LANGUAGE

APPLIED LINGUISTICS
MICHAEL McCARTHY

CAMBRIDGE {33° ee
UNIV ERSiEY. hea,
Spoken language and
applied linguistics

Michael McCarthy

AMBRIDGE
6) UNIVERSITY PRESS
PUBLISHED BY THE PRESS SYNDICATE OF THE UNIVERSITY OF CAMBRIDGE
The Pitt Building, Trumpington Street, Cambridge CB2 1RP, United Kingdom

CAMBRIDGE UNIVERSITY PRESS

The Edinburgh Building, Cambridge CB2 2RU, United Kingdom
40 West 20th Street, New York, NY 10011-4211, USA
10 Stamford Road, Oakleigh, Melbourne 3166, Australia

© Cambridge University Press, 1998

This book is in copyright. Subject to statutory exception

and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without
the written permission of Cambridge University Press.

First published 1998

Typeset in 9.5/13.5pt Swift Regular [cz]

A catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication Data

McCarthy, Michael, 1947-
Spoken language and applied linguistics /Michael McCarthy.
Dec:
Includes bibliographical references.
ISBN 0-521-59769-2 (pbk.: alk. paper). - ISBN 0-521-59213-5 (hardback: alk. paper}
1. Colloquial language. 2. Language and languages - Study and teaching.
I. Title.
P 408.M38 1998 98-39755 CIP
401’.41-dc21

Printed in the United Kingdom at the University Press, Cambridge

ISBN 0 521 59213 5 Hardback

ISBN 0 521 59769 2 Paperback

%
be| Ghee sd
Miruire Gass Soe
ba Leuae ee
4
}
To the dead and living of old Splott, Cardiff
 Contents

Acknowledgements

The author

Introduction
+

Spoken language and the notion of genre

What should we teach about the spoken language?

When does sentence grammar become discourse grammar?

Some patterns of co-occurrence of verb-forms in spoken

and written English

Vocabulary and the spoken language 108

Idioms in use: a discourse-based re-examination of a

traditional area of language teaching 129

‘So Mary was saying’: speech reporting in everyday

conversation 150

Glossary 176

References 182

Index 203
Acknowledgements

This book owes thanks to many colleagues and friends who have given
intellectual and personal support during its production. Firstly, as
always, the book would not be what it is without the productive partner-
ship I have enjoyed over the last decade with Ronald Carter, colleague
and friend. Ron has read the whole book, and offered comments and
insights both formally and informally, for which I am extremely grateful.
Four other colleagues in the Department of English Studies at the
University of Nottingham who have been instrumental in my thinking
also deserve special thanks: Rebecca Hughes, Roger Smith, Martha Jones
and Julia Harrison. At Cambridge University Press, Jean Hudson, who
provided the initial design for the CANCODE corpus, which this book is
based on, has provided me with many insights on the data, as well as the
flow of data itself. The one million words of the first phase of the
CANCODE corpus would not have existed without the contributions
made to it by fellow-researchers and students at Nottingham: Bethan
Benwell (now at the University of Stirling), who recorded hours of small-
group university tutorials and who understands spoken genres better
than most, deserves especial thanks. Other colleagues and friends around
the world whose comments (in formal contexts such as conferences as
well as informal ones) have inspired many of the statements in this book
include the late David Brazil, David Nunan, Nik Coupland, Justine Coup-
land, Guy Cook, Robert Cockcroft, Paul Drew, Doug Biber, Sue Conrad,
Michael Lewis, Mike Baynham, Joan Cutting, Nkonko Kamwangamalu,
Joanna Channell, John Sinclair, Norbert Schmitt, Mike Hoey, Merrill
Swain, Mike Makosch, Jim Lantolf, Tony Fitzpatrick, Aria Merkestein,
Geoff Tranter, Almut Koester, Bruce Pye, Matilde Grunhage-Monetti,
Barry O'Sullivan, Jeff Stranks and Nurdan Ozbek. Without the encourage-
ment of Alison Sharpe, Commissioning Editor at CUP, and the continued,
generous support of CUP, and Colin Hayes, ELT Group Director, in
particular, for the CANCODE project, I would not now be writing these
words. I put the finishing touches to the manuscript while enjoying a

vi
 Acknowledgements - vii

semester as Visiting Professor in Linguistics at Cornell University, USA,

where colleagues in the Department of Modern Languages (especially
Hongyin Tao and Linda Waugh) have offered invaluable intellectual
support in the book’s final stages of revision. As always, the book would
probably never have happened without the personal and domestic
support of my partner Jeanne McCarten. All of these people deserve a big
thank you. An anonymous reviewer for CUP was very patient with earlier
versions of the manuscript and offered detailed suggestions for improve-
ment. In the final editing and preparation for publication, under the
guidance of Mickey Bonin of CUP, comments and suggestions for final
revisions from him, Geraldine Mark and Jane Clifford have proved
invaluable. Whatever faults remain in the book must be laid entirely at
my door.
The author

Michael McCarthy is Professor of Applied Linguistics in the Department

of English Studies at the University of Nottingham, UK. He entered the
English Language Teaching profession in 1966 and has since then taught
English in Spain, The Netherlands, Great Britain, Sweden and Malaysia.
He has taught at the Universities of Cambridge, Wales, Birmingham and
Nottingham. He has published several books, including Vocabulary and
Language Teaching (with Ron Carter, Longman 1988), Vocabulary (Oxford
University Press 1990), Discourse Analysis for Language Teachers (Cambridge
University Press 1991), Language as Discourse (with Ron Carter, Longman
1994), English Vocabulary in Use (with Felicity O’Dell, Cambridge University
Press 1994), Exploring Spoken English (with Ron Carter, Cambridge Uni-
versity Press 1997), Second Language Vocabulary (with Norbert Schmitt,
Cambridge University Press 1997), as well as many academic articles on
vocabulary and on spoken discourse. He is also Series Editor of the
Cambridge University Press Word Routes/Word Selector series of bilingual
thesauruses. |

Vili
1
Introduction

1.1 Genesis of this book

This book is the result of ten years of study of the spoken language and
its importance to language teaching. Initially this involved transcribing
and analysing brief conversational extracts, and latterly (but with the
same aim always in mind) examining large numbers of conversational
extracts brought together in the CANCODE (Cambridge and Nottingham
Corpus of Discourse in English) corpus in the Department of English
Studies at the University of Nottingham (see 1.2 below). Over those years,
I have presented and published papers and written books, sometimes of
my sole authorship, often with my close colleague and co-researcher
Ronald Carter (and recently also with my colleague Rebecca Hughes).
Those papers and books have led me more and more into questions
concerning everyday spoken language as a model for language teaching,
how different types of spoken language can be classified, and what status
the spoken language should have as an object of study within applied
linguistics in general. That is essentially what this book is about, and the
CANCODE corpus has been an invaluable tool in getting answers to (some
of) these questions.

1.2 Overview

The book brings together revised versions of papers published over the
period 1988-1996 and some new, previously unpublished chapters, all
drawing on corpus data, occasionally quantitatively, but mostly qualita-
tively, for it is in the latter that I see the greatest potential for gathering
useful pedagogical insights from close observation of how people ‘do’
everyday talk. In this first chapter I outline the CANCODE corpus project,
upon which most of this book is based. I have tried to contextualise it with
reference to other corpus projects. The chapter also takes a historical
glance at the status of spoken language in language study and the

1
2 + 1 Introduction

teaching of languages over the centuries. There is a tendency in applied

linguistics nowadays to consider anything published more than five years
ago as ancient, and a historical perspective is often absent from MA course
syllabuses and professional textbooks. I firmly believe that we have much
to learn from our scholastic predecessors, not least in that their achieve-
ments may urge humility upon us when we are tempted to think that our
‘science’ has made massive steps forward. Chapter 1 then, concludes with
a note on the contributions made by speech-act theory and discourse- and
conversation-analysis to our improved understanding of spoken language,
though with notes of caution expressed along the way.
Chapter 2 attempts to construct the outlines of a theory of spoken
genre. In particular it focuses on the variation present in speech events
which, nonetheless, still share a lot of common features. Chapter 2 also
stresses the importance of looking for evidence that participants
themselves are aware that they are engaging in the creation of genres,
and that we are not just indulging in analytical fancies. We find such
evidence partly in terms of the participants’ orientation towards past
events and upcoming ones, the need for agreement among participants as
to ‘where they are’ in the talk and the need to bring into effect procedures
whereby events assume their final shape that we as analysts can recognise.
Chapter 2 uses the CANCODE corpus matrix outlined in Chapter 1 to
show how extracts display similarities at the lexico-grammatical level
which correspond to higher-order features of generically-oriented activity.
The chapter concludes that genre will always remain a difficult notion to
pin down because social activity is prone to so much variation. What is
apparent is that seen as a whole, behaviour is integrated: the transac-
tional, the interactional, the goal-orientation, the relationships among
participants, and the local lexico-grammatical details all complement
each other. Chapters 1 and 2 together are an attempt to sketch out a
theory of spoken language that will have pedagogical relevance.
Chapter 3 raises the question of just what should and can be taught
about the spoken language. I take a number of concepts from discourse
analysis and conversation analysis and examine existing research to try
to get answers to questions such as ‘Do these features matter?’, ‘Are they
universal?, ‘Are they likely to be transferred, and if so, under what sorts
of learning conditions?’, ‘How can syllabuses and methodologies take
such features into account?’. The features examined include exchanges
and adjacency pairs, discourse markers, ellipsis, openings and closings,
and in each case I look for the lexical and grammatical realisations of the
 1.2 Overview - 3

features and their cultural import, using CANCODE data examples to

illustrate the way speakers encode the features. The chapter concludes
that there is a good deal to be incorporated into teaching from the
research available in the areas looked at, but that traditional, presenta-
tion-based methodologies are inadequate to the task. I propose an
alternative methodology outlined by Carter and McCarthy (1995b) based
on what we call ‘the three I’s’, in contrast to the traditional ‘three Ps’ of
presentation, practice and production.
Chapters 4 and 5 are both devoted to issues of grammar. Chapter 4
argues for a discourse-based view of grammatical choice, and uses spoken
data to substantiate the claim that certain aspects of grammar are best
understood when examined in context. I try to demonstrate that some-
times we need to re-think how we group items in grammatical para-
digms, taking the words it, this and that as examples. I then move the
discussion on to the treatment of the past perfect, and attempt to show
how choice of that tense is discoursally motivated, rather than by some
deterministic rules about time. The chapter continues in this vein,
examining features that are differently distributed in spoken and
written data, and a typically spoken feature that has long caused
problems for grammarians, the English get-passive. Central to the thread
of discussion in the chapter is the advocacy of the usefulness in language
teaching of probabilistic statements (statements about the most likely
conditions of occurrence of a form), rather than only viewing grammar
as deterministic rules. I conclude that a discourse-based view of
grammar, backed up by spoken data, is extremely illuminating and has
direct consequences for what and how we teach in grammar classes.
Chapter 5 pre-empts two possible dangers: (a) that we may rush off and
assume that everything is different in spoken grammar and that nothing
we Say about written language has any validity for the description and the
teaching of spoken language, or (b), equally dangerously, that we should
assume that descriptions of the written grammar can simply be imported
wholesale into spoken grammars. Taking a discoursal framework built on
the work of the late Eugene Winter, the chapter looks at how sequences of
verb tense choices can ‘frame’ whole written texts and spoken episodes,
and how these framing functions are found in both spoken and written.
The data shows that the realisations of these functions are in some cases
the same in spoken and written, but also that their realisations are
sometimes crucially different. This evidence is used to support the
arguments that contextual grammars as advocated by Winter are indeed
4 + 1 Introduction

very powerful descriptive instruments, and that spoken grammatical

features cannot always be assumed to follow the same realisations as their
written functional counterparts. It is also an exercise in showing the
usefulness of direct comparisons between written and spoken texts.
Chapters 6 and 7 shift the focus to vocabulary, but still within the
perspective of a spoken discourse-based approach. Chapter 6 begins by
laying out some of the constraints inherent in multi-party interaction that
are likely to influence vocabulary selection, and then examines various
lexical features in spoken language extracts. The question of lexical
density is discussed, and especially the sorts of vocabulary found in low-
density, language-in-action episodes. These types of vocabulary present the
teacher with challenges that traditional, content-based notions of vocabu-
lary do not need to concern themselves with. From there, the chapter
shifts to looking at how speakers take up one another’s vocabulary
choices, either to progress or to hold up the smooth development of the
talk, and how the role of listener is just as important as that of main
speaker in building the vocabulary of an interaction, especially in oral
narratives. Wider issues of the differences between spoken and written
vocabulary lists generated from corpora are then examined, and the sorts
of problems frequency-counts raise for teaching are touched upon. The
chapter concludes that training learners’ observational powers might be
just as useful as pumping them with large numbers of new words.
Chapter 7 is about idioms, a subject most language teachers can
identify with as part of the stock-in-trade of vocabulary teaching, espe-
cially at higher levels of proficiency, and yet one that has rarely been
dealt with from a discourse perspective, and one where available infor-
mation on usage in everyday spoken language is scant. After discussing
problems of definition, I offer data from spoken narratives to support the
claim that idiom selection is not random, but plays a key evaluative role
in storytelling. The chapter then casts its net wider, with examples from
discourses where speakers are commenting on aspects of their world in
general; once again, idioms have an important evaluative role to play.
The data shows that speakers often ‘unpack’ the literal meanings buried
in idioms and create extended puns and images that thread through the
text. Idioms also have a role in reinforcing cultural membership. The
chapter ends with a discussion of the problems of teaching idioms, and
the kinds of exercises and contexts that can be exploited while remaining
faithful to a discourse-oriented approach.
The last chapter of the book returns to a grammatical area that usually
 1.3 The CANCODE corpus and its context + 5

has its place in any language-learning syllabus: reported speech. We

begin by comparing two extremes in speech-reporting: classic literature
and everyday conversational storytelling. The conversational reporting
looks at first glance to be rather impoverished when compared with
great literature, but I attempt to show its richness and flexibility. I then
look at other spoken reporting strategies, such as past continuous
reporting verbs and the narrative historical present. With the quantified
evidence from the CANCODE corpus, the chapter then compares what is
frequent in the corpus with what is usually found in literary reporting,
and shows that the corpus examples are very varied, serving a variety of
strategic functions. The chapter ends with a general set of recommenda-
tions for teaching speech reporting, as well as a general conclusion on
the book as a whole. (There is a glossary of terms on pages 176-81.)

1.3 The CANCODE corpus and its context N

The CANCODE (Cambridge and Nottingham Corpus of Discourse in

English) project was established in the Department of English Studies at
the University of Nottingham, UK, with the help of a generous research
grant from Cambridge University Press, the publisher of the present
work and (to date) of two other works (Carter and McCarthy 1997a, and
McCarthy and Carter 1997a) which draw on CANCODE material. A
complete list of publications based on CANCODE is given at the end of
this book. Cambridge University Press’s support enabled Ronald Carter,
me and our research assistant Jean Hudson to assemble an initial corpus
of 1,061,274 running words of informal spoken English, completed in
1996. Further support from Cambridge University Press is, at the time of
writing, enabling us (with the help of our new additional assistant, Julia
Harrison) to expand the corpus to five million words.
It must seem, to a profession dazzled by ever larger language corpora, a
rather puny enterprise to be working with only a million (or even five
million) words. Corpora now regularly consist of hundreds of millions of
words and the race to be first to hit a billion is undoubtedly one that will
run in tandem with the tick of the clock towards the new millennium.
Technology seems boundless and gigantic corpora can now be marshalled
with an ease that would have been considered pure science fiction when I
entered the ELT profession in 1966. CANCODE, therefore, is numerically
small by today’s standards. It is also relatively modest even by the
standards established some time ago by pioneers in corpus linguistics.
6 - 1 Introduction

It would be unfair, however, to judge CANCODE’s size only in relation

to huge written or mixed written/spoken corpora. More important is to
consider it in relation to the development of spoken corpora in general,
and to the problems that make spoken corpora much more difficult to
assemble than written ones. Spoken corpora are not new: anthropologists
and dialectologists have long used tape-recorded data as a major source
of evidence (Biber 1990), and such work continues (for example, the
Northern Ireland spoken corpus described by Kirk 1992). Some of the
earliest spoken investigations were carried out within the study of child
language acquisition (an example of this is the child-language word-
frequency analyses described in Beier, Starkweather and Miller 1967).' A
notable early spoken corpus project of the kind that has since become
quite common was the Oral Vocabulary of the Australian Worker (OVAW)
(Schonell et al. 1956 gives a full account of the data and its collection).
The OVAW corpus consisted of some 500,000 words of spoken language
and is still very useful for anyone interested in idiomatic words and
phrases in speech (e.g. Schonell et al. 1956: p. 67), a subject I pursue
further in Chapter 7 of this book using CANCODE data. The OVAW also
records the ubiquitous discourse markers found in everyday spoken
language. A decade after OVAW, the Davis-Howes Count of Spoken
English (Howes 1966) in the USA brought together half a million words of
interviews with university students and hospital patients, and produced
some interesting statistics for spoken usage.
Also in the 1960s a spoken-word count for Russian was published
(Vakar 1966), which was small (based only on 10,000 words taken from
drama texts), but which offered useful statistics about text coverage of
high-frequency words. The University of Leuven Drama Corpus (approxi-
mately one million words from contemporary plays; see Engels 1988 for
details) continued this tradition of using drama texts as a model for the
spoken language (see also the reference to Vanrespaille 1991 in Chapter
4). Further evidence that literary (or at least written-fiction-based)
corpora can also be useful for the comparative study of spoken language
may be seen in the work of researchers investigating the 1.5 million-word
TOSCA corpus at the University of Nijmegen (e.g. see Oostdijk’s 1990
study of fictional dialogue, based on TOSCA).
Meanwhile, in Great Britain, the half-million-word spoken segment of
the London-Lund corpus (itself a half of the one-million-word spoken/
written Survey of English Usage (SEU) at University College London; see
Svartvik 1990), and the conversational transcripts available in Svartvik
 1.3 The CANCODE corpus and its context + 7

and Quirk (1980) have been instrumental in some very important

investigations of the vocabulary of everyday spoken English (see
McCarthy and Carter 1997a for further examples). Nowadays, as noted
above, gigantic corpus projects such as the COBUILD Bank of English (see
Moon 1997 for a recent description of the project) and the British
National Corpus (see Crowdy 1993 and Rundell 1995a and 1995b for
details of design and content) contain considerable amounts of spoken
English data, including broadcast speech as well as everyday unrehearsed
conversation. Broadcast data of many different types form the basis of
the British English Lancaster/IBM spoken corpus (see Knowles 1990). In
the USA, work by Chafe and his colleagues, initially based on the British
London-Lund spoken corpus design (Chafe, Du Bois and Thompson 1991),
has developed into larger enterprises such as the five-million word
Longman Spoken American Corpus (see Stern 1997), which Biber and
others are investigating to great effect, and which is planned to feed
directly into language-learning resources.
Australian English has been subjected to corpus analysis in the
Macquarie University corpus project (see Collins and Peters 1988), and by
Eggins and Slade (1997), who look at everyday conversational activities
such as gossiping. Also important is the ICE (International Corpus of
English) project, which plans to bring together parallel corpora of one
million words from 18 different countries where English is either the
main language or an official language. The samples in the ICE corpus
include 300 spoken texts, although these include many scripted samples,
and broadcast interviews and discussions, with only 90 samples being
face-to-face informal conversations (see Nelson 1996; also Fang 1995).
Spoken corpora have, therefore, come of age, and many lessons have
been learnt along the way. What all the projects mentioned so far have
contributed to our understanding of spoken language and of corpora in
general is massive. Amongst other things, we have gained a better
understanding of the types of talk people engage in, instead of simply
assuming that the text typologies of written language applied the same
to spoken. We have gained a great deal of experience about transcription
of speech (see 1.4 and 1.5 below). We have also learnt the advantages (e.g.
time/cost) and disadvantages (e.g. lack of naturalness) of spoken corpora
taken from broadcast or drama-text sources. Technology has enabled
huge improvements in sound quality and unobtrusiveness of equipment.
Last but not least, being aware of previous corpus projects helps to
prevent the continual re-invention of the wheel, and builds up a body of
8 + 1 Introduction

evidence against which new corpora can be evaluated. Things have come
a long way since the early recordings of dialectologists and anthropol-
ogists, but the design of spoken corpora is still often opportunistic (‘get
whatever data you can’). The CANCODE corpus has tried to avoid
opportunism, and to follow design principles that will make its material
maximally useful to teachers, pedagogically-oriented researchers and
materials writers (see 1.4 below).

1.4 The CANCODE corpus and generic features

In setting up CANCODE, the research team decided to attempt to cover as

many useful speech-types (useful in terms of language learners’ perceived
needs) as was feasible. However, the immediate problem was that no
satisfactory classification of ‘text types’ for spoken language was avail-
able to parallel existing text typologies for written language. Chapter 2
discusses this problem within the framework of genre theory and
demonstrates the theoretical stance the CANCODE team finally adopted.
Here I shall outline the more practical steps that were necessary for the
operationalisation of the notion of speech-types in the corpus design.
In the gathering of spoken data for corpora two approaches seem
prevalent.” The first of these may be termed the ‘demographic’ approach,
where a population of speakers is targeted and where that population
records its spoken output over a given period of time. Biber (1993)
stresses that the targeting of the population is much more important
than sampling size. A well-chosen population sample would certainly
seem, intuitively, to generate a qualitatively better corpus than mere
opportunistic sampling or dumping of huge amounts of undifferentiated
text simply to compete in the corpus ‘numbers game’ that is gripping the
profession at the time of writing. Crowdy (1993) also argues for a
demographic approach, which was used by the British National Corpus
researchers. The other approach may be termed the ‘genre’ approach, in
that it attempts to target not only a population of speakers but particular
environments and contexts in which spoken language is produced. This
approach does not simply rely on a pre-ordained notion of what a speech
‘text’ is. As Atkins et al (1992) have argued, there are various ways of
defining textual boundaries in speech (e.g. the moment interlocutors
come together or part, or the marking of opening and closing features
linguistically). The genre approach tries to seek a balance between
speaker, environment, context and recurrent features. It has the advan-
 1.4 The CANCODE corpus and generic features + 9

tage that the corpus can be analysed from different perspectives (e.g.
types of speakers, emergent text types, situation types, etc.). It has the
disadvantage that genre is an ill-defined notion in the study of spoken
language in general (see Chapter 2). In the genre approach, decisions
have to be made about situational/contextual types as well as population
types, and these decisions are by no means straightforward. The
CANCODE project, which is based on a genre approach, confronted the
problem of generic coverage by attempting to control contextual vari-
ables of different kinds in the collection of data. The data once collected
could then be examined for ‘episodes’ (or linguistically marked speech
events) that displayed similar linguistic patterning at both the global
levels and the local levels. This will be exemplified below. The resultant
model provides a proposed classification that brings out the commonality
of spoken episodes; it does so in a way that offers the possibility of
linking their contextual and social features directly with the lexico-
grammatical ‘nuts-and-bolts’ of their step-by-step creation. The model
eschews categories such as ‘written-to-be-spoken’ or ‘rehearsed spoken’,
which have traditionally informed the study of variation in speech and
writing (see Crystal 1995 for a good recent discussion). This is because (a)
it is very difficult to know whether something is one thing or another
(e.g. a radio interview, or a university tutorial) and (b) the CANCODE team
decided to focus on, wherever possible, unrehearsed, non-formal talk.?
Five broad contexts for data collection based on the type of relation-
ship among participants were identified (principally by the team’s corpus
manager, Jean Hudson):

Transactional
Professional i
Pedagogical
Socialising
Intimate

Transactional relationships were defined as those where speakers display

needs or imperatives and move towards satisfying those needs in a goal-
oriented fashion outside of the contexts of professional, socialising or
intimate relationships. A clear example would be day-to-day service
encounters in shops, restaurants, etc., between servers and customers,
transacting goods, information or services. Professional relationships are
displayed in talk between professional colleagues in professional situa-
tions (e.g. informal company meetings, staff meetings, desk-to-desk talk).
10 - 1 Introduction

Pedagogical relations are those between teachers and their students and
student-student (e.g. informal tutorial conversations, pair- and group-
work). Socialising* relations accord with social or cultural activities
entered upon by participants but not in professional or intimate settings
(e.g. a group of friends preparing a party, talking with a stranger on a
train). Socialising is thus one of the most common categories, covering
much of our day-to-day activity. Intimate relations pertain between family
members or close friends in private, non-professional settings.
For each of these categories, three typical goal-types were posited:

Provision of information
Collaborative tasks
» Collaborative ideas

Provision of information is predominantly uni-directional, with one party

imparting information to others. The role of information-giver may, of
course, rotate among participants, but the dominant motivation for the
talk is information-giving (e.g. an enquiry at a tourist information office).
Collaborative tasks show speakers interacting with their physical environ-
ment while talking (e.g. two people packing a car prior to a journey).
Collaborative ideas are concerned with the interactive sharing of thoughts,
judgements, opinions and attitudes. Just as the context-types are broad
and refer to predominant rather than exclusive traits, so too are the goal-
types.
Some examples of how the categories were operationalised are given
in figure 1:

Context-type Goal-type Example

Transactional Information provision Tourist information office requests

for information
Professional Information provision Company sales conference, informal
informational talks
Pedagogical Collaborative ideas University small-group tutorial
Socialising Collaborative task Relatives and friends preparing food
for a party
Intimate Collaborative ideas Mother and daughter discuss family
matters
88NSeeoeoaoasewc““~“@@oTYoOownmwooma~Ow—S—oODODODo
oo

Figure 1: Examples of operationalised categories

 1.4 The CANCODE corpus and generic features - 11

The five broad context-types and the three goal-types for each one yielded
a matrix of 15 cells, each of which were targeted to be filled with data
samples complying with the context- and goal-types. Although the initial
target was to gather approximately 65,000 running words per cell, it did
not prove feasible to fill each cell with the same amount of data. All
kinds of data can be very sensitive and participants reluctant to release it
(e.g. intimate conversation, business plans). However, wherever possible,
a balance has been sought to cover the broad types within the corpus,
and adjustments were, and are being, constantly made to the make-up of
individual cells. The progress from one million to five million includes
the target of filling some of the ‘less full’ cells of the matrix. If this proves
difficult or impossible, it will be a useful evaluation of the corpus design,
and this, in CANCODE’s terms, is important. In the past, corpora have
tended to become fossilised, either because the initial design is rigidly
and uncompromisingly held to, or because a particular numerical target
has been achieved. The corpus thus becomes a ‘finished object’. The
CANCODE project has in-built, ongoing evaluation of its structure and of
its size, the former relating to the viability of the cellular model as it
stands, the latter in relation to the recoverability of information with
sufficient power of generality. This second problem includes the fact
that, for many high-frequency grammatical structures, even one million
words yields too much information, and sub-samples have to be ex-
tracted, while other, low-frequency words and structures (e.g. -ing
clauses) suggest that larger amounts of data might be necessary. This
continuous evaluation process is central to CANCODE’s development,
and the corpus becomes more of what Sinclair (1995) calls a ‘flow of
data’.®
The data samples assembled as the corpus builds provide possibilities
for comparisons over and above the particular settings in which they
were recorded. For instance, what does a ‘socialising’ (more public)
collaborative task have in common with one conducted in an intimate
setting? How does decision-making in the home (a goal-sub-type) reflect
the same generic activity as decision-making in the workplace? If only
partial answers can be found to such questions, then the possibility of
useful genre-oriented classification translatable into typological frame-
works for spoken language pedagogy would be within grasp. All along,
though, it must be remembered that the products of the classifications
exemplified in figure 1 are samples of spoken text; they are not themselves
the speech genres. We are moving towards building a bridge between
12 + 1 Introduction

genre and text-typology rather than ‘capturing’ generic activity when we

pluck out recurrent features that have left their traces in the spoken
transcripts. It is such textual samples that are the raw data for the
subsequent chapters of this book. Their recurrent features reflect, in
different ways, the generic activity their participants are engaged in;
Chapter 2 goes into greater detail as to how such activity is manifested in
lexico-grammatical and discoursal features of the extracts that comprise
the cellular structure of the corpus. As the corpus builds, there may be
adjustments to make to that structure,® but its initial output has been
promising in terms of pedagogical relevance, as the rest of this book
attempts to show.

1.5 Continuing problems with spoken corpora

Despite the many successful projects undertaken to date, no-one could

deny the continuing difficulties of assembling natural spoken data. Even
the best recording equipment is apt to produce tapes with frustrating
amounts of distracting noise (miraculously filtered out in real life by the
human ear) when used anywhere outside of a soundproof recording
studio. The catch is that good technical quality of sound recordings
almost certainly means the recordings were made with a microphone
obtrusively near the speakers, creating an unwanted degree of artifici-
ality. The most natural, uninhibited data inevitably comes from record-
ings where technical quality has had to be sacrificed to unobtrusiveness
of the equipment and its human operator. These problems have led some
spoken language projects to opt for broadcast data, which is compara-
tively easy to collect. The CANCODE corpus has eschewed this temptation
(though Carter and McCarthy 1997a include a local radio phone-in
transcript to achieve maximum variety in their selection of spoken texts).
We believe that a corpus biased towards large amounts of broadcast data
would, in the main, not be the best model for the spoken language in
relation to our overriding purpose. Our aim is the development of a
description of spoken English that is relevant and useful to language
teachers, especially in connection with the teaching of everyday listening
and speaking skills in informal situations.
Transcription is also a (wickedly expensive) problem in terms of time
and cost, as well as in terms of the system adopted. On average, an hour
of recorded spoken data can take 20 hours to transcribe to a satisfactory
degree of detail and accuracy, and, even then, there will inevitably be
 1.6 Transcribing: the black hole on the trail to infinity + 13

inaudible segments and segments undecipherable even to the original

speakers (when these can be brought in as informants). So one is always
dealing with an imperfect product, especially compared to the accuracy
with which the latest optical text scanners can quickly gobble up vast
amounts of written text and deposit them in machine-readable form into
the maw of even a modestly powered desk-top computer. Good tran-
scribers of spoken language have to be trained to the task. Even the best
audio typists often simply ‘miss’ relevant details (highly relevant to the
spoken discourse analyst) such as repetition and overlapping speaker-
turns, and can be (as Watts 1989 most memorably demonstrated)’ deaf to
the presence of discourse markers and other ‘little’ words which become
important the moment one starts to analyse the work they do in the
creation of interaction.

1.6 Transcribing: the black hole on the trail to infinity

Cook (1990) speaks of the problems of transcribing real spoken data as

potentially infinite, to the extent that one could, theoretically, include
in transcription any amount of contextual data, from intonation and
body language to what the participants were wearing, or the colour of
the wallpaper in the room where they were talking. A veritable black
hole lies at the end of such a quest. The CANCODE corpus has attempted
to transcribe what we consider to be relevant to our research aims and
has left unstated a number of phenomena which may well have played
a role in the configuration of the original utterances. The principle has
been a mix of perceived usefulness to language researchers and
teachers, machine- and human-readability of the text, time and cost,
and large dashes of common sense. Thus overlaps and interrupted
utterances are marked, as are truncated words and (where feasible) the
number of syllables occupied by a sub-audible segment (i.e. where the
speech can be heard but not sufficiently to make sense of it) along with
extralinguistic information where this helps the interpretation of the
text (e.g. ‘a baby is heard crying in the background’). Even with these
restrictions on what ends up in the final transcription, the text as
entered in the computer can look extremely messy and daunting to the
reader who wishes to focus mainly on content. The following extract is
(fortunately) not how things always look, but such segments do occur
very frequently:
14 + 1 Introduction

(1.1)
[Two men talking about gardening]

<$1> Worms they’re good.

<$2> Well balance of opinion on that is that er worms are generally good
excepting <$=> er the wor= <\$=> the casting worms on lawns.
<$1> I don’t think they <$01> do any harm at all <\$01>.
<$2> <$01> <$=> That they on balance <\$=> on balance
<\$01> they do more harm than good.
<$1> I don’t think I’d go to the trouble of getting rid of them.
<$4> <$G?> <$E> pause <\$E> No actually although we don’t profess to be
green we don’t use pesticides.
<$3> Mm.
<$2> We don’t use <$02> fungicide either really do we <\$0O2>.
<$4> <$02> <$=> And we don’t <\$=> Fungicide <\$02> and+
<$2> Mm.
<$4> +only weed killer on the paths.
<$1> Mm.

The symbols <$1>, <$2>, etc. at the beginning of turns tells the computer
which speaker is speaking and enables researchers to get back to archived
information about individual speakers (e.g. age, gender, etc.) if required.
Symbols such as <$O1> and <\$01> indicate the onset and end of overlaps
in competing speakers’ turns (in this case the first such overlap in this
particular conversation). The ‘equals’ sign (=) indicates an unfinished
word or a unit truncated in some way, with <$=> and <\$=> showing
truncated clauses. <$G?> means an (uncountable) number of sub-audible
syllables. <$E> means ‘extralinguistic information’. The + signs indicate
‘latched on’ talk (in this case <$4>’s final utterance is continuous, broken
only by <$2>’s Mm. A certain degree of ‘normal’ punctuation is included,
but with specialised meanings (e.g. a full stop means end of speaker turn
or low pitch termination of a unit within a turn).® (See glossary for an
explanation of terms.) Such information on the transcript makes extracts
virtually unreadable when they are being examined for their content,
and so the extracts reproduced in this book simplify the transcripts and
indicate overlaps in a more reader-friendly, visual form such that our
extract above might become:
 1.7 The status of spoken language in applied linguistics + 15

(1.2)
[Two men talking about gardening|]

<S 01> Worms they’re good.

<S 02> Well balance of opinion on that is that er worms are generally
good excepting, er the wor, the casting worms on lawns.
<S 01> I don’t think they do any harm at all.
<S 02> LThat they on balance, on balance they do more
harm than good.
<S 01> I don’t think I'd go to the trouble of getting rid of them.
<S 04> [inaudible] ... no actually although we don’t profess to be green
we don’t use pesticides.
<S 03> Mm.
<S 02> We don’t use fungicide either really do we.
<S 04> lAnd we don’t, fungicide and
<S 02> LMm.
<S 04> Lonly weed killer
on the paths.
<S 01> Mm.

In addition, we shall use square brackets to indicate inaudible segments,

and also to enclose back-channel utterances, that is when a listener says
something such as mm, yeah, uhum or suchlike simultaneously with (and
without interrupting the flow of) the speaker’s talk, for example:

(1.3)
<S 05> The nicest pizza I’ve ever had was in Amsterdam [<S 03> oh yeah] I
had a brilliant pizza

We use three-point ellipsis (...) to indicate pauses between one and two
seconds; pauses of two seconds or longer are given in round brackets, in
seconds. Commas will be used where there is a truncated structure or re-
casting of the structure, and, for practical purposes, in any other place
where serious ambiguity might arise. These conventions are purely for
the readability of extracts in this book; for research, the original tapes
and transcripts were referred to.

1.7 The status of spoken language in applied linguistics

Although it may seem obvious that spoken language is primary and that
written language is secondary in terms of their occurrence in human
16 + 1 Introduction

societies, written language became, over the centuries, quite under-

standably, what linguists and applied linguists utilised as their baseline
data. Written language was easy to observe and to codify, and the
codifications themselves, in grammars and dictionaries, being written,
took on a life of their own and acquired the status of ‘correct’ bench-
marks to which any question or dispute about usage could be referred.”
That is not to say that the spoken language was ignored in the study and
description of language or within language teaching before the advent of
tape-recorded corpora. In Tudor times, in Britain, for example, the
Vulgaria (the textbooks from which grammar school pupils learnt Latin)
were concerned with spoken Latin (see White 1932) and contained in
their Latin teaching examples and their English translations many
surprisingly colloquial utterances, such as I was beten this morning (i.e.
beaten by the schoolmaster), and Thou stynkest (you stink) (White 1932: 19)
and This bredde is moulled or hore for longe keping (this bread has gone off
from being kept too long) (Horman 1519: 142). Likewise, Ben Jonson’s
(1640) English Grammar declared itself to be:

For the benefit of all Strangers, out of his observation of the

English Language now spoken, and in use. (front matter)

Jonson further stated that:

Grammar is the art of true, and well speaking a language: the

writing is but an accident. (ibid.)

Admittedly, Jonson went on to describe the English Language in Latin

grammar terms, which had their origin in the study of written texts, but
at least in Jonson’s mind there was an important link between linguistic
description, language use and speech. The sixteenth century onwards in
Britain also saw a steady flow of manuals of rhetoric and eloquence in
speech which lay emphasis on oratory, good pronunciation and enuncia-
tion, etc. (see e.g. Sherry 1550; Peacham 1577; Holmes 1738; Herries
1773).'° In the seventeenth and eighteenth centuries, there was, too,
much scrutiny of the links between pronunciation and orthography in
English (e.g. Robinson 1617; Watts 1740). The nineteenth century saw a
veritable explosion of manuals describing the spoken languages of
speech communities newly contacted by Europeans in locations such as
New Zealand, the Orient, Africa, etc., where the absence of written texts
or their inaccessibility made a practical necessity of confronting spoken
tongues.'’ All of this goes to show that speech was never out of the
 1.7 The status of spoken language in applied linguistics - 17

minds of the community of scholars who described, taught and encour-

aged use of language(s) over the centuries, our applied linguist ancestors.
It is nonetheless true that the dominance of classical models meant that
the written code continued to overshadow how the language was
conceived, with an underlying written bias.
In the twentieth century, some early advocates for the inclusion of
spoken language within the purview of English teaching (both L1 and L2
teaching) may be found. Baker (1924) urged that mother-tongue English
teaching should not just be grammar and theoretical principles of the
language (and possibly essay writing), but should focus on language in its
social context, as a tool for spoken communication. Baker’s work was
based on a survey of business men and women, and she concluded that
the overwhelming language need was for spoken communication. The
1920s and 1930s also saw the rise of recorded media in the form of radio
and talking cinema pictures, and this influenced the debate on the status
of spoken language in education. Trueblood (1933), for instance, saw the
growth of spoken mass media as a positive thing, promoting ‘good’
speech amongst the population. Displacements of peoples during the
Second World War in Great Britain also contributed to the debate
amongst mother-tongue teachers over the status of speech in education:
evacuations from cities under threat from bombing to the countryside
more acutely focused the differences in accent and possible communi-
cation problems among dialect groups. Within this climate, Compton
spoke of the importance of training young people to speak in an ‘easy,
clear, reasonably exact, and friendly’ way, without the influence of the
‘dead hand of the elocutionary tradition’ (1941: vi-vii). It was not long,
with the advent of tape recorders, before voices were heard re-evaluating
existing assumptions about grammatical description and advocating the
collection of natural spoken data to counteract the written bias of
existing grammars (e.g. Dykema 1949). Clark (1946) in a book entitled
Spoken English, promises that grammar rules are illustrated with collo-
quial examples; however, the examples all seem still to have a strongly
written flavour, and it was to be some time before the real influence of
recording technology was felt.
From the point of view of English as a foreign language teaching,
perhaps the most significant publication of the early part of the twen-
tieth century was Palmer et al’s (1924) A Grammar of Spoken English (revised
and rewritten by Kingdon in the third edition of 1969). This extra-
ordinary book had much in it which a grammar of spoken language
18 + 1 Introduction

should ideally have, most notably its willingness to depart from single-
sentence examples and to admit what we would now call (with the
benefit of a terminology elaborated by discourse analysts in the last two
decades) exchanges and whole utterances. Its examples often take more
than one speaker into account, as in this illustration of ellipsis:

A: It must be mended.
B: What must? (p. 193)

or in auxiliary tagging:

A: I must leave early.

B: Yes, you must (p. 193)

The authors also acknowledge (ibid.: 283) that informal speech prefers co-
ordination of clauses over embedding, and that some grammatical
phenomena are the result of ‘real-time’ processing. The work of early
pioneers such as Palmer is often forgotten nowadays amid the torrent of
publications on language teaching that has rained on us since the early
1970s, but it should not be ignored, and its influence should not be
underestimated.

1.8 Spoken language and communicative pedagogy

Despite the examples cited above of voices raised in favour of the spoken
language, there is no gainsaying that spoken language remained the
poor relation of the written in most foreign language teaching right up
to the 1970s, even in those methodologies, such as the audio-lingual, that
did encourage learners to speak and drill target usages. It was not that
learners were not allowed to speak - the ‘conversation class’ and oral
tests of various kinds have a long pedigree - but that the models of
grammar and vocabulary that were the raw input were firmly based on
the written code. This was not assisted by the dominance of Chomskyan
views of language in studies of first- and second-language acquisition,
views that happily sanctioned the use of invented data and that demoted
real speech to the level of ‘performance data’ or held it to be downright
corrupt (for a trenchant critique of Chomskyan-influenced applied
linguistics, see Beaugrande 1997). The growing interest in speech-act
theory in the 1970s and 1980s may have been hoped to change all that;
indeed the communicative revolution that overtook English as a foreign
language teaching has, ultimately, made it possible to talk about an
 1.8 Spoken language and communicative pedagogy - 19

applied linguistics more at ease with the study of spoken language as one
of its important components. But the communicative enterprise has not
been without its problems. Notional-functional approaches to language
(e.g. Wilkins 1976) offered the possibility of wedding a view of language
that focused on ‘doing things with words’ (i.e. language as social action
rather than as abstract system) with the perceived need for ‘real com-
munication’ in and out of the language classroom, both in writing and,
more and more, in speech. One cannot help concluding, though, that
mistakes were made along the way. Most notably, there was an overly
simplistic tendency to equate speech-acts with particular linguistic
formulae, a sort of ‘phrasicon’ of speech acts, or ‘functions’, as they were
often popularly called, and there was a tendency simply to invent such
formulae rather than examine real data.'? This kind of reductionism
inevitably led to generations of learners who were taught, for example,
that in English, to disagree with someone, one might say I disagree with
you, or some such formulaic utterance. Real data usually show speech-
acts to be far more indirect and subtle in their unfolding. Disagreement
is a good case. In the CANCODE corpus, there are only eight occasions
where someone says I disagree, and none where with you follows. All eight
occasions have some sort of modification which suggests a reluctance on
the part of the speaker to utter such a bald statement; these include I just
disagree, I beg to disagree (context: semi-formal meeting), you see now I do
disagree, I’m bound to disagree, I’d er, I’d disagree. Where the verb-form
' disagree occurs, the contexts mostly either ‘report’ (or predict) disagree-
ment with someone, or disagree with ideas and propositions, rather than
people. This is just one of the kinds of problems a simplistic advocacy of
speech-act theory can create with regard to raw material for language
teaching. It would perhaps be reasonable to assume that other speech-
acts behave in this way too, unfolding indirectly and in negotiation, with
due sensitivity to interlocutors’ personal face. It may also turn out that
speech-act ‘performative’ verbs such as disagree, complain, invite, etc., may
be more useful as a means of reporting, predicting or in some way
talking about speech-acts, rather than performing them. Another point
about speech acts that needs to be taken into account is their staging, in
what Cohen (1996) refers to as ‘semantic formulas’, in that an act such as
‘apology’ may be composed of a number of phases (e.g. acknowledging
responsibility, promising non-recurrence, etc.). What is more, in Cohen’s
view, speech acts imply sociocultural choices (concerning the appro-
priacy of particular acts in particular situations) as well as sociolinguistic
20 + 1 Introduction

choices (which linguistic form is the most appropriate). The point is that
we can only appreciate the delicacy and subtlety of how speech acts are
realised in spoken interaction by examining real data, and the early
advocates of functional syllabuses and early investigations of learners’
performances of speech acts signally failed to do so. None of this should
surprise us, since the status of the spoken language as an object of study
in the applied linguistic enterprise had yet to reach the level it deserved.
Indeed, it is the feeling that things are still out of kilter with regard to
our ambitions to teach ‘real language’, ‘language for communication’
and ‘spoken skills’ on the one hand, and our readiness (or otherwise) to
accept revised descriptions of target languages based on what their
speakers actually say on the other hand, which is a principal motivation
of this book.

1.9 Discourse analysis, conversation analysis, and conclusion

The influence of discourse analysis and conversation analysis on lan-

guage teaching and language acquisition studies has helped improve the
status of real spoken data, not least in our endeavours to record and
analyse the spoken output of our own learners, and Chapter 3 of this
book refers to many such studies. Discourse analysts, with the useful
insights they provide of language patterning above and beyond the
sentence (see Cook 1989, and McCarthy 1991 for general surveys), offer
frameworks for higher-order structuring of syllabuses and materials (see
Aston 1988b; McCarthy and Carter 1994: ch. 5). The American traditions
of conversation analysis and ethnomethodology, on the other hand, offer
the possibility of fine-grain descriptions of how participants orient
themselves towards mutual goals and negotiate their way forward in
highly specific situations (see Psathas 1979, 1995; Atkinson and Heritage
1984; Boden and Zimmerman 1991).. These insights can have direct
applications in the construction of materials and activities for language
teaching. Conversation analysts study ‘talk in interaction’; all aspects of
interaction (including the non-verbal and non-vocal) are considered
relevant, and the recurrence of ‘order’ in talk is the crucial object of
investigation (Psathas 1995: 3). Much of this sort of work, however,
remains little known to many practising language teachers, and sus-
picions that studying the details of real spoken language, with all its
stops and starts and faltering utterances is the study of illiteracy, slang
and lazy speech habits abound. Even more wortying is the accusation
 1.9 Discourse analysis, conversation analysis, and conclusion + 21

that studying real speech is part of yet another conspiracy to impose

southern British speech habits (or American speech) on a global commu-
nity desirous of using English in culturally independent international
contexts. Such views still abound in the language teaching profession
and may be heard at any large language teaching conference (see also
Prodromou 1997). This book has no such conspiracy in mind, but its
author is curious to know just what ‘global’ or culture-free varieties of
English would look like as a syllabus. Whatever work has been done on
spoken British English, in corpus projects such as the Survey of English
Usage, COBUILD, The British National Corpus, the ICE corpus or
CANCODE, can (and ideally should) be duplicated for any dialect or
sociolect of English (including ‘global/international’ English) and for any
other language or dialect where there is a need for pedagogical model-
ling. The problems in deciding what can be called ‘standard features’ of
spoken (or written) language will likely be the same whether it is British
English or some other variety that is being described, as Mair (1992) has
shown in relation to a corpus of ‘standard’ Caribbean English. The point
is one has to start somewhere, and start with an open mind. Despite
continuing problems, the prospects for better spoken corpora are good,
and the contributions to language pedagogy and other applications thus
made possible by both the quantitative and qualitative study of spoken
language can only increase. This book respects the spoken language, sees
it as our most important raw material in understanding language in its
social context, and wishes its place within pedagogical models to be at
the forefront. We will not jettison the written language; it too must
remain the raw material, but we will more and more be able to compare
the written and spoken in a way that will complement our under-
standing of both ways of communicating.
One final point that needs to be made is that it is not my position that
only by examining corpora can language teachers come to understand
how spoken language works. Corpora are useful, and are a good short-cut
to seeing how language forms are used across a variety of users, but even
the biggest spoken corpus pales into insignificance compared with the
number of words ‘processed’ over the years by an adult language user. It
must also be constantly remembered that computers may ‘have knowl-
edge’ of what has been spoken, but cannot ‘use’ that knowledge.
Proficient users of a language may not be so good at reflexive ‘knowing’,
but use their knowledge whenever they speak. It is thus only when good
observers of language combine their talents with the display and analysis
22 + 1 Introduction

of data by the computer that the optimum gains can be made. We must
not, therefore, diminish the importance of experienced applied linguists’
and language teachers’ experience, observation and intuition (see Owen
1996 for persuasive arguments in favour of teachers’ intuitions being
respected). What is more, the absence of any feature from a corpus that
actually does occur in use does not mean it is a ‘freak’, or any less worthy
of study. Finally, it is my experience that ‘reading’ the corpus like a book
or like a living drama script cannot be substituted by mere number-
crunching by the computer. The right balance between quantitative and
qualitative analysis of the corpus is crucial. We might illustrate this with
the word just.
If we examine the more than 6,000 occurrences of this word in
CANCODE, we see that, numerically, it overwhelmingly collocates to the
‘right-hand side’ with high-frequency verbs, such as go, have, said, think,
come, put, want and take. These represent its ‘semantic preferences’
(Sinclair, personal communication). However, equally important are its
‘left-hand’ collocates, which reveal an unexpectedly high occurrence of
modal verbs (can, could, would, might, should). These left-hand phenomena
tell us much about the semantic ‘prosody’ (ibid.; see also Louw 1993) of
just, in that it seems to occur in contexts of tentativeness/indirectness/
face-saving (CANCODE examples include: Could you just pass the gravy; Can
you just take me through stage by stage; Could you just tell me about the
department). How this actually operates can then only really be fully
appreciated by examining individual conversation segments in detail.
Just therefore becomes a significant pragmatic particle, operating as a
marker of politeness/indirectness, as well as retaining its traditional
place in the language-teaching lexicon of combining with have + past
participle to express recent events.'? These qualitative judgements
are
crucial, and come only from close observation allied to the number-
crunching power of the computer.'* Being open to what the power of the
computer can reveal and not approaching the data with any prejudice
about how words work is part of that philosophy that Tognini-Bonelli
(1996) refers to as a ‘corpus-driven’ philosophy. This she contrasts with a
‘corpus-based’ approach, where the philosophy and ideas are taken for
granted beforehand, and the corpus is simply used to reinforce those
ideas. The ‘corpus-driven’ way is the one which demands even more
qualitative work, since the insights available are neither preconceived
nor do they always simply leap out of the statistics. CANCODE is corpus-
driven in this sense. However, we also reserve the term ‘corpus-informed’
 Notes - 23

for what we do with the insights in pedagogy, since insights alone are no
guarantee of good teaching, and must be mediated in some way to create
models that are meaningful and useful to language learners. This may
include editing corpus extracts before using them in class, or con-
structing a role-play activity based on the phases of a particular spoken
genre, rather than using an actual transcript as part of the activity. One
final remark that needs to be made here is that we should also be aware
of the dangers of taking over wholesale the metalanguage of written
corpus studies when analysing spoken data. It is not for nothing that I
put ‘left-hand’ and ‘right-hand’ in scare-quotes above. ‘Left-hand’ and
‘right-hand’ are page-driven metaphors, based on the written output of
concordancing programs. If we are really true to spoken language, we
should be talking about ‘prior-’ and ‘post-’collocates, not left and right,
since spoken language exists in time, not space. I return to this problem
in section 3.2.2, in relation to grammatical terminology.
As long as we keep a cool head in the face of the exhilaration of
computer power and vast arrays of text, we will not fall into the
temptation of substituting cold numbers for the real people who actually
produced the words.!° I can identify best with Chafe’s (1992) definition of
the corpus linguist as someone attempting to understand language:

‘... by carefully observing extensive natural samples of it and

then, with insight and imagination, constructing plausible
understandings that encompass and explain those observations.’

