(Human Cognitive Processing, V. 26) Graham Low - Et Al - Researching and Applying Metaphor in The Real World-John Benjamins Pub. Co (2010)

Researching and Applying Metaphor in the Real World
Human Cognitive Processing (HCP)

Human Cognitive Processing is a book series presenting interdisciplinary
research on the cognitive structure and processing of language and its
anchoring in the human cognitive or mental systems.
Editorial Advisory Board

Melissa F. Bowerman Eric Pederson
Nijmegen Eugene, OR
Wallace Chafe François Recanati
Santa Barbara, CA Paris
Philip R. Cohen Sally Rice
Portland, OR Edmonton, Alberta
Antonio Damasio Benny Shanon
Iowa City, IA Jerusalem
Morton Ann Gernsbacher Lokendra Shastri
Madison, WI Berkeley, CA
David McNeill Paul Thagard
Chicago, IL Waterloo, Ontario
Volume 26
Edited by Graham Low, Zazie Todd, Alice Deignan and Lynne Cameron
Researching and Applying
Metaphor in the Real World
Edited by
Graham Low
University of York
Zazie Todd
University of Leeds
Alice Deignan
University of Leeds
Lynne Cameron
The Open University
John Benjamins Publishing Company

Amsterdam / Philadelphia
TM
The paper used in this publication meets the minimum requirements of
8
American National Standard for Information Sciences – Permanence of

Paper for Printed Library Materials, ansi z39.48-1984.
Library of Congress Cataloging-in-Publication Data
Researching and applying metaphor in the real world / edited by Graham Low ... [et al.].
p. cm. (Human Cognitive Processing, issn 1387-6724 ; v. 26)
Includes bibliographical references and index.
1. Metaphor. I. Low, Graham.
P301.5.M48R48 2010
808--dc22 2010026150
isbn 978 90 272 2380 7 (Hb ; alk. paper)
isbn 978 90 272 8777 9 (Eb)
© 2010 – John Benjamins B.V.

No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any
other means, without written permission from the publisher.
John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands
John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa
Table of contents
Editors’ introduction vii

Graham Low, Zazie Todd, Alice Deignan and Lynne Cameron
1. The wonderful, chaotic, creative, heroic, challenging world
of Researching and Applying Metaphor: A celebration of the past
and some peeks into the future 1
Raymond W. Gibbs, Jr
Section 1. Metaphor and language learning

2. Can people be cold and warm? Developing understanding
of figurative meanings of temperature terms in early EFL 21
Ana M. Piquer-Píriz
3. Grasping the point: A study of 15-year-old students’ comprehension
of metaphorical expressions in schoolbooks 35
Anne Golden
4. “Drugs, traffic, and many other dirty interests”:
Metaphor and the language learner 63
Gill Philip
5. The gaps to be filled: The (mis)treatment of the polysemous senses
of hand, cool and run in EFL text books 81
Elisabet Amaya Chávez
6. A cross-cultural study of metaphoric understanding 105
Chongying Wang and Ann Dowker
Section 2. Capturing and analysing metaphors

7. Love, metaphor and responsibility: Some examples
from Early Modern and Present-Day English corpora 125
Heli Tissari
8. A critical look at the desktop metaphor 30 years on 145
Marina Terkourafi and Stefanos Petrakis
vi Researching and Applying Metaphor in the Real World
9. Pragglejaz in practice: Finding metaphorically used words

in natural discourse 165
Gerard J. Steen, Ewa Biernacka, Aletta G. Dorst, Anna A. Kaal,
Irene López-Rodríguez and Trijntje Pasma
10. Mapping principles for conceptual metaphors 185
Kathleen Ahrens
11. Systematicity in metaphor and the use of invariant mappings 209
Alan M. Wallington
12. Attitude, style and context: Matching cognitive and aesthetic
accounts of poetic interpretation 245
Elsbeth C. Brouwer
13. A genre approach to imagery in winespeak: Issues and prospects 265
Rosario Caballero and Ernesto Suarez-Toste
Section 3. The function of metaphor in discourse

14. Wot no similes? The curious absence of simile in university lectures 291
Graham Low
15. Metaphor marking and metaphor typological and functional ranges
in business periodicals 309
Hanna Skorczynska Sznajder
16. Critical analysis of creative metaphors in political speeches 321
Ralph Mueller
17. Metaphor in physical-and-speech action expressions 333
Lynne Cameron
18. The evaluative properties of metaphors 357
Alice Deignan
Index of names 375

Index of terms 381
Editors’ introduction
Graham Low*, Zazie Todd**,
Alice Deignan** and Lynne Cameron***
*Universityof York, UK / **The University of Leeds, UK /
***The Open University, UK
Few researchers in applied language studies and related disciplines can be un-
aware of the explosion of interest in metaphor in the last three decades. Metaphor
is now studied in a wide range of academic disciplines, using data from many di-
verse contexts, and in many languages. This current interest in metaphor was ini-
tially driven by work within the cognitive linguistics tradition, still probably the
dominant strand within metaphor research today. With the publication of Meta-
phors we live by (Lakoff and Johnson, 1980) and related articles, George Lakoff
and his fellow cognitive linguists advanced the framework known as ‘conceptual
metaphor theory’, or ‘cognitive metaphor theory’. Much metaphor research still
either works within this framework, or, increasingly, is testing its boundaries as
an explanatory and predictive theory.
Although many of the stronger claims put forward by the developers of con-
ceptual metaphor theory are still the topic of debate and challenge, the theory has
offered insights that are central to metaphor research, and which underpin most
current work, across many disciplines. These insights include:
– The ubiquity of conventional metaphor;
– The central importance of metaphor to the expression of abstract ideas;
– The ideological role of metaphor;
– The systematicity of much metaphor.
However, cognitive linguists do not – directly – set out to offer explanations for
the social aspect of human behaviours, discourses and language, and many ap-
plied researchers therefore view their work as partial, albeit immensely valuable.
In recent years, the gap in research that takes a social as well as a cognitive ap-
proach is gradually being reduced. There has been a slow but steady increase in
the number and range of studies that explore metaphor in use in social interaction
viii Researching and Applying Metaphor in the Real World
and studies that are concerned with the implications and applications of meta-
phor. This developing strand is termed here ‘real-world metaphor research’.
What characterises real-world metaphor research is its determination to take
account of the importance of two key factors in theorising and empirically inves-
tigating metaphor: use in context, and language. Firstly, metaphor is assumed to
be shaped by its use in contexts of human communication; how people use meta-
phor, for particular purposes and in specific situations, gives rise to the nature of
metaphor. Furthermore, because use in context often involves other people, either
directly or indirectly, social factors must take their place alongside cognitive fac-
tors. Secondly, the language of metaphor is assumed to be much more than simply
the verbal expression of a conceptual mapping; the language resources available
to a language user in a particular context will influence how metaphor is formu-
lated and what can be done with it. The commitment of real-world metaphor
research to these two assumptions influences the kind of research that is carried
out. Real-world metaphor researchers often select research questions or problems
in order to bring about positive change in contexts of use. They will try to collect
naturally occurring data in the context of use, or, if doing more quantitative or
experimental work, will attempt to maximise ecological validity. Research reports
will include justification for the choice of data, methods of data collection, analy-
sis and interpretation in terms of maintaining connections with the context of
use. Metaphor identification procedures will be designed or chosen to cope rigor-
ously with metaphor as it is actually used. The commitment to the importance
of language in metaphor use leads real-world metaphor researchers to question
some of the assumptions of conceptual metaphor theory, particularly those about
the pre-existence of conceptual metaphors in the minds of individuals. No single
united response to conceptual metaphor theory has yet emerged, however; as will
be clear from this volume, real-world metaphor researchers deal with the issues
raised by conceptual metaphor theory in different ways. A further implication of
taking language seriously is the need to investigate metaphor at all levels of lan-
guage, from genre to lexeme, and in all types of language, from poetry to the most
prosaic, since it cannot be assumed that metaphor takes the same shape or works
in the same way. Cross-linguistic studies of metaphor in use may suggest helpful
interventions in language-learning contexts.
The eighteen chapters in this collection reflect this drive in the range and va-
riety of their approaches to real-world metaphor research. They were written by
researchers working within their various contexts to address practical problems
in their own disciplines. The researchers are from diverse academic and applied
fields, including language teaching, applied linguistics, psycholinguistics, literary
studies, and computational linguistics.
Editors’ introduction ix
While the chapters by Cameron, Deignan and Low were specially written
for this collection, the remaining fifteen started life as conference presentations
at the sixth conference on Researching and Applying Metaphor (RaAM), held at
the University of Leeds, England, in April 2006. This was the tenth anniversary
of a conference that has grown from small beginnings, and it is worth briefly
summarising its history at this point. The first event in the RaAM series was also
held in Yorkshire, England, at the University of York, in January 1996. A seminar,
which brought together about 20 metaphor scholars and was made possible by
a grant from BAAL (the British Association for Applied Linguistics) and Cam-
bridge University Press, was organised by Lynne Cameron and Graham Low.
The collection Researching and Applying Metaphor, edited by Lynne Cameron
and Graham Low (1999), was developed from some of the papers from that first
event. The second RaAM conference was held at Copenhagen, in May 1997, fol-
lowed by the third at the University of Tilburg, the Netherlands, in July 1999.
RaAM ventured outside Europe, to the University of Manouba, Tunisia, for the
fourth conference, in April 2001. The fifth conference was held at the University
of Paris in September 2003. Each successive conference has attracted more del-
egates and presenters; RaAM 6 had six parallel sessions, with over 100 delegates.
A number of publications have resulted from RaAM conferences after 1996, in-
cluding Collin (1999), Steen (2004) and Maalej (2005).
Raymond Gibbs has given a plenary paper at each of the six conferences up
to and including 2006. He is the author of Chapter 1 in this collection, where he
establishes a set of basic criteria for real-world metaphor research and hazards
a glimpse into the future concerning research problems that need solving or
which researchers are starting to explore. It is perhaps more usual to preface an
edited collection of papers with a longish overview of the field by the editors,
but in this case it was felt that Gibbs’s chapter, based on his plenary talk, covered
much of the relevant ground and could serve as the intellectual springboard for
the other chapters.
Researching and Applying Metaphor in the Real World is then divided into
three sections, each reflecting central strands of interest. The first comprises six
papers on aspects of language learning. The processes involved in learning to use
metaphor in a foreign or second language are explored, as are the selection and
presentation of metaphors in language teaching materials and cross-cultural re-
sponses to metaphor. Researchers with experience of English Language Teach-
ing (ELT) practice have contributed importantly to the development of metaphor
within applied linguistics (for example, Low, 1988; Littlemore and Low, 2006).
They have a concern with central and typical language use, as that is what their
students need to learn. They are also concerned with the processes involved in
learning metaphors in a foreign language; a deeper understanding of these will

help teachers to support learning more effectively. As well as the more familiar
ELT world of the teenage and adult learner (Chapters 3 and 4 by Golden and
by Philip), the newer and expanding field of young foreign language learners is
represented in these papers (Chapter 2 by Piquer-Píriz, with Amaya Chávez in
Chapter 5 covering both primary and secondary levels). This section also includes
a comparative study of Chinese vs English speakers (Chapter 6 by Wang and
Dowker). The chapters also cover a range of approaches. While Piquer-Píriz and
Golden report studies which involved working directly with learners, Wang and
Dowker compare children with adults, Philip uses examples of learners’ writing as
evidence for her thesis and Amaya Chávez examines textbooks that learners use.
The second section consists of studies that in different ways develop research
methodology and classifications. They attempt to identify and analyse metaphors
in a way that is robust, explicit, systematic and reliable, and they point to the inher-
ent difficulties that are raised. They also reflect the interest of metaphor scholars
in cross-linguistic and cross-cultural issues, here interpreted in the widest sense.
The section begins with comparative studies of metaphor use between cultures
separated by time (Tissari in Chapter 7 on Early vs Modern English; Terkourafi
and Petrakis in Chapter 8 on the evolution of the computer ‘desktop’ metaphor
over 30 years), and between different modes of communication (Terkourafi and
Petrakis). The use of introspective studies with a small number of informants has
sometimes led to comparative metaphor studies which are simplistic and even in-
accurate. These contributions avoid that danger by employing data from language
in use, as well as experimental data; and their findings point to the key role of
context in shaping metaphor use.
Ahrens (Chapter 10) notes methodological problems with going ‘below’ the
level of conceptual metaphor (and using, for example, Grady’s primary metaphors),
while Wallington (Chapter 11) argues conversely that analysts would be better off
restricting the complexity of the sets of correspondences they attach to individual
conceptual metaphors. Studies in this section contribute to addressing a range of
weaknesses in some previous work in metaphor, including the fact that the frame-
works put forward were rarely falsifiable, and that reliability checks have, until
recently, rarely been undertaken outside psycholinguistic research (both points,
as Steen and colleagues argue in Chapter 12, simply reinforcing the need for trans-
parent, theory-neutral metaphor identification procedures). Caballero and Suarez-
Toste in Chapter 13 pursue questions of appropriateness by suggesting the value of
pairing insider and outsider perspectives in the analysis of specialised discourse,
thereby removing some of the tension between -emic and -etic research (see Lillis,
2008). Brouwer too in Chapter 14 is concerned about appropriate analysis of
Editors’ introduction xi
s pecialist discourse, but for her the problem is how to isolate and describe the po-
etic, if conceptual metaphor theory sees ‘poetic’ metaphor everywhere, including
in non-specialist discourse.
The third and final section attends to metaphor in specific types of discourse,
with each chapter making use of corpus data to understand more about the specific
nature of metaphor in use. Although conceptual metaphor theory contributed to
opening up research into metaphor in discourse and in particular the ideological
role of metaphor, it fails to offer a sufficiently specific theoretical account for the
intricate patterns revealed by new data analyses. Low in Chapter 14 finds no sup-
port for the assumption that metaphoric similes would appear frequently and in
explanatory analogies when he examines four spoken university lectures. Similes
in fact occurred very rarely, mainly in the more conversational style lectures, and
when they did occur they were not global or overarching, but used to manage local
interaction problems. Skorczynska Sznajder (Chapter 15) investigates the mark-
ing of metaphor in the discourse of business periodicals, in which various expres-
sions around metaphor appeared to be designed to signal or anticipate metaphor
use. Metaphor marking is found not to be related to the novelty or conventionality
of metaphors but rather to their function in the specific discourse context and
cotext. Mueller (Chapter 16) explores metaphor creativity in the genres of politi-
cal speeches, and finds it often generated by creative combinations of metaphors.
His close examination of particular metaphors reveals the difficulty of identifying
mappings from discourse data, in turn raising doubt about theory which requires
metaphors to be based exclusively on underlying mappings. Cameron’s chapter
(17) examines how people engaged in face-to-face conversation make use of a
particular kind of expression related to metaphor that she calls “physical-and-
speech-action” expressions. She traces the complex relations between language
and context, moving between particular discourse settings and a larger corpus.
Like Deignan in the following chapter (18), she finds conceptual metaphor theory
too broad to account for the particulars of real-world language use, and suggests
an alternative theoretical account. Deignan investigates the discourse phenom-
enon of metaphor evaluation using corpus techniques that she has developed for
metaphor research. Corpus analysis shows multiword figurative expressions to be
important numerically, and possibly more important than individual words as the
locus of metaphoricity. Again, theoretical explanations of metaphor in discourse
are required to take account of the highly specific characteristics of the evaluative
force that metaphorical expressions develop through use.
Together, the collection provides a snapshot of real-world metaphor studies that

attempt to uphold the highest standards of empirical research and to address,
xii Researching and Applying Metaphor in the Real World
without compromise, data of real-world phenomena, qualities that we hope char-

acterise RaAM conferences and the recently established association that the con-
ferences have led to. As ever, successful research studies, such as those reported
here, open up possibilities for further work. The dynamic and ever-changing na-
ture of human communication means that, while we are unlikely ever to reach
a full understanding of it, researching metaphor use in human communication
remains an intriguing and exciting venture.
References
Cameron, Lynne and Low, Graham (Eds.) (1999). Researching and applying metaphor. Cam-
bridge: Cambridge University Press.
Collin, Finn (Ed.) (1999). Special edition of Metaphor and Symbol, 14 (1).
Lillis, Theresa (2008). Ethnography as method, methodology, and “deep theorizing”: Closing
the gap between text and context in academic writing research. Written Communication,
25, 353–388.
Littlemore, Jeannette and Low, Graham (2006). Figurative thinking and foreign language learn-
ing. London: Palgrave Macmillan.
Low, Graham (1988). On teaching metaphor. Applied Linguistics, 9, 125–147.
Maalej, Zouhair (Ed.) (2005). Metaphor, cognition and culture: Selected papers from the fourth
conference on researching and applying metaphor. Manouba, Tunis: Faculty of Letters, Arts,
and Humanities.
Steen, Gerard (Ed.) (2004). Researching and applying metaphor across languages, special issue
of Journal of Pragmatics, 36 (7).
chapter 1
The wonderful, chaotic, creative,

heroic, challenging world of Researching
and Applying Metaphor
A celebration of the past and some peeks
into the future
University of California, Santa Cruz
This chapter provides a brief overview of the many intellectual accomplish-

ments of real-world metaphor scholars during the past ten years, and offers my
personal speculations about the course of future research on metaphor in the
real world. I pay particular attention to methodological advances and the will-
ingness of current metaphor scholars to address difficult problems in trying to
study metaphor in various real-world contexts. This attention to methodological
problems has direct implications for contemporary theories of metaphor, and
several questions will likely attract the attention of metaphor scholars over the
next ten years, many having to do with understanding the complexity of meta-
phoric meanings in context.
Keywords: metaphor, real world, future research
1. Introduction
Soon after I entered graduate school in cognitive psychology back in the late
1970s, I decided to study the topic of metaphor. Although I had support on this
from my academic advisor, there were many people, both at my university and
from elsewhere, who tried to dissuade me from starting out on this journey.
“Metaphor is too hard”, “Metaphor is just poetry”, or even “Metaphor is a career-
killer” were among the many pessimistic phrases I heard from others. But there
was something about the topic that I found fascinating and I was convinced that
understanding how metaphor worked and was interpreted could offer significant
challenges to contemporary theories of language processing and human cognition,

more generally. So I persevered, and after my first year, discovered the first bibli-
ography that listed virtually every notable article that had ever been published on
metaphor (Shibles, 1974). I took it upon myself to find every one of those articles,
read them, and learn what I could about the topic. At this time, I spent hundreds
of hours alone in a special area of my university library, digging deep, and think-
ing hard about what I soon discovered was indeed a difficult topic.
During this time “I wandered lonely as a cloud”, finding intellectual friend-
ship only through the metaphor scholars whose work I read. But my reading of this
work gave me several distinct impressions about the nature of metaphor research.
First, people in my own field of experimental psycholinguistics were primarily
studying people’s learning and understanding (and sometimes memory for) iso-
lated metaphorical expressions apart from realistic social and linguistic contexts.
Philosophers were also interested in isolated metaphorical expressions, yet even
worse, in my view, focused their attention on only a few treasured examples such
as “Man is wolf ” or “Juliet is the sun”. Linguists generally studied the ways that
individual verbal metaphors did, or did not, violate selectional restriction rules.
In all of these cases, the debates concentrated mostly on the ways that the tenor
and vehicle terms contrasted, compared, or interacted, as well as on the ways that
verbal metaphors literally deviated from potential pragmatic contexts of use. The
rise of cognitive linguistic research in the early 1980s significantly reoriented the
study of metaphor and alerted people to the possibility that systematic patterns
of metaphorical language directly reflected enduring patterns of metaphorical
thought. Yet here too, the strong emphasis was on individual linguistic expres-
sions apart from their real-world contexts, or what speakers/writers might intend
to communicate by the use of these expressions.
Most generally, the interdisciplinary study of metaphor treated the topic as if
it were (in Eliot’s words) “like a patient etherized upon a table”. Like pathologists
hovering over a corpse, metaphor scholars would poke at “Man is wolf ” or “kick
the bucket” wondering if these were really dead or alive, sometimes turning the
body over or around, to see if it looked different; does “Man is wolf ”, for example,
mean something different than “Wolf is man”? But, again, there was little concern
with where metaphor came from, or what metaphor actually does when bouncing
around in the real world of human speakers and interaction. Even as I, and oth-
ers, began to conduct experimental studies in the late 1970s and 1980s, looking at
the effect of context on different aspects of figurative language interpretation, the
emphasis was really on how short discourse contexts facilitated people’s compre-
hension of some statement (like “Regardless of the danger, the troops advanced”)
as having metaphoric as opposed to literal meaning. What metaphors actually
communicated in real-life situations, the roles that metaphors had in structuring
Chapter 1. Some peeks into the future
certain domains of thought, and how metaphors shaped and reflected culture,
were not topics that attracted much attention.
In 1995 Lynne Cameron and Graham Low invited me to attend a conference
in York, England entitled “Researching and applying metaphor”, that gathered
together a small set of metaphor scholars from a variety of fields and perspectives
to share their common interests in understanding metaphor in research and ap-
plied contexts (Cameron & Low, 1999). I was excited to receive this invitation and
was delighted when I attended the conference, met these other metaphor scholars,
and realized that there were others out there who shared my interest in studying
metaphor in the real world, and wanted to openly discuss some of the challenges
that an ecologically-valid study of metaphor entailed.
My aim in what follows is to review some of the accomplishments within this
tradition, note some of the enduring issues that metaphor scholars continue to
struggle with, and offer some speculations on the future directions in interdisci-
plinary and real-world metaphor research.
2. The scope of real-world metaphor research
I want to begin by suggesting a number of general propositions about what is

needed to research the way(s) in which metaphor is used in the real world, or can
be applied to real-world issues.
Firstly, there is the simple point that speakers/writers use a whole variety of
metaphor types when they speak or write; real-world metaphor research cannot
be limited to the analysis of just certain types of metaphor.
Secondly, there is a similar need to expand theoretical perspectives to ac-
count for empirical data that is encountered (I shall deal further with this point
below). It is fair to say in recent years that a significant amount of discussion and
criticism has been focused around Conceptual Metaphor Theory (with about as
much negative evidence presented against this perspective, as positive evidence
shown to support it). Many real-world metaphor research programs, have how-
ever, embraced more than one theoretical perspective, or at least included some
discussion of how particular data, whether they be linguistic examples/corpora,
or psychological results, relate differently to different theoretical perspectives.
Both the above propositions derive from the more general requirement that
real-world metaphor research should aim to be ‘ecologically valid’. By this I mean
that it should not base its conclusions purely on constructed, decontextualized
examples. What is good is that real-world researchers have begun to study meta-
phor in virtually every type of discourse, in fields as diverse as politics, the law,
music or food. At the very least, such research collectively provides evidence on
the incredible ubiquity of metaphor in language, and just as importantly, shows the
particular functions that metaphor has, often in different ways, across a variety of
discourses. In a related way, this research reveals the incredible extent that meta-
phor infiltrates, and indeed significantly structures, a large variety of linguistic
and non-linguistic domains, from conversation and musical lyrics, to dance and
synaesthesia.
Thirdly, there is a need to examine metaphor across and within many differ-
ent languages and cultures, and here the interests of real-world metaphor research
coincide with those of ‘traditional’ research perspectives. The good news is that a
vast number of languages have in fact been studied. Indeed, no language has ever
been found to not use metaphor, and this research provides additional evidence
on the ubiquity of metaphor in language and thought. A real-world approach has
also made clear the extent to which language differences are rooted in varying
conceptual or thought patterns, or conventional aspects of culture.
Fourthly, a real-world approach to metaphor needs to relate psychological
states and processes, such as ‘understanding’, closely to the actual contexts in
which those understandings are constructed, are used and/or are constrained.
I argued in 1994 that, as regards understanding at least, researchers need to dif-
ferentiate between four things:
Metaphor processing: The very fast, mostly unconscious processes that lead to
metaphor comprehension in real-time listening and reading.
Metaphor interpretation: The slower, sometimes conscious, reflective process-
es associated with richer, deeper metaphoric meanings being understood.
Metaphor recognition: Both the processes by which ordinary people some-
times, but not always, recognize that a particular word or phrase conveys met-
aphorical meaning, and the strategies by which analysts identify metaphors in
speech and writing.
Metaphor appreciation: The processes that lead to metaphors being appreci-
ated or evoking affective responses.
It has sometimes been mistakenly assumed that the study of just one of the above
can provide a comprehensive overview of metaphor understanding, and that a
theory of metaphor recognition is equivalent to one of metaphor understanding.
Perhaps the most serious error occurs when theorists embrace their own con-
scious intuitions about metaphor interpretation as if they were a direct insight
into all parts of metaphor understanding. Real-world metaphor studies have,
however, begun to examine all four aspects of how metaphors are understood,
recognized, and appreciated.
Lastly, real-world metaphor research needs to explore situations which are as

much social as psychological and to try and examine how both aspects interact.
Thus there needs to be, and has been, research into: the neuropsychology of meta-
phor comprehension and use, metaphor use in different bi- or multilingual con-
texts (including the role of metaphor in translations), the development of meta-
phor understanding and use in children, and developing control over metaphor
in second/foreign language acquisition.
3. Ongoing struggles
As researchers have struggled to put some or all of the above propositions into
practice over the last decade or so, a series of themes have emerged, involving
questions about the following six contrasts:
Diachronic–synchronic
To what extent are linguistic analyses of metaphors in speech and writing
reflective of diachronic as opposed to synchronic processes?
Language as system – language as use
To what extent are linguistic analyses of metaphor reflective of generaliza-
tions about the language system as opposed to actual language use?
Idealized speaker/hearer – real speaker/hearer
Similar to the previous contrast, to what extent are analyses of metaphor in
speech/writing reflective of idealized speaker/hearers or actual speaker/hearers
in the real world?
Context and conceptual metaphor
How does one identify conceptual metaphors in real discourse? What repli-
cable, reliable procedure can be developed to do this?
Intuitions of the analyst – objective assessments
Should judgments about metaphor (e.g., the identification of metaphorically-
used words and phrases, the identification of conceptual metaphors) be per-
formed relying on the intuitions of individual metaphor analysts, as opposed
to through the application of a more objective procedure or assessment?
Mind–language/language–culture
What kind of inferences about human cognition and culture can analysts
make from linguistic analyses of metaphor?
The enduring nature of these issues do not reflect any sort of failure on the part
of real-world metaphor scholars, but demonstrates the care with which they typi-
cally do their work and their resistance to making glib generalizations, sometimes
seen in other areas of metaphor studies.
4. Important analytic advances
Despite the continued existence fundamental questions, real-world metaphor re-

search has, I would argue, made advances in at least seven areas:
4.1 The messy reality of metaphor use
Metaphor is evident throughout language, thought, and culture. But metaphor

does not always appear in nice, neat packages that can be easily plucked out from
some context for analysis. Speakers use metaphorical language, and engage in
metaphorical thought, in complex, often-contradictory patterns that make sim-
ple conclusions about both the ubiquity and structure of metaphor difficult to
make. Rather than retreat back to made-up, isolated examples, real-world meta-
phor scholars have exhibited great enthusiasm for uncovering the messy reality
of metaphor use, and, again, focused attention on the benefits and limitations of
different methods for doing such analyses.
4.2. Attention to discourse, not single utterances
The study of metaphor has advanced considerably through studies that give
proper attention to metaphor in real discourse, and to see how metaphor gets
extended across discourse (e.g., across several conversation turns or throughout
and across texts).
4.3 Importance of corpus research
One of the greatest advances, in my view, in metaphor research over the past ten
years is the development of corpus research. Many claims about the existence,
prevalence, and structure of different metaphorical patterns seen in the interdis-
ciplinary metaphor literature, have been shown, through analyses of metaphor in
real corpora, not to be true, or at the very least to be less credible than initially
thought (Deignan, 2005). Metaphor scholars are increasingly turning to avail-
able corpora to better test various hypotheses about the frequency and form of
etaphors, which has resulted in greater caution in scholars relying solely on

m
their own intuitions within the metaphor community. Indeed, I would urge that
all metaphor scholars incorporate corpus analyses within their respective work.
4.4 Attention to reliability of analyses
An increasing concern with the variability of analysts’ intuitions in making judg-

ments about linguistic matters has been a major worry for metaphor researchers.
Many metaphor scholars now seek to establish more objective criteria for deter-
mining instances of metaphor, and for drawing links between patterns of meta-
phoric language use and metaphorical thought, to take one notable example. The
work of the Pragglejaz Group (2007), for instance, has been devoted to trying to
find a reliable procedure for identifying metaphorically-used words in discourse.
4.5 Greater attention to quantitative analyses
Claims about the importance or ubiquity of particular metaphorical patterns, in

either language or thought, are often made without adequate empirical support.
One trend in recent real-world metaphor research is that more scholars are trying
to show data in a quantified form by either, for example, reporting the frequen-
cies with which different metaphors are seen in particular texts, or comparing the
findings from one’s own textual analysis with those seen in large corpora.
4.6 Questions about language and thought links
The major development in metaphor research over the past 25 years has been
the idea that metaphors are not just linguistic devices, but are reflections of fun-
damental structures of human thought and cognition. But most of these claims
have been made from linguistic analyses, where the existence of a conceptual
metaphor is postulated as a motivating link for different systematic patterns of
linguistic expressions. Real-world metaphor scholars are clearly interested in
such language and thought links, but are often critical of asserting associations
between language and thought, especially in terms of ‘thinking’ in the minds of
actual speakers producing conventional and novel metaphors, without some fur-
ther evidence to support such a contention. Among the questions discussed have
been (a) what kind of linguistic evidence is sufficient for drawing links between
language and thought, and (b) how does a scholar infer whether actual speakers
are thinking metaphorically, and not just using a metaphoric scheme of thought
without thinking that way?
4.7 Complexity of body, mind, language and culture interfaces
Finally, in the interdisciplinary metaphor community, scholars have become in-

creasingly interested in the bodily foundations for metaphorical language and
thought, and try to infer the contribution of cultural norms and knowledge to the
creation and use of metaphorical language. Cross-linguistic studies, particularly,
have been very helpful in showing what aspects of metaphor are motivated by
universal aspects of bodily experience and those that are specific to individual
cultures and cultural communities.
5. Some peeks into the future
For the second part of the paper, I offer my own thoughts on the future of meta-
phor research over the next ten years. These comments are clearly speculative,
and personal to an obvious extent. But at least some of the themes mentioned
below address important issues that I sincerely believe all metaphor scholars will
have to acknowledge, and struggle with, in their own future work.
5.1 How do you do your analyses?
Perhaps the most controversial challenge metaphor scholars face today concerns
the limitations of individual linguistic intuitions and the increasing desire for
more objective methods for identifying metaphorical language in real discourse
and drawing inferences from metaphoric language patterns to possible concep-
tual (primary) metaphors. Cognitive linguistic analyses of conceptual metaphor,
for instance, infer the presence of an underlying metaphorical mapping, or con-
ceptual metaphor that presumably motivates the existence of different conven-
tional expressions. Hundreds of articles on metaphor posit literally hundreds of
different conceptual metaphors for dozens of target domains. Consider one clas-
sic example, ideas are people:
(1) He is the father of modern biology.
The father corresponds to the person who had the idea / creative insight.
(2) Einstein gave birth to the theory of relativity.

To originate an idea corresponds to give birth to a child.
(3) Those ideas died off in the Middle Ages.

Ideas correspond to the body alive.
One question that immediately arises is how does one get ideas are people from
seeing these three different conventional expressions? Thus, when saying “Those
ideas died off in the Middle Ages”, why must we think of “those ideas” as people as
opposed to some other biological entity? For the first expression, “He is the father
of modern biology”, why is it that the idea of “modern biology” is assumed to be
a person, rather than the creator of modern biology being the only metaphorical
element here, namely the “He” being referred to as “the father”? Questions like
these arise for many classic cognitive linguistic analyses, which raise additional,
general questions, about the specification of the source domain: how many in-
dividual expressions must systematically relate to one another to be sufficient to
posit an individual conceptual metaphor, and how do other conceptual metaphors
for ideas (e.g., ideas are plants, ideas are objects) relate to one another?
Judging the empirical adequacy of any individual analysis is usually done as a
matter of belief in one’s own intuition, something supported by counter-examples.
Yet what is needed is a clearer description of how the metaphor analyst came up
with their classification. What were the criteria for determining what counts as a
metaphorical expression in the language? Following this, what are the criteria for
positing that a conceptual metaphor of some sort underlies the creation and use of
a set of systematically-related linguistic expressions? With a few exceptions (e.g.
Cameron, 2003; Pragglejaz Group, 2007), metaphor scholars have not provided
criteria or guidelines by which they conduct their individual analyses of meta-
phor in language and thought. Providing such criteria will be essential toward
placing interdisciplinary metaphor research on a firm empirical footing, which
implies that the results of our analyses can be verified and replicated.
5.2 Different metaphors – different theories?
A decade ago, I raised one issue that still today hampers interdisciplinary debate
in theories of metaphor. One of the great differences in approaches to metaphor
lies in the type of metaphoric language scholars wish to account for. Although
many traditional theories of metaphor typically study classic A is B or resem-
blance metaphors, such as “Lawyers are sharks” or “My job is a jail”, cognitive
linguists, following the pioneering work of Lakoff and Johnson (1980), have fo-
cused on metaphors that have implicit source domains, often ones rooted in cor-
relations with bodily experience, such as “My marriage is on the rocks” or “I don’t
see the point of your argument”. The primary emphasis in understanding resem-
blance metaphors is to recognize, usually for the first time, the way that the source
and target domains interact to give rise to novel metaphorical meaning. On the
other hand, understanding a conventional expression like “I don’t see the point
10 Raymond W. Gibbs, Jr
of your argument” is thought to depend on accessing a well-established concep-

tual metaphor, or in this case a primary metaphor such as knowing is seeing.
Not surprisingly, then, work on resemblance metaphor emphasizes the novelty
of metaphorical mappings, while work on correlational metaphors focuses on the
possible existence of enduring patterns of metaphorical thought.
The difficulty with theoretical debates about metaphor is that scholars embrac-
ing different positions are really looking at rather different kinds of metaphori-
cal language, with quite different messages about the possibility of metaphorical
thought. But these debates rarely acknowledge these differences and incorrectly
assume that certain theories (e.g. Conceptual Metaphor Theory) can be refuted
because they are incapable of explaining how, for example, novel resemblance
metaphors are understood. There are at least two possibilities for how this situa-
tion should be dealt with. First, scholars must begin to more openly acknowledge
the limits of their preferred theories in accounting for only one kind of meta-
phoric language. Under this view, there will never be a single theory capable of ac-
counting for how all metaphorical language is used, or all metaphorical meanings
are expressed, or all metaphorical thought is conceived. Different theories will
be required to handle different types of metaphor. Second, there may be a single
theory of metaphor, and anyone advocating such a position must openly account
for how different aspects of metaphor (e.g., resemblance and correlational) meta-
phors are understood. Many cognitive linguists contend that Conceptual Blend-
ing Theory (Fauconnier & Turner, 2002) has the greatest power and flexibility to
provide this kind of comprehensive account. Yet most psychologists view Con-
ceptual Blending Theory with some scepticism, given its lack of falsifiable predic-
tions, and insist that a theory grounded in experimental evidence is needed (e.g.
Career of Metaphor Theory, or Attributive Property Theory). Whether the first
(multiple solutions) or second (single solution) of these two general possibilities
becomes most evident, a key challenge for the future is to address the ‘different
metaphor – different theory’ issue seriously. I hope that the next 10 years will see
more progress in this area than the last 10 years.
5.3 What do metaphors mean?
One of the most remarkable aspects of much contemporary metaphor research

is the fact that little attention is given to what metaphors actually communicate in
discourse. Scholars from many different fields are concerned with issues related to
how metaphors are understood, and their possible functions in language, but there
is not sufficient attention to the range of meanings readers/listeners infer when
understanding various kinds of verbal metaphor. For instance, psycholinguists
Chapter 1. Some peeks into the future 11
often study how context determines whether a particular statement expresses lit-
eral or metaphorical meaning, but the description of the literal or metaphorical
meaning is rarely specified in any detail other than through rough simple para-
phrases. Metaphor is known to be ‘pregnant with meaning’ but it is often unclear
which of the many potential meanings a metaphor may communicate are actually
inferred by people in different circumstances. This is surely an important issue to
study, especially, in my view, because of the trade-off in metaphor understanding
between maximizing cognitive effects, or meanings, while simultaneously mini-
mizing cognitive effort (from Sperber & Wilson, 1995).
Let me give an example from one ongoing research project in psycholinguis-
tics that shows differences in the amount and possibly kinds of meanings people
infer from simple metaphorical expressions in context (Lonergan & Gibbs, 2010).
Consider what the speaker intends to communicate by “Marriage is an ice box” in
the following conversational exchange:
Mary said to John,
“We exchanged marriage vows ten years ago.”
“We have been married a long time.”
Mary continued,
“We are still hanging in there.”
She then said,
“Marriage is an icebox.” (1492 msec.)
(metaphorical assertion)
Mary’s final statement conveys a declarative assertion that compares her marriage
to an icebox, from which a listener presumably draws a variety of inferences, such
as that Mary’s marriage is unemotional, confining, and perhaps lacking in sex.
Now compare the meaning of this expression when it is used in a slightly different
context:
Mary said to John,
“We exchanged marriage vows ten years ago.”
“We have been married a long time.”
John then asked,
“Are you happy in your marriage?”
Mary then said,
“My marriage is an icebox.” (1403 msec.)
(metaphorical assertion + implicature)
In this situation, Mary’s utterance about her marriage not only conveys certain
things about her marriage, but also provides an indirect answer to John’s question
about whether she is happy in her marriage. Thus, Mary’s final utterance conveys
both a metaphorical assertion and a conversational implicature. On the surface, it

appears that the meaning of “My marriage is an icebox” conveys more meanings
in the second context than in the first, by virtue of the added conversational im-
plicatures in the second case, which was set up by John’s question. One reasonable
expectation is that people should take more effort, and more time, to comprehend
the metaphorical utterance in the second context than in the first. But in fact,
the results of one study showed that this was not the case. Readers actually took
slightly and significantly less time to read “My marriage is an icebox”, and many
similar metaphors, in contexts like the second context (1403 milliseconds) than
in the first type of context (1492 milliseconds).
What explains this seemingly contradictory result that metaphors conveying
conversational implicatures can take less time to comprehend than metaphors
with no implicatures? The answer to this question, of course, lies in closer exami-
nation of the discourse contexts in which metaphors are understood. In the case
above, where Mary’s metaphor also conveys an implicature, providing a “yes”
or “no” response to John’s question, it seems likely that the implicature comes
across so strongly as to make that interpretation immediately relevant, thereby
short-circuiting processing of the various possible meanings of the metaphorical
assertion “My marriage is an icebox”. Under this view, once listeners infer suf-
ficient cognitive effects from the metaphor to promptly answer John’s question,
processing will cease, as that interpretation will satisfy the principle of optimal
relevance. In Relevance Theory terms, the ostensive stimulus is relevant enough
for it to be worth the addressee’s effort to process it, and the ostensive stimulus is
the most relevant one compatible with the communicator’s abilities and prefer-
ences (Sperber & Wilson, 1995: 270). Understanding “My marriage is an icebox”
in the first context, where no question is asked, allows listeners to derive more
metaphorical meanings, in order to derive an optimally relevant interpretation
in that situation.
The important point here is that listeners appear to be drawing different num-
bers of metaphorical meanings or inferences when interpreting “My marriage is
an icebox” in the two contexts considered here. We often have the intuitive sense
that some metaphors express more complex meanings than others, especially
in varying discourse situations. One difficulty, however, is that scholars do not
have a good way of assessing this complexity, or more specifically, of being able
to individuate and count metaphorical meanings This is true for both the use
of metaphorical words in context (e.g. “I can’t stand working for my boss”) and
metaphorical expressions (e.g. “My marriage is an icebox”). I personally think
trying to address this complex problem is a major challenge for future metaphor
studies, although the issue of individuating and quantifying any sort of meaning
is relevant to all aspects of language, not just metaphor.
In general, the issue of how to individuate and count metaphor meanings

is critical for any attempt to characterize metaphor understanding as being con-
strained by the trade-off between maximizing cognitive effects while minimizing
cognitive effort.
5.4 How are metaphors used?
My discussion about trying to understand which meanings of a metaphor are

actually understood in real-world contexts suggests that the pragmatic role that
metaphors serve in discourse influences the kinds of meanings people derive
from these statements. Much real-world metaphor research over the years has
examined the various conceptual, ideological, and rhetorical roles that metaphors
may play in context. Let me now briefly describe some ongoing psycholinguistic
research that shows that people can indeed recognize the pragmatic roles that
metaphors can play in context, which also appear to affect the processing time
needed to interpret these expressions (Gibbs & Tendahl, 2006). There have been
no published studies that specifically investigated how different contexts give rise
to different cognitive effects when reading or listening to linguistic metaphors.
Relevance Theory, however, provides several suggestions on this. According to
Relevance Theory, cognitive effects are achieved by one of the following three
types: (1) new information provided by a contextual implication, (2) strengthen-
ing of an existing assumption, and (3) a contradiction and possible elimination of
an existing assumption. How might these different cognitive effects be manifested
with metaphor? Consider the following three contexts, each of which ends with
the metaphorical statement “Lawyers are sharks”.
Strengthening context
Tom said to Peter:
“Lawyers support malicious people.”
“They don’t care about the victims.”
“They just care about the money”
“Do you have anything to add, Peter?”
Peter replied:
“Lawyers are also sharks.”
New information context – contextual implication

Tom said to Peter:
“Lawyers work in a court.”
“They went to a law school.”
“They specialize in different fields.”
Peter replied:
Contradiction context
Tom said to Peter:
“Lawyers support people in need.”
“They care about their client’s troubles.”
“They are not concerned with money.”
Peter replied:
These different contexts evoke different readings of “Lawyers are sharks”. Each of
these different meanings is related to the basic metaphorical understanding of the
comparison between “lawyers” and “sharks”. But the cognitive effects one draws
from reading this metaphorical utterance in the three contexts nevertheless dif-
fer quite a bit. We have just confirmed these impressions in a study in which US
college students read one of the three contexts and final utterance above (there
were 24 contexts and metaphorical expressions presented overall). One half of the
metaphors evoked positive associations for the target concept (e.g. “Operas are
feasts”), and one half of the metaphors evoked negative associations for the target
(e.g. “Lawyers are sharks”). After reading each story and metaphorical ending,
participants rated their agreement with the following statements:
a. Peter thinks negatively about lawyers.
b. Peter thinks that Tom thinks negatively about lawyers.
c. Peter is trying to convince Tom of something about lawyers that Tom does not
already believe.
d. Peter’s remark expresses complex meanings.
The findings from this study indicated that college students are quite capable
of understanding the general metaphorical meanings of the speakers’ final ex-
pressions, as well as recognizing that these metaphors convey additional cog-
nitive effects that differ across the three types of contexts. First, participants
correctly rated the positive metaphors as expressing positive views about the
metaphor topic, and negative views about the topics of the negative metaphors.
This shows that people were sensitive to the general positive or negative mean-
ings of the metaphorical statements. Second, participants saw the metaphors in
the strengthening contexts as agreeing more with the addressees’ beliefs than
when new contextual implications were communicated. This was especially the
case for positive metaphors. Third, participants recognized, in the contradictory
contexts, that speakers were trying to convince addressees of something they did
not already believe about the metaphor topic, but they did not make the same
assumption when reading the contextual implication or strengthening contexts.
Finally, there was a linear increase in participants’ complex meaning ratings
across the strengthening, contextual implication, and contradictory contexts for
both the positive and negative metaphors.
A second study in this series examined the speed with which people read
these metaphors in the three types of context. Participants read these stories one
line at a time on a computer screen, pushing a button once they had read and un-
derstood each statement. The results showed that people took significantly longer
to read the metaphors in the contradictory contexts (1939 milliseconds) than they
did either the strengthening (1717 milliseconds) or contextual implication (1709
millisecond) contexts.
These new experimental results are both interesting and important. They
provide empirical support for Relevance Theory’s assertion that context criti-
cally determines cognitive effort and effects. Of course, the rating task does not
cover an exhaustive test of the different cognitive effects that participants may
receive in response to the different metaphors and different contexts. However,
the results clearly indicate that the cognitive effects of metaphors vary widely ac-
cording to the context, and specifically show an increase in the cognitive benefit
of metaphors from strengthening, via contextual implication to contradiction
contexts. At least in this case, there is a strong association between understand-
ing more complex cognitive effects for a metaphor and the time needed to un-
derstand those meanings.
As natural as these results may seem, they point out something important
about metaphors that is mostly ignored in theories of metaphor. For instance,
psycholinguistic studies, again, traditionally explain understanding of metaphori-
cal versus literal meaning, but do not examine the complex pragmatic effects that
arise when metaphors are comprehended. If metaphors are more or less complex
according to their context, it will be difficult to say that metaphor, as a specific
type of language, is more or less complex than literal paraphrases. Moreover, psy-
cholinguists often suggest that metaphors should be compared with literal para-
phrases in default contexts. But what is a default context? Is it the context in which
a metaphor can be understood most easily, which would be a strengthening con-
text, or the context in which a metaphor has the highest communicational value,
which would be a contradictory context according to the results of our study? We
can only conclude that metaphors do not have meanings per se and they do not
have an invariant degree of complexity. Instead, metaphors exist in contexts and
are more or less complex accordingly, express more or fewer meanings, express
stronger or weaker explicatures and implicatures depending on the complex spe-
cifics of the real-world situation.
5.5 Mixed and multiple metaphors
Language educators often warn against the production and use of mixed met-
aphors, especially when these seem to be contradictory or come from different
source domains. But real speech and writing is full of such examples, and people
seem able to deal with them without terrible difficulty. Consider the following
newspaper headline and first two paragraphs from a March 2006 story in the San
Francisco Chronicle:
GOP is in ‘deep funk’ over Bush spending
The Republican rebellion that President Bush smacked into with the Dubai ports
deal was the tip of the iceberg of Republican discontent that is much deeper and
more dangerous to the White House than a talk radio tempest over Arabs run-
ning U.S. ports.
A Republican pushback on Capitol Hill and smoldering conservative dis-
satisfaction have already killed not just the ports deal but key elements of Bush’s
domestic agenda, and threatens GOP control of Congress if unhappy conserva-
tives sit out the November midterm election.
These two long sentences are full of metaphors (e.g. “smacked into”, “tip of the ice-
berg”, “much deeper”, “tempest”, “running”, “pushback”, “smoldering”, “killed”, and
“sit out”) and metonymies (e.g. “The Republican”, “The White House”, “Capitol
Hill”) that most people can easily interpret. Psycholinguists face the challenge of
describing how readers comprehend phrases like “smacked into”, “tip of the ice-
berg”, “pushback” and “The White House” as conveying different figurative mean-
ings that differ from talk about physical actions like smacking into something, or
pushing back something, and entities like icebergs in the ocean or large buildings in
Washington, D.C. We now know a good deal about how people comprehend some
of these words/phrases, and more generally recognize that the frequency of these
figurative forms reveals important insights into the ways people ordinarily con-
ceptualize of the concepts to which this language refers. It is fair to say that meta-
phor research has not yet addressed how people combine their understandings
of various figurative phrases to achieve more global interpretations of speakers’
complex utterances or discourse, such as seen in the above newspaper excerpts.
But this type of metaphor phenomena, and its interaction with metonymy, repre-
sents a rich, although complex, topic for future interdisciplinary research.
6. Metaphorical aesthetics
My final speculation on a future trend in metaphor research is the study of meta-

phorical aesthetics. How do people emotionally and aesthetically respond to dif-
ferent aspects of metaphorical language, as well as non-linguistic metaphor, such
as gesture and pictorial metaphor? One of the greatest complaints from tradi-
tional metaphor scholars in philosophy, literature and art is that the move toward
conceptual metaphor ignores some of the transcendental, aesthetic functions of
metaphor. Reducing metaphorical aesthetics to patterns of conventional thought
and embodiment appears, in some people’s view, to rob metaphor of its special
qualities to emotionally move us, and see old topics in a new light. Although advo-
cates of cognitive perspectives on metaphor sometimes resist this characterization
(Lakoff & Turner, 1989), there remains a gap in metaphor studies between those
interested in metaphorical thought and those studying metaphorical aesthetics.
It is now time to close this divide, and my suggestion is that scholars turn their
attention to ways of linking linguistic and psychological studies of metaphor use
with both theoretical and applied studies of aesthetics.
7. Conclusion
In this chapter, I have tried to take a step back and systematically assess the most
important research findings and analytic advances seen in recent years, with re-
spect to exploring how metaphor is used in the ‘real world’, or has applications
to real-world issues. In doing so, I have especially emphasized the attention that
many real-world metaphor scholars have given to methodological issues, stressing
clarity, transparency and reliability. In the second part of the chapter, I outlined
a set of potential but key topics for future research that should, in my view, at-
tract the attention, both empirically and theoretically, of metaphor scholars from
a broad range of academic fields.
References
Cameron, Lynne (2003). Metaphor and educational discourse. London: Continuum.

