political Discourse in Transition in Europe
1989–1991. 1998. xi, 272 pp.
35 Jucker, Andreas H. (ed.): Historical Pragmatics. Pragmatic developments in the history of English. 1995.
xvi, 624 pp.
34 Barbe, Katharina: Irony in Context. 1995. x, 208 pp.
33 Goossens, Louis, Paul Pauwels, Brygida Rudzka-Ostyn, Anne-Marie Simon-Vandenbergen
and Johan Vanparys: By Word of Mouth. Metaphor, metonymy and linguistic action in a cognitive
perspective. 1995. xii, 254 pp.
32 Shibatani, Masayoshi and Sandra A. Thompson (eds.): Essays in Semantics and Pragmatics. In honor of
Charles J. Fillmore. 1996. x, 322 pp.
31 Wildgen, Wolfgang: Process, Image, and Meaning. A realistic model of the meaning of sentences and
narrative texts. 1994. xii, 281 pp.
30 Wortham, Stanton E.F.: Acting Out Participant Examples in the Classroom. 1994. xiv, 178 pp.
29 Barsky, Robert F.: Constructing a Productive Other. Discourse theory and the Convention refugee hearing.
1994. x, 272 pp.
28 Van de Walle, Lieve: Pragmatics and Classical Sanskrit. A pilot study in linguistic politeness. 1993.
xii, 454 pp.
27 Suter, Hans-Jürg: The Wedding Report. A prototypical approach to the study of traditional text types. 1993.
xii, 314 pp.
26 Stygall, Gail: Trial Language. Differential discourse processing and discursive formation. 1994. xii, 226 pp.
25 Couper-Kuhlen, Elizabeth: English Speech Rhythm. Form and function in everyday verbal interaction.
1993. x, 346 pp.
24 Maynard, Senko K.: Discourse Modality. Subjectivity, Emotion and Voice in the Japanese Language. 1993.
x, 315 pp.
23 Fortescue, Michael, Peter Harder and Lars Kristoffersen (eds.): Layered Structure and Reference
in a Functional Perspective. Papers from the Functional Grammar Conference, Copenhagen, 1990. 1992.
xiii, 444 pp.
22 Auer, Peter and Aldo Di Luzio (eds.): The Contextualization of Language. 1992. xvi, 402 pp.
21 Searle, John R., Herman Parret and Jef Verschueren: (On) Searle on Conversation. Compiled and
introduced by Herman Parret and Jef Verschueren. 1992. vi, 154 pp.
20 Nuyts, Jan: Aspects of a Cognitive-Pragmatic Theory of Language. On cognition, functionalism, and grammar.
1991. xii, 399 pp.
19 Baker, Carolyn and Allan Luke (eds.): Towards a Critical Sociology of Reading Pedagogy. Papers of the XII
World Congress on Reading. 1991. xxi, 287 pp.
18 Johnstone, Barbara: Repetition in Arabic Discourse. Paradigms, syntagms and the ecology of language. 1991.
viii, 130 pp.
17 Piéraut-Le Bonniec, Gilberte and Marlene Dolitsky (eds.): Language Bases ... Discourse Bases. Some
aspects of contemporary French-language psycholinguistics research. 1991. vi, 342 pp.
16 Mann, William C. and Sandra A. Thompson (eds.): Discourse Description. Diverse linguistic analyses of a
fund-raising text. 1992. xiii, 409 pp.
15 Komter, Martha L.: Conflict and Cooperation in Job Interviews. A study of talks, tasks and ideas. 1991.
viii, 252 pp.
14 Schwartz, Ursula V.: Young Children's Dyadic Pretend Play. A communication analysis of plot structure and
plot generative strategies. 1991. vi, 151 pp.
13 Nuyts, Jan, A. Machtelt Bolkestein and Co Vet (eds.): Layers and Levels of Representation in Language
Theory. A functional view. 1990. xii, 348 pp.
12 Abraham, Werner (ed.): Discourse Particles. Descriptive and theoretical investigations on the logical,
syntactic and pragmatic properties of discourse particles in German. 1991. viii, 338 pp.
11 Luong, Hy V.: Discursive Practices and Linguistic Meanings. The Vietnamese system of person reference. 1990.
x, 213 pp.
10 Murray, Denise E.: Conversation for Action. The computer terminal as medium of communication. 1991.
xii, 176 pp.
9 Luke, K. K.: Utterance Particles in Cantonese Conversation. 1990. xvi, 329 pp.
8 Young, Lynne: Language as Behaviour, Language as Code. A study of academic English. 1991. ix, 304 pp.
7 Lindenfeld, Jacqueline: Speech and Sociability at French Urban Marketplaces. 1990. viii, 173 pp.
6:3 Blommaert, Jan and Jef Verschueren (eds.): The Pragmatics of International and Intercultural
Communication. Selected papers from the International Pragmatics Conference, Antwerp, August 1987. Volume
3: The Pragmatics of International and Intercultural Communication. 1991. viii, 249 pp.
6:2 Verschueren, Jef (ed.): Levels of Linguistic Adaptation. Selected papers from the International Pragmatics
Conference, Antwerp, August 1987. Volume 2: Levels of Linguistic Adaptation. 1991. viii, 339 pp.
6:1 Verschueren, Jef (ed.): Pragmatics at Issue. Selected papers of the International Pragmatics Conference,
Antwerp, August 17–22, 1987. Volume 1: Pragmatics at Issue. 1991. viii, 314 pp.
5 Thelin, Nils B. (ed.): Verbal Aspect in Discourse. 1990. xvi, 490 pp.
4 Raffler-Engel, Walburga von (ed.): Doctor–Patient Interaction. 1989. xxxviii, 294 pp.
3 Oleksy, Wieslaw (ed.): Contrastive Pragmatics. 1988. xiv, 282 pp.
2 Barton, Ellen: Nonsentential Constituents. A theory of grammatical structure and pragmatic interpretation.
1990. xviii, 247 pp.
1 Walter, Bettyruth: The Jury Summation as Speech Genre. An ethnographic study of what it means to those
who use it. 1988. xvii, 264 pp.