That is what I shall try to live up to in the rest of this book.

Notes

1 Another example, which includes informal spoken language by adults, as well

as by selected age groups of children from 6 years upwards, in a corpus of
some 84,000 words, may be found in Carterette and Jones (1974).
2 I refer here to methodologies of data collection, not end use. For a discussion
of the different motivations for the use of corpora, see Nelson (1992).
w Inevitably, some more rehearsed types of discourse creep in here and there,
for example (semi-) prepared classes and tutorials in the ‘Pedagogical’ cate-
gory, or indeed, personal anecdotes, which may well often represent the
umpteenth telling of a particular story, and therefore cannot truly be called
‘unrehearsed’.
4 1 am grateful for assistance with the choice of the term socialising for this
category to Merrill Swain of OISE, University of Toronto, Canada. Merrill
24 - 1 Introduction

persuaded me that our earlier name for the category, sociocultural, ran the risk
of being too general and of overlapping confusingly with current work in
socio-cultural theory and second language acquisition.
ul Sinclair actually has a slightly different idea in mind, that of the continuously
selfupdating corpus which takes advantage of an automatic inflow of new
data (e.g. newspaper texts electronically imported); this is unrealistic with
regard to spoken data in the present state of technology.
Atkins et al (1992) make the point that balance can only be achieved after an
initial corpus has been built. One might suppose this to be as true of written
corpora as it is of spoken, and one should not assume that existing text
typologies for written language are a reliable reflection of the output of a
speech community, most of whom rarely if ever indulge in the kinds of
written production that inform some written corpora, even if they may have to
read such material.
See also Stubbe and Holmes (1995) on public perceptions of discourse marker
usage.
co For information on the British National Corpus transcription methods for
spoken language, see Crowdy (1994).
No) Mitchell (1957) states: ‘It is certainly the common view that the written form is

the only one deserving serious attention, study and cultivation.’

10 See also Hale (1903) for a summary of 16th century views of rhetoric. Courtly
eloquence was also considered a selling-point for at least one language
textbook for learning French ‘... as it is now spoken in the court of France’,
which included ‘familiar dialogues, the niceties of the French tongue and
twelve discourses.’ (Boyer 1694).
il The reader is referred to the Nineteenth Century bibliography on CD-ROM
published by Chadwyck-Healey (1994), where numerous examples of language
manuals based on ‘exotic’ (for the Westerners of that time) spoken languages
may be found. This was true of Japanese, for instance (e.g. Mutsu 1894), but
was also true the other way round (e.g. Coningham’s 1894 course in business
conversation ‘Specially written for Japanese merchants dealing with for-
eigners.’).
12 Even more fundamental problems are inherent in speech-act theory when
applied across cultures, since the Western-based model of the individual act
fails to take cultural and social contexts into account. See especially Rosaldo
(1982) for a good critique of speech-act theory from a cross-cultural viewpoint.
13 Aijmer’s (1985) study of just in 170,000 words of spoken language notes that it
is many times more frequent in spoken rather than written language, but she
too takes the matter much further than numerical proportions and interprets
several different functions for the word, the most common being, in her data,
an emphatic particle stressing the truth of the utterance. Sinclair’s approach,
involving semantic prosody, looks at a more general level of context, implying
an even greater degree of qualitative analysis.
 Notes - 25

14 Another example of the need for qualitative interpretation is Tottie’s (1983)

discussion of the numerical fact that spoken English contains much more
negation than written English, which she puts down to the pragmatics of
interaction and the presence of acts such as explicit denials and rejections in
the spoken.
15 Biber (1990) rightly urges that we should not be ‘intimidated’ by large corpora,
and that smaller corpora are quite adequate for many purposes.
2
Spoken language and the notion of genre

2.1 Introduction

As stated in Chapter 1, in the design of the CANCODE corpus, it was

decided to aim for adequate coverage of everyday genres that would be
useful for pedagogical purposes. However, the notion of genre was not an
easy one to define or to put into operation. Much good research has been
done on genres in more specialised varieties of written language (most
notably Swales 1990; also Christie 1986; Reid 1987; Martin 1992), and a
broad definition can be gleaned from such work, which emphasises the
socially-rooted nature of genres and their recognisability for participants
within ‘discourse communities’. In this respect, the fact that native
language users can label written and spoken discourses with genre-
names is clearly significant from the point of view of the recognisability
of genres for participants themselves (see Walter 1988: 6); one does not
have to be a linguist to recognise a ‘story’ or ‘an argument’. Degrees of
institutionality are also an important factor, and are most noticeable in
highly conventionalised contexts such as academic writing, scientific and
technical reports, literary genres such as the novel or the sonnet, and so
on. Stubbs (1996: 12) stresses the mutually defining nature of institutions
and genres (though he alternates between ‘genre’ and ‘text-type’ to refer
to characteristic modes of expression): as institutions change and
develop, so do the text-types that give voice to their activities. And yet the
question remains as to how we recognise the relevant linguistic facts
that reveal the presence of different genres, how participants orient
towards them and how they show their recognition of them. These
problems are particularly acute in spoken language, where, apart from
well-studied genres such as service encounters (Merritt 1976; Hasan 1985;
Ventola 1987; Aston 1988a; Iacobucci 1990) and narratives (Labov 1972;
Jefferson 1978; Polanyi 1981; Goodwin 1984), many of the everyday forms
of talk we engage in remain unclassified in generic terms. There have
been a number of studies of ‘register’, i. the relationship between

26
 2.2 Theories of speech genre + 27

language features and their context of utterance (most notably Halliday

1978), much of which research focuses on levels of formality, inter-
personal aspects of meaning, and spoken/written differences. Within the
context of register studies, Biber’s (1988) seminal work on the distin-
guishing features of written and spoken texts and how key language
features cluster in different types of texts overlaps to some extent with
the study of genre. In his later work, Biber uses the term ‘register’ for ‘all
aspects of variation in use’ (Biber 1995: 9), and his research covers aspects
that we shall here relate to genre. Register studies contribute greatly to
an understanding of the different factors that influence linguistic choice,
but do not offer a clear model of what, for example, a service encounter
is, and how participants show their recognition of being engaged in such
a generic activity.

2.2 Theories of speech genre

Perhaps the most notable early example of the type of study that will
interest us in this section was Mitchell’s (1957) investigation of the
language of buying and selling at markets and shops in Cyrenaica.
Mitchell’s study was concerned with how different aspects of the context
of situation (the participants, the setting, purposes, etc.) shaped the
language that was used between buyers and sellers into a recognisable,
patterned form of interaction. In a manner that has since seen echoes in
the work of spoken genre analysts such as Hasan (1985) and Ventola
(1987), Mitchell identified stages in service encounters he observed.
Service encounters are interactions concerned with the transaction of
goods, information and services, most typically exemplified by conversa-
tions in shops, restaurants, etc. The stages Mitchell identified included
salutation > enquiry as to the object of sale > investigation of the object
of sale — bargaining — conclusion. Within each stage, considerable
variation is possible (for example, generated by different spatial relations
between the participants - open air markets created different proxemics
to those in closed markets), and whole transactions proceeded dynami-
cally to construct texts, summed up by Mitchell in a memorable piece of
imagery:

A text is a kind of snowball, and every word or collocation in it is

part of its own context, in the wider sense of this term; moreover,
the snowball rolls now this way, now that.
28 + 2 Spoken language and the notion of genre

Since Mitchell’s seminal research, the notion of genre in spoken language

has been explored by various other linguists, most notably Hymes (1972),
who sees genre as a higher-order feature of speech events. Hymes stresses
the dynamic characteristic of genres, and separates them from the speech
event itself: a genre may coincide with a speech event, but genres can also
occur within speech events, and the same genre can show variation in
different speech events. Much of the subsequent debate has centred
round this question of dynamism and local variation in the actual
realisation of genres. Most linguists entering the debate seem to accept
the (theoretical) existence of genres as recognisable norm-governed
activities comprising varying degrees of institutionalised linguistic and
non-linguistic behaviour. Coupland (1983) stresses this variability and
offers the contrast between buying and selling stamps or newspapers and
buying a holiday at a travel agent’s: both are ‘service encounters’ but the
latter is less likely to be played out as a ritual, following a template, and
is likely to offer more possibilities for interactional/relational talk (i.e.
talk oriented towards the establishment or maintenance of social rela-
tions) alongside the transactional talk that gets the business done.
Bargiela-Chiappini and Harris (1995) make a similar point in distin-
guishing between the more institutionalised roles and goals in settings
such as service encounters as compared with (in their case) business
meetings, where activity may be more fluid and open to variation. Ylanne-
McEwen (1997) has shown in great detail just how significantly the
transactional and relational elements intertwine dynamically in service
encounters (once again in the travel agency context); so much so that any
modelling of genre would be fruitless without at least equal regard for
the interactional/relational process as well as the transactional pro-
cess(es) that are realised in typical encounters. Underlining this, Linden-
feld’s (1990) study of small talk in urban French market places shows that
relational ‘small talk’ is far from random, with vendors’ small talk
focusing mainly on utilitarian concerns, while customers’ small talk
tends strongly towards personal topics. Vendor and customer are both
constructing and re-affirming their identities in the discourse. And,
although predominantly expressed in ethnographic/conversation-
analytic terms, Komter’s (1991) study of job interviews also considers the
small talk that usually takes place at the beginning of an interview as a
legitimate phase in the unfolding interview process, echoing Mitchell’s
(1957) seminal study by talking of the ‘phase structure’ (p. 54) of the
interview as a whole. Thompson (1997) in an in-depth study of university
 2.2 Theories of speech genre + 29

oral research presentations, also stresses the building of an appropriate

relationship between speaker and audience and the co-construction of
roles as being an integral part of the research presentation genre. The
research presenter usually constructs him/herself as ‘the modest, self:
deprecating expert’ (p. 334) and engages in complex facework, that is to
say, protecting him/herself and the listeners from potential damage to
mutual esteem, in order to establish and maintain that relationship. All
this is important, since it suggests that models of genre that look upon
relational episodes (episodes concerned with establishing and reinforcing
social relations, as opposed to ‘doing business’) as things that merely
‘crop up’ on the record in an unpredictable way and disturb the
normative ‘flow’ of the transactional elements of the genre are mis-
leading and misguided. As we shall see below, a view of genre that gives
equal attention to the relational unfolding must have significant implica-
tions for models of teaching the spoken language.
Some studies focus more on the variable sequences and mixing of
activities that characterise genres. Duranti (1983) argues for a dynamic
view of genre: the same genre can be performed in different ways
according to the particular event, depending on who the speakers are,
what their purposes are, where the genre occurs in sequence in the speech
event, etc. Walter (1988: 2-3), in describing the generic characteristics of
jury summations, also lays emphasis on the setting in which speech
occurs as a crucial variable. Fairclough (1995: 167ff) highlights the way
genres are sequenced and intermixed, using broadcast political discourse
to exemplify how generic characteristics can change over time, in this
case the process of what he calls the ‘conversationalisation’ of political
discourse. The question of sequencing of elements, separating obligatory
and optional elements, and, above all, how such elements are recognised
amid the variation common in genre-building activity is stressed by
Eggins and Slade (1997: 230-5), who also, in the way that the present book
attempts to do, recognise the importance of lexico-grammatical analysis
(e.g. lexical evaluation devices) as well as the analysis of elements beyond
clause-level such as turn-taking or adjacency pairs.
Bakhtin’s work (especially Bakhtin 1986) has been very influential in
promoting an understanding of genre. Bakhtin’s notion centres on the
‘utterance’, a somewhat abstract unit of talk which may vary in length
from one speaker-turn to a whole monologue or (in written language) a
whole novel, the defining feature being its termination at a point where
an interlocutor may respond.’ Bakhtin sees utterances as reflecting
30 + 2 Spoken language and the notion of genre

specific conditions and goals of different areas of human activity, which

reflect those conditions and goals not only through their lexical, gram-
matical and phraseological configurations, but through their ‘composi-
tional structure’ (1986: 60). Whilst utterances are locally configured and
individual, ‘each sphere in which language is used develops its own
relatively stable types of these utterances’ (ibid.), and these stable types
are what constitute genres. Also central to genres for Bakhtin, as well as
their relative stability, are their interpersonal aspects: ‘each speech genre
in each area of speech communication has its own typical conception of
the addressee, and this defines it as a genre’ (ibid.: 95). Kelly Hall (1995)
reinforces the Bakhtinian perspective on interaction by exploring the
interplay between the conventionalised/socio-historical meanings of the
generic resources available to interactants and their practical strategies
in each new situation they encounter. The importance of Bakhtin’s work
(and indirectly, too, that of Vygotsky; see Emerson 1983; Wertsch 1985) is
that it breaks down the distinction between language as the product of
the individual psyche and language as a social construct. Any theory of
genre needs to include that perspective.

2.3 Goal-orientation in interaction

A picture of dynamism, fluidity, variability, mixing and negotiation

emerges as the current consensus on spoken genres. This is to be
expected, since conversationalists, especially in less rigidly institution-
alised settings, are social animals with practical goals, and it is their
goals that drive interaction, rather than some sense of obedience to
generic norms. What is more, goals may not be fixed or pre-ordained, and
may be emergent within the discourse (i.e. become apparent as a result
of the interaction, rather than shaping it from the start), and multiple in
nature (Tracy and Coupland 1990). Seeing genre in terms of participant
goals is paradoxically both fruitful and problematic. If we take a goal-
orientation view, we are able to integrate more satisfactorily the trans-
actional elements of conversations and the relational/interactional
elements: Iacobucci (1990), for example, shows how relational episodes in
customer calls to a phone company are far from marginal or simply to be
seen as ‘side sequences’, but are often clearly oriented towards fulfilling
the transactional goals of the discourse more efficiently. What is often
termed ‘casual conversation’ is perhaps the prime example of how useful
the study of goal-orientation is. Casual conversation displays a variety of
 2.3 Goal-orientation in interaction - 31

features which have led some to the view that it is too vague a notion to
qualify for the label of genre, or else that it is defined by the very fact
that, in terms of genre-mixing and embedding, ‘anything goes’. Yet
casual conversation is no less goal-driven than any other type of talk,
even though the goals may be multiple, emergent and predominantly
relational. In (2.1), a group of female students are chatting casually over a
cup of tea on a Sunday evening. The first topic is ‘Sundays’, but it is clear
that the overall purpose of the chat is to reinforce the camaraderie the
students have built. Then there is a sudden, apparently incoherent
switch in topic as one of the girls notices a piece of jewelry. Such switches
are unproblematic in casual conversation, since the underlying rela-
tional goals can provide coherence, here taking the form of compli-
menting, admiring, approving, etc.

(2.1)
<S 03> I like Sunday nights for some reason, I don’t know why.
<S 02> [laughs] cos you come home.
<S 03> I come home.
<S 02> You come home to us.
<S 01> And pig out.
<S 02> Yeah yeah.
<S 03> Sunday’s a really nice day I think.
<S 02> It certainly is.
<S 03> It’s a really nice relaxing day.
<S 02> It’s an earring it’s an earring
<S 03> LOh lovely oh, lovely.
<S 02> It’s fallen apart a bit but
<S 03> It’s quite a nice one actually, I like that I bet, is
that supposed to be straight.
<S 02> Yeah.
<S 03> Oh I think it looks better like that.
<S 02> And there was another bit as well another dangly bit.
<S 03> What, attached to
<S 02> lThe top bit
<S 03> lThat one
<S 02> Yeah ... so it was even
<S 03> |Mobile earrings
<S 01> What, that looks better like that it looks better like.
that
32 + 2 Spoken language and the notion of genre

So, looking at relational goals can often help us understand casual

conversation better than simply pursuing notions such as ‘topic’ or
‘business in hand’. On the other side of the coin, the problem with
studying participant goals is how one actually determines what the goals
are, since these will often not be explicitly stated by participants, and the
evidence for the analyst is usually indirect, and available only in the
shape of phenomena such as ‘formulations’, and other similar kinds of
linguistic evidence we shall consider below. Formulations are para-
phrases of previous talk or summaries of positions reached in the
ongoing talk, whereby participants articulate their view of the directions
and goals of the unfolding discourse (Heritage and Watson 1979), which
the analyst can use as direct evidence for statements about the way the
discourse is progressing.
Goal-orientation is part and parcel of a view of genre that ties it closely
to action. Dolz and Schneuwly (1996) see the link between genre and
social action as a defining characteristic, and the ability to use the
generic resources to achieve goals as inseparable from the ability to act
in the immediate social situation. As always, however, such concepts are
elusive in real data, and the analyst is working with imperfect records of
what actually happened. Nonetheless, attempting to see things from the
participants’ viewpoint and how they articulate their own understand-
ings at least avoids the worst excesses of the imposition of an order by
the analyst, using rhetorical frameworks that may not reflect in any way
the reality of the conversational encounter for those involved in it.

2.4 Genre as social compact

Here I shall take the line that genre is a useful concept that captures the
recurrent, differing social compacts (i.e. co-operative sets of behaviour)
that participants enter upon in the unfolding discourse process, whether
writing or speaking. Sociologists provide frameworks whereby such
social practices can be seen to create and reflect or are constrained by
social structure. Sociolinguists and ethnographers in their turn observe
how participants orient themselves towards necessary actions such as
establishing roles and identities, protecting face, achieving goals, and so
on. The genre analyst stands somewhere between these and actual texts;
it is a series of textual extracts of recurrent events which are the genre
analyst’s hard evidence. The texts are not the genres in themselves, they
are simply patterned traces of social activities. It is the task of the genre
 2.4 Genre as social compact + 33

analyst to construct the bridge between texts and socially constituted

activity so that texts become meaningful and can yield clues (probably no
more) to their original, real-time processes of unfolding. To this end, the
genre analyst may work with a variety of types of evidence, some more
immanently linguistic than others. In this way, different types of ‘generic
activity’ become the focus, rather than a static notion of ‘genres’.
But we must be sure that we can adduce at least some evidence that
generic activity has a socio-psychological reality for language users and
that we are not just creating an edifice for the satisfaction of analysts.
Amongst the many linguistic traces of genre-oriented behaviour which
may be found in texts are features that may be classified along four
dimensions, which do suggest that generic activity has a real basis in
language use. In addition to ‘formulations’, already mentioned above, I
shall refer to these under the headings of ‘expectations’, ‘recollections’
and ‘instantiations’.

Expectations

Speakers may signal expectations regarding the kind of generic activity

that is to be negotiated prior to or in the course of realising social
compacts. This corresponds more or less to what Kelly Hall (1995) refers
to as ‘expectations of the use of the [generic] resources’, and is prospec-
tive. Most obviously one finds this in spoken narratives in the form of
bids to embark upon story discourse:

(2.2)
[Speaker is talking about recent earthquakes in Greece.]

<S 01> About seven people died outside Athens, more down in Corinth
but this one guy, [laughs] this is true, there was a guy down down in
Mega and that town got you know it got really badly hit

Here, complex expectational signalling occurs in the transition from

‘reported fact’ to ‘true story’, including the topic-shift marker but, and
the narrative-specific use of this instead of a (see Wald 1983), which opens
up the expectation that a story is about to happen (with the usual
suspension of casual turn-taking routines, a particular kind of listener-
ship, etc.). Also there is laughter (despite the tragic context, this is going
to be a funny anecdote), and a claim to truth, all of which express an
expectation of the kind of activity that will unfold. There is, of course, no
guarantee, in spoken interaction, that that is what will unfold. But not
34 + 2 Spoken language and the notion of genre

only narratives and institutionally-sequenced activities such as service

encounters show evidence of participant expectations: the following is
the opening talk between two people who are about to prepare a punch-
bowl for a family party:

(2.3)
<S 01> Right.
<S 02> Right ... so where.
<S 01> Oh we’re going to need more than that [laughs].
<S 02> More orange juice.
Here we have again quite complex expressions of expectations. The
discourse-marker right sets up the expectation that a new phase of activity
is about to commence, the so suggests order and considered action, we’re
going to need more signals an expectation of further actions which will be
collaborative, not authoritarian, the oh and the laughter introduce an
expectation of more relational/interactional elements of affect and social
enjoyment. <S 02> converges with the expectation set up, echoing
<S 01>’s use of more. This can be seen as a kind of communicative
accommodation, as Giles et al (1991) define it, the ‘strategy of adapting to
each other’s communicative behaviours’. We can sense generic activity
emerging in the form of convergence on the level of actions and social
relations, particular alignments of the participants, and an emerging
compact (they are deciding as they go along how things will proceed; they
are not pre-determined) that anticipates the activity that will follow. The
text we are left with as analysts at the end of the activity (when the
punch-bowl is full) may well be a ‘language-in-action’ generic type (i.e. an
interaction where the language is generated directly by the actions of the
participants; see section 6.2), but at its incipient stage it is simply
working its way towards the product we as analysts call language-in-
action, by setting up expectations along different social parameters that
make the emergence of a language-in-action text most likely.

Recollections

Recollections refer to participants’ past experiences of social activities

and are manifest in the ‘voices’ (in the Bakhtinian sense)? from other
discourses that echo in our discourses. These may be ritualised expres-
sions that overtly mark activity as recurrent and patterned (e.g. Did I tell
you the one about ...? indicating repeated joke-telling encounters, or Okay
 2.4 Genre as socialcompact + 35

as usual we're going to start by talking about ... as an opener to an informal

meeting), or more oblique references to related or relevant other dis-
courses. In (2.4) a group of students are chatting informally during a
university small-group tutorial at a point where the tutor has left the
room to allow the students to get on with some group-work uninhibited
by his presence:

(2.4)
<S 01> It’s really embarrassing, he always comes back into the room
when nobody’s talking! [laughter]
This marking of the context as recurrent, and (in this case) unsatisfact-
orily so, manifests at least some desire to orient towards an ideal pattern
of behaviour known and familiar to the participants (perhaps from other
tutorials with different tutors): to the students, the tutorial is seen as an
occasion when they should do the talking. That tutorial group-work
should be a collaborative activity accompanied by discussion is recog-
nised in the participants’ amusement and/or discomfort (the laughter) at
what they see as their past failures to play their proper roles. Their
recollections are foregrounded here and clearly constrain their orienta-
tion towards the current activity. In (2.5), recorded at a hairdresser’s, the
activity that begins with the customer sitting in the cutting chair evokes
previous occurrences, and the hairdresser acknowledges her own re-
peated ‘voice’:

(2.5)
[<S 01> hairdresser <S 02> customer]
<S 01> Now, are you all right?
<S 02> I’m fine thanks and you?
<S 01> I’m fine thank you, yes [<S 02> [laughs]] are we cutting it as
normal or anything different or?
<S 02> Erm any suggestions, or [laughs]? I always ask you that
<S 01> Without touching the back.
<S 02> La.
<S 01> I mean you could go very wispy into the neck and sort of have a
wedge.
<S 02> Yeah.
<S 01> Keep that back wedge, keep that very into the neck like sort of
wedge.
36 + 2 Spoken language and the notion of genre

Here it is as if the participants are trying to break free from their past
pattern of behaviour. The use of we to the customer invokes informality
and signals the regularity and echoic nature of things, but both parties
then freely enter a more negotiative phase that results in a different
discourse (we suppose) from the normal one, and new possibilities for
convergence (that is to say the ‘meeting of minds’ that assists the
discourse in efficiently achieveing its goals) are opened up.

Formulations

Formulations comment on the current, ongoing activity in terms of its

present progress, with speakers periodically summing up where they
think the discourse is. They are different from recollections, in that
recollections refer to past, related discourses, linking current activity
with previous experience. (2.6) carries on from (2.5), with the hairdresser *
and customer continuing to talk about what is to be done to the
customer’s hair:

(2.6)
<S 01> Would you like height?
<S 02> I do like height yes it has gone a bit sort of flat on the, yeah.
<S 01> That is the only thing.
<S 02> Yeah.
<S 01> I mean the top heavy here, but Iknow you like the height.
<S 02> You can do it as normal.
<S 01> Do you want it cut over your ears?
<S 02> You did that last time and I, it was alright but I wasn’t too struck
you know.
<S 01> So basically you want it cut shorter?
<S 02> Yes.

The hairdresser sums up the gist as she perceives it and offers it as a

formulation of the current activity. The customer is, of course, free to
reformulate the assessment. Such formulations enable participants to
take the conversation in collaboration from one staging post to another.

Instantiations

The first hairdresser extract, (2.5), simultaneously allows us to talk about

instantiations, the fourth and most frequent type of evidence crucial to
 2.4 Genre as social compact + 37

understanding how genres unfold, for, in recognising the possibly

socially stultifying nature of the repeated discourse, the hairdresser and
customer activate a procedure for instantiating a new set of goals and it
is such instantiations that give generic activities the fluidity and dyna-
mism that make them often seem unamenable to classification. Instan-
tiations are goal-oriented, both in the sense that they enable
transactional elements to proceed more efficiently (for example, signal-
ling a change of mind in ordering food in a restaurant, politely or
otherwise cutting someone short to produce closure, etc.), and in that
they may be interactionally-oriented (e.g. complimenting/joking), or both
simultaneously (as the hairdresser extract would seem to be, creating
both a better social bond and agreeing the business to be carried out). In
(2.7), recorded in a bookshop, note how the customer instantiates the
closure of the encounter, which takes several turns before it is accom-
plished and the assistant converges. The extract begins at a point where
the assistant is explaining that the book the customer wants is not at
present available but will be soon:

(2.7)
[<S 02> assistant <S 01> customer]

<S 02> It’s on a stock list so we should have it back in the space [<S 01>
mm [inaudible]] we don’t have it here now.
<S 01> Ili probably call here again, there’s no particular
<S 02> lSure, it’s
going on here Handy’s [inaudible] [<S 01> mm] space on here instead
of
<S 01> Erm.
(8 secs)
<S 01> Well Pll call back then
<S 02> l’m sorry about that
<S 01> |see whether it’s
<S 02> Lr
make sure, it will take to, going through again but it erm probably
will take a couple of days it’ll be back [<S 01> yes] in stock again
[<S 01> yes].
<S 01> Thanks very much indeed.
<S 02> Thank you.
<S 01> Thank you.
38 + 2 Spoken language and the notion of genre

Expectations, recollections, formulations and instantiations are not

always discrete entities, nor would we expect them to be, but they do
offer a suggestive classification of different orientations towards generic
activity. Such orientations are ever-present, but only manifest themselves
overtly at particular points in the discourse when the goals require it.
Thus what we have termed expectations might typically, but not necessa-
rily, surface during openings, whilst instantiations would include deci-
sions about topic change, closure, etc., and operationalising topic shift,
pre-closure, etc. But, whereas it seems possible to talk about participants’
more global orientation towards generic activity, considerable problems
still remain in attributing generic significance to the numerous parti-
cular, selected language forms that comprise speech events. For anyone
in search of a useful pedagogical model of genre, at some point the
observation of form-functional correlations, whether locally (i.e. lexico-
grammatically; see below), or also, ideally, on a more global scale (where
we can observe regular patterns of events such as openings and closings,
topic change, etc.), becomes a matter of great interest, and it is a problem
that genre theorists have only supplied partial answers to. Below we shall
consider some evidence for the generic significance of local, lexico-
grammatical features present in texts of different types in the CANCODE
corpus.

2.5 Integrating higher- and lower-order features

Comparisons of speech texts from different settings and with different

participants often yield lexico-grammatical similarities that enable us
to observe generic patterning wherein the lower-order features of lexis
and grammar correspond to higher-order features of goal-type and
context-type in the CANCODE matrix (see section 1.4). Equally, differ-
ences in distribution of lexico-grammatical features across texts may be
significant of different context-types as well as goal-types. Biber and
Finegan (1989) show this clearly with differences between written and
spoken texts which can be in many respects similar, but in crucial
respects different. For example, in their corpus, personal letters share
features with many non-conversational spoken genres, but are
distinguished by a greater number of ‘affect markers’ (e.g. I feel; see Biber
1988: 131-3 for further discussion). Let us compare some CANCODE
extracts to illustrate this type of variation. (2.8) continues the punch-
bowl filling task which opened in (2.3) above. More speakers now
 2.5 Integrating higher- and lower-order features + 39

participate, a mixture of close family relations and more distant ones, all
assembled for a family party. The problem under discussion is how to
cool the punch down quickly.

(2.8)
<S 01> If you put this in the freezer
<S 02> LYes
<S 01> LThat’ll cool it down quicker won’t
it.
<S 02> Yes and it won’t freeze
<S 01> LNo no
<S 02> LCos of the alcohol anyway so.
(4 secs)
<S 03> Oh yeah there should be room in the top here.
<S 01> Mm yeah that’ll cool it down.
<S 04> lErm there’s erm two of them.
<S 01> That'll cool it down very quickly.
<S 02> Orange ... oh.
<S 05> Too much orange in there is there Tone.
<S 02> No there wasn’t enough orange.
<S 05> Well the extra bit you’ve got there.

The goal-type is a collaborative task (see section 1.4) and the context is
‘socialising (i.e. not intimate at this moment, even though some of the
participants, being close family members, may sometimes behave inti-
mately). We might safely say that there is a high degree of shared
knowledge, both in terms of the goal and in the fact that the physical
setting is immediate and visible. The atmosphere is relaxed and informal.
Speech is not constant, and silences, intolerable in other contexts (e.g.
the four seconds indicated), are acceptable while action is ongoing. These
features are reflected at the lexico-grammatical level. Immediacy of
context and the close relationship between the words used and the task
being carried out (what Ure 1971 calls ‘language-in-action’) is reflected in
a high number of deictic items (this, the, here, there, etc.) that point to
things in the immediate environment, discourse markers (oh, well), and a
low number of full lexical words (since things do not need to be named).
These lexical words are sometimes repeated (cool, orange), so that the
lexical density of the text (a measure that counts the number of
contentful, lexical items as a percentage of all the words in the text) is
40 + 2 Spoken language and the notion of genre

only around 22% (see section 6.2 for a further example and discussion),
which is very low (the average density for all texts, written and spoken
together, usually comes out at round 40% for full content items).
Contractions of subject and verb abound, and there is also a possible
ellipsis of initial there is in Too much orange in there is there Tone (though this
could equally well be simply considered as a feature of flexible word-
order in highly informal speech).
If we shift the context to an intimate one (brother and sister, at home,
alone, packing things for sister to go away to university), we see some of
the same features again, with some accentuated:

(2.9)
[<S 01> sister <S 02> brother]

<S 01> Mm can we try and get this stuff downstairs.

<S 02> Well you take that, you give yourself a er hernia ... so is this this
styling brush going in or, it won’t go in will it.
<S 01> Which one... I’m gonna use it.
<S 02 Oh right so you want it left up here.
<S 01 Yeah.
(14 secs)
<S 01> Cos I got er things like plates and.
<S 02> I'll leave the stereo somewhere where you can’t find it okay.
<S 01> I’m not going without that.
(4 secs)
<S 01> Helen’s just gonna go right.
<S 02> Don’t they look after it, it was all, it was all dusty when I got it
back off holiday, looks cleaner again, only cleaned it before I came
back from university.
<S 01> I’m wearing this.
<S 02> Oh yeah I was gonna take them down actually.

The degree of shared knowledge is high here in terms of the immediate

physical setting, but lower in terms of the goal, i.e. precisely which items
must be packed, which the sister must decide. Unsurprisingly, the lexical
density is higher (28.5%), corresponding to the wider range of items in
the environment, from which important selections have to be made by
the sister, but it is still low, well below Ure’s 40% general average for all
texts. Two of the lexical items are vague, general ones (stuff, things),
appropriate to an intimate and relaxed relationship. There is still a high
 2.5 Integrating higher- and lower-order features + 41

degree of deixis (this, that, the, down, etc.), action-oriented discourse

markers (oh right, oh yeah), and subject ellipsis in (It) looks cleaner again and
(I) only cleaned it, as well as subject- verb contractions in quantity. There is
a directness that is unproblematic in the intimate relationship (in the
sound recording the you’s are stressed, as: well YOU take that, YOU give
yourself a hernia). Long silences are also unproblematic (e.g. 14 seconds).
(2.8) and (2.9) thus do seem to have much in common, enabling us to talk
of generic commonality in which activity, convergence towards trans-
actional and relational goals, along with the differences in the relation-
ships pertaining among participants, can be seen reflected in local
linguistic choice. The further example of ‘language-in-action’ given in
section 6.2 only serves to underline this view. ’
If we shift to a different goal-sub-type within the general heading of
‘collaborative tasks’ (see section 1.4), decision-making for future action
(i.e. not the simultaneous tie between language and action of the
previous two extracts), we see a different pattern of features which again
varies internally depending on relation-type. In (2.10), a married couple
are at home, planning their next holiday in the presence of two close
family relatives who are house-guests; the relationship is intimate.

(2.10)
[<S 01> the husband <S 02> the wife <S 03> female house guest]

<S 01> Mm that’s interesting [<S 01> what] there’s two places offering
deals at the same hotel let’s compare them.
(6 secs)
<S 02> Bit dear.
<S 03> Yeah one’s by boat and one’s by plane.
<S 01> Oh I see.
(7 secs)
<S 01> It’s even the same library shot they’ve got of the hotel.
<S 02> Why don’t you have a look and see if we should take the car or go
on the train have look at The Rough Guide and see.
<S 01> LI think we should go by
train.
<S 02> Do you think, take the car to Felixstowe and go over on the boat.
<S 01> Yeah we haven’t got to worry about parking then and finding our
way back.
42 + 2 Spoken language and the notion of genre

<S 02> Well have a look at The Rough Guide and see what it says about
train travel.
<S 01> I don’t want to at the moment I’m too tired.
<S 02> Oh come on do it.
<§ 03> I’ll do it I will plan your holiday [laughter].
<S 01> You two managers get together and leave me and Dave to have
sleeps.
<S 03> Alright where’s The Rough Guide ... erm.
<S 01> This is the same have you got two the same or something oh
you're there.
<S 03> Where’s The Rough Guide then.
<S 02> It’s in that bag you're sitting on it.
<S 03> You great elephant.
<S 01> I couldn’t feel it cos my bottom is so enormous [laughter].
<S 02> Yeah we'll just leave the car behind and go on the bus, go on the
trains.
<S 01> I reckon that’s what we should do [<S 02> yeah] the only problem
that we’ve got then is carrying luggage.
<S 02> Yeah I won’t take any.
<S 03> Just take a ruckie.
<S 02> Just take one, just take a Sainsbury’s bag.
<S 01> Well this is what we usually do I haven’t got a rucksack.
<S 02> Well at least take, we’ve got the cool bag so we'll take that and
just so we can keep, it’s only small so we can take things in that and
we can get the train down to Brugge and then cos it’s only a wee bit
south then take the train back up and go into Delft and up to
Amsterdam and there’s somewhere else nice on the way go up to
Amsterdam and then just get it back again is Delft in Holland or
Belgium?
<S 03> Delft Holland.
<S 01> Holland.
(5 secs)
<S 02> And it says in my book the train’s quite cheap.
<S 01> The only problem I suppose if we do go by train is the hassle of
finding the right buses and coaches and.
<S 02> Nah cos the train stations’ll be in the cities won’t they it’ll be fine.

Here, at the higher-order level, the goal is a series of decisions. These are
dealt with on a problem-by-problem basis, each problem converging to a
 2.5 Integrating higher- and lower-order features - 43

solution accepted by the participants and enabling decisions to be made.

Hoey (1983) describes ‘problem-solution’ patterns in written discourse in
terms of a regular sequence of discourse segments moving from the
situating of a problem, to responses to it, to evaluations of those
responses, and thence to a solution. Much the same is apparent in
problem-solving in spoken discourses. In this extract, shared knowledge
is high in terms of the overall goals but low in terms of <S 01>’s
possession of travel brochures, which only he can see at this point. The
atmosphere is relaxed and informal. Silences are acceptable, it would
seem, only when the brochures are being consulted (a language-in-action
phase); talk is otherwise continuous. The lexical density is 35.5%,
reflecting the higher number of necessary lexical references to places
and activities beyond the immediate situation. Deictic references to the
immediate environment are only high in number towards the end of the
extract, where the speakers are trying to find The Rough Guide travel book
(this, there, that), which is another mini-language-in-action episode. The is
frequent, referring to shared knowledge in the speakers’ world (the car,
the boat, The Rough Guide, etc.). Suggestions by individuals are an impor-
tant part of the goal-orientation, and are realised both in formulaic and
more indirect ways (Mm that’s interesting ... let’s ...; Why don’t you ...; I
think we should ...; Do you think ..., and imperative verb forms). The
formulation at the end of the extract projects convergence (Yeah we'll just
...5 T reckon that’s what we should do).
In comparison, let us look at another decision-making collaborative
task extract, this time in a publisher’s group meeting. The context-type is
professional, but informal. The speakers have a number of publication
plans for books on the table and have to decide key dates and actions to
ensure their efficient publication. Once again, problem-solution episodes
make up the higher-order framework. Shared knowledge is high in terms
of the goals (such meetings are a regular event), but low in terms of each
individual’s expertise and current perspective on the problems. The
extract (2.11) begins as the group have been discussing whether to
reprint a particular title and have made a number of decisions:

(2.11)
<S 02> That first six month’s going to be a killer ... not to worry erm any
other questions.
<S 01> No that’s all.
<S 02> Well I’ve got one [<S 01> yes] and that’s about the readers [<S 01>
44 + 2 Spoken language and the notion of genre

yes] can you just fill me in again [<S 01> mm] just very quickly
[<S 01> mm] how many and when are they likely to hit me.
<S 01> How many books [<S 02> yeah] how many titles.
<S 02> Was it sixty did you say.
<S 01> Erm we were talking about, well the adult series will be six levels
and ... er thirty initially.
<S 02> Right and six months after that another thirty [<S 01> yeah] and
that’s likely to hit me in a year.
(4 secs)
<S 01> Erm.
(3 secs)
<S 01> Yes ... year to eighteen months.
<S 02> Right that, and to what degree I mean that means that they'll
come into production thirty titles will come into production or er
that that will be the beginning of looking at an identity and pre-
planning.
<S 01> No I suspect they will actually be going into production in about
say about summer next year.
<S 02> Right.
<S 01> Lit’s what I would aim for.
<S 02> Right well then what we need to do is erm er sit down together
and have a planning meeting [<S 01> mm] I think again [<S 01> mm]
just go over it I know you've already spoken to me about it but er I'd
like to just go over it again and think what the issues are and see
erm ... I’m pretty confident about the extents which I gave you but
I'd like to have a look at the production issues involved and who’s
going to actually do the setting.
<S 01> Mm.

Silence is again tolerated when participants attend to the physical

environment (in this case the papers in front of them), but otherwise,
talk is continuous. Turns are longer than in the other extracts we have
been considering in this section. Lexical density is 35.7%, again reflecting
the necessary lexical references to entities beyond the immediate en-
vironment, but still firmly on the ‘spoken’ side of the overall textual
average. Although not officially chaired, the meeting is steered in phases
by individuals (e.g. <S 02> in the opening turns). Metalanguage con-
cerning meetings and decision processes is apparent in formulations
(again the opening turns, also That will be the beginning of looking at an
 2.5 Integrating higher- and lower-order features + 45

identity and pre-planning; What we need to do is erm er sit down together and
have a planning meeting I think again just go over it). Discourse markers
indicate decision phases and topic shifts (right, well). Spatial deixis (items
such as here, over there) is non-apparent. Participants signal their satisfac-
tion with (or worries over) the ongoing interaction (not to worry), and
suggestions are signalled collaboratively (It’s what I would aim for; Well then
what we need to do is ...). Politeness and indirectness are used (Can you just
fill me in again just very quickly; Erm we were talking about ...; No I suspect ...).
We have compared four extracts and seen that they fall naturally into
two pairs on the basis of not only activity type but in terms of their
lexico-grammatical features. They do, of course, vary within the two
pairs, but we can say that more unites them than divides them on the
parameters considered, despite their different settings and relation
types. We could broadly indicate the generic similarities thus:

high deixis low deixis

high shared =™ punchbowl directness

knowledge T
it @ packing to go away

@ planning a holiday

i w publisher's meeting {)
low shared indirectness
knowledge

low lexical. —§_

AAA _ high lexical
density density

Figure 2: Generic patterning of four extracts (2.8, 2.9, 2.10, 2.11)

Figure 2 is, to say the least, an idealised version of what are quite
complex differences, but it is intended to show how spoken texts may be
usefully positioned with regard to their similarities along just a few
dimensions. Other dimensions could also be brought into play. Plotting
texts in this fashion enables us better to capture the variation present in
texts which share similarities in generic-level activities, and enables us to
make at least some links between higher-order concerns and the basic
46 + 2 Spoken language and the notion of genre

lexico-grammatical choices which speakers make in line with their goals

and relationships in particular settings. The gradability of generic
features thus displayed is in line with Wikberg’s (1992) advocacy of the
recognition of gradience and variability in the classification of written
genres, which are often subject to over-simplified text typologies. As
Biber and Finegan (1991) advocate, genre study should include both a
characterisation of typical texts as well as a characterisation of the range
of variation. Within the narrow scope of this section, I have suggested
some directions in which a genre-based methodology may go in the
search for an adequate framework of description for spoken genres.

2.6 Conclusion

In this chapter I have attempted to draw a tentative theory of spoken

genre in light of the variation which is acknowledged to be present in
speech events which seem, nonetheless, to share common features. I have
tried to account for participants’ awareness of generic activity partly in
terms of their orientation towards past events, upcoming ones, the need
for formulation of ongoing activity and the need to instantiate those
procedures whereby events take on shapes and characteristics that
enable linguists to categorise them into different genres. I have used the
CANCODE corpus matrix to demonstrate how extracts controlled for
variables such as goal-type and context-type can be seen to display
similarities at the lexico-grammatical level which fit in with the higher-
order features of generically-oriented activity. Genre will always remain
a slippery notion precisely because human activity is open to so much
variation, driven by the common-sense purposes of interactants and the
need to build and maintain human relations. In all the extracts used in
this chapter, relational work is high on the agenda for participants, and
it is only rarely separable from the transactional business at hand,
however institutionalised that business may be. What is apparent is that
at all levels, behaviour is integrated: the transactional, the interactional,
the goal-orientation, the relationships among participants, the local
lexico-grammatical selections, and that all we can hope to do as analysts
is occasionally to tease out those different exigencies and put them
under the spotlight. This chapter and Chapter 1 have attempted to lay
the foundations for a pedagogically-driven theory of spoken language in
applied linguistics. That theory may be said so far to be founded on the
following principles:
 2.6 Conclusion + 47

Spoken language pedagogy cannot simply just proceed from the same
assumptions as written language pedagogy with regard to language
use. The historical dominance of the written language in applied
linguistics has militated against good understanding of the spoken
language.
Spoken language has its own grammar and lexicon, which, although
coinciding in most cases with the written language, differ in crucial
areas that correspond to the goals and relations of interactants in
particular settings.
We can only begin to describe the special lexico-grammar of talk by
looking at real data.
The best data for a pedagogical theory of spoken language is everyday,
informal talk.
The text-typologies often used in the description of written language
should not be simply transferred to spoken language. We must first
research what the differences are in different types of spoken events
and come to understand spoken genres.
Central to a theory of spoken genre are participant goals and relation-
ships.
A corpus of spoken language for pedagogical purposes will be designed
with goal- and context-variation in mind, and will include goals and
contexts relevant to language learners (e.g. pedagogical relationships,
service transaction goals). These design features will be more important
than mere size of the corpus.
The corpus needs qualitative as well as quantitative analysis to be
pedagogically useful.
Observation of spoken language use by participants in real encounters
of various kinds is the first step in building language syllabuses and
teaching materials.

Against this background, the next chapter will ask and try to offer answers
to more detailed and practical questions about language pedagogy.

Notes

1 See Hasan (1992) for a critique of the ambivalence of some of Bakhtin’s

categories. Hasan is right to criticise Bakhtin’s work as being difficult to
operationalise. However, as long as one does not regard Bakhtin’s ideas as a
model or instrument of analysis, but rather as a thought-provoking set of
48 + 2 Spoken language and the notion of genre

theories, their value in assisting our understanding of and our ability to

construct the nature of the spoken language remains intact.
N Young (1990: 75ff), in her study of written and spoken academic language,
refers to the ‘generic situation’ that influences selections of field, tenor and
mode (in the sense of those terms as laid out by Halliday 1978), rather than to
‘genres’ as such. Young is attempting to make a separation between the
linguistic ‘code’ (where field, tenor and mode are represented by lexico-
grammatical selections) and language as ‘behaviour’, which is a social/contex-
tual level, where the generic situation is located.
i) Bakhtin (1986: 89) talks of the multiple voicing of discourse: ‘Our speech ... is
filled with others’ words’; all language, for Bakhtin contains echoes of previous
discourses: any speaker ‘is not after all the first speaker, the one who disturbs
the eternal silence of the universe’ (ibid.: 69), and each utterance is ‘filled with
echoes and reverberations of other utterances’ (ibid.: 91).
3
What should we teach about the
spoken language?!

3.1 Introduction

Chapter 1 outlined how the status of the spoken language in applied

linguistics has developed over the centuries and how- it has grown in
importance in recent decades with corpus projects coming to the
forefront. Chapter 2 looked at a number of studies that point towards a
model for spoken genres. Nowadays, a vast amount of descriptive
literature is available on spoken language, for English and for other
widely-used languages. Sociolinguists, discourse analysts and conver-
sation analysts, coming at spoken language from their several perspec-
tives, all offer the interested language teacher something to get excited
about, and often present findings that seem to be (or which seem as
though they ought to be) relevant to second or foreign language
teaching. Chapter 2 abounds in such references. Further studies of this
kind are reviewed and discussed in McCarthy (1991), McCarthy and
‘Carter (1994), Schiffrin (1994), and Stenstr6m (1994). These studies are
many and varied in their general characteristics, ranging from major
attempts to model the spoken language in terms of structure (e.g.
Sinclair and Coulthard 1975), prescriptions of socio-cultural norms to
which conversational participants orient themselves, such as turn-taking
(e.g. Sacks et al 1974), prescriptions of generic features that define whole
spoken episodes such as narratives (e.g. Labov 1972), descriptions of
complex surface manifestations such as discourse-marking (e.g. Schiffrin
1987), and a host of other concerns. We have seen, in Chapter 2, how
these and other resources combine to characterise the genres we are
familiar with as language-users; on the practical level, we need to
examine each type of linguistic resource in detail, in order to be able to
relate them to our higher-order concerns such as goal-orientation and
relationships. In later chapters, we shall examine lexico-grammatical
resources of particular kinds; in this chapter we concentrate more on
discourse-level features. For this reason, it may be helpful from the

49
50 + 3 What should we teach about the spoken language?

outset in this chapter to divide the somewhat amorphic world of spoken

language studies into very broad types, and to discuss each type in terms
of its possible implications for teaching. By ‘teaching’ here I mean
intervention in the learning/acquisition process in the classroom and
designed input in the form of syllabuses, materials and methods. The
broad classification which follows is essentially synthetic and pedagogy-
driven, since this chapter is concerned with examining the spoken
language from the point of view of creating what Widdowson (1980) calls
‘relevant models’ of language description.

3.2 The spoken language: key descriptive areas

3.2.1 Structural features

Three structural units fundamental to all spoken interaction emerge

from a wide range of studies in discourse and conversational analysis: the
transaction, the exchange and the adjacency pair.”