Deignan, Alice (2005). Metaphor and corpus linguistics. Amsterdam & Philadelphia: John
Benjamins.
Fauconnier, Gilles & Mark Turner (2002). The way we think: Conceptual blending and the mind’s
hidden complexities. New York: Basic Books.
Gibbs, Raymond W. (1994). The poetics of mind: Figurative thought, language, and understand-
ing. New York: Cambridge University Press.
Gibbs, Raymond W. & Marcus Tendahl (2006). Cognitive effort and effects in metaphor com-
prehension: Relevance theory and psycholinguistics. Mind & Language, 21, 379–403.
Lakoff, George & Mark Johnson (1980). Metaphors we live by. Chicago: University of Chicago
Press.
Lakoff, George & Mark Turner (1989). More than cool reason: A field guide to poetic metaphor.
Chicago: University of Chicago Press.
Lonergan, Julia & Raymond W. Gibbs (2010). How many meanings can a metaphor express?
Manuscript in preparation.
Pragglejaz Group (2007). MIP: A method for identifying metaphorically used words in dis-
course. Metaphor and Symbol, 22, 1–40.
Sperber, Dan & Deidre Wilson (1995). Relevance: Communication and cognition. Oxford:
Blackwell.
section 1
Metaphor and language learning

chapter 2
Can people be cold and warm?

Developing understanding of figurative meanings
of temperature terms in early EFL
Ana M. Piquer-Píriz
University of Extremadura
The understanding and production of figurative language in childhood has been

the subject of a great deal of research. Most of the research literature on this
topic has concentrated on the L1 but the insights gained can also be relevant to
the process of learning a foreign language. This chapter reports on a study car-
ried out with young learners of EFL in order to test their understanding of the
semantic extensions of the lexemes cold and warm when referring to personality
at ages 6, 8 and 10. This area of metaphorical language seems to be particularly
problematic for children, and yet the figurative uses of adjectives such as ‘warm’,
‘cool’ and ‘cold’ are conventional in English. As such, these uses will need to
be grasped by EFL learners. The results of this study are discussed in relation
to the role played by the children’s understanding of the two domains involved
(temperatures and personality) and some of the possible implications for the
teaching of these lexemes and their semantic extensions in the EFL classroom.
Keywords: figurative meanings, young learners, EFL, temperature terms
1. Introduction
Children’s ability to understand and produce figurative meanings in their native

language has been the subject of a great deal of research. From the perspective of
developmental linguistics and psychology, the interest has mainly concentrated
on idioms (Cacciari & Levorato, 1989; Levorato & Cacciari, 1992; Abkarian, Jones
& West, 1992) and metaphor (Gardner et al., 1975; Vosniadou & Ortony, 1983;
Nippold & Sullivan, 1987; Winner, 1988; Zurer Pearson, 1990; Glicksohn & Yafe,
1998). These studies start from the premise that in order for a metaphor to exist,
there must be some pre-existing similarity between the two elements involved. In
fact, they tend to focus their analysis on a very specific type of figurative utterance
22 Ana M. Piquer-Píriz
that often takes the linguistic form of a simile expressed as ‘A is like B’ or ‘A is

B’, for instance, ‘a cloud is a marshmallow’ or ‘a camera is like a tape-recorder’
(Gentner & Stuart, 1983). The children’s answers and, consequently, their capacity
to deal with figurative language are analysed in terms of their ability to perceive
this similarity.
In recent years, cognitive semanticists have developed a new framework in
which figurative meanings can be fitted. According to this account, the differ-
ent senses of a polysemous word are motivated by the mental mechanisms of
metaphor and metonymy and grounded in our bodily experience and the interac-
tion with the environment that surrounds us. This view has given rise to models
for lexical networks (Lakoff, 1987; Langacker, 1990) based on the notion that the
different meanings of a given lexeme form radially structured categories which
consist of a central member connected to the others via metonymy, metaphor or
image schemata. Thus, cognitive linguists do not conceive metaphor as being mo-
tivated by pre-existing similarity.1 Rather, in their view, metaphor and metonymy
themselves organize concepts with respect to one another. In order to understand
an abstract concept, we use another concept which is more concrete and our ex-
periences with the physical world serve as a natural and logical foundation for the
comprehension of more abstract notions (for further details on the importance
of the theoretical distinction between ‘resemblance’ and ‘correlational’ metaphors
and its implications for empirical studies, see Gibbs in his introduction to this
volume). This idea has had important implications for (a) applied linguistics, in
particular, for the teaching of vocabulary in an L2 (for a detailed overview see
Boers & Lindstromberg, 2008) and (b) it has also begun to serve as a theoretical
basis for research into children’s language acquisition and development (Johnson,
1999 and Nerlich, Todd, Herman & Clarke, 2003). This chapter attempts to ben-
efit from the research into these two fields and it aims to explore young learners’
understanding of the figurative meanings of the English words warm and cold
when referring to personality. It could, thus, be situated in the context of the last
two situations (both social and psychological) established by professor Gibbs in
his introduction to this volume: the development of metaphor understanding and
use in children and developing control over metaphor in second/foreign language
acquisition. And, following some of the contrasts he establishes, this chapter is a
synchronic analysis of the understanding of figurative language (language use) by
young learners in a real-world context (an EFL classroom).
Chapter 2. Understanding cold and warm in early EFL 23
2. Warm and cold and domain knowledge
The interest in children’s understanding of the meaning of warm and cold when
referring to personality can be traced back to the seminal article by Asch & Ner-
love (1960). In this paper, they explored children’s understanding of what they call
“double-function terms”, that is, adjectives that refer both to physical properties of
things and psychological properties of people, for instance, hard, deep, bright
or cold and warm. Their interest in these particular lexemes had partly arisen
from the observation that such terms exist in languages belonging to different
families and the figurative meanings assigned to them are very similar. They car-
ried out a study with five groups of children, ranging in age from three to twelve
in order to explore whether the children mastered these terms in both senses
and how they understood the nexus between the physical and the psychological
meaning. Their main results can be summarised as follows: (a) the physical sense
of the term is acquired first, (b) the psychological meaning comes later and, ap-
parently, as an independent one and (c) the dual property of the lexeme is realised
at last and not spontaneously as a rule.
In the eighties, Winner (1988) also dealt with dual-property adjectives which
she classed as psychological-physical metaphors. She established a classification
of metaphors in which she distinguished between sensory and non-sensory meta-
phors depending on whether they are based on similarity that is apprehended by
our senses or not. Non-sensory metaphors can be sub-divided into relational and
physical-psychological metaphors. The former are based on similarities between
objects, situations or events that are physically dissimilar, but, owing to parallel
internal structures, function in a similar way (e.g. clouds and sponges which func-
tion to hold and then release water). Psychological-physical metaphors are based
on a resemblance between the sensory attributes of a physical object, perceived
through any sensory modality, and psychological, non-sensory attributes of a per-
son (e.g. cheerful people are described as sunny and cranky people as sour). Ac-
cording to Winner, children have problems to interpret non-sensory metaphors
because they try to find a sensory similarity between topic and vehicle when a
non-sensory similarity is at issue.
The evidence offered by these studies and others in the same line (e.g.
Gentner & Stuart, 1983; Vosniadou & Ortony, 1983 and Evans & Gamble, 1988)
suggests that children’s perception of the properties that relate topic and vehicle
in a metaphorical comparison changes with development, with perceptual or
sensory links coming to action first, followed by relational or functional links
and then physical-psychological links. There are two possible accounts for these
findings. On the one hand, children may not be able to perceive the kinds of
similarities on which non-sensory metaphors are based. On the other hand,
failure may be due simply to insufficient knowledge of the vehicle and/or topic
domain, thus limiting the kind of connections between topic and vehicle that
children perceive. Support for the first hypothesis would lead to the conclusion
that metaphorical ability is not fully developed until a certain age. Children could
only grasp metaphors based on sensory similarities. In contrast, support for the
second hypothesis would suggest that when the child has sufficiently developed
knowledge of the elements being linked, they are able to interpret the metaphor
even if the ground is non-sensory (Keil, 1989).
This latter issue has been recurrently examined in the literature on children’s
understanding of figurative language. Several scholars (Carey, 1985; Winner, 1988;
Keil, 1989; Vosniadou, 1989; Cameron, 1996) have argued that children’s ‘misun-
derstanding’ of metaphors would not be due to their lack of metaphorical capacity
but to their developing ‘knowledge of the world’, particularly, the lack of ‘domain
knowledge’. These two notions refer to how children’s knowledge of concepts and
events and the relations among them become enriched in response to different
inputs and experiences (their own social experiences with other people and the
world that surrounds them, including all sort of inputs they are exposed to: the
explanations provided by their parents, teachers, siblings or friends but also sto-
ries, cartoons, TV programmes, computer games or the Internet). For example,
the knowledge of ‘dog’ may include the following notions: four-legged animal, it
barks, it can be a pet, it may also be aggressive, there are very different breeds, etc.
There is some share-knowledge (cultural within a community) about a concept
but also each individual may have a particular view of ‘dogs’ due to a specific
experience with them. Children are in the process of developing this knowledge
and, therefore, although even very young children may have the competence to
perceive all the kinds of similarities that adults perceive, they may lack articulated
knowledge of the domain from which either the topic or vehicle is drawn, and
thus fail to see the similarity between them.
Studies on children’s development of conceptual domains will therefore clarify
the domain differentiations which children are able to establish at different stages.
This point is particularly relevant if we attempt to apply the cognitive linguis-
tic view of polysemy to foreign language instruction. If we accept the cognitive
premise that our abstract reasoning is based on our understanding of concrete
concepts via metonymical or metaphorical projections from source/concrete to
target/abstract domains, then any methodology that aims to enhance this pro-
jection needs to take into account what kind of domain differentation a child is
capable of at various ages.
So far, the scant studies that have explored children’s acquisition of polysemy
from this perspective have focused on monolingual children. In the case of ap-
plied cognitive linguistics studies, although they are far more numerous, they
have mainly concentrated on adolescent or adult learners of EFL. However, EFL

is expanding and is introduced at ever younger ages in the educational systems
of many countries and these notions may be also applied to young learners. Re-
search carried out with young (5-, 7-, 9- and 11-year-old) Spanish learners of EFL
has shown that they were able to understand some semantic extensions of body
part terms in English. In general, these children were able to grasp the figura-
tive meaning and they often explained the links among the polysemous senses of
these lexical items by resorting to metonymy and metaphor (Piquer-Píriz, 2005a,
2008a, 2008b), as predicted by cognitive linguistics. Nevertheless, the children’s
preference for certain reasoning strategies varied at different ages, and this had an
impact on their successful understanding of the figurative meanings (MacArthur
& Piquer-Píriz, 2007). Furthermore, it was also shown that children’s reasoning
was heavily influenced by their developing knowledge of the world. For example,
when presented with some semantic extensions of head (the head of a bed, ham-
mer, stairs and a line of cars), the younger learners mostly based their reasoning
on the knowledge about the human body. In contrast, older children showed a
better differentiated and more developed knowledge of domains and were more
flexible in the use of other schemas, for example, animal schemas which led them
to a more successful identification of extensions such as the head of a line of cars
(Piquer-Píriz, 2005b).
Thus, although the application of the cognitive linguistics view of polysemy
to the learning of vocabulary in an L2 may offer an appealing alternative to tra-
ditional approaches to the foreign language lexicon from very early ages, issues
such as the kind of domain differentiation a child is capable of at various ages
or the use and deployment of certain reasoning strategies need to be taken into
account at least at this early period of life. By means of exploring the field of
temperature terms, the study presented below aims to shed some further light
on these issues.
3. Method
Two research questions were addressed in this study:

1. Are 6-, 8-, and 10-year-old Spanish learners of EFL able to identify the se-
mantic extensions of the core lexical items warm and cold, the prototypical
meaning of which they know from their English lessons, when they refer to
personality?
2. What kind of reasoning is involved in the children’s (non) recognition of these
figurative senses?
3.1 Participants
A hundred and forty children (75 girls and 65 boys) in their first (mean age: 6;1),
third (mean age: 8;5) and fifth year (mean age: 10;5) of primary school took part
in this study. Apart from a girl from Uruguay and a boy from Ecuador, the rest
came from Spain and they all were native speakers of Spanish. They attended two
state schools in Extremadura (western Spain) and they came from a variety of
social and economic backgrounds.
3.2 Stimuli and procedure
A short story in English in which friendly and unfriendly types of behaviour were
illustrated was devised (see Appendix 1).2 Dolls representing the characters in the
story were used to facilitate children’s understanding of the contents.
The sessions took place in a quiet area of the school with groups of five chil-
dren. The same procedure was followed in each session. First of all, the children’s
knowledge of the core meaning of the lexemes warm and cold was checked and
reinforced. In general, they did not have any problems to recognise and produce
these words that they often use to refer to weather conditions in their EFL class-
room. Secondly, the children were told the short story in English twice and they
were asked to retell it in Spanish to check comprehension. Finally, they had to
decide which character in the story was cold and which was warm and to justify
their decisions. Their explanations were tape-recorded and transcripts were made
of all the recordings to be analysed later from a quantitative and a qualitative
perspective.
3.3 Results and discussion
A quantitative analysis of the children’s identification of the friendly (warm)

and unfriendly (cold) characters in the three age groups offers the results shown
in Table 1.
As can be seen in the graph, in the 6-year-old group, 46% of the children
made the right choice, that is, they identified the unfriendly character in the story
as cold and the friendly one as warm. The percentage of right identifications in the
8-year-old group is very close to 70% whereas in the case of the 10-year-olds, this
percentage stays just below 60%. At first sight, it seems surprising that older chil-
dren (10-year-olds) had more problems with the task than their younger counter-
parts (8-year-olds). A qualitative analysis of the results was also carried out due
UNDERSTANDING COLD AND WARM IN EARLY EFL 1
Table 1. Quantitative
results of the study
Quantitative results
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
6-year-olds 8-year-olds 10-year-olds
UNDERSTANDING
right 46%
COLD AND
68%
WARM 59%
IN EARLY EFL 2
wrong 43% 32% 41%
don't know 11% 0% 0%
Table 2. Types of answers

Table 1. Quantitative results of the study
Types of answers
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Bodily Bodily
Personality sensation Weight sensation +
clothes weight
6-year-olds 6% 66% 17% 11%
8-year-olds 18% 69% 12% 2%
10-year-olds 76% 10% 14% 0%
Table 2. Types of answers

to the nature of the study with the children freely explaining their responses. This
analysis complemented and rectified some of the quantitative data, explaining the
surprising success of the 8-year-olds. The individual examination of the children’s
answers shows that there are three main justifications used by the children to de-
cide why the two characters in the story are cold or warm and they are related to
(1) bodily sensations, (2) weight and (3) personality (for details, see Table 2).
The preferred justification in the case of the 6- and the 8-year-olds is that
related to bodily sensations, that is, these children concentrated mainly on the
prototypical meaning of warm and cold as temperature terms and particularly
in the bodily sensation of being cold or warm. In the context of the short story,
this was explained in terms of the clothes worn by the dolls used to illustrate
the story:
(1) C: “porque tiene más frío – porque lleva menos ropa [se refiere a ‘the fat
lady’] ‘warm’ (R: ésta [thin lady] es ‘warm’ ¿por qué crees?) tiene más
calor porque lleva más cosas”
C: Because she feels colder – because she’s got less pieces of clothing on [she
refers to the fat lady] warm (R: this one [thin lady] is warm why do you
think so?) she feels warmer because she’s got more things on.3
Thus, the 8-year-olds’ quantitative success on the task was only apparent. Most of
them do not relate coldness and warmth with unfriendliness and friendliness but
with their prototypical temperature senses.
This justification based on bodily sensations is also used by some of the 10-
year-olds but its presence in only marginal in this age group (10% of the answers
as opposed to over 60% in the other two groups). In contrast, most of the 10-year-
olds’ explanations are based on the correlations warm-friendly and cold-unfriendly,
that is, the figurative meanings of warm and cold referring to personality:
(2) C: ‘the fat lady’ es la ‘cold’ porque es fría y severa y que no quiere ayudar
a nadie y ‘a thin lady’ pues es más buena y creo que es ‘warm’ porque le
ayuda – no sé – que es más acogedora o como se diga.”
C: the fat lady is the cold one because she’s cold and severe and she doesn’t
want to help anybody and ‘a thin lady’ is better-natured and I think she’s
warm because she helps him – I don’t know – she’s more welcoming or
whatever the word is.
The numbers of responses based on this justification in the three age groups seem
to show a developmental pattern: 6%, 18% and 76% in the 6-, 8- and 10-year-old
groups, respectively. Interestingly, not all the 10-year-olds that used this moti-
vation qualified the unfriendly character as cold and the friendly one as warm
(what explains why only 59% of their responses are correct). Some of the children
showed a richer knowledge of people’s personalities and of temperature terms and
focused on other aspects that can be related to warm and cold:
(3) C1: “la ‘fat lady’ warm porque, repre, creo que representa la – el calor a la
maldad porque se ponen a hacer así ‘uuuuu’ y le sale humo y la ‘thin lady’
es ‘cold’ porque – porque representa al frío
C2: porque el calor es así como más fuerte
C1: como el demonio, el fuego del infierno es más malo
R: y el frío ¿por qué?
C1: porque el cielo es azul – porque el frío se representa en un color azul y el cielo
es azul pero bueno – y se representa también el caliente rojo por el infierno
que es rojo”
C1: the fat lady warm because – she repre – I think she represents the – heat
represents evil because they do like this ‘uuuu’ and smoke comes out and
the thin lady is cold because – because she represents coldness
C2: because heat is stronger
C1: like the devil, the fire in hell is worse
R: and coldness why?
C1: because heaven is blue – because coldness is represented as blue and the
heaven is blue but it is good – and heat is represented as red because hell
is red.
For this particular child, warm has not positive connotations but it is rather re-
lated to fire and hell whereas heaven is blue which is related to coldness. This
interpretation of good and evil with strong religious connotations shows how
children can be heavily influenced by the cultural values of their communities.
It is quite frequent that Spanish children of that age have just taken or are about
to take their First Communion and are developing and being inculcated in these
religious notions.
Finally, a marginal justification in all age groups is based on the idea that fat
people are more ‘calurosas’, that is, they tend to feel warmer than thin people (a
clearly embodied explanation):
(4) C: “Eh – ‘warm’ y ‘cold’ (R: ‘warm’ ‘the cold’ – ‘the fat lady’ ‘warm’ – ‘warm’ –
y ‘the thin lady’ ‘cold’ ¿por qué?) porque al ser tan gorda tiene muchas
calorías y al ser tan flaca tiene mucho frío”
C: mmm – warm and cold (R: warm the cold – the fat lady warm – warm –
and the thin lady cold why?) because she’s so fat that she’s got lots of
calories and she’s so skinny that she feels very cold
4. Conclusions
The results of this study seem to indicate that there is a development in these
young EFL learners’ understanding of the lexemes warm and cold and their fig-
urative senses in relation to personality. Further research is obviously needed. It
would be interesting, for example, to have data for older children and adults to see
how these patterns evolve. But, there seems to be a sequence from the concrete,
prototypical meaning which is more salient to the youngest learners (6-year-olds)
and that they clearly relate to bodily sensations to the abstract, figurative sense
related to personality which is accessible to most of the 10-year-olds.
The numbers of responses based on this justification in the three age groups
seem to show a developmental pattern: 6%, 18% and 76% in the 6-, 8- and 10-year-
old groups, respectively. These results are in line with those obtained by Asch and
Nerlove (1960) and Winner (1988). This sequential pattern (from prototypical to
figurative) also coincides with the acquisition order of some meanings of get in
monolingual children (Nerlich, Todd & Clarke, 2003).
It seems, then, that in the process of grasping the figurative sense of poly-
semous words children go from concrete to abstract via analogical reasoning in
their L1 and also when learning an L2. However, there are other factors which
also seem to be involved in this process. Children’s analogical reasoning is in-
fluenced by their growing knowledge of concepts, as illustrated in the different
conceptions of cold and warm people reflected in the answers of the youngest and
oldest children that participated in this study. The youngest children (6-year-old)
mainly stay in the concrete realm and relate cold and warm with their own bodily
sensations. This does not mean that they lack figurative capacity (see Piquer-Píriz,
2008b) but it seems that they have not made the links between temperature terms
and personality yet. The 8-year-olds seem to be on their way to do so but the
process does not seem to be completed yet. In contrast, most of the 10-year-olds
relate both domains but in the process of developing their world knowledge, some
notions are not clearly conventionalised yet and they make their own interpreta-
tions, as shown in Example (3).
When designing pedagogical materials aiming to foster the use and compre-
hension of the multiple senses of polysemous items in the EFL classroom, it seems
clear that which and when different semantic extensions should be introduced are
important issues. Analysing the domain differentiation which children make at
different ages can help to provide some answers.
Notes
1. Some cognitive linguistic approaches recognise the existence of some metaphors in which
the two elements are cognitively linked due to their actual similarities or to the human capac-
ity to impose resemblance between them. In this sense, Grady (1999) distinguishes between
this type of metaphor, which he calls ‘resemblance’ metaphor, and ‘correlation’ metaphors. The
latter are experientially motivated, that is, directly grounded in aspects of our experience and,
therefore, more primary and universal, according to Grady.
2. I would like to thank Fiona MacArthur for making up this story.
3. The explanations provided by the children illustrated in this chapter have been translated
into English as literally as possible to maintain the original sense. It should be born in mind that
they are spoken utterances produced by children and they have not been edited for grammar
or style neither in Spanish or English in order to preserve the children’s original words which
convey their ideas.
References
Abkarian, Gene G., Alison Jones & Gretchen West (1992). Enhancing children’s communica-
tion skills: Idioms ‘fill the bill’. Child Language Teaching and Therapy, 6, 246–254.
Asch, Solomon & Harriet Nerlove (1960). The development of double function terms in chil-
dren. In B. Kaplan & S. Wapner (Eds.), Perspectives in psychological theory (47–60). New
York: International Universities Press.
Boers, Frank & Seth Lindstromberg (2008). How cognitive linguistics can foster effective vo-
cabulary teaching. In F. Boers & S. Lindstromberg (Eds.), Cognitive linguistic approaches to
teaching vocabulary and phraseology (1–61). Berlin & New York: Mouton de Gruyter.
Cacciari, Cristina & Maria Chiara Levorato (1989). How children understand idioms in dis-
course. Journal of Child Language, 16, 387–405.
Cameron, Lynne (1996). Discourse context and the development of metaphor in children. Cur-
rent Issues in Language and Society, 3, 49–64.
Carey, Susan (1985). Conceptual changes in childhood. Cambridge, MA: MIT Press.
Evans, Mary A. & Dianna L. Gamble (1988). Attribute saliency and metaphor interpretation in
school-age children. Journal of Child Language, 15, 435–449.
Gardner, Howard, Mary Kircher, Ellen Winner & David Perkins (1975). Children’s metaphoric
productions and preferences. Journal of Child Language, 2, 125–141.
Gentner, Dedre & Patricia Stuart (1983). Metaphor as structure-mapping: What develops? (Tech.
Rep. 5479). Cambridge, MA: Beranek & Newman.
Glicksohn, Joseph & Tal Yafe (1998). Physiognomic perception and metaphoric thinking in
young children. Metaphor and Symbol, 13, 179–204.
Johnson, Christopher (1999). Metaphor vs. conflation in the acquisition of polysemy: the case
of see. In M. Hiraga, C. Sinha & S. Wilcox (Eds.), Cultural, psychological and typological
issues in cognitive linguistics (155–208). Amsterdam & Philadelphia: John Benjamins.
Keil, Frank (1989). Concepts, kinds and cognitive development. Cambridge, MA: MIT Press.
Lakoff, George (1987). Women, fire and dangerous things. What categories reveal about the mind.
Chicago: The University of Chicago Press
Langacker, Ronald (1990). Concept, image and symbol. The cognitive basis of grammar. Berlin:
Mouton de Gruyter.
Levorato, Maria Chiara & Cristina Cacciari (1992). Children comprehension and production of
idioms: The role of context and familiarity. Journal of Child Language, 19, 415–433.
MacArthur, Fiona & Ana M. Piquer-Píriz (2007). Staging the introduction of figurative ex-
tensions of familiar vocabulary items in EFL: Some preliminary considerations. Ilha do
Desterro: A Journal of English Language, Literatures in English and Cultural Studies, 53
(Metaphor in language and thought: Contemporary perspectives), 123–134.
Nerlich, Brigitte, Zazie Todd, Vimala Herman & David Clarke (Eds.) (2003). Flexible patterns
of meaning in mind and language. Berlin: Mouton de Gruyter.
Nerlich, Brigitte, Zazie Todd & David Clarke (2003). Emerging patterns and evolving poly-
semies: The acquisition of get between four and ten years. In B. Nerlich et al. (Eds.), op.
cit., 333–360.
Nippold, Marilyn & Michael Sullivan (1987). Verbal and perceptual analogical reasoning and
proportional metaphor comprehension in young children. Journal of Speech and Hearing
Research, 30, 367–376.
Piquer-Píriz, Ana M. (2005a). La Comprensión de algunas extensiones semánticas de los lexemas
‘hand’, ‘mouth’ y ‘head’ en las primeras etapas del aprendizaje del inglés (Young EFL learn-
ers’ understanding of some semantic extensions of the lexemes ‘hand’, ‘mouth’ and ‘head’).
(Doctoral dissertation, Universidad de Extremadura, 2005). Servicio de Publicaciones de
la Universidad de Extremadura.
Piquer-Píriz, Ana M. (2005b). The age factor in the understanding of figurative meaning in
EFL classrooms. Paper presented at the 2º Congresso sobre Metáfora na Linguagem e no
Pensamento (Niterói, Brasil).
Piquer-Píriz, Ana M. (2008a). Reasoning figuratively in early EFL: Some implications for the
development of vocabulary. In F. Boers & S. Lindstromberg (Eds.), Cognitive linguistic ap-
proaches to teaching vocabulary and phraseology (219–240). Berlin & New York: Mouton
de Gruyter.
Piquer-Píriz, Ana M. (2008b). Young learners’ understanding of figurative language. In
M. S. Zanotto, L. Cameron & M. C. Cavalcanti (Eds.), Confronting metaphor in use. An
applied linguistic approach (183–198). Amsterdam & Philadelphia: John Benjamins.
Vosniadou, Stella (1989). Similarity and analogical reasoning: a synthesis. In S. Vosniadou &
A. Ortony (Eds.), Similarity and analogical reasoning (1–18). Cambridge: Cambridge Uni-
versity Press.
Vosniadou, Stella & Andrew Ortony (1983). Intelligence of the literal-metaphorical-anomalous
distinction in young children. Child Development, 54, 154–161.
Winner, Ellen (1988). The point of words: Children’s understanding of metaphor and irony. Cam-
bridge, MA: Harvard University Press.
Zurer Pearson, Barbara (1990). The comprehension of metaphor by preschool children. Journal
of Child Language, 17, 185–203.
Appendix
Short story
I’m going to tell you about a little boy. His name was Sam. He was very nice, he was (6, 8, 10)
years old, just like you. Sam’s mummy was always getting cross with Sam. Do you know why?
Well, it was because Sam was always losing things. If he had a ball, he lost it, and he said to his
mum(my): “Mum, where’s my ball?.” “I don’t know, Sam” said his mum(my). “Look for it”. And
Sam looked for it. If he had a book, he lost it, and he said to his mum(my): “Mum, where’s my
book?.” “I don’t know, Sam”; said his mum(my). “Look for it”. And Sam looked for it. And Sam
lost his shoes, and he said to his mum(my): “Mum, where are my shoes?” And Sam’s mum(my)
said: “I don’t know. Look for them.” Sometimes he found his ball or his books and sometimes he
didn’t. Sometimes his mother helped Sam find his ball or his books, and sometimes she didn’t.
Sometimes she was nice and sometimes she wasn’t nice. One day, Sam lost something VERY
important. Do you know what he lost? Well, he lost his nose. Yes, his NOSE. No, don’t ask me
how he lost his nose – but he did. One moment he had a nose on his face, and the next moment
it was gone. He put his hand on his face and he said “Oh no! My nose! Where is it? I’ve lost my
nose!” Well, Sam’s mum(my) wasn’t there. Sam was in the park. He couldn’t say: “Mum, where’s
my nose?” She wasn’t in the park. But Sam saw a fat lady, and he said: “Excuse me, I’ve lost my
nose. Where is it? Can you help me find my nose?. ” But the fat lady said: Stupid boy. You’ve
lost your nose. Well, that’s your problem. YOU look for it. I won’t help you.” Poor Sam began to
cry. What a horrible lady! No nose and no help! Then he saw a thin lady, and he said: “Excuse
me, I’ve lost my nose. Where is it? Can you help me find my nose?” And the thin lady said: “Oh
poor boy. You’ve lost your nose. I’ll help you.”. So Sam and the thin lady looked everywhere for
his nose. They looked on the ground, they looked under the benches, they looked behind the
trees (expand as necessary). And in the end they found the nose! Do you know where Sam’s
nose was? No? Well, it was in his pocket! Yes, Sam had used his hankie and he put his hankie
AND his nose in his pocket. Silly boy! But wasn’t he lucky the thin lady helped him? Because
she helped him, Sam now has a nose on his face! What a nice lady!
chapter 3
Grasping the point

A study of 15-year-old students’ comprehension
of metaphorical expressions in schoolbooks
Anne Golden
University of Oslo
Metaphorical expressions as defined by Conceptual Metaphor Theory are

frequently used in all sorts of texts, including real-world school books. This
study investigates the comprehension of different types of metaphorical expres-
sions by different groups of 15-year-old students in Norway. 50 metaphorical
expressions were selected from nine textbooks in lower secondary school
and presented to 400 students in a multiple choice task. About 40% of these
students had Norwegian as their second language. A questionnaire relating to
language practice and school experience provided the basis for categorizing the
students into several groups. A comparison of the results from different groups
of students showed that the linguistic minority students as a group under-
stood substantially fewer expressions than their peers with Norwegian as their
mother tongue, and that some of the language minority groups understood
fewer than others. The metaphorical expressions were categorized into differ-
ent sets according to several variables, relating to the expressions themselves
and to the contexts in which they appeared in the introductory phrases of the
multiple choice items. Some sets of expressions turned out to be more difficult
than others. The minority students’ choice of wrong alternatives seemed to be
influenced by the resemblance in form between single words in the introduc-
tory phrase and in the distracter. This study shows the need a) to take the di-
versity of the metaphorical expressions into account in studies of metaphorical
comprehension as urged by Gibbs (this volume) and b) to focus on vocabulary
of all sorts in education.
Keywords: comprehension, metaphors, second language vocabulary, reading,

school text book
36 Anne Golden
1. Introduction
Understanding written texts is essential for young people in school. In order to glean
knowledge, students need to understand the written texts presented to them. Some
students are poor readers, in the sense that they do not fully grasp what they have
read. This is a frequent experience with language minority children in school because
they are reading in their second language. There may be a number of reasons for their
reading difficulties. Kulbrandstad (2003) claims that students’ problems in reading
in their second language are mainly due to three factors: (1) incomplete knowl-
edge of the language (the grammar, the vocabulary), (2) inadequate background
knowledge (use of inappropriate schema) and (3) poorly developed metacognitive
skills (lack of control of their own comprehension). Challenges presented by vo-
cabulary have indeed been pointed out by several researchers (Oakhill & Garnham,
1988; several articles in Huckin, Haynes & Coady (Eds.), 1993; Urquhart & Weir,
1998), and hence, the need to teach vocabulary in class has been strongly advocated
(see e.g. Sökmen, 1997; Beck, McKeown & Kucan, 2002). But the relation between
reading comprehension and vocabulary is complex (Anderson & Freebody, 1985)
and in need of further investigation (Pearson, Hiebert & Kamil, 2007). In recent
years, there has been an increase in the awareness of the extent of metaphoricity
in language in general, and consequently also in the language used in textbooks.
Vocabulary studies should therefore include the comprehension of vocabulary in
its figurative sense. Studies of figurative language in textbooks as well as of students’
mastery of this type of language are therefore needed.
2. Background
2.1 Research on vocabulary in Scandinavia
Research on vocabulary has been an important part of second language research

in Scandinavia, mainly due to Åke Viberg’s extensive work in Sweden (see e.g.
Viberg, 1993, 1999, 2002, 2005). Part of his research has been on lexical typol-
ogy and contrastive studies, especially focusing on verbs. By categorizing and
comparing the 20 most frequent verbs into semantic fields in eleven European
languages he found a regular pattern: in each field one or two verbs were by far
more frequent. These verbs were named nuclear verbs and these verbs in Swedish
had an equivalent in all or almost all of the eleven different languages he studied.
Examples of nuclear verbs from the motion field are come and go, from the field
of verbal communication say, and from the field of possession take and give, e.g.
they are all among the most polysemous verbs in the language.
Chapter 3. Grasping the point 37
2.2 Research on vocabulary in schoolbooks in Scandinavia
Vocabulary in textbooks in physics, geography and history in upper primary

school and lower secondary school in Norway was carried out by Golden and
Hvenekilde in the 1980s. The vocabulary in these books was categorized as high-
ly common school words, subject-specific words and non-subject-specific words
(Golden, 1984). The highly common school words were frequently used function
words and everyday school words (like those introduced in beginners’ textbooks
when studying Norwegian as a foreign language). The subject-specific words were
selected by subject teachers and included words like breddegrad [Eng: latitude],
samfunn [Eng: society], fordampe [Eng: evaporate]. The remainder were classified
as non-subject-specific words. A comparison of the vocabulary1 in these three sub-
jects’ textbooks showed – as expected – that the highly common school words were
found in all disciplines, whereas the subject-specific words were unique to their
particular academic disciplines in about 93% of the cases. The unexpected re-
sult was that about 55% of the non-subject-specific words turned out to be subject
specific as well; they were found only in one of three subjects.2 Examples of this
category are gni [Eng: rub] that only appeared in physics, kvist [Eng: twig] that
only appeared in geography and frykt [Eng: fear] that only appeared in history.
This shows that vocabulary as a general rule is more diverse and topic-centred
than one would think.
2.3 Research on learners’ vocabulary in Scandinavia
Viberg has also studied second language speakers’ lexical production on several
occasions (Arnberg & Viberg, 1991; Viberg, 1993; Viberg, 1998) using different
methods, including free conversation mixed with the retelling of short video clips
and informal play with flannel boards. Different age groups were studied. One
pattern emerged again and again: Certain verbs were overused while others were
underused by second language users, as compared to Swedish native speakers
of the same age. The overused verbs were the nuclear verbs (e.g. go, take). Verbs
which are typologically marked verbs in Swedish and Scandinavian languages
(e.g. set, lay) were avoided.
Golden studied the vocabulary of five students with different mother tongues
(Spanish, Turkish, Polish and Farsi) when retelling a fairy tale which had been
previously read in Norwegian (unpublished study, but reported in Golden, 2009).
Their verb vocabulary was compared with the verb vocabulary of five Norwegian
students. The foreign students used a higher proportion of nuclear verbs and text-
specific verbs (the latter normally being low-frequency verbs, but having a rather
38 Anne Golden
high frequency in the text, like grave [Eng: dig] and hugge [Eng: chop]. However,
the semantic range of the nuclear verbs used by the foreign students was narrower
compared to the range of the native speakers, very few verbs were used in a figu-
rative sense, and the collocations of the text-specific verbs were also less varied.
The foreign students did not use the typologically-marked Scandinavian durative
construction with verbs like stå, sitte and ligge [Eng: stand, sit and lie] that is typi-
cal in this text genre as well as in most oral narratives in Norwegian. These were
frequently used in the Norwegian students’ retellings.
2.4 Vocabulary comprehension research in Scandinavia
In Denmark, Golden and Hvenekilde’s vocabulary study was duplicated by Gimbel

(1995) using school books from history, biology and geography. This research was
followed up by a comprehension study of the non-subject-specific words (or the
‘pre-subject vocabulary’ as they were called in Gimbel, 1998). Turkish and Dan-
ish fifth graders were asked to explain the meanings of 50 words. The difference
between the two groups was substantial, as the Danish students attained an aver-
age of 42 words correct and the Turkish students 15.6. The answers offered by the
two groups of students when they misunderstood an item in question were also
dissimilar. The Danish students’ wrong answers displayed a semantic similarity to
the correct ones, as when the Danish noun frost [Eng: frost, cold] was explained
as something like slippery ice. The Turkish students’ answers had usually a phono-
logical similarity, as when the Danish noun grænse [Eng: border] was explained as
Danish græs [Eng: grass].
The same strategy of focusing on the form of the unknown word was found
in Haastrup’s (1991) study of Danish learners of English in the ninth grade in a
think-aloud task. But Haastrup saw a difference between the students’ guessing
strategies which related to their proficiency levels. The bottom-up strategy relat-
ing only to the orthographic /phonological aspect of the word was primarily used
by the low-proficiency learners whereas the high-proficiency learners tended to
use a top-down strategy and their guesses were guided by the theme in question.
Metaphorical comprehension research in Scandinavia

Research on metaphorical comprehension has been scarce in Scandinavia. In
Lise Iversen Kulbrandstad’s study of minority students’ reading comprehension
in Norway, a metaphorical expression medaljens bakside [lit: the back side of the
medal] appeared in the title of one of the texts read (Kulbrandstad, 1998). None
of the minority students (as opposed to their Norwegian peers) had any reason-
able idea as to what a text with a title like this would be about. However, they all
had some idea of what a medal was, and their interpretation of the title indicated
that they relied on a concrete understanding of the word. Although this study
included only one metaphorical expression, it supports the experience of many
teachers and statements from many students of a second language: certain ele-
ments of vocabulary, notably metaphorical expressions, are difficult to master in
a new language.
Birgitta Hene (2004) studied the comprehension of eight basic adjectives
in Swedish by 48 children aged 10–12, where half of the children were born in
Sweden with Swedish as their mother tongue and the rest were foreign-adopted,
some at an early age (before the age of two) and some later (as old as six years).
The adjectives belonged to two categories: the category ‘dimension’ (i.e. stor [Eng:
big]) where the basic meaning involves some sort of spatial extension primar-
ily apprehended by vision, and the category ‘physical properties’ (i.e. tung [Eng:
heavy]) where the basic meaning is primarily apprehended through perception or
bodily sensation. The children were interviewed individually about the use and
meanings of the adjectives, including their metaphorical meanings (e.g. Can you
name something that is heavy? Can people be heavy? What are they like?). The
results showed an increasing acceptance of metaphorical use as age increased,
and the responses of the children adopted at an earlier age resembled those of the
younger Swedish children. Most of the children adopted later gave only concrete
answers or different answers from the others. Even if the researcher underlines
the special situation of adopted children that makes any generalisation difficult,
she concludes that there seems to be a growth in the comprehension and aware-
ness of adjectives that develops from concrete to abstract meaning, from visible
mental state to more subtle interior properties, from feelings to social properties,
and from local to more distant environment.
3. Comprehension research: Focus on methods
Outside Scandinavia, research on metaphorical comprehension has been a schol-

arly concern for some time, the main activity – at least until recently – has been
in the field of first-language acquisition as part of the research on general devel-
opment process. The traditional view of Piaget has been that children do not un-
derstand metaphors until late in the “concrete-operational period” between the
ages of 10 and 12. This theory has been challenged by researchers like Vosniadou
(1987), Gentner (1977), and Özçalışkan (2002). They all find that younger children
understand metaphors, provided the methods used in the studies are adequate. As
Gibbs (this volume) points out, an awareness of the influencing factors is therefore
of interest in all kind of studies on metaphorical comprehension. The variables in
40 Anne Golden
question when studying the comprehension of metaphorical expressions can be

classified according to three main categories: (1) different elicitation methods used
including different types of response, (2) different ages and backgrounds of the
individuals and (3) different types of metaphorical expressions used.
3.1 Variables in comprehension tasks
The type of elicitation method is one obvious variable when testing comprehen-
sion in general and in comparing results obtained by different researchers. The
main questions concern how items are presented in the tests and in this case: how
the metaphorical expressions are presented to the participants. Are the expres-
sions presented in a context? If yes, how rich is the context? Is it written or oral?
What kind of response is demanded? Are the participants given any kind of sup-
port (cf. the examples in Gibbs (this volume) that show how context influences
the processing of metaphors and hence comprehension)? Such considerations
constitute general variables whenever a particular item in language is tested, and
certainly when testing vocabulary.
Another variable concerns the individuals being tested, especially regarding
variation in age as well as different background variables such as language back-
ground and competence. Young children have less experience with the world,
they have a smaller vocabulary than adults and they have different approaches to
testing in general. Most second language students also have a smaller vocabulary,
they might have other experiences (especially when certain culturally-specific
practices are referred to), and their knowledge schemata might be different. How-
ever, many experiences are similar across cultures – even universal (for example,
we are all humans with more or less the same body shape). And people with dif-
ferent education and occupations will likewise be familiar with different subjects,
something which might aid contextual understanding.
A third variable concerns the type of metaphorical expression presented in
the test. A glance at some well-known examples illustrates the variety of expres-
sions used. In Winner and Rosenstiel’s (1976) study, the sentence the prison guard
was a hard rock was presented for interpretation, and the informants below the
age of 14 did not fully understand the metaphorical expression involved. But to
draw a conclusion about children’s general metaphor comprehension from such
an example is not possible. There are several possible explanations for such a lack
of comprehension. One is that the personality trait in question is not one that
children have experience with, i.e. the target domain is new. The source domain
is, however, a natural object, a rock, which is well-known to everybody. The con-
text in which this personality trait is presented could have clued in the children
to the meaning of the metaphorical expression. But in this case the person in
question has an occupation that few children are familiar with. So we might ask:
when children fail to understand this sentence, what is it that they do not under-
stand? Vosniadou et al. (1984) propose that children’s lack of understanding of a
metaphorical expression might be due to factors other than being too young to
understand metaphors. They might understand the metaphorical expression hard
rock in another context, or they might understand a similar expression about a
personality trait that is more familiar, such as happiness or sadness, like being a
bouncing bubble (used in Waggoner & Palermo, 1989). The understanding of a
metaphor includes – as Gibbs (this volume) emphasizes – an interaction of the
source and the target domains. The source domain of the metaphors must thus
be familiar, and the target domain must be if not familiar, then at least recogniz-
able. And basic level categories, being acquired early, offer easily comprehended
metaphors (Cameron, 2003).
Background knowledge is certainly relevant as shown in Keil’s study (1986).
In this study, 5-year-old children could explain metaphorical expressions when
the items compared belonged to conceptual domains the children already had
distinguished.
Evans and Gamble (1988) used a metaphorical expression in the sentence her
skirt was a balloon as she walked, which was interpreted by some children as the
skirt was bright red. Evans and Gamble explain this by theorizing that children
link the attribute red with balloons, red being a salient color in children’s lives. In
discussing this example, Cameron (2003) adds that it might be due to children’s
lack of experience with skirts in the wind, i.e. lack of experience with the target do-
main. The children needed a day-with-wind schema, which might have been trig-
gered had the context been more informative. In Evans and Gamble’s example only
the words skirt, balloon, and walk were available and this was not rich enough.
Both Vosniadou (1987) and Cameron (2003) also mention the different at-
tributes that are mapped, and the nature of these mappings might affect the inter-
pretation. Perceptual or physical similarity seems to be easier than physiological
state. Cameron uses the example dancing dinghies (referring to boats moving up
and down on the sea) to illustrate this. Here the interpretations may depend on
the similarities in shape or in movement. But she adds that psychological links
established as connotations in the socio-cultural group (e.g. happiness in the sun-
shine reflected on moving water) also matters.
However, in order to gain access to the source domain and the context, the par-
ticipant must be guided by the vocabulary. So the question is whether the vocabu-
lary is of a kind that is likely to be known by the participant. Is it appropriate for the
age? Is it a basic level term or does it belong to another level in the hierarchy? Does it
belong to a topic that the participant is familiar with? A dinghy is a rather infrequent
42 Anne Golden
noun and is probably unknown even to intermediate second-language students un-

less they happen to be particularly interested in boats. A discussion of adequate
methods is thus relevant for research on metaphor comprehension in general and
particularly for research involving minority and second language students.
4. Research questions
The aim of my study is to investigate students’ comprehension of metaphorical

expressions found in real-world books, i.e. textbooks in lower secondary edu-
cation.3 The minority students are in focus. How great is their level of compre-
hension compared to that of their Norwegian peers? This question is seen from
two angles: (1) whether there is a difference between various subgroups of stu-
dents (groups that emerge from different variables) in their understanding of the
expressions and (2) whether certain sets of metaphorical expressions (sets that
emerge from different variables) are easier to understand than others. If so, to
what extent is this difference related to the way in which the metaphorical expres-
sions are presented in the test?
5. Participants
400 students in the tenth grade of lower secondary school in Norway, all approxi-
mately 15 years old, participated in the study. This grade is the highest one in
the obligatory component of the Norwegian school system. No specialization is
offered to students at this stage and they are still mixed as far as interests and abili-
ties are concerned. 230 of these students claimed Norwegian to be their mother
tongue, 170 students claimed another language to be their mother tongue. 23 dif-
ferent mother tongues were listed by the various minority students, the largest
minority group by far being the Pakistani group (61 students). Other groups were
those with Turkish as their mother tongue (16 students) and Vietnamese as their
mother tongue (12 students).
The students filled out a questionnaire asking about their language and educa-
tional background. These answers provided the information needed to categorize
the students into several subgroups. One of the questions in the questionnaire was
which language the minority students considered their best. The answers to this
question gave rise to three subgroups, those students who claimed that Norwe-
gian was their best language (70 students), those who claimed that their mother
tongue was their best (56 students), and those who claimed equal fluency in both
languages (44 students).
Another question concerned their entering in the Norwegian school system.

The answers showed that the majority of them (136 students) started in the first
grade, whereas 34 started later.
Several questions concerned their language practice at home. All of them
mentioned their mother tongue in this regard and most of them (148 students)
also mentioned Norwegian. Only 22 students did not mention Norwegian as part
of the language environment of the family.
The students were also categorized into three groups according to their results
from testing about the literal meaning of 31 items. The scores divided the students
into three tiers: students with a high vocabulary score, students with a medium
vocabulary score, and students with a low vocabulary score.
6. Procedure
The students were tested in class by their teachers. A questionnaire related to

their school and language practices was first completed. Then the students were
given a multiple choice test. The test consisted of 31 items with a lexical unit
used in a non-figurative sense and 50 items with a lexical unit used in a meta-
phorical sense. The tested words and expressions were embedded in a phrase for
contextualization. Four different options were presented that all offered an ex-
planation, an example or a description of the lexical item in question, with only
one of them being appropriate. The metaphorical expressions were all extracted
from the students’ school books. The lexical units tested in their non-figurative
sense were derived from the words used in the metaphorical expressions. Ex-
amples were presented by the teacher in order to ensure that the students had
understood the task.
7. Comprehension testing
7.1 Metaphorical expressions
The metaphorical expressions were extracted from three series of social science
textbooks used in the eighth through tenth grades, giving a total of nine text-
books. Four of the books were scanned while the rest were skimmed through.
To identify which expressions were considered metaphorical, the Concep-
tual Metaphor Theory as presented by Lakoff and Johnson (1980, 1999), Grady
(1997), Lakoff and Turner (1989), and Gibbs (1994, 1997) was used. According to
this view, there is a basic meaning of words, i.e. words belong to certain domains.
44 Anne Golden
When a word is used in reference to a new domain, there is a mapping from one
domain to another and the lexical item (word or expression) is used metaphori-
cally. Metaphors are thus seen as cross-domain mappings from the source to the
target domain, the source domain supplying the language and imagery, and the
target domain providing the actual topic referred to. The term metaphor is thus
reserved for the mappings, indicating that metaphor is a conceptual entity. The
lexical items appearing in the texts are realizations of these metaphors and are
called metaphorical expressions.
The identification of a lexical item’s basic meaning is not unproblematic,
a topic on which I will not elaborate in this article. The identification of the
metaphorical expressions employed in my study is slightly different from the
Pragglejaz method, where one criterion is that basic meaning of the lexical item
must be currently in use (Steen, 2005, this volume; Pragglejaz Group, 2007).
Pragglejaz, for example, rules out the inclusion of words like ardent as meta-
phorical expressions because its historic basic meaning relating to sense of
temperature has fallen out of use in favour of its metaphorically-derived sense
relating to emotions. Moreover, compound words like braindrain, where there
is no distinction between a basic and metaphorical meaning, are also not iden-
tified as metaphorical expressions by the Pragglejaz procedure. In my method
of identification, metaphorical words with historical derivations are considered
metaphorical if they belong to a conceptual metaphor that is still active. Com-
pound words are considered metaphorical when the individual elements are
used in a figurative sense.
As the metaphorical expressions in the textbooks were collected, the source
and target domains were noted. An electronic search was then made through all
of the books, using words belonging to the source domain in question as a search
string. The expressions for the test were chosen in a pragmatic manner. Namely,
I wanted a variety of what I considered to be normal expressions that were found
more than once and in more than one book. Special attention was however given
to collecting expressions from two conceptual metaphors. One of them, under-
standing is seeing, is both a common and much discussed conceptual metaphor
in the metaphorical literature with many realizations in Norwegian. The other, to
agree is to be co-located, turned up with several realizations in the first book
studied. Both of these metaphors provided the basis for seven items each in the
final version of the test.
7.2 Task
A multiple choice task was used to tap the students’ understanding of the meta-
phorical expressions. The metaphorical expressions were presented in phrases in
order to contextualize them (see below). Then an explanation, an example, or a
description of the particular expression was proposed, along with three distract-
ers and the students were asked to mark the answer they found appropriate. They
were given 45 minutes to complete the task.
The words that constituted the metaphorical expressions were also tested with
their literal meanings. The format was the same as with the metaphorical expres-
sions. Examples of the tasks are (translated into English):
(1) A medal is
a. a kind of award
b. a kind of stamp
c. a piece of cloth
d. a little stone
(2) Light helps us
a. to see the colour of things
b. to listen if the radio is on
c. to smell if the food is burnt
d. to taste if the food is good
(3) When we talk about the back of the medal [colloquial Eng: the flip side
of the coin], we mean that
a. a lot of people get medals without deserving them
b. more people should get medals
c. something negative might go along with something positive
d. it is difficult to be just
(4) That something comes to light means that

a. we put it in the sun
b. we hear about it
c. we think that it is nice
d. we believe in it
An original task with 111 items was given as a pretest to three girls of the same
age. The final version of 50 metaphorical plus 31 literal items was based on the
girls’ results and comments. Three sample sets of questions and answers were pre-
sented by the teacher, followed first by the items which focused on a word with a
literal meaning and then by the items focusing on a metaphorical expression.
46 Anne Golden
8. Findings
As expected there was an overall difference between the students with Norwegian
as their mother tongue and the students with Norwegian as their second language
both in the task with the literal meanings and particularly in the task with the
metaphorical expressions.
The majority students achieved a very high score on the 31 items with a literal
meaning (mean score 96.9 %) indicating that these words are part of the vocabu-
lary of 15-year-old native speakers of Norwegian. The minority students also had
a high score on this part of the test (mean score 85.6%), but there was a much
larger variation in this group with scores ranging from 29% to 100% correct. Still
it means that most of the minority students were familiar with this vocabulary.
As for the 50 items with a metaphorical expression, the group of majority
students chose the correct alternative in 88.4% of the items, while the minority
students selected the correct answer in barely 67% of the items (see Table 1).
Table 1. Results of the multiple choice task: Different subgroups of students
Student Subgroups Group Number Score
selection based on of mean s.d. mean
students %
Total Linguistic majority students 230 44.2 5.1 88.4
number of Linguistic minority students 170 33.4 9.9 66.9
students
Linguistic Self- Norwegian best 70 37.3 7.8 74.6
minority evaluation Norw. & Mother tongue equal 44 32.3 8.6 64.5
students Mother tongue best 56 29.5 11.5 59.1
Linguistic Score on High vocabulary score 76 39.6 6.2 78.2
minority items with Medium vocabulary score 44 34.4 6.3 68.9
students literal
Low vocabulary score 50 23.2 8.9 46.4
meaning
Linguistic Language Mention Norwegian 148 33.7 9.8 67.4
minority practice Not mention Norwegian
22 31.3 10.3 62.6
students at home
Linguistic Start in Norwegian school in 1 grade 136 34.6 9.3 69.2
minority Norwegian Norwegian school later 34 28.9 10.9 57.8
students school
Part of the Mother Mother tongue: Urdu/Panjabi 61 32.7 8.7 65.5
minority tongue Mother tongue: Turkish 16 31.8 11.4 63.6
students Mother tongue: Vietnamese 12 32.3 10.4 64.7
When the group of minority students is divided into subgroups according to

which language they considered their best, the group that claimed Norwegian as
their best language achieved the highest score (just below 75%). The group that
claimed equal fluency in Norwegian and their mother tongue, achieved the next
highest score (64.5%), and the group that considered their mother tongue to be
their best language achieved the lowest score (slightly more than 59%). In other
words, the students’ self-evaluation of their best language correlated with their
mastery of the language.4
The subgroups defined by the relative size of their vocabulary (as defined
above5) differed even more in their scores than the subgroups defined by their
self-evaluation of best language. The group with presumably the most prolific
vocabulary attained the highest score on the metaphorical items, around 78%
correct, the intermediate group achieved a score of around 69% correct, and the
group with the lowest score attained as low as slightly more than 46% correct. This
indicates a strong correlation between the size of the student’s receptive vocabu-
lary and the comprehension of metaphorical expressions.
The subgroups defined by the students’ mentioning of Norwegian as part
of their home language environment also show a difference in score. The stu-
dents that listed Norwegian as part of their language environment at home scored
slightly more than 67%, while those that did not list Norwegian in this regard
(only slightly more than 12% of the minority students) scored slightly less than
63%, but this difference is not significant.
The subgroups defined by the students’ entrance into the Norwegian school
system show a difference in score as well. The students that started in first grade
scored slightly less than 70%, those that started later (20% of the minority stu-
dents) scored slightly less than 58%. This difference is significant.6
There were no significant differences between the three biggest minority
groups; the Pakistani, the Turkish, and the Vietnamese.
The participants who had Norwegian as their mother tongue scored signifi-
cantly higher on the multiple choice task than those students who had Norwegian
as their second language. This is not surprising at all. What is surprising is than
none of the subgroups of the minority students, not even those who considered
Norwegian to be their best language, reached the average of the native speakers.
If the students’ self-evaluation mirrors their actual competence in their mother
tongue as well as seems to be the case in their second language where they scored
rather low, then the language situation of the minority students is serious. It means
that their competence in both languages is low. So how do they learn?
The students’ self-evaluation as a group correlated with their mastery of the
language. But what does this mean? What part of their language skills do students
48 Anne Golden
think about when they consider one of their languages better than the other?
When the same multiple-choice test was carried out later with 79 students (61
minority students and 18 majority students) in the first year of higher elementary
school, this question was extended (Golden & Larsen, 2005). The minority stu-
dents were asked to evaluate their proficiency in the four areas of listening, speak-
ing, reading and writing in both Norwegian and their mother tongue, by giving
themselves points from one to six. The answers showed that it was their oral pro-
ficiency that seemed to influence their overall evaluation of their best language.
The students that considered their mother tongue to be their best language gave
themselves higher points in both listening and speaking in their mother tongue,
but not in reading and writing. An assumption is that this also was also the case
for the students in the tenth grade as well.
As expected, the results obtained by the students that had entered the Norwe-
gian school system in first grade scored higher than those that had started later.
This underpins the view that learning a second language takes time and that the ac-
quisition of a suitable vocabulary takes much more time than most people realize.
When the minority students chose a wrong alternative, then their incorrect
choice frequently consisted of a word with a form which resembled a word in the
introductory phrase. This agrees with Gimbel’s previously mentioned study on
vocabulary comprehension.
8.1 Relation between metaphorical and literal lexical items
As mentioned above, the 31 words included in the task testing the students’ com-
prehension of the literal meanings were derived from the words in the meta-
phorical expressions. Even if there was a positive correlation between the scores
on the literal test and the metaphorical test for both the majority students and
the minority students7, the study did not reveal a straightforward link between
the individuals’ understanding of the literal and the metaphorical understanding
of an item. This goes along with Gibbs’ view of metaphorical processing as direct
and not through an initial literal interpretation. Gibbs and colleagues performed
several experiments which show that the processing of metaphorical expressions
does not take more time than the processing of words and expressions used in
a literal sense (e.g. Gibbs, 1997; Gibbs & O’Brian, 1990). For the native students
there was no such link at all. For the minority students there were a certain cor-
respondence between the literal and the metaphorical comprehension of some
of the lexical items. For about half of the items tested both in the literal and the
metaphorical part of the task,8 significantly more students chose a wrong alter-
native on the item with a metaphorical expression when they had also chosen
a wrong answer on the same item used with the literal meaning. For the other
half, the difference was either insignificant or there were more students that un-
derstood the metaphorical expression without understanding the words in their
literal sense.
9. Variability in expressions
9.1 Types of metaphorical expressions in the task
As already mentioned, the type of metaphorical expressions used is one important

variable in testing comprehension. The 50 metaphorical items in my study were
categorized into sets and assigned a value according to certain factors. The theo-
retical basis for the categorization into sets is derived from other studies in second
language acquisition and reading comprehension as well as studies in cognitive
linguistics, i.e. from research concerning factors influencing comprehension. In
the introductory phrases that provided context for the expressions, three factors
were assumed to influence the comprehension task: the structural complexity of
the phrase, its main theme, and the characters involved in the settings. In the meta-
phorical items themselves the following variables were isolated: length of the ex-
pressions, vocabulary frequency at different levels, word class and imagery. In this
article, I will present four of these variables that seem to be the most influential:
the theme in context, the length of the metaphorical item, the general frequency of
the item, and the imagery. A fifth factor is the presence or absence of an equivalent
expression in the students’ mother tongue. This was investigated for three language
groups, the Urdu/Punjabi, the Turkish and the Vietnamese group.
9.2 Theme
The 50 metaphorical items in the task were assigned a certain value according to
the variable of theme, meaning the theme evident in the context of the metaphori-
cal expressions. The value was either ‘young’, ‘adultish’ or ‘neutral’. The values were
assigned after an analysis of the actual topics in the context of the selected items.
Topics like school, relations, and games were given the value ‘young’ and topics
such as work, economy, and politics were given the value ‘adultish’. Some items
were difficult to categorize and were assigned a ‘neutral’ value and discarded from
the comparison. Hence a prototypical approach to the categorization was applied
along with the Conceptual Metaphor Theory, and awareness of the fuzzy edges
of the categories was maintained. The comparison is thus between the clearest
50 Anne Golden
e xamples in each category, with the ambiguous or inappropriate examples exclud-

ed. Examples of items in the introductory phrase with the two values are (trans-
lated word-by-word into English):
a. Young context: To break out of a marriage means … (theme: ‘relations’)
b. Adultish context: If we say that the Labor party is in hard weather, it means…
(theme: ‘politics’)
Length of the metaphorical items

The metaphorical items tested consisted of three types:
a. Single words: To say that the co-habitation bursts means …
b. Verb + particles: To break out of the marriage means …
c. Multiword expressions: Peter being in fire and flame means…
Since particles (prepositions or adverbs) are short and non-salient, these items
were categorized with the single words, leaving two sets to be compared, the ‘short
set’ and the ‘long set’. The rationale for a categorization according to number of
words is the hypothesis that an expression consisting of several words would
cause a greater possibility of misunderstanding as it could evoke more informa-
tion which required interpretation.
9.3 Frequency
Vocabulary acquisition in both the first and second language depends on sev-
eral factors, one of these being frequency. The higher frequency an item has, the
more likely it is to be learned. The reason for this stems from the assumption that
an item becomes more entrenched in the memory each time it is heard. But the
frequency of the individuals’ input is of course impossible to collect. A corpus
consisting of texts which young people are likely to have read is unfortunately
not available in Norway. However, even if this corpus had existed, it would only
have given an indication of the frequency of the different words and expressions
that young people are actually exposed to. The corpus chosen to indicate the fre-
quencies of the metaphorical items was the Bokmålskorpus, one of the biggest
corpora in Norway (Bokmålskorpuset, 2004). This corpus consists of 16 million
words and is assembled from among three genres: novels, newspaper and maga-
zine articles, and factual prose. The frequency of the words and expressions with
a metaphorical meaning, the frequency with a literal meaning, and the frequency
of each word in the multiword expressions were collected. Three or four values
were assigned, depending on these frequencies. A confirmation of the values was
undertaken by using a frequency dictionary (Vestbøstad, 1989).
9.4 Imagery
Several researchers (e.g. Ellis & Beaton, 1993; Svanlund, 2001) have advocated the
advantage afforded by imagery in the interpretation procedure, but the practical
realization of this concept in an acceptable way is not straightforward. In order
to complement my own evaluation on the image strength of the items as ‘strong’
and ‘weak’, two groups of native speakers – students and teachers – was asked to
sort the items into these categories as well as a ‘don’t know’ category. This is not
an unusual procedure; see e.g. Ellis & Beaton (1993). Even though people differed
in their way of visualizing things, there was strong agreement on the items at each
end of the scale. The expressions for which there was the least consensus (or high
percentage of ‘don’t knows’) were assigned a neutral value and discarded from the
comparison. Examples of items with strong and weak imagery are:
a. Strong imagery: When the school gives the green light to something…
b. Weak imagery: When a teacher has a certain viewpoint, it means …
9.5 Equivalence in the mother tongue
The 50 items were translated into the three languages of Urdu, Turkish and Viet-
namese by people with both native competence in these languages and a back-
ground in language studies (interpreters, professors, language students). They
marked those items which they considered to have equivalents in their native
language, both in wording and in meaning. These translations were later given
to three other people with about the same qualifications and retranslated into
Norwegian. This procedure gave an opportunity to verify the categorization into
equivalent expressions and non-equivalent expressions. If the two people did not
agree or were unsure, the item in question was assigned the value ‘neutral’ and
discarded from the comparison.
10. Results from the sets
10.1 Theme
For both the majority group and the minority group of students, the average score
was much higher for expressions within a context that was assigned the value
‘young’ than for expressions that were assigned the value ‘adultish’: near 90% cor-
rect answers versus about 78% for the group of the majority students and about
82% correct answers versus 51% for the group of minority students (see Table 2).
52 Anne Golden
To verify this result, a statistical test – the paired sign test – was used, which simply
compared the pair of scores on the items with the ‘young’ and the ‘adultish’ value
for each individual and tested whether there were significantly more than 50% of
the students who attained a higher score on the former value. 86.5% of the students
in the majority group did indeed score higher on the expressions which where
classified as ‘young’ in theme. The group of minority students had an even more
unified score as 96.5% scored higher on these items. This is shown in Table 3. Both
results are statistically significant (p < 0.0005).
Table 2. Result of different sets: The group of majority students and the group
of minority students
Sets Result
Majority students n = 230 Minority Students n = 170
Correct answer % s.d Correct answer % s.d
Theme Young n = 16 95.8 7.4 81.9 19.7
Adultish n = 16 78.1 16.3 51.0 22.2
Length Short n = 27 91.7 8.9 73.0 19.4
Long n = 23 84.5 13.1 59.6 21.5
Imagery Strong n = 23 91.8 9.6 70.9 20.6
Weak n = 8 83.3 15.8 62.0 19.8
Table 3. Pair sign test of different sets: The group of majority students and the group
of minority students
Sets Results
Majority students n = 230 Minority students n = 170
Students scoring Students scoring
higher on the set higher on the set
n % n %
Theme Young n = 16 199 86.5 164 96.5
Adultish n = 16 31 13.5 6 3.5
Length Short n = 27 184 80.0 153 90.0
Long n = 23 46 20.0 17 10.0
Imagery Strong n = 23 155 67.4 126 74.1
Weak n = 8 75 32.6 44 25.9
10.2 Length of the metaphorical items
Both groups of students scored higher on the ‘short’ set than on the ‘long’ set (see
Table 2). The comparison of the scores on these sets at the individual level (the
paired sign test) shows that 80% of the majority students had a higher score on
the ‘short’ set. The results for the minority students were even more unified, as
90% scored higher on the ‘short’ set (see Table 3). The differences are statistically
significant (p < 0.0005).
10.3 Frequency
The students’ results for the sets based on the frequency of the metaphorical mul-
tiword expressions were as expected; the sets with the most frequent expressions
were also those which were usually correctly interpreted by both groups. The dif-
ference between the scores on the set with medium frequency and the set with the
low frequency was, however, very small. The sets with single-word metaphorical
items revealed a more surprising result. Examples of the values assigned to the
single-word items are:
Extra high frequency: To stand for an opinion.
High frequency: When somebody turns in a case. (Meaning ‘to change opinion’)
Medium frequency: When people step on each other.
Extra low frequency: The accident welded people together.
The set consisting of words with extra high frequency (like stå [Eng: stand]) did
not produce the highest correct score, but the lowest. It was the set consisting
of words with medium frequency (like the verb tråkke [Eng: trample]) that had
the highest rate of correct interpretation. This was the result both when the total
frequency of these verbs and the frequency with the metaphorical meanings were
taken into account. This result was common to both the majority and minority
groups of students (but there was, as expected, a big difference between the scores
in the two groups). The scores are presented in Table 4.
Table 4. Expected and actual order of sets defined by word frequency
(for both metaphorical use and total) among majority and minority students
Groups of students Frequency – metaphorical use Total frequency
Ex high High Med Low Ex high High Med Low
n=7 n=9 n=5 n=6 n=7 n = 10 n = 5 n=5
Expected order 1 2 3 4 1 2 3 4
Majority n = 230 84.3 92.7 97.4 93.8 84.3 93.2 97.7 92.7
Actual order 4 3 1 2 4 2 1 3
Minority n = 170 65.2 74.8 83.8 70.6 65.2 75.1 86.6 66.4
Actual order 4 2 1 3 4 2 1 3
54 Anne Golden
10.4 Imagery
The result was as expected, both the majority and the minority group scored
higher on the set with ‘strong’ image value (see Table 2). The comparison of the
scores with the paired sign test at the individual level gave the following results:
about 67% of the majority students and about 74% of the minority students had a
higher score on the ‘strong’ image set (see Table 3). The differences are statistically
significant (p < 0.0005).
10.5 Equivalence in the mother tongue
The Urdu group and the Vietnamese group scored higher on the items which
had a mother tongue equivalent, but not the Turkish group (see Table 5). Further
analysis indicated that the results for the Turkish group might have been caused
by inadequate knowledge of their mother tongues. There was a smaller percentage
that considered either their mother tongue alone to be their best language or con-
sidered themselves equally fluent in both their mother tongue and Norwegian.
More than half of the Turkish students considered Norwegian to be their best
language. The comparison of the scores on these sets by Urdu and the Vietnamese
students at the individual level (the paired sign test) shows that in both language
groups there were significantly more students (around two third of each groups)
that chose a correct answer on the items that had an equivalent expression in the
mother tongue (see Table 6).
Table 5. Result of the sets defined by the equivalence of the expressions in the students’
mother tongue: Three groups of students
Mother Urdu/Punjabi Turkish Vietnamese
tongues N = 61 N = 16 N = 12
Num Correct St.d Num Correct St.d Num Correct St.d
item answers item answers item answers
% % %
Equiva- Yes 13 69.6 17.2 20 58.1 22.4 13 68.6 18.9
lences No 14 60.4 21.1 18 67.4 23.5 15 61.7 19.3
Table 6. Pair sign test of the sets defined by the equivalence of the expressions in the
students’ mother tongue: Two groups of students
Results
Mother tongues Urdu/Punjabi Vietnamese
N = 61 N = 12
Students scoring Students scoring
Number higher on the set Number higher on the set
Sets
of items of items
N % N %
Equivalency in the Yes 13 41 67.2 13 8 66.7
mother tongue No 14 20 32.7 15 4 33.3
11. Summary – The variability of the expressions
The context in which the metaphorical expressions were embedded was important.
When the topic was of a kind associated with young people, the items were easier
to grasp than when the topic concerned an adult theme. The explanation might
be related to the students’ familiarity with these themes, i.e. they are part of their
real-life discourse. Hence the students’ background knowledge makes it easier to
guess the meaning correctly even if they were unfamiliar with the vocabulary. The
results also showed that metaphorical expression with words of medium frequency
seemed easiest, while expressions with words of extra high frequency seemed most
difficult. This is not as surprising as it may first appear. Most of the high frequency
words were verbs, all strongly polysemous with a meaning that is vague and hard to
capture. Even if second language students use these nuclear verbs more often than
native speakers (cf. Viberg’s studies), it does not mean that they also understand
the different meanings of these verbs. The fact that single words are easier than
multiword expressions corresponds with my hypothesis, namely that longer expres-
sions give rise to more associations that need sorting out. An explanation as to why
expressions with strong imagery are easier than weak imagery might be that there
is a correlation between strong imagery and transparency. At least the expressions
with weak imagery seemed to be among the least transparent. The last result, that
expressions resembling the wording in expressions in the mother tongue are easier
than those without a resemblance, is in line with the cognitive approach to language
learning – it is easier to understand things that are similar. That this was the case for
the Pakistani and the Vietnamese students only, is probably due to a difference in
the Turkish group’s relative mastery of their mother tongue as compared to the two
other groups. In other words, there were fewer Turkish students that reported their
mother tongue to be their best language.
56 Anne Golden
12. Discussion
The metaphorical expressions were presented to the students in a multiple choice

task. This method has the advantage of being time effective; hence it is feasible to
test a great number of items within the available time. It is also less demanding on
the students as they do not have to produce language and their potential produc-
tion difficulties will not interfere. There are however several disadvantages. The
main one is that the distracters are often hard to produce in a satisfactory way, re-
sulting in some distracters possibly being too simple with others being potentially
too difficult. But this problem should not have affected the difference in scores
between the various groups and subgroups because all students took the same
test. Any difficulty caused by poorly-worded distracters would thus have been
common for all groups. Another objection is that multiple choice tasks measure
results on a nominal scale only, tapping an either/or comprehension which can-
not reveal a partial understanding of the expressions.
In my study the age of the participants was the same, but their languages and
to a certain extent their cultural backgrounds were different. But because all of
the students had attended the Norwegian school system for several years, many
of them for nine years or more, they were familiar with the Norwegian school
culture and everyday school tasks. Multiple choice tasks to test vocabulary acqui-
sition are frequently used in Norwegian schools.
The hunt for authenticity – i.e. metaphorical expressions that students meet
in real-life textbooks – resulted in finding a variety of expressions in different con-
texts, which were thus in need of categorization. This revealed a difference in the
comprehension of certain sets that was not only due to the metaphorical expres-
sion in question, but also to other factors, like the contextualization of the lexical
units. Some of the sets did, however, contain only a small number of items, which
is a methodical weakness that ought to be followed up by further research which
investigates these variables. But this study clearly supports the view that testing
of metaphorical comprehension is a complex enterprise and one must take this
diversity into consideration when selecting metaphorical expressions for testing.
13. General results and remedies
The problem which minority students experience in comprehending meta-

phorical expressions found in textbooks most probably presents an obstacle to
their understanding of the overall content of those textbooks, thus contributing
to a poorer school performance. This is a problem that has to be treated seri-
ously. Some researchers focusing on language learning research in general have
advocated teaching the students various vocabulary acquisition strategies. Such

strategies are particularly useful for low-proficiency students, who tend to de-
pend on formal similarity (orthographic/ phonological) instead of consider the
immediate context or overall theme when guessing the meaning of an unknown
word (see e.g. Haastrup, 1988). This approach stems from the view that students
will benefit from consciousness-raising activities that serve to raise their meta-
linguistic awareness. My proposal is therefore to give students an introduction
into metaphor theory as presented in the Conceptual Metaphor Theory in order
to provide them with the awareness needed to experiment with patterns that exist
in the vocabulary. Several recent trials where students were made aware of Con-
ceptual Metaphor Theory have yielded positive results for students’ understand-
ing of metaphorical expression (see e.g. Boers, 2000, 2004; Li, 2002).
The categorization of the metaphorical expressions according to their im-
age strength carried out by the native-speaking students in this study revealed
a difference in the ability of the various expressions to trigger an image. These
categorizations were performed in a few classes at the University of Oslo, before
the Conceptual Metaphor Theory was presented to the students. Later on in the
same course when discussing these expressions again, the students were able to
give an explanation of the mapping behind even some of the expressions in the
weak imagery set, such as to have a certain viewpoint, connecting it to the con-
ceptual metaphor understanding is seeing. This indicates that an awareness of
the structure of metaphors influences the ability to form an image. Such an image,
according to the results of this study, yields the correct interpretation.
Such consciousness-raising experiments could easily be expanded to include
comparisons between the realization of a conceptual metaphor in the students’
mother tongue and in the language being learned. This would allow the students
themselves to become the experts, as they would have to furnish the teacher and
their classmates with examples from their mother tongue. In my study, the Nor-
wegian metaphorical expressions that had an equivalent in the students’ mother
tongue seemed to be easier to understand for the students that considered their
proficiency in their mother tongue either equal to or better than their proficiency
in Norwegian. Making the students aware that languages often utilize the same
conceptual metaphor even though the actual metaphorical expressions may dif-
fer (see e.g. Kövecses, 2005) would contribute to students’ experience of pat-
tern and similarity, thereby most likely enhancing their perception and learning
power. The analysis of the seven metaphorical expressions belonging to each of
the two conceptual metaphors understanding is seeing and to agree is to be
co-located did not reveal that the students had discovered the underlying met-
aphor by themselves. In both these metaphors, there are expressions that seemed
to be rather easy and others that seemed to be rather difficult. A further analysis
58 Anne Golden
of the number of mappings that had to be done did not give any indication of
the degree of difficulty of the expressions either. An additional effect of using
the minority students’ mother tongue in the school setting would also boost the
perceived status of other languages, as the tendency – at least in the Norwegian
educational discourse – is to consider minority languages as a hindrance rather
than a resource.
My proposal complements Graves’s (2000) recommendations for word con-
sciousness to be included as a component of a vocabulary curriculum. The vocab-
ulary should, however, be extended to include not only metaphorical expressions
but also metaphorical thinking (Littlemore & Low, 2006), so that students become
aware of the links between metaphorical expressions and their underlying con-
ceptual metaphors. In addition, I support Nagy’s (2007) call for more research on
the connection between rich vocabulary instruction and text comprehension. The
concept of rich vocabulary instruction must include metaphorical expressions of
different types. And – according to the results from this study – the teacher and
the teaching material should be aware of the importance of the contextualization
of the metaphorical expressions. Some themes favor the students’ interpretation,
others do not. The exercises and activities in class should take this into account.
Notes
1. The vocabulary was root-lemmatized, i.e. the comparison was done at word family level.
Word (instead of root lemma or word family) is used here as a shorthand.
2. In this comparison, the very low-frequency root lemmas were excluded. The frequency
limit for excluding root lemmas varied from frequency 2 up to frequency 6 depending on the
size of the text.
3. The study is part of my Dr. Philos. dissertation (Golden, 2005).
4. The difference is significant between the first and second group, but not between the second
and third.
5. The groups do not have exactly the same size, as several students achieved the same score
and the delimitations were set between students’ scores.
6. p < 0.0005.
7. r = 0.543 for the majority students and r = 0.718 for the minority students, according to
Spearman’s test.
8. Words very frequently used in everyday language in school were not tested in their literal
sense. This is the reason for the unequal number of items in the literal and metaphorical part
of the test.
References
Anderson, Richard C. & Peter Freebody (1985). Vocabulary knowledge. In H. Singer &
R. B. Rudell (Eds.), Theoretical models and processes of reading (343–371). Newark, DE:
International Reading Association.
Arnberg, Leonore & Åke Viberg (1991). Utvärdering av skolförberedelsesgrupper i Rinkeby.
Sammenfattning. Stockholm University: Center for Research on Bilingualism.
Beck, Isabel, Margaret McKeown & Linda Kucan (2002). Bringing words to life: Robust
vocabulary instruction. New York: Guilford.
Boers, Frank (2000). Metaphor awareness and vocabulary retention. Applied Linguistics, 21,
553–575.
Boers, Frank (2004). Cross-cultural variation as a variable in comprending and remembering
figurative idioms. European Journal of English Studies, 8, 375–388.
Bokmålskorpuset (2004). Data retrieved Jan 3, 2004, from http://www.tekstlab.uio.no/norsk/
bokmaal/
Cameron, Lynne (2003). Metaphors in educational discourse. London: Continuum.
Ellis, Nick C. & Alan Beaton (1993). Psycholinguistic determinants of foreign language vocabu-
lary learning. Language Learning, 43, 559–617.
Evans, Mary & Dianna Gamble (1988). Attribute saliency and metaphor interpretation in
school-age children. Journal of Child Language, 15, 435–449.
Gentner, Dedre (1977). Children’s performance on a spatial analogical task. Child Development,
48, 1034–1039.
Gibbs, Ray W. (1994). The poetics of mind: Figurative thought, language and understanding.
Cambridge: Cambridge University Press.
Gibbs, Ray W. (1997). Taking metaphor out of our heads and putting it into the cultural world.
In R. Gibbs & G. Steen (Eds.), Metaphor in cognitive linguistics (125–144). Amsterdam &
Philadelphia: John Benjamins.
Gibbs Ray W. & Jennifer E. O’Brien (1990). Idioms and mental imagery: The metaphorical
motivation for idiomatic meaning. Cognition, 36, 35–68.
Gimbel, Jørgen (1995). Bakker og udale. Sprogforum. Tidsskrift for sprog- og kulturpædagogik,
3. København: Danmarks Pædagogiske Bibliotek.
Gimbel, Jørgen (1998). Tyrkiske børns fagrelevante danske ordforråd i femte klasse. In J. Møller,
P. Quist, A. Holmen & J. N. Jørgensen (Eds.), Tosproget udvikling. Københavnerstudier i to-
sprogethed. Køgeserien K4 (95–111). Danmarks Lærerhøjskole.
Golden, Anne (1984). Fagord og andre ord i o-fagsbøker for grunnskolen. In A. Hvenekilde &
E. Ryen (Eds.), Kan jeg få ordene dine, lærer (170–175). Oslo: LNU/Cappelen.
Golden, Anne (2005). Å gripe poenget. Forståelse av metaforiske uttrykk fra lærebøker i samfunns-
kunnskap hos minoritetselever i ungdomsskolen. Oslo: UniPub Acta Humanoria, nr 227.
Golden, Anne (2009). Ordforråd, ordbruk og ordlæring i et andrespråksperspektiv. Oslo:
Gyldendal Akademisk.
Golden, Anne & Vibeke Larsen (2005). Egenvurdering av språkferdigheter og metaforforståelse
blant minoritetselever i videregående skole. NOA, 1–2.
Grady, Joseph (1997). Foundations of meaning: Primary metaphors and primary scenes. Ph.D.
dissertation, University of California, Berkeley.
60 Anne Golden
Graves, Michael F. (2000). A vocabulary program to complement and bolster a middle-grade

comprehension program. In B. Taylor, M. Graves, & P. van den Broek (Eds.), Reading for
meaning: Fostering comprehension in the middle grades (116–135). Newark: International
Reading Association.
Haastrup, Kirsten (1991). Lexical inferencing procedures or talking about words. Language De-
velopment Series. Tübingen: Gunter Narr Verlag.
Hene, Birgitta (2004). Adjektivs metaforiska betydelser – utlandsadopterade och svenska barns
tolkningar. In K. Hyltenstam & I. Lindberg (Eds.), Svenska som andraspråk – i forskning,
undervisning och samhälle (277–296). Lund: Studentlitteratur.
Huckin, Thomas N., Margot Haynes & James Coady (Eds.) (1993). Second language reading and
vocabulary learning. Norwood: Ablex Pub. Corp.
Keil, Frank C. (1986). On the emergence of semantic and conceptual distinctions. Journal of
Experimental Psychology: General, 112, 357–389.
Kövecses, Zoltán (2005). Metaphor in culture: Universality and variation. Cambridge: Cam-
bridge University Press.
Kulbrandstad, Lise (1998). Lesing på et andrespråk. En studie av fire innvandrerungdommers
lesing av læreboktekster på norsk. Oslo: Acta Humaniora, Universitetsforlaget.
Kulbrandstad, Lise (2003). Lesing i utvikling: teoretiske og didaktisk perspektiver. Fagbokforla-
get. LNUs skriftserie, nr 153.
Lakoff, George & Mark Johnson (1980). Metaphors we live by. Chicago: University of Chicago
Press.
Lakoff, George & Mark Johnson (1999). Philosophy in the Flesh. New York: Basic Books.
Lakoff, George & Mark Turner (1989). More Than Cool Reason. Chicago: The University of
Chicago Press.
Littlemore, Jeannette & Graham Low (2006). Figurative thinking and figurative language learn-
ing. Palgrave Macmillian.
Li, Fuyin T. (2002). The acquisition of metaphorical expressions, idioms, and proverbs by Chinese
learners of English: A conceptual metaphor and image schema based approach. Ph.D. dis-
sertation, The Chinese University of Hong Kong.
Nagy, Willliam (2007). Metalinguistic awareness and the vocabulary development and reading
comprehension. In R. K. Wagner, A. E. Muse & K. R. Tannenbaum (Eds.), Vocabulary ac-
quisition: Implications for reading comprehension (52–77). New York: The Guifford Press.
Oakhill, Jane & Alan Garnham (1988). Becoming a skilled reader. Oxford: Basil Blackwell.
Özçalışkan, Şeyda (2002). Metaphors we move by: A cross linguistic-developmental analysis of
metaphorical motion events in English and Turkish. Ph.D. dissertation, University of Cali-
fornia, Berkeley.
Pearson, P. David, Elfrieda H. Hiebert & Michael L. Kamil (2007). Vocabulary assessment:
What we know and what we need to learn. Reading Research Quarterly, 42, 282–296.
Steen, Gerard (2005). What counts as a metaphorically used word? The Pragglejaz experience.
In S. Coulson & B. Lewandowska (Eds.), The literal-nonliteral distinction (299–322). Ber-
lin: Peter Lang.
Svanlund, Jan (2001). Metaforen som convention. Graden av bildlighet i svenskans vikt- och
tyngdmetaforer. Acta Universitatis Stockholmiensis Stockholm Studies in Scandinavian
Philology nr 23. Stockholm: Almqvist & Wiksell International.
Sökmen, Anita J. (1997). Current trends in teaching second language vocabulary. In N. Schmitt
& M. McCarthy (Eds.), Vocabulary: description, acquisition and pedagogy (237–257). Cam-
Urquhart, Sandy & Cyril Weir (1998). Reading in a second language: Process, product and prac-
tice. London: Longman.
Vestbøstad, Per (1989). Nynorsk frekvensordbok: Dei vanlegaste orda i skriftleg norsk. Bergen:
Alma Mater.
Viberg, Åke (1993). Crosslinguistic perspectives on lexical organization and lexical progres-
sion. In K. Hyltenstam & Å. Viberg (Eds.), Progression and regression in language (245–66).
Viberg, Åke (1998). Crosslinguistic perspectives on lexical acquisition: the case of language-
specific semantic differentiation. In K. Haastrup & Å. Viberg (Eds.), Perspectives on lexical
acquisition in a second language (343–376). Odense: Odense University Press.
Viberg, Åke (1999). Polysemy and differentiation in the lexicon. In J. Allwood & P. Gärdenfors
(Eds.), Cognitive semantics. Meaning and cognition. John Benjamins Publishing Company.
Viberg, Åke (2002). Polysemy and disambiguation cues across languages. The case of Swedish
få and English get. In B. Altenberg & S. Granger (Eds.), Lexis in contrast (119–150). Am-
sterdam & Philadephia: John Benjamins.
Viberg, Åke (2005). The lexical typological profile of Swedish mental verbs. Languages in Con-
trast, 5, 121–157
Vosniadou, Stella (1987). Children and metaphors. Child Development, 58, 870–885.
Vosniadou, Stella, Andrew Ortony, Ralph E. Reynolds, & Paul T. Wilson (1984). Sources of dif-
ficulty in the young child’s understanding of metaphorical language. Child Development,
55, 1588–1606.
Winner, Ellen, Anne K. Rosenstiel, & Howard Gardner (1976). The development of metaphoric
understanding. Developmental Psychology, 12, 289–297.
Waggoner, John E. & David S. Palermo (1989). Betty is a bouncing bubbles: Children’s compre-
hension of emotion-descriptive metaphors. Developmental Psychology, 25, 152–163.
chapter 4
“Drugs, traffic, and many other dirty interests”

Metaphor and the language learner
Gill Philip
Università di Bologna
Existing empirical research into the role of metaphor in the foreign language
learning process focuses primarily on comprehension and recall. Yet stu-
dents’ ability to produce conventional metaphor in their speech and writing is
considered one of the measures of advanced proficiency in a foreign language.
While Danesi (1994) argues that “conceptual fluency” is fundamental if stu-
dents are to achieve naturalness in their language production, Charteris-Black
(2002) and others stress that conceptual knowledge does not necessarily lead
to the production of acceptable linguistic forms. There is a gap to be bridged
between learning the concepts and learning how they are realized linguisti-
cally. In this chapter, figurative language produced by advanced learners of
English is examined with reference to general language corpora, both for the
students’ mother tongue, Italian, and their foreign language, English. This
mode of investigation makes it possible to identify when unusual phraseol-
ogy can be ascribed to language transfer alone, and when other factors appear
to be involved. The data presented in this chapter illustrate how conceptual
knowledge formed in the mother tongue can interfere with the acquisition of
foreign language conceptualisations, and highlight the importance of phrase-
ology in fixing conceptual meaning.
Keywords: collocation, delexicalisation, figurative language, language learning,

phraseology
1. Introduction
What kinds of figurative language do learners produce in their discursive writing?