The transaction

The term transaction is here used broadly in the sense that Sinclair and
Coulthard (1975) use it, to label stretches of talk identified by certain
types of activity at their boundaries. For example, in the classroom,
teachers will typically divide the business of a lesson up by marking the
transitions to new phases with some sort of conventional marker (right,
okay, now,? so, etc.) characteristically uttered with falling intonation and
often followed by a pause or ‘silent stress’. Around these markers,
metalinguistic activity may also take place, for example in a phone-call:
‘Okay, well, that was the main thing I was ringing about, but there was
one other thing’ (attested). The transaction, like the paragraph in written
language, has no pre-defined length, and is only recognisable by its
boundaries.
It is hard to imagine talk proceeding efficiently without participants
signalling in some way or other and recognising such boundaries, and
the transaction, as a structure, is probably a discourse universal. What
we can say as language teachers, however, and this statement will apply
equally to a number of other features, is that the transaction as a unit of
discourse may present us with a problem on two distinct levels. On the
first, there may be a problem of awareness, among both teachers and
learners, that transaction signalling is an important part of behaving
 3.2 The spoken language: key descriptive areas + 51

linguistically in the target language (for example, there is evidence that

in certain types of more formal talk, lack of metalinguistic signalling
can affect comprehensibility; see Williams 1992). On the second level,
the problem is principally a lexical one: how does the target language
realise such marking (compare English well with Spanish bueno and pues,
French bien, and Swedish nd)? Are the L2 literal lexical equivalents for
items like good! and now then! also those used for marking purposes?*
Bilingual dictionaries are notoriously poor sources for such cross-
linguistic information, and the teacher may best be able to attack both
the awareness problem and the lexical problem via observations of real
data, a point to which we shall return. But even if the awareness
problems and the lexical problems can be resolved, there still remain
the problems of generating classroom activities that offer the learner
opportunities for transaction-marking in naturalistic settings, which the
teacher-fronted classroom is unlikely to be able to do. In the teacher-
fronted classroom it is usually the teacher, and the teacher alone, who
marks the transactions. I have elsewhere presented limited data to show
that task-based contexts where students carry out group activities can
yield natural use of boundary marking by participants when the teacher
is no longer dominant (McCarthy 1991: 131), but such manifestations of
ideal and natural discourse are not guaranteed, and the many factors
bearing upon task success will have to be taken into account (see Skehan
1996).
I have focused on the question of the transaction and its ‘teachability’
because, in many ways, it sums up the questions we may ask of all the
discourse features which are examined in turn below, and these are:

1 Are discourse features automatically transferred from L1 behaviour to

|
4
2 Can such transfer occur effectively without lexico-grammatical input
(in other words, are things like markers usually just ‘picked up’ along
the way or do they need to be taught)?
3 Is teacher/learner awareness of the spoken language sufficient to pre-
empt and address the possible problems raised by questions 1 and 2?
4 Is communication in L2 without the performance of features such as
transaction-marking adequate, satisfactory and satisfying for learners?

A related question, one that Scarcella (1983) takes up, is whether there
are developmental factors at play in the learner’s ability to be discours-
ally competent in the target language. Scarcella is concerned with
52 +» 3 What should we teach about the spoken language?

whether features such as conversational strategies increase with general

language proficiency and whether some strategic areas are likely to be
acquired before others, offering a qualified yes to both of these questions.
The same questions could be asked about structural features such as
transaction-marking, and it is likely that a qualified yes is the answer
there, too, though little hard evidence exists.
Questions 1 and 2 above are linked to each other. Most features of
linguistic behaviour do, sooner or later, carry over from one language to
another, given sufficient exposure and motivation, and considerable
amounts of language are just ‘picked up’ within and without the
classroom, as all experienced language teachers know. But it would be
an odd pedagogy that refused to teach, say, the past tense, or the
vocabulary of food and drink because these would be ‘picked up anyway’
sooner or later. What is important is to identify those features which are
natural in L1 performance and desirable in the learner’s L2 performance,
to offer short-cuts to the necessary lexico-grammatical knowledge to
realise such features, and, above all, to enable the appropriate contextual
environments to be created in the classroom and in the teaching
materials so that it may be observed whether or not such behaviour is in
fact transferred and is part of the learner’s repertoire. Question 3 under-
lies the motivation of this entire book. I do not believe that a teacher
who is unaware and a bad observer of the spoken language can really
give the kind of input necessary to raise the learner’s awareness of the
key, defining features of talk. I do firmly believe that learners trained to
be good observers of data have taken an important step towards
facilitating uptake of those very features, however long-term or delayed
the actual output of such features may be in terms of actual use.
Question 4 above surfaces again later when we consider discourse-
marking in general.

The exchange

Similar problems to those raised for transactions apply to the notion of

the exchange, although there are specific features that make it worth
discussing separately from the transaction. The exchange is the minimal
structural unit of interaction, consisting of an initiation and a response
(for example, a question and its answer, or a greeting and a return
greeting). But this minimal condition is typically elaborated in casual
conversation to include a third function, the follow-up, and is in fact
 3.2 The spoken language: key descriptive areas - 53

often realised in quite complex configurations (see Hoey 1991b). A

canonical example of the structure Initiation-Response-Follow-up (IRF) is
this CANCODE extract:°

(3.1)
[<S 01> is asking her great-niece about a forthcoming trip to London]

<S 01> What part of London would you be actually in? I

<S 02> Well I would be going from Paddington to Victoria. R
<S 01> I see yes. F

However, more complex sequences (for example, where checks are felt
necessary) can sometimes mean that the follow-up is delayed, though
still present:

(3.2)
<S 01> What time is it?
<S 02> Twenty to six.
<S 01> Is that all? S
<S 02> Yeah.
<S 01> OhI thought it was later. “www

Learner repertoires often range from only performing the Response

function (especially in early stages or in the traditional, teacher-fronted
classroom, where the teacher commands both IJ and F slots), to Initiating
but still not making any follow-up. This latter behaviour is sometimes
noticeable in classroom pair-work where the teacher is often hovering
and monitoring performance (but see Kehe and Kehe 1989 for how the
teacher may respond constructively to such difficulties). The follow-up
very frequently has a relational/interactional function, where social,
cultural and affective meanings are encoded in relation to responses, in
addition to acknowledging the response and its information, and where
key conversational processes such as convergence are effected. It is thus a
crucial structural element in our global, generic framework, where
relational features are given the same priority as transactional ones, as
we advocated in Chapter 2. Making it part of the learner’s repertoire is
therefore very important. The reactions encoded in the follow-up are
often formulaic, and can be viewed as a lexical problem across languages
(compare the British English reactives Really!, Oh, right, That’s nice!, You
don’t say!, I guessed as much! and How interesting/awful! with their realisa-
54 + 3 What should we teach about the spoken language?

tions in other languages and/or other varieties of English). But as before,

solving the lexical problem alone may not be sufficient to foster natural
interaction complete with the follow-up function. Contextual conditions
for its natural occurrence must also be assured, and where these are non-
existent, the learner will never have the proper opportunities to practise
this basic, core function of the spoken language. The awareness problem
in this case is usually related to learner expectations that it will be the
teacher who follows up, normally by evaluating the linguistic quality of
the learner’s response (whether a response to a teacher-initiation or to an
initiation by a fellow-student in pair- or group-work). The learner needs
to be made aware that follow-up is not just for teachers, and not just for
evaluating correct or incorrect performance. Textbooks with dialogues
that do not have follow-ups certainly do not help in this task, and where
natural data is inaccessible or unavailable, the next best thing may be
the editing of textbook dialogues to ensure the inclusion of at least some
follow-ups. To be a good editor, though, presupposes an appropriate
awareness of when and where and how follow-ups are used. Careful
observation and/or access to a corpus gives a great advantage in this
respect.

The adjacency pair

The adjacency pair is a unit associated with American conversation

analysis (e.g. Schegloff and Sacks 1973), usually conducted from an
ethnomethodological standpoint, typically concerned with how partici-
pants behave in interaction in terms of alignment (i.e. how they position
themselves socially in relation to their interlocutor(s)), achieving goals,
negotiating outcomes, etc. Speakers naturally orientate themselves to
bring together in the discourse utterances that mutually condition one
another, such that the sequence in (3.3) is considered to be in line with
participants’ expectations:

(3.3)
[at a dinner-table, <S 02> is the guest]:

<S 01> D’you want some olive oil Dennis?

<S 02> Mm ta.

In (3.4), later in the same conversation, a remark by <S 05> is responded

to by his wife, <S 04>, but the response is problematic:
 3.2 The spoken language: key descriptive areas - 55

(3.4)
<S 05> The nicest pizza I’ve ever had was in Amsterdam [<S 03> Oh yeah]
Thad a brilliant pizza.
<S 04> lin Cyprus.

The incoming information has not prompted convergence, but is openly

challenged, and it will require work on the part of the speakers to resolve
the possible hold-up in the talk. The first part of the adjacency pair
predicts the occurrence of a second part, and the second part is seen to
fulfil that prediction in some way. In the case of (3.4), we might expect a
response showing interest and encouraging <S 05> to go on and describe
the wonderful pizza. Instead the unpredicted happens: the information is
contradicted. The adjacency pair, in its concern with local decisions by
speakers, does overlap with the notion of exchange; the difference is that
the exchange is primarily seen as a structural unit building up into
higher order units, while adjacency is concerned more with local conver-
gence between participants. Another way of putting it is that discourse
analysts working with the exchange are much more interested in the
presence of the pattern as a trace in the text for the analyst’s purposes,
while conversation analysts who work with adjacency are trying to under-
stand local, individual choices from the participants’ viewpoint.
There is no evidence to suggest that learners do not orientate them-
selves to create appropriate adjacency, however imperfectly they may
‘actually realise it, and thus adjacency, as a global notion, is interesting
but may not be an essential component of a relevant linguistic model for
pedagogy, and probably does not need to be ‘taught’. However, once
again, a number of adjacency pairs are highly formulaic and can be
treated as an aspect of the lexico-grammatical content of the syllabus.
One may not need to be taught to ‘be adjacent’, but one may well benefit
from learning a number of ready-made formulae (what Nattinger and
DeCarrico 1992 and Lewis 1993: 94 refer to as ‘lexical phrases’ and
‘institutionalised expressions’, respectively) which will enable fluent,
natural and culturally and pragmatically appropriate adjacency pairs to
be realised. Examples that readily spring to mind are reactions of
condolence, congratulation-sequences, seasonal greetings, telephone
opening conventions, phatic exchanges, etc. In a related way, there is the
problem of ‘dispreferred’ second parts of adjacency pairs (Pomerantz
1984), that is to say responses that do not fit in with expectations, as in
the wife’s challenge over the pizza in (3.4). For example, how does one
56 + 3 What should we teach about the spoken language?

disagree with an assertion (see the discussion in section 1.8 on ‘dis-

agreeing’) or refuse an invitation without causing offence or making the
receiver feel threatened? What tends to happen in native-speaker speech
is considerable elaboration of the second pair-part (the second part of the
adjacency pair) to include reasons for the divergence, and often a preface
to the actual ‘dispreferred’ utterance. These conditions, of course, apply
to generally co-operative discourse; people will always exercise their
right to be conflictual and rude when they choose to be, or at least fairly
direct, as <S 04> feels she can be to her husband in example (3.4). Two
further corpus examples illustrate typical co-operative behaviour, the
second more direct than the first:

(3.5)
[A young daughter, <S 01>, is being helpful and offering to make
everyone toast. Most family members accept two slices. She then
addresses her father.]

<S 01> Dad, one piece or two?

<S 02> One’ll do for me, Jen, if you
<S 01> LRight, okay.
<S 02> Cos I’ve gotta go in the bath in a minute, love.

(3.6)
[<S 01> addresses his sister-in-law, <S 02>, and tells her he never realised
she made her own Swiss rolls.|

<S 01> And I’ve never realised that you’ve made it, I thought this was
probably, I probably thought it was bought.
<S 02> Oh, you're joking! It’s our speciality of the house!

The dispreferred utterances are here indicated in bold. Clearly, speakers

generally wish to avoid over-blunt refusals, divergences and contradic-
tions, and, once again, considerable lexical effort is expended in elabor-
ating the dispreferred response. In (3.5), there is a risk that the daughter
will interpret her father’s choice of only one piece of toast as a snub to
her efforts to help in the kitchen. We have what looks like an aborted
polite conditional and a reason from the father. In (3.6) we have a
conventional informal preface of disbelief that softens the confirmation
that the first speaker was saying something wildly wrong. Here discourse
and culture overlap, and notions of politeness, threats to face (Brown and
 3.2 The spoken language: key descriptive areas + 57

Levinson 1987) and acceptable behaviour in different cultures come to

the surface, for example, customs connected with the acceptance or
refusal of food, cultural expectations as to whether it will be home-made
or shop-bought, etc. Once again, though, it is questionable whether one
can effectively ‘teach’ L2 culture via a presentational methodology, or
whether language and cultural awareness activities may be preferable,
where the goal is to observe, discuss and come to understand features of
interaction rather than to ‘learn’ or imitate them. Indeed, there are
several arguments for advocating such an approach. One is that current
models of language and culture have moved away from the notion that
culture is simply ‘there’ in language (the ‘colliding cultures’ view), and
more towards interactional models, where culture is seen as context- and
situation-dependent and as something negotiated among interlocutors
(Blommaert 1991). The other argument is a pedagogical one: learners
probably gain better awareness from exploratory and problem-solving
encounters with real data rather than having it interpreted for them and
(re-)presented by teachers (see Jackson 1990 for an account of students’
own language analyses and a good discussion of this issue).

3.2.2 Interactional features

Under this heading we include those areas of linguistic choice where

speakers manage the interaction in less obviously structured ways (in the
’ sense of regular patterned occurrences of restricted choices), in order to
proceed effectively towards their goals whilst maintaining relational
equilibrium. Amongst many such features, turn-taking, discourse-
marking and information-staging are central, and present similar ques-
tions as to their ‘teachability’ as did the more overtly structural features
in section 3.2.1.

Turn-taking

Turn-taking, although a universal feature, may cause problems on the

cultural plane. Indeed, one might argue that this is one area where
English language teaching materials have erred somewhat on the side of
seeing the issue as primarily a lexical one. Three problems seem fre-
quently to arise. One is that some cultures seem to tolerate longer
silences between turns, for example as is often perceived by other
Europeans during conversations in Finnish. Another such example is the
58 + 3 What should we teach about the spoken language?

rule-conflicts that can cause conversational tensions between ‘silence-

filling’ discourse styles such as that found in American English, and
cultures where silence is permissible thinking-time and face-protecting,
such as Japanese (Noguchi 1987; Lebra 1987). A second problem is related
to acceptable forms of interruption across cultures, for example the
preference for direct interruption with markers such as ma (but) in
Italian compared with the indirect use of well and/or agreement contrasts
in English (Testa 1988). The third problem is that of differing styles of
‘back-channel’ (the noises and verbalisations made by listeners to show
understanding, continued interest, etc.) (Yngve 1970). Spanish speakers
often acknowledge incoming talk with what translates to English as a
machine-gun-like ‘Yes-yes-yes!’, and which frequently indicates im-
patience or irritation with the speaker in British English. There seems
again, here, to be a mix of lexical and cultural problems which need
addressing in the classroom. And yet turn-taking (and especially interrup-
tion) is one area in English language teaching where there is no shortage
of lexical advice. Course-books frequently offer conventional phrases for
interrupting, such as Sorry to interrupt and Can I just say something?. Useful
though such phrases may be, they are no substitute for close observation
of data. When one does this using data from informal conversational
settings (and even in some semi-formal contexts) one finds, in the case of
English, that an interruption often simply occurs and is then followed or
is itself broken off by the interruption marker, as in (3.7) and (3.8):

(3.7)
[<S 02> interrupts a conversation in a corridor between two women]

<S 01> with Carol and he doesn’t even realise

<S 02> lHelen, sorry to interrupt, erm did
you manage to ring Patrick?
(attested)

(3.8)
[<S 02> interrupts a colleague during a semi-formal discussion]

<S 01> well it would have to go to a later meeting

<S 02> LI don’t ... for this document,
sorry to interrupt, but I don’t think you can ignore the importance
of [etc.]
(attested)
 3.2 The spoken language: key descriptive areas - 59

In (3.8) the interruption has in fact already taken place before the marker
is used. Once again, it is the combination of lexical items and culturally
acceptable behaviour which should be of interest to the teacher and
learner, and, as before, observation and awareness of what happens in
real data must be added to the act of providing the lexical resources. The
lexical resources alone are simply insufficient without culturally sensi-
tive insights from direct observation. In terms of our over-arching genre-
model, interruption may be more permissible in some genres than
others. Overlap and interruption in language-in-action collaborative
tasks among intimates may be seen as practical and goal-facilitating; in
discussion and argument genres in professional or socialising settings, it
may be seen as aggressive or downright rude.

Discourse marking

Discourse marking has already been referred to in connection with

transaction-boundary marking, but discourse markers are widely used to
signal many different functions in conversation (e.g. well, mentioned
above as a marker of disagreement or divergence) (see Schiffrin 1987;
Fraser 1990). All languages seem to have a finite set of lexical items
(single-word and phrasal) to signal functions such as shared knowledge
(e.g. English you know), proclaimed knowledge (English you see), topic
shifting (English but and so), pre-closings and returns to interrupted
topics (English anyway), boundary marking, and many other interactive
and structural functions. What is more, hardly any stretch of informal
conversation is without markers: extract (3.9) is by no means untypical of
natural conversational data:

(3.9)
[<S 01> is explaining a consumer report scheme her mother got roped
into.]

<S 01> She gets a pound or something, you know, a month, but it was
something that, I remember I was a kid, and, well, sort of, about
sixteen seventeen or something, and this woman came to the door
and erm I agreed to it [laughs] and my mother kept, you know, my
mother did it and she kept it on, you know, for about the last
twenty years doing this.

Watts (1989) has shown just how unconscious native speakers can be of
60 + 3 What should we teach about the spoken language?

their own use of discourse markers, and markers do seem to display an

automaticity characteristic of the more routine aspects of speech. This
may be one of the reasons why they are so often absent from concocted
dialogues in language text-books, and indeed from many dictionaries.
Research evidence also suggests that where spoken discourse lacks a
normal distribution of markers, it can create problems of comprehension
as well as sounding unnatural (Tyler, Jeffries and Davies 1988).
There seems to be no obvious reason why the set of discourse markers
for any language should not be part of the most basic lexical input in the
syllabus and materials, for they are indeed very useful items and,
lexically, usually quite simple and straightforward and often familiar to
learners from their basic semantic meanings (e.g. good may already be
known as the opposite of bad). Of all the features considered in this
chapter, discourse markers, on the face of it, lend themselves most easily
to a presentational methodology.® However, a note of caution must be
added in that much work remains to be done in actually establishing
how and when native speakers use markers. Aston (1995), for example,
has shown that English thank you has an important phase-marking
function in service encounters over and above its meaning of ‘expressing
gratitude’. It may thus be premature to teach the set of markers as
lexical formulae. As always, though, one would hope that raising aware-
ness of their widespread role in spoken language through discussion and
exemplification might proceed in tandem with learning them as items,
and observation of natural data, as much as word-learning, is again
desirable wherever feasible. Furthermore, the fact that a presentational
methodology might work for inputting discourse markers is no guaran-
tee that they can be successfully produced in pair- or group-work, and it
may be better simply to allow production to be delayed until suitable
natural opportunities arise.

Information staging

Information staging brings us to some interesting observations on the

grammatical level, as opposed to the lexically- and culturally-oriented
interpretations offered above for some of the other discourse phe-
nomena. The canonical word-order for the clause in English of Subject-
Verb-Object-Adverbial is frequently manipulated in informal speech to
produce a variety of re-arrangements of the information encoded. These
rearrangements enable foregrounding of entities, signalling of topically-
 3.2 The spoken language: key descriptive areas - 61

prominent items and interactive features such as tags and tails (see
glossary). As well as front-placed objects for foregrounding or contrast,
we find items placed before the core clause elements and after them. Two
examples follow. (3.10) shows how a pre-posed item copied in the main
clause can supply useful information to anchor a topic in the listener’s
consciousness (see Geluykens 1992 for further examples; see also Carter
and McCarthy 1995b). (3.11) shows a typical interactive function, the end-
copying of items singled out for evaluation (see Aijmer 1989; McCarthy
and Carter 1997b):

(3.10)
<S 01> Well, Karen, where I’m living, a friend of mine, she’s got her
railcard and | was telling her...

(3.11)
<S 01> It’s very nice that road up through Skipton to the Dales.

All evidence suggests that word-order phenomena (including cleft struc-

tures) and information-structuring of this kind exist in many languages
(e.g. for French, see Lambrecht 1988; Italian, Duranti and Ochs 1979;
Swedish and Yiddish, Kallgren and Prince 1989; Japanese, Ono and Suzuki
1992; Spanish, Guitart 1989). It is interesting to ponder whether features
of this kind are automatically carried over from L1 to L2 (e.g. see Trevise
1986), and precise grammatical realisations may vary across languages.
‘The grammatical inventories of language teaching syllabuses often
ignore these phenomena, as do concocted text-book dialogues. This would
certainly seem to suggest a revision of the grammatical categories which
need to be addressed if a course is adequately to cover conversational
phenomena. A significant level of linguistic encoding of relational func-
tions is lost if the learner has neither the opportunity, nor the gramma-
tical apparatus, nor the confidence to transfer such features from L1, to
realise the kinds of functions we have illustrated. A revision of the
grammatical component of the syllabus in this way will probably mean
far greater emphasis on front- and end-placing in the clause and manipul-
ability of the core elements than tends to happen at present. Such an
admission of previously ignored grammatical phenomena is not without
its problems, though, not least of which is that syllabuses above all, and
teaching materials in the main, tend overwhelmingly to be written, and
one is faced with the conundrum of raising awareness of spoken forms
through written-down examples and exercises. Another related problem
62 + 3 What should we teach about the spoken language?

is the power of the written sentence to affect our view of the nature of the
grammatical information. It is no mere chance that linguists who do
tackle phenomena such as topicalising and ‘tails’ speak of them in a
metalanguage that is ‘page-driven’, referring to ‘left-dislocated’ elements
(e.g. see Geluykens, 1992 on English, Blasco 1995 on French, Rivero 1980
on Spanish) or ‘right’ dislocation (e.g. Fretheim 1995; Valiouli 1991).
Clearly, spoken language has neither ‘left’ nor ‘right’; if anything, it has
‘before’ and ‘next’.”? Equally, metaphors of ‘dislocation’ have a tendency
to suggest that something is wrong or ‘out of place’, rather than perfectly
normal, acceptable and significant in conversational terms. This does
suggest that spoken grammar methodology must lean heavily towards
audiotaped support or at the very least adequate contextualisation to
enable teachers and learners to reconstruct the original utterances. It
also suggests that the spoken language pedagogy which emerges from
our overall genre-based concerns needs a carefully forged metalanguage
which the language teaching profession does not yet share.

3.2.3 Generic features

The notion of genre in spoken language as described in Chapter 2 tries to

capture the sense that participants have of involvement in particular
language events, which unfold in predictable and institutionalised ways
and move, stage by stage, towards a recognisable completion. Much work
exists in the characterisation of written genres (see Swales 1990), while in
the spoken language, the notion of genre has tended to be poorly
described (see Chapter 2). Obvious examples of well-circumscribed genres
are church sermons, wedding speeches, jokes, lectures, service en-
counters and stories. It is possible to define the elements or phases which
are usually present and to recognise these in their linguistic realisations
in different contexts (McCarthy and Carter 1994: 24ff).8 Genres are
typically more sharply defined at their openings and closings, for obvious
reasons, since speakers and listeners need to know just what sort of
language event they are involved in. Chapter 2 presented evidence for
speakers’ awareness of the expectations of upcoming activity and how
their recollections of previous experiences influenced their view of the
unfolding genre. Thus a joke may be conventionally signalled by some
marker such as Have you heard the one about ...?, while a true-life journal-
istic report of a terrorist bomb outrage would never be expected to begin
in this way. Genres are subject to embedding: for example, narrative
 3.2 The spoken language: key descriptive areas - 63

anecdotes may occur during a university lecture, and casual conversa-

tional episodes may occur during a service encounter. Genres, too, as we
have argued all along, depend as much on relational elements as they do
on purely transactional ones; the building and maintenance of relation-
ships simply cannot be separated from other business at hand.
Two principle problem areas seem to occur with regard to genres:
firstly, openings and closings are usually highly conventionalised and
secondly, learners may be more able to realise some elements better than
others. Oral anecdotes are a good example of this second problem, and I
have suggested elsewhere (McCarthy 1991: 137-142), with support from
data, that lower-level learners may often have great difficulty in realising
what Labov (1972) would call the ‘evaluation’ function:in narrative, that
is to say those lexical embellishments (e.g. intensifiers, hyperbole) and
grammatical configurations (e.g. I might easily have been killed; if only I
hadn’t been so stupid!) which give the story interest and ‘tellability’. In
Labov’s model, evaluative elements are just as important as the bare
bones of the narrative structure, which are ‘abstract’ (flagging the story
or giving a kind of ‘headline’), ‘orientation’ (telling the listener who,
where, when, etc.), ‘complicating events’ (the unusual, funny, scary, etc.
events that make the story a story), ‘resolution’ (how the events worked
out), and ‘coda’ (rounding off the story and bridging back to the present
moment). The challenge to provide adequate evaluation often applies to
listeners as well as tellers, for listeners, in natural settings, usually
‘contribute evaluation in some form, and the storytelling is rarely a one-
sided monologue, as Duranti (1991) has demonstrated with Italian data.
Helping learners to master evaluative techniques (e.g. expressing sur-
prise, horror, empathy, disgust, etc.) helps them to build better relation-
ships with their narrating interlocutors and to present a better self-
image as narrator. In the classroom, learners can be encouraged to re-tell
the same story, adding better evaluation each time (a not unnatural
activity, since in real life stories are usually told and re-told, with
embellishments built up over time). Re-telling of stories is a departure
from the usual ‘one shot only’ constraint of the typical oral classroom.
Openings and closings, I have suggested, are more likely to be formulaic,
but there are often interesting cultural differences, as, for instance,
between Greek and British English telephone calls (Sifianou 1989) , Dutch
and American English ones (Houtkoop-Steenstra 1991), or Chinese ‘and
English opening greetings in conversation (Hong 1985). Edmondson, *
House, Kasper and Stemmer (1984) contrast content! and interpersonal-
f
\
64 + 3 What should we teach about the spoken language?

orientation (i.e. the way language is sensitive to the interlocutor rather

than just to the content of the message) for conversational openings and
closings in German and English, as Jaworski (1990) does for native speakers
and American learners of Polish. We return, as always, to the need for
observation, awareness and sensitivity in confronting the characteristic
creation of genres in an L2, rather than any notion that such intercultural
phenomena can simply be ‘taught’ in a presentational sense. Fortunately,
more and more research is being conducted to produce accurate descrip-
tions of genres across cultures, and knowledge of cultural similarities and
differences can only be of positive assistance in language teaching and
learning. The potential of the CANCODE corpus for classifying texts along
generic lines (see Chapters 1 and 2) renders it a particularly useful tool for
work towards the creation of such pedagogical frameworks.

3.2.4 Contextual constraints

Since spoken language is more likely to be immediately context-bound

than written language, we find a separate set of concerns which emerge
from the way speakers encode contextual features. Situational ellipsis is
a good example of a typical spoken-language context-driven feature
(Carter and McCarthy 1995). In English, where entities can be retrieved
from the immediate situation, normally expected items of structure may
be omitted, as in these extracts:

(3.12)
[<S 01> is assembling things for her friend before they go out.]

<S 01> Handbag is it, what else then?

(3.13)
[<S 01> is commenting on the listener’s participation in a consumer
survey, which brought the listener the unexpected benefit of a new
telephone line, at the expense of the survey organisation]

<S 01> Put the phone in as well for you, did they?

(3.14)
[<S 01> is commenting in a friendly, joky way on the listener’s use of an
item in <S 01>’s flat]

<S 01> Think it’s your house or something?

 3.2 The spoken language: key descriptive areas - 65

In (3.12) your (handbag) is ellipted, in (3.13) we have a typical spoken

English structure of initial subject-dropping (they) which is later realised
in a tag, and in (3.14) we have the auxiliary verb and subject omitted (do
you). All the ellipted items are obvious in the context. This type of ellipsis
gets surprisingly scant attention in currently popular British pedagogical
grammars (one notable exception being Swan 1980/1995), whereas
CANCODE figures suggest one occurrence of this type of ellipsis per 300
words (approximately) of casual conversation (excluding narrative seg-
ments, where situational ellipsis is rare). Scarcella and Brunak (1981)
note an absence of this kind of ellipsis in their non-native-speaker data,
and it may well require direct intervention by the teacher to create
awareness of its existence and its contexts of use, as -well as a set of
guidelines for structural restrictions (e.g., in the case of the examples
above, characteristically first and second person, often with verbs of
mental process, etc.). Much more research is needed in this area, and
better descriptions of all the types of ellipsis that commonly occur in
conversation would be a great step forward. In the meantime, teachers
and learners have to rely on their own inductive abilities to derive rules
and guidelines, with as much help from real data as is possible.
Another context-bound feature of talk is its variable lexical density, as
the examples in section 2.5 and further examples in section 6.2 attempt
to demonstrate. Extremely context-bound types, such as ‘language-in-
action’ sequences, where language is generated by some task being
undertaken at the time (e.g. cooking, shifting furniture), may display
very low lexical density (i.e. a high proportion of function-words as
opposed to full vocabulary items; see section 2.5 for some examples of
variability across different genres). (3.15) shows sequences from data of a
family cooking food in their kitchen:

(3.15)
<S 01> What are you going to do with that?
<§ 02> Oh, it’ll go in in a minute. I can taste it as I go along and then add
the same amounts again.
(7 secs)
<S 02> Yeah, I’ll just give that a stir and see where we are first.
(6 secs)
<S 03> If you put this in the freezer, that'll cool it down quicker, won’t
it?
<S 01> Yes, and it won’t freeze.
66 + 3 What should we teach about the spoken language?

Notably, such sequences are usually high in demonstratives, pronouns,

and deictic words in general, and lower in content lexis. Such raw data,
by definition, can be difficult to interpret without considerable contex-
tual information, and, for this reason, concocted text-book dialogues
rarely occur in such a heavily context-bound form. However, I would
argue that interesting (and entertaining) work can be done in the
classroom with tantalising data of this kind precisely to raise awareness
of the fundamental importance of items such as deictics in oral com-
munication. Learners can be asked to imagine what might be going on,
and to offer lexical items as possible referents. Also worth noting in this
context is that deictic items do not necessarily translate one-to-one
across languages (e.g. demonstratives between English and French, or the
tripartite demonstrative system of Spanish) and they may need to be
focused upon in their natural contexts of occurrence. Language-in-action
texts are usually good vehicles for demonstrative usage.

3.3 Bringing the arguments together: pedagogical modelling

Much of what I have said in the latter part of this chapter seems to have
veered away from the discourse/conversational-analytical and genre-
based concerns of the present book towards lexical and grammatical
preoccupations and cross-cultural comparisons,
but it would be a
mistake to separate the three areas. We do need to have recourse to
discourse analysis and conversational analysis and we do need an over-
arching genre-oriented framework to establish how and when lexico-
grammatical forms occur. As I have tried to demonstrate, discourse
structures and generic patterns of interaction so often crucially depend
on strings of particular lexico-grammatical choices for their realisation.
In the final analysis, discourse is realised and made possible through the
network of lexico-grammar that responds to contexts and to the needs,
goals and relationships of participants, as subsequent chapters will
continue to argue. Moreover, this chapter has not addressed a further
level of encoding which has been shown increasingly to be discourse-
sensitive, that is to say intonation, and real data again force us to re-
assess sentence-grammar models of intonation and models which rele-
gate intonation to imponderables such as ‘attitude’ and ‘emotions’.
Bradford (1988) has shown how a combination of awareness-raising and
more traditional exercises can bring a discourse-sensitive approach to
intonation into the classroom, basing her work on Brazil’s (1985) model.
 3.3 Bringing the arguments together: pedagogical modelling - 67

Thus materials can incorporate facts about the spoken language while
still offering practical exercises that give learners a feeling of ‘doing and
learning’. Equally, there is no reason why a syllabus as a whole should
not have as its primary headings discourse- and genre-based categories
(for an example see McCarthy and Carter 1994: ch. 5), even though its
actual items may be lexico-grammatically organised. A syllabus is not a
methodology, but, rather, reflects a view of language and priorities for
teaching.
This chapter asks the question: what should we teach about the spoken
language? The answer is, clearly, quite a lot, but how and mediated by
what, are different questions. I have argued that a good deal of what
concerns discourse- and genre-analysts is culturally motivated, and that
cultural awareness is the key to avoiding inappropriate transfer of
discourse features across languages and to fostering appropriate transfer.
But I have also argued that such transfer is unlikely to occur (a) if the
lexico-grammatical repertoire is inadequate to the task, (b) if the
teaching materials and syllabus ignore or underplay those very features
of lexico-grammar that give the spoken language its naturalistic flavour,
and (c) if classroom management and methodology militate against the
learner ever being presented with natural opportunities to realise the
kinds of discourse functions we have looked at.
Follow-ups in exchanges, transaction boundaries and interruptions, to
name but a few features, cannot simply be taught via the traditional
‘Three Ps’ methodology (Presentation-Practice-Production). How does one
‘practise’ narrative evaluation? How does one ‘present’ transaction
boundaries? How does one ‘produce’ discourse markers naturally? An
alternative methodology to supplement the ‘Three Ps’ may provide the
answer, one based on the convenient mnemonic of the ‘Three Is’ (Illustra-
tion-Interaction-Induction). Illustration means looking at real data where
possible, or at the very least texts carefully concocted on the basis of
observations of real data. Interaction means talk among learners and
teachers about language (carried out in L1 if necessary), sharing and
forming views, breaking down cultural barriers and stereotypes, etc., in
an environment where discourse awareness activities are brought to the
fore (e.g. activities which focus on particular discourse patterns of L1, or
comparisons between L1 and the target language). Induction means
drawing conclusions about the way in which L2 realises its discourse
patterns and genres and the meanings encoded in particular instances of
lexico-grammar. If this is done in tandem with a syllabus where the
68 + 3 What should we teach about the spoken language?

lexico-grammatical and intonational components are discourse sensitive

and not merely sentence-based abstractions, then teaching about the
spoken language may have an unexpectedly powerful pay-off in the more
rapid acquisition of fluency and naturalistic conversational skills. Such a
claim would not be out of line with the general drift of argument of
second-language acquisition researchers who in recent years have at-
tempted to show that ‘noticing’ phenomena in L2 is a crucial step
towards effective acquisition. But it is what the target for ‘noticing’ is
that matters most, and if the input is impoverished, there will not be
much worth noticing.

Notes

This chapter is a revised version with expanded references of a paper published

in the Australian Review of Applied Linguistics (1995) 17 (2): 104-20.
N House (1985) comments on the universality of the exchange and the adjacency

pair, but suggests that there may be cross-linguistic differences in the degree of
routinisation occurring in them.
ie) On the use of now as a discourse marker, see Aijmer (1988).

4 Work on the translatability of discourse markers across languages includes

Fraser and Malamud-Makowski (1996) (English and Spanish), and Fischer and
Drescher (1996) (English/German/French).
ul Antaki, Diaz and Collins (1996) look upon the third part of the exchange as a

confirmation or refutation of the interlocutor’s contribution in the second

part.
6 Contrastive studies of discourse markers and lists of markers in different
languages that can be used as a basis for lexical input are particularly useful in
the pedagogical context. For an excellent example, see Fraser and Malamud-
Makowski’s (1996) study of English and Spanish discourse markers. See also the
papers in Jaszczolt and Turner (1996) on English and Spanish and English and
Japanese discourse strategies, including marker usage.
N I acknowledge too, in section 1.9 of this book, that corpus linguists have also
inherited the ‘left/right’ metaphors of the written page, and that spoken
corpora force us to revise corpus-analytical terminology as well as grammatical
terms.
8 This is not to say that genres are always true to their stereotypes, and generic
episodes adjust themselves to the exigencies of real-time interaction and the
need to move towards goals, as Chapter 2 has attempted to demonstrate (see
also Benwell 1996).
4
When does sentence grammar become
discourse grammar?!

41 Introduction

The first three chapters have been concerned with fairly broad questions.
In this chapter, we begin to look at the lower-order choices that speakers
make in the discourse process and the creation of genres. In this chapter
and the next, we shall consider grammatical choices within a discourse
framework; there then follow two chapters which look at lexical features
in spoken discourse.
The term ‘discourse grammar’ is frequently heard nowadays and what
it means is not always clear. In this chapter I shall try to define and
exemplify what I mean by the term, and to put it forward as an essential
element in a discourse-based applied linguistics. I shall also attempt to
show that applying discourse-grammatical criteria to the spoken language
is not just an optional extra, but a necessary tool for an adequate analysis
and explication of real spoken data. The arguments in this chapter owe a
great deal to the ongoing work on spoken language conducted with my
colleagues Ronald Carter and Rebecca Hughes, at the University of Net-
tingham, in the CANCODE project (see Chapter 1; also see Carter, Hughes
and McCarthy 1995 for a further discussion of some of the issues raised in
this chapter). Here I shall set out a list of criteria for analysing gram-
matical choice as an aspect of discourse rather than as a phenomenon
confined to the bounds of the clause or sentence. The corpus examples
will, Ihope, demonstrate that the kinds of choices contained within them
depend on contextual features which cannot be ignored and without
which descriptive statements about such features are inadequate.

4.2 Paradigms and actual choice

In his work on intonation, the late David Brazil spoke of ‘existential

paradigms’, that is to say, sets of actual choices open to speakers at any
given place within a discourse (Brazil 1985: 41). Brazil was concerned

69
70 + 4 When does sentence grammar become discourse grammar?

with how lexical items, although often apparently belonging to large,

open sets, occur in real discourse in a manner constrained by context
such that thinking of their occurrence as a choice from a wide range of
possible alternatives is unhelpful. In real situations, the selection of an
item may be from only a small range of plausible alternatives or indeed
may not represent a real choice at all (as with the of in the compound
item Queen of Hearts). In McCarthy (1988; 1992) I attempted to build on
Brazil’s work by examining how speakers create kinds of local synonymy,
antonymy and hyponymy that represent paradigms valid only for those
particular discourses in which the relations are realised, and which may
be different from the words grouped as synonyms, antonyms or hypo-
nyms in a dictionary or thesaurus. Something similar occurs with
grammatical items: the traditional paradigms of choice of items in
grammatical sets may be replaced by paradigms of actual choices in real
discourses.” One such striking example is the relationship between the
pronoun it and the demonstratives this and that, which come together on
many occasions to form a new paradigm of three members (instead of
the two-member demonstrative set with their plurals). This paradigm
offers a significant choice to a speaker or writer who wishes to refer to
entities in the text or in the immediate situation. Viewing the actual
choice as a different paradigm is a very clear example of when ‘grammar’
(understood in the traditional sense as the rules that underlie structural
configurations and choices from closed sets) becomes ‘grammar-in-dis-
course’, that is to say phenomena only fully explicable with reference to
contextual features and speakers’/writers’ choices of form.
In McCarthy (1994) I used written newspaper and magazine data to
show how it, this and that perform similar, yet distinguishable functions
in the organisation of focus and topicality in texts. I argued that (1) it
signalled continued, ongoing topics, (2) this signalled new or significant
focus, and (3) that signalled a variety of distancing or marginalising
functions (e.g. other-attribution, emotional distance, rejection of proposi-
tions or ideas, downgrading or de-focusing, reference across topics, etc.).
Though these three choices were exemplified in written extracts in the
1994 paper, the same can be done with spoken passages. The following
data extracts illustrate the three types of signalling:

(4.1)
It as simple topic continuation?:
[A customer <S 01> asks an assistant <S 02> for help in a bookshop.]
 4.2 Paradigms and actual choice + 71

<S 01> I wonder erm if you could help me.

<S 02> Yeah.
<S 01> I’m looking for two books, one’s a book on organization Schools as
organizations by Charles Handy [<S 02> oh yes] Can you tell me where
it might be.
<S 02> Yes there would be one or two places we’ve got it on stock
[<S 01> yes] it might be on the business section because all his books
are generally at the business section.

(4.2)
This as focus on a new or important topic’:
[<S 03> is recounting the first time he heard the phrase pin money, when
he went to get a library ticket from a British Council office overseas.

<S 03> I worked there just a very short time erm it was when I was doing
my PhD I went out there to do some research and erm I always
remember there was a Director there at that time he was called
Macnamara and I went along there to get to get him to sign fora
library ticket so that I could use the university library or the public
library or something and he said erm erm I always remember this I’d
never heard this phrase before he said how would you like to earn
some pin money and I was sort of, young fellah I didn’t know what
this meant [laughs]

(4.3)
That as distancing”:
[A university tutor is talking about Jane Austen]

<S 01> She’s looking at people who pretend for instance to have good
manners and the essence of good manners is concern for other
people and she can see that within that society many many people
have outwardly excellent manners but that is something of veneer
it’s shallow it’s pretence, appearance

In (4.2) and (4.3), the speakers could have just continued their topics with
it instead of this or that, but they choose instead to focus or distance the
discourse entities referred to. The choices available operate at a level
beyond that of the clause or sentence and are to do with signalling the
speaker’s stance towards the message. When grammatical choices
72. » 4 When does sentence grammar become discourse grammar?

operate in this way, ‘grammar becomes discourse’, that is to say,

grammatical phenomena require discoursal explanations.
But why should we bother to seek a discoursal context for pedagogy if
the traditional explanations of items such as pronouns and demonstra-
tives stand us in good stead for most teaching needs? The question is a
common-sense one, but has its answer in the fact that the traditional
explanations often cannot capture adequately the selection procedures
the learner will need to engage with in constructing larger segments of
text (which we surely hope our learners will do in ‘writing skills’ or
‘speaking skills’ programmes). Traditional, sentence-based grammar
teaching methods (a) tend to focus on only spatio-temporal meanings of
the demonstratives, and (b) do not bring together for illustrative
purposes the pronoun it and the demonstratives, typically keeping them
in separate compartments, despite the evidence of data that they
frequently exist as viable alternatives in real discourse. ‘Re-arranging the
paradigms’ is not just re-arranging the furniture to liven up the living-
room; its purpose is more faithfully to represent actual language use.
This (It? That?) should always be the prime motivation for treating
grammar as discourse.

4.3 Rules that do not say enough

Pedagogical grammarians simplify the grammatical facts and provide

rules of thumb about second language grammar that work for most
learners in most situations, and we shall always continue to do that.
However, simplification that fails to generate appropriate grammar may
be counter-productive. One such simplified rule might be ‘the English
past perfect is used when an event happens in a past time before another
past time’. This enables the learner of English to construct well-formed
sentences such as I spoke to Brian Thorne yesterday for the first time. I had
actually met him once before, many years ago. However, the same two
sentences would clearly be as well-formed if the second were in past
simple, although with a different emphasis. We have, then, a rule that
explains a grammatical choice that is certainly well-formed, but which
does not offer sufficiently rigorous guidelines to generate the choice in
many other types of situations where it is equally appropriate. When this
problem arises, there is no substitute for looking at choices that speakers
and writers have made in real contexts and examining the contextual
features that would seem to have determined one choice or another.
 4.3 Rules that do notsayenough - 73

Contextual awareness then underpins choice, and grammar becomes

discourse once more. Such an approach does not exclude simple, clear
guidelines where these can be formulated. In the case of the past perfect,
for example, a significant number of occurrences of the tense form are
found in indirect speech reports, such as:

(4.4)
[<S 01> is talking about an accident.]

<S 01> And I chipped a bone at the end of, on the end of my elbow, I
didn’t know it was broken it was two weeks before I went to the
hospital it just seemed to get worse and worse.
<S 02> Right.
<S 01> And then when I went finally went they said that I'd chipped this
bone.

(4.5)
[<S 01> is recounting a story about a tramp who deliberately got into
trouble by creating traffic accidents just to get himself a bed in the police
cells for the night; she tells how he caused a collision with her car but
escaped unhurt]

<S 02> So it was a bit of a miracle he wasn’t hurt wasn’t it.

_ <S 01> l Apparently it
was his party no it was his party-piece because the police told me
that he’d done it very often this ’cos it got him a bed for the night
you know it got him in hospital.

Observation of real data that results in a simple rule or guideline is

extremely helpful, but sometimes the occurrences are more complex and
need closer scrutiny. In the case of the past perfect, the most interesting
instances are those where there is little formal evidence (such as the
presence of a reporting verb) to explicate the choice, and one is forced to
fall back on contextual features only. In (4.6) below, several past perfect
forms occur, none of which could be said to be conditioned by any
structural or lexical constraints.®

(4.6)
[Two young women are talking about mutual friends from their days
together at Brunel University, in the south of England. They were driving
74 +» 4 When does sentence grammar become discourse grammar?

along a motorway when the recording was made, with <S 02> driving,
which accounts for her minimal contributions to the talk!]

<S 01> I got on better with Glynbob I think and John Bish let me and
Trudie sleep in his bed last time we went up to Brunel or the one
time when we stayed in Old Windsor with them cos erm Ben had
given us his room cos he’d gone away for the weekend and erm it
was me and Trudie just in Ben’s room and John Doughty had a
double bed so he, John Bish had a double bed so he offered us his
double bed between us and then slept in Ben’s room cos Ben and PQ
had gone away for the weekend but they tried to get, they'd gone
away and tried to get back like to catch me and Trudie before we left
[<S 02> Yeah] and they just missed us by half an hour they were
really pissed off because apparently they'd been driving really fast
like trying to get back but erm I mean we didn’t know they were
trying to get back we didn’t leave until like very late we went to the
Little Chef for breakfast on the Sunday cos it was only over the road
from where they were living and Andy Symons the bar manager like
came back with us and stayed the night at Glynbob’s house as well
so he came to Little Chef with us in the morning as well.
<S 02> Oh God.
<S 02> There was like loads of us in the Little Chef ... and we got there
and we had to wait like ages for them to do the food and stuff and
we were going oh we don’t mind we don’t mind ... 1remember
going to the Little Chef after the Valedictory and erm we took the
minibus down and Cooksie drove cos he’d been driving all night
and he drove the minibus down and it was in the morning it was
after like the ball and PQ still had some wine left ...

What these examples have in common is that they give a reason or

justification for the events that are recounted. They are not the main
events themselves, they are rather something the speaker feels to be a
necessary background to what happened. Note the prevalence of cos (and
in one case apparently). This use of cos/because to justify or explain events
has been commented on by Ford (1994) as a feature of conversation,” and
it does seem to co-occur with the past perfect in a noticable number of
cases (76 times in the one million word CANCODE sample). Another clear
case of backgrounding is (4.7):
 4.3 Rules that do not sayenough + 75

(4.7)
[<S 01> is telling a story about getting drunk when he was a young man.
Bass is a make of beer popular in Britain. The crew is his way of referring
to his gang of friends/colleagues.]

<S 01> During the war they lowered the specific gravity of the beer.
<S 02> Aha.
<S 03> Yeah they did.
<S 01> And er round about 1947 or so it was back to normal and I’d gone
out it was after some exam results for the bank and, celebrating
with the crew and unbeknownst to me they’d got the new deadly old
Bass on you see and they finished up draping me over a hedge.

Examining the discourse conditions under which these past perfects

occur enables us to posit a broad macro-function for the tense form at a
level beyond the sentence, in terms of how clauses recounting events
relate to one another, some being backgrounded, others being fore-
grounded as main events.’ In other words, a discourse grammar focuses
on the kinds of preoccupations that speakers routinely deal with in
speech, that is, how can I best organise my message to make it clear,
coherent, relevant, appropriately organised, etc? This accords with
radical proposals to re-think the way we construct grammars such as
Brazil’s (1995) ‘grammar of speech’, where he pertinently sums up the
position with regard to grammar and the language user:

In other words, we do not necessarily have to assume that the

consideration of such abstract notions as ‘sentences’ enters into
the user’s scheme of things at all. (p. 15)

By the same token, even though they necessarily may have a more
conscious emphasis on grammatical form, it should not be assumed that
foreign-language learners do not need or want to pay the same attention
to choices reflecting organisation, staging and coherence of the overall
message. However, the hypothesis that past perfect coincides frequently
with explanations and justifications reflects a probabilistic view of
grammar, and one which says no more than ‘this is a fact of occurrence
in a number of cases sufficiently large to warrant pedagogical attention
and to provide a useful, probabilistic guideline’. It is distinct from
deterministic grammatical statements, which are open to confirmation
or falsification through structural tests of well-formedness. Many state-
76 + 4 When does sentence grammar become discourse grammar?

ments about grammar as discourse will be probabilistic, but that does

not make them any the less useful pedagogically. Such statements may
be open to modification, but if they can be shown to be reliable across
different contexts, then they stand as good examples of grammar
becoming discourse, and can often suggest explanations of grammatical
choices that the conventional, more deterministic rules explain only very
inadequately.2 Grammar becomes discourse when the conventional,
sentence-based rules simply do not say enough to enable the learner to
make appropriate grammatical choices in the way native speakers do to
organise and ‘stage’ their messages, and where choice can only be
explained in context.