Despite its importance for language pedagogy and lexicography, as well as for lin-
guistics in general, this question has not been adequately addressed in the existing
literature. Research into learners’ language production tends to focus more on
64 Gill Philip
‘normal’ aspects of the language – organizational markers, collocation errors, and

terminological mismatches – with errors of a more abstract, conceptual nature all
too often relegated to the rag-bag category of ‘language interference’.
While it would be futile to contest the existence of language interference, the
term itself is somewhat abused. It is all too easy to explain away learner-produced
oddities by stating that they are caused by the influence of patterns from another
language, especially when the oddity itself seems to evade definition in terms of
grammar or conventional syntax. The vague explanations that often accompany
the indication of such an error, “it doesn’t sound quite right”, “we wouldn’t say it
like that”, and so on, do little to illuminate the matter. Having failed to identify
the cause of the error, the instructor cannot offer students advice on how to avoid
making similar mistakes in the future, and the popular notion that language mas-
tery is acquired, not learned, is reinforced.
Yet there must be some basis underlying the identification of a linguistic pe-
culiarity, even if that reason proves difficult to pinpoint. In this chapter, I consider
the problem in terms of conceptual mismatches between L1 and L2, and describe
its workings using examples from assignments produced by advanced learners of
English at an Italian university. Gibbs (this volume) stresses the need to base meta-
phor interpretation on empirical data. In this chapter, extensive use of corpora is
made: in the first case, the examples presented and discussed are drawn from a cor-
pus of around 80,000 words which I compiled from my advanced (C1) students’
homework assignments between 2003 and 2005. This corpus contains a range of
text types and tasks, unlike the majority of learner corpora which are primarily
composed of assessed essays. The anomalous uses that emerge from the students’
writing are then compared against concordance and collocation data from general
reference corpora for the target language (English) and also for the students’ L1
(Italian) so that patterns attributable to language transfer can be ascertained.1 On
the basis of the evidence provided by these corpora, I argue here that our concep-
tual knowledge of a word or expression’s meaning range is forged from the sum of
the conventional collocational and phraseological patternings of that word or ex-
pression in the L1, and that it is inadequate knowledge of the word’s phraseological
behaviour in the L2, rather than incomplete L2 conceptual knowledge, that results
in the production of the “it doesn’t sound right” type of interlanguage error.
2. Metaphor and language learning
Metaphor is occupying an increasingly prominent position in language teach-

ing and in pedagogical lexicography. Lakoff and Johnson’s (1980) Conceptual
Metaphor Theory is only now gaining ground in applied linguistics, as it filters
Chapter 4. Metaphor and the language learner 65
down through university studies into teacher training courses and pedagogical
resources. Abstraction is attractive to the language learner and teacher alike, as it
shifts the emphasis away from the nitty-gritty of word-perfect utterances towards
a more generalised impression of how the language communicates ideas, i.e. from
knowledge of the language to knowledge about the language.
Several existing studies attest to the utility of appealing to students’ concep-
tual awareness during the language learning process. In vocabulary acquisition in
particular, it seems that language items are more successfully learned when a spe-
cific focus is directed on the relation of figurative meanings to their correspond-
ing literal meaning (Boers, 2000; Charteris-Black, 2000); it has also been shown
that encouraging students to make use of their powers of visualisation (Boers &
Stengers, 2005; Stengers et al., 2005) aids the comprehension of new items in text
and also facilitates the recall of the same items in subsequent vocabulary tests.
Despite the success that raising students’ metaphorical awareness has had,
most researchers remain sanguine about the effects of such knowledge on lan-
guage production.
Knowledge of the conventional metaphoric themes of a given language does not
guarantee mastery of its conventional linguistic instantiations. As it is impossible
to predict exactly how a particular language will instantiate identified metaphoric
themes, learners cannot employ their awareness of those metaphoric themes to
“generate” figurative expressions in the target ... (Boers, 2000: 569)
Charteris-Black (2002) also notes that knowledge of the new language’s concep-
tual norms is of limited service for students wishing to produce native-like utter-
ances. As he states, “where linguistic forms are quite different, activation of an
equivalent first language conceptual basis does not always lead to the correct L2
linguistic form” (2000: 125). Charteris-Black (ibid.) repeatedly draws the reader’s
attention to the fact that although conceptualisations may be shared across lan-
guages, the precise linguistic instantiations related to the concept can differ con-
siderably. Ultimately, it is the linguistic form that carries the meaning.
The fundamental role of phraseology is also noted by Deignan et al. (1997),
who stress that “the exact words and phrases which express this conceptual link
in L2 cannot be guessed by reference to L1, so these need to be discussed and
learned” (1997: 354). In other words, the abstract knowledge which can be drawn
on successfully for decoding is insufficient for encoding purposes. Holme (2004)
too reminds us of the relatively arbitrary nature of a conceptual metaphor sche-
ma, being “a principle of meaning extension whose destination cannot always be
predicted” (2004: 97).
It therefore becomes apparent that there is an important relationship holding
between concepts and the conventional phraseology with which they are realised,
66 Gill Philip
but this interaction of form and meaning is often overlooked or downplayed.

Metaphorical schemata are generalisations, and as such are minimally concerned
with details. We have read that knowledge of the L2 conceptual frame of reference
does not seem to be enough to ensure the production of acceptable linguistic
renditions: there seems to be a gulf between drawing on a concept to aid com-
prehension, and encoding the concept in a satisfactory way. The claim that meta-
phor is “an important vocabulary-building skill for the language learner” (Lazar,
1996: 44), and that metaphor and metonymy are “hugely productive forces within
the lexicon” (Moon, 2004: 200) may be justified enough, but the ways in which
metaphor is exploited and exploitable require more detailed investigation.
It is easy to over-generalise the range of application that the concept actually
has when abstracting out from linguistic expressions to concept. Such over-gen-
eralisation is difficult to spot in a monolingual setting: counter-examples are no-
toriously difficult to invent, and the same can be said of unconventional phrase-
ology. Learner language however provides a wealth of evidence for the priority
of linguistic form over concept, because it illustrates how apparently innocuous
changes to conventional phraseology can result in a failure to transmit the in-
tended meaning. Philip (2005a) has shown how students’ expression of the con-
cept life as valuable commodity – common to the students’ L1 – is dependent
on particular phraseological renderings, and if these are altered, the result is only
partial transmission of meaning. As this breakdown in meaning can occur even
when the concept seems have been applied correctly and in the absence of gram-
matical or syntactical errors, it must be explained as a phraseological phenom-
enon caused by collocational incongruity.
A knowledge of how words typically combine (in collocations and conven-
tional phraseology) helps to shape the corresponding understanding of concepts.
Should the necessary linguistic knowledge be incomplete or inaccurate, so too
will be the understanding – and expression – of those concepts. This observation
runs contrary to the accepted view that concepts are drawn on in the creation of
new expressions. The reality is that word forms do not combine promiscuously.
While the generation of new expressions can be ascribed to conceptual force, the
precise forms that these expressions can take is entirely determined by norms of
linguistic usage, i.e. the accepted ways in which words combine with one another
into preferred phraseological patternings or “lexical networks” (Gibbs & Matlock,
1999). Viewed from this standpoint, it becomes apparent that encoding in the
L2 requires considerable knowledge of how concepts are lexicalised, rather than
knowledge or awareness of the concept alone. In fact, the greater the students’
repertoire of conventional collocations and phraseology, the more proficient they
appear to be in expressing concepts effectively. This can be contrasted with the
observations cited above, which point out that knowledge of the concept does
not lead to the production of appropriate linguistic forms. Perhaps when Danesi
speaks of “conceptual fluency” (1994: 454) he is picking up on this greater sensi-
tivity to native-speaker norms of phraseology in which form, meaning and gen-
eral conceptual trends are interwoven.
3. Encoding idiomatic meaning in the L2
Because Italian and English are quite closely related both linguistically and cultur-
ally, it comes as no surprise that the languages enjoy a similar outlook on the world
(shared conceptual schemata) and often express this in similar ways (shared lin-
guistic expressions). Cultural and lexical similarities make it relatively easy for an
Italian student to become reasonably proficient in English, as there is so much in
common. Yet the corollary of such linguistic and conceptual proximity is that stu-
dents often rely more on their powers of deduction and intuition than on explicit
learning. This is especially true once students move beyond simple, concrete con-
structions and start to use turns of phrase and more abstract language. It is all too
easy to fall into the trap of assuming that words correspond on a one-to-one basis,
to remain unaware that most words have more than one meaning (in lexicographi-
cal and translation terms), and to overlook the fact that the meaning of words in
combination may not correspond to the sum of those words’ individual meanings
(idiomaticity and phraseological meaning). This lack of language awareness can be
remedied for the L2 by making explicit reference to metaphor in teaching, as the
studies cited in section 2 have affirmed. But unless students are particularly sensi-
tive to the workings of their L1, they will tend to prefer familiar word combinations
in their L2 encoding. As a result their language production is often characterised
by anomalous collocations and – even worse – word-for-word renditions of idiom-
atic phrases such as those illustrated in Examples (1) and (2).2
Idiomatic language is not only notoriously difficult to decipher in the L2, but it
can also pose a problem in the L1 – L2 encoding process. Casting the very obvious
cases of idiom aside (those which violate truth conditions, such as raining cats and
dogs), it should never be forgotten that most language learners are not linguists
by profession, and as such they are less inclined to break down and categorise the
language they use. It is understandable that non-compositional expressions and
terminology can be considered ‘literal’ by non-experts, because the lay person’s
perception of what counts as ‘figurative’ is much closer to literary metaphor than
to the much more pervasive dead metaphor: as Gibbs & Matlock remind us, “ex-
perts’ intuitions often differ from those of ordinary individuals who have no pre-
conceived notions about the phenomenon of interest” (1999: 263). Examples (1)
and (2) constitute fairly typical instances of word-for-word calquing of Italian
68 Gill Philip
figures of speech into English. The highlighted phrase in Example (1) corresponds
to the Italian la fuga dei cervelli (‘the brain drain’); and the phrase in Example (2)
corresponds to provare sulla mia/propria pelle (‘to experience first-hand’).
(1) …the incredible “escape of the brains” and the difficulties in which the
scientific research is left.
(2) As I could experience (on my own skin), research in Italian universities is

very scarcely promoted.
Mistakes such as these are often put down to laziness on the part of the student, who
is probably aware that the phrase is not correct in English. Students at a lower level
of proficiency than those whose work is discussed in this chapter often leave direct
translations or even untranslated text in their compositions when they do not know
the equivalent and have not been able to find it (or simply have not bothered to look
it up). As far as these examples are concerned, it is safe to say that Example (1) is a
case of laziness, as the stimulus text used the term brain drain, and the student failed
to recognise and re-use it. Instead, he inadvertently created a humorous expression:
the sensation that the incredible “escape of the brains” sounds facetious or ironic is
confirmed by corpus data. The only modifier found before escape of the is luckiest
(the luckiest escape of my life occurs 4 times in BNC), and, in more general terms,
the string the escape of the appears to favour the company of wild and dangerous
things, a category to which brain (intelligent person) is not normally assigned. Us-
ing this word to complete the string flouts the expression’s normal combinatorial
preferences and creates a humorous effect (see Louw, 1997).
Example (2) is a different matter, however, as this particular use of pelle is not
listed in the large bilingual dictionary that the students use (Ragazzini, 1995), nor
is it listed in the same publisher’s corpus-based monolingual Italian dictionary
(Zingarelli, 2001), suggesting that it is not really thought of as having a different
sense to the established (in Italian) metaphorical ones of ‘life’ (experience) and
‘proximity/intimacy’. This being the case, the student would not have been able
to locate an appropriate translation even had she looked for one, so she fell back
on translation.
Examples (1) and (2) above are, thankfully, quite uncommon in advanced
learner writing. Students are more likely to shy away from phraseological turns
and figurative language than to attempt to recreate them in the L2 (Philip, 2005b).
They are unwilling to cause offence or unintentional humour, and avoid situations
that are liable to end up in a loss of face. These examples have been extracted from
coursework assignments which were not graded for assessment, so the students
have been less conservative than they might have been in an exam setting. By far
the most frequent type of interlanguage anomaly is caused by errors of colloca-
tion, and these are examined in the next section.
4. Collocation and conceptualisation
While collocation errors could be considered by some to provide evidence of in-

adequate conceptual knowledge in the L2, they can be comprehensively account-
ed for in linguistic terms alone. Conceptual knowledge does not come out of thin
air – it is created and sustained through linguistic forms. The study of unsuccess-
ful approximations of conventional linguistic forms makes it apparent that con-
ceptual mapping is selective and highly dependent on, and sensitive to, particular
lexical realisations. This section deals with a number of collocation anomalies
related to non-literal word senses in a bid to reveal their linguistic origin, and how
this might relate to Danesi’s (1994) notion of conceptual fluency.
4.1 The meaning of delexicalised words
Collocations are a headache for the language learner because they are word-form
specific and resist generalisation. While common noun-verb collocations are in-
troduced at a very early stage in the language learning process, the collocations
that vex advanced learners are those more relevant to academic and other dis-
cursive writing, particularly verb-adverb and noun-adjective collocations. These
often appear to be arbitrary because they differ in inexplicable ways from the
equivalent patternings in the L1. Furthermore, the fact that such word combina-
tions do not exert their full meaning potential (they are at least partially delexi-
calised) is often not perceived by learners, who tend to favour a compositional
interpretation of language.
Delexicalisation entails two principal aspects of meaning which act in tan-
dem. In the first case, delexicalised content words lose some of their salient
meaning, and function words lose some of their grammatical functional val-
ue. Secondly, their status as autonomous orthographic units (character strings
surrounded by white space) is weakened. As a result, the meaning that del-
exicalised words convey is created and bolstered by their co-occurrence with
habitual collocates: the words work together to create meanings which are not
necessarily present when the same words are being used compositionally. A
similar phenomenon can be observed in idioms, but with one crucial differ-
ence: the meaning of an idiom extends beyond the meanings of its component
parts, while the meaning of a delexical form is restricted and delimited by its
collocates (Philip, 2007).
The examples brought forward in this section are all instances of collocation
transfer, and illustrate how L1 delexicalised chunks are broken down and re-
formulated verbatim in the L2. An analysis of the errors and the L1 patterns that
have influenced them makes it clear that students stick to lexical combinations
70 Gill Philip
that are familiar to them; it is difficult to find any evidence which might support
there being a conceptually-driven approach to encoding.
(3) …you can meet people belonging to different cultures, nationalities, races;
you have the chance to enlarge your views.
(4) Even in Italy the Government is elaborating measures for the introduction
of e-learning in higher-education institutions in order to catch up with
standards in other countries.
(5) My nerves broke down and I went into a heavy depression.
Enlarge one’s views, elaborate measures and heavy depression are all fairly typical
collocation errors in Italian natives’ production of English. Enlarge occurs fre-
quently as a mistranslation of ‘broaden/widen’, because allargare is formally simi-
lar to large. Obviously with the two words referring to different kinds of space –
horizontal for allargare, but both horizontal and vertical for enlarge – the meaning
referent is anomalous for English. Elaborate (mistranslated from elaborare, ‘to
process [information]’) again appears to violate English conceptual norms, in that
elaborate is a synonym of embellish, not devise. Measures are introduced or taken,
but apparently not formulated so, as was the case for Example (3), both of the col-
locates are inappropriate for the context of use.
When the meaning of the offending collocate is figurative but has been trans-
lated by the equivalent normally reserved for the literal sense, this problem is
exacerbated. The student who produced Example (3) has used heavy to translate
pesante, a highly polysemous word whose senses extend well beyond the range of
its English equivalent. Heavy is fine as a translation for the literal sense, but is in-
appropriate for most of the figurative meanings; here the translation should read
deep depression. As it seems not to have occurred to the student that the meaning
of pesante in this example is not the same as the meaning of pesante in una valigia
pesante (‘a heavy suitcase’), he had no reason to double-check the meaning in
his dictionary. By ‘knowing’ that these words are translation equivalents, he has
simply transferred the patternings of pesante onto heavy. Yet even had he known
that English prefers to express emotions in terms of depth rather than weight, he
may have still produced an anomalous collocation, such as profound depression
(profound being close to the Italian equivalent, profondo).3 Collocations are not
compositional and therefore difficult to predict or second-guess.
4.2 Literal and figurative meanings in translating collocation
Failure to recognise different senses of a word, and the role and function of any
given word in a chunk of language, is a major problem in language acquisition for
all but the few students who are linguistic experts. Just as most users of a comput-
er have little idea of how the components are built and how they interact with one
another, for most people language is simply a tool for communicating with. Our
L1 forms our frame of reference for the world, and part of the pleasure of learning
a foreign language is discovering new ways of viewing the world as expressed in
and through the new language.
When learners set about acquiring an L2, they will find that some of their
existing L1 concepts are meaningless while others seem to have the same value,
so can be transferred successfully. Although inappropriate or irrelevant concepts
can be suppressed during the use of the L2, it is not so easy a matter to understand
how similar concepts match up in appropriate and/or conventional linguistic pat-
terns. We have already seen what happens when idiomatic phrases and colloca-
tions are translated verbatim from L1 to L2; but the problem is as relevant to fully
lexical language, especially when the literal/figurative boundary is reached.
Concepts are often perceived to be shared across languages and the existence
of direct (or nearly direct) equivalents reinforces the illusion of similarity. De-
spite these felicitous correspondences, no two languages are translations of each
other. However, the precise ways in which the L1 and L2 uses diverge can easily be
missed by learners, by their teachers (unless their command of both languages is
excellent), and also by lexicographers: monolingual lexicography is not concerned
with contrasting languages, and most bilingual lexicography is still structured on
native monolingual models. As a result of this under-emphasis on contrast, and
over-generalisations of collocational patternings, the documentation of fine levels
of distinction is usually inadequate to prevent interlanguage from seeping into
learners’ speech and writing.
One of the most difficult aspects of meaning for learners to grasp is that the
translation of a L1 word in its literal sense may not be an appropriate translation
for the same L1 word when used figuratively. For the non-expert language learner,
words are not split up into sub-senses unless they are homographs and thus quite
clearly ‘different words’. As a result, if a translation equivalent is known for the
literal meaning, it is likely that this L2 expression will serve as an all-purpose
equivalent for that L1 word. From this simplistic view of equivalence, it is easy to
overlook the fact that what is ostensibly the same string of characters may in fact
represent distinct meanings, with their own rules of syntactic patterning.
Example (6) illustrates what can happen when the differences in figurative
extensions of a common word have not been identified.
(6) If you live in a condominium conflicts and discords can be born with others.
The concept of birth as beginning is very closely related to the literal sense of
birth, and it is used in both English and Italian. At first glance, Example (6) looks
72 Gill Philip
like an attempt to be creative which results in a conceptual near miss, though, as

with so many errors, its apparent creativity stems from L1 norms: the student has
transferred the conventional collocational patterning of nascere, in a way which
is alien to English.
In Italian, nascita (birth) collocates with emozioni (emotions, feelings) dif-
ficoltà (difficulties), equivoci (misunderstandings), guai (trouble), problemi (prob-
lems) – the general category to which “conflicts and discords” belongs – over and
above the range that English expresses with birth. BNC data for English indicates
that the metaphorical sense of birth applies to nations, businesses, organisations,
political movements, social trends and academic disciplines, but nowhere is it
used for emotional or mental states. For this reason, then, the collocation of con-
flict and discord with born is inappropriate (i.e. atypical and hence anomalous).
It also violates English norms of usage and, by extension, the conceptual range
ascribed to birth. It needs to be stressed that this is a matter of collocation error,
and not one of conceptual incompatibility. The error is not caused by the particu-
lar conceptual ranges that birth or nascita have, but because when English speaks
of the creation of nations, businesses, organisations, political movements, social
trends and academic disciplines, birth is acceptable, yet it is not normally used to
describe emotions or troubles (troubles start, problems arise, and conflicts and
discord are caused/provoked). The overriding concept (beginnings) is basically
the same for all these expressions, but while Italian can use nascere to lexicalise all
these ideas, English chooses from a range of collocates depending on the entity
that is being mentioned.
Further cases of erroneous metaphorical transfer are found in Examples (7)
and (8). Again, Italian collocations are translated in ways which are not acceptable
nor particularly comprehensible in English.
(7) It might be better if we slacken our way of life and if we learn from the nature!
(8) Summing up, I prefer to live in a city like Bologna because of the many-sided
opportunities that I can find in it.
Once again, the errors presented here can be read from a conceptual standpoint
or a linguistic one. If we consider them as compositional choices which privilege
the salient meanings of slacken and many-sided respectively, then we have to try
to justify the choices in terms of what these words’ figurative meanings imply.
If, instead, we consider them as the direct translations of non-compositional or
formulaic expressions, then it is the meaning of the whole expression that is of
interest, and not the individual values of the components.
As a literal translation of allentare, slacken (Example 7) is the best choice, but
here the meaning is not literal (i.e. collocating with screw, knot, etc.), but rather
the figurative sense ‘to slow down or relax’. Does the student mean to unloosen in
its fully salient sense, which would trigger off metaphorical associations such as
life being tense and our feelings constrained by it; or does she mean ‘slow down
the pace’, ‘take it easy’, ‘relax’? My impression is that it is the second option, based
on the fact that students are on the whole very reluctant to create novel figurative
language (Philip, 2005b). Students’ reliance on L1 norms of phraseology seems to
be a way of avoiding the use of expressions whose unfamiliarity emphasises their
figurative nature. Learners view them as being more figurative than native speak-
ers do, and thus seek shelter in native forms which are familiar to them, even if
they are not conventional in the L2.
Example (8) is characterised by the same type of error: many-sided (Exam-
ple 8) is the literal translation of poliedrico, the adjective derived from poliedro
(‘polyhedron’). To his credit, the student has recognised that the meaning is fig-
urative, avoiding transliteration and opting for the translation provided for the
figurative sense ‘many and varied’. However, the choice of translation still reflects
the literal meaning of poliedrico (the correct translation in this context would be
‘[great] variety of ’). Once again we are forced to decide if he is trying to express a
particular mental image, or if he is using a familiar L1 expression in translation. It
cannot be ruled out that the student might have in mind an image of opportunity
as an object with many facets, like a diamond, but this is impossible to ascertain.
Linguistically, however, it can be verified that poliedrico collocates principally with
two recurring common nouns – figura (‘figure’), and attività (‘activity/ies’) – and
with proper names, in particular names of artists, musicians and other creative
people. This conventional use of the adjective in the student’s L1 contributes to
and reinforces the conception that a person or thing described as poliedrico is
characterised as having many aspects, faces or sides. This interaction between fa-
miliar language and familiar concept is one of the factors responsible for colloca-
tion errors such as many-sided opportunities. The student may feel that a different
adjective does not quite convey the right sense, and so prefers the L1 rendering
regardless of its lack of currency in the L2. The reason why the collocation sounds
strange is that many-sided collocates in English with questions, debates and prob-
lems, all of which have contrasting characteristics. Opportunities do not share this
element of contrast, and are therefore defined by number and quantity (many, a
lot of) rather than by their disparity. This student’s repertoire of conventional L1
collocations has contributed to his conceptual knowledge of what an opportunità
is, and he applies this semantic information to the L2 encoding process with-
out stopping to think that his conceptual knowledge is language-specific, and so
might not carry over to the L2.
It is here that we see how language and concepts interrelate. Concepts do
not exist independently of language, and contrary to the opinion that concepts
generate new linguistic metaphors, cross-linguistic comparisons demonstrate
74 Gill Philip
that conceptual schemas are linguistically determined and language-specific. As a

final illustration of mismatch between L1 and L2 concept/ conventional linguistic
expression, let us examine the example which also appears in the title to this chap-
ter: drugs, traffic and many other dirty interests.
(9) And in the end one of the biggest problems that affects big towns is the
criminality that frightens especially women and people in general. It’s a
plague that sometimes is connected to drugs, traffic and many other dirty
interests.
As well as referring to things that are physically soiled, dirty is used to describe
dishonesty, unfair dealings, negative evaluations of sex, and bad things in general.
These abstract categories are informed by collocates such as jokes, words, business,
and lies, but the precise ways in which these concepts are lexicalised in language
are far more specific than might be imagined. For example dirty business is a con-
ventional, non-compositional expression, but the apparently synonymous dirty
interests (Example 9) is not; in fact it is almost meaningless in English. Being a
compositional pairing, it is difficult to figure out which particular sense of dirty is
being alluded to, and by grouping together drugs (dirty = ‘illegal’), traffic (dirty =
‘polluted’) and interests (dirty = any of the established senses, including ‘sexually
deviant’, ‘illegal’, ‘morally questionable’), the resulting chunk reads as an opaque
metaphor, or a rather zany and imaginative zeugma. In contrast, the Italian ex-
pression which this student has translated, interessi sporchi, is a conventional col-
location which draws on the corruption sub-sense of sporco (‘dirty’). Whereas the
near-synonyms affari (‘business’) and interessi both collocate normally with spor-
co, it is interesting to note that although business can collocate with dirty, interests
cannot. Even if the student had already come across the expression dirty business
in text or in a dictionary, he would have had no means of discovering that the
collocation he produced should be unacceptable. If we abstract out from dirty’s
collocational patternings to the concept that it represents, there is no reason why
dirty interests should be unacceptable. It is simply an illustration of the fact that
“linguistic behaviour among users of a language is highly stereotypical, even in
matters of fine detail” (Hanks, 2004: 246). As was true of Examples (6)–(8), it is
doubtful whether any figurative meaning was intended, especially as the phrase
is introduced by a metaphor proper, a plague, which although conventional, still
exerts some degree of metaphorical life in both languages.
4.3 Discussion
The data presented in this chapter lends support to the claim that conceptual er-
rors in the L2 are ultimately caused by the inappropriate use of linguistic forms.
However, attributing all errors to language interference is somewhat simplistic
and not particularly illuminating. If we start from the premise that our conceptual
knowledge is built up from the sum of the linguistic expressions that we know,
then figurative language studies must examine the role of phraseology in consid-
erably greater detail. When figurative language is studied from a cross-linguistic
perspective, it becomes all too apparent that conceptual sets are only partially and
selectively exploited. The meanings conveyed by conceptually-related figurative
expressions are not governed by abstract thought, but by collocational tendencies,
and with the very precise and detailed phraseological patterns in which those col-
locates co-occur.
Conceptual knowledge in the L1 is an abstraction of the language patterns of
the L1. Proficient learners tend not to find decoding difficult because they have
already amassed a considerable store of conventional language forms in the L2 on
which they can draw. However, even if the recognition of form is relatively trou-
ble-free, the memorisation of new language items or new uses of familiar items
is rather more problematic, with conventional phraseological patterns seeming
to get distorted at some point between recognition and recall. This is most likely
due to the prioritising of salient meanings, with the result that these will be re-
called with greater ease than will their contextual, phraseological meanings. Yet
conventional expressions are not typified by salience but by delexicalisation and
idiomaticity, which function in close collaboration with regular phraseological
patternings. Even the smallest change to the established wording of a phrase can
interfere with the transmission of the intended meaning.
5. Encoding L1 concepts in the L2: The creation of opaque metaphor
While the phraseology of Example (9) saves it from total incomprehensibility (the
pattern ‘x, y and other z’ indicates a logical connection between dirty interests and
its collocates drugs and traffic), the final examples to be presented in this chapter
are not so fortunate. When L2 lexical and conceptual mapping is inadequate or
erroneous, the inevitable result is communicative failure. There are clearly differ-
ent gradations of incomprehensibility, and context can do much to ease the pas-
sage of information. When the collocation is at fault, as in previous examples, the
context helps the reader to pass over the error and select an interpretation based
on native norms, and this helps to neutralise the disjointedness that arises from
76 Gill Philip
the non-standard phraseology. Text is predictive, and textual meaning is partly

created by the reader’s expectation of what will come next. Sometimes, however,
text does not do what we expect it to, and if the language does not follow familiar
patterns, deciphering its meaning can be challenging. This was the case with Ex-
amples (1) and (2), which transported L1 idiomatic meaning word-for-word into
L2. A reader unfamiliar with the L1 patterns would recognise that the meaning
was idiomatic, but may not be able to understand what is meant.
(10) Recently, Britain’s young have been questioned about several issues so as
to try to inquire which their interests, expectations, ambitions are and how
they relate to society. Yet this attempt to define clearly these features has
ended up leaving us with the same puzzled and confused frown.
(11) We had better understand the young and elderly without starting off from
a biased point of view, whereby they are separated by so deep a grave, but
rather by watching how their perspectives on reality can change when they
face up reality.
Examples (10)–(11) show what appear to be deliberate attempts to use figurative

language for rhetorical purposes: they do not follow standard L1 patterns, nor do
they tally with L2 norms, and they appear to be compositional. Meaning can be
extracted from these phrases, but by failing to adhere to L2 phraseological norms,
fluency is compromised. Puzzled and confused (Example 10) does not appear in
the BNC (even though the near-synonymous phrase bewildered and confused
occurs 3 times); and although puzzled, perplexed and worried all modify frown,
confused does not. Additionally, puzzled and… follows the verb look in 25% of in-
stances, and is typically located in post-modifying position. Were frown to have
been replaced by look (on our faces), the expression would have passed virtually
unnoticed. As it is, however, the non-standard version requires reprocessing. As
a facial expression, frown is related to look, but it forms different phraseological
patterns; and this is enough to impede the flow of the meaning.
Using terms which represent vertical space (i.e. depth) is an unusual way of
speaking metaphorically about distance. In both Italian and English, separato and
separated tend to collocate with terms which represent horizontal space: this is
true for time spans, viewpoints and physical distance, while the only examples
of vertical separation offered in corpus data refer to physical divisions effected
by the use of plate glass or metal. So when we find the expression separated by so
deep a grave used to explain divergence in opinions (Example 11), it strikes us as
odd because it refers to vertical space. Although it is beautifully constructed in
grammatical and rhetorical terms, the conceptualisation is anomalous. To further
confuse the reader, the choice of grave here is rather infelicitous because it occurs
in a context where the young and the old are being compared. The proximity of
elderly and grave triggers the literal meaning of grave, which may not be the one
intended. With no other indicators provided, the expression remains opaque and
open to variable interpretation.
The sorts of language mismatch illustrated in this chapter suggest that famil-
iarity with collocational patterning is ultimately more influential than conceptual
knowledge in achieving fluency in a foreign language. Students bring their L1
conceptual knowledge with them when they work in the L2, and if the languages
share common cultural and linguistic ground, over-generalisations abound re-
garding the applicability of conceptual and lexical information. The ‘same’ word
is thought to have the same meaning and sphere of reference, and because of this
students are apt to use the L2 equivalent in the same phraseological patternings as
those used in the L1. Similarity makes students reluctant to consult dictionaries
at the advanced level, because they are quite convinced that they already ‘know’
the word. And because many students have managed to get by in much of their
language learning by falling back on translation and approximate renderings of
what they believe they have seen, they often fail to develop the necessary degree
of sensitivity to phraseology required for them to master the L2.
6. Conclusions
While attention to metaphor in foreign language pedagogy is indisputably helpful

in the learning process, some of the issues raised in this chapter require further
attention. In the first place, it is apparent that a great deal of awareness-raising is
required in the language classroom if students are to fully appreciate how their L1
knowledge is to be encoded in the L2. It is not apparent to most students that their
world knowledge is structured in terms of their L1, and it comes as a surprise to
find that the L2 lexicalises concepts in palpably different ways. It is therefore im-
portant that students be encouraged to compare and contrast the two languages,
even though this runs somewhat contrary to the preferred monolingual approach
to foreign language teaching.
One of the problems with teaching and learning figurative expressions in the
L2 is the risk of over-estimating the metaphorical vividness. Decoding from the
L2 favours salient meanings, reading phrases compositionally when in most cases
the language is non-compositional, delexicalised and metaphorically dead. As a
consequence, figurative expressions whose wordings are different to those used in
the L1 are often perceived as being more figurative than they really are. It is not
easy for a learner to appreciate delexicalised, phraseological meanings because
they sound unconventional compared to the patterns they are familiar with in
their L1. The relatively low incidence of this sort of language in learner writing
78 Gill Philip
can probably be attributed to a desire to avoid sounding ‘foreign’, when in fact the
use of these conventional phrases would have precisely the opposite effect.
Figurative and metaphorical senses of words do not exist in isolation, but are
created and fixed in context. Form and meaning interact in very delicate and de-
tailed ways, as discussion of the data above has highlighted, so if priority is given
to content words alone, the link between wording and meaning is seriously com-
promised. Encouraging students to remember lexical information conceptually
or visually, while advantageous to the learning and decoding of new vocabulary,
may cause interference between recognition of a language item and its recall for
encoding process (as opposed to recall in elicitation tests). Errors and inaccura-
cies in the phraseology can interfere with meaning even when no fault can be
found with the collocation of content words.
Divorcing content and structure causes meaning to disintegrate. Meaning is
wholly dependent on form, and if learners are to incorporate conventional figura-
tive language into their productive repertoires, they will have to focus at least as
much on the finer points of phraseology as they currently do on the semantic and
conceptual content.
Notes
1. The corpora consulted were (for English) the British National Corpus http://www.natcorp.
ox.ac.uk/ and (for Italian) CORIS http://corpora.dslo.unibo.it/CORISCorpQuery.html
2. Here and in subsequent examples: any errors in the examples are original; all emphasis is
editorial.
3. There were only two occurrences of heavy + depression in the British National Corpus
(http://www.natcorp.ox.ac.uk/) both of which referred to the atmosphere (ambience); compare
to twenty-four occurrences of deep depression, of which twenty refer to the emotional state, two
to the weather, and two to the economy.
References
553–571.
Boers, Frank & Hélène Stengers (2005). Adding sound to the picture: An exercise in motivating
the lexical composition of idioms in English, Dutch and Spanish. Paper read at Phraseology
2005: The many faces of Phraseology. Louvain-La-Neuve, October 2005.
Charteris-Black, Jonathan (2000). Metaphor and vocabulary teaching in ESP economics. Eng-
lish for Specific Purposes, 19, 149–165.
Charteris-Black, Jonathan (2002). Second language figurative proficiency: A comparative study

of Malay and English. Applied Linguistics, 23, 104–133.
Danesi, Marcel (1994). Recent research on metaphor and the teaching of Italian. Italica, 71,
453–64.
Deignan, Alice, Danuta Gabryś, & Agnieszka Solska (1997). Teaching English metaphors using
cross-linguistic awareness-raising activities. ELT Journal, 51, 352–360.
Gibbs, Raymond W. & Teenie Matlock (1999). Psycholinguistics and mental representations.
Cognitive Linguistics, 10, 263–269.
Hanks, Patrick (2004). The syntagmatics of metaphor and idiom. International Journal of Lexi-
cography, 17, 245–274.
Holme, Randall (2004). Mind, metaphor and language teaching. Basingstoke: Palgrave MacMil-
lan.
Lakoff, George & Mark Johnson (1980). Metaphors we live by. Chicago: Chicago University
Press.
Lazar, Gillian (1996). Using figurative language to expand students’ vocabulary. ELT journal,
50, 43–51.
Louw, Bill (1997). The role of corpora in critical literary appreciation. In A. Wichmann,
S. Fligelstone, T. McEnery & G. Knowles (Eds.), Teaching and language corpora (240–251).
London: Longman.
Moon, Rosamund (2004). On specifying metaphor: An idea and its implementation. Interna-
tional Journal of Lexicography, 17, 195–222.
Philip, Gill (2005a). From concept to wording and back again: Features of learners’ production
of figurative language. In A. Wallington, J. Barnden, S. Glasbey, M. Lee, & L. Zhang (Eds.),
Proceedings of the Third Interdisciplinary Workshop on Corpus-Based Approaches to Figura-
tive Language (46–53). Birmingham: University Press Birmingham.
Philip, Gill (2005b). Figurative language and the advanced learner. Research news: The newslet-
ter of the IATEFL Research SIG, 16, 16–20.
Philip, Gill (2006). Metaphor, the dictionary, and the advanced learner. In E. Corino, C. Marello,
& C. Onesti (Eds.), EURALEX Atti del XII congresso internazionale di lessicografia (895–
905). Alessandria: Edizioni dell’Orso.
Philip, Gill (2007). Decomposition and delexicalisation in learners’ collocational (mis)behaviour.
In Online Proceedings of Corpus Linguistics 2007. Retrieved from http://ucrel.lancs.ac.uk/
publications/CL2007/paper/170_Paper.pdf
Stengers, Hélène, June Eyckmans, Arnout Horemans, & Frank Boers (2005). Optimising mne-
monic strategies through CALL: A tool called Idiom Teacher. Paper read at Phraseology
2005: The many faces of Phraseology. Louvain-La-Neuve, Belgium. October 2005.
Ragazzini, Giuseppe (Ed.) (1995). Il Ragazzini: dizionario inglese italiano – italiano inglese.
Bologna: Zanichelli.
Zingarelli, Nicola (Ed.) (2001). Lo Zingarelli 2002 Vocabolario della lingua italiana. Bologna:
Zanichelli.
chapter 5
The gaps to be filled

The (mis)treatment of the polysemous senses
of hand, cool and run in EFL text books
Elisabet Amaya Chávez

Universidad de Extremadura
Over fifteen years have elapsed since the publication of Low’s (1988) article on
teaching metaphor; this has been followed by a number of experimental studies
on the teaching/learning of figurative language in an FL. Some of these have
been able to make recommendations about the treatment of such difficult as-
pects of English for the non-native speaker as polysemy or phrasal verbs. How-
ever, little is known about how much impact this research activity has actually
had on the real world of ELT, or on what actually goes on in EFL classrooms.
As Gibbs (this volume) suggests, “[…] real-world metaphor research needs
to explore situations which are as much social as psychological and to try and
examine how both aspects interact.”
In order to gain some insight into what learners of EFL gather about the
meaning potential of the words they are taught, I examine the text books used
in two educational settings with learners between the ages of six and eighteen.
Focusing on three highly polysemous words (hand, cool and run), I examine
what different senses of these words are introduced, practised and recycled over
the twelve years in which English is an obligatory subject, and what activities
are used to foster understanding of the figurative uses of these words. This study
reveals that applied metaphor research has had virtually no impact on the text
books used in these classrooms.
Keywords: applied metaphor research, EFL, text book, polysemy, metaphor,

metonymy
82 Elisabet Amaya Chávez
1. Introduction: Polysemy in the EFL classroom
Taylor (2002: 323) has noted that most lexical items are polysemous to a greater or
lesser extent, which means that most words a student learns to master in a foreign
language will have more than one sense. As Lazar (1996: 44) puts it,
[…] the figurative extensions of a word’s meaning […] form an intrinsic part of
the lexicon of the native speaker, even if the relationship to the literal meaning
of a word has been forgotten or obscured. Being able to understand and generate
figurative extensions for a word in English would thus appear to be an important
vocabulary-building skill for the language learner.
It is also known that polysemous terms may pose problems for non-native speak-
ers. This may be one of the reasons why vocabulary acquisition is one of the
main areas of interests for Applied Cognitive Linguists (e.g. Boers, 2000; Boers &
Demecheleer, 1998; Deignan et al., 1997; Dirven, 2001; Kövecses & Szabó, 1996;
Lazar, 1996; Lindstromberg, 1996; Verspoor & Lowie, 2003), who have shown
that making learners aware of the figurative motivation of target language words
can help them in various ways: in fostering comprehension and retention, for
example. In other words, it would be beneficial for learners to become aware of
the processes involved in the semantic extension of core vocabulary items. How-
ever, whether these insights inform the teaching of polysemous words in FL class-
rooms is largely unknown. Certainly, little is known about what other strategies
and methods may be adopted in fostering learners’ grasp of the semantic possi-
bilities and referential flexibility of the target language words they know, if indeed
this is done in any principled way.
A very important source of information both for teachers and learners is the
text book used in the classroom. English language teachers tend to rely heavily
on the mediation of course book writers in the application of new research into
ELT. This does not mean that all teachers limit themselves to the texts they use
(the approach proposed in the teaching materials may well be supplemented by
other activities, and is therefore only a partial indication of what is going on
in the EFL classroom), but the approach to the teaching of polysemy adopted
by teachers may be conditioned by that of the text book used in the classroom.
Thus, looking at the text books that a learner uses in the course of his or her lan-
guage learning life can shed some light on what the learner discovers about the
principles behind polysemy.
This chapter analyses the treatment of polysemy in the text books used in two
Spanish educational settings. The study pursues three main aims. First, to explore
what happens in a language learner’s life aged 6 to 18, that is, in “obligatory edu-
cation”. Second, focusing on three lexical items of different grammatical classes,
Chapter 5. Hand, cool and run in EFL text books 83
this chapter studies how many and what different senses and forms of these words
are presented in the course books chosen and at what stages these are introduced,
practised and/or recycled. The third major aim is to discover whether the two
hypothetical learners receive any instruction (explicit or embedded) on principles
of meaning extension via metaphor and metonymy. These findings are compared
with that of research in this field.
2. Method
In exploring what, if anything, the learner discovers about the principles behind
polysemy during the course of his or her language learning life, I will be refer-
ring, for ease of exposition, to two hypothetical students (Student A and Stu-
dent B). These abstract figures should not be taken to represent any real learn-
ers, but rather a representation of part of the typical learning experience of two
youngsters attending schools in Cáceres (Spain) from the age of 6 to 18. In order
to gain some insights into how these hypothetical learners might be helped to
come to grips with the meaning potential of the target language lexicon, I have
selected for this study particular educational settings, a range of different EFL
text books and a small number of lexical items. These are described in turn in
the following sections.
2.1 The educational settings
The Spanish system is divided into “Educación Infantil” (from 3 to 6), “Primaria”
(from the age of 6 to 12), “E.S.O.” (from 12 to 16) and “Bachillerato” (from 16 to
18). Education is obligatory from ages 6 to 16, but the last two years of second-
ary education have been included in this study for two reasons. First, because
English is obligatory in “Bachillerato”, and second, because it allows a wider view
of the presentation of the three items analysed. Nonetheless, in this chapter I
will not be employing the terminology used in Spain (i.e. “Primaria”, “E.S.O.” or
“Bachillerato”), but rather, the learning life of two hypothetical students has been
considered in terms of years, from years 1 to 12, when English is an obligatory
subject. So, year 12, for example, is “Segundo de Bachillerato”, or the last year at
school. Text books play a minor role in “Educación Infantil” and have thus not
been included in this study.
The two hypothetical students in this educational system attend schools in
Cáceres, a town of about 95,000 inhabitants in the west of Spain (Extremadura).
Learners in this setting tend to go to both the Primary and Secondary school in
their area. Thus, two educational settings in which an English language learner’s
life takes place were chosen in order to examine the text books used in each one.
The settings are: Primary school ‘A’, and Secondary school ‘A’, and Primary school
‘B’ and Secondary school ‘B’.1
2.2 The text books
The two Primary and Secondary schools attended by these hypothetical learn-
ers use a different text book for each academic year. Thus, a selection of the text
books used in these four schools should result in a total of twenty-four different
text books (or twelve per student) over the twelve-year period of education con-
sidered here. However, the number actually obtained is twenty-one, because the
same texts are employed by schools A and B in years 7, 11, and 12.
I read this corpus of 21 EFL texts (each ‘text book’ comprising the student’s
book, activity book, tape, and Teacher’s book) taking note of all the examples
they contain of hand, cool and run in order to examine the forms presented, their
senses and when they are introduced (i.e. the first time the term is presented to
the learner), as well as the frequency of these forms and senses over the twelve-
year period. Finally, I studied the teaching/learning activities and strategies pro-
posed by material writers to promote the understanding of figurative uses of these
words over the twelve years, comparing them to those offered by applied cogni-
tive linguists.
The books used in each school are the following (see References for more
details): in Primary School ‘A’ (from 6 to 12 years old or years 1 to 6) I exam-
ined My Friends (2001 [MFA1],2 2003 [MFA2]) and Best Friends (2000 [BFA4],
2001 [BFA3], 2002 [BFA5], 2002 [BFA6]) published by Oxford University Press;
in Secondary School ‘A’ (from the age of 12 to 18-years 7 to 12) English Zone (2002
[EZA9], 2003 [EZA10], 2004 [EZA7], 2004 [EZA8]) and Citizens (2003 [CA12],
2004 [CA11]) published by Longman. In Primary School ‘B’, the books used are
Zoom (2003 [ZB1,2,3,4,5]) produced by Richmond Publishing and Oxford Ex-
change (2002 [OEB6]) published by Oxford University Press, and for Secondary
School ‘B’ English Zone (2004 [EZB7]) published by Longman, Oxford Exchange
(2002 [OEB8,9,10]) published by Oxford University Press and Citizens (2003
[CB12], 2004 [CB11]), another Longman publication.
These text books were used in the EFL classrooms described during the aca-
demic years 2003 to 2005 and can thus be regarded as a reasonably up-to-date
source of information for teachers and learners. It should be noted, likewise,
that they were produced at least twelve years after Low’s (1988) influential ar-
ticle on teaching metaphor, and between four and eight years after other notable
contributions to the field (e.g. Boers & Demecheleer, 1998; Deignan et al., 1997;
Kövecses & Szabó, 1996; Lazar 1996; Lindstromberg, 1996). Although practis-
ing teachers may well not know of the research reported in these publications,
it seems reasonable to expect that the authors of published teaching materials
would keep abreast of developments in the field of applied linguistics and ELT
methodology. At the same time, it should be noted that such texts are produced
in response to the demands of the educational settings in which they are used,
and all those examined were written in accordance with the Spanish curriculum
for English. Thus, although written by different authors and published by dif-
ferent companies, the content is very similar in each book, the focus being on
the four skills. Overall, the emphasis in the curriculum is on skills development,
with a focus on grammar and phonology, while vocabulary input is not specified
(although topic areas are). Nevertheless, lexico-grammatical aspects of English
such as ‘phrasal verbs’ are the focus of attention in years 11 and 12, and it might
be expected that the treatment of this difficult area of English at least might be
affected by recent research.
2.3 Items chosen for analysis: hand, cool and run
In a study of the type described here, it was not possible to study a large number
of polysemous lexical items and, at the same time, obtain the kind of detail about
their treatment that was considered desirable. Thus, some selection was neces-
sary. The three items chosen for the study are three highly polysemous words
(hand, cool and run), which, using the Collins COBUILD Dictionary’s frequency
bands (2001) as a guide, are also very frequently used in English. According to
this authority, hand and run are among the 680 most frequently used words in
English. That is, along with function words like and or the, they are an essential
part of a speaker’s word stock. Cool, which is included in the COBUILD Fre-
quency Band 4, is not quite as frequent as the former, but is nevertheless still very
commonly used by native speakers.
The choice of these items was not motivated solely by their frequency, how-
ever. Another factor was that the core form of each word belongs to a different
grammatical class: hand is a noun, run a verb, and cool an adjective. Insights into
what treatment the polysemous senses of words of different grammatical classes
receive in these text books might also reveal something about how the different
senses of these words are learned in context.
Finally, each of the items selected lends itself not only to sense extension with
the same form, but also to grammatical shift or conversion. Thus, hand can be used
as a verb (e.g. Hand it to me) and cool as a verb or noun (e.g. Cool it! or He lost his
cool) and run can also be used as a noun (e.g. a run of good luck). Such grammatical
shifts may be regarded, from the point of view of the learner of English as a foreign
language (whose first language may not allow such conversions), as motivated by
conceptual metonymies like entity for activity or attribute for entity (see
Kövecses, 2002), for example, and will thus be one of the aspects of polysemy gen-
erally in English with which learners will need to become familiar.
The different possible forms and senses of these words and examples of how
they may be used in English have been taken from a variety of sources, espe-
cially standard dictionaries: Oxford Dictionary of Phrasal Verbs (1993); Collins
COBUILD, English Guide 7: Metaphor (1997); Collins COBUILD Dictionary of
Idioms (2000); Oxford Advanced Learner’s Dictionary (2000); Collins COBUILD
English Dictionary for Advanced Learners (2001), and The Penguin Dictionary of
English Idioms (2002). When the examples of the words are from the text books
analysed, this is clarified in the text by using the abbreviation for each text de-
scribed above. For example, Hands up! (MFA1: 6).
3. Description of results
This core section gives an overview of the main findings in this study, that is, the
presentation of the three lexical items in the text books examined.
3.1 Analysis of hand
Hand is a highly polysemous word that can be used in different grammatical

forms. This may be one of the reasons why it appears so frequently both in the
language of speakers and in the text books examined, although not all the gram-
matical forms are present in the teaching materials.
The prototypical meaning of hand is introduced at an early stage of both lan-
guage learners’ lives: it appears in the text books used in the first year for student
A and the second year for student B. Although this central meaning can be ex-
tended by means of metonymy (for example to denote a worker or a set of
playing cards), without changing its form, or metaphor (as in the case of the
hands of a watch), neither of these extensions is found in the text books used in
student A’s learning life. In contrast, student B encounters extensions of the core
meaning in the tenth year, metonymic (1a) and metaphoric (1b), these being the
only examples that appear in the books:
(1) a. Do you want a hand? (OEB10: 22)
b. The future is in your hands (OEB10: 52)
Hand compounds are introduced in the sixth and seventh years respectively, most
of them related to sports in both cases, for example:
(2) handball (EZA/B7: 64)
As a verb alone and denoting activity, hand is very frequent in spoken English
(the Collins COBUILD Dictionary (2001) includes the verbal use of this word in
its Frequency Band 4, which is proof of its importance for learners of English).
However, this form is not introduced at any stage in any of the text books exam-
ined; that is, neither student A nor student B comes across the verb in their text
books. Nevertheless, both students are introduced to one use of hand followed by
a particle in the eleventh year:
(3) hand out (CA/B11: extra material).
In multi-word units hand can appear in habitual collocations, which can be mo-
tivated by conventional knowledge (Kövecses & Szabó, 1996), that is, people’s
knowledge, in a given culture, of certain conventionalised gestures that involve
the human hand (e.g. hands up! meaning ‘surrender’); as an adverb, such as by
hand (manually for unaided) or at hand (hand for proximity). It can also
appear in idioms, motivated either by metonymy, as in give me a hand, or meta-
phor, for instance, when something gets out of one’s hands, s/he loses control over
it; in proverbs and disjuncts, which are on the one hand and on the other hand and
where the metaphoric mapping is sides are alternatives. As regards the year
when all these forms and senses of hand are introduced, habitual collocations are
presented in the second year in both cases, the examples being
(4) a. clap your hands (MFA2: 12)
b. put up your hands (ZB2: 8)
Adverbs are only introduced to student A, in the tenth year, in just one example:
(5) Is the fruit picked from the tree by hand? (EZA10: 61).
Figurative idioms are introduced in the eleventh and eighth years respectively, but
there is one single instance in each case:
(6) a. Will you give me a hand with my English homework? (CA11: 71)
b. Perhaps you could give us a hand? (OEB8: 102)
Proverbs appear only twice, in the eleventh year in both cases – as the same edi-
tion is used, one of them being Spanglish (7a):
(7) a. What they give with one hand se lo llevan con la otra (CA/B11: 312)
b. A bird in the hand is worth two in the bush (CA/B11: extra material)
Finally, the disjuncts are introduced earlier to student B (year 9) than to student
A (year 11).
In sum, hand is absent in text books as a verb alone. In addition, the senses
extended from the core meaning as a noun are not introduced throughout stu-
dent A’s learning life. In contrast, as an adverb it only appears in the course books
used by student A.
3.1.1 Frequency of hand