4.4 Spoken and written grammar

Grammar must become discourse in order to answer questions about

differences between spoken and written language, for it is only by
observing actual discourses that we can properly describe the distri-
bution of forms in the spoken and written mode. Much grammar is, of
course, shared by both modes, and it would be an exaggeration to
suggest that huge differences exist in distribution for every grammatical
feature. The crucial point is that we should never assume that if a
grammar has been constructed for written texts, it is equally valid for
spoken texts. Some forms seem to occur much more frequently in one
mode or the other, and some forms are used with different shades of
meaning in the two modes. Since almost all grammars in the past have
based their descriptions on written norms (and even those which have
used spoken data, such as Quirk et al 1985, have often neglected common
features of spoken grammar; see Carter and McCarthy 1995b for exam-
ples), it thus becomes important to look at what really happens in the
spoken language to detect significant differences. Carter and McCarthy
(1995b) and McCarthy and Carter (1997a), using the CANCODE corpus,
identify a selection of key areas where differences occur. These include
the prevalence of certain types of ellipsis in spoken language (e.g. of
subject pronouns, auxiliary verbs, articles, initial elements of fixed
expressions), different formal types of speech reporting in spoken and
written (see Chapter 8 of this book for a full account), and the occurrence
of pre- and post-posed items (topics and tails) in conversation. These last
phenomena deserve closer scrutiny and discussion. Examples (4.8) to
 4.4 Spoken and written grammar - 77

(4.11) represent structural configurations almost exclusively associated

with informal speech:

(4.8)
<S 01> That woman who’s a verger at church, her husband, his parents
own that butcher’s shop.

(4.9)
<S 01> Paul in this job that he’s got now when he goes into the office he’s
never quite sure where he’s going to be sent.

(4.10)
[Two friends deciding what to eat in a restaurant]

<S 01> I'm going to have Mississippi Mud Pie I am.

<S 02> I’m going to have profiteroles. I can’t resist them I can’t ... just
too moreish.

(4.11)
[Students chatting in doctor’s surgery waiting-room.]

<S 01> You got a cold too?

<S 02> Can’t seem to shake it off... everyone’s going down like flies.
<S 01> Trouble is can leave you feeling weak for so long it can flu.

(4.8) and (4.9) utilise a slot before the ‘clause proper’ (in the traditional
sense) begins. Such phenomena have been variously termed ‘left disloca-
tions’, ‘themes’, and ‘topics’ (see e.g. Aijmer 1989; Geluykens 1989 for
English, French and Italian; Geluykens 1992 for English; Blasco 1995 for
French; Rivero 1980 for Spanish). This ‘topic slot’ or ‘head’ (McCarthy
and Carter 1997b) plays an important role in how the speaker orientates
the listener, it is an act of consideration to the listener, for example in
(4.8) taking him/her from a convenient anchor already familiar to some
entity that is new and which is to be the topic of the upcoming clause.
As such it is a quintessentially spoken feature, reflecting the demands of
face-to-face interaction and the real-time synthesis of talk. Its mirror-
image counterpart, the ‘right-displaced/dislocated’ (see Ashby 1988, 1994
on French; Heilenman and McDonald 1993 on French; Fretheim 1995 on
Norwegian) or ‘tail’ slot (Aijmer 1989) corresponds regularly with
contexts which are evaluative, as it does in (4.10) and (4.11); it signals
78 +» 4 When does sentence grammar become discourse grammar?

more than the mere core clause without a ‘right-hand’ dislocation, and
the overlay is heavily interpersonal. The prevalence of quotation marks
around the unfamiliar terminology in these last few sentences is not
only because the structures themselves are rarely dealt with by gram-
marians but above all because much of the terminology itself is locked
in a written, sentence-based perspective on language. Spoken language
has no ‘left’ or ‘right’ in the way that characters on a page do, and
topics and tails pass without note and so naturally that it seems odd to
suggest that anything is ‘dislocated’ at all (see also section 3.2.2). The
metaphors of written text should not be transferred uncritically when
grammar confronts spoken discourse, and the fact that spoken language
is produced in time rather than space, for a here-and-now listener
rather than a temporally displaced reader, becomes paramount in
explaining grammatical phenomena. Discourse drives grammar, not the
reverse.
Another example which underlines the power of context over grammar,
and not vice-versa, is the distribution in spoken and written texts of the
structure be to with a future meaning. Many teachers will recognise
themselves (as I have done myself in the past) teaching sentences such as
You are to be at the airport at eight-thirty, where the be to denotes a firm and
unshiftable fact about the future. Yet in reality, this structure is so rare in
everyday spoken language that it would, in my opinion, have no signifi-
cant place in a speaking-skills grammar course. In one million words of
CANCODE data, only four examples of future be to occur, one in a
university small tutorial group, and three (one of which is repeated and
co-ordinated) in semi-formal meetings as (4.12) and (4.13) illustrate:

(4.12)
[Tutor speaking in university tutorial on Pride and Prejudice.]

<S 01> And there’s also of course the famous first sentence of Pride and
Prejudice from which this section has received its name It is a truth
universally acknowledged that a single man in possession of a good
fortune must be in want of a wife. That this statement is meant to
have ironic qualification is shown both by the orotundity of the
diction and by contrast with what is said in the following sentence
that the concern is to be not for the universe but with the
neighbourhood not with the totality of mankind but with the
surrounding families. Em that’s all it says about that bit.
 4.5 Units of description in grammar and discourse - 79

(4.13)
[business meeting]

<S 01> Oh no oh right well no but it’s taken two months to do that.
<S 02> LRob is to look at it
and Ann Pascoe to look at it and formal comments to be
collated and sent back to David.
<S 03> LAnd one month for that.
<S 02> Yeah.
<S 03> I think that’s safest.

The reason for the rarity of this feature is quite clear: be to in its future
meaning is a distancing form, suggesting external-and impersonal
authority of some kind that may appear pompous or face-threatening in
face-to-face talk, where speakers overwhelmingly prefer to realise the
same meanings with ‘softer’ expressions such as supposed to or going to.
Hardly surprising, then, that we only find it in a meeting and a tutorial,
where ‘authority’ and ‘distancing’ characterise the genres. In many
written contexts, on the other hand, authoritative statements may be
put forward without face threat, especially in journalistic reporting
registers (for examples of be to in newspaper reports announcing
decisions, events, changes, etc., see McCarthy and Carter 1994: 126; also
Chapter 5 of this book) or in texts denoting regulations, obligations, etc.
Grammar becoming discourse here means, once again, seeing gramma-
tical choice and structural configuration in talk as motivated by inter-
personal factors, not just ideational or transactional ones. Distinguishing
the key aspects of written and spoken contexts is crucial to the obser-
vation and understanding of their widely differing distributions of forms
and meanings and to the prioritising of those forms and meanings in
teaching.

4.5 Units of description in grammar and discourse

Several linguists have noted the relative absence in spoken data of forms
common in written texts which are
often seen as core features of
grammar and central to grammatical description. Anyone who looks at
large amounts of informal spoken data, for example, cannot fail to be
struck by the absence of well-formed ‘sentences’ with main and subordi-
nate clauses. Instead we often find turns that are just phrases, incom-
plete clauses, clauses that look like subordinate clauses but which seem
80 + 4 When does sentence grammar become discourse grammar?

not to be attached to any main clause, etc. Hockett (1986) makes the
point that, while analysts have long ignored such phenomena, ‘speakers
and hearers do not ignore them - they carry a sizeable share of the
communicative load’. Therefore, the fact that well-formed sentences are
not the norm urges a re-assessment of the usefulness of notions such as
‘main’ and ‘subordinate’ clauses. (4.14) amply illustrates the problem of
‘subordinate’ clauses:

(4.14)
[Two students are talking about what people are going to wear to a
forthcoming ball.]

<S 01> I really am I’m so pissed off that everyone’s erm everyone’s going
to be wearing erm
<S 02> Cocktail dresses.
<S 01> I don’t, I really don’t see the point the whole point of a ball is that
you wear like a proper dress
<S 02> Wear a ball dress Iknow I mean my dress is huge.
<S 01> So is Nicola definitely going to, erm is Nick definite
<S 02> lWell she she
says she is but if she sees everybody else wearing a cocktail dress
she’s bound to fork out the money she’s got loads of money.
<S 01> Cos mum said to me you know that she would buy me like aa
little black dress but the thing is then I wouldn’t feel right you
know.
<S 02> Well I mean you wear a
<S 01> lShe, but you know she means like something from like erm
Miss Selfridge or something.
<S 02> Yes if, Imean you wear a little black dress just to, you know
<S 01> Clubbing or something.
<S 02> To a party.
<S 01> Yeah exactly.

The first ‘sentence’ seems to be spoken by two speakers, with <S 02> pro-
viding the object of <S 01>’s verb wearing. In the next pair of utterances,
<S 02> repeats <S01>’s direct object in a slightly reworded form
(<S 01>... wear like a proper dress <S 02> Wear a ball dress). In the final
utterances of the extract, both speakers ‘complete’ the same clause, but
with different constituents (Clubbing or something/To a party). How do we
analyse ‘other-completed clauses/sentences’? How do we analyse ‘other-
 4.5 Units of description in grammar and discourse - 81

repeated grammar’? Are <S 02>’s utterances here part of <S 01>’s ‘sen-
tences’, or units of their own? Where both complete the clause, are the
two different constituents of equal status in its structure? A grammar-in-
discourse approach sees structure as a collaborative/negotiative process
rather than as a deterministic product. Within the relevant factors of a
description it includes real-time contextual features such as turn-taking,
repetition and joint construction by more than one party.!°
There are some items in (4.14) which do not seem to be main clauses:

that everyone’s erm everyone’s going to be wearing cocktail dresses

that you wear like a proper dress
(she says) she is
if she sees everybody else wearing a cocktail dress
(mum said to me) that she would buy me like a a little black dress

The first two are complement-clauses (thus not subordinate in the true
sense), which are frequent in conversational language. Two are reported
clauses within speech reports (also frequent; see Chapter 8), which are
arguably ‘main clauses’ transferred from other discourses. Only one
seems to be a conventional conditional subordinate clause with if. The
clause Cos mum said to me you know that she would buy me like a a little black
dress would seem superficially to be a candidate for subordination, but (a)
it is separated from its ‘main’ clause by an intervening turn, and (b) it
seems simply to add background information rather than place any
restriction or contingency on the main portion. It means ‘this is the
reason I’m asking about Nicola’ and does not exercise any causal
conditioning on the main clause. It seems more reasonable, therefore, to
treat the cos clause as non-subordinate, but discoursally as providing
background/supporting information. Something
similar operates with
certain types of if clauses, as in these three further examples from the
same conversation:

(4.15)
<S 01> Oh there’s orange juice in the fridge as well if you want a drink
...erm no if we have this and go back to your house.

(4.16)
<S 01> Yeah help yourself... there’s scissors in the drawer if you need to
cut it open.
82 +» 4 When does sentence grammar become discourse grammar?

The first and third if-clauses are best considered, from a discourse point
of view, as non-subordinate explicating units, making explicit the
speaker’s reasons for uttering the main clause. The second example (if we
have this and go back to your house) seems to operate as a polite suggestion
without any element of conditionality. In a discoursal sense it is clearly
‘main’ information and resembles the kind of ‘topicalisation’ referred to
in section 4.4 above (see also Haiman 1978 for a discussion of con-
ditionals as ‘topics’).
This kind of problem of classification has led some linguists to propose
abandoning the notion of ‘subordination’ altogether when it comes to
describing and explaining spoken grammar (e.g. Blanche-Benveniste
1982; Schleppegrell 1992) and to advocate substituting the clause as a
more viable basic unit for spoken language than the sentence (Miller
1995). Another good reason for advocating this is that some sentence-
configurations, for example non-finite clauses / prepositional clauses plus
main clause (such as On leaving the building she noticed a black car), which
are found in formal written language, are extremely rare in conversation.
Blanche-Benveniste (1995) gives examples for spoken French, and much
the same applies to spoken English (e.g. see Esser 1981 on the absence of
embedded clauses). Grammar becomes discourse when conventional
sentence-based units of description fail to account for the facts, sug-
gesting an alternative descriptive model based more on units of in-
formation and interpersonal considerations generated within real
contexts.

4.6 Grammatical puzzles

By no means has all of the grammar of a language such as English yet

been satisfactorily described and explained, and, where areas are fuzzy
or indeterminate or apparently unsusceptible to rule-creating, seeing
grammar as discourse can often help. Let us take the case of get-passives
in English (as in His bicycle got stolen). These have been studied for a long
time, though mostly on the basis of concocted sentences, with the
notable exception of Collins (1996), which this present look at the
phenomenon parallels, but with the intention of drawing wider implica-
tions, which Collins does not.'! Linguists have considered the interchan-
geability (or otherwise) of get- and be-passives, or the place of get and be on
a ‘passive gradient’ (Svartvik 1966), and the typical absence of explicit
human agents in get-passive sentences (see especially Hatcher 1949, also
 4.6 Grammatical puzzles + 83

Stein 1979; Granger 1983; Gnutzmann 1991; Collins 1996). The debate
has also homed in on the potential of get for focusing simultaneously on
actions and their end results (see Stein 1979; Vanrespaille 1991). Other
considerations that have come into play are whether get-passives corre-
spond to contexts where (usually ill) fortune plays a role in the event (He
got killed, It got burnt, etc.), the historical development of the get-passive in
relation to other meanings of get (Givén and Yang 1994), and different
distributions across different varieties of English (Collins 1996). And yet
there is a general feeling that get-passives have continued to elude
satisfactory description.!* Once again, the way forward seems to be to
look at real data, to consider grammar as discourse and to bring to the
fore that discourse involves speakers and listeners, not just messages.
The CANCODE corpus contains 139 clear get-passives, from which a
very consistent (but not exclusive) pattern emerges.!? Of the 139
examples, 124 refer in some way or another to an ‘adversative’ context
(i.e. a state of affairs that is signalled by the conversational participants
as manifestly undesirable, or at the very least, problematic); these
include verb phrases such as:

got flung about in the car

got killed
got locked in/out
got lumbered
didn’t get paid
got picked on
got sued
got burgled
get intimidated
get criticized
get beaten

Of the 139 examples, 130 have no agent explicitly stated (i.e. 93%, which
sits well with Collins’ 1996 figure of 92%). Among the examples which do
have stated agents are:

most things got written up by scribes

the whole bus got stripped by the Italian police
and got sued by the owners
she’s going to get eaten by the wolf
you get intimidated by the staff on the labour ward
84 +» 4 When does sentence grammar become discourse grammar?

These agents are somewhat impersonal or, in the case of the wolf, non-
human. The absence of agency or the presence of impersonal, non-
specific agency fits in with what linguists have previously noted (e.g.
Granger 1983: 194).
But in a number of the ‘adverse’ cases, we are faced with circumstances
that are not inherently adverse or undesirable, for example:

(4.17)
[A customer in a village shop has just realised that the shopkeeper has
remembered a neighbour’s fish order but forgotten her order of fish for
her cat. She addresses the neighbour humorously.]

<S 01> So you got remembered and our cat got forgotten.

(4.18)
[Students talking about upcoming hectic social timetable.]

<S 01> I’ve got invited to the school ball as well.

<S 02> Are you.
<S 03> Don’t really fancy it.

(4.19)
[Students discussing job prospects.|

<S 01> Do you know how much lawyers get paid for an hour the best
ones
<S 02> l] don’t I don’t care.
<S 01> Six hundred pound an hour.
<S 02> I don’t care.

It is the speaker’s stance towards the situation that signals it as undesir-

able or problematic rather than the situation in itself (see Chappell 1980;
see also Sussex 1982 for a critique of over-simplifying the adversative/
beneficial dichotomy). In (4.17), although the situation is clearly benefi-
cial for the neighbour whose fish order was not forgotten, the speaker
sees it as part of her own general misfortune. In one of the relatively few
occasions where the context is clearly beneficial rather than adversative,
we can still see the speaker overlay that can be effected by the get-passive,
in this case suggesting a downplaying of self-praise when reporting
success:
 4.6 Grammatical puzzles + 85

(4.20)
[The speakers are talking about <S 02>’s past successes as a tennis player.]

<S 01> And were those like junior matches or tournaments or county
matches.
<S 02> Er both county and er, well I played county championships and
lost in the finals the first year and er I got picked for the county for
that and then so II played county matches pretty much the same
time.
<S 01> Right, good.

The key to understanding the get-passive is that it reflects the stance of

the speaker rather than the content of the message. It is yet another case
where examining sentences and jettisoning the people who produce
them is inadequate. The get-passive might indeed be a linguistic puzzle,
but it is considerably demystified the moment we look upon it as
something the speaker overlays onto events to reflect his/her stance
towards those events. Some linguists have recognised this, most notably
Lakoff (1971) and Stein (1979), but the benefit of examining real spoken
data is that intuitions on that score can be supported by facts.'*
We can say that the get-passive coincides overwhelmingly with adverse
or problematic circumstances (89% of occurrences) but that these are
adverse/problematic as judged by the speaker. Get also coincides overwhelm-
ingly with the absence of an explicit agent (93% of occurrences),
suggesting that emphasis is on the person or thing experiencing the
process encoded in the verb phrase, rather than its cause or agent.
These two descriptive statements are directly useful for teaching, as
are many of the actual contexts and examples found in the corpus. In
this case, and, one suspects many others, the linguistic puzzle can best
be illuminated by treating grammar as discourse rather than merely as a
feature of the internal structure of sentences. But the kinds of statements
we have ended up making about the get-passive are different from
statements of structural prescription (e.g. that the passive is formed with
the past participle, not the base-form of the verb); we can therefore
distinguish once again between deterministic grammar and probabilistic
grammar, the latter being statements of what is the most likely context
of use. Grammar-as-discourse leans more often towards probabilistic
grammars, and sees probabilistic statements as no less useful for
language learners than deterministic ones.
86 + 4 When does sentence grammar become discourse grammar?

4.7 Conclusion

Whether one looks at individual grammatical items such as pronouns or

demonstratives, or whether one considers wider structural phenomena
such as topics and tails, subordination, tense choice, etc., in spoken
discourse, grammar is often most adequately explained by referring to
contextual features and, above all, by taking into account interpersonal
aspects of face-to-face interaction. I have argued that a grammar-as-
discourse approach includes:

e The appropriate re-arranging of existing paradigms in line with real

choices in discourse.
e The re-examination of conventional, deterministic rules in the light of
real contexts and probabilistic correlations between forms and con-
texts.
e The re-assessment of traditional units of description to reflect what
kinds of units are actually (differentially) manifested in spoken and
written discourse.
e The unwillingness to accept linguistic puzzles as insoluble mysteries
without subjecting them to scrutiny in actual occurrence.

The more one pursues these lines, of course, the more one realises that
grammar can become discourse whenever the (applied) linguist wants it
to and that the two are not separate levels of language that brush
shoulders only when awkward problems need solving. Grammar only
really exists in discourse, and it is best viewed as the regular traces left
behind by myriad conversational encounters where the exigencies of
person-to-person communication are paramount, always taking prece-
dence over mere ‘content’. In the words of Hopper and Thompson (1993),
grammar is ‘sedimented conversational practices’. This whole book takes
the view that grammar should ideally be seen, in Fox and Thompson’s
(1990) words:

... aS necessarily including the entire interactional dimension of

the communicative situation in which conversationalists
constitute the people and things they want to talk about.

Grammar-as-discourse thus confronts the negotiation of social identities,

in which formal choices are fully implicated. For the applied linguist
interested in language teaching applications, a grammar-as-discourse
approach is certainly challenging. Some aspects of that challenge may be
 Notes + 87

unrealistic at times (for example, in a spoken language awareness lesson,

just how does one bring out ‘the entire interactional dimension of the
communicative situation’ if one does not have access to natural data in
the classroom?). However, it has its unexpected rewards too in the new
light that is often shed on age-old and apparently intractable gram-
matical problems. It also offers a way of bridging the unfortunate gap
that frequently exists between the ‘grammar’ class and the ‘speaking
skills’ class, too often perceived by teachers and learners alike as
different worlds that have no agreed meeting place.

Notes

1 This paper is based on a presentation with the same title given by the author
and Rebecca Hughes at the TESOL Convention, Long Beach, California, 1995. It
repeats arguments put forward there, but uses different data sets for exempli-
fication. A different and more extended version of the paper may be found in
Hughes and McCarthy (1998).
The idea of re-grouping items based on their discoursal properties rather than
just their formal characteristics is echoed in Hoffmann’s (1989) grouping of
words such as I, you, here and now as ‘means of referring to elements of the
situation of utterance ... they form a field of their own that is in opposition to
traditional categorisation (pronoun or adverb)’. Similarly, Crymes (1968:
64-70) brings together do so, do it, do this and do that as a discoursal set of
substitutes and examines their usage.
w It in this example refers to a thing (a book), but, as can this and that, may also
refer to events, facts, propositions, etc. (see Peterson 1982; Ehlich 1989;
McCarthy 1994).
There is also, of course, the ‘new this’ common in informal spoken narratives
where this is in opposition to the indefinite article (e.g. ‘Then this policeman
suddenly appears and everybody runs’). See Wald (1983) for examples and a
discussion.
ul There is another sense of that which occurs in spoken discourse and sometimes
in popular journalism, which refers to entities as given, known but not
topical, exemplified in the opening segment of a story told by an elderly,
disabled ex-member of the British Women’s Air Force who had the ‘honour’ (in
her eyes) of being invited onto the flight deck during a holiday flight. The story
opens with a characteristic ‘new this’ (see note 4):
<S 01> Well, I don’t know how I got this honour really. I had all the badges,
you know, I used to be in the air force, and I’d spoken to two or three
people, but when we got on the plane, they took me on first, because they
had to lift me on, you know, with that lift, and the pilot was sat in one of
the seats. There was nobody on the plane but me, and, I don’t know, I
88 + 4 When does sentence grammar become discourse grammar?

must have said, either said something funny about flying, or he’d said
something, noticed my badges, I don’t know which it was...
That lift here seems to mean ‘one of the typical lifts they have on aircraft for
getting disabled passengers on and off, which I can assume you know about’
(reinforced by the shared knowledge marker you know).
The kind of constraint I refer to here is that exercised by words like before,
already or just, which often provide compulsory contexts for the use of past
perfect. For example:
(1)
[<S 01> is recounting how he found it difficult to be off work on sick leave.]
<S 01> Cos then at the time all I wanted to do was get back to work and being
stuck in the, you know in for six months [<S 02> mm] you know I just I
just couldn’t believe that they could leave anybody, I mean wasn’t it
better to have me back at work instead of paying me sick pay
<S 02> Yes
<S 01> Em and because I’d always worked and I'd never ever been off work
before you know I just couldn’t handle it

(2) Here there are a number of past perfects but the italicised one is
constrained by already, while the ones before it are not:
[<S 01> is talking about Christmas.]
<S 01> Well yeah I mean Christmas was really good for us this time. I
mean we’d done a lot of pre-planning for it hadn’t we Mary
you know we'd er
<S 02> lSaved
<S 01> lsaved money for the, obviously to to cut the costs
down towards, er we'd saved you know a fair a fair bit for presents and
we'd already saved a hell of a lot of money for the food
<S 02> Mm
NI See also McTear (1980) on the pragmatics of because.
On foregrounding and backgrounding in general through grammatical
aspect, see Hopper (1979).
9 The coincidence of past perfect with contexts of explaining/justifying events
seems to be true of many cases in written texts too (Hughes, personal
communication).
10 Harris (1990) argues for the integration of strictly linguistic and such non-
linguistic features of communication, though Fleming’s (1995) critique of
Harris’s position rightly points up some of the difficulties of integrating
temporality (e.g. the real-time constraints of speech) into linguistic descrip-
tion. Harris’s perspective is clearly one which supports viewing grammar as
discourse.
11 Vanrespaille (1991) is another exception, in that her study of the get-passive is
corpus-based. Although her corpus includes spoken material from the Survey
of English Usage, she does not indicate precisely what proportion of her
 Notes + 89

approximately 700 instances of the get-passive are natural spoken. However,

from her tables of results, it would seem that a large number of her examples
come from (written) drama texts.
12 Chappell (1980) refers to ‘tantalizingly unresolved questions’ about the get-
passive.
13 The figure of 139 occurrences in one million words is two-and-a-half times as
frequent as Granger’s (1983) figure of nine occurrences in her spoken corpus
of 160,000 words. Our figure matches only slightly better with Collins’ (1996)
figure of 96 per million. His data are taken from the London-Lund and ICE
corpora, which, I would argue, are less representative of everyday informal
conversational genres than the CANCODE corpus. What is more, the 65
occurrences of core get-passives in a mixed written corpus of one million
words (created for spoken/written comparisons with CANCODE) does not
either accord completely with Collins’ figure of 43 per million for written,
though both corpora coincide in suggesting that get-passives are twice as
frequent in spoken as in written.
14 Lakoff uses concocted sentences. Stein uses a corpus, but it is written (novels
and plays).
a)
Some patterns of co-occurrence of verb-forms in
spoken and in written English

5.1 Introduction

In this chapter we continue to focus on grammar, and take further the

discussion of written and spoken parallels and differences embarked
upon in section 4.4. Here we redress the temptation to conclude that
everything is different between spoken and written grammars, and show
that features of patterning at the discourse level are often shared by the
two modes of communication. However, the fact that the chapter also
shows patterns that are not shared between the two modes is very
important indeed, for it underscores one of the central arguments of this
book, that we cannot assume that grammars modelled on written
language can simply be imported wholesale into the description of
spoken language. Spoken grammar must always be elaborated in its own
terms, using spoken data. If, at the end of the exercise, spoken and
written are shown to have many features in common, then this is a
convenience to be thankful for, and not something that can be prejudged
without careful research. Our task is to identify precisely those areas that
are shared and those which are not.
Amongst the linguists who have always argued for discoursal inter-
pretations of lexico-grammatical phenomena, the late Eugene Winter
stands out, and his most enduring achievement has without doubt been
the demonstration of the way the several levels of linguistic form work
in harmony to create the text. Elements of form, whether lexical or
grammatical, share a potential in contributing to the process of inter-
pretation of the larger form: the coherent, cohesive and patterned
artefact that is the text (Winter 1982). For our purposes, we could add the
words ‘or genre’ to the last sentence. In Winter’s conceptual framework
the grammar of any individual clause or sentence is involved in creating
higher-order patterns. Grammar and context work in synergy, and any
proper study of text involves a close scrutiny of the patterning of
grammatical forms. Winter’s work is normally associated with written

90
 5.1 Introduction - 91

text analysis, but in this chapter I shall try to show that the principles of
textual-grammatical patterning apply equally to spoken discourses. The
spoken textual product is no less a record of the harmonious collabora-
tion of grammatical form with other levels of form than is the written
text. I shall attempt to demonstrate that the kinds of grammatical
patterns found in certain types of written texts (narratives and reports)
have their functional counterparts in spoken discourse and that the
business of signalling the macro-level development of the discourse is of
the same importance and given similar prominence in the spoken as in
the written. But where there are differences between spoken and written,
these will be noted as important. I shall, in particular, hope to develop
Winter’s notion of ‘situation’ utterance and its ‘sequence’ in subsequent
utterances. ‘Situation’ and ‘sequence’ are two concepts which Winter,
working with written sentences, sees as fundamental in the interpreta-
tion of sequences of clauses in text (Winter 1982: 2).
The written examples in this chapter are taken from British popular
journalistic sources, in the tradition of Winter’s own work (ibid.), along
with some literary examples. Spoken examples are taken from the
CANCODE corpus. Written and spoken examples taken side by side offer
the researcher a number of possible avenues of exploration and ques-
tions to answer, including:

1 What sorts of recurring grammatical patterns can be observed over

_ clause-boundaries in the written and in the spoken mode?
2 How determined are the patterns? Is the ordering of grammatical
elements within them relatively fixed or flexible?
3 What are the functions of such patterns within the text or utterance
sequence?
4 Do patterns found in one mode occur in the other with the same
realisations?
5 Where patterns occur in one mode but do not seem to occur in the
other, is there evidence that the same functional patterns exist but
with different formal realisations in each mode?

Not all of these questions can be addressed or answered in this short

chapter, but where particularly interesting evidence presents itself,
relevant issues from (1) to (5) will be taken up in turn.
92 + 5 Patterns of co-occurrence of verb-forms in English

5.2 A clear case of parallel: the narrative ‘situation-event’ pattern

Within the field of grammar-and-discourse studies of written texts,

published work already exists on correlations between grammatical
features and text-types and macro-functions (i.e. functions beyond the
level of sentence, such as paragraph organisation, reference across
sentences, etc.). An appropriate example to begin with is Zydatiss’ (1986)
paper on so-called ‘hot news’ texts such as are often found in newspapers
and on news broadcasts. A hot news text typically has an attention-
grabbing headline and/or first sentence, operating rather in the way of
an ‘abstract’ (to use Labov’s (1972) term for the narrative-initial utter-
ance(s)) with which a teller gives a broad outline of what the upcoming
story is about and why it might be worth listening to (on the other
elements of Labov’s model, see section 3.2.3). In clause-analytical terms
(see Hoey 1983: 138ff), the abstract has a parallel in the ‘preview’ segment
of a ‘preview-detail’ textual pattern; it answers the question: ‘What,
roughly, is this text going to be about? What is its main point?’ and prompts the
question ‘What are the details?’. On a more general level, it corresponds to
Winter’s (1982) notion of ‘situation’.
In the hot news texts examined by Zydatiss (ibid.), the ‘abstract/
preview’ sentence of the text typically occurs with present perfect tense
on the main verb(s). Subsequent sentences then typically have past
simple form. Two British newspaper examples follow, with relevant
main-clause verb forms in bold type:

(5.1)
SAM DIES AT 109
The oldest man in Britain has died aged 109 — six weeks after
taking the title.
Sam Crabbe, a former sugar broker, from Cadgwith, Cornwall, did
not give up smoking until he was 98 and enjoyed a nightly tot of
whisky. He was taken ill just hours before his death.
Sprightly Sam became Britain’s longest living man when 112-year-
old Welshman John Evans died last month.
(Daily Mirror, 26.7.1990: 8)

(5.2)
INVASION OF THE CRAWLIES
Poisonous black widow spiders have invaded Britain by plane.
 5.2 A clear case of parallel: the narrative ‘situation-event’ pattern - 93

They stowed away in crates of ammunition flown from America

to RAF Welford, Berks.
AUS airman at the base near Newbury captured one of the
spiders in a jar after it crawled out of a crate.
(Daily Mirror, 27.7.1990: 19)

This pattern of present perfect followed by past simple is quite common

in short reports in British popular newspapers. It is immediately notice-
able that the pattern is not easily reversed or re-ordered. In British
English (though perhaps less so in American) the first sentences with past
simple would sound odd:

*The oldest man in Britain died aged 109 — six weeks after taking
the title.
*Poisonous black widow spiders invaded Britain by plane.

Although there is no inherent restriction on the use of past simple

without a temporal marker indicating separation from the now-present,
the only way we could rescue the anomalous versions of the opening
sentences is by inclusion of such a marker:

The oldest man in Britain died yesterday aged 109 — six weeks
after taking the title.
Poisonous black widow spiders invaded Britain by plane when a
cargo of bananas arrived from South America last week.

The original versions, with present perfect and no marker, dislocate the
temporal reference of the main event and signal it as an event ‘relevant
to now’. In the case of (5.1) the actuai time of Sam Crabbe’s death is not
revealed until the end of the text, and even then only by indirect
reference to a separate event (the death of John Evans). Not only are the
initial present perfect forms unreplaceable, but the whole sequence of
tenses cannot be reversed. We could interpret the pattern as the fronted
present perfect supplying a higher-order temporal frame (at the level of
the text), within which the individual events are realised in their normal
(unmarked) narrative/report tense: viz the past simple. The present
perfect, with its ability to signal ‘now-relevance’ for events in the past,
signals the now-relevance of the text itself, of all that is to follow. Each
detail or additional fact (or events and their orientations in narrative
terms) simply unfolds within the usual narrative conventions governing
chronological and non-chronological interpretation. The tense-aspect
94 + 5 Patterns of co-occurrence of verb-forms in English

choices, therefore, fulfil important textual and interpersonal functions

rather than purely temporal-ideational ones in the developing text. In
Winter’s terms, the power of the initial sentence to establish an over-
arching ‘situation’ plays a major role in the interpretation of the
relationship between it and subsequent clauses. Further support for this
view comes from (5.1)’s headline: as we move up to an even higher order
in the text (the general statement, for which the whole text provides the
particulars), we conventionally shift to another tense, the so-called
historical present (Sam dies), this form itself conventionally restricted to
the headline position.
In spoken discourse, we find a close parallel, though without the
separate option of a present tense ‘headline’. The narrative or report
preview/abstract often occurs in present perfect and the subsequent
details are told in past simple. The ‘newsworthiness’ element is still
present in the preview/abstract, and its function seems to be to signal
that (a) the report has now-relevance, and (b) the established situation
extends over subsequent clauses until transaction-termination is sig-
nalled (or negotiated among conversational participants). Some
CANCODE examples follow, with relevant items in bold:

(5.3)
[<S 01> is a young female, aged 24, whose father is a railway employee.
While she was a student she enjoyed free and reduced-cost rail travel.
Now that she herself is employed, she has lost her privilege travel. She
reports this to the other speakers.]

<S 01> I’ve started paying now

<S 02> |{inaudible}
<S 01> LAs soon as I started working | lost it, I’m
no longer dependent on dad and mum.
<S 03> lYou’re only you only get that
when you're a student Mary.
<S 01> It was while I was dependent on dad [<S 03> yeah] as soon as I’m
earning [<S 02> well] I've had my lot.

(5.4)
[<S 01> is recounting how she and her partner have experimented with
pasta-making. <S 02> and <S 03>, who are also partners, report their
experience doing the same.]
 5.2 A clear case of parallel: the narrative ‘situation-event’ pattern + 95

<S 01> We did this, erm, quite a lot of ravioli, didn’t we, but it was fiddly,
very fiddly, I imagine this is
<S 02> Yeah we've done that yeah
<S 03> We've done that ... we started off trying to use ravioli moulds
and then soon discovered that actually they’re more trouble than
they’re worth so we just
<S 02> We cut them out.

(5.5)
[<S 01> is encouraging <S 02> to make her choice from a restaurant
menu. He suggests scampi; she rejects it.]

<S 01> You don’t want scampi, no, oh you’re calorie watching are you.
<S 02> Yeah I’ve been I’ve been going to the weight-watchers but wait till
you hear this I went on, first time and I’d lost three and a half pound
[<S 01> yeah] and I went last week and I'd lost half a pound so I went
down to the fish shop and got fish and chips I was so disgusted
[<S 01> laughs] but I’ve been all right since.

Examples (5.3) and (5.5) are interesting as instances of symmetry in the

tense-aspect framework, beginning with present, perfect, going on in past
simple and then coming back again to end with present perfect. (5.5) is
also a complete narrative. The tense-aspect pattern fits well with the
Labovian (1972) notion of narrative initial abstract (the story ‘headline’)
and narrative-final coda, the coda being a ‘bridge’ from story-time back to
conversation-time. Both present perfect clauses may therefore be seen to
function at a macro-level, framing the text and establishing contextual
relevance. In (5.4), <S 01> is using typical narrative past tense to recount
what she and her partner did. <S 02> takes up the topic and flags it as a
new conversational focus, raising it to a more important status in the
conversation by his use of present perfect we’ve done, in contrast to
<S 01>’s we did. The most important question for all our examples (5.1) to
(5.5) is not ‘When did these events happen?’ but ‘How do these utterances relate
to one another as a report/narrative and how should the receiver interpret the
clauses that realise them in their context of utterance?’!
Narratives and reports of this kind are a good place to begin the
exemplification of the main argument of the present chapter, not only
because of the topic-initial phenomena which are parallelled in the
written and the spoken, but also because discourse analysts have long
been preoccupied with the correlations, consistent over large amounts of
96 + 5 Patterns of co-occurrence of verb-forms in English

data, between particular narrative functions and tense-aspect shifts.

Narrative analysts (of which Fleischman (1990) is the foremost example)
have examined the coincidence of so-called ‘historical present’ tense (i.e.
the use of present tense to report past events) and crucial shifts in the
narrative, e.g. to peaks and climaxes, from orientation to complicating
events, or, in the case of Johnstone (1987), to indicate shifting participant
roles in speech-reporting segments (for more on tense in speech re-
porting, see Chapter 8). In both spoken and written mode, tense-aspect
choices seem to play a significant role in creating the conditions for
interpretation of the relationships among sequences of clauses, and
patterns of such sequences have become conventionalised for narratives
within our culture.

5.3 Used to and would: another narrative pattern

We shall stay with narrative for this section in order to highlight another
fairly common pattern in spoken data, one which has not escaped the
attention of pedagogical grammarians (e.g. Alexander 1988: 235). This
pattern is found in sequences where speakers are reminiscing or are
reporting typical or recurrent events in the past. Once again we may
observe a preview or abstract clause establishing the general situation
(with used to), followed by subsequent detail or extending clauses using
would. Four examples follow. The two uses of past simple was in (5.7) will
be commented on below:

(5.6)
[<S 01> is recounting a story about a friend of hers who wanted to learn
to bake cakes.]

<S 01> She wanted to bake them herself and she never really knew how
and her gran always used to bake cakes and she’d go and watch

(5.7)
[<S 01> is recounting how he and his wife, <S 02>, were involved in a
market-research exercise where the research organisation tapped into an
electronic machine on which he and his wife recorded their weekly
purchases. This was done automatically through the telephone line,
usually at night]
 5.3 Used to and would: another narrative pattern - 97

<S 01> Cos it used to ring about three o’clock in the morning, you used
to go down and answer it and there was no-one there.
<S 02> And that was the computer.
<S 01> You'd hear beep beep beep beep.

(5.8)
[Later in the same conversation; <S 03> is <S 02>’s uncle|]

<S 02> They used to you know ring up early hours of the morning, well
you would, the phone wouldn't ring they’d ring that computer.
<S 01> And they read it.
<S 03> Yeah.
<S 01> And it'd go through the phone.

(5.9)
[<S 01> tells a ghost story round the dinner table.]

<S 01> When I lived in Aberdeen years ago erm we were in a cottage in
the country my then wife and I you know and erm the people that
lived there before used to see apparitions.
<S 02> Oh.
<S 03> Did they.
<S 01> Yeah ten o’clock on a Friday night regularly they would hear
somebody and they'd be sitting in the living-room watching telly
and at ten o’clock every Friday they'd hear someone walking up the
stairs.
<S 03> Yeah.
<S 01> They’d go out there and there’d be nobody there you know.

The whole used to + would sequence may function as the ‘orientation’

(again in Labov’s (1972) terms), i.e. background concerning time, place,
situation, etc., for a narrative of a particular, one-off event or set of
events, as in (5.5), and in (5.10):

(5.10)
[<S 01> is reminiscing about his first work experiences; Canton is a
district in the city of Cardiff, UK]

<S 01> And er I got a job with an Irish milkman and he had er a pony
and cart [<S 02> yeah] and his stables was in the lane at the back
98 + 5 Patterns of co-occurrence of verb-forms in English

of Albert Street in Canton where I [<S 02> yeah] lived [<S 02> yeah
yeah yeah] and I used to drive this horse and cart and deliver his
milk ... the only snag was he was rather er fond of the booze
[<S 02> yeah] and of course as he got paid his money it would
go across the bar
<S 02> lAcross the bar yeah yeah
<S 01> lAnd one Saturday I, as a matter of fact I’ll tell
you my wage was twelve and sixpence a week and I was up at half
past four in the mornings with his milk as you know [<S 01 my
goodness yeah yeah] and on one Saturday he never had the money to
pay me...

Again we may note a problem with reversibility in this pattern. Used to

contains within it the meaning of ‘not true any more’ (a point made by
Downing and Locke 1992: 364), and can stand without a time adjunct.
Used to can refer to states as well as habitual events in the past (Swan
1995: 633),2 and thus is the ideal candidate for the ‘scene-setting’
function in narrative, where stories often commence with a ‘that was the
state of things then’ framework. Would, on the other hand, needs to be
anchored temporally, usually by the co-presence of a time expression
(Quirk et al 1985: 228-9), with wording such as:
When we were involved in the survey the phone would ring at three o’clock
in the morning.
Quirk et al (1985: 228-9) note this restriction but, surprisingly, con-
sidering that their grammar does use spoken sources, confuse the issue
by suggesting that past habitual would is more formal than used to. Our
spoken data show both forms occurring in the same, highly informal
discourses, with functions which seem to have nothing to do with
formal/informal distinctions. Used to clearly has the potential to situate
the discourse in a way that would alone does not. This seems particularly
true of (5.10), where used to occurs in the story-orientation, followed by
more detailed elaboration. Such an observation underlines the above-
clause function of the initial verb and, once again, enables us to shift the
emphasis away from a purely temporal-ideational interpretation to a
‘staging’ function with textual and interpersonal implications.
It would appear uncontroversial to suppose that the used to + would
sequence should appear in the written mode too, and we do find it in
literary narrative, as these examples from Sean O’Faoldin’s novel Bird
Alone show:
 5.3 Used to and would: another narrative pattern - 99

(5.11)
[The author is describing a man called Christy Tinsley. Note that the
occurrence of used without to is an acceptable Hiberno-English form]

He came out of gaol that May ... we used meet from time to time
and go wandering out the fields or around the city. Always he
would be chewing the boiled sweets that he loved...
(Oxford University Press edition, 1985: 284)

(5.12)
[The author is describing how he fought against the desire to see a
female with whom he enjoyed a fraught relationship.]

I used try to interpose other things between us; wandering and

drinking with my grander, and he hinting his sympathy and
trying to plot for me; I would even turn to praying when the asp
hurt cruelly. (ibid.: 163)

But worth noting here also is another pattern that occurs in literary text
and for which some corresponding examples occur in our spoken corpus.
The pattern is used to followed by simple past tense form, as in these
examples from Joseph Conrad’s Under Western Eyes; the form greeted has a
habitual meaning in the context of the story:

(5.13)
[The author is describing two ladies.]

I used to meet them walking in the public garden near the

university. They greeted me with their usual friendliness but I
could not help noticing their taciturnity.
(Higdon and Bender, 1983: 105.23)

(5.14)
Some of them used to charge Ziemianitch with knowing
something of this absence. He denied it with exasperation;
(ibid.: 272.16)

In spoken example (5.7) above, there was some evidence for this same
pattern, with two past simple occurrences (was/was) between used to and
would. Both patterns (used to + would and used to + past simple) are also
found in example (5.15):
100 + 5 Patterns of co-occurrence of verb-forms in English

(5.15)
[<S 01> is talking about a female cat the family once owned.]

<S 01> But er em Mrs Baker used to look after her and [<S 02> mm] and
so on but when, as soon as I arrived home [<S 02> yeah] and it, I put
put the car away in the garage you know [<S 02> mm] and er and
when I was walking halfway down from the garage to the house see
erm she used to [<S 02 mm mm] run up there you know yes up up
there she’d run up and make sure they’re [<S 02> mm] like that and
she’d walk down with me to the house like you know [<S 02> yes]
[<S 03> mm] [<S 02> ah] I used to give her a little bit of something to
eat and [<S 02> yeah] and then er if er ifI had a letter, take up a
letter up to er to post [<S 02> mm] and Id say to her, see show her
the letter ‘Coming up the post now?’ see.

Used to still seems to have its situating function and the subsequent past
simple verb forms must still be interpreted as repeated events, as with
those preceded by would. But we note one difference here which is
relevant: the past simple forms occur in subordinate clauses introduced
by when / as soon as, not main clauses as in (5.7), (5.13) and (5.14). The
pattern used to + past simple in main clauses is rare in the spoken corpus
compared with the much higher frequency of used to + would. The
difference may simply be one of conventionality, or there may also be
something characteristically literary in the choice of the more implicit
expression of habitualness that past simple in the main clause conveys.

5.4 Narrating the future: further problems concerning

parallel functions

In section 4.4, we looked at the be to construction with future meaning,

and noted its extremely low frequency in informal spoken data. Here we
shall consider its frequent use in journalism and what parallels there
might be in the spoken mode for the functions it performs in the
written. McCarthy and Carter (1994: 126, 129) draw attention to the
grammatical pattern common in British popular newspaper reports of
future events, where a preview-detail text pattern, parallel to the ‘hot-
news’, past-events text, claims the reader’s interest and attention in a
headline or text-initial sentence(s) employing the be to form for the main,
‘hot-news’ future event(s). Subsequent sentences, expanding on the
preview, employ the will future. Be to, just like our other examples so far,
 5.4 Narrating the future: further problems - 101

seems to be conventionally used in this kind of journalistic report to

establish general situation and newsworthiness. As with the present
perfect, be to anchors the event in a context of present relevance but
projects into the future. Two examples of this are (5.16) and (5.17):

(5.16)
ELECTRICITY CHIEFS TO AXE 5,000
Five thousand jobs are to be axed by electricity generating firm
National Power, it was announced yesterday.
Smaller power stations will close but bosses pledged no
compulsory redundancies over the next five years.
(Daily Mirror, 27.7.1990: 2)

(5.17)
‘KING’ ARTHUR’S BATTLEGROUND TO BE WIPED OFF THE
FACE OF THE EARTH
The battleground where Arthur Scargill and 7,000 miners took on
the government is to be razed.
Orgreave coking plant - scene of Britain’s worst industrial
violence - will be bulldozed and turned into a landscaped wood.
(Daily Mirror, 27.7.1990: 6)

Any informal, conversational report of the same events would almost

certainly eschew the be to form. Indeed, in one million words of the
CANCODE corpus, be to future simply never occurs in ordinary casual
conversation. Where it does occur is in formal and semi-formal contexts
such as business meetings, where three of the four occurrences in
CANCODE are to be found:

(5.18)
[British National Health Service business meeting]

<S 01> You’ve probably heard that er a lot of the recommendations that
are coming out of central government now about complaints
handling are precisely about the the jargon word empowering the
staff [<S 02> mm] to be able to deal with things at the front line like
that ... they’ve got a big task on their hands if this is to be in place
by April nineteen-ninety-six
102 + 5 Patterns of co-occurrence of verb-forms in English

(5.19)
[Publisher’s business meeting]

<S 01> Right yeah sorry I thought you were I mean for turning it round
[<S 02> no no no no] and getting it
<S 02> LOh no oh right well no but it’s
taken two months to do that.
<S 03> Rob is to look at it and Ann Pascoe to look at it and
formal comments to be collated and sent back to David.

(5.20)
[Publisher’s business meeting, as (5.19)]

<S 01> Then there’s a couple of things written in pencil there in the
schedule for the estimate for, can’t remember what they are.
<S 02> Which one’s that.
<S 01> lWhich I am to check to confirm them.

Quirk et al’s major grammar makes no reference to any restrictions on

the occurrence of be to nor suggests any parallel informal spoken form
(1985: 217), which is the kind of problem that is likely to arise if the
spoken language is not rigorously examined on its own terms. Celce-
Murcia (1991), however, offers a possible spoken parallel in the occur-
rence of be going to (initial) plus will (subsequent) sequences when
speakers are reporting future events. In other words, we have, once
again, the suggestion of an over-arching tense-aspect framework which
establishes situation and signals newsworthiness and/or relevance, ex-
tending over a series of following clauses. In the CANCODE corpus, there
is some limited evidence for this pattern. Examples (5.21) and (5.22) offer
some support for the argument but also raise further interesting
questions that need to be addressed:

(5.21)
[The three speakers are part of a group of eight people celebrating New
Year. It is coming up to midnight and they discuss the problem of
knowing exactly when it is 12 o’clock.]

<S 01> What are we going to get the actual chimes from Red have we got
something that will tell us exactly.
<S 02> Anything on TV.
<S 03> Well my watch says it’s coming up anyway.
<S 02> There’ll be something on TV.
 5.4 Narrating the future: further problems - 103

(5.22)
[<S 02> is making punch for a party. <S 01> is watching him add each
ingredient.]

<S 01> [laughs] what you going to do with that.

<S 02> Oh it'll go in in a minute I can taste it as I go along and then add
the same amounts again.

On the face of it, the pattern of be going to + will occurs in examples (5.21)
and (5.22). In (5.21), the over-arching situation is ‘establishing the correct
time’, and the predictions about the television provide the ‘sequence’
(Winter, 1982: 2). But to be more precise, the sequence is also a ‘problem-
solution’ sequence (see Hoey 1983; see section 2.5 for a brief summary of
the components of a typical ‘problem-solution’ pattern), and as such
must be distinguished from the preview-detail reports of the newspaper
examples (5.16) and (5.17). This may not invalidate the potential for be
going to and will to co-operate at the macro-level; it simply suggests that it
would be too simplistic to pair the spoken sequence functionally with
our written examples using be to and will. Be going to + will as a framework
for problem-solution sequences may well be a common pattern in spoken
language, and is intuitively appealing. Hoey (1983: 76-7) has already
noted the correlation of present-perfect + present tense sequences for the
‘response’ segments of problem-solution reports (i.e. those which look
back at problems now solved); a similar correlation might be posited for
be going to + will with current problems and their projected solutions, at
least in informal spoken mode. However, more contrastive spoken and
written data would be needed to validate the claim.
Additionally, be going to + will are used in the spoken mode for
‘announcements’, with going to signalling newsworthy items of informa-
tion. (5.23) and (5.24) below illustrate such contexts:

(5.23)
[<S 01> is a health-service worker informing <S 02> about a new ‘patient’s
handbook’ that they are producing.]

<§ 01> I’m sort of chairing the working group em [laughs] a document
that that it’s official name is going to end up being something like
Patient Handbook [<S 02> yeah] but at the moment it it’s lovingly
known as the alternative Gideon [<S 02> laughs] you'll find it on the
locker next to the bed or something yeah [<S 02> laughs] that, well
104 + 5 Patterns of co-occurrence of verb-forms in English

that’s right yeah I mean that’s literally where we want it to be it’s

going to be in every clinic

(5.24)
[Speakers are discussing a currency devaluation.]

<S 01> It’s the import bill’s going to rise usually it'll double overnight
cos exports’ll be half the price they are now [<S 02> uh huh] so
they'll have to export a lot more to get the same amount of money
in [<S 02> mhm] so their costs have risen but their revenue’s fallen

In the New Year’s party example (5.21) above, the speakers co-operatively
negotiated the problem of establishing exactly when it would be mid-
night by discussing current possibilities for action, and this raises the
interesting point that grammatical patterns in the spoken mode may be
created by more than one participant. Patterns do not only extend across
clauses; we should not be surprised to find them across turn-boundaries
too. The punch-bowl text (5.22) further underlines the present orienta-
tion of be going to (<S 01>’s question What you going to do with that? could
be paraphrased as ‘What is your current determination for action?’), in
contrast to the point-in-the-future reference of will. (5.23) and (5.24)
appeared to have a ‘headlining’ function. (5.21) to (5.24) together would
seem to justify claiming a situating potential for be going to, even if not in
exact parallel to be to future. Indeed, Quirk et al. (1985: 214) talk of the
semantics of be going to as indicating ‘future fulfilment of the present’,
thereby suggesting a ‘now-relevance’, not unlike that posited for present
perfect tense.* Moreover, they tantalisingly claim that be going to tends
not to be repeated in a text referring pervasively to the future’ (ibid.: 218),
giving a (concocted?) example of a spoken weather-forecast sequence
with initial be going to plus sequential wills. While this may,be so in
weather forecasts, repeated be going to is by no means eiipdckibiel as
example (5.23) shows:

(5.25)
[[wo young women are getting dressed to go out for an evening on the
town.]

<S 01> What are you going to wear then.

<S 02> I’m just going to put my jeans on and a black top I'm not going
to think about it Sha don’t let me think about it I’m going to be
really boring.
 5.5 Bringing the arguments together - 105

It is possible to interpret (5.25) as a stylistically marked version of what

otherwise might be a be going to (situation) + will (sequence) pattern. This
interpretation would rest on the notion of conventionality, rather than
determinism, in the kinds of patterns we have been examining. But there
is certainly little disagreement among those who have closely scrutinised
the be going to versus will choice that be going to projects from a present
orientation into the future, whilst will does not (see especially Lakoff
1970; Aijmer 1984a; Haegeman 1983a: 65; 1983b; 1989; Binnick 1991:
389-90).

5.5 Bringing the arguments together

When we look back over the examples of both written and spoken
patterns, a common thread emerges. The situating sentences and utter-
ances, as we have called them, characteristically contain verb-forms
whose semantics have something to say about ‘now-relevance’. Even used
to, we have argued, implicitly carries a ‘not true now’ meaning which sets
it apart from would. This link across the forms enables us to posit a
general phenomenon at the discourse level whereby the verb-forms
under investigation may be distinguished by their ‘situating’ versus
‘sequencing’ potential, and their patterns of co-occurrence may be
brought together under these general headings:

Situating Sequencing

present perfect past simple

used to would
used to past simple (literary examples)
be to will (written and formal spoken)
be going to will

The patterns do not have any deterministic length, and may be short,
with just one or two sequencing verb phrases, or they may extend over
longer stretches of discourse. They are important, though, as illustrations
of the potential for describing conversational ‘episodes’ of talk in the
sense of functional units of varying length at levels between the clause/
sentence and the whole conversation (Van Dijk 1982; Benwell 1996),
some of which be identifiable by the occurrence of discourse
may
markers, some by intonational criteria. Some of the spoken examples we
106 + 5 Patterns of co-occurrence of verb-forms in English

have looked at also raise interesting questions about the correlation of

grammatical choice and adjacency pairs and exchanges.

5.6 Concluding remarks

Winter’s work on contextual grammar supports the view that co-occur-

rence, pattern and sequence of grammatical forms has fundamental
significance in the interpretation of consecutive clauses in written texts.
What I have tried to do in this chapter is to demonstrate that such
sequences occur in spoken data too, and that there is every reason to
suppose that similar (even if not always the same) functions are being
performed at the macro-level by the particular juxtaposition of tense-
aspect choices and the regularity of the ordering of particular elements,
but that there may be differences in realisation in the spoken and
written. What all the patterns we have noted in this chapter seem to
have in common is the creation of a situation anchored to the moment
of utterance, followed by details or events which are governed by that
situational framework. The situations and their sequence sentences/
utterances operate within different temporal frameworks, whether past
or future, and represent textual units, both in the written and the
spoken mode.
Such contextualised grammatical statements are of great relevance to
language teachers for the following reasons:

e They assist in more clearly distinguishing notoriously problematic

choices for learners (e.g. going to versus will).
e They make the study of grammar in texts more truly related to how
the grammar builds text itself, rather than just picking out examples
of tense/aspect usage in texts for atomistic, local interpretation.
e They can be used to highlight fundamental, generalisable meanings of
tenses and aspects in a way that lists or rules, sub-rules (i.e. sub-
divisions of rules, as are often found in big grammars) and ‘exceptions’
often fail to do.
e They correspond to everyday functions that learners are likely to want
to realise in their use of the target language, such as narrating,
reporting, talking about the future.
e They can assist us in distinguishing what is different between spoken
and written grammars, and what features they share. Throughout, I
have argued that we cannot simply assume where the differences or
 5.6 Concluding remarks - 107

overlaps will be, or, even worse, assume that written-based descriptions
can be transferred wholesale to the spoken mode.
They can assist us in cross-linguistic comparisons. English is not
unique with regard to the kinds of functions we have discussed for
some of the verb forms. For example, other languages with future be
going to equivalents seem to use them with a ‘present-rooted’ function:
see Haegeman (1983) for a comparison between the English, French
and Dutch cognate forms; also on French see Wales (1983); on Spanish,
Bauhr (1992). It may be possible to tap discoursal knowledge of the
learner’s L1 as a foundation for discussion of similar phenomena in the
target language.

This chapter has dealt with grammar in the written and spoken
language, pointing out parallels and differences between the two modes
of communication by observing grammatical phenomena operating at
the level of discourse. The next chapter will shift the focus, but not the
approach, to questions concerning vocabulary.

Notes

il Hopper (1979) shows how the choice between perfective and imperfective
aspect in Russian signals the relation between ‘new’ narrative events and
already established ones (i.e. a discourse-level function), and I should certainly
not want to suggest that the kinds of discourse-level patterning in verb tense/
aspect choice discussed in this chapter are a speciality of English (see also
section 5.6).
This may be seen as part of a more general overlapping between the expression
of states and habits found in the English verb-phrase (see Brinton 1987).
3 Joos (1964: 139ff) sees be going to as a mirror image of the present perfect, in
that both emanate from the present. Fleischman (1983) reiterates this, seeing be
going to as the ‘prospective’ mirror-image of the present perfect’s ‘retrospective’
viewpoint.
6
Vocabulary and the spoken language’

6.1 Introduction

In this chapter, we shift the focus from grammar to lexis, for if the
central thesis of the book is to be adequately supported, then not only
grammatical choices should be the subject of our investigation of the
ways in which discourse- and genre-constraints underlie the choices at
the formal, lexico-grammatical level. Our central concern with the way
discourse-level phenomena such as goals and relationships motivate
generic activity must have its reflection in lexical choice.
Work on the patterning of lexis in written text, such as the studies of
lexical cohesion associated with Halliday and Hasan (Halliday and Hasan,
1976; Hasan 1984) and the study of the significance of multiple ties
between words by Hoey (1991a), have been in sharp contrast to the rather
scant amount of research into the kinds of vocabulary patterns that
occur in everyday spoken language. The absence of a proper body of
studies of vocabulary represents a serious drawback for anyone wishing
to pursue the central argument of this book: that applied linguistics and
language teaching stand to benefit greatly from discourse-based lan-
guage descriptions and attention to real spoken data. Grammatical
structure interpreted from a discourse viewpoint does not seem to have
suffered the same lack of attention, as Chapters 4 and 5 of this book have
attempted to demonstrate. It is thus a pity that vocabulary should be so
often seen as beyond the purview of discourse analysts. I have attempted
to fill this lacuna to a limited extent and have carried out small-scale
research into patterns of lexical reiteration and relexicalisation in
conversation (see McCarthy, 1988; McCarthy, 1991: ch. 3; McCarthy,
1992a; McCarthy and Carter, 1994: ch. 3). Other researchers have looked
at lexical repetition (e.g. Persson 1974, who uses spoken and written
data; Schenkein 1980; Blanche-Benveniste 1993; Tannen 1989; and most
notably, Bublitz 1989), and at formality in vocabulary choice in spoken
language (Scotton 1985; Powell 1992), but their efforts represent only a

108
 6.1 Introduction - 109

relatively small body of literature in comparison with that now available

on the role of grammar in spoken interaction (see, for example, Ochs et
al. 1996 and the copious references given therein). In this chapter we
shall consider a number of general features of vocabulary in informal
spoken language and point to areas of significant interest to applied
linguists and language teachers concerned with achieving a clearer
understanding of the character of spoken language and the creation of
pedagogically relevant models.
When looking at vocabulary in conversational language (as opposed to
single-authored written text), certain key points have to be borne in mind:

1 Speakers operate under the constraints of ‘real-time’ planning; careful

composition and selection of words is the exception rather than the
rule.
2 More than one speaker normally contributes to the vocabulary which
occurs.
3 Roles vary in conversation; speakers may not be participating as
equals, and one or more speakers may dominate vocabulary selection
(see Thomas 1984).
4 Roles may shift as the conversation progresses; this may be reflected in
vocabulary choice.
5 Topic - what is being talked about - is neither pre-determined nor
singularly defined, but shifts and develops, often without sharp
boundaries between topics. Vocabulary choice reflects such shifts and
contributes to connectivity and ‘shading’ between topics (i.e. moving
from one topic to another without a sudden jump, making connec-
tions between topics, etc.).
6 Speakers have no automatic rights to have their topics addressed, but
have to negotiate them.
7 Spoken language is usually more implicit (Chafe, 1982) and situation-
dependent. The lexical density of spoken texts may differ considerably
from that of written texts on the same topics (Ure, 1971; Stubbs, 1986).
8 Conversation contains a large amount of vocabulary whose function is
mainly ‘relational’ or ‘interactional’ (i.e. in the service of establishing
and reinforcing social relations), rather than ‘transactional’ (i.e. func-
tioning principally in the transmitting of information, goods or
services). Issues such as convergence and communicative accommoda-
tion between speakers (see section 2.4) are thus relevant to the study of
lexical patterning. (See glossary for definitions of terms.)
110 + 6 Vocabulary and the spoken language

9 Conversation contains a significant number of prefabricated lexical

expressions which facilitate fluency and which are often idiomatic in
structure and meaning (see section 3.2.1; Chapter 7).

With these factors in mind, I propose here to look at a number of pieces

of natural conversational data and some features of their vocabulary.

6.2 Speech function and lexical density

(6.1)
[Five friends, all university students, three female: <S 01>, <S 04>, <S 05>;
two male <S 02>, <S 03>, around the dinner-table.]

<S 01> Well, I’ve got the other camera so ... if Dave ... then we can load
that and have lots of jolly photographs in the pub.
<S 02> Mm.
<S 03> Mm.
<S 02> Mm.
<S 04> [to <S 02>] It won’t rewind.
<S 02> What ... the batteries are flat.
<S 04> lThey’re not, they’re brand new I
put them in the other day maybe it’s just the way it rewinds.
<S 02> That’s the film speed type.
<S 04> I may have over taken one.
<S 02> Well it should have recovered by now.
<S 05> Oh he’s licking my feet.
<S 02> What? [laughs] it’s only a dog.
<S 05> It’s a dog.
<S 01> Woof.
<S 04> Right that should be it, well Amanda.
<S 01> Yeah?
<S 04> Why don’t I put your film in here.
<S 01> Okay and I can
<S 04> Lor put those batteries in the other camera?
<S 01> I can’t take it out half way through though and
<S 04> lWell have you
started it? What is it then a thirty-six? ... Well why don’t I put my
batteries in your camera.
<S 01> Yeah ... Idon’t ...1 mean I don’t mind putting my film in there.
<S 04> No well yeah if you want to use the film at some other time.
 6.2 Speech function and lexical density + 111

<S 02> I'm sure it’s got no batteries in, it feels extremely light to me [to
<S 04>] just put the batteries in that camera.
<S 04> Yes that’s what I’m doing.
<S 02> Yeah.?

This is a rather typical example of the genre of ‘language-in-action’ (that

is to say language being used principally in support of actions that are
taking place at that moment, in this case solving a problem with a
camera). As such, the text is lexically quite ‘light’. Of its 227 orthographic
words, only 56 (just under 25%) are lexical nouns, verbs, adjectives or
adverbs with content meaning accounting for the topic of solving the
camera problem. If we examine these 56 tokens (individual occurrences
of words), we find there are 40 types (i.e. different vocabulary items); for
instance, camera occurs four times, batteries four times, put six times, and
film four times.
Words which are not fully ‘lexical’ in the sense of content nouns,
verbs, adjectives and adverbs include:

e 27 items (11.9% of the total words) that refer exophorically to the

immediate situation, such as ‘the other camera’, ‘that’s the film speed
type’, ‘your film in here’.
e 13 discourse markers (Schiffrin, 1987; Fraser, 1990) (accounting for
5.7% of the total words), small words and phrases that contribute to
the management and development of the discourse and which perform
important structural and interactive functions such as marking phases
in the activity and projecting the assumed state of shared knowledge.
Examples are: well, right, I mean.
e 9 items (3.9% of the total words) that mark modality, expressing
degrees of certainty about the propositions stated by the content
words, and expressing the speakers’ perceptions of what ought to
happen or what is a desirable state of affairs. Examples are: can, may,
maybe, should, I’m sure.
e A large remaining number of other function words such as preposi-
tions, determiners, particles, conjunctions, etc..

Does the fact that there are only 40 different content words (repre-
senting 25% of the total orthographic words) mean that the vocabulary
load for a learner wishing to achieve this kind of naturalness in this
kind of situation is likely to be light? Ure’s (1971) study suggested that
an average of 40% lexical density could be expected across a wide range
112 + 6 Vocabulary and the spoken language

of written and spoken texts, and that a figure less than 30% is lexically
light. If vocabulary to be taught/learnt is thought of in the traditional
sense of content words, then the answer would be ‘yes’ to the question
of lightness for language-in-action situations such as this one. The 40
types contain many basic-level words (put, want, day, dog, start, etc.) and
the more topic-specific words are easily arrangeable into interrelated
sets (camera, film, battery, etc.) which can be taught in a conventional
manner. However, there is no doubt that the words referring directly to
the immediate situation need to be carefully considered. Notorious
problems arise between languages over the usage of items such as
demonstratives; deixis with this and that in English does not always
correspond with cognate forms in other languages (e.g. Spanish/German]/
Danish; see also section 4.2). Furthermore, if we consider the relational/
interactional dimension, then the discourse markers, the modal items
and such features as intensifiers/downgraders (e.g. just, only) will also
enter into the learning equation. These may require a methodological
approach quite distinct from the ‘presentation-practice-production’ one
(the ‘three Ps’) traditionally used for teaching vocabulary. Without this
relational/interactional dimension to vocabulary learning, we run the
risk of an over-emphasis on the transactional features of spoken
language and of creating a ‘reduced’ personality for the learner, who
may well be able to achieve transactional goals but who may be unable
to project his/her personality and create appropriate relationships with
interlocutors (for a discussion of the term ‘reduced personality’, see
Harder 1980).

6.3 Repetition, relexicalisation, negotiation of topic

Extract (6.2) raises different issues connected with its vocabulary:

(6.2)
[<S 01> is father in the family. <S 02> is his daughter, <S 03> his wife.
They are getting dressed for a family wedding. <S 01> comes in holding
the coat of his suit.]

<S 02> That looks very nice put it on and let’s have a look at you.
<S 01> I don’t like the two buttons I didn’t know it had two buttons I
thought it had three.
<S 03> Well it’s the style of the coat Ken.
 6.3 Repetition, relexicalisation, negotiation of topic + 113

<S 02> Nick’s has only got two buttons.

<S 03> Lit’s a low cut.
<S 01> LAll right?
<S 02> lVery nice.
<S 03> Lit’s
beautiful.
<S 02> Lovely lovely.
<S 01> Does it look nice?
<S 02> Yeah it goes very well with those trousers there’s a colour in the
jacket that picks up the colour in the trousers.
<S 03> Them others he wears are striped but they clashed, too much
alike
<S 01> lTwo different stripes.
<S 03> LBut
not matching each other if you understand what I mean.
<S 02> Yeah yeah ... yeah.
<S 01> Lit’s all right then eh?
<S 02> It’s very nice Dad it looks very very good.
<S 01> I don’t like the I like three buttons you see.
<S 03> Ken it’s the style of the coat.

This extract underlines the point made in the prolegomena to our

examination of data: that topics are negotiated in spoken discourse.
While the overall topic of the father’s jacket is agreed
among the
participants, it is clear that the father himself would like to launch a
discussion of the problematic nature of its having only two buttons (he
opens with the subject and tries to resurrect it again at the end). The
mother and daughter apparently wish to dismiss such talk and are bent
on persuading the father that the jacket looks fine. Where participants
are trying to agree a topic, we often find a significant variation between
exact repetition of vocabulary and what may be called ‘relexicalisation’,
where content is re-cast in different but near-synonymous words. Firstly,
it is important to note that exact repetition is not always pragmatically
appropriate; the following concocted exchange would be considered by
most people as odd:

(6.3)
<S 01> Hi! Freezing cold today!
<S 02> (with exact same intonation) Hi! Freezing cold today!
114 + 6 Vocabulary and the spoken language

<S 02> is much more likely to say something like Hi! Yes, bitter! or Yes, it is
freezing! (see McCarthy 1984 for further discussion). Exact repetition
(including syntax, lexis and intonation) often suggests a non-increment
to the topical progression of the discourse (increments are things which
push the topic forward; a non-increment deliberately stalls it), and can
be interpreted as staying with one’s present position in the talk, of a
refusal to converge or communicatively accommodate (Giles et al 1991),
with whatever local implicatures that may carry, which can be seen in
the wife’s repetition of It’s the style of the coat. This holds up the discourse
from reaching its goals, in this case, those of the genre of ‘checking/
approving appearances’, and places a strain on the relationships of the
participants. The socially co-operative norm in non-ritual, non-copying
adjacency pairs is typified by (6.4) below, where repetition is accompan-
ied by relexicalisation (as indicated by the items in bold face), and
convergence is achieved:

(6.4)
[Iwo middle-aged male teachers are gossiping about a female ex-
colleague.|

<S 01> There was this guy that she was really madly in love with that
went on and ended up working on an oil rig somewhere.
<S 02> Really.
<S 01> Oh yes she really was really loyal, very struck on him.
<S 02> Smitten.
<S 01> Smitten with him, had he, had he asked her at that particular
time er I think she would have probably married him.
[later in the same conversation]
<S 01> And this is going back to the time when she was living in oh
<S 02> Southampton.
<S 01> Southampton yeah.
<S 02> In that big house.
<S 01> In that huge house, I mean she’s got an awful lot to offer,
tremendous amount, I mean what a personality.

In (6.2), it is precisely the cases of exact lexical repetition which focus

most sharply the stubborn stance of the conflicting parties and the
frustration of the goals of the episode. This may be contrasted with the
more solidary, supportive and convergent exchanges, where repetition
gives way to relexicalisation:
 6.3 Repetition, relexicalisation, negotiation of topic - 115

<S 01> All right?

<S 02> Very nice.
<S 03> It’s beautiful.
<S 02> Lovely lovely.
<S 01> Does it look nice?

Here the speakers vary one another’s lexis and also pick up one another’s
words (note how <S 01> takes up nice: repeating across three speaker-
turns). Notable too is the occurrence of paraphrase (which is, by defini-
tion, a type of relexicalisation, but which affects longer stretches of
language than single words or phrases) within speaker-turns, rather than
exact repetition. This can be seen as a co-operative gesture by the
speakers, ‘explaining themselves’ for their listeners:

<S 02> Yeah it goes very well with those trousers there’s a colour in the
jacket that picks up the colour in the trousers.
<S 03> but they clashed too much alike ... but not matching each
other...
<S 02> It’s very nice Dad it looks very very good.

What we may observe in (6.2) and (64) is the importance of the

interweaving of lexical repetition and relexicalisation. The ability to vary
one’s lexis while still saying more or less the same thing pushes the
discourse forward and gives out important interactional signals. It is a
fundamental feature of lexical competence and is one of the basic
characteristics of vocabulary patterning in everyday talk. Learners often
need practice in variation, but they cannot be expected to vary their
choice of word if their word-stock is impoverished. Here lies the opportu-
nity for motivated vocabulary building, bringing abstract notions such as
synonymy, antonymy and hyponymy to life. If we can see how these
lexical relations are exploited interactively in real contexts, then there is
a good reason for learning synonyms and other sense groupings; if we do
not have them available as speakers for just the kinds of everyday
functions highlighted here, we may end up sounding very unnatural
indeed and may become victims of socio-pragmatic breakdowns in
communication, even though, on the purely transactional level, we may
survive.
The ability to vary one’s lexis applies not only across speakers, it will
be noted. When speakers are unsure of the best wording or are offering
particular meanings for words in context, they will often explore various
116 + 6 Vocabulary and the spoken language

lexical possibilities, including negated opposites, in an attempt to put

across what they want to say:

(6.5)
[<S 01> is describing the landscape of his part of England to a Welsh
person who has never been there.]

<S 01> It’s flat you know it’s not er hilly like Wales but [<S 02> mm] you
get used to that strangely enough after a while, I mean it’s not as
flat as a pancake it’s kind of undulating and lots of little villages

Exploiting lexical alternatives in this way is only possible if one has an

adequate vocabulary of synonyms and antonyms, and the ability to
retrieve them in real time. If possible, the vocabulary class, especially at
the advanced level where many words are known but often in an
atomised way, should offer activities where learners have the opportu-
nity to string alternatives together to produce meanings. So often, the
opposite is the case, with learners being unable or unwilling to utter
words unless they feel they are the ‘right’ ones, and with so many
exercises demanding just one, right word as the answer. Dialogue games
where repetition is ‘forbidden’ can force learners to look for alternatives,
as Can games where contradiction or agreement (without repetition) is
required.

6.4 More than one speaker: the listener’s contribution to

lexical patterning

In this section, we shall consider an area often underplayed in discus-

sions of spoken language: how listeners behave. We shall use a typical
situation where the contribution of listeners comes most to the fore -
everyday storytelling. (6.6) is a typical professional anecdote of the kind
exchanged among friends and colleagues:

(6.6)
[A woman is telling a funny story about a conference she attended|]

<S 01> Well, the conference theme was the 1990s and they did this talk
and there was this amphitheatre that seats 2,000 they started off this
sort of slide and sound sequence
<S 02> Ah Son et Lumiere.
<S 01> Son et Lumiere.
 6.4 More than one speaker: the listener’s contribution + 117

<S 02> I see and they zapped it to you.

<S 01> It starts off with this tiny black, we’re all in the dark you see and
tiny little ... and we hear this click-click and you see this little
coloured pattern and this coloured pattern gets bigger and bigger.
<S 02> What was this projected by, a movie projector or video or what?
<S 01> No it was ... erm ...a slide sequence but it was one after another
... anyway the very funny bit was that the sound went.
<S 02> [laughs] That’s the trouble when you rely on technology.
<S 01> Yeah and that was very very funny and we're all sitting there in
the dark and this picture thing going on you see, obviously going
ahead of the sound and him saying "Why can’t we hear any sound?
Why is there no sound, technician?’ you know, chaos and a great big
smile on everybody’s face.
<S 02> Especially the other companies.
<S 01> And anyway they got it going again and you heard this ’click-click’
again and this coloured thing suddenly reveals itself to be the Berlin
Wall with people on the top of it.
<S 02> That’s a novel idea. Was Pavarotti singing as well?
<S 01> Oh yeah yeah.

Much of the literature on spoken language focuses somewhat unduly on

what speakers do, and those whose main role in the discourse is to listen
to and support current speakers are often neglected (see, however,
Bublitz 1988; McGregor and White 1990; Duranti 1991). One good
context in which to examine the role of listeners is oral narrative, since
the speaker who takes the role of storyteller has the main speaking role.
Those on the receiving end of the story have reduced turn-taking rights
and may only interrupt at specified places (Houtkoop and Mazeland
1985). Nonetheless, listeners are by no means expected to remain passive
and silent throughout.
It may be helpful first to view the anecdote in (6.6) in terms of Labov’s
(1972) generic model of narrative, one of the most important components
of which is evaluation (see section 3.2.3). The story must be made
interesting, funny, shocking, appalling, or whatever, in order constantly
to keep at bay the possible objection ‘well, so what?’ on the part of the
listener(s). The teller therefore must work hard to evaluate the story. At
the lexical level, this may include choices of expressive vocabulary,
intensification, hyperbole, onomatopoeia, all of which may stretch the
teller’s lexical talents.2 Note some examples of these phenomena in (6.6):
118 + 6 Vocabulary and the spoken language

tiny little
click-click
great big smile
very very funny

as well as the explicit marking of the high point of the story by the teller:
the very funny bit was. But the listener is not passive in evaluating the story
either. He comments on the story as it progresses:

and they zapped it to you

that’s the trouble when you ...
that’s a novel idea

Another characteristic that may be observed in (6.6) is the use of several

items by the teller which seem to be deliberately vague and to make her
account less precise:

this sort of slide and sound sequence

this picture thing
this coloured thing

Vague and rather general words like those highlighted are frequent in
everyday talk. Channell (1994) has made a thorough study of some
aspects of vague language, confirming its widespread occurrence (see
also Aijmer 1984b and many annotated examples in the conversational
extracts in Carter and McCarthy 1997). Thing is certainly a frequent and
very useful word in spoken language; it can substitute for a wide range
of names of objects, processes, entities and even persons in discourse (see
Fronek 1982). Despite their vagueness of denotation, such words rarely
cause problems for listeners and pass unnoticed, but they do seem to
make an important contribution to naturalness and the informal,
convergent tenor of everyday talk. Indeed, the listener would be consid-
ered irritating and unco-operative if he/she constantly demanded clarifi-
cation and specification of vague language items.
Listeners can also show convergence with speakers by predicting what
the speaker is about to say:

(6.7)
[Two men are discussing domestic pets.]

<S 01> Well of course people who go to the vet’s are [<S 02> mm]
interested in the cats and dogs ain’t they?
 6.5 Fixed expressions - 119

<S 02> Yeah but the people that first have pets, kit-, pets er don’t realise
what’s involved do they?
<S 01> LCare,
well it sorts them out you know those that don’t care that’s
it so [<S 02> mm] but [<S 02> mm] if you wanna you know,
somebody that’s keen on having [<S 02> mm] a pet [<S 02> mm]
and want it in good order.
<S 02> lDone ... done properly that’s right yeah.*

Here we see that it is not only in ‘monologic’ modes such as storytelling

that listeners have a role in creating the lexical patterns of a text. The
effective listener is constantly predicting the upcoming discourse, and
quite often these predictions are verbalised, as if to say: ‘I know what you
are going to say and I think this is what you will say’. Each of the speakers
does this in this extract:

<S 02>... pets er don’t realise

<S 01> LCare
<S 01>... and want it in good order.
<S 02> ldone ... done properly that’s right yeah.

Usually the predicted words are near enough to the speaker’s intended
ones not to impair communication. The phenomenon of prediction is
very common; we may note here that each speaker does it to the other.
Once again, this is evidence that listeners are not passive in constructing
the lexical fabric of a conversation; conversational convergence (in terms
of both goals and relationships) is necessarily a joint responsibility, and
the lexical choices are significant signals of attempts to converge.

6.5 Fixed expressions

In (6.8), we may observe another dimension of vocabulary choice in

conversational language, namely the number of fixed (including idio-
matic), ready-made expressions the speakers use (idioms are investigated
in greater depth in Chapter 7):

(6.8)
[<S 01> has just come back from a holiday where he had trouble with his
luggage going astray. He is about to go off on another trip. <S 02> is his
neighbour,]
120 + 6 Vocabulary and the spoken language

<S 02> When are you heading off again Bob?

<S 01> A week today ... I shall be off to Munich this time ... so I’m just
wondering where the luggage is going to go and looking at my case
now, I find that it’s burst open and whether it’s fair wear and tear I
don’t know, because last time I saw it it was in perfect nick.
<S 02> You reckon it might have suffered from its journey.
<S 01> Oh they get slung about you know, I never used to get a decent
case I buy a cheap one.
<S 02> Mm.
<S 01> Because they just get scratched.
<S 02> Mm.
Informal conversation is marked by a high occurrence of fixed expres-
sions and idioms of various kinds, some more transparent than others,
though still quite fixed in their syntax and lexical form. We may note
some examples here:

when are you heading off (phrasal verb)

a week today (fixed adverbial frame)
I shall be off to Munich (phrasal verb)
it’s burst open (phrasal verb)
this is fair wear and tear (irreversible binomial)
it was in perfect nick (restricted collocation)
they get slung about (phrasal verb)

Idiomatic fixed expressions are by no means equally distributed in

discourse (see Chapter 7; also McCarthy and Carter, 1994: ch.3) and an
over-concentration on written text in language learning may well
present an unbalanced view of their significance in communication.
They are always evaluative in the sense that they are not neutral
alternatives to their literal counterparts but include some attitude or
comment on the entities and phenomena they describe and project an
informal relationship between conversational participants. Observing
how, when and where such expressions occur is a major challenge to
research in the spoken language (see Nyyssénen 1992), and available
evidence so far is scarce (exceptionally, see Strassler, 1982; Powell 1992).
Chapter 7 will attempt to remedy that situation by looking at idioms in
spoken data. But apart from opaque idioms, a good deal of semantically
transparent vocabulary is also fossilised into fixed, multi-word expres-
sions and restricted collocations (see Aisenstadt 1981) and compounds.
Some of these are highlighted in (6.9):
 6.5 Fixed expressions + 121

(6.9)
[<S 01> has been telling a story about a road accident involving her car
and a tramp.]

<S 01> And of course the police officer came and I was a bit shocked and
he said get in the passenger seat and he drove me to the police
station you see, somebody sent for an ambulance and there was all
activity going on this man was propped up at the side of the wall
he looked pretty you know he wasn’t bleeding or anything
<S 02> Wasn’t he?
<S 01> No no
<S 02> How fast were you going then?
<S 01> I wasn’t going very fast you see you know it was I'd only just
turned the corner more or less here, and there was a bit of a line of
traffic and then
<S 02> So it was a bit of a miracle he wasn’t hurt wasn’t it
<S 01> Apparently it was his party no it was his party-piece because the
police told me that he’d done it very often this ’cos it got him a bed
for the night, you know it got him in hospital [<S 03> [laughs]] and
when he were getting a bit fed up, he’d already had them there that
morning apparently saying the IRA had put a bomb under his bed
but then he picked on me and er it got him a bed for the night in
hospital and that was his he did it regular
<S 03> Good grief
<S 01> You know but the thing to laugh
<S 02> LIt’s a wonder he didn’t break every bone in his
body isn’t it
<S 01> It was just, well the police woman rang up I was there because I
were very upset at the thought I’d hurt him you know and she said
oh he’s only cuts and bruises they’re used to him and you mustn't
bother about this sort of thing but I was pretty upset at the time
<S 02> Yeah yeah.

This kind of ‘off-the-peg’ vocabulary assists fluent production in real

time, and seems to be just as significant as the single-word elements that
go to make up the text. Indeed, it is hard to envision the notion of
fluency having any real validity if language could not be (at least in part)
produced ready-assembled in this way (see Bolinger 1976). Thus spoken
language pedagogy cannot afford to neglect the multi-word items that
are found in data. Only close observation of real data can raise awareness
122 + 6 Vocabulary and the spoken language

of the centrality of such phenomena, and confronting data may be a

necessary preamble to the more conventional teaching and learning
strategies that are found in typical vocabulary lessons.

6.6 The spoken vocabulary syllabus

The question often arises as to how teachers with little or no access to

large amounts of real data can get access to relevant lexical facts about
the spoken language. One solution (though limited in what it can tell us,
as much of the necessary qualitative interpretation in this chapter
underlines) is the use of ever-increasingly available computer-generated
word lists. Computational analysis enables the creation of automated
word lists which can be used alongside the manually-compiled word lists
that often appear in syllabuses as a checking mechanism, or to compare
with computer-generated word lists for written language. Frequency lists
for spoken language do differ considerably from those dependent on
written databases, especially journalistic ones (such as Zettersten 1978).
The following lists are each based on approximately 100,000 words of
data, and reveal interesting differences (for a fuller discussion, see
McCarthy and Carter 1997a):

Written Spoken

= aa

=ONOTRON=-OOMNOGAHRAWDND
ee
ee
ee
 6.6 The spoken vocabulary syllabus + 123

19 at
-20. by
the
she
21 ‘have ao
2s do
28 from
24 are
wel
on

25 |
Zon etins
ined
there
27 ‘they what
288 snot he
29 an for
30 will eo
31 who this
82 been all
383 their be

34 had
35 = one
‘dori
not
36 which just
3 vou
388 alll
389 last
40 ~=+her
41 said
42 were one
43 we
44 when
45 more
46 ‘there
47 would
48 she
49 or
50 up

Figure 1: The 50 most frequent words from 100,000 words of written data
(newspapers and magazines)° and 100,000 words of spoken data (CANCODE). The
shaded cells in the spoken column indicate forms which occur significantly more
frequently in the spoken than in the written

Notable in these lists are both the similarity of rankings of basic words
and some differences which give the two modes of language their
characteristic qualities. The written list consists of mostly function
words, but the spoken list seems, on the face of it, to include a number of
lexical words such as know, well, get/got, go, think, right. However, most of
124 + 6 Vocabulary and the spoken language

these prove to be elements of discourse markers (e.g. yout+know, I+think) or

single-word markers (well, right) (see Stenstr6m 1990 for a discussion and
further examples). This suggests that any teacher wishing to incorporate
insights from vocabulary analysis of the spoken language has to decide
the status in the syllabus of discourse markers. As well as these lexical
words, a number of items which are often classified as conjunctions in
sentence-grammar may have to be re-assessed owing to their extremely
high frequency of occurrence as markers in the spoken language (e.g. so;
see also the discussion of subordination and words such as because, in
section 4.5). Other items will need re-examination too. What are the
commonest functions of the frequent spoken uses of get? Is get/got used
differently in spoken language as compared to written? To answer this
last question, let us consider some statistics for the form got. Got occurs
approximately 14 times more frequently in our spoken sample than in
the written. By far the most frequent use of got in spoken is in the
construction have got as the basic verb of possession or personal associa-
tion with something. Some examples follow, concordanced for the
subject pronoun I plus ’ve:

4684 03 enorsomething Yes cos I’ve got the cross-London transfer anyway A
2028 01 ipe it Erm not yet cos I’ve got to make the bread when I've finish
7782 01 is born in July, ’cos_ I’ve got so many birthdays in July. All
DoL03 know. I’ve got it down I’ve got it somewhere that outside the er c
481 02 um I tell you what else I’ve got Chris do you know we made an album
8552 02 West. Yes soam I. Er I’ve got an agreed overdraft limit of five
102 01 aying about the fellah I’ve got you She would marry him if he wor
1986 01 you know a sore finger I’ve got a great big bloody hole It’s not
4544 02 called Hearts of Fire [I’ve got that on video But they took off a
404701 omeofthe upper fours I’ve got erm a magazine and it had like sui
8899 02 1 got them. Yes Ihave I’ve got them they must be around out here.
3627 02 hat’s why it’s so heavy I’ve got like That’s why cos cos you got,
95001 = ildren I don’t know how I've got it unless you don’t go to the danc
6644 01 ewed. .I’ll tell youl I’ve got a choice between three months in t
482 02 Switzerland you did it I’ve got that upstairs. That was dreadful
990 04 eah it does doesn’t it I’ve got two now yes it does always disappe
6604 01 ve got jobs. Go for it. I’ve got ajob. Not yet. Do you want one? M
2579 04 rop Sorry Warwick No I’ve got some thanks Cheers Nice Mm Ver
6686 01 you got? Sweden’s not. I’ve got eleven. Norway. Norway isn’t eithe
1768 02 ah That’s the only one I’ve got Yeah that’s fine Yeah d’you mind
6794 01 t . That’s the only one I’ve got I haven’t got any of the small one
478 02 some I'll get some out I’ve got some up in the cupboard haven’t I

Figure 2: Sample concordance lines for I’ve got (spoken)

 6.6 The spoken vocabulary syllabus - 125

Not only got displays differences in distribution between written and

spoken; we can expect many other words to display significant differ-
ences too, especially apparently synonymous words such as start and
begin (Rundell 1995) and too and also. Other individual words too will
occur differently. For example, in our samples, the modal adverb probably
occurs six times more frequently in the spoken data than in the written,
and is thus one of those words that may have far greater importance in
the spoken vocabulary syllabus than it might have attracted in a syllabus
based only on written data. Meaning and use as well as frequency may be
different between written and spoken data. In the concordance for I’ve
got, above, the line ‘cos I’ve got so many birthdays in July is typical of the
rather vague types of meaning that can be expected to be more frequent
in informal spoken language. Another example of contrast between
spoken and written is the verb tend, which occurs nine times more
frequently in our spoken sample than in our written one, and there is
evidence to suggest that (especially among younger speakers when the
speaker designations are traced back to the archived information on
speaker age/gender, etc. in the CANCODE corpus) it is becoming gramma-
ticised, and being used to express habitualness or regular occurrence
rather than proclivity or bias towards an action. Again, the spoken
concordance reveals this pattern:

4783 02 chool Erm yeah but we don’t tend to go very often because it I mea
4789 02 quite far away Mm but I tend to like to save my money and spen
5768 37 up the drift The thing is I tend to borrow things off Tim and het
3073 04 on I tend not to use names I tend to use direct names very little b
297 01 that? Rock seaside rock. I tend to buy it and then wait a year so
7842 01 straight to bed Yeah What I tend to do is read or watch television
7281 01 to look too Do you like it, tend to like it slightly sort of forwa
7026 01 ke that with parties, people tend to not turn up Mm until after
4676 01 That’s right Yeah the shops tend to open about eleven o’clock
2050 02 ionally if I do buy bacon we tend to have it for a lunch you know w
5761 37 got six good glasses but we tend not to use them She was sayings
2151 02 ty seven Well that’s how we tend to go every fortnight and we spen
4771 02 re a couple of times and you tend to find that a lot of the London
7276 01 se, yeah Yep, How do you tend to like to dry it, do you like it
5763 37 e about half past six and he tends to clean the windscreen then Su
3002 03 in Cardiff it tends to be it tends to be quite wet it wasn’t too ba
7848 02 off straight back because it tends to flop, urm certainly tightene
6884 02 I think when it’s shorter it tends to, you notice it growing more a
7283 02 ith that a bit down, or that tends to go backI don’t quite know wh

Figure 3: Sample corcordance lines for tend (spoken)

126 + 6 Vocabulary and the spoken language

Many of these concordance lines suggest a focus on regularity or

habitualness rather than on the meaning of bias towards an activity, or
‘hedging’ of a proposition (as Low 1995 interprets his questionnaire
respondents’ use of tend). One might especially note we don’t tend to go
very often, what I tend to do is read or watch television, we tend to go every
fortnight and he tends to clean the windscreen. It is not that grammarsor
dictionaries necessarily fail to pick up these distinctions of sense (e.g.
see the excellent sense distinctions for tend in the COBUILD 1995
dictionary, where corpus evidence informs the dictionary, and where
the ‘habitual’ sense is foregrounded). The importance for the present
discussion is rather the prominence of tend with this sense in the spoken
corpus, and, above all the much higher frequency of tend in the spoken,
which suggests that its marginalisation in most grammars may give less
than a true picture of its use in everyday language. For example, Quirk
et al (1985: 236) relegate tend to a mere mention in a category labelled
‘other’ among the marginal auxiliaries, and yet tend is three times more
frequent in the spoken sample of 100,000 words than ought, which gets
much more attention in large grammars and in English language
coursebooks.
Computational analysis of language corpora can point up some inter-
esting and pedagogically useful differences between spoken and written
vocabulary use, and even relatively small samples (by today’s standards)
can yield original insights or can raise awareness for future observation
and verification in the field. But computers are not very good at picking
up the interactive features of lexical choice and are no substitute for
keen observation and qualitative interpretation. Having said that, the
combination of good quality observation and sophisticated computing
power, along with carefully targeted corpora, can be an extremely
effective tool in assisting progress towards our goal of discourse-oriented
language descriptions for pedagogical purposes. And yet the corpus
should not exercise a tyrannical hold on the syllabus. For example, it is
unlikely that the names of all seven days of the week will occur with
exactly equal frequency in any corpus (see Martin 1988), yet no-one in
their right mind would seriously doubt their equal status (from a
psycholinguistic point of view) as target items in the lexical syllabus.
Many other intuitively related word-sets may have to be considered in
this way too, and the syllabus should be corpus-informed rather than
corpus-driven (for discussion of this distinction, see section 1.9). It is also
worthy of note that, although we have concentrated on high-frequency
 6.7 Conclusion + 127

spoken items, there is still a huge lexical load of infrequently occurring

words in everyday conversation: the Davis-Howes word-count of spoken
English (Howes 1966) (see section 1.3) has already shown that getting on
for half of all the words listed in its 250,000-word corpus occur only once.
What is important in the case of automated analysis of spoken language
is to achieve a degree of objectivity of observation that real-time
participation in conversation denies us most of the time to a far greater
extent than reading written text does, and it is here that a corpus of
spoken language comes into its own.

6.7 Conclusion

Our conversational extracts have illustrated different aspects of the

lexical characteristics of everyday spoken language, and one does not
need much data to see the same features constantly recurring. A view of
language which starts from the premise that language forms in contexts
create the discourse process, helps us to locate and explain those
features in terms of the kinds of constraints which differentiate con-
versational language from the composed, single-authored written text
mentioned at the beginning of this chapter. The differences are many,
and here we have only considered the implications for lexical patterning
and, in turn, what broad implications for vocabulary teaching and
learning such a view of lexical patterning holds. We have also seen in
Chapters 4 and 5 that the same applies to grammar. The grammar of
spontaneous spoken language is different in crucial aspects from that of
carefully composed written text, and operates under the same con-
straints related to participants, goals and settings as lexical selection. A
discourse-based view of language forces us to change our understanding
of formal patterns but also raises important questions about how one
communicates such knowledge to learners and, above all, equips them
with the appropriate resources to enable them to converse naturally in
conversation in the target language. I have suggested that the kinds of
features highlighted here may provoke a reconsideration of methodolo-
gical assumptions in vocabulary teaching. It may not be enough to
present, practise and produce words (or even words in sentences) when
we are dealing with the kinds of lexical features we have argued to be
central to conversational language. A language-awareness approach may
be more effective and appropriate at the outset, and encouraging the
‘learner-as-researcher’ may be the best long-term strategy for empower-
128 + 6 Vocabulary and the spoken language

ing the learner to become a natural user of the target spoken vocabu-
lary. The lexical load may not necessarily be greater in spoken language
programmes, but it will certainly have different priorities and emphases,
and will be based on what real data can tell us, rather than intuitively
constructed word lists and sentences. Above all, the syllabus will
recognise that the vocabulary of a language is an integrated resource (as
we have argued for features such as synonymy and antonymy) which
serves the progression and development of topics and participant goals,
and just as importantly, the construction and maintenance of social
relations.

Notes

1 This chapter is a much revised version, with new data and expanded references,
of a paper published in Longa, H. P. (ed) (1994) Atti del Seminario Internazionale di
Studi sul Lessico, Forli - San Marino, 1992. Bologna: CLUEB, 119-30.
2 1 am grateful to Faye Wadsworth, formerly of the Department of English
Studies, University of Nottingham, for permission to use this piece of data,
recorded in 1990.
3 Lexical creativity in oral narrative and other genres is further discussed, with
corpus examples of morphological creativeness, in Carter and McCarthy
(1995a).
4 Data kindly supplied by Jim Lawley, formerly of the Department of English
Studies, University of Birmingham, 1987.
5 Sampled from the 100-million-word Cambridge International Corpus; © Cam-
bridge University Press.
6 The question of optimum corpus size is largely irresolvable, and seems to be
dependent on the current size of major competing corpus projects (see the
discussion in Chapter 1). We may confidently predict that the present-day rush
for corpora of tens and even hundreds of millions of words will be ridiculed
within a decade by those who will argue that anything less than a billion words
is inadequate. Carter and McCarthy (1995b) demonstrate that even relatively
small samples of 20-30,000 words of spoken language, if carefully targeted to
particular goal- and relation-types, can yield grammatical insights overlooked
by analysts of corpora many times that size, and can be pedagogically useful
and relevant.
7
Idioms in use: a discourse-based re-examination
of a traditional area of language teaching

7.1 Introduction

In an earlier paper on this subject,t and in McCarthy and Carter (1994:

109), we noted that idiomatic expressions of various kinds have long
preoccupied language teachers and learners, and that publishers regu-
larly devote special supplementary teaching materials to idioms and put
out dictionaries of idioms aimed at language learners (e.g. Cowie and
Mackin 1975; Longman 1979; Cambridge University Press 1998). We also
noted that, compared with the amount of research into features of
spoken language in general, including such areas as turn-taking (Sacks
et al. 1974), topic-management (Gardner 1987), and lexical patterning in
general (small though work in this last area has been, as was examined
in Chapter 6), we still seem to know very little about how idioms are
actually used in everyday talk. Instead, we tend to continue to teach
them as curiously disembodied items, detached from any contexts other
than imagined and contrived ones which may or may not bear any
resemblance to their real use. Learners tend to treat them as an oddity, a
whimsical feature of the target language, something to collect and
hoard, list-wise, in the vocabulary notebook. In this book, I have consis-
tently argued that language teaching and applied linguistics in general
can benefit from a spoken discourse-based orientation towards language
description, and that language teachers should concern themselves
critically with the descriptive approaches that generate raw material for
teaching. But the temptation is often to attempt to incorporate the more
exciting insights of discourse analysis, such as models of turn-taking and
the like (see section 3.2.2), and to leave traditional areas of interest, such
as pronunciation teaching and vocabulary teaching to the traditionalists.
However important and fruitful that line may be, an even greater
challenge exists in the re-modelling of traditional areas of teaching,
especially those embedded in largely unquestioned assumptions. This
chapter, therefore, attempts to look at some natural spoken contexts of

129
130 - 7 Idioms in use

use for idioms, and to examine what sorts of discoursal functions idioms
perform. It extends the work reported in McCarthy (1992b) and McCarthy
and Carter (1994), and tries to draw the threads together into a coherent
theory of idiom use.
In the earlier work referred to, we used the word ‘idiom’ to mean
strings of more than one word whose syntactic, lexical and phonological
form is to a greater or lesser degree fixed and whose semantics and
pragmatic functions are opaque and specialised, also to a greater or
lesser degree. An example of a string where all elements are fixed is the
expression rough and ready. The expression must be uttered with that
particular word-order, those particular words and with one single tone-
unit (/7ROUGH and ‘REAdy/ [' = primary stress; * = secondary stress; | | =
tone-unit boundary]); its meaning is fixed and largely non-negotiable (see
Cowie 1988 for an extended discussion of this feature of fixedness). Other
expressions may be more flexible in one or more respects along the
possible continua of lexico-grammatical, phonological and semantic/
pragmatic fixedness. The expression to turn a blind eye (to something) was
recently nominalised and pluralised by an interviewee on a BBC radio
programme, when he replied that ‘Blind eyes have been turned all the way
along to breaches of safety regulations ...’, revealing a degree of
’

acceptable syntactic flexibility. The cut-off point where fixed expressions

become open, freshly synthesised lexico-grammatical configurations
(created anew each time, as it were), and where opaque idiomatic
meaning becomes transparent and more and more literal is problematic
and ultimately impossible to pinpoint. But a definition of idioms that has
blurred boundaries has advantages as well as disadvantages: it usefully
enables us to incorporate within the term a wide range of fixed expres-
sions over and above the clause-idioms (of the type verb + complement, e.g.
hit the sack in English, meaning ‘to go to bed’) and idiomatic phrasal verbs
(e.g. take (sb) off, meaning ‘to mimic’), which are traditionally most often
focused upon in language teaching materials. These extra categories of
idioms include:?