The frequency of tokens of hand in the 21 text books varies not only in the total
number (student A encounters the term 93 times, whereas student B is presented
with a total of 139 tokens), but also in each year, the frequency is different, partic-
ularly significant being the dissimilarity in year 2 (8 tokens for student A and 19
for student B) and years 9 and 10 (12+11 vs. 27+28). It is also worth mentioning
that neither for student A nor for student B is there a progression in the learning
of this lexeme as presented in the books, as there are gaps (year 4 and year 1) and
bunching (year 12 in the case of student A and 9 and 10 for student 2). In other
words, the tokens appear frequently at certain points in the textbooks but not at
others (see Figure 2).
3.1.2 Frequency of forms of hand

As regards the frequency of forms of hand, the two students are presented with al-
most the same grammatical forms at Primary school, these being hand as a noun
(Student A: 12 tokens and student B: 25 tokens) and habitual collocations, moti-
vated by conventional knowledge in examples such as Hands up! and Clap your
hands. The interesting difference regarding the form is that student A encounters
a compound in a reading text, which is an extension of the core meaning:
(8) It’s the hands-on science museum in the centre of Oxford (BFA6: 49)
This use of hand is well employed in the book, since it appears to be a not uncom-
mon form, as a search of the BNC shows (118 matches have been found).
From years 7 to 12, noun compounds and disjuncts are the most frequent
forms (the noun has the highest frequency for student A – 34 tokens – and com-
pound for student B – 33). In contrast, hand as a verb alone does not receive any
attention in the books used by student A, and these plus adverbs are also absent in
student B’s books. To sum up, from years 7 to 12, students A and B are introduced
to new forms: verb plus particle, proverbs and disjuncts, adverbs (in the case of
student A) and compounds (in the case of student B).
3.1.3 Frequency of senses of hand

Regarding the senses of hand found in the text books, the literal or prototypical
meaning is the most frequent from years 1 to 6 in both schools (12 literal uses
for student A and 24 for student B), followed by the idioms mentioned earlier
motivated by conventional knowledge, and which are also literal. There is just one
figurative use of hand in student A’s books (hands-on) and none in student B’s.
The opposite happens from years 7 to 12: the number of figurative exten-
sions of hand is higher than the literal uses (student A: a total of 40 figurative
tokens – out of 75, 25 of which are metonymic and 15 metaphoric; student B:
a total of 70 – out of 107, of which 40 are metonymic and 30 metaphoric). This
frequency is due, above all, to the high number of compounds and disjuncts that
appear in the texts.
3.2 Analysis of cool
The second item analysed, cool, shows a similar picture to that of hand, though it
reflects even more clearly the unsystematic treatment of polysemy in the course
books examined.
The core meaning of cool, which has to do with temperature (i.e. ‘fairly cold;
not hot or warm’), has a number of metaphoric senses.3 For example, the adjec-
tive cool can be used to talk about ‘feelings’, ‘emotions’ or ‘behaviour’, denoting
‘calmness’ or ‘control over emotions’, the most frequent metaphoric sense ac-
cording to Deignan (1999). It can also denote ‘lack of friendly feelings’ (e.g. He
was cool and arrogant). Besides, cool is also used with the meaning of ‘fashion-
able’ connoting ‘approval’, as in He thinks it’s cool to do heroin; it can also refer to
a ‘large amount of money’ (e.g. The car cost a cool thirty thousand) or to ‘quality
of colours’, as in A room painted in cool greens and blues. The adjective can also
be used of ‘tastes’ or ‘scents’.
Student A does not appear to meet the core or literal meaning of cool at any
time in his learning life, a somewhat surprising omission if one considers the fact
that these text books have been prepared for learners who live in a part of the
world where being literally ‘cool’ is of utmost importance if one is to feel com-
fortable during the heat of the summer. More importance seems to be given to
the notion of ‘fashion’, for this learner is introduced to the metaphoric sense of
‘fashionable’ in the ninth year, in the following examples:
(9) a. I’m tall, good-looking, cool, intelligent, funny… (EZA9: 8)
b. He’s really funny and cool but… (EZA9: 8)
The case of student B is even more striking: he is introduced to the metaphoric

sense of approval in the third year in examples such as
(10) Cool! (ZB3: 9, 35 & 57)
but does not encounter the core sense until the ninth year, and only in one
example:
(11) […] cool, cool ice cream (OEB9: 79)
One might expect the metaphoric extension of cool to follow the more basic tem-
perature sense, but it does not. The different senses of this word are thus treated by
course book designers as though they were unrelated or homonyms.
In addition, although cool has many other semantic extensions and has other
functions apart from the adjectival (it can be used as a noun, verb, or appear in
idiomatic expressions), these are all absent in the text books examined.
3.2.1 Frequency of cool

As far as the frequency is concerned, the treatment of cool is different for students
A and B. In the case of student A, there is bunch coverage (0 tokens from year 1 to
8 and from 11 to 12, 2 tokens at year 9 and 13 tokens at year 10 – all figurative in-
stances). Student B, on the other hand, is introduced to almost the same number
of tokens (12 vs. 15 for student A) throughout his learning life, but these appear
at wider intervals (4 tokens in the fourth year, 4 at year 8 and 2 at years 9 and 10
respectively). In short, similar to hand, the treatment of this lexeme is unsystem-
atic, providing the learner with no indication of its semantic possibilities. For an
overview, see Figure 2.
3.2.2 Frequency of forms of cool

As regards the frequency of forms of cool, the adjective only appears in the books
used by student B from years 1 to 6. Student A, in contrast is not using cool at
this stage. From years 7 to 12, however, student A is presented with 15 tokens of
adjectival cool (note that he had not been introduced to any token from years 1 to
6), in a similar way to student B’s experiences. He or she comes across another 8
tokens from years 7 to 12.
3.2.3 Frequency of senses of cool

Finally, coverage of the senses of cool is extremely poor: student A only learns, as
presented in the text books, metaphoric uses which are associated with the idea of
being fashionable or attractive, in examples such as the following:
(12) a. I’m tall, good-looking, cool intelligent, funny (EZA9: 8)

b. That’s the coolest hat in the shop (EZA10: 38)
and s/he is not introduced to the core meaning at any point. Student B, on the
other hand, is presented with literal and metaphoric senses, but with no progres-
sion: the prototypical meaning is learnt at years 9 and 10, but only two examples
have been found in each year:
(13) a. […] cool, cool ice cream (OEB9: 79)
b. At night it’s cooler […] (OEB10: 76)
whereas the metaphoric extension, which is ‘approval’ in all cases, is introduced

in the third and eighth years respectively:
(14) a. Cool! (ZB3: 9)
b. That’s cool! (OEB8: 82)
In sum, there is only one form of cool in all the text books examined, the adjective,
and two senses, literal (only in the case of student B) and denoting ‘approval’.
3.3 Analysis of run
The last item, run, is, like hand, highly polysemous and frequent: for example, the
Collins COBUILD Dictionary (2001) records 58 senses as a verb alone and includes
it in its Frequency band 5, so its importance for learners of English is obvious.
The prototypical meaning, which according to Gries (2006) is ‘fast pedestrian
motion’, is introduced in the third year for student A and in the first year in the case
of student B, that is, at early stages. However, extensions of the core meaning are
absent. That is, although intransitive uses of the verb has other senses apart from
the core meaning (e.g. ‘motion of liquids’; ‘diffusion of colours’; ‘operate a system’;
‘publish’, among others), these are ignored in the text books examined. This means
that neither student A nor student B encounter any senses of the intransitive verb
apart from the prototypical meaning in the course of their learning life.
The treatment of the transitive use (‘run sth’) is slightly different from the in-
transitive. Both students encounter the prototypical meaning at late stages (years
10 and 8), for example:
(15) Run a marathon (EZA10: 90)
In addition, they are presented with a semantic extension of the core meaning in
the eleventh year, ‘be in charge of ’, but only one example has been found at ran-
dom in a reading text:
(16) Many [immigrants] run shops (CA/B11: 77)
A lack of progression or sequencing in the teaching of the different senses can also
be observed: there is a clear jump from the core meaning to the figurative sense of
‘being in charge of ’. Other senses such as ‘cause motion’; ‘smuggle’; ‘give a test’ or
‘cost’, among others, are not introduced at any point.
Similarly, intransitive phrasal verbs are introduced at late stages to both stu-
dents: the basic meaning, which is ‘fast pedestrian motion’ (e.g. It’s raining, run
in!) is encountered in the tenth year by student A and in the eighth year by stu-
dent B. But the gaps in terms of senses are still present – the only figurative sense
taught is ‘become used up’, and only to student A:
(17) The battery [of a mobile] may run out (EZA10: 51)
Other phrasal verbs (e.g. run away or run away with) are absent.
In a similar way, the only transitive phrasal verb present is ‘run out of ’, which
is introduced to both students in the eleventh year. In other words, although run
combines with different particles (by, down, in, off, out of or up) in many transitive
patterns, these are all absent in the corpus of books analysed. Furthermore, run
can also appear in idiomatic expressions, such as run deep, run its course, run riot,
run the gauntlet, run the show or run one’s eye over something, but none has been
found in the books.
In short, despite being highly polysemous, most of the possible senses of run
are absent in the books: among the different senses of the intransitive verb, only
the core meaning is present. One further sense is added in the case of the transi-
tive verb (‘be in charge of ’). In phrasal verbs, apart from ‘pedestrian motion’, the
only sense introduced is ‘use up’. Moreover, idioms are ignored.
3.3.1 Frequency of run

The number of tokens with run is higher for student B than for student A (85 vs.
138), but the coverage is more balanced as presented in the books used by stu-
dent A: there are gaps in terms of years, but these appear at the beginning (years
1 and 2). For student B, the dissimilarity is obvious (e.g. run is absent in the third
and sixth years). As in the case of hand and cool, the tokens appear frequently at
certain points in the text books but not at others. See Figure 2 below.
3.3.2 Frequency of forms of run

The gap in the treatment of run is clearer when looking at the forms both stu-
dents encounter in the course books from years 1 to 6: they are only introduced
to the intransitive verb form, whereas transitive verbs, phrasal verbs and idioms
are absent.
From years 7 to 12 more forms are introduced to both students, though the
intransitive verb has the highest frequency compared with the other forms (37
tokens – student A and 53 – student B), followed by the verb used transitively in
the case of student B (13 tokens vs. 3 for student A) and the transitive phrasal verb
in the case of student A (6 tokens of run out of). In other words, there is bunch
coverage in terms of forms.
3.3.3 Frequency of senses of run

As far as the frequency of senses of run found in the text books is concerned, its
coverage is different from hand and cool: figurative senses of hand (in the case of
student B) and cool are more frequent than the literal ones. The opposite happens
to run: the number of literal senses clearly exceeds the number of figurative uses
for both students (10 figurative senses out of 85 for student A, and 10 out of 138
for student B – introduced at late stages in both cases).
In sum, the treatment and coverage of run in the text books seems to be even
poorer and more unsystematic than the treatment and coverage of hand, and even
of cool bearing in mind its frequency in the English language, especially in terms
of the different senses introduced: there is a clear lack of progression, and the
figurative uses are almost neglected – all of them being associated with the idea
of ‘manage’ and ‘use up’.
4. Summary of results
The study reveals that, of the three cases studied, hand has the highest frequency
in the text books, followed by run. In addition, except in the case of cool, student
B encounters more tokens of the words than student A. Figure 1 shows the total
number of tokens for both students.
150
100
50
0
HAND COOL RUN
Student A 93 15 85
Student B 139 12 138
Figure 1. Summary of frequency of tokens of hand, cool and run

Yet both students share the uneven distribution of tokens of the words through-
out their language learning life, as shown in the graphs below, that is, the tokens
appear frequently in some years but are absent in others.
Figure 2. Summary of frequency of hand, cool and run in the life-span
of students A and B
As regards the senses, semantic extensions of the three items analysed are present
in the materials examined, though there are unexpected omissions. Nonetheless,
figurative uses are more frequent in some cases. Table 1 presents a summary of the
percentages of literal and figurative senses of the three items in the life-span of the
two hypothetical students:
It can be observed that literal senses are more frequent in the case of run (88%
for student A and 93% for student B) and hand, only for student A (56%). In contrast,
student B encounters more figurative uses of hand (51%) than literal ones, and cool
appears more frequent figuratively for both students (100% and 67% respectively).
Table 1. Summary of frequency of senses

Student A
Literal Figurative
Rate % Tokens Rate % Tokens
hand 56 52 44 41
cool 0 0 100 15
run 88 75 12 10
Student B
Literal Figurative
Rate % Tokens Rate % Tokens
hand 49 69 51 70
cool 33 4 67 8
run 93 128 7 10
However, the relation between the core sense and the more peripheral senses seems
to be ignored in the two sets of texts books. Rather, the different senses of the three
items are treated as though they were different words. The clearest example of the
lack of systematicity (although all three items reflect the mistreatment of polysemy
in the text books examined) is cool. This means that there is not embedded instruc-
tion on principles of meaning extension via metaphor and metonymy.
5. The presentation and practice of hand, cool and run
This section explores the types of activities present in the 21 text books, which
may usefully be contrasted with recent research into the teaching/learning of fig-
urative language in EFL in order to draw attention to the mismatch between the
two. I first provide a brief overview of recent research into the teaching/learning
of polysemous words; I then describe the activities.
5.1 Applied cognitive linguistics approaches to polysemy in FLT
According to cognitive linguists, metaphor and metonymy are not simply decora-
tive devices used in poetry, but rather, they are pervasive both in language and
thought. In short, they are responsible for the creation of new words and expres-
sions and reflect the way people ordinarily conceptualise themselves, events and
the everyday world. This leads applied cognitive linguists to consider metaphor
and metonymy useful tools in the ELT classroom.
Applied metaphor research carried out since Low’s paper (1988) has focused
mainly on the teaching of idioms (Boers et al., 2004; Kövecses and Szabó, 1996;
Lazar, 1996), prepositions (Boers and Demecheleer, 1998; Lindstromberg, 1996;
MacLennan, 1994), and phrasal verbs (Dirven, 2001; Kurtyka, 2001). These pro-
posals point to the need to raise learner’s metaphorical awareness, as it may be
useful for expanding learners’ vocabulary. The main objectives of raising learners’
awareness of metaphor are, according to Boers (2000: 566):
– recognition of metaphor as a common ingredient of everyday language;
– recognition of metaphoric themes behind many figurative expressions;
– recognition of the non-arbitrary nature of many figurative expressions;
– recognition of possible cross-cultural differences in metaphoric themes;
– recognition of cross-linguistic variety in figurative expressions.
An efficient strategy proposed by applied cognitive linguists for helping learn-
ers to cope with polysemy is to make the learner aware of the core sense and of
how different senses are extended from it, rather than treating them as unrelated
or homonyms, since, according to Boers and Lindstromberg (2006), this would
ignore the pedagogical potential of creating meaningful associations between the
senses. In this regard, their proposals are:
1. Point out and explain where there is an extension of meaning or a chain of
meaning.
2. Encourage learners to hypothesise on their own about the semantic motiva-
tion of a target use.
3. Provide learners with the means to corroborate or falsify their hypothesis.
Nevertheless, materials writers (at least of the books examined in this chapter) seem
unaware of the proposals offered by applied cognitive linguists. As has been ex-
plained, there is no embedded instruction of the motivation or semantic extensions
of hand, cool and run. In addition, an analysis of the types of activities proposed
by the course book designers for the teaching/learning of the three items shows
that the text books also lack explicit attention to motivation. That is, the materials
writers seem to avoid making explicit reference to the fact that a large portion of
language is figurative, paying more attention to the formal aspects. For example,
phrasal verbs are not explained from a semantic point of view in the text books,
but rather, writers focus on how they work (e.g. where the direct object should be
placed). Further, although there are experimental studies that have tested the bene-
fits of providing students with the core sense of polysemous words first (e.g. Nerlich
et al., 2003; Verspoor & Lowie, 2003), and others that show learners of English as a
foreign language are able to comprehend and use figurative language by resorting to
their figurative capacity (e.g. Piquer-Píriz, 2005), vocabulary is presented in the text
books examined as random lists to be memorised, which means that students are
not encouraged to reflect on the meaning potential of these words.
MacArthur and Piquer-Píriz (2007) suggest that, just as learners can be
helped to discover the rules governing the grammar of the target language, so
they can also be helped to discover the principles involved in the meaning exten-
sion of core vocabulary items. While native-speaking children and even learners
of English as a Second Language may do this largely on their own in response to
the large and varied exposure they have to the target language, it seems unlikely
that EFL learners will be able to do this without the aid or guidance of teachers
and learning materials given the somewhat restricted input they experience. If,
as happens with the lexical items examined here, little or no attention is paid to
this aspect of language learning, it is likely that learners will fail to appreciate the
semantic potential of the words they know and will either have to resort to using
a much larger number of words than those used by native speakers in different
discourse contexts (c.f. Nation, 1990; or Peyawary, 1999) or they will not be able
to communicate on a large number of topics, as their knowledge of what these
words can be used to refer to will be severely restricted.
5.2 Activities designed for the teaching/learning of figurative uses of hand,

cool and run in the text books
Most of the time, the three words analysed appear at random in reading texts, lis-
tening tape scripts or songs with no explanation about their meaning or exercises
devoted to them. And when there is some work dedicated to the terms, the focus
is, in most cases, not on the semantic possibilities of such terms, but rather on the
form or grammar, which is especially frequent in run (e.g. run/running), but hap-
pens also with hand and cool. This finding contrasts with the proposals made by
applied cognitive linguists, who point to the need to encourage students to think
about the metaphoric nature of language.
As regards the activities proposed by the materials writers for the learning of
the figurative senses of the three lexemes, the commonest tasks in the two sets of
text books are gap fills, matching exercises and translation.
In the case of hand, most of the assignments found are dedicated to the teach-
ing of disjuncts, followed by an exercise on proverbs and another on phrasal verbs.
Regarding the former, the first time the disjunct on the other hand is introduced to
student B is in the ninth year (OEB9: 91), but s/he is simply asked to make a trans-
lation into Spanish in this case. This exercise is followed by a matching activity,
where the learner has to link part of a sentence with another part that includes the
above mentioned expressions. Similarly, student A is asked to match some linkers
with the idea they express (CA11: 53).
In the exercise on proverbs (CA/B11: extra material), students are asked to

translate them and find equivalent versions in their own language when possible,
making use, if necessary, of the dictionary. This may be easy in the case of the hand
proverb, as the image evoked is similar in English and Spanish (A bird in the hand
is worth two in the bush / Más vale pájaro en mano que ciento volando), but not
in those where the image is different or are culture-specific. Deignan et al. (1997)
offer interesting proposals in this respect, such as awareness-raising through dis-
cussion and comparison of metaphors in L1 and L2, and maintain that,
If students are unaware of the metaphorical nature of their L1, there may be a
tendency to translate such items literally into L2, which may produce a marked
effect, or even be unintelligible to native L2 speakers. (p. 355)
The activity on phrasal verbs (CA/B11: extra material) is an exception in the sense
that the learner has to match the phrasal verbs with their meanings and then com-
plete some sentences with them:
4. hand out → B) distribute
– The teacher…………copies of the exam to his pupils.
As far as cool is concerned, in the case of student A there are only two activities fo-
cusing on the figurative sense, which are gap fills (EZA10: 38). On the other hand,
student B, who is introduced to the figurative sense at early stages – third year – is
presented with a reading text first, but with no activity on the term. In Unit 8, in
contrast, cool appears again in a reading text, but this time, the learner is asked to
look for known words and act out the dialogues (ZB3: 57). This is followed by a
listening activity where the student has to put a tick next to the term or expression
from the text he hears. Nonetheless, there is no reference to the figurative sense of
cool or to its relation with the core meaning in the text book. Other activities, at
later stages, are based on translation (OEB8: 82–83).
Interestingly, it is in the ninth year that student B encounters an activity fo-
cused on the core sense of cool (OEB9: 79). And, curiously, one of the instructions
found (p. 79) in the Teacher’s book for this activity is to “check students under-
stand the vocabulary in the box”. These exercises prove how the various senses of
cool are treated as different words.
Finally, in the case of run there are only activities related to the phrasal verb
run out of and only in the eleventh and twelfth years. The tasks are the following:
a. Matching. Students have to match the senses with the phrasal verbs, e.g.
I’ll go to the supermarket. We have run out of sugar → Finish
(CA/B11: 325 [Teacher’s book])
b. Writing sentences using some phrasal verbs, e.g.

Run out of → I can’t give you a lift because I’ve run out of petrol.
(CA/B12: 329 [Teacher’s book])
5.3 Strategies offered in the text books examined
The commonest strategies for vocabulary learning proposed by materials writ-

ers are the aforementioned, namely translation and the use of the dictionary. But
there are others, for example in the book used in the seventh year by both stu-
dents. There are “Teaching Tips” such as the following:
Grouping words by thinking of headings can help with memory (by association)
and recall. Letting learners think of their own headings themselves is a useful
part of the learning process. (EZA/B7: 137 [Teacher’s book])
It has already been shown that most figurative expressions can be traced back to
a common metaphoric theme or source domain. Thus, the idea proposed above
could be more useful if the learner is helped to organise or group the vocabulary
under their metaphoric themes or source domains, rather than as random lists
(Boers, 2000).
The books used in the eleventh and twelfth years (op cit) also offer some strat-
egies to help learners to understand difficult words: (1) Decide if the word is an
adjective, verb or noun, and (2) use the context.
However, the strategy proposed by applied cognitive linguists explained
above, namely the development of links between new vocabulary items and their
core sense, is not mentioned.
5.4 Summary
On the one hand, although semantic extensions of the three items analysed can be
found throughout the text books, it seems materials writers try to avoid dealing
with them, as most of the time the words appear at random with no explanation
or exercise, or the latter is focused on grammatical rather than on semantic as-
pects. One of the reasons why this is so may be because of the difficulty figurative-
ness can pose for learners.
The belief that figurative language may be difficult to teach and learn, could
also explain why most figurative senses of run are absent in the corpus of text
books examined, literal senses being more frequent (88% and 93% respectively –
see Figure 3). Phrasal verbs are one of the hardest areas of vocabulary for learners
of English and interestingly, they are almost absent in the text books examined,
especially in the case of run (bearing in mind it can combine with lots of particles),
and when they do appear, the emphasis is on syntactic behaviour. Further, they are
semantically disorganised and out of context (see Gibbs, this volume, for the im-
portance of context in the interpretation of metaphor). With this presentation the
learner is left with the impression that it is all very arbitrary and that there is not
an obvious “rule” (Rundell, 2005) underlying their use. Rudzka-Ostyn’s cognitive
approach (2003) could be a useful way of approaching phrasal verbs, although
it (and empirical studies showing the benefits of presenting phrasal verbs with
the conceptual metaphors that underlie them, e.g. Kövecses and Szabó, 1996) has
hitherto been associated with advanced learners.
On the other hand, there are plenty of exercises to teach the literal meaning
of hand and run at early stages. However the learner is not explicitly directed
to making connections between the core sense learnt at the beginning and the
semantic extensions he/she encounters in later years. The lack of explicit (and
embedded) instruction of motivation is even clearer when analysing the exer-
cises proposed for the teaching/learning of cool, since the activities focusing on
figurative senses precede those focusing on the core sense (in the case of stu-
dent B; student A is not presented with any activity to foster learning of the core
sense). These findings clearly reflect the mismatch between the proposals offered
by course-book designers, who do not include metaphor in the teaching materi-
als and skirt the issue of figurativeness, and those offered by applied cognitive
linguists, who argue for the inclusion of metaphor in the curricula and propose
explicit reference to principles of meaning extension (see Lazar (2003) for activi-
ties to teach hand and cool and Lindstromberg (2001) and Lindstromberg and
Boers (2005) for run).
6. Conclusions
Most people learn a foreign language, in this case English, when at school. In
other words, their language learning life can be expected to stretch from the age
of six to 16 or 18. So what have these two hypothetical learners gathered about
polysemy from the text books they have used during this period? Very little, it
would seem.
Although there are some differences between students A and B, mainly regard-
ing the year when the items are introduced and the frequency, the presentation
of the three items analysed is the same in the two sets of text books: the coverage
of hand, cool and run is unsystematic and extremely poor. For instance, there are
unexpected omissions of forms and senses; there is no progression; students are

not encouraged to make meaningful links from the core meaning of the lexemes.
Rather, the different senses of the three items analysed are treated as unrelated or
homonyms. In sum, there are gaps and bunching in terms of years (e.g. the items
analysed are introduced one year, then are absent and are presented again), forms
(e.g. hand does not appear as a verb in the text books) and senses (e.g. there is a
clear gap in the presentation of cool as only one metaphoric sense is introduced,
and the literal meaning is only introduced to student B at late stages). This means
that apart from a lack of explicit attention to the relation between the core sense
and the more peripheral senses, learners are unlikely to establish these relations
through embedded instruction on principles of meaning extension, which would
be reflected in the sequencing and frequency with which the forms and senses of
these three words are presented.
This study thus reveals the scant impact of all the research activity carried out
by applied metaphor researchers on English language teaching materials if the
treatment of the samples examined in these two sets of text books is representa-
tive of the way that polysemy is dealt with in textbooks generally. Our two hypo-
thetical learners, if they are to have any notion of the semantic potential of words
such as hand, cool or run, are going to have to look elsewhere for this knowledge,
for sadly they do not find it in the course books they use.
Acknowledgements
I wish to express my most sincere and heartfelt thanks to my supervisor, Fiona

MacArthur, for her endless help and support, for her patience, for making this
study possible. I also want to thank the anonymous reviewers for their useful
comments and suggestions. All remaining errors are my own.
Notes
1. The names of the schools have been disguised.
2. Each text book has the following code: abbreviation of its name, name of school (A or B)
and the year when it is used. For example, “MFA1” is My Friends, used in school ‘A’ in year 1.
3. Deignan considers that the different senses of cool are metonymic-based metaphors. For a
further discussion on this see Deignan (2005).
References
553–71.
Boers, Frank & Murielle Demecheleer (1998). A cognitive semantic approach to teaching prep-
ositions. ELT Journal, 52, 197–204.
Boers, Frank, Murielle Demecheleer, & June Eyckmans (2004). Cross-cultural variation as a
variable in comprehending and remembering figurative idioms. European Journal of Eng-
lish Studies, 8, 375–388.
Boers, Frank & Seth Lindstromberg (2006). Cognitive linguistic approaches to second or foreign
language instruction: Rationale, proposals and evaluation. In G. Kristiaensen, R. Dirven,
& F. J. Ruiz-Mendoza (Eds.), Cognitive linguistics: Current applications and future perspec-
tives (305–358). Berlin & New York: Mouton de Gruyter.
Deignan, Alice (1999). Metaphorical polysemy and paradigmatic relations. A corpus study.
Word, 50, 319–338.
Deignan, Alice (2005). Metaphor and corpus linguistics. Amsterdam & Philadelphia: John Ben-
jamins.
Deignan, Alice, Danuta Gabrys, & Agnieszka Solska (1997). Teaching English metaphors using
cross-linguistic awareness-raising activities. ELT Journal, 51, 352–360.
Dirven, René (2001). The metaphoric in recent cognitive approaches to English phrasal verbs.
Metaphorik.de. Retrieved May 15, 2005 from http://www.metaphorik.de/01/dirven.htm
Gries, Stefan Th. (2006). Corpus-based methods and cognitive semantics: The many meanings
of to run. In S. Th. Gries & A. Stefanowitsch (Eds.), Corpora in cognitive linguistics: Corpus-
based approaches to syntax and lexis (57–99). Berlin & New York: Mouton de Gruyter.
Kövecses, Zoltán (2002). Metaphor: A practical introduction. New York: Oxford University
Press.
Kövecses, Zoltán & Peter Szabó (1996). Idioms: A view from cognitive semantics. Applied Lin-
guistics, 17 (3), 326–55.
Kurtyka, Andrzej (2001). Teaching English phrasal verbs: A cognitive approach. In M. Pütz, S.
Niemeier and R. Dirven (Eds.), Applied cognitive linguistics II: Language pedagogy (29–53).
Berlin & New York: Mouton de Gruyter.
Lazar, Gillian (1996). Using figurative language to expand students’ vocabulary. ELT Journal,
50, 43–51.
Lazar, Gillian (2003). Meanings and metaphors: Activities to practise figurative language. Cam-
Lindstromberg, Seth (1996). Prepositions: Meaning and method. ELT Journal, 50, 225–236.
Lindstromberg, Seth (2001). (Sometimes) Against the grain: Total physical response for teach-
ing metaphorical language. HLT Magazine, 3/5. Retrieved December 22, 2005, from http://
www.hltmag.co.uk/sep01/lind.htm
Lindstromberg, Seth & Frank Boers (2005). From movement to metaphor with manner of
movement verbs. Applied Linguistics, 26, 241–261.
Low, Graham (1988). On teaching metaphor. Applied Linguistics, 9, 125–147.
MacArthur, Fiona & Ana M. Piquer-Píriz (2007). Staging the introduction of figurative ex-
tensions of familiar vocabulary items in EFL: Some preliminary considerations. Isla do
Desterro: A Journal of English Language, Literatures in English and Cultural Studies, 53
(Metaphor in language and thought: Contemporary perspectives), 123–134.
MacLennan, Carol H. G. (1994). Metaphors and prototypes in the learning and teaching of
grammar and vocabulary. IRAL, 32, 97–110.
Nation, I. S. P. (1990). Teaching and learning vocabulary. New York: Newbury House.
Nerlich, Brigitte, Zazie Todd & David D. Clarke (2003). Emerging patterns and evolving poly-
semies: The acquisition of get between four and ten years. In B. Nerlich, Z. Todd, V. Ger-
man, & D. D. Clark (Eds.), Trends in linguistics. Polysemy: Flexible patterns in mind and
language (333–357). Berlin & New York: Mouton de Gruyter.
Peyawary, Ahmad S. (1999). The core vocabulary of basic English: A corpus approach. Bergen:
The Humanities Information Technologies Research programme. HIT-senterets publik-
jonsserie, 2 (99).
Piquer-Píriz, Ana (2005). La comprensión de algunas extensiones semánticas de los lexemas
“hand”, “mouth” y “head” en las primeras etapas del aprendizaje del Inglés. Doctoral thesis.
Servicios de publicaciones de la Universidad de Extremadura.
Rudzka-Ostyn, Brygida (2003). Word power: Phrasal verbs and compounds. A cognitive ap-
proach. Berlin & New York: Mouton de Gruyter.
Rundell, Michael (2005). Why Are phrasal verbs so difficult? HLT Magazine, 7/3. Retrieved
December 22, 2005, from http://www.hltmag.co.uk/may05/idea.htm
Sinclair, John (Ed.) (2001). Collins COBUILD dictionary for advanced learners. Glasgow: Collins
Cobuild.
Taylor, John R. (2002). Category extension by metonymy and metaphor. In R. Dirven & R.
Pörings (Eds.), Metaphor and metonymy in comparison and contrast (pp. 323–347). Berlin
& New York: Mouton de Gruyter.
Verspoor, Marjolijn & Wander Lowie (2003). Making sense of polysemous words. Language
Learning, 53, 547–586.
Corpus data
British National Corpus (BNC)
Text books
Echevarría, Carmen, Emma Trelles & Elvira Menéndez (2004). English zone. 1 E.S.O. Madrid:
Longman.
Longman.
Longman.
Longman.
Holderness, Jackie (2001). Best friends. 3 Primaria. Madrid: Oxford University Press.
Holderness, Jackie & Wendy Superfine (2000). Best friends. 4 Primaria. Madrid: Oxford Uni-
versity Press
Holderness, Jackie & Wendy Superfine (2002). Best friends. 5 Primaria. Madrid: Oxford Uni-
versity Press.
Holderness, Jackie, Wendy Superfine & Paul Shipton (2002). Best friends. 6 Primaria. Madrid:
Oxford University Press.
Mark, Geraldine & Ben Wetz (2002). Oxford exchange. 1 E.S.O. Madrid: Oxford University Press.
Montañés, Elvira, Jim Lawley, Rodrigo Fernández & J. Arredondo (2004). Citizens. 1 Bachille-
rato. Madrid: Longman.
Montañés, Elvira, Jim Lawley, Rodrigo Fernández & J. Arredondo (2003). Citizens. 2 Bachille-
rato. Madrid: Longman.
Reilly, Vanessa (2001). My friends. 1 Primaria. Madrid: Oxford University Press.
Reilly, Vanessa (2003). My friends. 2 Primaria. Madrid: Oxford University Press.
Zanatta, Theresa (2003). Zoom. 1 Primaria. Madrid: Richmond Publishing.
chapter 6
A cross-cultural study
of metaphoric understanding
Chongying Wang*,** and Ann Dowker**

*Nankai University / **University of Oxford
This chapter was an attempt to investigate the similarities and differences in the
metaphoric understanding in Chinese and English children and adults. 115 and
123 participants took part in two experiments respectively. Participants’ inter-
pretations were classified as perceptual, psychological, behavioural, functional,
no-response, evaluative, descriptive, cross-sensory or associative, etc. Psycho-
logical and perceptual explanations were found to be the two main explanations
people gave. Results showed that adults gave a larger number of psychological
interpretations than children in the explanation task, but similar in multiple-
choice task which indicated that there were fewer obvious differences between
children and adults in the understanding of psychological interpretations than
in their production; Chinese people gave more psychological interpretations
than English in each age group; and English adults gave more perceptual inter-
pretations than Chinese adults.
1. Introduction
There have been numerous studies of the development and early use of metaphor
(Gardner, Winner, Bechhoffer & Wolf, 1978; Billow, 1981; Fourment, Emmenecker
& Pantz, 1987; Gentner, Falkenhainer & Skorstad, 1988; Caramelli & Montanari,
1995). Winner (1995) noted that “the studies that began to burgeon in the 1970s
continued to reveal a picture of metaphor as a relatively late-developing skill (e.g.
Cometa & Eson, 1978; Winner, Rosenstiel & Gardner, 1976), a view consistent
with Piaget’s (1959) discovery that children had considerable difficulties inter-
preting proverbs.” However, later studies suggested that metaphoric production
and comprehension begin very early in childhood, and become more sophisti-
cated in adolescence and adulthood, possibly after a temporary dip in middle
childhood (Gardner et al., 1978; Winner, 1988).
106 Chongying Wang and Ann Dowker
Some studies suggest that young children tend to use predominantly percep-
tual metaphors, while older children and adults use a relatively larger number
of cross-sensory metaphors and psychological / physical metaphors (Gardner et
al., 1978; Dowker, 2003; Gentner, Falkenhainer & Skorstad, 1988). Children fre-
quently use metaphor in the context of pretend play (Stross, 1975; Billow, 1981),
while adults associate it strongly with poetry, an association that is possibly less
strong for young children (Dowker, 1986; Dowker, Krasowicz, Pinto, Roazzi &
Smith, 1998).
Culture might be expected to influence the ways in which metaphor devel-
ops, as metaphors are known to vary cross-culturally. Gibbs points out in the
first chapter of this volume (4.7) that “Cross-linguistic studies, particularly, have
been very helpful in showing what aspects of metaphor are motivated by univer-
sal aspects of bodily experience and those that are specific to individual cultures
and cultural communities”. However, there have so far been relatively few psy-
chological studies of cross-cultural differences and similarities in metaphor use,
especially regarding development of metaphor.
Basso (1976) regarded metaphor as a crucial concept in understanding the
relation between culture and language:
for it is in metaphor – perhaps more than in any other form of symbolic expres-
sion – that language and culture come together and display their fundamental
inseparability. A theory of one that excludes the other will inevitably do damage
to both. (Basso, 1976: 93)
Some studies have been done on the ways in which differences in metaphoric usage
may reflect cultural differences in attitudes to and concepts of certain emotional,
social and cognitive domains, such as anger (Gibbs, 1994; Kövecses, 2000a), body
parts (Yu, 2000, 2001, 2002), emotion generally (Palmer, Bennett & Stacey, 1999;
Kövecses, 2000b), and time (Zhou & Huang, 2000; Moore, 2001). Such studies
have found considerable commonalities, revealing universal cognitive structures,
but have also indicated some cultural differences. Kövecses (2000a) suggested that
on the basis of linguistic evidence from English, Chinese, Japanese, Hungarian,
Zulu and Wolof, that there are universal aspects of the concepts of anger, but he
also claimed that some of the aspects are culture specific. Yu (1995) compared
the metaphoric expressions for anger and happiness in English and Chinese, and
noted that they are primarily based on common bodily experience, with surface
differences across languages explainable from cultural perspectives. For example,
Chinese tends to utilize more body parts, especially internal organs, than English
in its metaphors for anger, happiness, and other emotional states.
Fewer studies have been carried out on the ways in which language or cul-
ture may affect the actual frequency of metaphor. The objective of this chapter
Chapter 6. A cross-cultural study of metaphoric understanding 107
is to examine the similarities and differences in the understanding and use of

metaphor in Chinese and English children and adults from the perspective of
developmental and cognitive psychology. It was predicted that Chinese children
and adults would interpret metaphors very differently from their English counter-
parts. For instance, Chinese people might prefer to give more explanations based
on conventions, proverbs, idioms, etc. which they have been required to recite a
lot since early childhood. The English education system at least in theory places
more emphasis on original thinking.
2. Experiment 1
Much earlier research suggests that young children tend to use predominantly
perceptual metaphors, while older children and adults use a relatively larger num-
ber of cross-sensory metaphors and psychological / physical metaphors (Gardner
et al., 1978; Dowker, 1986; Gentner et al., 1988). Furthermore, Winner, Rosen-
tiel and Gardner (1976) noted that “cross-sensory metaphors posed less difficulty
than psychological-physical ones”, and “the relative difficulty of the psychologi-
cal-physical metaphors for younger subjects”.
In this experiment, it was predicted that adults would give a higher rate of
psychological interpretations than children for the same metaphors. Children
were expected to give more perceptual interpretations than adults. There were
expected to be more cultural differences between adults than children.
2.1 Method
2.1.1 Participants
115 participants took part in this experiment. They were equally divided ac-
cording to gender at each age level: eight- to eleven-year-old children and
eighteen- to forty-year-old adults (see Table 1). English child participants were
studying at St. Barnabas Primary School in Oxford. English adult participants
were all students at the University of Oxford. All the English participants were
native English speakers. Chinese children were selected at random from the Af-
filiated Primary School of Taiyuan University of Technology, P. R. China. Chi-
nese adults were students at Taiyuan University of Technology, P. R. China, and
had never been to English speaking countries. All the child participants were
selected at random from their classrooms and the adults were all unpaid vol-
unteers in universities. None had participated in any prior studies of figurative
language comprehension.
Table 1. Numbers of participants

Cultural group Age group Number
Chinese Adults 22
Children 27
Total 49
English Adults 37
Children 29
Total 66
Total Adults 59
Children 56
Total 115
2.1.2 Materials
The participants were given a metaphor explanation task in which they were re-
quired to supply their own interpretations of metaphoric sentences. Four meta-
phors were drawn from the literature (Winner et al., 1976; Searle, 1979) and three
were constructed by the researcher.
The authors of this chapter chose to use the decontextualized metaphors as
the stimulus materials because context might give clues about interpretation. For
example, if we add a context to the sentence, The moon was an orange last night,
it will be: It was full moon yesterday. The moon was an orange last night. Then the
explanations would be It was a full moon last night and it was orange in colour,
rather than the psychological explanations in Appendix (see Example 3). Another
example is, Sam is a dog and Sam always follows his boss. He is a dog. (with a con-
text) In this example, the explanations of the metaphoric sentence is obvious, Sam
is faithful to his boss, just like a dog to his master. Or if we add another context,
Sam was born in Dog year. Sam is a dog, the explanation will be totally different.
Therefore, the authors chose the decontextualized metaphors as stimulus material
in order to let participants focus on metaphors while interpreting rather than be-
ing influenced by different contexts which always give clues about explanations.
2.1.3 Procedure
All participants were instructed before they started the experiments that:
This is part of a linguistic study to compare the ways in which Chinese and English
people at different ages interpret different metaphors. There is no single right or
wrong answer to any question. It does not aim at any form of assessment of the
individual participant.
Then, the participants were told:

Sentences sometimes have different meanings and people do not always agree about
what they mean. Please explain the following sentences based on what you think
they mean.
The researcher would either ask the participants to write down their interpreta-
tions or record their explanations. If the participants were too young to write
accurately, the researcher usually asked them to do the task orally; if the child re-
sponded to only part of the sentence or merely repeated the words, the researcher
would require further explanations of the metaphors using nondirective probes.
To illustrate, in the example Sally is a block of ice:
Child Participant A: Sally is cold.
Researcher: What do you mean by ‘cold’? Can you tell me a little more about that?
Child Participant A: Sally is freezing, physically cold.
All the participants were tested individually.
2.1.4 Scoring
Participants’ interpretations were classified as perceptual, psychological, behav-
ioural, functional, no-response, evaluative, descriptive, cross-sensory or associa-
tive. Their interpretation could either be one type or a mixture of several types.
Whenever the researcher was not sure about a particular response, she would
have a discussion with other judges until agreement was reached. The responses
require minor editing. A sample metaphorical sentence together with an example
of each type of response and the definitions or guidelines by which they were
scored is presented in Table 2.
Table 2. Responses to sample metaphors

Name Example Definitions or guidelines for scoring
Perceptual Tom is big and fat. Based on, or involving perception.
Or: Tom is noisy.
Psychological Tom never forgets. Relating to, or arising from the mind or
emotions.
Behavioural Tom moves clumsily. Based on or involving action or behaviour
Functional Tom is a gardener, who Relating to a function.
waters plants the way an
elephant splashes with
its trunk.
Evaluative It is nice. Or: He is bad. Simple evaluation, no exact meaning
Descriptive Tom’s surname is elephant. Involving or characterized by description;
serving to describe. Or concerned with
classification or description.
in the year of the dog.
No response Do not understand. The participants give no response.

*The sample metaphor is: Tom is an elephant.
Table 2. (continued)
Name Example Definitions or guidelines for scoring
Cross-sensory(not in this example, but Different senses were used across. In this
(2.2) Resultsin: The smell of my mother’s example, odour is smelt, while warm is
perfume was bright sunshine.) what people feel; that is, tactual perception
My mother’s odour is very was used to describe olfactory perception.
warm.
Associative (not in this example, but in Of, characterized by, resulting from, or
Paul is a dog.) Paul was born causing association.
Figure 1 shows the frequency of different types of explanations for five
in the year of the dog.
metaphoric
No response Do not understand. The participants give no response.
sentences. It was found that psychological and perceptual explanations are the two
* The sample metaphor is: Tom is an elephant.
main explanations people gave. Therefore, the analyses here focus on these types o
2.2 Results
explanations.
Figure 1 shows the frequency of different types of explanations for five metaphor-
ic sentences. It was found that psychological and perceptual explanations are the
Figure Frequencypeople
two main1:explanations of thegave.
Different Types
Therefore, theofanalyses
Explanations
here focus on these
types of explanations.
Frequency of Different Categories
300 258 261

250
200
150 Total
100 51
50 16 9 14 15
0
Psychological
Behavioural
Associative
No response
Cross-sensory
Perceptual
Evaluative
Figure 1. Frequency of the different types of explanations
For each participant the number of items in each task using each type of meta-
phor was calculated and two-way ANOVAs were carried out with age and culture
For
as the each participant the number of items in each task using each type of metap
factors.
Highly significant effects of age and culture were found for psychological
explanations of metaphors (Figure 2). Chinese people gave more psychological
(F(1,111)=.325, MS=.303, p=.57, ns).
Figure 2: Marginal
explanations Means
than English of (F(1,111)
people Psychological Explanations
= 24.555, MS = 22.892,usedp < by Chinese and
.0001).
Children gave fewer psychological explanations than adults (F(1,111) = 36.778,
MS = 34.287,
English p < .0001).
Children and There
Adultswere no significant effects for age × culture inter-
actions (F(1,111) = .325, MS = .303, p = .57, ns).
3.5
2.5
Chinese
2
English
1.5
0.5
0
Adults Children
Figure 2. Marginal means of psychological explanations used by Chinese

and English children and adults
*Error bars denote the standard error of the mean.
* Error bars denote the standard error of the mean.
Figure 3 illustrates the significant effects of age, culture and age × culture interac-
tions for perceptual explanations of metaphors.
There were3 significant
Figure illustratescultural differences
the significant in the of
effects number of perceptual
age, culture and ageex-× culture
planations given. Chinese adults gave almost no perceptual explanations, but
English children and adults gave many perceptual explanations (F(1,111) =
interactions
9.553, for perceptual
MS = 6.570, explanations
p < .01). An age differenceofwas
metaphors.
also found in the Chinese
group: perceptual explanations were much more frequent among Chinese chil-
dren than Chinese adults (F(1,111) = 10.853, MS = 7.464, p < .001). There was
no significant difference between English adults and children for perceptual ex-
planations of metaphors.
Highly significant age × culture interactions were also found for perceptual
explanations of metaphors (F(1,111) = 14.111, MS = 9.704, p < .0001).
Chinese adults rarely gave perceptual interpretations, Chinese children and
both English children and adults gave far more perceptual interpretations than
Chinese adults.
Figure 3: Marginal Means of Perceptual Explanations used by Chinese and
112 English
Chongying Wang and
Children AnnAdults
and Dowker
2.5
2
Chinese
1.5
English
0.5
0
Adults Children
Figure 3. Marginal means of perceptual explanations used by Chinese

2.3 Discussion
The results showed that Chinese people gave more psychological interpretations
than English people. A typical example was given by an eleven-year-old Chinese
child for the metaphor “The smell of my mother’s perfume was bright sunshine”. She
said that “Mother is always selfless. As an old proverb says, the one who gives will
gain. Since mother has been giving a lot, she, in her child’s mind, is the greatest and
the most blest. Furthermore, mother always gives a hand when her child is helpless.
So child’s feeling to mother is hard to describe. He / She is so familiar with every cell
in mother’s body, even the smell of mother’s perfume. It seems that, smelling mother’s
perfume, child can easily grasp the bright sunshine in the darkness and then find
her/his mother.” By contrast, English people always interpreted it as, “my mother’s
perfume smells pleasant / like flowers / strong”. The differences observed between
people with the two cultural backgrounds may show that certain conventions,
such as proverbs, idioms, famous stories, or those related to traditional (Chinese)
culture and thoughts are more deeply rooted in Chinese people than in English
people. For instance, the above-mentioned Chinese girl first gave a statement,
“mother is always selfless”. “Always” here and in “mother always gives a hand when
her child is helpless” emphasized the general representation of mother in her mind
and perhaps also in her culture. She also referred to an old proverb “the one who
gives will gain” in her explanation. By contrast, English people tended to give an
interpretation based on the sentence itself without referring too much on stories
or idioms. This does not mean that idioms or common expressions are not rooted
in English people, but they may be more deeply rooted in Chinese people. In any
case Chinese people tend more than English to refer to them while interpreting
metaphoric sentences.
A comparison of the performances between the two age groups revealed that
adults tended to give more psychological interpretations than children. This is
in accordance with the findings of Winner, Rosenstiel and Gardner (1976) who
concluded that “the relative difficulty of the psychological-physical metaphors
for younger subjects may have been due to their unfamiliarity with psychologi-
cal domains”.
Chinese adults tended to give fewer perceptual interpretations than either
Chinese children or English children and adults. For example, Chinese adults
tended to understand the metaphoric sentence “the moon was an orange last night”
from the psychological perspective while Chinese children and English people
gave predominantly perceptual interpretations. For instance, a Chinese adult said
“Moon should be pure white. Orange moon implies the writer’s unmerciful mind
or strong emotion.” Chinese children tended to respond “The moon was round
last night.” and English children and adults to respond “The moon was full and
orange.” or “It was so bright. It was orange in colour.” This finding also supported
the hypothesis that there would be more cultural differences between adults than
children in the interpretations of metaphors.
Furthermore, in comparison with children, adults tended to give relatively
conventional interpretations to metaphors that were based on common expres-
sions. For instance, in responding “Sophie is a little cat”, a Chinese adult said
that “Sophie is cute and eager-beaver; loveable, but not faithful. Cat is the animal
who is cute but not faithful or reliable.” while an English adult said that “Little
cats invoke an image of cute, fun, friendly, and playful. Sophie displays a few or
all of these qualities and therefore is like a little cat.” By contrast, children usually
responded that “Sophie is cute and spoiled.” or “Sophie is small and agile.” In this
example we found that adults tended to start with a conventional interpretation
of a cat’s characteristics, associated it with Sophie and then interpreted Sophie’s
features. Thus, in comparison with children, adults tended to give relatively
conventional interpretations to metaphors that were based on common expres-
sions. Children’s interpretations are more cross-culturally similar. For the same
metaphoric sentences “Sophie is a little cat”, both Chinese and English children
tended to respond that “Sophie is cute / sweet / spoiled / small / fluffy / docile /
gentle / softy / cuddly.”
2.4 Conclusion
These findings supported the hypotheses that adults would give a larger number
of psychological interpretations than children; Chinese adults would give more
psychological interpretations than English; and English adults would give more
perceptual interpretations than Chinese adults. These hypotheses were investi-
gated further in Experiment 2 using a different methodology.
3. Experiment 2
Some studies have suggested that children do better in multiple-choice tasks

than in tasks in which they need to state the grounds of metaphor themselves
(Winner, Engel & Gardner, 1980a). Therefore, a multiple-choice experiment was
designed to test people’s understanding of metaphoric sentences. Moreover, such
tests may be easier to control experimentally. It was predicted that there would
be no obvious differences between children and adults in the frequency of psy-
chological interpretations in a task that did not require them to construct the
interpretations themselves.
3.1 Method
3.1.1 Participants
123 participants took part in this experiment. They were equally divided accord-
ing to gender at each age level: eight- to eleven-year-old children and eighteen- to
forty-year-old adults (see Table 3). English child participants were studying at
St. Barnabas Primary School in Oxford. English adult participants were all stu-
dents at the University of Oxford. All the English participants were native English
speakers. Chinese children were selected at random from the Affiliated Primary
School of Taiyuan University of Technology, P. R. China. Chinese adults were
students at Taiyuan University of Technology, P. R. China, and had never been
to English speaking countries. All the child participants were selected at random
from their classrooms and the adults were all unpaid volunteers in universities.
None had participated in any prior studies of figurative language comprehension.
None of them had participated in Experiment 1.
Table 3. Numbers of participants

Cultural group Age group Number
Chinese Adults 30
Children 29
Total 59
English Adults 41
Children 23
Total 64
Total Adults 71
Children 52
Total 123
3.1.2 Materials
The participants were given a multiple-choice task in which each metaphoric
sentence was followed by three to five possible interpretations. The metaphoric
sentences were same as those in Experiment 1. The possible interpretations were
constructed from the results in Experiment 1 while the main responses in differ-
ent categories were chosen (See Appendix).
3.1.3 Procedure
All participants were instructed before they started the experiments that:
This is part of a linguistic study to compare the ways in which Chinese and English
people at different ages interpret different metaphors. There is no single right or
wrong answer to any question. It does not aim at any form of assessment of the
individual participant.
On the basis of participants’ responses in Experiment 1, multiple choice tasks

were designed, in which participants were offered different paraphrases of meta-
phors, or could give their own interpretations of the metaphor if they agreed with
none of the choices. Participants were told:
Sentences sometimes have different meanings and people do not always agree about
what they mean. There are different things that some people think that the sentences
might mean. Please choose which one you think the sentence means. If you agree
with none of them, please write down your own explanations of the sentences.
The participants were all tested individually.

against English.
(3.3) Results
3.1.4 Scoring
The multiple choices for the metaphoric interpretation task were classified in the
same way as in Experiment 1. A sample metaphorical sentence together with an
example of each type of response and the definitions or guidelines by which they
were scored
Figure as presented
4 shows in Table 2.
the frequency ofThe participants’
different responses
categories have not been
of metaphoric explanations. I
corrected for errors against English.
accordance with findings in Experiment 1, the psychological and perceptual
3.3 Results
explanations are the two main explanations.
Figure 4 shows the frequency of different categories of metaphoric explanations.
In accordance with findings in Experiment 1, the psychological and perceptual
Figure 4: Frequency of the Different Types of Explanations
explanations are the two main explanations.
Frequency of Different Categories
350
300
250
200
150 Total

100
50
0
Cross-sensory
Behavioural
Association
Descriptive
Evaluative
Psychological
Perceptual
Figure 4. Frequency of the different types of explanations
In Experiment 1, participants showed an overwhelming preference for psycho-

logical and perceptual interpretations, so analyses focused on these two catego-
ries. Two-way Analyses of Variance were computed as in Experiment 1. Figure 5
shows very significant cultural differences (F(1,119) = 41.162, MS = 34.511, p <
.0001), but there were no significant effect of age (F(1,119) = .443, MS = .371, p =
.5, ns) or age × culture interactions (F(1,119) = .200, MS = .168, p = .655, ns).
The results demonstrated that Chinese people had a greater preference for
psychological explanations than English people (F(1,119) = 41.162, MS = 34.511,
p < .0001), but there was no significant difference between the performance of
children and adults (either English or Chinese).
Figure 5: Marginal Means of Psychological Explanations used by Chinese and
English Children and Chapter

Adults6. A cross-cultural study of metaphoric understanding 117
3.5
3
CROSS-CULTURAL STUDY OF METAPHORIC UNDERSTANDING
2.5
2 Chinese
frequency of perceptual explanations of metaphors but there was no significant age ×
1.5 English
culture
1 interaction (F(1,119)=.269, MS=.107, p=.605, ns).
0.5
English people chose more perceptual explanations than Chinese people
0
(F(1,119)=32.343,
AdultsMS=12.903, p<.0001).
Children Adults chose more perceptual explanations
Figure 5. Marginal means of psychological explanations used by Chinese

than children (F(1,119)=7.683, MS=3.065, p<.005).
Figure 6: Marginal Means of Perceptual Explanations used by Chinese and
Figure 6 illustrates that there were significant effects of both age and culture on
frequency
EnglishofChildren
perceptual
andexplanations
Adults of metaphors but there was no significant
age × culture
Figureinteraction (F(1,119)
6 illustrates = .269,
that there MS significant
were = .107, p = .605, ns).of both age and culture on
effects
2.5
1.5
Chinese
English
1
0.5
Adults Children
Figure 6. Marginal means of perceptual explanations used by Chinese

(3.4) Discussion
English people chose more perceptual explanations than Chinese people (F(1,119) =
32.343, MS = 12.903, p < .0001). Adults chose more perceptual explanations than
children (F(1,119) = 7.683, MS = 3.065, p < .005).
3.4 Discussion
The results supported our findings in Experiment 1 that Chinese people tended to
choose more psychological interpretations than English people. For example, in
responding to “her hair is straw”, fourteen Chinese people chose “She is a raunchy
female and might have lots of worries.” but only one English person selected this
one. Another example is that, almost 50% of Chinese people said that, “The smell
of my mother’s perfume was bright sunshine.” meant that “The smell of mother’s per-
fume makes children feel warm and amiable.” while only six English people among
sixty-four chose this one. By contrast, there was no obvious age difference in the
frequency of psychological interpretations. The explanation may be that children
find easier to understand than to provide psychological explanations.
English people chose more perceptual interpretations than Chinese people.
In the same example “her hair is straw”, English people tended to chose “Her hair
is blonde, dry and messy.” or “Her hair looks like straw.” more than Chinese people.
This experiment also revealed an increasing occurrence of perceptual interpreta-
tions with advancing age. In contrast many other findings, adults tended to chose
more perceptual interpretations than children. For instance, in comparison with
children (82.7%), adults (90.1%) tended to choose more interpretations as “Her
hair is blonde, dry and messy.” or “Her hair looks like straw.” It may be due to dif-
ferences in the nature of the stimuli between different studies.
4. General discussion and conclusion
Results of both Experiments 1 and 2 demonstrated that Chinese people gave

more psychological interpretations than English people, which may indicate that
conventions, such as proverbs, idioms, traditional stories, etc., are more deeply
rooted in Chinese people than in English people. English has a relatively low use
of proverbs compared to that in Chinese culture. The figurative language sensitiv-
ity might be enhanced by such a cultural experience. Similarly the visual iconic
role shaping the understanding of language of the Chinese characters contributes
significantly to metaphoric perception (Hiraga, 1994), e.g. the “guiding hand”
metaphor is well-known in Chinese (and Japanese) as a metaphor of learning.
A comparison of the two experiments also revealed that adults gave more
psychological interpretations than children in the explanation task but there was
no significant age difference regarding psychological interpretations in the mul-

tiple-choice task. The relatively lower frequency of psychological interpretations
in the explanation task given by children may have been due to unfamiliarity
with expressions relating to psychological domains, which made them harder to
retrieve and produce but did not affect comprehension in the same way.
Winner, Engel and Gardner (1980) found that children did better in multiple-
choice tasks than in tasks in which they had to state the grounds of the metaphor
themselves. This may explain why more age differences for psychological inter-
pretations were found in Experiment 1 than in Experiment 2. Children may be
better at understanding / selecting than providing psychological interpretations.
Chinese children gave fewer perceptual explanations than English children in
the multiple-choice task but gave similar numbers of (very slightly more) percep-
tual explanations than English children in the explanation task. This may suggest
a greater tendency by Chinese than English children to give psychological inter-
pretations to metaphors.
Contrary to many other findings (for example, the findings in Experiment 1),
we found that adults tended to choose more perceptual explanations than chil-
dren in Experiment 2, the multiple-choice task.
These results may suggest that “embodied metaphor” interpretations devel-
opmentally precede those that are more culture-specific. This would support the
view that cultural influences on metaphor increase as children get older. This is to
the prediction that cultural differences in metaphor interpretation and use might
increase with age, which could be tested in further studies. It is, however, impor-
tant to exercise caution in drawing conclusions, as these results might be specific
to a particular set of stimuli, and other stimuli might yield different patterns of
performance. Further research and experiments comparing different stimuli are
needed to check this possible explanation.
As well as comparing different stimuli in similar experiments to the present
one, it would also be desirable to carry out a wider variety of types of study to
investigate the extent to which they converge on similar results. Gibbs in the first
chapter of this volume (Section 2) listed four types of metaphor-related cogni-
tion: metaphor interpretation, metaphor processing, metaphor recognition and
metaphor appreciation. The present study dealt with metaphor interpretation. It
would be of great interest to see whether similar cultural differences and similar-
ities were found for the other three types. Moreover, as Gibbs points out, “a real-
world approach to metaphor needs to relate psychological states and processes…
closely to the(ir) actual contexts”. In the future we should supplement experi-
mental studies with observational studies of the contexts in which metaphor is
used in the two cultures, beginning with its origins in very young children’s lan-
guage development.
References
Basso, Keith H. (1976). “Wise words” of the Western Apache: Metaphor and Semantic Theory.
In K. Basso & H. Selby (Eds.), Meaning in Anthropology (pp. 93–121). Albuquerque: Uni-
versity of New Mexico Press.
Billow, Richard M. (1981). Observing spontaneous metaphor in children. Journal of Experi-
mental Child Psychology, 31, 430 – 445.
Caramelli, Nicoletta & Angela Montanari (1995). Animal terms in children’s metaphors. Lund
University, Sweden: Cognitive Science Research, 53, 3–36.
Cometa, Michael & Morris Eson (1978). Logical operations and metaphor interpretation:
A Piagetian model. Child Development, 49, 649–659.
Dowker, Ann (Ed.) (1998). Language play in children. Special issue of Cahiers de Psychologie
Cognitive / Current Psychology of Cognition, 17.
Dowker, Ann (2003). Young children’s and adults’ use of figurative language: How important
are cultural and linguistic influences? In B. Nerlich, Z. Todd, V. Herman & D. Clarke
(Eds.), Polysemy: Flexible Patterns of Meaning in Mind and Language (317–332). Berlin:
Mouton de Gruyter.
Dowker, Ann, Grazyna Krasowicz, Giuliana Pino, Antonio Roazzi and Anita Smith (1998).
Phonological and semantic devices in very young children: A cross-cultural study.
Cahiers de Psychologie Cognitive / Current Psychology of Cognition, 17, 389–416.
Fourment, Marie-Claude, Nelly Emmenecker & Valerie Pantz (1987). Etudes de la Production
de Métaphores chez les Enfants de 3 à 7 ans. Année Psychologique, 87, 535–551.
Gardner, Howard, Ellen Winner, Robin Bechhoffer & Dennie Wolf (1978). The development
of figurative language. In K. E. Nelson (Ed.), Children’s language. Volume 2. (1–37). New
York: Gardner Press.
Gentner, Dedre, Brian Falkenhainer & Janice Skorstad (1988). Viewing metaphor as analogy. In
David Helman (Ed.), Analogical reasoning: Perspectives from artificial intelligence, cognitive
science and philosophy (171–75). Dordrecht: Kluwer Academic.
Gibbs, Raymond W., Jr. (1994). The poetics of mind: Figurative thought, language, and under-
standing. Cambridge, UK: Cambridge University Press.
Hiraga, Masako K. (1994). Diagrams and metaphors: Iconic aspects in language. Journal of
Pragmatics, 22, 5–21.
Kövecses, Zoltán (2000a). The Concept of Anger: Universal or Cultural-Specific? Psychopathol-
ogy, 33, 159–170.
Kövecses, Zoltán (2000b). Metaphor and emotion: Language, culture and body in human feeling.
Press.
Moore, Kevin (2001). Spatial experience and temporal metaphors in Wolof: Point of view, con-
ceptual mapping and linguistic structure. Dissertation Abstracts International, Section A:
Humanities and Social Sciences, 61, 2685.
Palmer, Gary, Heather Bennett & Les Stacey (1999). Bursting with grief, erupting with shame:
A conceptual and grammatical analysis of emotion-tropes in Tagalog. In G. Palmer &
D. Occhi (Eds.), Languages of sentiment: Cultural constructions of emotional substrates
(171–200). Amersterdam: John Benjamins.
Piaget, Jean (1959). The language and thought of the child. London: Routledge and Kegan Paul.
Searle, John (1979). Metaphor. In Ortony, A. (Ed.), Metaphor and thought (92–123). 1st ed.
Stross, Brian (1975). Metaphor in the speech play of Tzetal children. Anthropological Linguis-
tics, 17, 305–323.
Winner, Ellen, Anne Rosenstiel & Howard Gardner (1976). The development of metaphoric
understanding. Developmental Psychology, 12, 289–297.
Winner, Ellen, Matthew Engel, & Howard Gardner (1980). Misunderstanding metaphor:
What’s the problem? Journal of Experimental Child Psychology, 30, 22–32.
Winner, Ellen (1988). The point of words. Cambridge, MA: Harvard University Press.
Winner, Ellen (Ed.) (1995). Developmental perspectives on metaphor: A Special Issue of Meta-
phor and Symbolic Activity, 10, (4).
Yu, Ning (1995). Metaphorical expressions of anger and happiness in English and Chinese.
Metaphor and Symbolic Activity, 10, 59–92.
Yu, Ning (2000). Figurative uses of finger and palm in Chinese and English. Metaphor and
Symbol, 15, 159–175.
Yu, Ning (2001). What does our face mean to us? Pragmatics and Cognition, 9, 1–36.
Yu, Ning (2002). Body and emotion: Body parts in Chinese expression of emotion. In N. Enfield
& A. Wierzbicka (Eds.), special issue “The body in description of emotion: cross-linguistic
studies,” Pragmatics and Cognition, 10, 341–367.
Zhou, Rong & Xiting Huang (2000). A cross-cultural study of the metaphorical representation
of time. Psychological Science – China, 23, 141–145.
Appendix
Selected metaphors in the multiple-choice task
1. The smell of my mother’s perfume was bright sunshine.
(1) Mother’s perfume is very bright. (Perceptual)
(2) The smell of mother’s perfume makes children feel warm and amiable. (Psychological)
(3) My mother’s perfume smelled warm, bright, light, fresh and sunny – like summer.
(Cross-sensory)
(4) The fragrance of the perfume was pleasant. (Evaluative)
2. Her hair is straw.
(1) Her hair is blonde, dry and messy. (Perceptual)
(2) She is a raunchy female and might have lots of worries. (Psychological)
(3) She is a farmer. (Association)
(4) Her hair looks like straw. (Repetition)
3. The moon was an orange last night.
(1) It was a full moon last night and it was orange in colour. (Perceptual)
(2) It means that the writer has a very good mood and has a good memory of the past
times, so s/he describes it as round and big orange. (Psychological)
(3) It has two meanings. Firstly it means that last night’s moon was round and bright;
secondly it means that the writer has good mood. (Perceptual and psychological)
(4) The moon was pleasant last night. (Evaluative)
4. Sally is a block of ice.

(1) Sally is cold-hearted, unemotional and hard to get close to. (Psychological)
(2) Sally is freezing. (Perceptual)
(3) Sally is an ice Queen. (Psychological)
(4) Sally is still. (Behavioural)
5. Sophie is a little cat.
(1) Sophie is cute, sweet, cuddly and playful. (Psychological)
(2) Sophie walks quietly and talks softly. (Behavioural)
(3) Sophie is small. (Physical)
(4) Sophie is furry. (Perceptual)
section 2
Capturing and analysing metaphors

chapter 7
Love, metaphor and responsibility

Some examples from Early Modern
and Present-Day English corpora
Heli Tissari
University of Helsinki
Diachronic corpus linguistics can help us trace changes in the usage of meta-
phors within various stages of one and the same language. This article deals with
conceptual metaphors occurring with the verb and noun love in Early Modern
and Present-Day English. The aim is to describe the seeming paradox that while
frequency counts based on corpus data yield similar metaphors for both peri-
ods, reading the data, one cannot help thinking that there are important cultural
differences between these periods. A close reading of selected examples suggests
that the conceptual metaphors love is a (fluid in a) container and love is
a valuable commodity/economic exchange are so general that they can
be used for many arguments and purposes which may even conflict with each
other. The concept of responsibility is invoked in this article as a means of show-
ing how this works.1
Keywords: cultural differences, diachronic corpus linguistics, emotion, love,

responsibility
1. Introduction
1.1 Using large-scale data for metaphor research
In his plenary talk at RaAM6, Gibbs (in this volume) emphasized both the im-
portance of using corpus data and of explicating one’s method of identifying
the metaphors in it. The modest aim of this chapter is to illustrate the complex-
ity of the linguistic data which we retrieve from corpora and the fact that, like
metaphors, our methods both highlight and hide aspects of the abstract world
126 Heli Tissari
(cf. Lakoff & Johnson, 1980: 10–13). The article will focus on the metaphors love
is a (fluid in a) container and love is a valuable commodity / economic
exchange in particular.
My work on the semantics and conceptual metaphors of love (Tissari, 2003),
which inspired this further article, dealt with metaphors combining with the Eng-
lish word love. The data came from several corpora representing Late Middle,
Early Modern and Present-Day English. Stefanowitsch (2006) did something
similar with several emotion words (anger, fear, happiness, sadness, disgust), using
the BNC to retrieve what he calls metaphorical patterns. These are multi-word
expressions which attest, for example, the noun anger, a specific lexical item from
the target domain (TD) anger, and some word(s) yielding a source domain (SD),
such as hot fluid in a container (e.g., X boil/seethe with anger, Stefanowitsch,
2006: 66–67, 74). Koller suggests that
… metaphor researchers can, and indeed should, make the best of what com-
puter-aided analysis has to offer. Combining the social and the cognitive in an in-
terdisciplinary fashion is best done by looking at data on a large scale, and corpus
analysis is a promising means to this increasingly important end.
(2006: 254, italics added)
Large-scale data nevertheless presents many challenges to the analyst. In spite of

her recommending large-scale corpus studies, Koller also points out that “… the
large amounts of data dealt with in corpus studies are anathema to the idea of
meticulously going through the motions of metaphor identification in each and
every instance” (2006: 242). Stefanowitsch (2006) does not discuss what would
happen if, in addition to the nouns, he also looked at verbs, adjectives, and ad-
verbs (cf. Deignan, 2005: 144–167; Deignan, 2006; Goatly, 1997: 82–106).
Tissari (2003: 325–422, 428–433) attempted to generalize on changes in met-
aphors between Early Modern and Present-Day English, instead of paying too
much attention to a single text or writer. The main idea was to classify occurrenc-
es of conceptual metaphors in order to see what types were likely to occur, partly
depending on kinds of love. Most of the types were found in previous cogni-
tive linguistic literature, which seemed to confirm the accuracy of the theory. The
study was rewarding as an exercise in cognitive corpus linguistics and historical
semantics. It was nevertheless clear that the data could be seen from other angles
as well. This article provides one such angle.
Chapter 7. Love, metaphor and responsibility 127
1.2 Background for the present article
The data
The examples discussed below were originally retrieved from two diachronic (Late
Middle and Early Modern English) and two synchronic corpora (1991), the Early
English Correspondence Sampler (1417–1681), the Early Modern English period
of the Helsinki Corpus of English Texts (1500–1710), the Freiburg-Brown Corpus of
American English (ca. 1991), and the Freiburg-LOB Corpus of British English (ca.
1991), by searching for the verb and noun love and other words derived from them,
resulting in a data set of 2,296 hits (Tissari, 2003: 330).2 The original analysis sug-
gested that the verb and noun love most often occurred in expressions conveying
the metaphors love is a valuable commodity / economic exchange, or love
is a (fluid in a) container, regardless of period (Tissari, 2003: 380). The results
concerning various kinds of loves seemed to agree with Stefanowitsch’s sugges-
tion that “given a large enough corpus, all metaphors will be instantiated for all
emotions” (2006: 91), taking into account nevertheless that some metaphors are
more typical of certain emotions or certain kinds of loves, such as romantic love
or friendship (Tissari, 2003: 367–369, 397–399).3
However, while the rough frequencies presented in Tissari (2003: 353–387)
did not reveal great differences between Early Modern and Present-Day English,
there still seemed to be genuine differences between data from the two periods,
even as regards metaphors. If the main metaphors of love remained the same,
what made the data from these periods so different?
To shed more light on this issue, I re-read part of the original data more
closely, retyping expressions which suggested responsibility connected with
love, both metaphorical and non-metaphorical, with as much co-text as seemed
necessary for a better understanding of the moral values associated with love.
The subset of the data which was analyzed for this article consisted of the type
love in each corpus, totalling 1,260 items, and produced a number of metaphors
not discussed below.4 The selection of examples for this article was based on
three criteria: (1) Each of them attests either the metaphor love is a valuable
commodity / economic exchange or the metaphor love is a (fluid in a)
container, or both. (2) Each seemed to illustrate something important about
the period in question. (3) Each seemed to illustrate a combination, even conflict
of values with respect to the concept of love. The second and third criteria are
clearly more subjective than the first, and the discussion of the examples below
is qualitative rather than quantitative.
128 Heli Tissari
1.3 Love and responsibility
The first sense given to the noun love in the Oxford English Dictionary (OED) runs
as follows:
That disposition or state of feeling with regard to a person which (arising from
recognition of attractive qualities, from instincts of natural relationship, or from
sympathy) manifests itself in solicitude for the welfare of the object, and usually
also in delight in his or her presence and desire for his or her approval; warm af-
fection, attachment.
Some of the words in the definition emphasize that love is a feeling, or an emo-
tion: feeling, sympathy, delight, desire, warm affection, attachment (note also the
possible metaphorical readings of warm affection and attachment, love is fire/
heat (Kövecses, 1988: 44–47, 64) and love is a unity of two complementary
parts (ibid. 18, 20) / love is a magnetic force (ibid. 50), but others suggest
that love is also something else, a disposition, and perhaps even action, because
one’s solicitude for the welfare of the object is likely to involve choices concerning
one’s verbal and physical behaviour. This is important because it concerns the
relationship between love, emotion and other concepts, i.e. placing love in a
larger conceptual framework.
That love is about such choices is a commonplace in academic discourse on
love. In a recent conference on “Emotions, Others and the Self ”, people discussed
love and virtue (Neiders, 2005), love and respect (Larson, 2005), and various oth-
er aspects of how people “use” their emotions, to quote Solomon (2005: 6). In a
plenary talk, Solomon summarized it all by saying that “love is a whole series of
decisions”. A recently published dissertation on love as a literary theme treated
the ethics of love between the sexes from a feminist point of view, claiming that
even our research into linguistic expressions of love can attest loving qualities
(Korsisaari, 2006).5
One sense that the OED gives to the adjective responsible is “morally ac-
countable for one’s actions; capable of rational conduct” (3.b.). Most academic
discourses on how to understand and employ human emotions in one way or
other presuppose that human beings are responsible for their emotional lives. For
example, Solomon (1994: 333) suggests that “[t]he waning of love can be attrib-
uted, in many cases, to lack of attention – lack of attention to the other person,
lack of attention to the relationship itself.”
In the following, we will discuss selections of text and what they have to say
about moral accountability. Each of these texts attests the verb and/or noun love
and either the metaphor love is a valuable commodity/love is an economic
exchange, or love is a container, or both. Lakoff and Johnson (1980: 30–32)
include the sentence “He’s in love” among their examples of metaphors of contain-
ment. Kövecses (1990: 144–159) refines the description of container metaphors
for emotions, focusing on the “body container”, which is affected by the increase
in the fluid in it, as in “She was filled with emotion” (ibid. 146). Kövecses (1986: 95)
introduces the metaphor love is a valuable commodity (in an economic ex-
change), for example “I gave her all my love”. Later, he also talks about the love
is an economic exchange metaphor (Kövecses, 2002: 46). This metaphor im-
plies reciprocity and thus also moral accountability, but the metaphor love is a
container is apparently quite neutral as regards moral or ethical choice, or at
least seems to require a very context-dependent reading.
This article distinguishes between several types of responsibility: for one’s self,
what one does, other people, and what one thinks and feels. This division is not
meant to be analytic as much as practical, used for the sake of the argument.6 The
aim is twofold; firstly, to show how an analysis of even a few lines of text instead
of a sentence (like “He’s in love”) adds to our understanding of the valuable
commodity or container in question and, secondly, to show how important
the moral dimension can be for understanding the data. In previous research on
the conceptual metaphors of emotion, responsibility has mainly been addressed
in terms of controlling the body which is the container for emotions (e.g., Lakoff,
1987: 380–415; Kövecses, 1990: 146–151).
2. Examples from the data
2.1 Responsibility for one’s self: The philosophical example
(1) And this loue or appetyte that euerye thynge hathe to it selfe, procedeth not,
nor cometh of the mocion of the soule: but by naturall intencion. For the
prouydence, or wysdom of God, hath giuen vnto thynges that he hath creat
this, that is to saye: a great cause to contynewe styll, in as moche as they
desyre naturally to lyue as longe as they may.
(The Early Modern English period of the
Helsinki Corpus of English Texts: Boethius’ Consolation of Philosophy,
translated by George Colville, 1556: 80–81)
This excerpt suggests that there is something that makes people take care of them-
selves: their love, or appetite, for themselves. The verb form hathe suggests posses-
sion, which receives a positive evaluation, since the ultimate “giver”, mentioned
in the next sentence, is God. In other words, this example represents the favourite
metaphor love is a valuable commodity.
130 Heli Tissari
This text is not overtly about responsibility, but the words naturall intencion
and desyre naturally are normative, especially when coupled with wysdom of God.
This text in fact says that God has put a desire to continue their lives even in the
face of adversity in all living beings (euerye thynge), and that this desire is a good
thing. Conversely, it would be bad not to desire to live. Life itself, then, is valu-
able, a gift a person receives from God and that one is likely to be responsible
for maintaining. The metaphor love is a valuable commodity does not itself
suggest that love is so crucial to our lives, but should be rephrased as love is an
indispensable commodity.
However, it is not any love, but ‘love of self ’, that is so indispensable. This is
interesting, because in other Early Modern English texts, loving one’s self may be
roundly condemned. A book entitled Practical Rhetoric includes the ideas that
“Love overcometh all things” and “Self-love is blind”, which the student of rheto-
ric is taught to express in various ways (Poole, 1972 [1663]). Not only do we see
then that love of self is indispensable for one’s life, but also that there are likely to
be two kinds of ‘love of self ’, one of which is recommendable, and another which
is to be condemned.7 If this is the framework for ‘love of self ’ in Early Modern
English society, then it concerns one’s responsibility for distinguishing between
these two and acting accordingly, taking care of one’s own life, but not loving one-
self too much. This is an inbuilt tension in the very concept of ‘love of self ’, and
merely saying that the metaphor self-love is a valuable / an indispensable
commodity applies to Early Modern English would distort the picture quite as
much as saying that self-love was then always considered a bad thing.
To obtain meaningful statistics for “good” and “bad” love of self is difficult,
because few hits for love in the data set discussed by Tissari (2003) concern ‘love
of self ’, which is consequently hardly discussed at all. This concept is likely to be
more salient than its linguistic frequency might suggest.8
Furthermore, this example suggests that the metaphors for each pole of an
emotion concept may prove to be asymmetrical on a positive-negative axis. It is
possible that the valuable commodity metaphor indeed was the favorite op-
tion when Early Modern speakers of English discussed “good” love of self, but
the metaphor love is blind was equally the favorite option in discussing “bad”
love of self. The image of the blind Cupid was a cultural topos, and reason, which
tends to be associated with sight, was often recommended for controlling one’s
emotions. Note that the very definition of responsible quoted above involved the
capacity for “rational conduct” (OED, a.3.b.).
Note also that while the discussion concerning one’s responsibility to love
oneself still continues in Present-Day English, looking for corresponding passag-
es in our contemporary texts should perhaps involve discussions of whether or
not a person can have the right to wish to die. This would take us from the realm
of love into the realm of rights, the understanding of which has also changed dur-
ing the past five hundred years.
2.2 Responsibility for what one does: The court house
(2) That he wished the E. to flye wth 2 or 3 gent. and for his owne goinge to the
consultacon at Drewrie house, he was drawne there vnto, by the love he bare
to the E. of South: to whome he ought his lyfe.
(HC: The Trial of the Earl of Essex: 13)
This brief quote again illustrates a tension between notions of good and bad. Here,
the Earl of Essex (Robert Devereux) is accused of treason, and the “consultacon at
Drewrie house” is something he should not have participated in.9 The metaphors
he uses in the indirect quote to defend himself against these charges may actually
reveal that he was himself hesitant about whether he had done the right thing.
The love is a magnetic force metaphor he was drawne there vnto suggests that
it was not entirely in his control to make the choice, which allows room both for
concluding that it was the right choice, or that it was an error. The next clause, by
the love he bare to the E. of South(ampton, Henry Wriothesley), differs by not be-
ing ambiguous.10 The love is a (gravitational) force/burden metaphor there
(to bear love to somebody) can also be interpreted as another love is a valuable
commodity metaphor, because the verb to bear (to carry, to hold) implies posses-
sion as well as action. The valuable commodity metaphor underlines that even
if the consultation was not the legitimate thing to do, it was motivated by some-
thing that was certainly legitimate. In effect, the whole quote suggests that the Earl
simply did his duty. The final metaphor, with its appendix, to whome he ought his
lyfe, downplays, even attempts to erase, the idea that the Earl had lost his ability to
behave in a sound and rational manner, while it also conveys the idea that love is
something that cannot always be controlled, and may control the lover.
Kövecses (1988: 49–55) talks about the uncontrollability of romantic love in
a chapter entitled “Passivity, lack of control, pleasantness”. This chapter discuss-
es “the presence of responsibility and choice in love” (ibid. 51). It includes what
could be read as a commentary on Example (2) (ibid. 53): “ … a person who
acts upon an emotion is considered less blameworthy or answerable for an ac-
tion than a person not acting on an emotion.” However, Kövecses also notes that
not all such actions on an emotion are evaluated similarly. There is a difference
between acting in love and acting in anger (1988: 53). There is even a difference
between acting on love that is evaluated negatively as against acting on love that
132 Heli Tissari
is evaluated positively. Love and friendship among men who were responsible
for and to each other tended to be seen positively in early modern England. The
phrase to whome he ought his lyfe is certainly supposed to mark the Earl of Essex’s
love for the Earl of South(hampton) as “good and respectable love”. However, its
force is diminished by the fact that the two earls were accused of conspiring to-
gether against the crown.
Example (2) diverges from the tendency of the Early Modern English data in
general to only foreground responsibility by suggesting that love as a force may
overrule what would otherwise be considered the right thing to do. There are
more hits which suggest that this tension is a question in the air, but the Early
Modern English writers usually emphasize somehow that they are acting respon-
sibly and that others should also behave likewise. This Early Modern English
tendency goes against the Present-Day English tendency to value the experience
of love in itself, especially romantic love. The metaphor love is a valuable
commodity receives different readings according to the value system(s) applied.
It is more important for Early Modern English writers to persuade their readers
that their love follows generally accepted moral standards than that it is expe-
rienced as pleasant. The forces which people employ in metaphors of love in
Early Modern as against Present-Day English are consequently different as well
(cf. Kövecses, 1988: 49–55).
Tissari (2003: 353–387) includes the love is a (gravitational) force/
burden metaphor (to bear love to somebody) among other force metaphors,
which downplays the difference between the data from our two periods. This met-
aphor potentially combines two ideas; that love can be a heavy duty, and that it is
nevertheless a good thing (a valuable commodity). Metaphors expressing the
idea that love can be a heavy duty hardly occur in the Present-Day English data.
Note, however, that this difference between specific types of metaphor has to be
spelled out. It would not help to call the Early Modern English metaphorical pat-
tern to bear love to somebody a valuable commodity metaphor, because these
abound both in the Early Modern and Present-Day English data. Moreover, a
speaker of Present-Day English might well say something like “The responsibility
weighed heavily on me” in discussing a responsibility which could be subsumed
under love. According to the OED, the noun responsibility only begins to occur in
eighteenth- and nineteenth-century English.
2.3 Responsibility for other people
Example 3: Showing love for one’s parents

(3) Sir
I cannot exprese the thankes I owe to so good a Father, first in showing my
fault so rivelie, and then in forgiving my fault so frilie: albeit I cannot love
and honnor youre Matie more then I did, yet this shall learne me heerafter not
to be so foulish as to sende anie sutche message as shall beare anie comment
but my owen, which shall be as full of love and respect as a dewtifull sone can
owe to so loving a Father, thinking I shall prosper no longer then I deserve to
be cald
Your Mties most humble and obedient sone and servant
Charles.
(The Corpus of Early English Correspondence Sampler:
1613 CHARLES I: 102)
The third and last Early Modern English example further underlines the value
system with which the Early Modern English metaphors of love are usually as-
sociated, exemplifying both a metaphor of containment (which shall be as full of
love and respect), and the metaphor love is an economic exchange / a valu-
able commodity (owe). Love is something Charles is expected to express (to
love, honour and respect his father) and which his letters to his father are expected
to contain. It involves Charles owing his father other things as well (thankes), and
hierarchical reciprocity, in which the father is prior to the son. Beside love, the
value system expressed in this letter emphasizes obedience (obedient sone), hu-
mility (most humble), and duty (dewtifull sone, owe). The address form so good a
Father explicitly assigns the letter to a framework of moral values.11
More specifically, the address form so good a Father can be seen to involve a ref-
erence both to Charles’s father as a representative of the category of fathers in gen-
eral and as a representative of “fathers (=kings) of the country” and, on a religious
plane, both sets of fathers can be seen to convey some of the love which derives from
“our Father in Heaven”. The letter is thus situated within a rich network of associa-
tions, beliefs, and social rules, reflecting the idea that in the world everybody except
God is subject to a superior (the “Great Chain of Being”, discussed by Lovejoy, 1978
[1936], Tillyard 1973 [1942]: 25–82, and Musolff, 2005, 2006, among others).12
The letter is especially interesting as regards the theory of conceptual meta-
phor, because it is both a metaphorical and a literal “container” for love. It contains
Charles’s writing which contains words which incorporate the love which Charles
is expected to show towards his father. A letter can also be seen as a “commodity”
which one exchanges with somebody else, since the nature of correspondence is
134 Heli Tissari
bi- or even multilateral. More generally, the nature of talk between people who love
each other is ideally reciprocal, especially as concerns their expressions of love for
each other. Consequently, the metaphor love is an economic exchange can be
seen to characterize discourse between lovers in a very general sense. Investigat-
ing this further should also take into account metonymy as an intermediary stage
between the physical aspects of people’s expressions of love (for example, letters)
and the “abstract” concept of love – it would actually be very difficult if not im-
possible to conceptualize love without ever having experienced any of its physical
manifestations.13
A question which may arise is the extent to which Charles’s letter is formulaic
as against spontaneous, or conventional as against innovative, but although this
question is of interest in itself, it does not necessarily influence the interpretation
of the letter as a carrier of the metaphors love is an economic exchange / a
valuable commodity and the letter is a container for love. The question
rather concerns the way in which these metaphors exist, the extent to which they
are social constructions, and how far they belong to the idiolect of particular writ-
ers (or speakers). These concerns in their turn can be distinguished from further
possible concerns about the writer’s intentions, emotions, or honesty.
Example 4: The sexes