1 Prepositional expressions: in two shake’s of a lamb’s tail (meaning (to do

something) very quickly).
2 Binomials and trinomials: usually irreversible combinations with and
or other conjunctions
whose order may vary from language to
language: black and white film (cf. Spanish blanco y negro, ‘white-and-
black’), ready, willing and able, give or take, safe and sound (see Malkiel
 7.2 Idioms in use: evaluative force + 131

1959; Gustafsson 1975; Norrick 1988; Fenk-Oczlon 1989 for further

examples and discussions).*
3 Frozen similes; usually formally identified by the removability of the
first as: (as) keen as mustard, (as) cold as charity (see Tamony 1982; Norrick
1986).
4 Possessive ’s phrases: a king’s ransom, the cat’s whiskers.
ul Opaque nominal compounds: blackmail, a mish-mash, the back of beyond.*

6 Idiomatic speech routines, gambits and discourse markers, which are

very frequent in conversational data: by the way, how’s it going?, that’s
that, mind you.
7 Cultural allusions; these include a wide range of quotations, slogans,
catch phrases, proverbs, all instantly identifiable to those who share
the cultural context: to be or not to be ..., sock it to me, every cloud has a
silver lining.

What is important to note from this list is that idioms cannot always be
uniquely identified by their formal properties, especially the phenomena
listed in 7, where a wide variety of formal types are found (phrases,
clauses, sentences, whole texts) and which are held in common by the
members of speech communities as references to their shared culture.
Ultimately, intuition also has to play a role, especially in borderline
cases, in the identification of idiomatic strings of words (see Bressan 1979
for an extended discussion of the problems).

7.2 Idioms in use: evaluative force

All of the types of idiomatic expression listed above may be found in

greater or lesser number in real spoken and written language. In the
literature, discussion usually centres upon the semantics, the syntax, the
cross-linguistic differences and the universality of such expressions (e.g.
see Makkai 1978; Fernando and Flavell 1981; Reagan 1987), and I shall
not gainsay the pedagogical use to which such insights may be put. What
is almost always lacking, though, is an attempt to examine function and
distribution in real contexts of use. Conversational routines, gambits and
discourse markers are an exception (see Coulmas, 1979, 1981a; Schiffrin
1987) and need little further comment here. There is often an underlying
assumption that idiomatic expressions are merely rather informal or
colloquial alternatives to their semantically equivalent literal free-forms.
This may well be true insomuch as the kinds of data where idioms occur
132 + 7 Idioms in use

often reflect a high degree of informality at the interpersonal level

between speaker and listener, but this does not go anywhere near far
enough to tell us why speakers choose idioms instead of literal, trans-
parent counterparts at particular points in a discourse. The question to
answer is: why should languages ‘duplicate’ ways of saying things,
offering the literal and idiomatic options that seem to operate (theoreti-
cally at least) in free variation at many places in discourse? Certainly,
real data seem to indicate, as I shall attempt to demonstrate below, that
idiom selection is not random and unmotivated, and can be linked to
features of language choice at the discourse level. Considerable support
for this claim is already available from computational analysis of written
texts of a wide variety of types, where the function of idioms as
evaluative devices, correlating with authorial comment segments in
texts, seems to be a regular pattern (see Moon, 1992). In spoken data, the
distribution and functioning of idioms is even more fascinating and less
well-researched. A spoken corpus enables us to observe idiom use in
everyday talk at close quarters across a wide range of speakers of diverse
ages and social classes, in different situational contexts.
One of the very few analysts to attempt to describe idiom use in
naturally-occurring spoken English is Strassler (1982) (and more recently,
see Powell 1992).° Strassler does what the present chapter advocates and
departs from the traditional way of analysing idioms as a semantic
problem and looks at the pragmatics of idiom use. Strassler found that
idioms were, relatively speaking, infrequent (they occur on average once
per 1,150 words in his data). This might immediately suggest, for the
learner needing to concentrate on spoken skills, that they are less of an
obstacle to native-like fluency than they are often believed to be. But
where idioms did occur in Strassler’s data, they did so with a degree of
predictability, not randomly. Idioms, Strassler maintains, are much more
likely to occur when a speaker is saying something about a third person
or about an object or other non-human entity, rather than about the
speaker him/herself or about the listener(s) (1982: 103). This he attributes
to the evaluative function of idioms and the risks to ‘face’ and inter-
personal relations which can stem from the self- or other-abasement
which idioms often entail (Strassler 1982: 103, 109) (‘face’ here is under-
stood in the sense propounded by Brown and Levinson 1987). To say to
someone I’m sorry to leave you twiddling your thumbs (instead of I’m sorry to
keep you waiting) expresses a certain dominance and confidence (and
flippancy?) on the part of the speaker and a potential offence and loss of
 7.3 Idioms in everyday stories and anecdotes - 133

face to the listener, which alternative, more transparent, ‘literal’ rendi-

tions seem to guard against. Indeed, when speakers do use idioms
directly about their interlocutors, they may explicitly gloss their usage to
guard against threats to face, as in (7.1):

(7.1)
[Speaker <S 01> is telling his listener how he envied him and another
colleague their ability in their job.]

<S 01> Well you and Aubrey used to make me sick actually, in the nicest
possible way.

Strdssler’s study and example (7.1) indicate the direction of the present
chapter: the potential for the integration of levels between lexical form
and communicative function (to include interpersonal elements such as
politeness and face). We shall also be interested to see whether genre has
any implications for the use of idioms.
(7.2) illustrates the evaluative function in action, with the use of an
idiom (in this case a frozen simile) to a third-person, non-participant
entity:

(7.2)
[The speakers are exchanging views about political dominance by the
Conservative Party in Devon, in the south-west of Britain.]

<S 01> But living down here in Devon there’s no way

<S 02> We’re disenfranchised completely.
<S 01> You know there’s such an enormous Conservative majority I
mean today I had to ring up one of the local councillors he’s as thick
as two short planks.®

Here we have a factual observation, followed by an evaluative comment.

This pattern, as we shall see later, is a recurring feature of idiom use.
This general ‘observation plus comment’ function, as I shall refer to it,
has a specific type of manifestation in oral narratives, to which we shall
now turn.

7.3 Idioms in everyday stories and anecdotes

In this section, we consider oral narrative as a clearly defined generic

activity with an identifiable discourse structure (see Labov 1972, and the
134 - 7 Idioms in use

brief summary of Labov’s model in section 3.2.3). Idioms do seem to occur

at important junctures in everyday stories, not just randomly. They often
occur in segments where the teller and listeners step back and ‘evaluate’
the events of the narrative, rather than in the reports of the events
themselves. Evaluation in Labov’s model refers to the function whereby
storytellers make the events they are telling worth listening to. Evalua-
tive clauses in narratives forestall the embarrassing question So what?
(Why should I want to listen to this story? What’s exciting/special/funny about it?).
Equally, listeners add their evaluation of events, commenting on the
story’s general worth, its effects on them, and so on. Evaluation is not an
optional extra in storytelling; without it there is no story, only a bland
report. In the following examples, the contrast between ‘events’ and the
evaluation of them is underlined by the shift from neutral expressions to
idioms:

(7.3)
[The teller, <S 01> introduces a new character into his story.]

<S 01> And Guss Hughes came along one day and we were always taking
the Mickey out of him he was you know he was one of these the lads
that always got taken so we all knelt down

(7.4)
[<S 01> (Mary) is recounting a story that happened when she was on
holiday with her friend, Dulcie.]

<S 01> I said what would you like to do this afternoon Dulcie she said oh
Mary let’s go to bingo now bingo is never ever my cup of tea [<S 02>
no] but seeing that I was supposed to be with her
<S 02> lsupporting her yeah
<S 01> I'd to fall in with her [<S 02> [laughs]] all right then Dulcie where
do we go now to bingo.

In (7.4) the transition from event segment to evaluative segment is

further reinforced by the discourse marker now. We may conveniently
represent the event line and evaluation line (which simply represent a
more specialised version of the ‘observation plus comment’ function)
diagrammatically:
 7.3 Idioms in everyday stories and anecdotes + 135
SS
looeee

Event line Evaluation line

a sd es ee
(7.3)
and Guss Hughes came along one
day
and we were always taking the Mickey out
of him he was you know he was one of
these the lads that always got taken
so we all knelt down

(7.4)
I said what would you like to do this
afternoon Dulcie she said oh Mary
let’s go to bingo
now bingo is never ever my cup of tea but
seeing that I was supposed to be with her
I'd to fall in with her
all right then Dulcie where do we go
now to bingo

In everyday stories, idioms often occur in codas (again in Labov’s sense;

the ‘coda’, at the end of a story, provides a bridge between the story
world and the real world of the teller and listeners; see also Chapter 3).
Here are some examples from corpus data, with relevant idioms high-
lighted:

(7.5)
[End of a story by speaker <S 02> of how a job opportunity she took
represented a big step in her career.]

<S 01> Still that’s the way it all started for you.
<S 02> That’s right.
<S 03> The big break wasn’t it.

(7.6)
{End of a story by speaker <S 01> in which she and a friend benefitted
from half-price food during a ‘happy-hour’ at a restaurant.|

<S 01> And that was drinks [<S 02> yeah] but that was half price it would
have been sixteen pounds each [<S 02> yeah] had it been later.
<S 03> Well that’s fair enough isn’t it.
<S 02> You can’t go wrong with that can you.
136 - 7 Idioms in use

(7.7)
[End of a story where speaker <S 01> has recounted a coincidence of
seeing once again a stranger she had felt attracted to but whom she
thought she would never see again.]

<S 01> I thought oh am I never gonna see you again and on the
Wednesday I was just walking past the bank and I saw him [laughs]
so he must have lived in Carmarthen.
<S 02> [laughs] that’s a bit odd
<S 03> lSmall world.
<S 02> lwhen things like that happen
isn’t it.
<S 01> I just sort of go I know him from somewhere and it clicked.

(For further examples see McCarthy and Carter 1994: 111.) In (7.5), the
idiom evaluates the whole narrative, summarising the main events and
relating them to the teller’s current career position. Idioms in (7.6) and
(7.7) similarly perform a summarising function. As I have argued else-
where (McCarthy 1991: 139-40), storytelling is normally a collaborative
enterprise, and listeners have the right (one might say responsibility) to
evaluate the events, and to ensure a smooth passage for all participants
from story world back to conversation world when the story has ended.
(7.5), (7.6) and (7.7) all have listeners using idioms to contribute to the
coda. But where listeners do use idioms for this purpose, they will have
to be ones that are careful not to abase the teller, unless the relationship
between teller and listener(s) is very relaxed and on equal and/or
intimate terms. The kinds of idioms we see in codas (often clichés,
proverbs, sayings of various kinds) partake of the ‘sheltering behind
shared values’ that Moon (1992) observed in idiom usage in her written
data (see also Loveday 1982: 83), and reinforces the importance of
observing the cultural contexts of idioms, in the broadest sense of the
word ‘culture’ (see McCarthy and Carter 1994: 114-17). Narrative codas
are essentially no more than a specialised example of a more general
class of points in discourses where gist is summarised, providing the
opportunity for participants to agree on what they have achieved so far,
and to move on to new topics (see the discussion on ‘formulations’ in
section 2.3). It is no surprise, therefore, that sociolinguists and conver-
sation analysts, coming at talk from an angle of trying to understand the
social significance of the precise moment of placement of particular
items, should find that idiomatic expressions occur regularly at topic-
 7.3 Idioms in everyday stories and anecdotes + 137

transitions (Drew and Holt 1995) and as summaries of gist in sequences

such as the formulation of complaints (Drew and Holt 1988). The motive
of these researchers is sociological; however, language teachers can learn
a great deal from such analyses. Drew and Holt’s work underlines the
non-random occurrence of idioms and strengthens the argument which
is at the heart of the present chapter: that we have much to gain by
closely observing how and when idioms are used, not just their formal
characteristics.
Another noteworthy feature of idioms in everyday talk is the way
speakers use them creatively, by a process of ‘unpacking’ them into their
literal elements and exploiting these. McGlone, Glucksberg and Cacciari
(1994) argue that speakers cannot ignore the non-idiomatic meanings of
individual words in idiomatic expressions, and that even in opaque
idioms, literal meanings of component words are in some sense acti-
vated, or at least are potentially available. In (7.8) two middle-aged
school-teachers are reminiscing about the early years of their careers;
one of them comments on a class he had:

(7.8)
<S 01> The second year I had, I started off with 37 in the class I know
that, of what you call dead wood the real dregs had been taken off
the bottom and the cream the sour cream in our case up there had
been creamed off the top and I just had this dead wood, I mean it
really was and he was so impressed with the job that I did with them
and the way that I got on with them and he immediately said right
how do you feel about taking a special class next year and I took one
from then on.
<S 02> Rather you than me.

The shift from the event line to the evaluation line is quite clear here
again, with idioms occurring in the evaluative segments. The idiom to
cream off and the idiomatic noun phrase the cream, are exploited by re-
literalising the notion of cream and adding sour as an ironic evaluation’.
The listener then characteristically adds his evaluation/coda with the
expression rather you than me. Items such as the cream and cream off in (7.8)
do admittedly raise problems concerning the borderline between fully
institutionalised, ‘fossilised’, opaque idioms and extended metaphors
which are perhaps not yet fully fossilised and retain some transparency
of meaning (on this fuzzy borderline see Choul 1982; Fernando and
138 - 7 Idioms in use

Flavell, 1981: 44-7). But this is not the issue of our present discussion,
and recent work on metaphor stresses the interpersonal and evaluative
functions of metaphors, so underlining their common ground with more
opaque idioms (see especially Low, 1988). Not least, the study of metaphor
must also confront the apparent duplication of meaning in the vocabul-
aries of languages and attempt to understand its functioning in exactly
the same way as the study of idioms should do.
(7.9) is an example of the exploitation of semantic connections
between two idioms to elaborate the coda:

(7.9)
[End of a story by a couple, speakers <S 02> and <S 03>, of how they were
involved in a consumer survey.|

<S 01> So you were Mr you were that Mr and Mrs Average they’re always
talking about then.
<S 02> Yeah.
<S 03> Yeah.
<S 01> The man and woman in the street.
<S 03> Yeah.

In (7.10), the teller and listeners create a series of puns to act as coda toa
spooky story about a ship being sunk in battle. The end of the event line
(the ‘resolution’ of the narrative in Labov’s terminology) is that all on
board the ship were killed, except the teller’s father, who had had a
premonition and refused to sail on the ill-fated voyage:

(7.10)
<S 01> Everyone, everyone died
<S 02> Anyway all hands lost but legs saved.
[laughter]
<S 03> Well sailors were always getting legless weren’t they anyway.
[laughter]
<S 01> Finding their sea legs.
<S 03> Yeah

The euphemistic all hands lost (= all crew dead) gives rise to association
with legs, in turn connecting with get legless (= get drunk) and with
finding their sea legs (= becoming accustomed to being on board ship).
Such punning and word-play memberships participants as ‘belonging’
 7.3 Idioms in everyday stories and anecdotes + 139

culturally, and may make ‘non-membershipped’ participants (e.g. non-

native speakers/speakers of other varieties with different idiom sets) feel
quite excluded. Indeed, idioms are often created among small groups or
those with shared interests (for example see Gibbon 1981), right down to
partnered couples, where intimacy is reinforced by ‘private’ sets of
euphemistic and humorous expressions (see the data in Hopper, Knapp
and Scott 1981). Humorous unpacking and semi- literalising of idioms is
by no means rare in the CANCODE corpus. Another example involves a
cultural allusion to the (at the time) popular media icons, the Teenage
Mutant Ninja Turtles:

(7.11)
[Two young women discussing having children. A ginger nut is a kind of
biscuit with a ginger flavour, and a nickname for people with ginger
hair.]

<S 01> There’s twins in our family.

<S 02> Is there.
<S 01> Yeah.
<S 02> Oh right so you might have a twin.
<S 01> And there’s ginger nuts.
<S 02> Nothing wrong with that.
<S 01> My mum reckons I’m gonna have ginger twins.
.<S 02> Mind you one of my best friends at home Ninj we call him
because he’s ginger you see, Ninja. [pron: /'nmd39/]

Other cultural allusions assume shared knowledge of whole expressions

by only saying part of them; this is a particularly exclusive form of
membershipping:

(7.12)
[The same two teachers as in (7.8), reminiscing about a room at their old
school, which <S 01> thinks now looks somewhat shabby,]

<S 01> I said Iremember this when it was a woodwork room, her room
and I said cor crikey it looks as though it could do with one or two
yes she said they rearrange the deckchairs round the edges every so
often but that’s as far as it goes.

It could do with one or two is understood as continuing ... coats of paint (i.e.
it needs redecorating), while rearrange the deckchairs is a partial rendition
140 - 7 Idioms in use

of to rearrange the deckchairs on the Titanic (meaning to make temporary

solutions that do not solve the major problems that lie ahead: the ship
Titanic sank, even though it was thought to be unsinkable).

7.4 Idioms in collaborative ideas discourses

One of the most frequent conversational activities is what we referred to

in section 1.3 as collaborative ideas, where participants share views of the
world, discuss matters of interest, etc. Often, conversations just grow out
of observations about the world and evaluations of those observations,
one observation following another. This type of conversational activity
differs from narrative in that there is not necessarily any kind of
chronological report. Speakers simply make statements of fact or of
perception about the world and accompany it with a comment indicating
their stance towards those observations. This often happens in talk
marked by topic drift, where participants may be ‘fishing’ for topics and
simply casting their minds around, making general and local observa-
tions. Not surprisingly, the place of idioms is firmly in the ‘comment’
segments in such discourses. Some examples are given in diagram form
to underline the functional shifts:

Observation Comment

(7.13)
<S 01> I wouldn’t come back
and live in a big town not at all they’re
dirty they’re noisy
<S 02> All this hustle and bustle

(7.14)
<S 01> and so I go into his bed
and he comes back in
so to my bed and his bed and chopping
and changing

(7.15)
<S 01> See, the folly of leaving
the company, you know,
you would have been jetting off
<S 02> Yeah left right and centre
[<S 01> yeah] Andorra one day Hong
Kong the next
 7.4 Idioms in collaborative ideas discourses + 141

(7.16)
<S 01> Left here at four,
[<S 02> oh yes yeah] ... three, three
and a half hours
He must have driven like the clappers

(7.17)
<S 01> I don’t know, I feel a bit
nervous now [laughs]
<S 02> Do you, stage fright is it
<S 01> I think so yes

(7.18)
<S 01> Mm what about something
like erm ... forensic linguistics
<S 02> Mm I mean, I think the thing is
<S 03> Kind of thing like who dunnit
on the trial [a who dunnit is an
idiomatic compound noun meaning a
mystery murder novel or film: i.e. ‘Who
has done this murder?’
<S 01> That’s right

(7.19)
<S 01> I don’t usually have chips
I usually have jacket potatoes
<S 02> Like mother like daughter

While the comment usually follows the observation, it may also precede
it, as in these examples:

Comment Observation

(7.20)
<S 01> Julie’s got a very cushy
number
she’s off to Mauritius

(7.21)
<S 01> Thomas is a bit of a pain
all sorts of things frighten him you
know, <S 02> Yeah] wakes up with
nightmares and that
(cont.)
142 - 7 Idioms in use

(cont.)
Nil nat err ee Se ee eee ee SS SS
seca

Comment Observation
ae a —Sacer Sa reg ah i
(7.22)
<S 01> I think she ought to be
told the time of day
when I was 21 I didn’t have a car
(7.23)
<S 01> I think there isn’t a
magic formula
it’s something that just happens

(7.24)
<S 01> you’re left to your own
devices that’s it
you get no, no further training nothing

Once again, it is to be noted that the ‘observation-comment’ function

may be split between participants in the conversation, creating cultural
solidarity between speakers and their listeners (on the supportive roles of
participants, see Bublitz, 1988).
In addition to the ‘observation-comment’ function in the more tempo-
rally displaced kinds of exchange of facts and perceptions illustrated in
the table, the ‘comment’ may be immediately referential, referring to the
here-and-now situation of the participants:

(7.25)
[<S 01> and <S 02> are house guests and are waiting to be called to table
for dinner]

<S 01> We're the privileged guests you know you and we’re
<S 02> lHow nice
<S 01> lWe’re
allowed to just sit here and swan it

(7.26)
[<S 01> is relaxing just before a family party, after working hard to
prepare it.]

<S 01> [yawns] well this is the calm before the storm isn’t it what time is
it
 7.5 Idioms, negotiation of meaning and convergence + 143

(7.27)
[<S 01> and <S 02> are looking at an old photograph of themselves from
1967]]
<S 01> It says it all doesn’t it.
<S 02> Absolutely 1967.

7.5 Idioms, negotiation of meaning and convergence

Another feature of idiom use that emerges from close examination of

conversational data is the role idioms play in segments where lexical
meaning is being negotiated among participants. In a way, this is also
true of the more specialised uses of idioms we have been concentrating
on up to now. Oral narrative functions such as evaluations and codas are
negotiated among participants, and attitude and stance do not simply
exist in discourse, but are usually negotiated and worked at, owing to
the potential threats to face, the interpersonal exigencies of real-time,
face-to-face interaction, and the normal desire for conversational conver-
gence (see glossary p. 177). Throughout our data we can observe idioms
being put to service to support the negotiation of lexical meaning. I have
discussed elsewhere (McCarthy 1988; see also section 6.3) how lexical
reiteration performs important negotiating functions. Idioms, with their
often rather vague and general meanings, can stand in for more precise
values in discourse where speakers work towards convergence and
‘mutual understandings. Some examples follow. (7.28) gives to the idio-
matic compound knick-knacks an instantial ‘equivalence’ to the candles
bought by <S 02>. (7.29) uses the idiomatic simile as flat as a pancake in a
relation of ‘opposition’ to the precise topographical description, and
(7.30) uses the binomial idiom life and death in opposition to the speaker’s
preferred term. These latter are characteristic uses of antonymous
expressions in the negotiation of lexical meaning (for further examples
see McCarthy 1988):

(7.28)
[Speakers are discussing <S 02>’s day’s shopping.]

<S 01> Sounds like it cos you bought your little knick-knacks there today
didn’t you.
<S 02> The candles.
<S 01> Yes yes.
144 - 7 Idioms in use

(7.29)
[Speakers are discussing the Eastern part of England]

<S 01> It’s flat you know it’s not er hilly like Wales but [<S 02> mm] you
get used to that strangely enough after a while, I mean it’s not as
flat as a pancake it’s kind of undulating.

(7.30)
[Speakers are discussing the decisions speech-therapists have to make in
their work.]

<S 01> That’s what Ana-Maria says, she’s sort of making not quite life
and death decisions but real life enhancing decisions among you
know dozens of priorities.
<S 02> Yeah.

Knick-knacks in (7.28) is typical of many idiomatic expressions which

enable speakers to avoid precise reference. In (7.31), a hostess at the
dinner-table wants her guests to take vegetables and condiments and to
start eating. Too precise a reference to these items may have sounded
rather commanding; the idiom she chooses is sufficiently oblique to
avoid any threat to her guests’ face:

(7.31)
<S 01> Look get started you know putting all the bits and pieces on.

The same binomial idiom occurs in (7.32), where the speaker is re-
counting how he got lavish presents at a job-leaving party, while another
colleague also leaving got nothing. Bits and pieces defocuses from the
precise meaning and tones down the possibly boastful (and thus face-
threatening to the teller) interpretation of the event:

(7.32)
<S 01> As I say I collected all these bits and pieces all these goodies and
everything and she got virtually nothing.
 7.6 Drawing the threads together + 145

7.6 Drawing the threads together: the overall functioning of

idioms in discourse

Although the division of this chapter so far may suggest rather discrete
functions for idioms, there is clearly an overlap between the notion of
narrative evaluation and the evaluations found in non-narrative observa-
tion-comment discourses. In section 7.5 I have also suggested that the
negotiation of lexical meaning has preoccupations in common with
other uses of idioms (e.g. the need to protect face, participant orienta-
tion, etc.). We have now reached a point where more over-arching
conclusions can be drawn, as a framework for pedagogy. What seem to
emerge most usefully from our data are the following insights:

e Idioms are never just neutral alternatives to literal, transparent

semantically equivalent expressions.
e Idioms always comment on the world in some way, rather than simply
describe it. They are evaluative and frequently involve potential threats
to face.?
e Speakers are aware of the face-threatening potential of idioms and are
careful to use idioms generally only for third-person reference, or to
mitigate threats to face in first- and second-person reference.
e Idioms are communal tokens that enable speakers to express cultural
and social solidarity; this is particularly so of those that are direct
cultural allusions such as proverbs, quotations, etc.
e Idioms may be relatively infrequent in occurrence, but it is possible to
predict where they are likely to occur (e.g. in particular genres and at
particular places).
e Idioms occur in a wide variety of forms, not just the verb + complement
type. In our data samples we have seen idiomatic nominal compounds,
frozen similes, binomials, phrasal verbs, and clausal idioms.

Another point that could be made, but which has been beyond the scope
of the present, necessarily limited discussion is that idioms are socio-
linguistically marked:
the overwhelming majority of idioms in the
CANCODE corpus are spoken by speakers over 25 years old. This may
make their teaching as productive vocabulary for younger age-groups
inappropriate.
The six main points above, and others raised along the way, have
implications for pedagogy. Firstly, it would be unwise to ignore idioms
just because they are not terribly frequent, for where they do occur
146 - 7 Idioms in use

seems to be in very ordinary, everyday contexts such as storytelling, and

at crucial junctures such as evaluations and codas, and other transition
points. What is more, in these and other contexts where they occur, we
have been able to observe how the workload of idiom selection is shared
among participants: in oral narrative, for example, it is certainly not
just tellers who use them, but listeners too. This suggests that idioms
are highly interactive items, and are therefore best looked at in context
and, if possible in interactions of the kinds I have exemplified. The
importance of looking at idioms in context actually has benefits for the
recognition of recurrent formal features too, as Coulmas (1981b) has
argued. In English, for instance, certain verbs such as take and get seem
to be ‘idiom-prone’, and regularly combine with many other words to
form expressions with specialised pragmatic functions which can only
be fully appreciated in context. Elsewhere Coulmas (1981a: 5) asserts
that all fixed expressions are better explained in terms of their use
rather than their semantics. More directly referring to pedagogy, Lattey
(1986) advocates contextual organisation of idioms on the basis of
recurring pragmatic functions related to the interaction of speaker and
listener, speaker and the outside world, positive evaluations and nega-
tive evaluations of people and phenomena, etc. Similarly, McCarthy and
O’Dell (1994) try to put into practice both a formal and functional
categorisation of idioms in the organisation of self-study units on
idioms for upper-intermediate learners of English. Our categories there
include ‘Idioms connected with praise and criticism’, ‘Idioms for
describing people’, ‘Idioms connected with problematic situations’. In
short, it is a matter of seeing idioms as a communicative resource,
rather than as a mere formal quirk of the language (for further
discussion, see Fernando 1996: ch. 6).
However, language teachers often quite justifiably ask the question:
how do I do controlled practice of items which are highly interactive?
One of the problems associated with any discourse-oriented approach to
language teaching is the implication that everything must be practised
in real interactive contexts or at the very least in well-designed simula-
tions. Teachers at the chalk-face know that such free production can
usually only work after an intensive period of controlled practice of
some sort. It is simply unrealistic (and likely to produce disastrous
results) merely to say ‘get into groups and use idioms with one another
in this activity’. But the answer to the plea is a complex one. Firstly,
discourse-oriented pedagogical approaches tend to demand something
 7.6 Drawing the threads together - 147

more than just a ‘presentational’ phase as the lead-in to practice.

Because discourse-based methods often go strongly against learner
expectations, awareness-raising may have to be the first step. In this
case the awareness-raising is concerned with why idioms are there, and
how speakers use them, before the ‘vocabulary-learning’ task is under-
taken. This may involve examining data extracts just as we have done in
this chapter, focusing on function rather than just meaning, observing
the tendency towards third-person reference, how face is protected, etc.
A controlled practice phase could then utilise just the sort of segmenta-
tion I have used in the tables that illustrate the division between, say,
event line and evaluation line in stories, or ‘observation’ and ‘comment’
in collaborative ideas discourses. Gap-filling and multiple-choice activ-
ities using idioms (provided as vocabulary input in the normal ways) in
such tables is then clearly controlled and related to function, avoiding
the scatter-gun effect of simply substituting idioms randomly through-
out texts and avoiding decontextualised focusing just on semantic
meaning. Narrative contexts can be re-created by providing summaries
of stories and giving alternative listener evaluations or codas, for
example:

A friend tells you a story about how she discovered that a

colleague she has worked with for ten years went to the same
school as her thirty years ago, even though they had never realised
this before. What could you say at the end of her story? Which of
these idioms would be suitable and why?
a ‘Oh well, that’s life’
b ‘It’s a small world, isn’t it’
c ‘I bet you were on cloud nine when you heard’
d ‘You live and learn, don’t you’
e ‘Well, would you believe it!’

During the more controlled phases of practice, it may be beneficial to

encourage learners to connect idioms with their own personal experi-
ences, as Bergstrom (1979) advocates, since we have seen how often they
occur in personal narratives. While it may never be possible to re-create
in the classroom the precise natural conditions under which idioms
occur, even the most traditional kinds of exercises and activities can
benefit enormously from an awareness of how idioms are used in real
discourse. Shying away from that fundamental question will get us
nowhere in the quest for more engaging and authentic contexts in which
148 - 7 Idioms in use

to teach an area of vocabulary that seems naturally to interest learners,

but which for too long has suffered from being considered a mere
problem of semantic and syntactic perversity. Chapters 6 and 7 together,
I hope, reinforce my argument that vocabulary choice is not something
that should be neglected as beyond the purview of discourse analysts.
Without it, there would be no discourse to analyse.
In Chapter 8 we turn again to grammatical issues, addressing an area
regularly included in pedagogical grammars and language coursebooks
in English language teaching: reported speech. There we shall see once
again that observation of spoken data, both quantitative and qualitative,
paints a somewhat different picture from the descriptions of reported
speech that have come to us via the study of written texts or from
concocted sentences.

Notes

1 This chapter grew out of an earlier paper (McCarthy 1992). The present version
is fundamentally revised and concentrates on spoken data only.
N Though it is worth noting that not all would agree with the inclusion of
nominal compounds and phrasal verbs in the category of idioms. Gottlieb
(1992), for example, would exclude them, for practical lexicographic reasons.
3 Iam aware of the debate on iconicity in the word-order of binomials, a recent,
useful contribution to which is Birdsong (1995). There do sometimes seem to be
some iconic principles at work, such that the first element is often speaker-
centred (e.g. here and there, now and then, back and forth), or moving from
unmarked to marked term in antonymous pairs (e.g. high and low, good and
bad), or displaying phonological ‘strength’ in the second element (e.g. wine and
dine, huff and puff). However, it is a matter of debate whether such principles
can usefully be incorporated into teaching, or whether binomials, because of
their idiomaticity, are best taught and learnt as unanalysable wholes, just like
monomorphemic words.
4 Here I have focused only on semantically opaque compounds. It has been
argued, however, that all compounds are idiomatic in the sense that they have
developed some sort of semantic or pragmatic specialisation in the process of
institutionalisation as compounds (see Kooij 1968).
5 Additionally, Norrick’s (1988) study of binomials uses real conversational data.
On the written side, Moon (1992) and Vorlat (1985) also examine real data.
6 I am grateful to Beth Sims, former student at the Department of English
Studies, University of Nottingham, for permission to consult and use her data,
from which this example is taken, as part of the CANCODE project.
7 See Ernst (1980) on the use of ‘extra’ adjectives in idioms.
 Notes + 149

8 In the study of written text, the clever, often humorous use of idioms ‘hidden’
in advertising, titles and headlines is well documented (see Moeran 1984; Diaz
1986; McCarthy and Carter 1994: 114-15).
9 Cerndk (1994) suggests that the subjective, evaluative and emotional aspects of
idioms are unparalleled elsewhere in the language.
8
‘So Mary was saying’: speech reporting in
everyday conversation

8.1 Introduction

In a small village in County Tipperary, Ireland, there dwelt a local

character whose nickname was ‘So-Mary-was-saying’, since this was his
stock response when greeted with any piece of news or village gossip; he
had always heard it already, from Mary, the hub of all information in the
small, tight-knit community.! News, gossip, stories, indeed the whole
fabric of everyday conversation depends heavily on quoting or referring
to the words of others, and it is hard to imagine a day of our lives when
we do not at some point support our discourse with direct or indirect
reference to someone else’s words. It is equally hard to imagine, there-
fore, any second language pedagogy claiming real adequacy that did not
take the matter of speech reporting very seriously and did not give it a
place in the syllabus. As I have argued throughout this book, central to
any investigation of language with a claim to pedagogical usefulness is
the close observation of how the most common, banal, everyday func-
tions of linguistic communication, such as speech reporting, are actually
carried out. Equally important is the willingness to approach real data
with an open mind, unencumbered by preconceptions arising from
sentence-based analysis or the evidence solely of written texts.
Hardly any stretch of casual conversational data is without reports of
prior speech, and it is hard to conceive of achieving any intermediate
level of competence in a foreign language without needing to know how
the speakers of that language make speech reports. Thus language course-
books can be relied upon to contain lessons on speech reporting and to
offer opportunities to practise reports. Yet there is evidence to suggest
that many books give an impoverished and inadequate coverage of what
really happens in everyday data, probably because of the continuing
influence, whether overtly acknowledged or not, of sentence-based gram-
matical models, of an over-reliance on written data, and because of a lack
of observation of everyday language. This chapter, like the other chapters,

150
 8.1 Introduction - 151

will base its evidence on the first one million words of the CANCODE
spoken corpus. That is not to say that such observations as are made
below cannot be arrived at by careful listening to native-speaker talk; as
we have discussed before, computers simply make it easier to look at a lot
of data in one go, but one usually needs to have some idea of what sort of
thing one is looking for in order to use the power of the computer most
efficiently. The particular strength of computerised corpora is that they
offer the researcher the potential to check whether something observed
in everyday language is a one-off occurrence or a feature that is wide-
spread across a broad sample of speakers. This book does not take the line
that experienced language teachers are stupid and need to have their
eyes opened by the findings of academics investigating huge corpora
vastly beyond the means of practising teachers to emulate.
The CANCODE corpus confirms the common-sense intuition that
speech reporting is exceedingly common in everyday language. It also
demonstrates that the ways in which speakers effect reports are many
and varied. These overlap to a considerable extent with those which
fiction writersrecreate in their stories (see Page 1973 for a seminal
discussion) and which journalists use to report the words of politicians
and other newsworthy figures (e.g. Zelizer 1989; Waugh 1995; Thompson
1996). But spoken data also exhibit choices which are rarely, if ever,
found in written-text reports. What is most striking is that everyday
conversational resources for reporting are much richer than is suggested
by sentence-based accounts of the structure of direct and indirect speech.
In this chapter I shall take the terminology proposed by Genette (1988)
as articulating a convenient framework for the differentiation of basic
types of speech reports, viz:

Type Characteristics

Direct: Reconstructions of quoted speaker’s words, usually form-focused,”

syntactically independent from the reporting clause, e.g.:
I said ‘What would you like to do this afternoon, Dulcie?’
The punctuation here is for the sake of illustration; the original
CANCODE corpus transcripts are not punctuated in this way, and, in
a spoken corpus, many types of audible and contextual cues have to
be taken into account in attributing the function of direct speech
report to any given string of words. For the purposes of the present
chapter, the original tapes were consulted to confirm the attribution
of speech-reporting.
152 + 8 ‘So Mary was saying’

Type Characteristics

Indirect: | Reconstructions syntactically dependent on the reporting clause,

frequently accompanied by changes in tense, deixis, pronouns (see
Banfield 1982: 25); heavily context-dependent for interpretation, e.g.:
Cos a friend of mine, he asked me if he could stay there and I said ‘Yes’
It is impossible to know whether the original words were ‘Can/could I
stay here/there?’, but it is usually unnecessary for successful commu-
nication for the original words to be reproduced.

Narratised: Reports of an act of speaking, without speaker’s words being quoted,

summarising the event, e.g.:
Oh my doctor came and she rang up and complained at the way that I'd
been treated
Baynham (1991), who uses a similar tri-partite division of speech-
report types, refers to this third type as ‘the lexicalisation strategy’.

Lucy (1993: 18-19) makes the further distinction that direct speech
reports are made within the framework of the ‘reproduced speech event’,
while indirect reports operate within ‘the perspective of the reporting
speech situation’ and are relevant to ‘the concerns of the current event’.

8.2 Speech and writing

Let us begin with two rather extreme examples of data containing speech
reports. Extract (8.1) is from the classic novel Ivanhoe, by Sir Walter Scott;
(8.2) is from an anecdote in the CANCODE corpus. Ivanhoe is an extreme
choice, but around the world, many learners of foreign languages get
much of their main exposure to speech reporting from classic literature,
as I myself did in Spanish and French as a student. What is more,
although modern novelists writing in English often experiment with
alternative ways of depicting speech, the rather stilted (to our ears)
manner in which it is done in Ivanhoe is still alive and well in many
examples of popular fiction published in magazines:

(8.1)
Rebecca again looked forth, and almost immediately exclaimed,
‘Holy prophets of the law! Front-de-Boeuf and the Black Knight
fight hand to hand on the breach, amid the roar of their followers,
who watch the progress of the strife - Heaven strike with the
 8.2 Speech and writing - 153

cause of the oppressed and of the captive!’ She then uttered a loud
shriek, and exclaimed, ‘He is down! He is down!’
‘Who is down?’ cried Ivanhoe; ‘for our dear Lady’s sake, tell me,
which has fallen?’
‘The Black Knight,’ answered Rebecca, faintly then instantly again
shouted with joyful eagerness — ‘But no - but no! - he is on foot
again, and fights as if there were twenty men’s strength in his
single arm — His sword is broken — he snatches an axe from a
yeoman ~ he presses Front-de-Boeuf with blow on blow - The giant
stoops and totters like an oak under the steel of the woodman -
he falls — he falls!’
‘Front-de-Boeuf!’ exclaimed Ivanhoe.
‘Front-de-Boeuf,’ answered the Jewess; his men rush to the rescue,
headed by the haughty Templar - their united force compels the
champion to pause - They drag Front-de-Boeuf within the walls.’
‘The assailants have won the barriers, have they not?’ said
Ivanhoe.
‘They have! - they have! — and they press the besieged hard upon
the outer wall; some plant ladders, some swarm like bees, and
endeavour to ascend upon the shoulders of each other - down go
stones, beams and trunks of trees upon their heads, and as fast as
they can bear the wounded to the rear, fresh men supply their
places in the assault.’
‘Who yield? — who push their way?’ said Ivanhoe.
‘The ladders are thrown down,’ replied Rebecca, shuddering; ‘the
soldiers lie grovelling under them like crushed reptiles - The
besieged have the better.’

(8.2)
[The speaker is recounting how, while on holiday with a friend, they
accidentally ended up as extras in a film, instead of playing bingo, which
they thought they were queuing up for.]

<S 01> So we’d been wandering round in the morning doing the usual
thing came back and had lunch and she, I said what would you like
to do this afternoon Dulcie she said oh Mary let’s go to bingo, now
bingo is never ever my cup of tea [<S 02> no] but seeing that I was
supposed to be, with her
<S 02> lSupporting her yeah.
154 + 8 ‘So Mary was saying’

<S 01> I’d to fall in with her [<S 02> [laughs]] all right then Dulcie where
do we go now to bingo, I don’t know she said but we'll find out so we
walked along and we saw this hall and she said I think that’s it so I
saw a lot of people and I said I don’t know Dulcie, doesn’t look like a
bingo hall so she said well go in the queue she said and find out
what’s happening so I go in this queue and I’m waiting so I saw
them taking names and writing things down so I had this feeling I
was in the wrong place [<S 02> [laughs]] so I thought to myself oh I’m
going from here but as I was stood at the table this person said er
now then you’re next so I said excuse me is this the bingo hall and
he said no my dear oh so I said oh I’m sorry and I started to walk
away but he said hang on a minute he said erm how would you like
to be an extra | said an extra for what [<S 02> [laughs]] he said for a
film he said [<S 02> no] we’re doing a film [<S 02> get away] so I said
me he said yes he said [<S 02> yeah] want lots of people [<S 02> yeah]
so I said oh I can’t really I’m sorry because I’ve got my friend with
me [<S 02> yeah] that’s all right he said ask your friend to come up
so I said well before you take any more details Ill ask her to come up
with me now, so I go back to Dulcie and she says all right Mary is,
will the bingo be starting soon I can’t see any chairs and tables no I
said we’re in the wrong place Dulcie [<S 02> [laughs]] and I said
they’ve asked us if you’d like to, like us to be in a film what d’you
mean she said well I said don’t know the story as yet [<S 02> mm] I
said but erm but I said I think it’ll be a laugh oh she said I'd love it.

Both extracts are rich in speech reports. Both in different ways stage the
speech reported as if it were happening before the reader’s/listener’s
eyes. Both are a conventional fiction in the sense that neither is likely to
be a true and accurate account of someone’s words.? The written text is
fictional because the characters are fictional anyway; the spoken text
(although told as a true story) is also a fiction, because we know the
speaker can do no more than attempt a reconstruction of what (s)he
considers the characters’ important original words to have been, even if
direct speech reports may be argued to be as less of a distortion of the
original words than the necessary paraphrasing of indirect speech (see
Coulmas 1985a). Indeed, it hardly matters whether the words are a true
re-statement; spoken storytellers’ reports are open to challenge, but are
usually only challenged if they stretch credibility or if another witness
to the original event(s) chooses to question their accuracy. The Ivanhoe
 8.2 Speech and writing + 155

text represents a particularly traditional and stylised type of speech

reporting, and I would reiterate that I certainly do not suggest that it is
typical of modern fictional literature. Most readers will be familiar with
writers who report speech by a mixture of direct speech and indirect
speech, and either by using the kinds of speech reporting verbs seen in
both extracts above or by simply indicating characters’ speech by using
quotation marks or paragraph indentations,
without any reporting
verbs. The two extracts are chosen purely for illustration of how wide
the gap can be between written and spoken speech reporting. Let us
consider what features make them so different. The following lists show
the use of reporting verbs and any accompanying modification in each
extract:

Ivanhoe text

exclaimed (3 times)
uttered a loud shriek
cried
answered faintly
shouted with joyful eagerness
answered
said (twice)
replied shuddering

Spoken story

said (29 times)

says
asked

The two extracts are not the same length (being approximately 300 and
just over 400 words respectively), but it is clear that the written text
avails itself of a range of reporting verbs and of adverbial modifiers
which add specific characteristics to the reported words (e.g. acoustic
parameters: loud, faintly). The spoken text seems to rely almost entirely on
the verb say, though it does vary its tense, using present simple on one
occasion, and an indirect report with ask embedded in a direct report.° If
we take the reporting verbs used in the Ivanhoe passage, we find that
their occurrence in the conversational corpus is quite different. In just
156 + 8 ‘So Mary was saying’

over one million words of CANCODE data, the verb exclaim does not occur
at all, and utter only occurs once in a quasi-speech reporting function,
when a speaker is attributing authorship to a pun that crops up in a
casual conversation.® Cry never occurs as a speech reporting verb, and
answer never occurs reporting direct speech. There are 250 examples of
reply/replying/replies/replied in the corpus, but all but two refer to written
replies to letters. The verb shout occurs 53 times in the corpus, reporting
speech both directly and indirectly. Five illustrative examples occur in
one conversation where a woman is recounting her experience giving
birth, surrounded by people urging her on:

(8.3)
<S 01> I can remember them all shouting at me to push and I was
getting so fed up with them all like, that girl, that, Imean she didn’t
even know me and she was telling me to push and I was thinking
what’s it got to do with you you know and Doctor Hill’s shouting in
the other ear and they were just, meant [<S 02> mm] nothing to me
... but like Nancy Carr was really, she was being horrible really cos
she had to to make me do it and afterwards she said like you done
really well and she said [<S 02> mm] sorry if I shouted at you ... but
<S 02> LWho who was
being horrible to you, Nancy Carr?
<S 01> Nancy Carr like shouting at me to push and that
<S 02> Mhm.
<S 01> And she said afterwards em sorry if I kept shouting at you but if
she just said push like that you wouldn’t bother would you you
wouldn’t try.

It will be noted that shout, as well as reporting indirect speech with the
original words reconstructed (e.g. shouting at me to push and that), also
occurs in narratised reports (and M’s shouting in the other ear; sorry if I
shouted at you). When used for direct speech reports (in seven cases), shout
is used in the continuous (-ing) aspect, which may be a reflection of the
greater vividness generally found in conversational speech reports
(Tannen 1986), for example:

(8.4)
[<S 01> and <S 02> are describing how one of their children fell off a wall
and was injured by a loose brick.]
 8.2 Speech and writing + 157

<S 01> The doorway opened and we heard somebody shouting help me
and then we seen him like crawling in but he had er pla, era plaster
thing on his arm didn’t he.
<S 02> Yeah for about a week.
<S 03> Yeah.
<S 02> To make sure there was nothing broken.

(8.4)
[<S 01> is describing a female neighbour who has learning difficulties. |

<S 01> She had somebody knocking at her windows shouting fire fire
and it was just a ruse to get her out the house you see.
<S 02> Mm.
<S 01> And er she was very sensible the old lady was she phoned.
<S 02> Good.
<S 01> And how we heard about this it was the following morning the
window cleaner came I told him about it he couldn’t clean the
windows detectives were there detectives come to that’s how we
heard about er that.

Thus we see that a spoken conversational corpus may well produce a

different set of usages compared with the kinds of speech reports found
in classic or popular fictional literature, with the Ivanhoe text displaying
well-documented literary devices (see Tannen 1988 for further examples
and a discussion). As noted above, although writing styles have changed
since the time of Sir Walter Scott, the kinds of graphic speech reporting
verbs and adverbial modifiers illustrated in the Ivanhoe extract are still
in common use, though perhaps more so in popular fiction (e.g.
magazine romances; see Nash 1990: 29-34 for examples; see also
Oostdijk 1990 for a general discussion on natural speech features versus
fictional speech) than in canonical literary works. As Hughes (1996: 49)
points out, literary conversations ‘are constructed by writers for
readers, rather than by interlocutors for one another’, and so we can
expect the literary author, with all the compositional time and reflec-
tion available, to exploit the full potential of reporting verbs and
adverbials to constrain the reader’s interpretation of characters’ spoken
words.
But does (8.2) above suggest that speech reporting in everyday conver-
sation may be expected to be an anodyne affair, endlessly repeating an
impoverished range of options? If so, then the teaching of speech
158 + 8 ‘So Mary was saying’

reporting in spoken contexts would seem a simple task, perhaps concen-

trating on a limited repertoire of declarative reports involving the verb
say, with further explorations into verbs such as tell, ask and shout at a
later stage. To answer the question, we need to look at more data, and to
take another, closer look at the strategies used by the speaker in extract
(8.2). We may list the following features of the speech reports in (8.2) as
potentially relevant:

1 The teller uses say not only to report statements, but questions too,
e.g.: I said what would you like to do this afternoon Dulcie (reported
question)
she said oh Mary let’s go to bingo (reported statement/response)
2 The teller uses the zero-quotative option (i.e. simply reporting speech
without any reporting verb or explicitly naming the speaker), e.g.:
all right then Dulcie where do we go now to bingo
We may note that this also occurs in the Ivanhoe passage, in the
paragraph beginning ‘They have! - they have!’. Zero quotatives are
usually not problematic: the simple rules of two-party turn-taking
which indicate ‘next speaker’ usually allow the receiver to interpret
who is being quoted (see Mathis and Yule 1994), along with general
semantic/pragmatic assumptions made by the listener that attribute
the speaker’s reports to sources already indicated, unless told other-
wise (see Palacas 1993 on ‘attribution’ semantics). In the written text,
Ivanhoe is attributed as speaker before and after the zero-quotative
turn, thus we assume it is Rebecca speaking, and in the spoken text,
the teller names the addressee (her friend, Dulcie), thus we assume it is
the teller quoting her own words.
3 The reporting verb may be placed initially, medially or finally, or in
more than one place in quoting any particular turn-at-talk:
she said I think that’s it (reporting verb initial)
I don’t know she said but we'll find out (reporting verb medial)’
what d’you mean she said (reporting verb final)
so she said well go in the queue she said and find out what’s happening
(reporting verbs initial and medial). Initial, medial and final place-
ment, though not combined, also occur in the Ivanhoe passage.
4 The reporting verb may be in historical present tense (i.e. a present-
tense form reporting a past event), just as other verbs in the narrative
may switch to historical present, e.g. so I go back to Dulcie and she says all
right Mary is, will the bingo be starting soon.
 8.3 Reports with -ing form reporting verbs + 159

5 The reported words attempt to resemble natural conversation as

closely as possible, for example, discourse markers are often included
in the report in the way they might be expected to occur in real
conversation:®
this person said er now then you’re next
I said oh I’m sorry
so I said well before you take any more details

So it is that, even in this short extract, we can see a variety of strategies

being adopted, even though only one verb is predominantly chosen to
launch the speech reports (say). We have also, in passing, observed how
shout has been used in natural conversational contexts. It now behoves us
to look wider and seek to account for whatever other strategies and
usages we can find in our conversational corpus.

8.3 Reports with -ing form reporting verbs

We have noted that shout occurs in the -ing form in a number of examples
of speech reporting. If we look for further examples of reports with the
-ing form, we find a surprising number - surprising, that is, given the fact
that almost all grammars, language-learning textbooks and research
articles dealing with speech reporting seem blissfully to ignore the
phenomenon. In just over one million words of CANCODE data, for
example, was/were saying occurs 136 times in speech reports, always
framing indirect reports. Examples include:

(8.5)
[Speaker is commenting on a reference to the death of British Labour
Party Leader John Smith in 1994, an event which caused great sadness
among Labour Party supporters.]