(4) Drumble is a large manufacturing town, concerned with business and with
money, both of which are here defined as exclusively male domains. We
know that money in Gaskell’s mind is associated with love. Here we learn
that men and women have different relationships to the image. Women
need love. Men have it to give. And since the possession of love is male, so is
the possession of money. The Cranford ladies know nothing of money.
(The Freiburg-Brown Corpus of American English:
Bonaparte, 1992: The Gypsy-Bachelor of Manchester.
The Life of Mrs. Gaskell’s Demon: 7)
Example (4) shows that writers of Present-Day English are often quite conscious
of the metaphors which they use (image). In the Present-Day English data, the
hits for love often occur in texts which deal with literary authors or texts (in the
“Belles Lettres, Biographies, Essays” category), or with the psychology of love. It is
thus a mistake to think of each of these metaphors as representative of metaphors
which people “live by”, hardly noticing them (cf. Lakoff & Johnson, 1980). A com-
monplace metaphor such as the love is an economic exchange / a valuable
commodity metaphor can be very conscious, something cultivated in academia
and by people with a university education and/or intellectual ambition. It can be
used to take a stand when discussing women’s rights, for example.
A discussion of love as an economic exchange could also appeal to people

who have read their Marx on the distribution of capital. This passage indeed em-
phasizes the difference between men and women as the givers and recipients of
love respectively, focusing on the material basis of it all.14 This involves a defini-
tion of love in terms of who is in control, instead of an emphasis on love as an
emotion. Some men may be less equal than others, but why is it that women are
so unequal as regards both love and money? Note that the concept of exchange
could be seen as neutral in itself when addressing these questions, and should it
be stated that women are responsible for changing such circumstances. Their re-
sponsibility cannot necessarily be seen as ensuing from their love for the opposite
sex, but rather as affecting it and as ensuing from their love for other women.
Comparing Examples (2) and (3) with Example (4), we see what appears to
be a shift from duty-based codes of behaviour to right-based codes of behaviour,
defined by Shweder and Miller (1985: 51) as follows:
In a rights-based ethical code, a right such as, “the right of all men to greatest
possible overall liberty” is taken as fundamental. In a duty-based ethical code, a
duty such as “the duty to obey God’s will as set forth in the Ten Commandments”,
is taken as fundamental.15
Early Modern English texts prefer an emphasis on people’s duties, while Present-
Day English texts often reflect the idea that people have certain individual rights
such as the right to possession, the right to choose their partner, or the right to
quit a relationship which does not work. These rights are often so self-evident
that they are not even mentioned.16 It is also common to discuss them as given in
order to raise more specific issues.
2.4 Responsibility for what one thinks and feels
Example 5: The closet

(5) There were other parts of me that I let go too. I love jazz clubs, but Daniel
hates them, so we never went. I love the mountains, but Daniel loves the
ocean, so we spent vacations in the Caribbean. Daniel never demanded that
I give up anything, but it seemed easier that way. We couldn’t afford a piano;
the Caribbean was cheaper; we couldn’t find a baby-sitter. I was his ‘wife’ and
it ended there. Daniel lied, too. He was my ‘husband,’ and he defined his life
as narrowly as I defined mine. We closed each other up in a closet of ‘love’
that nearly smothered us both. There was no spontaneity in our love – no
authenticity. So when I found out about Daniel’s affair with Shelia, things
changed forever between us. Maybe they couldn’t have gotten any worse.
(FROWN: Payne-Robinson, 1992: ‘Infidelity: A Love Story’, Essence: 18)
136 Heli Tissari
Example (5) includes the phrase a closet of ‘love’ which suggests containment, but
in contrast to the often positive connotations of such phrases denoting contain-
ment as to fall in love and to be in love, this closet of ‘love’ is considered fatal. A
small room may onset claustrophobia, but falling, for example, can be considered
risky as well. It is therefore necessary to look at the context to see what makes the
closet a bad place instead of a cosy or safe one in this case.
If one reads this passage as suggesting something about responsibility, then
both Daniel and his wife seem to have been responsible for the hollowness of
their relationship even before one became unfaithful to the other. Their mistake
was that they did not respect, express or act out their real emotions, but let fake
emotions constrain them to the point of suffocation.
If one considers Daniel and his wife’s bodies as containers, then these contain-
ers in such a relationship appear to accumulate but not vent bad emotions, which
can then only be expected to explode one way or another (Lakoff, 1987: 380–415;
Kövecses, 1990: 184–185). Alternatively, one may think of the fake emotions as
something that tightens around their bodies until they can no longer breathe.
Both images can very well be seen to apply simultaneously, so that both the inner
and outer pressures grow to a point where their joint effect is fatal to the couple
and their mutual love.
Several explanations suggest themselves for what makes a metaphor of con-
tainment so useful here. Firstly, one might say that experiences of containment,
both pleasant and unpleasant, are fundamental to our experience, beginning
from the womb and the mother’s arms, which sometimes may also well have pre-
vented us from moving towards a place where we wanted to be.17 Secondly, it
is a commonplace in texts about love that mere form (which can to certain ex-
tent be equated with a container) is not considered enough. People search for
genuine love (usually considered as existing inside the body as the container
for emotions, somewhere deeper than the surface, as suggested by Kövecses
(1990: 151–152)). The cultural differences between the early modern and present-
day world-views, for example, do not exclude this tendency.
Speakers of Present-Day English may nevertheless be more accustomed to
thinking in terms of such containment than their ancestors, because they are so
used to psychological and (psycho)therapeutic jargon. It would be interesting to
see how far this explanation holds, since such jargon must have its origins.
It is also a trait of the novel, a genre not yet established in the sixteenth and sev-
enteenth centuries, to have protagonists scrutinize their emotions and relationships
in order to understand their ensuing decisions. It is very difficult to determine the
degree to which such literary models affect the way people think and/or generate
language of a similar kind, because a diachronic analysis should also take into ac-
count both the growth in people’s literacy and the number of works in print.
Example 6: The hypnosis

(6) For some people, being in love has the disturbing side-effect of
anaesthetizing the rational mind to an extent where the person floats
through life on autopilot, hypnotized by the object of their desire. They do
not really see the other person, they only see what they want to see; they are
in love with a figment of their imagination. They will neither see nor admit
that their love is one-sided because they automatically ignore any signals
that do not fit into their concept of adoration. Even when incompatibilities
are pointed out to them, they will sweep them under the carpet or
reinterpret them into an acceptable version, finding all sorts of excuses to
explain unpleasant behaviour with which their partner confronts them.
This is a bit like doing a jigsaw puzzle and cutting the pieces with a pair of
scissors to make them fit – you will never get the true picture.
(FLOB: Pfeiffer, 1991: How to Cope with Splitting Up: 10)
This is another “negative” example of the container metaphor, concerning the

phrase (being) in love. Example (6) actually combines several metaphors, the main
ones being love is a container, love is a unity of two complementary
parts (Kövecses, 1986: 66), and understanding is seeing / intellection is vi-
sion (Jäkel, 1995: 225–226; Koivisto-Alanko & Tissari, 2006: 192–193; Sweetser,
1990: 37–40). The message seems to be that people looking for a partner are re-
sponsible for using their reason in order to choose well, to find a (truly) comple-
mentary part for themselves. They should not be guided simply by their emo-
tions, but be able to dismiss them in a case of “misfit”. In contrast to Example (5),
which pointed out a problem in not listening to one’s emotions, this passage thus
criticizes people who live by their emotions too much.
That one should let reason control one’s emotions is not a new idea, but one
that has been repeatedly reformulated by philosophers since ancient times and
was a commonplace during the early modern period as well. Consider, for ex-
ample, how Shakespeare’s comedies make us laugh at people’s follies in love. Re-
member also that reason was relevant for understanding Examples (1) and (2).
Self-love could be blind; reason could be misled by a debt of love.
Comparing Example (6) with Example (1), we see that both deal with a bal-
ance. Example (1) suggests that one should be able to love oneself in order to take
care of oneself, but not to love oneself too much. In its own way, Example (6) talks
about a very similar issue: one should allow oneself to be happy, but not at the
cost of future unhappiness. Example (6) also shares a theme with Example (2),
the power of emotions to move us towards both good and bad. The difference
from Example (3) is that while the obedient son needs to “make the pieces fit” by
apologizing to his father, Example (6) suggests that there is a limit to how much
138 Heli Tissari
one should work for such a fit, although in a different kind of relationship. While
people who are looking for a partner have the right to find one and be happy, they
do not have the right to tell their partner to change and adjust themselves beyond
a certain limit. Moreover, the partners found have a right to quit a relationship
which in their view does not work.
3. Conclusions
As obvious as it may seem, it is quite relevant with regard to diachronic and cor-
pus studies to point out that people may use the very same metaphors for op-
posite ends, such as saying that something is either good or bad. For example,
we have seen above that the metaphor love is a valuable commodity in an
economic exchange can be used both to emphasize the value of love and to
criticize people who control the exchange of love. Similarly, metaphors of con-
tainment can be used both to enhance the value and emphasize the amount of
love, and to suggest that love is something which suffocates a partner in a bad
relationship. This means that even if it is desirable to gather statistical informa-
tion about the actual usage of metaphors, figures suggesting the frequency of a
certain metaphor only tell part of the story. It is therefore also useful to approach
the metaphors from several, even cross-disciplinary angles, as suggested by Gibbs
(in this volume), among others.18 One way to better understand the metaphors
of love in Early Modern and Present-Day English is to see them as representing
a duty-based code of behaviour and a right-based code of behaviour (Dworkin,
1987 [1977]: 171–173; Shweder & Miller, 1985: 51) respectively; but again, while
this may help us understand the moral values of the metaphor users, it does not
predict when a particular metaphor is likely to show up.
To sum up, the six examples discussed above also suggest, or remind us of,
the following:
1. An infrequent mention of a concept does not mean that the concept is unim-
portant.
2. Metaphors for good and bad things need not be symmetrical.
3. The same emotion can be seen both as a good and a bad thing.
4. What can be thought of as an emotion can also be regarded as an attitude,
disposition, duty, responsibility, etc., depending on the context. This affects
the way the concept is evaluated.
5. Emotions can be contained not only in the human body but also in language
and its various manifestations.
6. Language is a means of exchanging love and other emotions.

7. The way we understand our emotions is closely tied to the way we understand
our position in the world as individuals / human beings.
Lakoff says in an interview that “moral systems are very complex” (Pires de
Oliveira, 2001: 35). I take this to mean that explaining moral systems using the
theory of conceptual metaphor produces a very complex explanation, as appar-
ently happens if we wish to deal with a great amount of data and to both provide
usage statistics on a number of metaphors and explain their use in various con-
texts. This is a fascinating pursuit in itself, because there appears to be no end to
the beautiful detail one can capture. However, this does not help us derive “gen-
eral principles governing all aspects of human language” (Lakoff, 1990: 40), unless
we take complexity to be such a principle, or unless we arrive at abstractions other
than the conceptual metaphors already identified.
Many conceptual metaphors listed, for example, by Lakoff and Johnson
(1980), seem to serve as good generalizations in that one can find repeated occur-
rences of them in linguistic data. However, it may lead to the mistake of consider-
ing them as somehow neutral parts of a system which keeps regenerating itself
with little purpose other than continued existence. Although science requires
detachment from the phenomena it studies, language as communication tends
to serve more specific aims as well, and while the values and evaluations it car-
ries may sometimes be concealed even from the language users themselves, they
nevertheless exist and tend to direct the way we think, speak and write. Studies
inspired by critical discourse analysis, such as Koller (2004), focus on this, but the
more one focuses on the categorization of metaphors per se, the greater the risk of
neglecting other issues. The commonplace emphasis on the experiential basis of
metaphors in cognitive linguistics (see, e.g., Evans & Green, 2006: 163–165, 309;
Johnson, 1987) may blind us to the fact that while experience shapes our beliefs
and values, our beliefs and values also shape our experience.
Notes
1. While writing this paper, the author was a member of the VARIENG Research Unit for
Variation, Contacts and Change in English at the Department of English, University of Hel-
sinki, and a research fellow at the Helsinki Collegium for Advanced Studies.
2. For more information on the corpora, see the manuals (Hundt, Sand & Siemund, 1998;
Hundt, Sand & Skandera, 1999; Kytö, 1996; Nurmi, 1998).
3. To be specific, the finding which concerns friendship involves the lexicon of love, as col-
lected by Coleman (1999).
140 Heli Tissari
4. This total consists of 292 hits for the Helsinki Corpus, 417 for the Corpus of Early English
Correspondence Sampler, 301 for the Freiburg-Brown Corpus, and 250 for the Freiburg-LOB
Corpus. The spelling variant loue in the early data was also included. My purpose in using such
a “random” sample was to gain a new understanding of the data by focusing on detail.
5. Korsisaari’s research (2006) was inspired, for example, by Luce Irigaray’s writings (see, e.g.,
Irigaray, 2002).
6. For a discussion of the concept of the “self ” in cognitive linguistics see, for example, Lakoff
and Johnson (1999: 267–289).
7. Such a distinction between potentially good or bad variants of the same emotion also applies
to Ojanen’s description of his grandparents’ humility and pride in their existence (2005: 55).
While they felt humble (=not proud) in relation to their God, they took pride in conducting
their farming tasks in the Finnish countryside.
8. The subset which I reanalyzed for this article includes an author who talks exceptionally
frequently about both love and self-love, a nun called Winefrid Thimelby. Her letters in the
Corpus of Early English Correspondence Sampler provide 56 hits for love, of which 7, or 12.5%,
concern self-love.
9. For more information, see Hammer (2006 [2004]).
10. See Honan (2006 [2004]).
11. Address in Early English correspondence has been researched in depth by Nevala (2004).
12. A mother’s letter to her daughter from the same period attests to the same idea:
(3b) … withall be carful that, whatsoever you doe, to love honer and obey your husband
in all things that is fitting for a resonable creture.
(The Helsinki Corpus: Mary Peyton, Letter: 87)
13. Any ontogenetic model of love needs to take this into account (see, e.g., Hatfield & Rapson
2004 [2000]).
14. Compare with: “ … things purchase the consent that perpetuates the gendered social con-
tract. They enable the prevailing power relationship between men and women by grounding
it in contractual terms, by lending it a logic: specifically, by interposing the economic logic of
exchange.” (Cowen Orlin, 1993: 187–188, emphasis added)
15. They base their discussion on Dworkin (1987 [1977]: 171–173), who talks about goal-
based, right-based, and duty-based theories of law.
16. Note, however, how the media plays with such rights, tending to praise brave individual
choices but simultaneously making money on rumours which both challenge and titillate peo-
ple’s sense of morality.
17. Gevaert’s research into medieval English expressions of anger (e.g., Gevaert, 2001) shows
that the idea of the body as a container for the emotions is familiar to speakers of Old and
Middle English. Indeed, it is probably common to most if not all cultures (Kövecses, 2005: 246–
257, Yu 1995).
18. Deignan (2005: 73–142) deals with corpus-linguistic, cross-linguistic, psycholinguistic and

discourse approaches to metaphor. Even within diachronic studies of metaphors in English, people
are looking for combinations of methods (e.g., Fabiszak, 2001/2002: 19–42; Kossmann, 2006).
References
Coleman, Julie (1999). Love, sex, and marriage: A historical thesaurus. Amsterdam & Atlanta:
Rodopi.
Cowen Orlin, Lena (1993). The performance of things in The Taming of the Shrew. The Yearbook
of English Studies, 23 (Early Shakespeare Special Number edited by A. Gurr, P. Hardman,
& L. Kelly), 167–188.
Benjamins.
Deignan, Alice (2006). The grammar of linguistic metaphors. In A. Stefanowitsch & S. Th. Gries
(Eds.) (106–122).
Dworkin, Ronald (1987 [1977]). Taking rights seriously. London: Duckworth.
Evans, Vyvyan & Melanie Green (2006). Cognitive linguistics: An introduction. Edinburgh:
Edinburgh University Press.
Fabiszak, Małgorzata (2001/2002). The concept of ‘joy’ in Old and Middle English: A semantic
analysis. Piła: Wyższa Szkoła Biznesu.
Gevaert, Caroline (2001). Anger in Old and Middle English: A ‘hot’ topic? Belgian Essays on
Language and Literature, 89–101.
Goatly, Andrew (1997). The language of metaphors. London & New York: Routledge.
Gustafsson, Ylva, Tom Kettunen, Camilla Kronqvist, & Michael McEachrane (Eds.) (2005).
Proceedings from the Conference Emotions, Others and the Self (August 25–27 2005 Åbo/
Turku, Finland). Åbo/Turku: Åbo Akademis tryckeri. (A CD-ROM.)
Hammer, Paul E. J. (2006 [2004]). Devereux, Robert, second earl of Essex (1565–1601). In
Oxford dictionary of national biography. Oxford: Oxford University Press, Sept 2004; on-
line edition, May 2006.
[http://www.oxforddnb.com/view/article/7565, accessed 6 June 2006]
Hatfield, Elaine & Richard L. Rapson (2004 [2000]). Love and attachment processes. In M. Lewis
& J. Haviland-Jones (Eds.) Handbook of emotions [2nd edition] (654–662). New York &
London: The Guilford Press.
Honan, Park (2006 [2004]). Wriothesley, Henry, third earl of Southampton (1573–1624). In
Oxford dictionary of national biography. Oxford: Oxford University Press.
[http://www.oxforddnb.com/view/article/30073, accessed 6 June 2006]
Hundt, Marianne, Andrea Sand, & Rainer Siemund (Eds.) (1998) Manual of information to ac-
company the Freiburg – LOB corpus of British English (‘FLOB’). Freiburg: Albert-Ludwigs-
Universität Freiburg, Englisches Seminar.
[http://khnt.hit.uib.no/icame/manuals/flob/, accessed 20 July 2006]
Hundt, Marianne, Andrea Sand, & Paul Skandera (Eds.) (1999). Manual of information to ac-
company the Freiburg – Brown corpus of American English (‘Frown’). Freiburg: Albert-
Ludwigs-Universität Freiburg, Englisches Seminar.
[http://khnt.hit.uib.no/icame/manuals/frown/, accessed 20 July 2006]
142 Heli Tissari
Irigaray, Luce (2002). The way of love (La voie de l’amour). (Translated by Heidi Bostic and
Stephen Pluháček.) London: Continuum.
Jäkel, Olaf (1995). The metaphorical concept of mind: ‘Mental activity is manipulation.’ In
J. R. Taylor & Robert E. MacLaury (Eds.), Language and the cognitive construal of the world
(197–229). Berlin & New York: Mouton de Gruyter.
Johnson, Mark (1987). The body in the mind: The bodily basis of meaning, imagination, and
reason. Chicago: The University of Chicago Press.
Koivisto-Alanko, Päivi & Heli Tissari (2006). Sense and sensibility: Rational thought versus
emotion in metaphorical language. In A. Stefanowitsch & S. Th. Gries (Eds.) (191–213).
Koller, Veronika (2004). Metaphor and gender in business media discourse: A critical cognitive
study. Houndmills, Basingstoke: Palgrave Macmillan.
Koller, Veronika (2006). Of critical importance: Using electronic text corpora to study meta-
phor in business media discourse. In A. Stefanowitsch and S. Th. Gries (Eds.) (237–266).
Korsisaari, Eva Maria (2006). Tule rakkaani! Naisen ja miehen välisestä etiikasta kirjallisuuden
rakkauskuvauksissa. (“Come, my lover! On the ethics of the relationship between a woman
and a man in literary accounts of love.”) Jyväskylä, Finland: Gummerus.
Kossmann, Bianca (2006). From evil riches to common fertilizer: Mucking in with seman-
tic change. In R. W. McConchie, O. Timofeeva, H. Tissari, & T. Säily (Eds.) Selected pro-
ceedings of the 2005 symposium on New Approaches in English Historical Lexis (HEL-LEX)
(83–94). Somerville, MA: Cascadilla.
[http://www.lingref.com/cpp/hel-lex/2005/index.html, accessed 3 January 2008]
Kövecses, Zoltán (1986). Metaphors of anger, pride and love: A lexical approach to the structure
of concepts. Amsterdam: John Benjamins.
Kövecses, Zoltán (1988). The language of love: The semantics of passion in conversational English.
Lewisburg: Bucknell University Press / London & Toronto: Associated University Presses.
Kövecses, Zoltán (1990). Emotion concepts. New York: Springer.
Kövecses, Zoltán (2002). Metaphor: A practical introduction. Oxford: Oxford University Press.
Kövecses, Zoltán (2005). Metaphor in culture: Universality and variation. Cambridge: Cam-
bridge University Press.
Kytö, Merja (Comp.) (1996). Manual to the diachronic part of the Helsinki corpus of English
texts: Coding conventions and lists of source texts. (3rd edition.) Helsinki: University of
Helsinki, Department of English.
[http://khnt.hit.uib.no/icame/manuals/HC/, accessed 20 July 2006]
Lakoff, George (1987). Women, fire and dangerous things: What categories reveal about the mind.
Chicago: The University of Chicago Press.
Lakoff, George (1990). Is abstract reason based on image-schemas? Cognitive Linguistics, 1,
39–74.
Lakoff, George & Mark Johnson (1980). Metaphors we live by. Chicago: The University of
Chicago Press.
Lakoff, George & Mark Johnson (1999). Philosophy in the flesh: The embodied mind and its chal-
lenge to western thought. New York: Basic Books.
Larson, Kate (2005). Love and respect: Some aspects of love’s ambiguous force in the thought of
Iris Murdoch. In Y. Gustafsson, T. Kettunen, C. Kronqvist, & M. McEachrane (Eds.).
Lovejoy, Arthur O. (1978 [1936]). The great chain of being: A study of the history of an idea.
Cambridge, Ma & London: Harvard University Press.
Musolff, Andreas (2005). Genetic information as part of the ‘Great Chain of Being’. metaphorik.
de, 08, 52–70. [http://www.metaphorik.de/08/ accessed 26 July 2006]
Musolff, Andreas (2006). How real are discourse traditions of metaphor use? Paper given at the
RAAM6 conference Metaphor in the Real World, University of Leeds, UK, 12 April 2006.
Neiders, Ivars (2005). Love and virtue: Learning to see what is important (Reflections
on a theme from Plato’s Symposium). In Y. Gustafsson, T. Kettunen, C. Kronqvist, &
M. McEachrane (Eds.).
Nevala, Minna (2004). Address in early English correspondence: Its forms and socio-pragmatic
functions. [Mémoires de la Société Néophilologique de Helsinki LXIV.] Helsinki: Société
Néophilologique.
Nurmi, Arja (Ed.) (1998). Manual for the Corpus of early English correspondence sampler
CEECS. Helsinki: University of Helsinki, Department of English.
[http://khnt.hit.uib.no/icame/manuals/ceecs/, accessed 20 July 2006]
OED = Oxford English Dictionary, the. The OED Online.
[http://dictionary.oed.com/entrance.dtl, accessed in June 2006]
Ojanen, Markku (2005). Elämän mieli ja merkitys. (Free translation: “Making sense out of life.”)
Helsinki: Kirjapaja.
Pires de Oliveira, Roberta (1991). Language and ideology: An interview with George Lakoff. In
R. Dirven, B. Hawkins, & E. Sandicioglu (Eds.), Language and ideology. Volume I: Theoreti-
cal cognitive approaches (23–47). Amsterdam: John Benjamins..
Poole, Joshua (1972 [1663]). Practical rhetoric. Ed. by R. C. Alston. Menston: Scolar Press.
Shweder, Richard A. & Joan G. Miller (1985). The social construction of the person: How is it
possible? In K. J. Gergen & K. E. Davis (Eds.), The social construction of the person (41–69).
New York: Springer.
Solomon, Robert C. (1994). About love: Reinventing romance for our times. Lanham, MA: Row-
man & Littlefield.
Solomon, Robert C. (2005). The politics of emotions, revisited. In Y. Gustafsson, T. Kettunen,
C. Kronqvist, & M. McEachrane (Eds.).
Stefanowitsch, Anatol & Stefan Th. Gries (Eds.) (2006). Corpus-based approaches to metaphor
and metonymy. [Trends in Linguistics Studies and Monographs 171.] Berlin & New York:
Mouton de Gruyter.
Stefanowitsch, Anatol (2006). Words and their metaphors: A corpus-based approach. In
A. Stefanowitsch & S. Th. Gries (Eds.) (63–105).
Sweetser, Eve (1990). From etymology to pragmatics: Metaphorical and cultural aspects of seman-
tic structure. Cambridge: Cambridge University Press.
Tillyard, Eustace M. W. (1973 [1942]). The Elizabethan world picture. New York: Vintage.
Tissari, Heli (2003). Lovescapes: Changes in prototypical senses and cognitive metaphors since
1500. [Memoires de la Société Néophilologique de Helsinki LXII.] Helsinki: Société
Néophilologique.
Yu, Ning (1995). Metaphorical expressions of anger and happiness in English and Chinese.
Metaphor and Symbolic Activity, 10, 59–92.
chapter 8
A critical look at the desktop

metaphor 30 years on
Marina Terkourafi and Stefanos Petrakis

University of Illinois at Urbana-Champaign / University of Zurich
The Desktop Metaphor (henceforth DM) was born in the late 1970s, when
members of the research group working on the Apple Lisa project used the
term ‘desktop’ to define a new graphical user interface between humans and
computers. We trace the evolution of this interface under the twin constraints of
the technology available at the time and the ever expanding target user group.
Specifically, we propose a series of four multi-modal blends and analyse the
conceptual processes by which each successive blend builds on previous ones
and further combines them with new conceptual domains, leading up to today’s
virtual desktop. Finally, we critically assess the impact of the DM on the quality
of human-computer interaction, from an intra- and a cross-cultural perspective.
Keywords: interface, multi-modal blend, virtual desktop, human-computer

interaction
1. Introduction
According to the 1989 edition of the Oxford English Dictionary (OED), the noun
‘desk-top’, first encountered in 1929, denotes “the top or working surface of a
desk.” This is the sense of ‘desktop’ that we will refer to as the ‘wooden desktop’ in
what follows. By 1997, however, it was judged necessary to include an additional
sense. Within the domain of computing this time, the noun ‘desk-top’ has come
to denote “the working area of a computer screen regarded as a representation of a
notional desktop and containing icons representing items such as files and a waste
bin, used analogously to the items they symbolise.” The first reported appearance
of this new sense in print dates from as early as 1982, when one reads in the April
issue of Byte magazine: “Every user’s initial view of Star is the ‘Desktop’, which
resembles the top of an office desk.”
146 Marina Terkourafi and Stefanos Petrakis
Attributive uses of the noun in expressions such as ‘desktop computers’

meaning computers “suitable for use at or on a desk” (ibid.) had preceded the
development of the new sense by almost fifteen years and may have provided a
bridge between this new sense and the original one. Nevertheless, as the OED
definition suggests, there is more than simply a metonymic link between wooden
desktops on which desktop computers sit on the one hand, and virtual desktops
on the other hand. The link is metaphorical, and it is explicitly invoked both by
pictorial and by linguistic means: software components and actions are named
after, and depicted as, objects and actions associated with the original wooden
desktop domain. Thus, on the virtual desktop, as around a wooden one, one finds
‘files’, ‘folders’, a ‘wastebasket’, rulers, scissors and glue, and is able to carry out
actions such as ‘open’, ‘edit’, ‘cut’, ‘paste’, ‘delete’, and ‘close.’ This metaphorical link
is what prompts the semantic shift resulting in the additional sense of ‘desk-top’
noted in the OED.
The metaphorical extension of the meaning of desktop to the domain of com-
puting is the outcome of a series of decisions taken by software developing com-
panies in the late 1970s. Their aim was to come up with a user-friendly interface
that would enable personal computers to break free from the confines of spe-
cialised research institutes and propel them into mainstream culture. With com-
mercial success in mind and graphical user interfaces a recent addition to their
armoury, software developers explored several possibilities of navigating a screen
before settling on our now familiar arrangement. Thirty years on we can attest to
the success of that original aim. Looking ahead, however, we may also wonder: to
what extent might the desktop metaphor be a victim of its own success? As today’s
virtual desktop forever expands in new directions, is the original association with
the wooden desktop adequate, or even appropriate, for the increasingly inclusive
set of target users across the globe, and the increasingly wide range of tasks that
today’s computers can perform?
In this paper, we first consider the desktop metaphor (henceforth DM) in
its functional dimension. Functionally, the DM is an interface enabling the in-
teraction of a (non-specialist) user with the computer. We distinguish two types
of interfaces, and associate each type with a set of constraints that a successful
interface must satisfy. Posing the question of what the DM is for helps deter-
mine the set of constraints it must satisfy. With this information in mind, we
distinguish four generations of computer interfaces, starting with the command
prompt, through Windows Icons Menus Pointers (WIMP) interfaces, leading up
to the first instantiation of the DM and finally to today’s virtual desktop. Further-
more, we illustrate how each interface in this series may be analysed conceptually
as a multi-modal blend using the preceding interface as one of its input spaces,
and assess the effectiveness of each blend as an interface in the light of the means
Chapter 8. A critical look at the desktop metaphor 30 years on 147
that software developers had at their disposal at each stage as well as their aims.
We complete our overview of the conceptual bases and historical development of
the DM by pointing out some limitations emanating from this historical lineage
that potentially render the DM a less than satisfactory solution to the problem of
human-computer interaction (or HCI) in today’s global context.1
Our research is thus pertinent to several of the issues raised by Gibbs in
the introduction to this volume, specifically those concerning the interface of
metaphor with cultural norms, the synergy of linguistic and pictorial metaphors,
and finally, the DM itself as a mixed metaphor composed of several more spe-
cific ones and facilitating the metaphorical construal of different tasks under
its global umbrella. What is especially interesting in this regard is that, in the
case of the DM, the culture-specific ways of interacting with the physical envi-
ronment of a Western-type office were treated as a natural – rather than con-
ventional – benchmark on which to model this metaphor. The resulting built-in
culture-specificity of the DM, we argue, poses some intrinsic limitations to the
quality of human-computer interaction involving the DM, providing at least one
reason to experiment with more intuitive interfaces such as touch screen scroll-
ing popularized most recently.
2. The Desktop Metaphor as an interface
To operate any piece of machinery, an interface is necessary. Common interfaces

include electricity switches, gas-cooker knobs, water taps, door handles, locks,
keys, hand brakes and foot brakes, bicycle pedals, piano pedals, piano keyboards
and typewriter keyboards, bicycle gears, gear levers, and remote controls. In fact,
a large part of our everyday lives consists in manipulating the objects around us
(from opening doors and cooking meals to erecting buildings and communicat-
ing with others in far away places) with the help of such interfaces.
Some interfaces, such as levers and wheels, occurred early on in the history
of humankind, and rely on well-understood principles taught in basic physics
classes. Some actions, such as transporting and lifting small weights, may still
be possible without levers and wheels, but they would be much more cumber-
some or even impossible for the individual agent to perform. Other interfaces,
like touch screens and mobile phones, rely on complicated principles of electron-
ics that require years of training to be well-understood and then only by experts.
The activities we use them for would simply not be possible without them.
Interfaces, then, are needed, to various degrees, to enable the interaction of
the individual (lay) user with the machine. They achieve this by replacing the
need for combined human effort or expert knowledge by a suite of a few simple
basic actions that the individual agent can perform to obtain the desired effect. In
this way, interfaces both multiply the potential of the individual agent and spare
one the need for specialised knowledge. Moreover, interfaces set standards of in-
teraction with machines of similar functionality: once I know how to operate a
gas cooker, I can operate most gas cookers, and once I know how to drive a car, I
know how to drive most cars. But it is still possible to be puzzled by water taps in
unfamiliar places until one figures out whether to twist them, press them, press a
foot pedal, or simply wave one’s hands in front of them to achieve the desired ef-
fect of water pouring out of them. In other words, basic knowledge of an interface
can be extended in new directions in order to be adapted to particular situations.
Interfaces may be metaphorical but need not be. Whether an interface will
rely on a pre-existing metaphor or not depends on the type of entity that it is de-
signed to enable interaction with. Natural entities may already be conceptualised
metaphorically. For instance, heat is conceptualised by combining two primary
metaphors, intensity is quantity and more is up, generating a vertical scale,
on which increased intensity is signalled by upward direction. It seems only
natural, then, for interfaces manipulating heat, such as thermometers, radiator
controls or the knobs of electric/gas cookers, to inherit this metaphorical vertical
representation, and adapt it to the particular needs human users fulfil using each
interface. Thus, thermometers, which represent levels of heat, simply reproduce
this vertical representation as is, but radiator controls and cooker knobs intend-
ed to affect levels of heat adapt it to a circular shape that is easily ‘twistable’ by the
human user. In this sense, there seems to be little choice with regard to the par-
ticular interface used to model interaction with a natural entity. Such interfaces
are endowed with intuitiveness in virtue of the underlying primary metaphors,
and may be expected to emerge in society in a ‘bottom-up’ fashion, exhibiting
commonalities across cultures, at least inasmuch as the primary metaphors they
build on are universal.
Pre-existing metaphors are, however, not always available for invented enti-
ties, such as telephones and computers – though they sometimes may be, as in
the case of cars, which are arguably metaphorically construed as pack animals.
Interfaces for manipulating invented entities are subject to (at least) two other
constraints: first, the technology available at the time and second, the target user
group.2 For instance, we may trace the development of telephone devices under
these two constraints from the original two-piece receiver and transmitter fixed
on the wall, through French telephones, that took human anatomy into account
by combining receiver and transmitter in a single unit, to today’s mobile phones,
which respond to the ever increasing possibility of, and hence need for, mobility
in contemporary societies.
While subject to the combined constraints of the technology available at the

time and the target user group and hence not entirely unconstrained, interfaces
for manipulating invented entities clearly allow significantly more scope for
innovation compared to interfaces for manipulating natural entities that may
already be conceptualised metaphorically. The particular choice of interface,
then, becomes a strategic choice on behalf of the developer that crucially affects
future interactions of users with the machine in a top-down fashion. The suc-
cess of this choice depends on the extent to which the interface calls up scripts
for interacting with the machine that are both appropriate to its functionality
and familiar to the users.3 The more successful the interface in these two re-
spects, the more ‘natural’ and intuitively intelligible it will be for users. In this
way, the interface opted for can ultimately feed back into our conceptualisation
of the original entities, such that the terms of the interface may be used to refer
to components of the original entity itself. The DM as outlined in 2.3 below
provides an example of this.
The various interfaces defining human-computer interaction over the last four
decades fall under the latter category of interfaces, modelling invented entities. In
other words, they are subject to the twin constraints of the technology available
each time and the target user group. It is indeed possible to follow the evolution
of these interfaces by keeping track of the ever expanding range of technological
capabilities and how these were put to use according to the group of users tar-
geted each time. Definition of a particular target group led to the selection of
cultural scripts that would be familiar to this group. In this way, four generations
of interfaces may be distinguished, leading up to today’s virtual desktop. What
became known as the Desktop Metaphor may more specifically be identified with
the third step in this series of interfaces. To theoretically account for it, however,
it is necessary to go back to a time before its appearance, and examine the circum-
stances out of which the DM arose.
3. From the command prompt to the virtual desktop
3.1 The command prompt
Until the late 1970’s, the command prompt provided the standard way of interact-
ing with computers. This consisted in typing formal commands in an artificial
language such as a DOS script into a box in order to achieve particular effects like
opening a file or saving it. The command prompt as an interface may be theoreti-
cally accounted for as a multi-modal blend (see Figure 1).
Figure 1. The command prompt as a multi-modal blend
The domain of giving commands in natural languages and the domain of typing
linguistic expressions (familiar from typewriters developed over the preceding
century) function as the input spaces. What these spaces have in common is the
philosophical notion of speech acts, that is, the possibility to affect the state of the
world by using language, or, traditionally put, to ‘do things with words’ (Austin,
1962). The notion of speech acts, implying a speaker who utters a proposition p
with an illocutionary force F (F(p)), functions as a common denominator be-
tween the two input spaces, providing the generic space which enables entities
from one domain (e.g., speaker, oral command) to be mapped onto entities in the
other domain (typist, written command). The two input spaces draw on different
modalities (spoken vs. written), hence the resulting blend, in the shape of the
command prompt, is defined as a multi-modal one.
The main advantage of the command prompt is its flexibility. This makes it ap-
pealing to expert users who can adapt the interface to their purposes, such as cre-
ating user-defined ‘scripts’ or macros, and interacting with networked computers
even with low bandwidth (Soegaard, 2003). This advantage is indeed so important
to expert users that the command prompt retains its popularity among their ranks
even today, when it has, to all other intents and purposes, been superseded by a va-
riety of more recent interfaces. However, this flexibility comes at a cost that is often
prohibitive for the lay user. Generally, learnability and retention of commands are
very poor, such that interaction with the computer is essentially impossible without
proficiency in the language of the operating system. Error rates are high, and error
messages and assistance are hard to provide because of the diversity of possibilities
for error plus the complexity of mapping from tasks to interface concepts and syn-
tax. All in all, the command prompt is not a suitable interface for non-expert users,
and this was realised by software developers early on. Personal computers are, after
all, commercial products, and, like all products, their success is measured – for
some, exclusively – by the numbers of their users. If personal computers were to
break away from the confines of specialised research institutes into mainstream
culture, an interface that was more ‘user-friendly’ had to be devised.4
3.2 WIMP interfaces
The next step in the evolution of human-computer interaction interfaces came

with the appearance of what would eventually come to be called WIMP graphical
user interfaces (GUIs), an acronym standing for Windows, Icons, Menus, Point-
ers, after the names of the components of the new interface. The technology avail-
able to human-computer interaction researchers in the 1970s enabled them to start
experimenting with graphics in a 2D or pseudo-3D (i.e. mapped to 2D) space.
Windows, a term referring to rectangular, floating resizable shapes, became the
prevalent means for framing an application or a document. Miniature bitmapped
pictures, generic icons of virtual objects, served as thumbnails of real-world ob-
jects. Primary shapes such as rectangles and simple text labels enabled the creation
of menus and lists of commands. An uncanny piece of hardware with one, two or
up to five buttons, the renowned mouse device invented by Douglas Engelbart in
1963 (Müller-Prove, 2002), provided a new way of navigating through a 2D plane
and interacting with graphically represented objects. This domain of 3D-to-2D
mapping was one of the two input domains to the new blend that would produce
WIMP interfaces. Being closely reminiscent of the domain of drawing, a phyloge-
netically as well as ontogenetically primary domain of human experience, 3D-to-
2D mapping ensured the intuitive appeal of the new interface beyond the group of
expert users to the wider target group of lay computer users.
The second input domain was the command prompt, itself a multi-modal
blend as outlined above.5 Contrary to WIMP interfaces, the command prompt
interface merely implies the existence of objects. The commands typed into a
command prompt alter the state of objects that are invisible or otherwise not
ersistently perceived by the user. Attaching to these objects a visual form (in
p
contrast to a linguistic one) and presenting these forms to the user implements the
existence of the manipulated objects, eliminating the indirect implication of their
existence: the objects achieve existence by being ever-present entities visible to
the eyes of the user. In turn, the commands that can affect these entities take the
form of visible actions chosen from a menu, or mouse gestures, like dragging an
icon or closing a window.
WIMP interfaces blend a name-based approach with a spatial one. Visual rep-
resentations of objects and actions function as visual signifiers, much as the origi-
nal linguistic forms standing in for them in the command prompt functioned as
linguistic signifiers of the corresponding signified objects or actions. The shift from
interacting through the syntax of an artificial language to motor-visual interaction
is accomplished through a generic space which highlights the common signifieds
of the visual and linguistic signifiers. This highlighting is achieved by means of
primary metaphors that help align the relevant topologies of the two inputs (cf.
Fauconnier & Turner, 2002: 339). These include metaphors pertaining to creation
(existence is visibility, existence is having a form, create is moving to
a location/here), representation (states are shapes, shapes are contain-
ers), and acting by moving (action is motion, change is motion, acting on is
transferring an object). The primary nature of these metaphors provides a fur-
ther guarantee of the intuitiveness of the new interface. The resulting blend once
more draws on different modalities, these being on the one hand typing linguistic
expressions (itself a hybrid modality combining language and motor action), and
on the other hand visually guided motor actions on a 2D plane, represented as a
canvas or screen (another hybrid modality combining vision and motor action).
Figure 2 illustrates how a primary metaphor is used to align the topologies
of the command prompt and of drawing on a 2D (pseudo 3D) plane. The rele-
vant metaphor is creation is visibility. To create a new file using the command
prompt, one must type in and execute the appropriate command (e.g., using the
copy con command in a DOS prompt). However, this does not lead to any visible
results, although the new item has been created. To see this item, the user must type
in an additional command (e.g., the dir command). When creating the graphical
representation of a file on a 2D drawing canvas, the painter can directly perceive
this: the drawing stands before her eyes. In the resulting blend of these two input
spaces, i.e. WIMP interfaces, creating a file is achieved via a visual-motor command
(e.g., clicking on a “New File” label in a menu) which makes an appropriate icon vis-
ible on the screen and available for further manipulation. The causal and temporal
relation between the user’s action and its effect are thus visually implemented and
can be immediately perceived by the user, providing a prototypical instance of what
Fauconnier and Turner (2002: 324ff.) call “compression of vital relations.”
Figure 2. WIMP interfaces as a multi-modal blend
The use of WIMP interfaces is a revolution in its own right. The cognitive burden
imposed by the complex syntax of command prompt interfaces is rapidly reduced,
a feature that is particularly appealing to non-expert users. Instead of memorising
or reading the instruction manual time and again, a simple gesture such as right-
clicking on a part of the screen or activating a menu, presents the user with avail-
able commands that apply to a given situation or state of interaction: memorisa-
tion is no longer required. Memory-related tasks are further enhanced by what
is called in psychology motor-memory reinforcement and the use of mnemonics
(Müller-Prove, 2002) through spatial relationships (method of loci). At the same
time, moving, copying, opening and other commands with a visual counterpart
can be reduced to simple mouse gestures, providing a cognitively economic, easier
and faster alternative for performing actions. This way of behaving in a computer
environment by interfering with objects (object orientation) was the most sig-
nificant gain introduced by WIMP interfaces. Though not yet fully implemented
by WIMP interfaces of that era, direct manipulation of objects, as later formally
described and defined by Ben Shneiderman (Shneiderman, 1983), is the core idea
behind their success and intuitive use. According to Shneiderman’s definition,
direct manipulation involves continuous representation of objects of interest

through time, and rapid, reversible, incremental actions and feedback. A very
successful implementation of a WIMP GUI was the Smalltalk programming envi-
ronment that was developed at Xerox PARC. However, word processors, applica-
tions still in use today in more or less the same form as at that time, remain the
archetypal example of a WIMP interface.
3.3 The Desktop Metaphor
The 1980s brought great glory and acknowledgement to the innovations born
inside the Xerox research labs. But Xerox was not to reap the financial benefits of
its inventions. Following the launch of the Smalltalk programming environment,
Xerox tried to commercialise a set of WIMP applications hosted by an integrated
graphical environment. These were brought together in the Xerox Star, a personal
computer presented to the public in 1981. However, the Star failed to be a com-
mercial success. During the same period, similar ideas were being developed by
Apple as part of two separate projects. The first of these to materialise as a product
was Lisa, a personal computer with its accompanying software. It is the Lisa proj-
ect that is credited with being the first to use and define the term “desktop meta-
phor” (Perkins et al., 1997). Following Star’s fate, Lisa did not reach commercial
success. It was the second of Apple’s projects that achieved that. That project was
the Macintosh and it is its entry into the market that marks the turning-point that
put the DM on the map.
What is known about the discussions that took place among the creators of
the DM confirms that the target audience for the Macintosh were white-collar
employees, office workers that operate on and around a physical office desktop.
Selection of this target group is clearly driven by considerations of commercial
success, and in turn justifies the selection of the wooden desktop as a source do-
main. All of the nameless, generic objects that were floating in a WIMP environ-
ment were now named after real-world objects that existed in the office domain.
Thus, icons, windows, pointers and menus were conceptualised in terms of the
wooden desktop, as was implied by their forms and labels, and verified by their
functionality. The effect of superimposing a metaphor from a familiar real-world
domain onto WIMP interfaces translated into a significant gain in usability and
intuitiveness. So much was made clear by the success of the Macintosh and all the
other players that subsequently entered the market following in Apple’s steps.
Conceptually, the desktop metaphor constitutes a blend between the input do-
mains of WIMP interfaces and of the wooden desktop, typically found in an office
environment. The generic space between these two input spaces comprises Agents,
Undergoers and Procedures. This triplet is responsible for the alignment that takes
place between the two input spaces, aligning the bureaucrat (white collar employ-
ee) with the mouse pointer, the real world desktop objects with bitmapped objects
(icons) and windows, and finally the actions that are performed on desktop objects
with the visual-motor actions that take place inside WIMP interfaces.
Figure 3 provides an illustration of the blending process that takes place in
this conceptual integration network. Whereas in a WIMP interface deleting an
item is accomplished by permanently making the corresponding icon invisible,
in the real-world desktop you would simply throw it in the bin. The mapping
that takes place between these two input spaces is captured in a generic space
encompassing the Agent, the Undergoer, and the Procedure, in this case destroy-
ing an item. In the WIMP interface input space, the Agent maps onto the point-
ing device, the Undergoer to a generic icon, and the Procedure to permanently
removing an item from visibility. In the real-world desktop, the Agent maps onto
the bureaucrat, the Undergoer to an office object (e.g., a paper document) and the
Procedure to throwing it in the bin. These are projected in the blend where the
Agent is now the cursor (e.g., an icon of a hand or an arrow), the Undergoer is a
document object (e.g., an icon of a paper document), and the Procedure is drag-
ging and dropping the icon over the icon for the wastebasket.
Figure 3. The Desktop Metaphor as a multi-modal blend