<S 01> Caroline was saying she still feels like shedding a tear when she
thinks of that.

(8.6)
[Speaker is commenting on the way her water-lily flower opens and closes
regularly.|

<S 01> I was saying to Kevin they’re a very unusual flower, they must
have some kind of time-clock.
160 + 8 ‘So Mary was saying’

(8.7)
[Speaker is passing on a bit of village news to a neighbour.]

<S 01> Brian was saying the village hall nearly caught fire last night.

The same occurs (13 times) with tell, as in (8.8) and (8.9):

(8.8)
[Woman commenting on the problem of tick infestation in dogs: the
‘border women’ is a reference to two women who own border terriers, an
English breed.]

<S 05> Those border women, I should know their names, she was telling
Colin one of them had a tick it was like a bluebottle.

(8.9)
[Speaker is commenting on the way firms respond to job applications.

<S 01> Er my son is with the Electricity Board

<S 02> Yes.
<S 01> And he was telling me that they have this sort of procedure as
well
<S 02> Mm.
<S 01> They have to reply er initially within, well I can’t remember the
number of days.
<S 02> So two or three working days.
<S 01> That’s right that sort of thing.

A similar pattern is found in (8.10) with read and suggest:

(8.10)
[Speakers are discussing French politics.]

<S 01> Yeah I mean I think it would be true to say as Steve’s indicated
there is er a hell of a battle still going on because it has been the
domain of the President although what the President’s up to er I was
reading in the paper yesterday that
<S 02> lLame duck President.
<S 01> Well it wasn’t so much lame duck it was close to that but they
were suggesting that if people looked at Mitterand’s private life er
you know the the sort of thing that’s been published in the British
press lately’d be pretty tame.
 8.3 Reports with -ing form reporting verbs + 161

What do these speakers mean by using the -ing form in their reporting
verbs, and why should it be that linguists and language pedagogues have
mostly ignored this uncontroversial, very natural-sounding usage? It
seems that our speakers are de-focusing from the actual words uttered by
the original speakers and focusing instead on the content, in terms of its
newsworthiness or topical relevance; the speech report may simply be
there to introduce a new topic or argument. (I am grateful to Jeff Stranks
for some illuminating insights here; personal communication) This may
help us understand why the types of written texts often chosen for
research on speech reporting are unlikely to exhibit such functions.
Firstly, one thing that needs to be considered is the differences in the
exigencies of speech reporting in different contexts. Waugh (1995) has
made the point that journalistic speech reporting focuses on ‘conveying
information and [is] concerned with issues of referentiality, truth,
reliability and accountability’. It is thus not surprising that a major study
of reporting such as Thompson (1994), which uses a good many journal-
istic examples, does not include past -ing reports, since the past -ing takes
the focus off ‘reliability’ and ‘accountability’. Some spoken speech
reporting is pre-occupied with the same demands as journalistic re-
porting, for example, the courtroom data in Philips (1985), which,
although the data contains indirect reports, have none in past -ing form.
Also, the vividness and ‘real-time’ staging of speech reports in oral
narratives (especially during narrative ‘peaks’ or climaxes; see Larson
1978: 68-76; Coulmas 1986), may push storytellers towards reinforcing
the fiction that they are reporting faithfully their protagonists’ words.
This may explain the fact that, in the narrative texts in the CANCODE
corpus, speech reports are overwhelmingly direct speech, and with
reporting verbs in past simple (said, told) or historical present says. But a
great deal of spoken language is not concerned with faithful reproduc
tion of speech, or even creating the illusion of it. Speech is often reported
indirectly in casual conversation either as a topic-opener (8.7), or simply
in support of some point being discussed or made (8.9), where the
demands of veracity and the faithful reproduction of words spoken is of
secondary importance. For that reason our Irishman in the anecdote at
the beginning of this chapter greets all news with ‘So Mary was saying’:
he wants to impress his audience that he has already heard the news, not
the words. The past -ing form reports in the CANCODE corpus are genre-
restricted, and seem to belong to more general, non-narrative, casual
conversational contexts where they serve to signal topic management in
162 + 8 ‘So Mary was saying’

the ways illustrated in extracts (8.5) to (8.10), and contrast with the
‘focus-on-words-uttered’ function of the past simple reports. It is unlikely
that a corpus biased towards journalistic or legalistic texts will yield the
kinds of -ing form indirect reports illustrated in any significant
number.’° It is even more unlikely that speech reporting approached as a
phenomenon of the sentence (as it is in the works of many researchers)
will be considered in this essentially interactive way. Much of the work
done on speech reporting has preoccupied itself with the syntactic
principle of ‘backshift’, i.e. how speech displaced from its original time of
utterance is normally reported in the past (or how it is understood: see
Boogaart 1996), such that direct quotations like ‘I’m going home’ tend to
be backshifted to ‘(s)he said (s)he was going home’. Classic studies such as
Coulmas (1985b), Comrie (1986), Goodell (1987) and Huddleston (1989),
all concern themselves with backshift and pay no attention to the
possibility of past tense -ing reporting verbs. From a pedagogical point of
view, there would seem to be no justification for excluding the -ing form
reports, especially since the same phenomenon occurs in other languages
(e.g. French, Spanish). From a broader theoretical viewpoint, the lesson of
this section is that discourse grammars should not be just an account of
the above-sentence behaviour of conventionally described structures, but
must also be prepared to encounter and explain structures not pre-
viously observed or discussed within the canon of grammar for any
particular language.

8.4 Other reporting verbs

So far we have concentrated mainly on say, and some additional

examples with tell, read, shout and suggest. Apart from say and tell, another
frequent reporting verb in the corpus is ask (which appeared in an
embedded report in extract (8.2), and it is worth considering it in some
detail, for it tells us a lot about conversational speech reporting in
general. There are more than 700 reports of various kinds with ask in the
corpus. Only four of them are direct speech reports;!! all the rest are
indirect or narratised. By far the most frequent category (just over 200) is
the speaker reporting his/her own acts of asking. The most frequent
structural collocations following asked/asking are for, to, if and about, in
descending order of frequency. Some examples are of the typical kinds of
patterns practised in language classrooms, especially those with wh- and
if reported clauses, where word order can be a difficulty for the learner:12
 8.4 Other reporting verbs + 163

(8.11)
<S 01> I phoned up the hospital and asked who I should address the
letter to.

(8.12)
<S 01> Then I saw Mark Porter and asked if he’d seen you he says yeah
he’s been playing snooker with me all afternoon.

(8.13)
[Speaker is talking about the arrangements for a funeral.]

<S 01> So I says em, well you don’t interfere do you I mean so I asked
him what the arrangements were oh there’s a chapel of rest in the
village em he says and I want to get him moved to the church, I said
but aren’t people going from the house ... all the wreaths came to
the house but there was no hearse.

Extract (8.13) makes the point once again that questions do not have to
be reported with ask: the speaker’s second question about the funeral
party going from the house is introduced by said, just as we saw in
extract (8.2).!° The other very frequent types of reports with ask, illus-
trated in the following examples, are of types that may well suffer from a
lack of attention in the language class or else not be considered as
significant aspects of speech reporting:

(8.14)
Feature(s): ask (sb) for sth

[Speaker is recounting a conversation with the doctor-]

<S 01> And I’ve asked him for water retention tablets.
<S 02> Mm mm.
<S 01> But they wouldn’t give ’em me, erm J had an operation my
stomach just kept going up and down didn’t it bloating up and then
going down.

(8.15)

Feature(s): (a) ask sb to do (b) passive voice

[Speaker talking about working in the health service.

164 + 8 ‘So Mary was saying’

<§ 01> And erm I'd heard one or two bad things about it all about these
on-calls and things and you got really tired and and er I sort of was,
the next thing I was asked to do this job and I didn’t have any
choice in the matter.

(8.16)
Feature(s): (a) ask about sth (b) unrealised (irrealis) speech reports

<S 01> You know, if you go to the doctor’s er for something and then
[<S 02> mm mm] you come out and you haven't asked about it
[laughs] you put off asking don’t you.

As well as the varied patterns of complementation in examples (8.11) to

(8.16), the occurrence of passive and of potential, non-realised or irrealis
speech reports has been mentioned. 31 passives or pseudo-passives with
ask occur in the corpus, and the most consistent pattern is I (be/get) ask to
(do), as represented in these concordance lines:

Initially the r= one of the main reasons that | [[was asked]] to sit on the group was em ev everybody
one of my objectives one of the things that | [[was asked] to achieve within the first year was to estab
Normally a manager does it but you know | [[was asked] to do it. And em it’s only comparatively rec
<\$=> and er | sort of was the next thing | [[was asked]] to do this job and | didn’t have <$O48> a
<$=> Em <\$=> <$?> Oh you know I’ve [[been asked]] to do some G C S E English <$H> next te
a university+ <$2> Ah yes. <$1> +and I’ve [[been asked]] to do this study. <$2> Mm. <$1> So my
<$1> +<$04> because what <\$O4> I've [[been asked]] to do is to talk to people who've had so
<$1> <$03> Yeah. <\$O3> <$2> And | [[got asked]] to do it cos they wanted er you know a go

Figure 1: Concordance lines for I (be/get) ask to (do):

With this glance at ask, and with the other verbs reviewed, it is now
becoming clear that the range of speech reporting in conversation is
both syntactically and pragmatically richer than our initial spoken
extract (8.2) might have suggested. As well as conventionally treated
direct and indirect reports we have seen reports with -ing, passive voice
reports, reports
of speech events not actually realised, variation in
positioning of reporting verbs, and so on.
Another reporting verb that occurs, in direct reports only, is go. Some
examples follow:
 8.4 Other reporting verbs - 165

(8.17)
[Young women talking about shaving their legs.]

<S 01> I didn’t shave mine for a week when I went to Crete cos I thought
I'd get browner if I had hairy legs ... and my sister told everybody
when we went out at night that I hadn’t shaved.
[3 turns later]
<S 01> Cos I was embarrassed and when we were out and I had a dress
she went look at her legs she’s got hairy legs.

(8.18)
<S 01> I can remember getting to the customs in America and this guy
went where are you staying, when, I went with with my friend
she went
<S 02> lHow much money have you got.
<S 01> lAre you going out with him.
<S 03> Yeah yeah.
<S 01> And I went no, are you sleeping with him I thought would be the
next question and she went, no she said are you planning to get
married I went no she said oh you're definitely leaving after the end
of the year.

(8.19)
[Speaker is recounting an incident in a bookshop when her friend only
vaguely knew what she was looking for.|

<S 01> And erm there’s a new map out or something accompanies this
book Sue was going in like we went in and it was just art books and
we said oh d’you do sort of fantasy books or something Sue said and
I was going oh God like you know and he was going oh what what
books did you want and it was kind of like bit embarrassing really
he was going oh what is it a medical book or something you know
like no no

All the speech reports with go in the corpus are by young speakers under
30 years of age,'* and all seem to be in contexts where the maximum
amount of dramatic/graphic representation is attempted, often with
mimicry in voice quality or other paralinguistic features. Tannen (1988)
reports that in her spoken narrative data, go was the most frequent
speech reporter (reporting verb) after say.
166 + 8 ‘So Mary was saying’

8.5 Tense and the reporting verb

Although say is by far the most frequent reporting verb in the corpus and
although its simple past form said (with just under 2,000 occurrences) is
by far the most frequent form, other tenses and forms also occur
frequently. There are 113 cases of the (grammatically anomalous) histor-
ical present form I says, characteristically occurring in narrative reports.
Typical of these are the following, one of which is extract (8.12):

(8.20)
[Speaker has contracted shingles and is recounting her interview with
the doctor.]

<S 01> I asked I said is it contagious? she says no she says no you know
it’s children’s stuff she says you know the chicken pox
<S 02> lYeah
chicken pox.
<S 01> I says well I haven’t been anywhere where there’s been any
children I don’t know how I’ve got it.

(8.21)
<S 01> Then I saw Mark Porter and asked if he’d seen you he says yeah
he’s been playing snooker with me all afternoon I was so mad these
chips went up this wall and the language was
<S 02> {laughs]

(8.22)
<S 01> Bumped into his mum coming out did you find him she says I
says no I didn’t [laughs] and she went barmy then cos you used to
didn’t you.

Johnstone (1987) suggests such switches to the historical present on the

reporting verb may not be random and may coincide with quotations of
authoritative speakers’ words, making them stand out from others’
words. It certainly seems that historical present signals the fore-
grounding of the quoted speech in a way that (unmarked) past simple
does not; it is thus an extremely useful discourse strategy in narratives
where speech reporting is frequent, especially where there is rapid
interchange of speakers’ words which includes the narrator’s reports of
his/her own words.
 8.6 Discoursal functions of as I/you sayreports + 167

Present tense say(s) also occurs frequently when reporting speakers’

words that relate to permanent facts or truths, as well as to things
speakers have said which are still relevant or important:15

(8.23)
[Speakers are assembling a portable cot borrowed from a friend.
<S 02> has been given verbal instructions by the friend (‘she’).]

<S 01> It’s not as difficult as it first seemed.

<S 02> She says you've got to twist these round and it makes them solid
or something.

In (8.23) there is no implication that the instructions were given more

than once, simply that what the friend said is relevant to what the
speakers are now doing. Similarly in (8.24), the woman’s reported words
are still (and permanently) relevant:

(8.24)
[Speakers are discussing someone who wants to try out archery just once;
the difficulty is that normally one has to sign up ior a whole course.]

<S 01> That was what I was trying to get over to her on, it’s something
that you can’t do one-off ... although this woman at Marksman
Bows says they will do an hour’s individual tuition for a one-off
visitor to give them a taste of it

8.6 Discoursal functions of as I/you say reports

I have argued that forms such as the past continuous, historical present
and present simple represent discourse strategies whereby speakers
exercise control over entities such as topic, foregrounding and relevance.
The expression as I/you say is also an important way of using speech
reports to manage the discourse. In the one million word CANCODE
corpus sample, as I say occurs 170 times, and as you say 29 times. The
expression uses present simple tense to refer to recent speech, either
168 + 8 ‘So Mary was saying’

(8.25)
[<S 01> has been talking about how he and his wife built their own
garden pond.

<S§ 01> Then you put your concrete on top of that.

<S 02> LConcrete on top of that ... but a thin
layer just for flooring.
<S 01> We put about four inches on.
<S 02> Yeah.
<S 01> Three to four inches.
<S 02> Oh right ... oh that’s interesting.
<S 01> But as I say we did it ourselves.

(8.26)
[<S 01> and <S 02> are telling <S 03> about a building that they feel
should be conserved but which is to be demolished and replaced by a
petrol station, despite the fact that there is already a disused, boarded-up
petrol station in the same road.

<S 01> And it should’ve been a listed building but nobody listed it ...
well the council have sold it to a garage company and it’s going to be
knocked down and a petrol station
<S 02> |Petrol station
<S 01> Lwas going to be put there
so.
<S 02> It seems we’ve got one five hundred yards down the road
that’s been
<S01> ‘| That’s empty
<S 02> lboarded up cos it doesn’t make no money.
<S 03> Mm.
<S 02> We don’t see the point in having it at the bottom of the street.
<S 01> But em the museum were very dischuffed that the council didn’t
let them know because I think if they had they could have had it
listed and nobody would have been able to touch it ... so that caused
a lot of trouble didn’t it I mean.
<S 02> lYeah.
<S 01> lwith the petition and everything and
it went to the council but the council still passed it.
<S 02> Well they said it was too late didn’t they.
<S 01> Yeah so it’s been passed and
 8.6 Discoursal functions of as I/yousayreports + 169

<S 02> Just waiting for it to be knocked down and built.

<S 03> But as you say the garage on the main road has been boarded up
for some time.
<S 02> It’s been boarded up now for nearly seventeen months.

Support for the interpretation of as I/you say as having a discourse-

marking function comes from the added presence of but in these two
examples. But is the single most frequent preceding collocate of as I say
(occurring 30 times), and is widespread in conversation generally as a
marker of topicsummarising or topic-shift.1° As I/you say differs from
similar expressions with I/you was/were saying, which tend to take the
conversation back to an earlier point in order to (re)develop a topic entity
or add to it, rather than summarise or close it, for example:

(8.27)
[Speakers are discussing the closure of railway lines in Britain and the
nostalgia for the age of steam railways.]

<S 01> There’s a, there’s a hell of a big following of steam trains.

<S 02> LYes oh imagine yes oh
yes oh yes what did I say just now, I, yeah but some of that’s
nostalgia cos they’ve gone away.
<S 01> Yes.
- <S 03> Yeah yeah.
<S 02> But I was saying em earlier on do you remember [<S 01> mhm]
that em it’s a criminal waste what they've done the destruction is it’s
criminal you know.
<S 03> lYeah yeah yeah
yeah yeah.
<S 02> It’s theft really.
<S 03> Yeah.

(8.28)
[Speakers are discussing the desirability of learning foreign languages. ]

<S 01> Mm well yeah because they’re going to be at a disadvantage aren’t

they in terms of the business world and er you know for, so from
that point of view I think it’d be a good idea that that people did
learn another language.
170 + 8 ‘So Mary was saying’

<S 02> Do you think it should be do you think they should

<S 01> lLanguage.
<S 02> Back to what I was saying before a bit although I’m I’m trying
not wanting to force that idea but would it be an idea if all
Europeans learned one second European language do you think, not
necessarily English.
<S 01> Well you can you can never tell what the circumstances are going
to be can you erm [sighs] yeah erm in, French and German seems to
be the the two languages that spring to mind as being used most
widely in Europe so I guess French if you can speak French German
and English maybe.
Close observation of real data can bring out the rich variety of discoursal
functioning of speech reporting strategies, and there are undoubtedly
many more which could be observed; the limits of this book mean that
we can only discuss the most frequent and salient types here. We shall
conclude this section with some examples which are presented simply
for observation. Extract (8.29) is taken from a conversation between two
young women who are checking and confirming arrangements made
with a third party for a social visit. The discourse strategy in focus here is
that of checking; what is of note is the variety of syntax and complemen-
tation which may not occur in formal written texts. Extracts (8.30) and
(8.31) reiterate the point (see also note 11) that canonical or unmarked
word-order (in this case that of indirect speech) may be disturbed in real-
time speech without any disturbance to communication:

(8.29)
<S 01> What we could try and do is ring Ali again try and ring her again.
<S 02> Yeah we will.
<S 01> When we're in Tunny Wells cos the thing is erm did you say to
her what time.
<S 02> No I never mentioned what time.
<S 01> So you said well
<S 02> |[Inaudible]
<S 01> What did you say, we'll be up in the evening?
<S 02> Yeah I think she thinks we're gonna be up in the evening.
[later]
<S 02> Yeah I think she thinks we’re gonna be up in the evening.
<S 01> Cos did you say to her about dinner or
<S 02> Yes.
 8.7 Drawing the arguments together - 171

(8.30)
<S 01> Yeah this woman also told me on Monday this chap came in and
he was hanging around for ages and he started asking me how you
got into tourism. You know what qualifications [laughs] you needed.
<S 02> [laughs] He was in the middle of an enforced career change he
said.

(8.31)
<S 01> Do you know how many students they’re expecting in July.
<S 02> Yeah thirty-something though we're likely next term to have
sixty-four students a lot of them may be six months and a lot of
them will be Japanese here in August from what I’ve heard, about
seventeen of them.
<S 01> Yeah.
<S 02> And Liz is going to join them she tells me.

8.7 Drawing the arguments together

We can now review the kinds of insights that might change our approach
to the teaching of speech reporting. The following points, I suggest,
would be important for a pedagogical model arising out of the observa-
tion of authentic spoken data:

1 The graphic speech introducers found in traditional literary reporting

styles (including contemporary popular fiction) may occur only rarely
in conversation, or not at all. Nonetheless, there is variety in the verbs
used in conversational reporting.
2 All our examples so far have indicated that the kinds of adverbial
phrases that specify the context of reporting verbs (as in the Ivanhoe
text) do not occur with any significance in conversation.
3 Past simple is not the only tense found on reporting verbs, even
though it may be the most frequent in some genres. Past continuous
for topic management in casual conversation and _historical present in
narration occur with a frequency that cannot be ignored.
4 Although the position of say in direct reports varies between initial,
medial and final in relation to the words reported, no other speech
introducer displays this mobility and all other direct-report verbs are
initial. However, word-order may vary from what is conventionally
considered to be correct or canonical in indirect reports (e.g. verb-final
172 + 8 ‘So Mary was saying’

indirect reports, retention of interrogative word-order in reported

questions).
5 The common literary device of inversion of reporting verb and subject
(e.g. ‘The Black Knight,’ answered Rebecca) does not occur in the spoken
corpus.
6 There is great syntactic variety in spoken reports with ask, including
passive voice and different complementation patterns.
Unrealised speech is reported, as well as actual speech.
In the most general sense, it is clear that speech reports are best
observed in their real contexts of utterance. It is there that we see their
roles in the discourse, realising functions such as foregrounding,
dramatisation (Mayes 1990; Baynham 1996), the creation of relation-
ships (Alvarez-Caccamo 1996) and the general management of the
interaction.

What these points mean for teaching may be translated into a set of
principles:

1 Teaching speech reporting should not be over-obsessed with backshift

and sequence of tenses with indirect speech at the expense of the rich
variety of tense and aspect forms that real data throw up.
Word-order (e.g. in reports of questions) should be seen as more flexible
than traditional prescriptions might suggest, and some word-orders
frequently marked as errors (e.g. ‘she asked me what was I doing’)
should not be considered erroneous since they regularly appear in
Nnative-speaker speech.
Once again, the principle of language awareness seems important
here: exposure to real data (without the necessary implication that it
has to be reproduced by the learner), will enable the learner to become
aware of forms such as go and the historical present which they are
likely to hear in the speech of native speakers, both in real-life
encounters and on the increasingly global media of films and tele
vision in English.
Frequent but tricky forms such as passive-voice reports (e.g. I was asked
to take part in a survey) may need a good deal of practice, and contrastive
exercises bringing out the different contexts of past simple and past
continuous reports may need to be designed to supplement what is
found in textbooks.
Although conversational speech reporting differs from literary speech
reporting, literary texts and conversational transcripts could be used
 8.7 Drawing the arguments together - 173

together, both in the teaching of literature and in teaching the spoken

language, their potential for contrast being a useful means of pro-
moting language- and literary-awareness.

As is the case with so many aspects of language surveyed in this book, the
willingness to confront spoken data wherever possible, and not simply to
rest on the assumption that what written texts tell us is sufficient for
pedagogy, is likely to pay dividends in terms of authenticity of material
in teaching and in a much better preparation for the learner for
encounters with users of the target language in its spoken forms outside
the classroom.
This book has come to its end. Over the eight chapters I have attempted
to trace out a landscape illuminated here and there by glimmers of
insight which the access to a spoken corpus has given me. A book such as
this can only exemplify from a relatively small amount of the language
that surrounds us each day, and can only base its conclusions on a
limited number of language features. But some things emerge time and
time again, whichever words or structures we look at, and these are that
face-to-face interaction brings the people who use language slap-bang
into the centre of investigation; it is simply impossible to idealise the
data away from who said it, to whom, at what point, with what apparent
goals and purposes, in the context of what relationship, and under what
circumstances. Equally present have been concerns such as the mutual
protection of face, the desire to converge socially, the joint construction
of meanings and of generic activity, and the active roles of listeners. Thus
the status I am advocating for the spoken language within applied
linguistics, whether pedagogical applied linguistics or the many other
branches of our profession that apply linguistic insights, is a fundamen-
tally humanistic one; it is also one that respects language learners’ status
as human interactants themselves, both in their L1 and in their target
language(s). Everyone is a ‘native speaker’ of something, and everyone
knows what interaction is like, when it is successful, when it is frus-
trating or when it fails. Every learner I have ever taught wants to succeed
in interaction. That does not necessarily mean they want to sound like L2
native speakers. It is rather to say that they want to do what L2 native
speakers do and what they themselves do in their L1, and to learn the
ways of doing in the target language, from which they can make their own
choices. The more we know about how interaction is done in the
language(s) we teach and/or learn, and the more we can clear our minds
174 + 8 ‘So Mary was saying’

of presuppositions that the spoken language is just like the written but a
bit sloppier and more prone to ‘mistakes’, the better it will be for all of
us.

Notes

1 I owe this information to Barry O’Sullivan, a native of Tipperary, and at the

time of writing a lecturer in English at Okayama University, Japan.
N Lucy (1993: 18) observes that direct speech reports typically foreground ‘the
original form of the utterance’ [his emphasis] as well as conveying ‘its
substantive message as well’.
Clark and Gerrig (1990) argue that speech reports are ‘selective depictions’ of
original utterances.
Reyes (1984: 65) makes the point that quoting others’ words in no way releases
the reporter from the normal responsibilities of communication; the
reportee’s and reporter’s voice are heard together:
Citar ... no exime de la responsabilidad de la intenci6n comunicativa;
suscitar otra voz no es perder la propia, repetir es decir, en la medida que
sea. [her italics]
ul Bauman’s (1986: 66) data confirms the overwhelming preference for say as the
reporting verb.
In the original data extract, reproduced here, <S 01> is telling <S 02> about a
recently received teaching job offer:
<S 01> She’s left me two days to decide.
<S 02> As in to do or not to do.
<S 01> That is the question.
<S 02> That’s a good line yes.
<S 01> Yeah [laughs].
<S 02> Yes yes it’s
<soi> lI thought of it first [laughs].
<S 02> It was Match of the Day nineteen sixty seven I think isn’t it as
uttered by Jimmy Hill when when commenting on Brentford against
Halifax in the er in the F A Cup
N Person (1996) makes some interesting observations on the literary device of
‘restarts’ in direct speech reports. These are not dissimilar to the medial-
placement of the reporting verb in spoken reports.
ie.) Bauman’s (1986: 67) data once again confirms the frequency of this phenom-
enon as a speech-reporting strategy. See also Aijmer (1987) on oh and ah in
natural conversation.
There are two exceptions to the neglect of the -ing form reporting verb that I
have been able to find. Leech (1987: 31) gives -ing form examples of indirect
reports with read, tell and say but is rather dismissive of them as colloquial
‘exceptions’ to the general rules of progressive aspect, and as representing a
 Notes + 175

stage of instability in the language. Eastwood (1994: 353-4) has examples of

ask (in past and present continuous) and wonder (in present continuous) in the
section on reported speech in his pedagogic grammar, but lets them pass
without commenting on the tense/aspect choice. Ely et al’s (1995) data includes
an occurrence of was telling, but it passes without comment.
10 Polanyi (1982) makes an interesting distinction between stories and reports;
the latter, considered as neutral reconstructions of verbal events, do not have
to justify their tellability.
11 For example:
<S 01> So I actually stood there, to the doctor and I says well no I’m not
happy you see cos, he asked me you know sort of are you happy
with the result and I says well no I says I don’t feel as if they’ve
improved or whatever I says they’re still aching and I've still got veins
there
All four direct reports with ask have initial-position reporting verb.
12 Though it must be noted that native speakers do not always observe the ‘rules’
of word-order in indirect reports, for example in this extract, interrogative
word-order is retained in an indirect report:
[Speaker is talking of an experience while staying in hospital.
<S 01> You know the sweetener, erm I asked one of the cleaners could
she get me something because it was on a very bad day and she had
to ring down for permission for somebody else to get one and then
when she brought it we all had it sort of shared at the top end of the
ward so that everybody could use it
13 Indeed, in one example, the speaker uses ask and then changes to say before
quoting; this may be a reaction of the non-preference for direct reports with
ask:
[Speaker is recounting the meal arrangements on a package tour holiday].
<S 01> And then on Friday night we asked the rep, I said by the way are
we getting a meal tonight she says no you don’t have a meal.
14 Another direct-speech report-marker used by the younger generation in the
corpus is like, which can occur without any reporting verb. For a full
discussion see Romaine and Lange (1991).
15 Leech (1987: 25) accounts for such usages as referring more to the result of the
communication rather than the communication itself.
16 See Schiffrin’s (1987:61) remarks on but as a discourse marker.
Glossary

+ = see separate heading for that item

Adjacency pairs

Adjacency pairs consist of two utterances that go together in an anticipated

way. A greeting (Hello!) anticipates a reply (e.g. another Hello). A question
anticipates an answer; an offer anticipates an acceptance or a rejection.
Adjacency pairs consist of two parts, a first pair-part and a second pair-part:

(First pair-part) A: Want a coffee?

- adjacency pair
(Second pair-part) _B: Er yes please.

Back-channel

This refers to noises (which are not full words) and short verbal responses
made by listeners which acknowledge the incoming talk and react to it,
without wishing to take over the speaking turn (=). Typical back-channels in
English are mm, uhum, yeah, no, right, oh, etc. In the transcripts in this book
they are shown as occurring during the speaker’s turn, though sometimes it
is difficult to distinguish between back-channels and full speaking turns, and
the decision to transcribe one way or the other is ultimately subjective. Here
is an example with back-channels from speaker <S 01>, shown within square
brackets [ ]:

<S 02> Oh yes, yes, yes mind you my parents were really quite well-off when we lived in
Ireland but the education in England was very expensive [<S 01> mm] and I can
remember my mother had jewellery and silver and she used to keep selling it
[<S 01> really] to pay for our extra music lessons and tuition in this and that
[<S 01> mm] and er I it was, must have been difficult.

Cleft structures

Cleft structures occur when the clause (-9) is ‘split’ and becomes two separate

176
 Glossary + 177

clauses but still only containing one message. Cleft structures can occur with
it and with wh-words:

Jeremy ate the cake. — It was Jeremy who ate the cake. (It-cleft)
We need more money.— What we need is more money. (Wh-cleft)

Clause

A clause is a unit of language based around a verb. All clauses, except

imperatives and clauses with subject ellipsis (+), have a subject, and many
have objects and adverbials too. Clauses may be ‘finite’ (i.e. with a verb that
changes for tense, person, number), or ‘non-finite’ (i.e. with a verb that ends
in an -ing form, or with a past participle, or with the infinitive form).
Examples of clauses:

She loves nursery school.

(finite: subject-verb-direct object)
He never laughs.
(finite: subject-verb)
I knew the answer, but didn’t tell her.
(two clauses: finite: subject-verb-object-linking adverbial + finite: subject-verb-
indirect object)
Listening to that music, I forgot all my troubles.
(two clauses: non-finite: -ing form of verb-adverbial + finite: subject-verb-direct
object)
To get there by six, you'll need to leave here at about 5.30.
(two clauses: non-finite: infinitive verb-adverbial-adverbial + finite: subject-
verb-verb-adverbial-adverbial)

Clauses are the building blocks of sentences in written language. In all

language, they are the most basic unit (-) of communication.

Convergence

As conversations progress, speakers’ contributions either converge with or

diverge from one another. There may, for example, be temporary disagree-
ment or misunderstanding, and so the speakers’ turns (>) will diverge. When
the misunderstanding is resolved, the turns converge again. Or one speaker
may be temporarily pursuing a different goal from another speaker. Conver-
gence is an ideal state where speakers’ minds mesh, where they are on the
same wavelength, pursuing the same goals, and each participant sees the
same need to co-operate and get to the desired outcome.
178 + Glossary

Deixis | deictic words

This term describes what may be termed the orientational features of

language; deixis involves words which point backwards and forwards in a
text as well as outside the text to a wider ‘out-of-text’ context. For example,
words like these/that/this/those locate an utterance in relation to space and to
the speaker’s sense of closeness or involvement with something; words like
now and then relate to the current moment of utterance; and words like we/
you/they/him/I relate to who is speaking, who is present, included, excluded,
etc. Thus, I’d like to pop in to that little shop over there before we leave contains
deictics which orientate a listener interpersonally, temporally and spatially
in relation to the proposition of the sentence. Certain contexts of language,
such as language used to do things like packing, cooking, moving furniture,
etc. involve a lot of deictics because the objects and other phenomena being
dealt with are normally immediately visible to all speakers and thus forms of
language such as Could we just move that into this corner here? are relatively
commonplace.

Discourse markers

Discourse markers are words or phrases which are normally used to mark
boundaries in conversation between one topic or bit of business and the
next. For example, words and phrases such as right, okay, I see, I mean, help
speakers to negotiate their way through talk indicating whether they want to
open or close a topic or to continue it, whether they share a common view of
the state of affairs, what their reaction is to something, etc. For example, in
telephone and other conversations the discourse marker anyway usually
serves to indicate that the speaker wishes either the current stretch of talk or
the whole exchange itself to be brought to a conclusion, or to resume an
interrupted topic. Similarly, right often serves to indicate that participants
are ready to move on to the next phase of business.

Ellipsis

Ellipsis is pervasive in spoken discourse. It occurs in writing where it usually

functions textually to avoid repetition where structures would otherwise be
redundant. For example, in the sentence We ran for the bus but missed it, it is
clear that we remains the subject of both clauses; or in the sentence The chair
was broken and the table too, where it is clearly unnecessary to repeat the verb
was broken. Ellipsis in spoken English is mainly situational (i.e. affecting people
and things in the immediate situation), and frequently involves the omission
 Glossary + 179

of personal subjects, where it is obvious that the speaker will remain

unambiguous. This feature is especially common with verbs of mental
process: for example, (I) think so, (I) wonder if they'll be coming to the party. Such
ellipsis also occurs with main or auxiliary verbs where meaning can be
relatively easily reconstructed from the context.

Fronting (or front-placing)

Fronting refers to the movement of an element from its ‘canonical’ position

and its relocation as the first element in a construction. Taking the sentence I
dedicated my life to that man and his music, we can front the indirect object as
follows: To that man and his music I dedicated my life. The process allows a focus
or emphasis to fall on the fronted element (-> topics).

Information staging

When speakers and writers create their message, they have choices as to how
to stage the information. Staging choices include whether to bring an item to
the front of the clause (> fronting), such as an object (e.g. The second part we
can leave till next week), or whether to use a tense that backgrounds an event
rather than foregrounding it (e.g. past perfect is a backgrounding tense, as is
past continuous, while past simple forgrounds events). Staging the informa-
tion means that the final message is not like a flat landscape, but has some
elements that are more appparent/important than others.

Interactional

= relational

Relational/Interactional

Relational or interactional language is language which is primarily personal

and social in orientation. Its effective use normally allows social and
interpersonal relations to be maintained. In some contexts such as service
encounters or even sometimes in formal interviews it is combined with
transactional (+>) language to soften and make less forbidding the business
of getting certain tasks done.

Speech acts

Speech acts refer to the communicative intention of what is said or written.

In speech-act theory, all language is seen as doing things. Examples of speech
180 - Glossary

acts are apologies, requests, denials, offers. Speech acts may be indirect, and
often it is not easy for the analyst to decide what ‘act’ is being performed by a
stretch of talk. Listeners decide in context what speech act is being
performed, for example, You mustn't worry will probably be heard by the
listener(s) as a comfort/reassurance, whereas You mustn’t come in here might be
heard as a prohibition.

Tails

The term ‘tail’ (sometimes called ‘right-dislocated item) describes the slot
available at the end of a clause (>) in which a speaker can insert grammatical
patterns which amplify, extend or reinforce what (s)he is saying or has said.
Examples of tails (in italics) include:

She’s a really good actress, Clare.

Singapore’s far too hot for me it is.
He’s quite a comic the fellow, you know.
It’s not actually very good is it that wine?

We may note also the extent to which tails cluster with different kinds of
tags, hedges and modal expressions and how they often serve to express
some kind of affective response, personal attitude or evaluative stance
towards the proposition or topic of the clause.

Topics

Topics (sometimes called ‘heads’, or ‘left-dislocated items’) perform a basi-

cally orienting and focusing function, identifying key information for
listeners and establishing a shared frame of reference for what is important
in a conversational exchange. Topics are a subcategory of fronting (+) and
involve, most typically, placement at the front of a clause (->) of a noun(s) or
noun phrase(s) that anticipates a structure which then forms the main
subject/object of the clause. Topics are in italics in the following examples:

The women in the audience, they all shouted in protest.

That bloke in the green, I can’t stand him.
This friend of ours, Carol, her daughter, she decided to buy one.

Topics are almost exclusive to informal spoken English. They parallel tails
(%), although tails generally serve a much more affective/evaluative
purpose.
 Glossary + 181

Transactional

Transactional language is language used in the process of conducting

business and generally getting things done. It contrasts with relational/
interactional (->) language.

Turns

A turn is a stretch of spoken language uttered by one speaker, before another

speaker takes over. In this extract there are three turns, one of which is
interrupted:

Turn 1 (interrupted) <S 01>Yeah it’s ready for erm er

Turn 2 <S 02> lThanks ever so much.
Turn 1 (completed) <S 01> Lfor
action ah it’s all right.
Turn 3 <S 03> Afraid I forgot the cheese.

Units

All language is analysed in units. Units are stretches of language which may
be classified in various ways. A word is a lexical unit. A clause is a
grammatical unit. A sentence is a unit common in written text, but not so
common in spoken language. In spoken language analysis, one of the most
difficult questions is to define what is the basic unit of communication.
Possible units for analysing spoken language include intonational tone-units
(ie. a stretch of language with one main rising or falling stress), speech turns
(-), and clauses (=>).

Vague language

Vague expressions are more extensive in all language use than is commonly
thought and they are especially prevalent in spoken discourse. When we
interact with others there are times where it is necessary to give exact and
precise information (for example, concerning departure times for trains); but
there are occasions where it would not be appropriate to be precise as it can
sound unduly authoritative and assertive. In most informal contexts most
speakers prefer to convey information which is softened in some way by vague
language, although such vagueness is often wrongly taken as a sign of careless
thinking or sloppy expression. Examples of vague language include phrases
such as or something, or anything, or whatever, all usually in final position:

Can you get me a sandwich or something?

Have they got mineral water or anything like that?
CANCODE bibliography (1994-98)

CANCODE stands for Cambridge and NottinghamCorpus of Discourse in

English. Cancode is a corpus-based project established in the Department of
English Studies, University of Nottingham and at Cambridge University
Press, Cambridge. All data collected by the project are the property of
Cambridge University Press. The bibliography is of recent work published in
whole or in part in connection with the project.

Carter, R. A. (1997) Investigating English Discourse: language, literacy and literature.

London: Routledge.
Carter, R. A. (1998) Orders of reality: CANCODE, communication and culture.
ELT Journal 52, 1.
Carter, R. A. and McCarthy, M. J. (1995a) Discourse and creativity: bridging
the gap between language and literature. In Cook, G. and Seidlhofer, B.
(eds) Principle and Practice in Applied Linguistics. Oxford University Press,
303-321.
Carter, R. A. and McCarthy, M. J. (1995b) Grammar and the spoken language.
Applied Linguistics 16, 2, 141-158.
Carter, R. A. and McCarthy, M. J. (1997a) Exploring Spoken English. Cambridge
University Press.
Carter, R. A. and McCarthy, M. J. (1997b) The English get-passive in spoken
discourse: description and implications for an interpersonal grammar.
Mimeo: University of Nottingham, Department of English Studies (sub-
mitted to Journal of English Language and Linguistics).
Carter, R. A., Hughes, R. and M. J. (1998) Telling tails: grammar, the spoken
language and materials development. In Tomlinson, B. (ed) Materials
Development in L2 Teaching. Cambridge University Press.
Hughes, R., Carter, R. A. and McCarthy, M. J. (1995) Discourse context as a
predictor of grammatical choice. In Graddol, D. and Thomas, S. (eds)
Language in a Changing Europe. Clevedon: BAAL/Multilingual Matters,
47-54.
Hughes, R. and McCarthy, M. J. (1998) From sentence to discourse: discourse
grammar and English language teaching. TESOL Quarterly, 32, 2: 263-287.

182
 CANCODE bibliography (1994-98) - 183

McCarthy, M. J. Conversation and literature: tense and aspect. In Payne, J. (ed)

Linguistic Approaches to Literature. Birmingham/University of Birmingham:
English Language Research, 58-73.
McCarthy, M. J. and Carter, R. A. (1994) Language as Discourse: Perspectives for
Language Teaching. Harlow: Longman.
McCarthy, M. J. and Carter, R. A. (1995) Spoken Grammar: what is it and how
do we teach it? ETL Journal 49, 3: 207-218.
McCarthy, M. J. and Carter, R. A. (1997a) Written and Spoken Vocabulary. In
Schmitt, N. and McCarthy, M. J. (eds) Vocabulary: Description, Acquisition,
Pedagogy. Cambridge University Press, 20-39.
McCarthy, M. J. and Carter, R. A. (1997b) Grammar, tails and affect: con-
structing expressive choices in discourse. Text 17, 3: 405-29.
Stanfield, C. (1996) English as she is spoke (conversation with CANCODE
researcher Jean Hudson). Cambridge Language Reference News 2, 2.
Tao, H. and McCarthy, M. J. (1998) Understanding non-restrictive relative
clauses, which is not an easy thing. Mimeo: Cornell University, Depart-
ment of Modern Languages.
References

Aijmer, K. (1984a) Go to and will in spoken English. In Ringbom, H. and

Rissanen, M. (eds) Proceedings from the Second Nordic Conference for English
Studies. Abo: Abo Akademi, 141-57.
Aijmer, K. (1984b) Sort of and kind of in English conversation. Studia
Linguistica 38: 118-28.
Aijmer, K. (1985) Just. In Backman, S. and Kjellmer, G. (eds) Papers on
Language and Literature. Gothenburg: Acta Universitatis Gothoburgensis,
t—10:
Aijmer, K. (1987) Oh and Ah in English conversation. In Meijs, W. (ed) Corpus
Linguistics and Beyond. Amsterdam: Rodopi, 61-86.
Aijmer, K. (1988) ‘Now may we have a word on this’: The use of ‘now’ as a
discourse particle. In Kyto, M., Ihalainen, O. and Rissanen, M. (eds)
Corpus Linguistics, Hard and Soft. Amsterdam: Rodopi, 15-34.
Aijmer, K. (1989) Themes and tails: the discourse function of dislocated
elements. Nordic Journal of Linguistics 12, 2: 137-54.
Aisenstadt, E. (1981) Restricted collocations in English lexicology and
lexicography. ITL Review of Applied Linguistics 53: 53-61.
Alexander, L. G. (1988) Longman English Grammar. London: Longman.
Alvarez-Caccamo, C. (1996) The power of reflexive language(s): code
displacement in reported speech. Journal of Pragmatics 25 (1): 33-59.
Antaki C., Diaz, F. and Collins, A. (1996) Keeping your footing: conversational
completion in three-part sequences. Journal of Pragmatics 25 (2): 151-71.
Ashby, W. (1988) The syntax, pragmatics, and sociolinguistics of left- and
right-dislocations in French. Lingua 75: 203-29.
Ashby, W. (1994) An acoustic profile of right-dislocations in French. French
Language Studies, 4 (2): 127-45.
Aston, G. (ed) (1988a) Negotiating Service: Studies in the Discourse of Bookshop
Encounters. Bologna: Editrice CLUEB.
Aston, G. (1988b) Learning Comity. Bologna: Editrice CLUEB.
Aston, G. (1995) Say ‘thank you’: some pragmatic constraints in
conversational closings. Applied Linguistics 16 (1): 57-86.
Atkins S., Clear, J. and Ostler, N. (1992) Corpus design criteria. Literary and
Linguistic Computing 7 (1): 1-16.

184
 References - 185

Atkinson, J. and Heritage, J. (eds) (1984) Structures of Social Action. Cambridge

University Press.
Bailey, C.J. (1985) Irrealis modalities and the misnamed ‘present simple
tense’ in English. Language and Communication 5 (4): 297-314.
Baker, E. (1924) Causes for the demand for spoken language. English Journal
13:595-7:
Bakhtin, M. (1986) Speech Genres and Other Late Essays. In Emerson, C. and
Holquist, M. (eds). Austin: University of Texas Press.
Banfield, A. (1982) Unspeakable Sentences: Narration and Representation in the
Language of Fiction. Boston: Routledge and Kegan Paul.
Bargiela-Chiappini, F. and Harris, S. (1995) Towards a generic structure of
meetings in British and Italian managements. Text 15 (4): 531-60.
Bauhr, G. (1992) Sobre el futuro cantaré y la forma compuesta voy a cantar
en espanol moderno. Moderna Sprak 86 (1): 69-79.
Bauman, R. (1986) Story, Performance, Event: Contextual Studies of Oral Narrative.
Cambridge University Press.
Baynham, M. (1991) Speech reporting as discourse strategy: some issues of
acquisition and use. Australian Review of Applied Linguistics 14 (2):
87-114.
Baynham, M. (1996) Direct speech: what’s it doing in non-narrative
discourse? Journal of Pragmatics 25 (1): 61-81.
Beaugrande, R. de (1997) Theory and Practice in Applied Linguistics:
Disconnection, Conflict, or Dialectic? Applied Linguistics 18 (3):
279-313:
Beier E., Starkweather, J. and Miller, D. (1967) Analysis of word frequencies
in spoken language of children. Language and Speech 10: 217-27.
Benwell, B. (1996) The Discourse of University Tutorials: An Investigation
into the Structure and Pedagogy of Small-group Teaching across a
Range of Academic Disciplines. Unpublished PhD thesis, University of
Nottingham UK.
Bergstrom, K. (1979) Idioms exercises and speech activities to develop
fluency. Collected Reviews Summer: 21-2.
Biber, D. (1988) Variation Across Speech and Writing. Cambridge University
Press.
Biber, D. (1990) Methodological issues regarding corpus-based analysis of
linguistic variation. Literary and Linguistic Computing 5 (4): 257-69.
Biber, D. (1993) Representativeness in corpus design. Literary and Linguistic
Computing 8 (4): 243-57.
Biber, D. (1995) Dimensions of Register Variation. Cambridge University Press.
Biber, D. and Finegan, E. (1989) Styles of stance in English: lexical and
grammatical marking of evidentiality and affect. Text 9 (1): 93-124.
186 + References

Biber, D. and Finegan, E. (1991) On the exploitation of computerized corpora

in variation studies. In Aijmer, K. and Altenberg, B. (eds) English Corpus
Linguistics. London: Longman, 204-20.
Binnick, R. (1991) Time and the Verb: A Guide to Tense and Aspect. New York:
Oxford University Press.
Birdsong, D. (1995) Iconicity, markedness, and processing constraints in
frozen locutions. In Landsberg, M. (ed) Syntactic Iconicity and Linguistic
Freezes: The Human Dimension. Berlin: Mouton de Gruyter, 31-45.
Blanche-Benveniste, C. (1982) Examen de la notion de subordination.
Recherche sur le Francais Parlé 4: 71-115.
Blanche-Benveniste, C. (1993) Repetitions de lexique et glissement vers la
gauche. Recherches sur le Francais Parlé 12: 9-34.
Blanche-Benveniste, C. (1995) De la rareté de certains phénomenes
syntaxiques en frang¢ais parlé. French Language Studies 5 (1): 17-29.
Blasco, M. (1995) Dislocation et thématisation en frangais parlé. Recherche sur
le Francais Parlé 13: 45-65.
Blommaert, J. (1991) How much culture is there in intercultural
communication? In Blommaert, J. and Verschueren, J. (eds) The
Pragmatics of International and Intercultural Communication. Amsterdam:
John Benjamins, 13-31.
Boden, D. and Zimmerman, D. H. (eds) (1991) Talk and Social Structure: Studies
in Ethnomethodology and Conversation Analysis. Cambridge: Polity Press.
Bolinger, D. (1976) Meaning and memory. Forum Linguisticum 1 (1): 1-14.
Boogaart, R. (1996) Tense and temporal ordering in English and Dutch
indirect speech. In Janssen, T. and Van der Wurff, W. (eds) Reported
Speech: Forms and Functions of the Verb. Amsterdam: John Benjamins,
Z3S=So-
Boyer, A. (1694) The Compleat French-Master for Ladies and Gentlemen. London:
Tho. Salisbury.
Bradford, B. (1988) Intonation in Context (student’s and teacher’s book).
Cambridge University Press.
Brazil, D. (1985) The Communicative Value of Intonation in English. Birmingham:
English Language Research, University of Birmingham.
Brazil, D. (1995) A Grammar of Speech. Oxford University Press.
Bressan, D. (1979) Idioms and second language teaching. Teanga 1: 31-40.
Brinton, L. (1987) The aspectual nature of states and habits. Folia Linguistica
XXI (2-4): 195-214.
Brown, P. and Levinson, S. (1987) Politeness: Some Universals in Language Usage.
Cambridge University Press.
Bublitz, W. (1988) Supportive Fellow-Speakers and Cooperative Conversations.
Amsterdam: John Benjamins.
 References + 187

Bublitz, W. (1989) Repetition in spoken discourse. In Mullenbrock, H.-J. and

Noll-Wieman, R. (eds) Anglistentag (1988) Gottingen: Vortrage. Tubingen:
Niemeyer, 352-68.
Cambridge International Dictionary of Idioms (1998) Cambridge University Press.
Carter, R. A., Hughes, R. and McCarthy, M. J. (1995) Discourse context as a
predictor of grammatical choice. In Graddol, D. and Thomas, S. (eds)
Language in a Changing Europe. Clevedon: BAAL/Mutilingual Matters,
47-54.
Carter, R. A. and McCarthy, M. J. (1995a) Discourse and creativity: bridging
the gap between language and literature. In Cook, G. and Seidlhofer, B.
Principle and Practice in Applied Linguistics. Studies in Honour of H. G.
Widdowson. Oxford University Press, 303-21.
Carter, R. A. and McCarthy, M. J. (1995b) Grammar and the spoken language.
Applied Linguistics 16 (2): 141-58.
Carter, R. A. and McCarthy, M. J. (1997) Exploring Spoken English. Cambridge
University Press.
Carterette, E. and Jones, M. H. (1974) Informal Speech. Berkeley and Los
Angeles: University of California Press.
Celce-Murcia, M. (1991) Discourse Analysis and Grammar Instruction. Annual
Review of Applied Linguistics 11 [1990] 135-51.
Cernak, F. (1994) Idiomatics. In Luelsdorff, P. A. (ed) The Prague School of
Structural and Functional Linguistics: a Short Introduction. Amsterdam: John
Benjamins, 185-95.
_Chadwyck-Healey, (1994) The Nineteenth Century on CD-ROM. Bibliographic
Records.
Chafe, W. (1982) Integration and involvement in speaking, writing, and oral
literature. In Tannen, D. (ed) Spoken and Written Language: Exploring
Orality and Literacy. Norwood, New Jersey: Ablex Publishing Corporation,
Bomb:
Chafe, W. (1992) The importance of corpus linguistics to understanding the
nature of language. In Svartvik, J. (ed) Directions in Corpus Linguistics.
Berlin: Mouton de Gruyter, 79-97.
Chafe, W., Du Bois, J. and Thompson, S. (1991) Towards a new corpus of
spoken American English. In Aijmer, K. and Altenberg, B. (eds) English
Corpus Linguistics. London: Longman, 64-82.
Channell, J. (1994) Vague Language. Oxford University Press.
Chappell, H. (1980) Is the get-passive adversative? Papers in Linguistics 13 (3):
411-52.
Choul, J.-C. (1982) Si muove, ma non troppo: an inquiry into the
non-metaphorical status of idioms and phrases. In Herzfield, M. and
Lenhart, M. (eds) Semiotics. New York: Plenum, 89-98.
188 - References

Christie, F. (1986) Writing in schools: generic structures as ways of meaning.

In Couture, B. (ed) Functional Approaches to Writing Research Perspectives.
London: Frances Pinter, 221-39.
Clark, A. (1946) Spoken English. Edinburgh: Oliver and Boyd.
Clark, H. and Gerrig, R. (1990) Quotations as demonstrations. Language 66:
764-805.
COBUILD (1995) Collins Cobuild English Language Dictionary. London: Collins.
Cohen, A. (1996) Developing the ability to perform speech acts. Studies in
Second Language Acquisition 18: 253-67.
Collins, P. (1996) Get-passives in English. World Englishes 15 (1): 43-56.
Collins, P. and Peters, P. (1988) The Australian corpus project. In Kyto M.,
Ihalainen, O. and Rissanen, M. (eds) Corpus Linguistics, Hard and Soft.
Amsterdam: Rodopi, 103-20.
Compton, J. (ed) (1941) Spoken English. London: Methuen and Co. Ltd.
Comrie, B. (1986) Tense in indirect speech, Folia Linguistica 20 (3-4):
265-296.
Coningham, C. G. (1894) Practical Business Conversation. Yokohama: Kelly &
Walsh Ltd.
Cook, G. (1989) Discourse. Oxford University Press.
Cook, G. (1990) Transcribing infinity: problems of context presentation.
Journal of Pragmatics 14: 1-24.
Coulmas, F. (1979) On the sociolinguistic relevance of routine formulae.
Journal of Pragmatics 3: 239-66.
Coulmas, F. (ed) (1981a) Conversational Routine. The Hague: Mouton.
Coulmas, F. (1981b) Idiomaticity as a problem of pragmatics. In Parret, H.,
Sbisa, M. and Verschueren, J. (eds) Possibilities and Limitations of
Pragmatics. Amsterdam: John Benjamins, 139-51.
Coulmas, F. (1985a) Direct and indirect speech: general problems and
problems of Japanese. Journal of Pragmatics 9 (1): 41-63.
Coulmas, F. (1985b) Nobody dies in Shangri-la: direct and indirect speech
across languages. Georgetown University Round Table on Languages and
Linguistics: 140-53.
Coulmas, F. (1986) Reported speech: some general issues. In Coulmas, F. (ed)
Direct and Indirect Speech. Berlin: Mouton de Gruyter, 1-28.
Coupland, N. (1983) Patterns of encounter management: further arguments
for discourse variables. Language in Society 12: 459-76.
Cowie, A. P. (1988) Stable and creative aspects of vocabulary use. In Carter,
R. A. and McCarthy, M. J. (eds) Vocabulary and Language Teaching. London:
Longman, 126-39.
Cowie, A. P. and Mackin, R. (1975) Oxford Dictionary of Current Idiomatic English.
Volume 1 Oxford University Press.
 References - 189

Crowdy, S. (1993) Spoken corpus design. Literary and Linguistic Computing 8 (2):
259-265.
Crowdy, S. (1994) Spoken corpus transcription. Literary and Linguistic
Computing 9 (1): 25-8.
Crymes, R. (1968) Some Systems of Substitution Correlations in Modern American
English. The Hague: Mouton.
Crystal, D. (1995) Refining stylistic discourse categories. In Melchers, G. and
Warren, B. (eds) Studies in Anglistics. Stockholm: Almqvist and Wiksell,
35-46.
Diaz, O. (1986) ‘Partir du bon pneu’: L’expression idiomatique a travers
l’expression publicitaire. Glottodidactica 18: 75-82.
Dolz, J. and Schneuwly, B. (1996) Genres et progression en expression orale
et écrite. Eléments de réflexions 4 propos d’une expérience romande.
Enjeux 37/38: 49-75.
Downing, A. and Locke, P. (1992) A University Course in English Grammar.
London: Prentice Hall.
Drew, P. and Holt, E. (1988) Complainable matters: the use of idiomatic
expressions in making complaints. Social Problems 35 (4): 398-417.
Drew, P. and Holt, E. (1995) Idiomatic expressions and their role in the
organisation of topic transition in conversation. In Everaert, M., van der
Linden, E.-J., Schenk A. and Schreuder, R. (eds) Idioms: Structural and
Psychological Perspectives. Hillsdale NJ: Lawrence Erlbaum Associates,
LYSE74
_Duranti, A. (1983) Samoan speechmaking across social events: one genre in
and out of a ‘fono’. Language in Society 12: 1-22.
Duranti, A. (1991) Four properties of speech-in-interaction. In Verschueren, J.
(ed) Pragmatics at Issue. Amsterdam: John Benjamins, 133-50.
Duranti, A. and Ochs, E. (1979) Left dislocation in Italian conversation. In
Givon, T. (ed) Syntax and Semantics, Volume 12: Discourse and Syntax. New
York: Academic Press.
Dykema, K. (1949) The grammar of spoken English: Its relation to what is
called English grammar. American Speech XXIV (1): 43-8.
Eastwood, J. (1994) Oxford Guide to English Grammar. Oxford University Press.
Edmondson, W., House J., Kasper, G. and Stemmer, B. (1984) Learning the
pragmatics of discourse: a project report. Applied Linguistics 5 (2):
LVS S27;
Edwards, J. (1992) Design principles in the transcription of spoken discourse.
In Svartvik, J. (ed) Directions in Corpus Linguistics. Berlin: Mouton de
Gruyter, 129-44.
Eggins, S. and Slade, D. (1997) Analysing Casual Conversation. London: Cassell.
Ehlich, K. (1989) Deictic expressions and the connexity of text. In Conte,
190 - References

M-E., Petofi, J. and Sézer, E. Text and Discourse Connectedness. Amsterdam:

John Benjamins, 33-52.
Ely, R., Gleason, J., Narasimhan, B. and McCabe, A. (1995) Family talk about
talk: mothers lead the way. Discourse Processes 19 (2): 201-18.
Emerson, C. (1983) The outer word and inner speech: Bakhtin, Vygotsky and
the internalization of language. Critical Inquiry 10 (2): 245-64.
Engels, L. (1988) The effect of spoken and written-to-be-spoken English on
word frequency counts of written English. In Klegraf, J. and Nehls, D.
(eds) Essays on the English Language and Applied Linguistics on the Occasion of
Gerhard Nickel’s 60th Birthday. Heidelberg: Julius Groos Verlag, 407-25.
Ernst, T. (1980) Grist for the linguistic mill: idioms and ‘extra’ adjectives.
Journal of Linguistic Research 1 (13): 51-68.
Esser, J. (1981) On the analysis of complex sentences: a study in the cohesion
of spoken English. In Esser, J. and Hubler, A. (eds) Forms and Functions.
Tiibingen: Gunter Narr Verlag, 163-74.
Fairclough, N. (1995) Critical Discourse Analysis. London: Longman.
Fang, A. C. (1995) Distribution of infinitives in contemporary British English:
a study based on the British ICE corpus. Literary and Linguistic Computing
10 (4): 247-57.
Fenk-Oczlon, G. (1989) Word frequency and word order in freezes. Linguistics
273917596:
Fernando, C. (1996) Idioms and Idiomaticity. Oxford University Press.
Fernando, C. and Flavell, R. (1981) On idiom: critical views and perspectives.
University of Exeter.
Fischer, K. and Drescher, M. (1996) Methods for the description of discourse
particles: contrastive analysis. Language Sciences 18 (3-4): 853-61.
Fleischman, S. (1983) From pragmatics to grammar: diachronic reflections
on complex pasts and futures in Romance. Lingua 60: 183-214.
Fleischman, S. (1990) Tense and Narrativity. London: Routledge.
Fleming, D. (1995) The search for an integrational account of language: Roy
Harris and conversation analysis. Language Sciences 17 (1): 73-98.
Ford, C. (1994) Dialogic aspects of talk and writing: because on the
interactive-edited continuum. Text 14 (4): 531-54.
Fox, B. and Thompson, S. (1990) A discourse explanation of the grammar of
relative clauses in English conversation. Language 66 (2): 297-316.
Fraser, B. (1990) An approach to discourse markers. Journal of Pragmatics 14:
383-95.
Fraser, B. and Malamud-Makowski, M. (1996) English and Spanish contrastive
discourse markers. Language Sciences 18 (3-4): 863-81.
Fretheim, T. (1995) Why Norwegian right-dislocated phrases are not
afterthoughts. Nordic Journal of Linguistics 18 (1): 31-54.
 References - 191

Fronek, J. (1982) Thing as a function word. Linguistics 20: 633-54.

Gardner, R. (1987) The identification and role of topic in spoken interaction.
Semiotica 65 (1/2): 129-41.
Geluykens, R. (1989) The syntactization of interactional processes: some
typological evidence. Belgian Journal of Linguistics 4: 91-103.
Geluykens, R. (1992) From Discourse Process to Grammatical Construction: on Left-
dislocation in English. Amsterdam: John Benjamins.
Genette, G. (1988) Narrative Discourse Revisited. Translated by J. E. Lewin. Ithaca
NY: Cornell University.
Gibbon, D. (1981) Idiomaticity and functional variation: a case study of
international amateur radio talk. Language in Society 10 (1): 21-42.
Giles H, Coupland, J. and Coupland, N. (1991) Accommodation theory:
communication, context and consequences. In Giles, H., Coupland, J.
and Coupland, N. (eds) Contexts of Accommodation Developments in Applied
Sociolinguistics. Cambridge University Press, 1-68.
Givon, T. and Yang, L. (1994) The rise of the English get-passive. In Fox, B. and
Hopper, P. J. (eds) Voice: Form and Function. Amsterdam: Benjamins,
119-49.
Gnutzmann, C. (1991) Linguistic and pedagogical aspects of English passive
constructions. Teanga 11: 48-65.
Goodell, E. W. (1987) The treatment of tense in indirect speech. TESOL
Quarterly 21 (2): 305-325.
Goodwin, C. (1984) Notes on story structure and the organisation of
participation. In Atkinson, J. and Heritage, J. (eds) Structures of Social
Action: Studies in Conversation Analysis. Cambridge University Press,
225-46.
Gottlieb, H. (1992) Idioms into Danish. In Nielsen, J. E. (ed) Words that Teem
with Meaning: Copenhagen Views on Lexicography. Copenhagen: Museum
Tusculanum Press, University of Copenhagen, 56-80.
Granger, S. (1983) The Be + Past Participle Construction in Spoken English.
Amsterdam: North Holland.
Guitart, J. M. (1989) On Spanish cleft sentences. In Kirschner, C. and
Decesaris, J. (eds) Studies in Romance Linguistics. Amsterdam: John
Benjamins.
Gustafsson, M. (1975) Binomial Expressions in Present-day English. University of
Turku.
Haegeman, L. (1983a) The Semantics of Will in Present-day British English: A
Unified Account. Brussels: Koninklijke Academie.
Haegeman, L. (1983b) Be going to, gaan, and aller: some observations on the
expression of future time. International Review of Applied Linguistics in
Language Teaching XXI (2): 155-7.
192 - References

Haegeman, L. (1989) Be going to and will: a pragmatic account. Journal of

Linguistics 25 (2): 291-317.
Haiman, J. (1978) Conditionals are topics. Language 54 (3): 564-89.
Hale, E. E. (1903) Ideas on rhetoric in the 16th century. PMLA 18: 424-44.
Halliday, M. A. K. (1978) Language as Social Semiotic. London: Edward Arnold.
Halliday, M. A. K. and Hasan, R. (1976) Cohesion in English. London: Longman.
Harder, P. (1980) Discourse as self-expression: on the reduced personality of
the second language learner. Applied Linguistics 1 (3): 262-70.
Harris, R. (1990) On redefining linguistics. In Davis, H. and Taylor, T. (eds)
Redefining Linguistics. London: Routledge, 18-52.
Hasan, R. (1984) Coherence and cohesive harmony. In Flood, J. (ed)
Understanding Reading Comprehension. Newark, Delaware: International
Reading Association, 181-219.
Hasan, R. (1985) The structure of a text. In Halliday, M. A. K. and Hasan, R.
Language, Context and Text: Aspects of Language in a Social-semiotic Perspective.
Oxford University Press, 52-69.
Hasan, R. (1992) Speech genre, semiotic mediation and the development of
higher mental functions. Language Sciences 14 (4): 489-528.
Hatcher, A. G. (1949) To get/be invited. Modern Language Notes 64: 433-46.
Heilenman, L. K. and McDonald, J. L. (1993) Dislocated sequences and word
order in French: a processing approach. Journal of French Language Studies
3 (2): 165-90.
Heritage, J. and Watson, D. (1979) Formulations as conversational objects. In
Psathas, G. (ed) Everyday Language. New York: Irvington Press, 123-62.
Herries, J. (1773) The Elements of Speech. London: E & C Dilly.
Higdon, D. L. and Bender, T. K. (1983) A concordance to Conrad’s Under Western
Eyes. New York: Garland Publishing.
Hockett, C. (1986) Grammar for the hearer. In McGregor, G. (ed) Language for
Hearers. Oxford: Pergamon Press, 49-68.
Hoey, M. P. (1983) On the Surface of Discourse. London: Allen and Unwin.
Hoey, M. P. (1991a) Patterns of Lexis in Text. Oxford University Press.
Hoey, M. P. (1991b) Some properties of spoken discourse. In Bowers, R. and
Brumfit, C. (eds) Applied Linguistics and English Language Teaching.
Basingstoke: Macmillan/MEP.
Hoffmann, L. (1989) Towards a pragmatically founded grammar. In
Graustein, G. and Leitner, G. (eds) Reference Grammars and Modern
Linguistic Theory. Tiibingen: Max Niemeyer, 111-32.
Holmes, J. (1738) The Art of Rhetoric Made Easy, or the Elements of Oratory, etc.
London (no publisher).
Horman, W. (1519) Vulgaria. London (no publisher).
 References - 193

Hong, B. (1985) Politeness in Chinese: impersonal pronouns and personal

greetings. Anthropological Linguistics 27 (2): 204-13.
Hopper, P. J. (1979) Aspect and foregrounding in discourse. In Givon, T. (ed)
Syntax and Semantics Volume 12: Discourse and Syntax. New York:
Academic Press, 213-41.
Hopper, P. J. and Thompson, S. (1993) Language universals, discourse
pragmatics, and semantics. Language Sciences 15 (4): 357-76.
Hopper R., Knapp, M. L. and Scott, L. (1981) Couples’ personal idioms:
exploring intimate talk. Journal of Communication 31 (1): 23-33.
House, J. (1985) Contrastive discourse analysis and universals in language
usage. Papers and Studies in Contrastive Linguistics 20: 5-14.
Houtkoop, H. and Mazeland, H. (1985) Turns and discourse units in everyday
conversation. Journal of Pragmatics 9: 595-619.
Houtkoop-Steenstra, H. (1991) Opening sequences in Dutch telephone
conversations. In Boden, D. and Zimmerman, D. (eds) Talk and Social
Structure. Oxford: Polity Press.
Howes, D. H. (1966) A word count of spoken English. Journal of Verbal Learning
and Verbal Behaviour 5: 572-606.
Huddleston, R. (1989) The treatment of tense in indirect speech. Folia
Linguistica 23 (3-4): 335-340.
Hughes, R. (1996) English in Speech and Writing. London: Routledge.
Hughes, R. and McCarthy, M. J. (1998) From sentence to discourse: discourse
grammar and English Language Teaching. TESOL Quarterly 32 (2):
263-287.
Hymes, D. (1972) Models of the interaction of language and social life. In
Gumperz, J. and Hymes, D. (eds) Directions in Sociolinguistics: The
Ethnography of Communication. New York: Rinehart and Winston Inc,
Smile
Iacobucci, C. (1990) Accounts, formulations and goal attainment strategies
in service encounters. In Tracy, K. and Coupland, N. (eds) Multiple Goals
in Discourse. Clevedon: Multilingual Matters Ltd, 85-99.
Jackson, H. (1990) OCP and the computer analysis of texts: The Birmingham
Polytechnic experience. Literary and Linguistic Computing 5 (1): 86-8.
Jaszczolt, K. and Turner, K. (eds) (1996) Contrastive Semantics and Pragmatics
Volume II: Discourse Strategies. Oxford: Elsevier Science Ltd.
Jaworski, A. (1990) The acquisition and perception of formulaic language
and foreign language teaching. Multilingua 9 (4): 397-411.
Jefferson, G. (1978) Sequential aspects of storytelling in conversation. In
Schenkein, J. (ed) Studies in the Organisation of Conversational Interaction.
New York: Academic Press, 219-48.
Johnstone, B. (1987) ‘He says ... so I said’: verb tense alternation and
194 - References

narrative depictions of authority in American English. Linguistics 25 (1):

33502.
Joos, M. (1964) The English Verb: Form and Meanings. Madison: University of
Wisconsin Press.
Kallgren, G. and Prince, E. F. (1989) Swedish VP-topicalisation and Yiddish
verb-topicalisation. Nordic Journal of Linguistics 12: 47-58.
Kehe, D. and Kehe, P. D. (1989) Maintaining teacher control during pair/
group work. Mijinteoir Teanga 2 (2): 35-9.
Kelly Hall, J. (1995). (Re)creating our worlds with words: a sociohistorical
perspective of face-to-face interaction. Applied Linguistics 16 (2): 206-32.
Kirk, J. (1992) The Northern Ireland transcribed corpus of speech. In
Leitner, G. (ed) New Directions in English Language Corpora. Berlin: Mouton
de Gruyter, 65-73.
Knowles, G. (1990) The use of spoken and written corpora in the teaching of
language and linguistics. Literary and Linguistic Computing 5 (1): 45-8.
Kooij, J. (1968) Compounds and idioms. Lingua 21: 250-68.
Komter, M. (1991) Conflict and Cooperation in Job Interviews: A Study of Talk, Tasks
and Ideas. Amsterdam: John Benjamins.
Labov, W. (1972) Language in the Inner City. Oxford: Basil Blackwell.
Lakoff, R. (1970) Tense and its relation to participants. Language 46: 838-49.
Lakoff, R. (1971) Passive resistance. Papers from the Seventh Regional Meeting.
Chicago Linguistic Society 149-62.
Lambrecht, K. (1988) Presentational cleft constructions in spoken French. In
Haiman, J. and Thompson, S. (eds) Clause Combining in Grammar and
Discourse. Amsterdam: John Benjamins.
Larson, M. (1978) The Functions of Reported Speech in Discourse. Dallas: The
Summer Institute of Linguistics and the University of Texas at
Arlington.
Lattey, E. (1986) Pragmatic classification of idioms as an aid for the language
learner. International Review of Applied Linguistics XXIV (3): 217-33.
Lebra, T. S. (1987) The cultural significance of silence in Japanese. Multilingua
6 (4): 343-57.
Leech, G. (1987) Meaning and the English Verb. 2nd edition. London: Longman.
Lewis, M. (1993) The Lexical Approach: The State of ELT and a Way Forward. Hove
UK: LTP.
Lindenfeld, J. (1990) Speech and Sociability at French Urban Market Places.
Amsterdam: John Benjamins.
Longman Dictionary of English Idioms. (1979) London: Longman.
Loveday, L. (1982) The Sociolinguistics of Learning and Using a Non-native
Language. Oxford: Pergamon.
Louw, B. (1993) Irony in the text or insincerity in the writer? The diagnostic
 References - 195

potential of semantic prosodies. In Baker M., Francis, G. and Tognini-

Bonelli, E. (eds) Text and Technology: In Honour of John Sinclair.
Philadelphia: Benjamins, 157-76.
Low, G. D. (1988) On teaching metaphor. Applied Linguistics 9 (2): 125-47.
Low, G. (1995) Intensifiers and hedges in questionnaire items and the lexical
invisibility hypothesis. Applied Linguistics 16 (4): 505-41.
Lucy, J. (ed) (1993) Reflexive Language. Cambridge University Press.
Mair, C. (1992) Problems in the compilation of a corpus of standard
Caribbean English: A pilot study. In Leitner, G. (ed) New Directions in
English Language Corpora. Berlin: Mouton de Gruyter, 75-96.
Makkai, A. (1978) Idiomaticity as a language universal. In Greenberg, J. H.
(ed) Universals of Human Language, Volume 3: Word Structure. Stanford CA:
Stanford University Press, 401-48.
Malkiel, Y. (1959) Studies in irreversible binomials. Lingua 8: 113-60.
Martin, J. (1992) English Text: System and Structure. Amsterdam: John
Benjamins.
Martin, W. (1988) Variation in lexical frequency. In Van Reenen, P. and Van
Reenen-Stein, K. (eds) Distributions spatiales et temporelles, constellations des
manuscrits. Amsterdam: John Benjamins, 139-52.
Mathis, T. and Yule, G. (1994) Zero quotatives. Discourse Processes 18: 63-76.
Mayes, P. (1990) Quotation in spoken English. Studies in Language 14 (2):
325-63.
McCarthy, M. J. (1984) A new look at vocabulary in EFL. Applied Linguistics 5
(1): 12-22.
McCarthy, M. J. (1988) Some vocabulary patterns in conversation. In Carter,
R. A. and McCarthy, M. J. Vocabulary and Language Teaching. London:
Longman, 181-200.
McCarthy, M. J. (1991) Discourse Analysis for Language Teachers. Cambridge
University Press.
McCarthy, M. J. (1992a) Interactive lexis: prominence and paradigms. In
Coulthard, R. M. (ed) Advances in Spoken Discourse Analysis. London:
Routledge, 197-208.
McCarthy, M. J. (1992b) English idioms in use. Revista Canaria de Estudios
Ingleses 25: 55-65.
McCarthy, M. J. (1994) It, this and that. In Coulthard, M. (ed) Advances in
Written Text Analysis. London: Routledge, 266-75.
McCarthy, M. J. and Carter, R. A. (1994) Language as Discourse: Perspectives for
Language Teaching. London: Longman.
McCarthy, M. J. and Carter, R. A. (1997a) Written and spoken vocabulary. In
Schmitt, N. and McCarthy, M. J. (eds) Second Language Vocabulary:
Description, Acquisition and Pedagogy. Cambridge University Press, 20-39.
196 - References

McCarthy, M. J. and Carter, R. A. (1997b) Grammar, tails and affect:

constructing expressive choices in discourse. Text 17 (3): 405-29.
McCarthy, M. J. and O’Dell, F. (1994) English Vocabulary in Use: upper-
intermediate and advanced. Cambridge University Press.
McGlone, M. S., Glucksberg, S. and Cacciari, C. (1994) Semantic productivity
and idiom comprehension. Discourse Processes 17: 167-90.
McGregor, G. and White, R. (1990) Reception and Response: Hearer Creativity and
the Analysis of Spoken and Written Texts. London: Routledge.
McTear, M. (1980) The pragmatics of because. In McCormack, W. and Izzo, H.
(eds) The Sixth LACUS Forum 1979. Columbia SC: Hornbeam, 455-63.
Merritt, M. (1976) On questions following questions in service encounters.
Language in Society 5: 315-57.
Miller, J. (1995) Does spoken language have sentences? In Palmer, F. R. (ed)
Grammar and Meaning. Cambridge University Press, 116-35.
Mitchell, A. G. (1957) Spoken English. London: Macmillan and Co. Ltd.
Mitchell, T. F. (1957) The language of buying and selling in Cyrenaica: a
situational statement. Hespéris XLIV: 31-71.
Moeran, B. (1984) Advertising sounds as cultural discourse. Language and
Communication 4 (2): 147-58.
Moon, R. (1992) Textual aspects of fixed expressions in learners’ dictionaries.
In Arnaud, P. J. and Béjoint, H. (eds) Vocabulary and Applied Linguistics.
Basingstoke: Macmillan, 13-27.
Moon, R. (1997) Vocabulary connections: multi-word items in English. In
Schmitt, N. and McCarthy, M. J. (eds). (1997) Second Language Vocabulary:
Description, Acquisition and Pedagogy. Cambridge University Press, 40-63.
Mutsu, H. 1894 A Japanese Conversation Course. Tokio [sic]: Z. P. Maruya & Co.
Nash, W. (1990) Language in Popular Fiction. London: Routledge.
Nattinger, J. R. and DeCarrico, J. S. (1992) Lexical Phrases and Language
Teaching. Oxford University Press.
Nelson, F. (1992) Language corpora B.C. In Svartvik, J. (ed) Directions in Corpus
Linguistics. Berlin: Mouton de Gruyter, 17-32.
Nelson, G. (1996) The design of the corpus. In Greenbaum, S. (ed) Comparing
English Worldwide: the International Corpus of English. Oxford University
PLESS 27 30:
Noguchi, R. R. (1987) The dynamics of rule conflict in English and Japanese
conversation. International Review of Applied Linguistics 25 (1): 15-24.
Norrick, N. (1986) Stock similes. Journal of Literary Semantics XV (1): 39-52.
Norrick, N. (1988) Binomial meaning in texts. Journal of English Linguistics 21
(ie WO. '
Nyyssonen, H. (1992) Lexis in discourse. In Lindeberg, A-C., Enkvist, N. E. and
 References +» 197

Wikberg, K. (eds) Nordic Research on text and Discourse. Abo Akademis

Forlag, 73-80.
Ochs E., Schegloff, E. and Thompson, S. (eds) (1996) Interaction and Grammar.
Cambridge University Press.
Ono, T. and Suzuki, R. (1992) Word order variability in Japanese
conversation: motivations and grammaticization. Text 12 (3): 429-45.
Oostdijk, N. (1990) The language of dialogue in fiction. Literary and Linguistic
Computing 5 (3): 235-41.
Owen, C. (1996) Do concordances require to be consulted? ELT Journal 50 (3):
219-24.
Page, N. (1973) Speech in the English Novel. London: Macmillan.
Palacas, A. (1993) Attribution semantics: linguistic worlds and point of view.
Discourse Processes 16: 239-77.
Palmer, H. E., Blandford, F. G. (1924) A Grammar of Spoken English. First
edition. Third edition 1969, revised and rewritten by R. Kingdon.
Cambridge: Heffer.
Peacham, H. (1577) The Garden of Eloquence etc. London: H. Jackson.
Person, R. (1996) Restarts in conversation and literature. Language and
Communication 16 (1): 61-70.
Persson, G. (1974) Repetition in English: Part I, Sequential Repetition. Uppsala:
Acta Universitatis Upsaliensi.
Peterson, P. (1982) Anaphoric reference to facts, propositions and events.
Linguistics and Philosophy 5 (2): 235-76.
Philips, S. (1985) Reported Speech as Evidence in an American Trial.
Georgetown University Round Table on Languages and Linguistics: 154-70.
Polanyi, L. (1981) Telling the same story twice. Text 1 (4): 315-36.
Polanyi, L. (1982) Linguistic and social constraints on storytelling. Journal of
Pragmatics 6 (5/6): 509-24.
Pomerantz, A. (1984) Agreeing and disagreeing with assessments: some
features of preferred/dispreferred turn shapes. In Atkinson, J. and
Heritage, J. (eds) (1984) Structures of Social Action. Cambridge University
IBECSS tSZatOile
Powell, M. J. (1992) Semantic/pragmatic regularities in informal lexis: British
speakers in spontaneous conversational settings. Text 12 (1): 19-58.
Prodromou, L. (1997) Global English and its struggle against the octopus.
IATEFL Newsletter 135: 12-14.
Psathas, G. (ed) (1979) Everyday Language: Studies in Ethnomethodology. New
York: Irvington Publications, Inc.
Psathas, G. (1995) Conversation Analysis: The Study of Talk-in-Interaction. London:
Sage Publications.
198 + References

Quirk, R., Greenbaum, S., Leech, G. and Svartvik, J. (1985) A Comprehensive

Grammar of the English Language. London: Longman.
Reagan, R. (1987) The syntax of English idioms: can the dog be put on?
Journal of Psycholinguistic Research 16 (5): 417-41.
Reid, I. (ed) (1987) The Place of Genre in Learning: Current Debates. Victoria:
Deakin University Press.
Reyes, G. (1984) Polifonia textual: la citacion en el relato literario. Madrid: Gredos.
Rivero, M. (1980) On left-dislocation and topicalisation in Spanish. Linguistic
Inquiry 11 (2): 363-93.
Robinson, R. (1617) The Art of Pronunciation. London (no publisher).
Romaine, S. and Lange, D. (1991) The use of Like as a marker of reported
speech and thought: a case of grammaticalisation in process. American
Speech 66 (3): 227-79.
Rosaldo, M. (1982) The things we do with words: Ilongot speech act and
speech act theory in philosophy. Language in Society 11: 203-37.
Rundell, M. (1995a) The BNC: A spoken corpus. Modern English Teacher 4 (2):
ils lisy
Rundell, M. (1995b) The word on the street. English Today 11 (3): 29-35.
Sacks H., Schegloff, E. A. and Jefferson, G. (1974) A simplest systematics for
the organisation of turn-taking for conversation. Language 50 (4):
696-735
Scarcella, R. (1983) Developmental trends in the acquisition of
conversational competence by adult second language learners. In
Wolfson, N. and Judd, E. (eds) Sociolinguistics and Language Acquisition.
Rowley MA Newbury House.
Scarcella, R. and Brunak, J. (1981) On speaking politely in a second language.
International Journal of the Sociology of Language: 59-75.
Schegloff, E. A. and Sacks, H. (1973) Opening up closings. Semiotica 8 (4):
289-327.
Schenkein, J. (1980) A taxonomy for repeating action sequences in natural
conversation. In Butterworth, B. (ed) Language Production: Volume 1:
Speech and Talk. London: Academic Press, 21-47.
Schiffrin, D. (1994) Approaches to Discourse. Oxford: Basil Blackwell.
Schiffrin, D. (1987) Discourse Markers. Cambridge University Press.
Schleppegrell, M. (1992) Subordination and linguistic complexity. Discourse
Processesmon()ad li7eSde
Schmitt, N. and McCarthy, M. J. (eds) (1997) Second Language Vocabulary:
Description, Acquisition and Pedagogy. Cambridge University Press.
Schonell, F., Meddleton, I., Shaw, B., Routh, M., Popham, D., Gill, G., Mackrell,
G. and Stephens, C. (1956) A Study of the Oral Vocabulary ofAdults. Brisbane
and London: University of Queensland Press/University of London Press.
 References +» 199

Scotton, C. (1985) What the heck, sir: Style shifting and lexical colouring as
features of powerful language. In Street, R., Capella, J. and Giles, H. (eds)
Sequence and Patterning in Communicative Behaviour. London: Edward
Arnold, 103-19.
Sherry, R. (1550) A Treatise of Schemes and Tropes ... Gathered out of the Best
Grammarians and Oratours. London: John Day.
Sifianou, M. (1989) On the telephone again. Differences in telephone
behaviour: England versus Greece. Language in Society 18 (4): 527-44.
Sinclair, J. McH. (1995) Corpus typology - a framework for classification. In
Melchers, G. and Warren, B. (eds) Studies in Anglistics. Stockholm:
Almavist and Wiksell, 17-33.
Sinclair, J. McH. and Coulthard, R. M. (1975) Towards an Analysis of Discourse.
Oxford University Press.
Skehan, P. (1996) A framework for the implementation of task-based
instruction. Applied Linguistics 17 (1): 38-62.
Stein, G. (1979) Studies in the Function of the Passive. Tubingen: Gunter Narr
Verlag.
Stenstrom, A-B. (1990) Lexical items peculiar to spoken discourse. In
Svartvik, J. (ed) The London-Lund Corpus of Spoken English. Lund: Lund
University Press, 137-75.
Stenstr6m, A.B. (1994) An Introduction to Spoken Interaction. London: Longman.
Stern, K. (1997) The Longman Spoken American Corpus: Providing an in-
depth analysis of everyday English. Longman Language Review 3: 14-17.
Strassler, J. (1982) Idioms in English: a Pragmatic Analysis. Tubingen: Gunter
Narr Verlag.
Stubbe, M. and Holmes, J. (1995) You know, eh and other ‘exasperating
expressions’: an analysis of social and stylistic variation in the use of
pragmatic devices in a sample of New Zealand English. Language and
Communication 15 (1): 63-88.
Stubbs, M. (1986) Lexical density: a computational technique and some
findings. In Coulthard, R. M. (ed) Talking About Text. Birmingham: English
Language Research, 27-42.
Stubbs, M. (1996) Text and Corpus Analysis. Oxford: Blackwell.
Sussex, R. (1982) A note on the get-passive construction. Australian Journal of
Linguistics 2: 83-95.
Svartvik, J. (1966) On Voice in the English Verb. The Hague: Mouton.
Svartvik, J. (ed) (1990) The London-Lund Corpus of Spoken English: Description and
Research. Lund University Press.
Svartvik, J. and Quirk, R. (1980) A Corpus of English Conversation. Lund:
Liberlaromedel.
Swales, J. (1990) Genre Analysis. Cambridge University Press.
200 - References

Swan, M. (1980/1995) Practical English Usage. Oxford University Press.

Tamony, P. (1982) ‘Like Kelly’s nuts’ and related expressions. Comments on
Etymology 11 (9-10): 8-10.
Tannen, D. (1986) Introducing constructed dialogue in Greek and American
conversational and literary narrative. In Coulmas, F. (ed) Direct and
Indirect Speech. Berlin: Mouton de Gruyter, 311-32.
Tannen, D. (1988) Hearing voices in conversation, fiction and mixed genres.
In Tannen, D. (ed) Linguistics in Context: Connecting Observation and
Understanding. Norwood, NJ: Ablex, 89-113.
Tannen, D. (1989) Talking Voices. Cambridge University Press.
Testa, R. (1988) Interruptive strategies in English and Italian conversation:
smooth versus contrastive linguistic preferences. Multilingua 7 (3):
2oomouze
Thomas, J. (1984) Cross-cultural discourse as ‘unequal encounter’: towards a
pragmatic analysis. Applied Linguistics 5 (3): 226-35.
Thompson, G. (1994) Reporting. Collins-Cobuild English Guides no. 5 London:
HarperCollins.
Thomspon, G. (1996) Voices in the text: discourse perspectives on language
reports. Applied Linguistics. 17 (4): 501-30.
Thompson, S. E. (1997) Presenting Research: A Study of Interaction in
Academic Monologue. Unpublished PhD Dissertation. University of
Liverpool.
Tognini-Bonelli, E. (1996) Corpus: Theory and Practice. Birmingham: TWC (also
forthcoming, to be published by Benjamins, Amsterdam).
Tottie, G. (1983) The missing link? or, why is there twice as much negation
in spoken English as in written English? In Jacobson, S. (ed) Papers from
the Second Scandinavian Symposium on Syntactic Variation 1983. Stockholm:
Almaqvist and Wiksell International, 67-74.
Tracy, K. and Coupland, N. (1990) (eds) Multiple Goals in Discourse. Clevedon:
Multilingual Matters.
Trevise, A. (1986) Is it transferable, topicalisation? In Kellerman, E. and
Sharwood-Smith, M. (eds) Crosslinguistic Influence in Second Language
Acquisition. New York: Pergamon, 186-206.
Trueblood, T. (1933) Spoken English. Quarterly Journal of Speech 19: 513.
Tyler, A. E., Jeffries, A. A. and Davies, C. E. (1988) The effect of discourse
structuring devices on listener perceptions of coherence in non-native
university teachers’ spoken discourse. World Englishes 7 (2): 101-10.
Ure, J. (1971) Lexical density and register differentiation. In Perren, G. E. and
Trim, J. L. M. (eds) Applications of Linguistics: Selected Papers of the Second
International Congress of Applied Linguistics, Cambridge, 1969. Cambridge
University Press, 443-52.
 References - 201

Vakar, P. (1966) A Word-Count of Spoken Russian. Columbus, Ohio: OSU

Press.
Valiouli, M. (1991) Right-dislocated anaphorically functioning nominals,
concord and referential/attitudinal perspective. In Proceedings: 5th
Symposium on the Description and/or Comparison of English and Greek.
Thessaloniki: Aristotle University, 159-70.
Van Dijk, T. A. (1982) Episodes as units of discourse analysis. In Tannen, D.
(ed) Analyzing Discourse: Text and Talk. Washington DC: Georgetown
University Press, 177-95.
Vanrespaille, M. (1991) A semantic analysis of the English get-passive.
Interface 5 (2): 95-112.
Ventola, E. (1987) The Structure of Social Interaction: A Systemic Approach to the
Semiotics of Service Encounters. London: Frances Pinter.
Vorlat, E. (1985) ‘Your marriage is going through a rocky patch’: on idioms
in the Lonely Hearts column. In Debusscher, G. and Van Noppen, J. (eds)
Communiquer et traduire: Hommages aJean Dierick. Brussels: Editions de
l'Université de Bruxelles, 103-8.
Wald, B. (1983) Referents and topics within and across discourse units:
observations from current vernacular English. In Klein-Andreu, F. (ed)
Discourse Perspectives on Syntax. New York: Academic Press, 91-116.
Wales, M. L. (1983) The semantic distribution of aller + infinitive and the
future tense in spoken French. General Linguistics 23 (1): 19-28.
Walter, B. (1988) The Jury Summation as Speech Genre. Amsterdam: John
Benjamins.
Watts, I. (1740) The Art of Reading and Writing English. London: R. Hett and
J. Bracstone.
Watts, R. J. (1989) Taking the pitcher to the ‘well’: native speakers’
perception of their use of discourse markers in conversation. Journal of
Pragmatics 13: 203-37.
Waugh, L. (1995) Reported speech in journalistic discourse: the relation of
function and text. Text 15 (1): 129-73.
Wertsch, J. (1985) The semiotic mediation of mental life: L. S. Vygotsky and
M. M. Bakhtin. In Mertz, E. and Parmentier, R. (eds) Semiotic Mediation:
Sociocultural and Psychological Perspectives. London: Academic Press, 49-71.
White, B. (ed) (1932) The Vulgaria of John Stanbridge and the Vulgaria of Robert
Whittinton. London: Kegan Paul, Trench, Trubner and Co. Ltd.
Wikberg, K. (1992) Discourse category and text type classification:
procedural discourse in the Brown and the LOB corpora. In Leitner, G.
(ed) New Directions in English Language Corpora. Berlin: Mouton de Gruyter,
247-61.
Widdowson, H. G. (1980) Models and fictions. Applied Linguistics 1 (2): 165-70.
202 + References

Wilkins, D. A. (1976) Notional Syllabuses. Oxford University Press.

Williams, J. (1992) Planning, discourse marking and the comprehensibility
of international teaching assistants. TESOL Quarterly 26 (4): 693-711.
Winter, E. O. (1982) Towards a Contextual Grammar of English. London: Allen &
Unwin.
Ylanne-McEwen, V. (1997) Relational processes within a transactional setting:
An investigation of travel agency discourse. Unpublished PhD
dissertation. University of Wales, Cardiff.
Yngve, V. H. (1970) On getting a word in edgewise. Papers from the 6th Regional
Meeting, Chicago Linguistic Society. Chicago: Chicago Linguistic Society.
Young, L. (1990) Language as Behaviour, Language as Code: A Study of Academic
English. Amsterdam: John Benjamins.
Zelizer, B. (1989) ‘Saying’ as collective practice: quoting and differential
address in the news. Text 9 (4): 369-88.
Zettersten, A. (1978) A Word-Frequency List Based on American English Reportage.
Kobenhavn: Universitetsforlaget i Kobenhavn.
Zydatiss, W. (1986) Grammatical categories and their text functions - some
implications for the content of reference grammars. In Leitner, G. (ed)
The English Reference Grammar: Language and Linguistics, Writers and Readers.
Tubingen: Max Niemeyer Verlag, 140-55.
Index

abstract 63, 95 corpus-informed approach 22

adjacency pair 29, 54, 106, 114 corpus-based approach 22
affect markers 39 corpus design 8
antonyms 116 corpus projects 1
as I/you say, in speech reports 167 cross linguistic comparisons 107
ask, in speech reports 162-165, 172 cultural allusions 131
awareness raising 66 cultural problems 58

backgrounding for past perfect 74 data collection 8

back-channel 58 demographic approach 8
backshift 162 genre approach 8
be to 78, 100 deixis 39, 43, 112
binomials 130, 143 demonstratives 112
broadcast data 7 direct speech 161
discourse analysis 20
CANCODE, 5-23 influence on language teaching 20
clause completion 80 discourse communities 26
coda 63, 95, 143 discourse features in L2 and L1 51
collocation, left-hand 22, 23 discourse grammar 69
collocation, post 23 discourse marker 34, 39, 45, 59, 112
collocation, prior 23 in vocabulary syllabus 124
collocation, right-hand 22, 23 dispreferred utterances 56
competence 51 divergences 56
complement clause 81
complicating events 63 ellipsis 65
computational analysis 122 ethnomethodology 20
context types 9, 11, 38 evaluation 117, 134, 137, 143
intimate 9 event line 137
pedagogical 9 exchange 52, 67,106
professional 9 existential paradigms 69
socialising 9 expectations 33, 34, 38
transactional 9 extended metaphors 137
contextual constraints 64
contradictions 56 face 56, 132
convergence 34, 54, 109, 143 fixed expressions 119-122, 130
conversation analysis 20 follow-up 52, 53,54, 67
co-operative behaviour 56 formulations 32, 33, 36, 38, 45
corpora, spoken 6 frequency lists 122-123
corpora, spoken and written 5-23 front placement 60
corpus analysis 7 frozen similes 131
corpus-driven approach 22 function of idioms in discourse 145-149

203
204 - Index

future 100 induction 67

information staging 60-61
generic features of spoken language 45, initiation 53
62-64 instantiations 33, 37, 38
generic patterning 45 interaction 67
generic type 34 interactional features 57-62, 109, 112
genre 26-48 interruptions 67
as social compact 32 IRF 53
closings 63 it, this, that 70
dynamic characteristic of 28
dynamic view 29 Labov’s narrative model 63
every day genres 26 language in action 34, 39, 41, 65, 111
openings 63 learner awareness 17, 20, 52, 127
theories of 27-30 left-dislocation 62, 77
get-passive 82, 83, 85 lexical density 40, 43, 44, 65, 109
go, in speech reports 165 lexical patterning 109
goal orientation 30, 32, 43 and vocabulary teaching 127
goal sub type 11, 41 lexical problems 58
goal types 10, 11, 31, 38, 39 lexis 108 see also vocabulary
collaborative ideas 10, lower order features 38
collaborative tasks 10
provision of information 10 main clause 79, 80
going to 100, 103, 104, 105 methodology, genre based 46
grammar, spoken and written 76-79 modals 112
grammar and discourse 70, 86 multi-word items 21
grammatical choice 69, 70, 105
grammatical classification 79-82 negotiation of meaning 143
grammatical forms news broadcasts 92
co-occurrence of 106 news-worthiness 94
patterning 104, 106 non-finite clause 82
sequence of 106 notional-functional theory 19
now relevance 105
head 77
higher-order features 38, 43, 90 observation-comment function 142
historical present 96, 158 nominal compounds, opaque 131
hot news 92 oral anecdotes 63
for future 100 orientation 63
idioms 120, 129-149, 131, 144
controlled practice 147 participant goals 30, 32, 47
cultural allusion 139 participant relationships 47
evaluative function of 132 past perfect 75
in coda 135 past simple 93, 95, 99, 162
in collaborative ideas 140 pattern, preview-detail 92
in everyday stories and anecdotes pattern, situation-event 92,
133-140 patterning
opaque idioms 137 of lexis 108
teaching 129-149 principles of textual grammatical
to save face 144 patterning 91
idiomatic expressions see idioms literary texts 99
idiomatic speech routines see idioms patterns of co-occurrence of verb forms
if-clause 82 90
illustration 67 pedagogical grammar 72
 Index - 205

pedagogical modelling 66 spoken and written grammar 76-79

phrases 79 spoken and written, patterns of co-
prepositional clause 82 occurrence of verb forms 90
prepositional expressions 130 spoken communication 17
present perfect 93, 95 spoken corpora, problems with 12
presentational methodology 60 spoken language and genre 26-48
problem-solution pattern 43, 103 spoken language v. written language 39,
progressive ’s phrases 131 47,125, 126
spoken language
qualitative analysis 22, 47 communicative pedagogy 18
quantitative analysis 22, 47 relational elements 28
status of in applied linguistics 15
real data, observation of 73 structural features 50
recollections 33, 34, 36, 38 transactional elements 28
recording equipment 12 standard features of language 21
refusals 56 story-orientation 98
register 26 structural features in L2 51
relational goals 31, 32, 63, 109, 112 subordinate clauses 79, 80, 81
relexicalisation 112-116 syllabus, for spoken vocabulary 122-127
repetition 81, 112-116 synonyms 116
reporting verbs 155, 171
resolution 63 tag 60
response 53 tails 60, 76
right-dislocation 62, 77 teacher awareness 52
role of listener 116, 119 tend 125-26
in oral narrative 117 tense
and aspect 93, 95
semantic preferences 22 speech reports 166-167, 171
sequence 91, 103, 105 text types 8, 26
service encounters 27, 28, 63 textbook dialogues 66
shared knowledge 40, 43 theme 77
silences 41, 43, 44 theories of spoken genre 26-46
situation 105 three Is methodology 67
situation 91, 92 three Ps methodology 67, 112
social compact 32 topic 31, 76, 77, 109
social context 17 negotiation 112-116
sound recordings 12 switch 31
spatial deixis 45 topicalisation 82
speech act theory 18, 19 transaction 50
speech function, and lexical density boundary 67
110-112 marking 52
speech genre, theories of 27-30 transactional goals 112
speech in writing 171 transcription 7, 12
speech reporting 150-175 of back-channel 15
in written language 76, 152 of pauses 15
in spoken language 76 of speaker turns 14
direct speech 151 of sub audible segments 15
indirect speech 151 problems with 13
natratised 151 trinomials 130
teaching 172 turn taking 29, 57, 58, 81
with -ing form verbs 159-162
speech types 8 used to 96, 98, 100
206 + Index

utterance 29 written and spoken grammar 76-79

written and spoken, patterns of co-
vague language 118 occurrence of verb forms 90
varieties of English 21 written language v. spoken language 39,
vocabulary 108 see also lexis 47, 125, 126
discourse analysis 108 word order 61
and spoken language 108 in speech reports 171-172
roles of vocabulary choice 109 word stress 130
would 96, 98
will 100, 103, 104,105 written language v. spoken language 16
 SPOKEN
LANGUAGE &
APPLIED LINGUISTICS
poken Language and Applied Linguistics brings
together a number of separate studies by the
author, all based on the Cambridge Nottingham
Corpus of Discourse in English (CANCODE) spoken
corpus, and weaves them together to illustrate the
central role the study of spoken language can play in
- _ applied linguistics. After an introduction to the corpus
and its make-up, McCarthy lays out the: main
components of a theory of spoken genres, with corpus
examples. There then follows a broad discussion of what
can/should be taught about the spoken language,
followed by chapters on discourse grammar and on the
sometimes parallel, sometimes different, grammatical
_ patterning of spoken and written texts. The author then
turns his attention to lexis, with a general overview of
the vocabulary of spoken language, plus a specific study
of idioms in the CANCODE corpus. The book closes
with a look at another central area of language teaching,
speech reporting. All the chapters are amply illustrated |
with real corpus examples, and McCarthy brings to
bear his more than 30 years’ experience in language
teaching to argue for putting spoken language right at
Bic centre ofthe syllabus.

“ISBN 0-52 -s57eb «