The intuitive appeal of the desktop metaphor lies in its calling up familiar scripts
for objects and actions from the real-world office domain, enabling knowledge
transfer – or, in Fauconnier and Turner’s words, “pattern completion” (2002: 328) –
between the two domains.6 In both domains, files may be opened, copied, and
stored away to be edited again later. Moreover, files are organised in a hierarchi-
cal system of folders that may be searched. Unwanted files are disposed of in the
wastebasket. In this way, which actions would be applicable to which objects, as
well as the effects of those actions on the particular objects, can be predicted,
drawing on the behaviour of their counterparts in the input spaces, such that new
users may be familiarised with the computer environment with minimal tuition
and maximal effectiveness.
Nevertheless, not all knowledge of possible actions on a computer desktop en-
vironment can be imported from the real-world office domain. Operating success-
fully on a virtual desktop further requires knowledge of two classes of conventions,
one class applying to particular actions, and another class applying to particular
objects. Under the first class one finds conventions such as, for instance, destroying
a file, which is realised by DELETing it, and the act for putting a file into a folder,
which is realised by DRAGging AND DROPping the icon of the file over the icon of
the folder. In both cases, the user is faced with a cross-modal mismatch. In the for-
mer case, the mismatch is between the name of the action and its significance (de-
leting in the real world amounts to erasing the contents of a file, not reducing it to
non-existence). In the latter case, the mismatch is between the visual representation
of the action and its real-world significance (dragging and dropping is not a pos-
sible action for putting files into folders in the real world). In addition to the novel
matching between (names of) actions and their effects, users are called to master a
class of conventions applying to particular objects that can override the former class
of conventions applying to actions. For instance, like all folders, MY COMPUTER
and the WASTEBASKET both can be OPENed but, unlike other folders, they can-
not be DELETEd. In another famous instance of human-computer interaction fail-
ure coming from the Macintosh implementation of the desktop metaphor, DRAG-
ging AND DROPping the floppy drive icon over the WASTEBASKET has the effect
of ejecting a diskette from the disk drive, rather than destroying it, although the
latter is the effect achieved by dragging and dropping any other file or folder into
the wastebasket. Fauconnier & Turner (2002: 341) explain the infelicity of this move
as a failure to satisfy the Integration, Web and Topology principles all at once. De-
spite the counter-intuitive effects of these actions, users learn these conventions as
emergent structure unique to the blend (cf. Fauconnier & Turner, 2002: 42–44). In
other words, users are prepared to accommodate some degree of departure from
the original office domain, provided one retains the feeling that the blend remains
overall functionally consistent, or ‘integrated.’
The DM provides several examples of compression (Fauconnier & Turner,

2002: 312ff.). Compression allows us to accommodate for the fact that the waste-
basket is now presented alongside other files and folders as being on the desktop,
rather than underneath it, and also for the fact that, contrary to its physical coun-
terpart, it never fills up (Fauconnier & Turner, 2002: 340).7 Similarly, it is “creation
by compression” that is responsible for how we are able to construe novel ways
of acting on virtual objects, such as ‘right-clicking’ or ‘dragging and dropping’, by
associating a sound and/or a gesture, respectively, with an effect that cannot be
achieved outside the blend (cf. ibid.: 319–320).
According to Fauconnier and Turner (2002: 329), “ [i]ntegration in the blend-
ed space allows its manipulation as a unit, makes it more memorable, and enables
the thinker to run the blend without constant reference to the other spaces in the
network.” It is indicative of the level of integration that the DM has attained as a
multi-modal blend in our days that it appears to have expanded beyond the do-
main of computing. The DM has achieved a reality of its own, such that real-world
actions can now be (sometimes anachronistically) described in its terms.8 Under
Forceville’s (2002: 7) criterion of the irreversibility of source and target, this brings
into question the extent to which the DM ought to be properly considered a (pro-
totypical) metaphor any more. Source and target seem to be quite reversible in this
case, as the description of real-world actions in terms of the DM attests. Observa-
tion of children’s acquisition of the conceptual content of “files,” “folders,” “saving”
and “deleting,” as well as dictionary definitions of these terms would seem to sup-
port the independence of the original DM from the input spaces that brought it
about as a blend. This makes the DM available as a conceptual domain in its own
right, able to provide one of the input spaces to the next, and final, multi-modal
blend in the family of desktop metaphors so far, today’s virtual desktop.
3.4 Today’s virtual desktop
The main reasons for change once more reside in the technology available at the
time and in the target user group. Technological progress since the 1980s occurred
at two levels. At the level of hardware, increased compactness has led to increased
mobility, such that a lot of the external interfaces (e.g., screen, keyboard, mouse)
are integrated in today’s laptops, while others (e.g., audio-speakers, drives provid-
ing additional storage) may be added as peripherals. At the level of software, devel-
opments have been toward increased versatility. Personal computers are no longer
just an office accessory for white-collar employees, or even academics. Computer
software addresses the needs of professionals in several disciplines, including med-
icine, lab-based professions, astrophysics, design, education, communications, and
even the arts, to mention but a few. Moreover, computers are no longer just part of
our professional lives, they are a part of our personal lives as well. Entertainment,
leisure, and communications-related activities, such as shopping, reading the press,
downloading and performing music, browsing through the world wide web, social
networking, communicating via text, speech and video, editing photographs, and
playing video games have been added to the suite of tasks that computers routinely
help us to perform. This expansion of technological capabilities means that an ever
expanding group of users can be targeted. They include professional users, who
use computers as part of their work, but also home users, who use computers for
non-work related activities, and include pensioners, disabled users, housewives,
and children.
As a result, different domains of real-world knowledge on the one hand are
blended with the original DM domain on the other hand, producing a multi-
modal blend with multiple inputs, the contemporary virtual desktop. Like its
predecessor, the original DM of the 1980s, the generic space for today’s virtual
desktop comprises Agents, Undergoers and Procedures. However, the mapping
mediated by this generic space is no longer simply between the bureaucrat on the
one hand and the computer user on the other, but between the computer user on
the one hand and the analyst, the consumer, the journalist, the music performer,
the gamer, the communicator, and so on, on the other hand. Today’s apparently
omni-potent user acts on Undergoers such as virtual photographs, virtual musi-
cal instruments, virtual lab equipment, virtual board or other types of games,
virtual books, and virtual communication devices, produced by the alignment of
their real-world counterparts with the products of software applications, via Pro-
cedures such as image editing, performing a musical instrument, playing a game
etc., which once more derive their significance from their counterpart actions in
the real world.
Despite these alterations and additions to its original purpose and format, to-
day’s virtual desktop still preserves the desktop metaphor at its core. This under-
lies the treatment of new objects, such as photographs and songs, as files that may
be opened, edited, and saved or deleted. Arguably, it is the co-occurrence of these
(often unrelated) objects on the screen, driven by compression (cf. Fauconnier
& Turner, 2002: 316), that prompts us to find a link between them, enabling us
to make this conceptual adjustment. Another instance of compression, this time
aided by topology (cf. ibid.: 327), is responsible for allowing us to conceptually
reconcile being ‘simultaneously’ at different locations when logging in remotely
(cf. ibid.: 317–319). Moreover, computer jargon, although continuously enriched
with new terms such as blogs, widgets, and wikis, essentially expands on the DM’s
original vocabulary.
Figure 4 illustrates the operation of today’s virtual desktop as a multi-modal

blend, using ACTIVATION of an object of interest as an example. In the original
virtual desktop, the object is the document (conceptualised in terms of a real office
document). In the contemporary virtual desktop, the same term “document” is
used to describe a vastly more generic notion. This genericity is better understood
by considering some of the possible input spaces to today’s virtual desktop. The
Agent may be a shopper, a photographer, or a music maker, the process of activat-
ing corresponds to visiting, developing, or composing, and the Goal or Undergoer
respectively may be a shopping site, a film, or a sound recording, and so on. All
these are projected (or ‘compressed’) onto the same terms inherited from the origi-
nal virtual desktop. These elements are part of the emergent structure in the blend:
one would not even dream of listening to, or watching a document in the original
virtual office, but this is precisely what happens in today’s virtual desktop.
Figure 4. Today’s virtual desktop as a multi-modal blend

However, the expansion both of the range of tasks performed using computers
and of the target user group raises the question of the adequacy of the original
office-like conceptualisation scheme. Simply put, how far is it possible to stretch
the conceptualisation away from the much narrower intentions and target group
of its original creators before the metaphor starts to break down? The continu-
ing accumulation of unrelated icons (from compasses for navigation software
to anchors for web-browsing to depictions of fire for virus-protection software)
on today’s virtual desktop challenges two aspects of the metaphor, its predictive-
ness, and its intuitiveness. Creators of the original DM had paid particular at-
tention to both aspects when selecting the real-world office domain as input to
the original DM. Indeed, it is the possibility of knowledge transfer between the
two domains (from the wooden desktop to the virtual desktop) that should be
credited with the immediate appeal of the DM to lay users that translated into its
huge commercial success. Knowledge transfer from the real-world office domain
is however no longer adequate for the range of tasks performed using computers
today. The predictiveness of the metaphor is thus curtailed as several knowledge
domains – some (such as photography, music, shopping) already existing in the
real world, and others (such as web-browsing) existing only within computing
environments – impose requirements that must be met by means of the same
suite of actions available on a virtual desktop. These actions, then, acquire a dif-
ferent significance depending on the object acted upon, increasing the amount of
special conventions that must be learnt to survive conceptually on today’s desk-
top. This threatens the Web principle, requiring instead constant reference to the
input spaces and thus pushing toward Unpacking, thereby counteracting Integra-
tion (cf. Fauconnier & Turner, 2002: 329ff.).9 The intuitiveness of today’s desktop
is also curtailed, as multiple cross-modal mismatches arise, in turn further en-
couraging Unpacking. As a result, one significant advantage of the original DM,
namely reducing the need for memorisation, appears to be threatened.
Clearly, the relative independence that the DM has achieved from its original
input spaces (WIMP interfaces and the wooden desktop) is a major advantage as
it is now faced with these challenges. The reason why users already familiar with
the DM may not be puzzled by the loss in predictiveness and intuitiveness as out-
lined above is because they have already mastered, to a greater or lesser extent, its
special conventions, and hence only need to adapt this knowledge to the require-
ments of the new era. A simple experiment could be devised to discover whether
‘naïve’ new users (to the extent that such can be found in highly technological so-
cieties) faced with today’s virtual desktop find it more puzzling than the original
DM of the early 1980s. Our guess is that they would find it more puzzling, and the
reason for that is that basic computer literacy (i.e. the DM having achieved rela-
tive independence as a conceptual domain in its own right) is a precondition to
extending its application to new domains. In other words, as pointed out above,
the fact that “basic knowledge of an interface can always be extended in new di-
rections in order to be adapted to particular situations” presupposes basic knowl-
edge of the interface.
If basic computer literacy provides at least a partial answer to the challenges
posed by predictiveness and intuitiveness, a final challenge, this time posed by
the global dimension of today’s virtual desktop, is not as easily addressed. To-
day’s virtual desktop has expanded its audience, both hierarchically, across social
classes, and horizontally, across geographical locations. This expansion presents it
with two additional problems. The first concerns the universality of the DM: the
wooden desktop domain is (or was) typical of a western, office-centric culture.
Its success built on exploiting cultural scripts that were already salient to its tar-
get group of users. Since salience varies by person/culture (Forceville, 1996: 101,
149; 2002: 11), salience of the corresponding scripts cannot be taken for granted
by other social or cultural groups, seriously hampering the desired knowledge
transfer, a necessary first step toward computer literacy, as outlined above. This
“cross-cultural” problem (where culture is broadly conceived as operating on a
common set of scripts) recurs every time knowledge transfer pertaining to an
office-related object or procedure is needed. The concept being transferred is ex-
pected to exist in the target culture; non-existence leads to non-transferability of
the DM across cultures in a compact and consistent way. Consequently, to call up
scripts that are salient to the expanded range of users, one needs to move away
from specific world knowledge and draw on general world knowledge.10 In other
words, to achieve universal “plug and play,” one needs to be rather generic. In this
respect, metaphors drawn from nature (e.g., a spider’s web) would seem to have
an inherent advantage.
The second problem posed by the hierarchical and geographical expansion of
its target user group concerns the translatability of the DM. Even in those cultures
where the wooden desktop domain is available, it may be tied up with (potentially
negative) connotations (e.g., red-tape) that are carried over in word-per-word
translations (e.g., of menus, commands, etc.) and limit its user friendliness, in
terms of both its applicability to new domains and appeal to new categories of us-
ers. This “intra-cultural” problem highlights the different ways in which speakers
conceptualise and think about the office domain, its objects and procedures: an
office object can be quite different as a concept between languages, depending on
its relationship with other concepts in the same language.
These points cannot be immediately accommodated during online produc-
tion of the blend. To integrate such unrelated components, the original fit with
the wooden desktop domain begins to loosen and the emergent structure starts
to be less and less dependent on it. At the same time, our ability to come up with
the appropriate emergent structure crucially depends on our having internalised

the components of the DM in its original form, i.e. on a minimum of computer
literacy. The danger here, as described above, is that not one, but several principles
of conceptual blending are simultaneously beginning to be less and less satisfied,
producing an outcome that is less than optimal. The impact of this is seen mainly
in terms of reduced intuitiveness and user friendliness for the ‘naïve’ new user (a
group whose numbers in less technologically developed parts of the world should
not be underestimated). The desktop metaphor has been superseded by the ex-
panding usage of the virtual desktop: it is simply not generic enough to play the
role it is called to play in the new millennium.
4. Summary
Our aim in this paper has been threefold. First, we have tried to provide a func-
tional motivation for the desktop metaphor. To this end, we outlined different types
of interfaces and defined the desktop metaphor as a family of interfaces for hu-
man-computer interaction enabling interaction with an invented entity, and hence
subject to the twin constraints of the technology available each time and the target
user group. Second, we provided a timeline of the development of these interfaces,
starting with the command prompt of the 1970’s, through WIMP interfaces, the
original DM, and leading up to today’s virtual desktop. Each of these interfaces
may be analysed as a multi-modal blend, using the previous blended space as one
of its input domains. This process of knowledge accumulation proves crucial to
successfully construing the final blend, today’s virtual desktop. Third, we provided
a critique of today’s virtual desktop as an appropriate interface for HCI given its
expanded range of users and tasks. Specifically, we argued that lack of familiarity
with the original DM can seriously curtail the predictiveness and intuitiveness of
today’s virtual desktop. Familiarity with the original DM may in turn be difficult
to achieve for the expanded group of users of today’s virtual desktop that go well
beyond the group of white-collar employees targeted by the original DM.
This inbuilt inadequacy of today’s virtual desktop, due to its historical lineage,
raises serious doubts as to its appropriateness to provide the conceptual bridge
that will propel less technologically advanced portions of the global population to
the IT-saturated twenty-first century. In other words, there is a real danger that,
instead of achieving increasing inclusiveness, as has been the aim of software de-
velopers to date, the DM may increase the rift between the technologically adept
and the technologically inept, forever excluding the latter from knowledge avail-
able to the former. This is a danger that software developing companies must seri-
ously consider if we are to avoid this undesirable social situation in future.
Notes
1. Our aim in this paper is to provide an integrated analysis of the Desktop Metaphor from a
functional, a historical and a conceptual perspective. This aim is different from that of Fauconnier
and Turner (2002), who, despite making a number of interesting remarks about the Desktop
Metaphor, use these remarks to illustrate theoretical points and do not to offer an analysis of
the DM as a phenomenon in its own right. More importantly, they do not distinguish between
the various stages that the DM went through historically, but instead conflate these in one blend
(most obviously when they write about “the frame of office work” and “the frame of traditional
computer commands” as being directly blended with each other; p. 131, cf. also p. 295). How-
ever, both the historical record (the sequence with which products were launched, as well as
discussions between experts at the time) and the conceptual analysis (blends building on the
level of familiarity with the computer domain achieved through previous blends) suggest that
the evolution of the DM is chronologically punctuated by a sequence of distinct stages that are
important to acknowledge as such in order to provide a more in-depth analysis of the DM.
2. When this target group includes no less than the whole of humankind, responding to this
latter constraint is achieved by taking into account features such as human anatomy, provid-
ing an opportune instantiation of Fauconnier and Turner’s (2002: 322ff.) overarching goal of
“achiev[ing] human scale.” More narrow definitions of the target user group (as we shall en-
counter for the DM in 2.3 below) may provide specific constraints that will be superimposed
on this general one.
3. This parallels the function of metaphors in advertising (Forceville, 1996: 81), where the
success of the advertisement depends on the extent to which the metaphor calls up scripts that
are memorable and/or desirable to the particular target audience.
4. Revealing something about developers’ concerns in those early days, the first appear-
ance of the term ‘user-friendly’ in print dates from 1977 and occurs in a related context,
specifically in the phrase “STDS-I does not provide the user with a sufficiently ‘user-friendly’
interface to allow noncomputer scientists to easily work with a data base” (Set Theoretic Data
Structures (STDS), vol. 31).
5. The process of knowledge boot-strapping implied by the re-entrancy of each blend as an
input space to the next blend may be explained as an instance of recursion achieving increasing
compression to human scale (Fauconnier & Turner, 2002: 334).
6. Fauconnier & Turner’s Pattern Completion principle states: “Other things being equal,
complete elements in the blend by using existing integrated patterns as additional inputs. Other
things being equal, use a completing frame that has relations that can be compressed versions
of the important outer-space vital relations between the inputs” (2002: 328).
7. This is an instance of compression winning over topology (Fauconnier & Turner, 2002: 340),
illustrating the optimality-theoretic way in which the principles are played against each other,
as Fauconnier and Turner predict (ibid.: 322).
8. For instance, in an exhibition of Michelangelo’s drawings, the caption to one of the great
master’s drawings in which he had glued strips of paper to the bottom and left of the original
piece of paper to expand the drawing surface referred to this action as “cut-and-paste editing”
(Michelangelo’s drawings, 23 March–25 June 2006, British Museum, London).
9. According to the Web principle, “[o]ther things being equal, manipulating the blend as a
unit must maintain the web of appropriate connections to the input spaces easily and without
additional surveillance or computation” (Fauconnier & Turner, 2002: 331), while according to
the Unpacking principle, “[o]ther things being equal, the blend all by itself should prompt for
the reconstruction of the entire network” (Fauconnier & Turner, 2002: 332).
10. The terms are taken from Schank and Abelson (1977: 37) according to whom knowl-
edge about the world forms a continuum, ranging from specific knowledge (“knowledge [we
use] to interpret and participate in events we have been through many times”) represented
by scripts, to general knowledge (“knowledge [which] enables a person to understand and
interpret another person’s actions simply because the other person is a human being with
certain standard needs who lives in a world which has certain standard methods of getting
those needs fulfilled”).
References
Austin, John (1962). How To Do Things With Words. Oxford: Clarendon Press.
Fauconnier, Gilles & Turner, Mark (2002). The Way We Think: Conceptual Blending and the
Mind’s Hidden Complexities. New York: Basic Books.
Forceville, Charles (1996). Pictorial Metaphor in Advertising. London/New York: Routledge.
Forceville, Charles (2002). The identification of target and source in pictorial metaphors. Jour-
nal of Pragmatics, 34, 1–14.
Müller-Prove, Matthias (2002). Vision and reality of hypertext and graphical user interfaces. MA
Thesis, Department of Informatics, University of Hamburg. Retrieved May 10, 2005 from:
http://www.mprove.de/diplom/index.html
Perkins, Roderick, Dan Keller & Frank Ludolph (1997). Inventing the Lisa user interface. In
S. Cherry (Ed.), ACM interactions 4 (pp. 40–53). New York: ACM Press.
Schank, Roger & Abelson, Robert (1977). Scripts, Plans, Goals and Understanding: An Inquiry
Into Human Knowledge Structures. New Jersey: Lawrence Erlbaum.
Shneiderman, Ben (1983). Direct manipulation: A step beyond programming languages. IEEE
Computer, 16, 57–69.
Soegaard, Mads (2003). Interaction-Design.org Encyclopedia: Interaction Styles. Retrieved May
11, 2006 from: Interaction-Design.org: edit http://www.interaction-design.org/ encyclo-
pedia/interaction_styles.html.
The Oxford English Dictionary. 2nd edition (1989). OED Online. Oxford University Press. Ac-
cessed May 10 2006.
chapter 9
Pragglejaz in practice
Finding metaphorically used words
in natural discourse
Gerard J. Steen, Ewa Biernacka, Aletta G. Dorst, Anna A. Kaal,

Irene López-Rodríguez and Trijntje Pasma
VU University Amsterdam
This paper aims to present the application of a new method for metaphor iden-
tification in discourse, the Pragglejaz Metaphor Identification Procedure, in two
corpus-linguistic projects. It provides background information to the method
and then demonstrates its basic workings by means of three examples. Then
some major issues with the method are discussed in more detail: the nature of
units of metaphor analysis, the identification of contextual and basic meanings
of words in lexical units in discourse, the role of etymology in determining
basic meanings, and the role of distinctness as well as similarity required for
metaphorical word meaning.
1. Introduction
As is also pointed out in the introductory chapter to this volume by Ray Gibbs, the
identification of metaphor in natural discourse is not unproblematic (Steen, 2007;
Steen et al., 2010; cf. Cameron, 1999). This is partly because of the rich variety dis-
played by metaphor in the real world, as opposed to the typically decontextualized
and constructed examples of metaphor in research that is less ecologically valid
(see Gibbs in the introductory chapter, again). But even researchers with great ex-
pertise in the study of metaphor as a natural language phenomenon frequently dif-
fer in their judgments of what counts as a metaphor. Their investigations typically
do not exhibit specific and fixed criteria for making decisions about which words
or phrases they consider as metaphorically used. Lack of such uniform criteria for
the process of metaphor identification leads to inconsistency within researchers
and lack of agreement between researchers, and hence reduces the validity and
reliability of claims about metaphoricity. As is also pointed out by Gibbs, this has
166 Gerard J. Steen et al.
been one of the on-going struggles in metaphor research. This is why a clearly for-
mulated procedure is crucial for scholars researching metaphor in discourse.
One solution to this problem has been presented by the Pragglejaz Group
(2007), who published their Metaphor Identification Procedure, called MIP. The
Pragglejaz procedure offers a flexible tool for metaphor scholars in various fields
of language study. It is the product of a group of metaphor researchers known by
the initial letters of their first names:
Peter Crisp, Chinese University Hong Kong, China
Ray Gibbs, University of California, Santa Cruz, USA
Alan Cienki, VU University Amsterdam, The Netherlands
Graham Low, University of York, UK
Gerard Steen, VU University Amsterdam, The Netherlands
Lynne Cameron, The Open University, UK
Elena Semino, Lancaster University, UK
Joe Grady, Cultural Logic LLC (Washington DC), USA
Alice Deignan, University of Leeds, UK
Zoltán Kövecses, Eötvös Loránd University (Budapest), Hungary
This group collaborated for six years with the specific aim of developing a tool for
metaphor identification in natural discourse that is both reliable, as indicated by
statistical tests, and valid, in that it attempts to make explicit how it makes use of
current empirical research in cognitive linguistics, discourse analysis, psycholin-
guistics, and applied linguistics (cf. Steen, 2001, 2002, 2005, 2007).
MIP is not the only metaphor identification procedure that is available (cf.
Cameron, this volume), but it is the only procedure that has been formally tested
with a view to making it available as a tool to a larger audience. Since its publica-
tion, MIP has been adopted by quite a few metaphor researchers who do empiri-
cal metaphor research on natural data: at RaAM 7, in Cáceres 2008, there were
over a dozen presentations that utilized or referred to MIP. Since practical experi-
ence with MIP has been limited, however, a report of hands-on experience with
its application may be useful in highlighting its strong and weak points, so that
researchers may make better use of the tool for their own purposes.
The Pragglejaz procedure has been adopted in two research programmes on
metaphor in natural discourse at the VU University Amsterdam. The first pro-
gramme is called “Metaphor in discourse: Linguistic forms, conceptual structures,
and cognitive representations” and runs from September 2005 through August
2010. It involves the first five authors of this contribution. The first stage of this
programme involves the analysis of four samples of 50,000 words from a publicly
. The second author and the fifth author were involved in the first year of the programme.
From the second year on, Berenike Herrmann and Tina Krennmayr replaced the second and
fifth authors.
Chapter 9. Pragglejaz in practice 167
available sample from the British National Corpus, called BNC-Baby. The second
programme is called “Conversationalisation of public discourse”, has the same tim-
ing, and involves the first and last author of this contribution. The first stage of this
programme has analysed three samples of in total 130,000 words from three Dutch
corpora in one coherent metaphor project. Both programmes have employed the
Pragglejaz procedure as part of a more encompassing method for metaphor identi-
fication which will be described elsewhere (e.g. Steen, 2007; Steen et al., 2010).
This presentation is a report of our practical experience with applying the
Pragglejaz procedure to our data. The aim of our contribution is to present the
method and its operationalization in some detail, in order to increase its availabil-
ity to researchers interested in metaphor identification and analysis in natural dis-
course. We will outline the particular steps of the procedure and demonstrate how
it is applied on the basis of a few examples taken from our database. Then we will
address and discuss in detail various issues that we have faced at each step of the
procedure in the metaphor identification process. But first we will provide some
background information on both the procedure as well as our own research.
2. Demonstration
The Pragglejaz procedure is a practical and systematic method for identifying

metaphorically used words in discourse. It consists of the following steps (Prag-
glejaz Group, 2007: 3):
1. Read the entire text–discourse to establish a general understanding of the
meaning.
2. Determine the lexical units in the text–discourse.
3. (a) For each lexical unit in the text, establish its meaning in context: that is,
how it applies to an entity, relation, or attribute in the situation evoked by
the text (contextual meaning). Take into account what comes before and
after the lexical unit.
(b) For each lexical unit, determine if it has a more basic contemporary mean-
ing in other contexts than the one in the given context. For our purposes,
basic meanings tend to be:
– More concrete (what they evoke is easier to imagine, see, hear, feel,
small or taste);
– Related to bodily action;
– More precise (as opposed to vague);
– Historically older;
Basic meanings are not necessarily the most frequent meanings of the
lexical unit.
(c) If the lexical unit has a more basic current–contemporary meaning in
other contexts than the given context, decide whether the contextual
meaning contrasts with the basic meaning, but can be understood in
comparison with it.
4. If yes, mark the lexical unit as metaphorical.
The practical application of this procedure may be illustrated by the analysis of
three individual words in their contexts from three different texts from the Brit-
ish National Corpus. For the sake of concise presentation at this initial stage of
the paper, we assume that, for each of the presented examples, we have read the
entire text (or a longer fragment), so that we have a general understanding of it
(Step 1 of the procedure), and that we have decided about the demarcation of
lexical units (Step 2).
2.1 Case study one – way
Consider the following excerpt, from BNC text AC2 (fiction):

(1) You’re sixty-one and have been looking forward to retirement. I’m fifty and
have a long way to go.
Step 3a. Contextual meaning

In this context, the meaning of the noun way is ‘a period of time’. It is preceded by
the adjective long, which adds the length quality to the abstract route or the time
of life ahead and it is followed by the verb to go, which indicates abstract advanc-
ing on this abstract journey.
Step 3b. Basic meaning

The basic meaning of way in the two advanced learners’ dictionaries used in our
project, the Macmillan English Dictionary for Advanced Learners (Rundell, 2002)
and the Longman Dictionary of Contemporary English, is ‘the particular road,
path, or track that you use to go from one place to another’. We take this meaning
to be basic because it is the more concrete sense.
Step 3c. Contextual meaning vs. basic meaning

This step raises the question whether the contextual and the basic senses are dis-
tinct. The answer to this question is affirmative: the contextual sense of way is ab-
stract and contrasts with the basic meaning which is concrete and physical. This is
reflected by our main source of reference, the Macmillan dictionary, which makes
a distinction between the main sense of road, path or track and the subsense of
time as two separate descriptions.
Step 3d. Contextual meaning vs. basic meaning

This step raises the question whether the two senses can be seen as understood
in comparison with each other. The answer to this question is also affirmative:
an abstract road along which you proceed in life is like a physical road along
which you go (this is the reason why some linguists see this as an instance of
LIFE IS A JOURNEY, Lakoff & Johnson, 1980, 1999, but to us this is another
type of analysis and decision; see Steen, 2007; Steen et al., 2010).
Step 4. Metaphorically used or not?

Yes, the contextual sense of way is distinct from the basic sense of way but it is
understood by comparison to the basic sense. Way is a metaphorically used word
in this sentence.
2.2 Case study two – bet
Now consider the following excerpt, from BNC file KB7 (a conversation):
(2) I bet you feel like a [?] now, give them in when they can’t truss them dunno.

The verb to bet in this sentence means that the speaker is sure about something
he or she is saying.

In its basic sense, the verb to bet is used with reference to money; to bet money
means to risk money by trying to predict the result of a race, game, competition.
This sense is basic because it is a concrete sense which involves, for instance, the
transfer of money.

The two senses are distinct: in the given context, the verb to bet has nothing to
do with money, so it is distinct from the basic sense. The Macmillan dictionary
lists the contextual use as a fixed phrase which is described as distinct from the
basic sense.

The two senses are understood by comparison: making a statement about some-
thing you believe is like betting money on the result of a race, game or competi-
tion (this is the reason why some linguists see this as an expression of LIFE IS
A GAMBLING GAME, Lakoff and Johnson, 1980, 1999, but again, to us that is
another matter; see Steen, 2007; Steen et al., 2010).

Yes, we can understand the meaning of the contextual bet by comparing it to the
basic sense of the verb, and the two senses are distinct.
2.3 Case study three – attack
Our third example is (3), from BNC news text A1H:

(3) He fearlessly attacked convention, which caused problems when he pitched
into established reputations.

In this context, the verb attacked indicates the expression of strong criticism to-
wards an idea.

The basic meaning of the verb to attack is to use violence to harm a person or to
use weapons to try to defeat an enemy. This involves concrete physical interaction,
whereas argument does not.

The two senses are distinct: the contextual sense of attack in this sentence dif-
fers from the basic sense of the verb. The dictionary lists these two senses as two
separate descriptions.

The two senses are understood by comparison: verbal attacking is like physical
attacking.

Yes, the contextual sense of to attack is distinct from the basic sense of this verb
but they are understood by comparison.
The three examples show how MIP should be applied in researching metaphors in
discourse. There are, however, specific complications which arise at each step of
the procedure. That is why the major aim of the detailed discussions in the sections
to follow will be to report on how we have made our own decisions when dealing
with these difficulties in order to make them explicit and open for discussion.
3. More about demarcating lexical units
The second step in MIP is to demarcate every lexical unit in the text as a precon-
dition for being able to establish the contextual and basic meaning and possible
metaphoricity of the unit. In practice, a single lexical unit more often than not
consists of only one word. In a number of instances, however, two or more words
can combine as one lexical unit, and will have to be analysed for metaphoricity as
such. The most important issue for classifying multi-word units as single lexical
units is whether a combination of words designates only a single referent in the
projected text world, and hence evokes one concept. This is most conspicuously
the case with phrasal verbs, compounds and so-called polywords, as the examples
below will show. For each of these categories, we will give some examples to il-
lustrate how we operationalize our ideas regarding units of analysis.
Phrasal verbs have to be demarcated as single lexical units, since, as the
Pragglejaz Group suggests, they cannot be decomposed without loss of meaning.
More importantly, the combination of verb and particle designates one referent in
the projected text world. The phrasal verb take off, in a sentence such as the plane
takes off, constitutes a meaning that cannot be arrived at simply by combining the
senses of the individual parts. This specific combination of verb and particle de-
notes one action, that of a plane gathering speed and going up into the air, which
cannot be deduced from the separate parts of the phrasal verb. The example sen-
tence shows that the parts combined designate one referent in the text world, and
activate one concept. Therefore, it will be demarcated as one lexical unit.
The same can be said of, for instance, show up. In a context such as my friend
did not show up at the party, the phrasal verb signifies one referent in the text
world, namely the action of not arriving at a particular place. We cannot arrive
at this specific meaning by simply combining the semantic meaning of the sepa-
rate words either. We therefore take this combination of verb and particle as one
lexical unit. To obtain further support whether combinations such as take off and
show up are indeed phrasal verbs, the part-of-speech tags inserted in the BNC-
files can be consulted. If the word class annotations in the file exhibit a combina-
tion of a verb and an adverbial particle, we may conclude that these function as
parts of a phrasal verb.
Not only phrasal verbs are demarcated as single lexical units. Conventional-
ized compounds that are spelt as two separate words are analysed in a similar
manner, and this is in accordance with the suggestions made by the Pragglejaz
Group. For us it is also important that compounds designate single referents in a
specific text world.
The most frequently occurring compounds are those made up of two nouns,
such as stock market. This compound refers to one specific phenomenon in the
domain of finance, and can be seen as designating one specific referent. If it

did not, market might become eligible for annotation as a metaphorically used
word. Compounds spelled as hyphenated words or solid words, such as alpha-
numeric, do not present a problem and can be demarcated as single lexical units
in the same manner.
It is often hard to tell just by looking at the meaning when we are dealing
with a compound, and thus one lexical unit, or with more than one lexical unit.
One sign of a combination being a compound is that these combinations have the
primary stress on the first part, in cases such as stock market. In order to obtain
support for the occurrence of a compound, we utilize the Macmillan English Dic-
tionary for Advanced Learners. This dictionary lists the pronunciation, and thus
also the stress patterns, for each word, compound and other expressions. There-
fore, when in doubt as to whether we are dealing with a compound or not, we can
consult the stress patterns listed in the dictionary. Whenever the primary stress is
on the first part, we are dealing with a compound and consequently with a multi-
word single lexical unit. Whenever the primary stress is on another part than the
first, we are simply dealing with two or more frequently combined words.
One last frequently occurring category of single lexical units consisting of
more than one word is that of so-called polywords, such as let alone and of course.
Following Pragglejaz again, we also treat polywords as expressions which are not
decomposable and designate one single aspect of the text world. Although a great
number of these expressions are not listed separately in the dictionary, they can be
found with the help of a finite list published as part of the BNC. Whenever there
is doubt as to whether an expression is a polyword, this list is consulted. If an ex-
pression is indeed listed as a polyword, it is demarcated as a single lexical unit. If,
however, an expression is not included in the list of polywords, the separate parts
of the expression will each be demarcated as lexical units and will have to be an-
alysed for metaphoricity individually.
As far as idioms and other fixed collocations are concerned, we will anal-
yse each individual word in the expression for possible metaphoricity. Although
some fixed collocations are explicitly included as an entry in the dictionary, we
will analyse each part in order to establish the semantic roles of each of the con-
tent words. In addition, idioms and fixed collocations are typically decomposable,
as has also been pointed out by the Pragglejaz Group, meaning that the senses of
the separate items to some extent contribute to the overall meaning of the expres-
sion. Each unit thus typically designates a referent in the text world, in contrast to
for instance compound nouns and phrasal verbs. For instance, the phrase treading
water figuratively refers to a life-saving activity (treading) in an environment (wa-
ter) by some agent (the person in the water). Therefore, each separate unit of the
expression will be demarcated as a single lexical unit, and will have to be analysed
for metaphoricity as such.
4. More about contextual meanings and basic meanings
Now that we have established what, according to our application of the Pragglejaz
procedure, counts as a lexical unit in a text or piece of discourse, we should re-
turn to the third step of the MIP in which the contextual meaning of a lexical
unit is juxtaposed to a more basic one, from which we then continue to analyze
whether these two meanings can be contrasted as well as understood in compari-
son with each other. This may result in a positive judgment on the metaphoricity
of a lexical unit.
4.1 Contextual meanings
Contextual meaning has been identified by the Pragglejaz Group as “how it ap-
plies to an entity, relation or attribute in the situation evoked by the text. [...] Take
into account what comes before and after the lexical unit” (p. 5). Thus, the con-
textual meaning is the referential meaning a lexical unit has considering context
and co-text. This may be a conventionalized meaning, which we do find in the
dictionary. It may also be a novel or specialized meaning, which we do not find in
the dictionary. Thus, in a straightforward sentence like,
(4) If necessary, you may attack us after reading this article
the contextual meaning of attack would most likely be ‘criticize’ or ‘comment at’.
As easy as this may seem, in practice MIP is confronted with quite a few prob-
lems. To begin with, there are cases of contextual ambiguity. Sometimes more dif-
ferent meanings can be applicable at the same time. In the following example,
(5) It was Paula’s job to show samples, parading slowly up and down in front of
the clients as they sat on the elegant spindle-leg chairs taking in every detail
of the garments with a critical eye. (BNC-Baby: BMW)
up and down may indicate a literal change in height but just as well a simple hori-
zontal movement, for example from side to side. The one interpretation is literal,
the other is not and may influence whether it is judged metaphorically used or
not at a later stage. Similar cases of ambiguity may be encountered with idioms
that can be taken both literally and metaphorically at the same time or unclear
anaphoric reference in which case e.g. pronouns can refer to both concrete and
abstract items. In these situations, we choose to keep possible metaphoricity in as
WIDLII (When In Doubt Leave It In).
Context is also unclear and hence potentially ambiguous in the case of un-
finished sentences, a characteristic of conversations. The Pragglejaz Group men-
tion that “when a word is spoken only partially, such as when one restarts one’s
tterance, and is transcribed as a word fragment, the researcher may be able to

u
infer from the context the entire word that was likely intended. Does one count
such a word as a lexical unit?” (p. 23). In our procedure, we avoid such a word or
stretch of text and write it off as DFMA (Discard For Metaphor Analysis) since it
seems unfair to analyze some incomplete material but not all of it.
A second problem arises when two different analyses seem equally plausible.
A good example is personification, e.g.
(6) Labour hopes to transform the situation. (BNC-Baby: A1F)
Since a political party cannot ‘hope’ like human beings can, the party is per-
sonified. However, analyzed through metonymy, Labour stands for the people
making up that party, through which the sentence would not be metaphorical
(cf. Low, 1999). Since this last argumentation appeals to analysis at a conceptual
level, our procedure prefers the first one and saves personification as possibly
metaphorical.
A final problem in establishing the contextual meaning of a lexical unit is a
possible lack of specialized knowledge, for example in academic texts. The Mac-
millan dictionary we use for our research project is a dictionary for general pur-
poses, which does not always contain these specialized definitions. In such cases
we might turn to our own common sense to interpret the specific meaning but we
may also decide to add a specialized dictionary to our project.
When establishing the contextual meaning, we are, thus, very much aware of
the different interpretations possible. When there is ambiguity without any dra-
matic lack of contextual knowledge (such as with aborted utterances) and when
one way of reasoning convincingly leads to metaphoricity, we keep the lexical unit
in as possibly metaphorical.
4.2 Basic meaning
In order to determine whether a word is metaphorically used, we need to compare

the contextual meaning to a basic one. In our procedure, we take a synchronic
notion of basic sense as our starting point, in which basic meaning is the most
concrete, specific and human-oriented sense you can find in the dictionary with-
in one grammatical category. In sentence (4) above, the basic meaning of attack
would then be ‘to use violence to harm a person, animal, or place’.
This definition of basic meaning is similar to the Metaphor Identification Pro-
cedure in all but one respect: where the Pragglejaz Group ignore grammatical
word class boundaries and subcategories, our own application of the procedure
does leave room for such a distinction and treats the noun and verb of e.g. dog as
distinct lexemes between which there cannot be a metaphorical relationship in

language use. In particular, when one uses the verb to dog, there is a conventional-
ized relationship between the lexical unit, the concept that it presumably activates
(an action), and the referent which this designates in the projected text world (‘to
follow obstinately’). This semantic analysis can be carried out without recourse to
the meaning of the noun. That there is this connection between verb and noun
in the language code (and in the dictionary), because of a conventionalized mor-
phological conversion, is another matter, which is irrelevant for the analysis of the
role of the verb in most contexts.
Similarly, where MIP treats “words with identical base (noninflected) forms as
the same lexical unit” (p. 28), we make a distinction between, for example, count
and uncount nouns, or intransitive, transitive and linking verbs. Consequently,
the intransitive use of a verb like to prove (to make an official request for a job or
a place in a college or university, or for permission to do or have something) can-
not have a metaphorical relationship in use with its transitive counterpart (to put
a layer of something such as paint onto a surface).
The reason for deviating from MIP is that we are studying metaphor in dis-
course, not in language in general. Our focus is on which distinct referent is ex-
pressed by a word within a context and topic which have been chosen by an author
or speaker who is aware of the different tools language offers. We are not denying
that the one grammatical subcategory or word class can be metaphorically de-
rived from the other. This is, however, a different kind of research, which has to do
with the role of metaphor in the morphological structure of a language.
A final remark concerns the importance of stylistic, dialectal, historical or
geographical variation on establishing the basic meaning. Specific dialectal and
scientific terms may often have only one obvious meaning within their genre or
register, appearing to be homonyms to specialized readers, but outside this con-
text, in everyday language, these same terms are polysemous. Here, we decided to
stick to our general criteria: the issue is not whether specialized readers do or do
not realize the metaphorical use of a lexical unit at a conceptual level; what mat-
ters is whether in our normal everyday language there would be a metaphorical
derivation involved within a grammatical category.
The synchronic notion of basic sense as the embodied sense is the boldest as-
sumption existing at the moment within metaphor theory. Other linguists might
prefer a diachronic point of view of basic as being the historically oldest sense of
the word. In the following section we will set out a comparison between these two
approaches and indicate the implications of our own preference for metaphor
identification.
5. The role of etymology: Diachronic versus synchronic

approaches to meanings
When basic meanings are defined as the most concrete and human-oriented ones,
they also tend to be the historically older meanings. In fact it is one of the main
tenets of Cognitive Linguistics that human abstract reasoning is grounded in
our concrete bodily experiences and our interaction with the environment that
surrounds us. Abstract or more complex things tend to be understood in terms
of more concrete and simpler models. This suggests that concrete and simpler
meanings precede abstract and complex ones, so that the most basic, that is, con-
crete and human-oriented, meaning of a lexical unit is likely to correspond with
its oldest sense.
Consider, for example, the body part name head which has extended its
meanings to denote more abstract phenomena (data from the Oxford English
Dictionary).
Head
c. 825 → t op part of the body containing the skull, mouth, eyes, ears, nose and
brain
c. 1000 → any rounded part of a plant (e.g. a head of cabbage, lettuce, garlic)
c. 1300 → the top or front part of something (e.g. the head of the bed, of the
table)
c. 1551 → leader or most important person of a group
Practically all body-part names can operate metaphorically and metonymically
in language. The head is not only the top part of the body containing the skull,
mouth, eyes, ears, nose and brain, but it also enables us to conceptualize some
plants as presenting a similar shape, so that head becomes a cabbage. The top and
front parts of some objects may also be called the head, such as the head of the
bed (due to its physical orientation). There may also be a correlation with placing
important things on top positions (e.g. vassals used to kneel before the king and
in most companies the directors’ offices are located in the upper part of a build-
ing), so that we can refer to the head of a company.
Nevertheless, older meanings do not always necessarily correspond with
the most concrete ones. Sometimes a lexical unit originates with a highly ab-
stract sense that, as time goes by, gives way to more concrete meanings. Such is
the case with the term reinforce, from Old French enforcier. A look at a diction-
ary based on a fairly recent corpus of contemporary English such as Macmillan
shows four senses:
1. to make an idea, belief, or feeling stronger;

1a. to make a situation, process, or type of behaviour stronger and more like-
ly to continue;
1b. to make a building, structure, or object stronger;
2. to make a group of soldiers, police etc stronger by adding more people or
equipment.
Following the embodiment view the most concrete human-oriented sense of rein-
force is (b), that is, the idea of making something physical stronger. Yet a histori-
cal dictionary like the OED shows a different story. The first sense attested is the
military one (c. 1600), whereas the more concrete meaning emerges almost one
century later (c. 1692).
Cases of counter-intuitive mappings in which abstract senses have led to con-
crete ones are not very frequent. Because most speakers are not familiar with the
history of the language, we rule out historical motivation as a separate factor in
our research. Its overlap with our current approach needs to be checked in inde-
pendent future work. For now, we favour the embodiment view when determin-
ing what counts as the basic meaning of a word.
Another complication in applying a historical criterion when establishing ba-
sic meanings results from language change. At times the basic meaning of a word
has got lost and all contemporary senses are simply metaphorical extensions of
that original sense. Such is the case, for instance, with issue and stage.
Issue
Macmillan
1. a subject that people discuss or argue about, especially relating to society,
politics, etc.
2. a magazine that is published at a particular time
3. a set of things, for example shares in a company, that are made available to
people at a particular time
4. the act of officially giving something to someone
5. someone’s children
OED: way out, exit

Taking into account the etymological sense of exit, it is not difficult to under-
stand the senses listed in the dictionary as metaphorical derivations. The notion
of ‘subject’ (sense 1) can be seen as ‘words or ideas coming from people’s mouths
or minds’, whereas the idea of ‘magazine’ (sense 2) may be interpreted as ‘a series
of papers coming out of a printer’. Likewise, the notions of ‘set of things available
to people’ and ‘the act of giving something’ can be understood through transfers
going from one person to another and, finally, ‘child’ might be explained in terms
of ‘a baby exiting the mother’s body’. However, all these metaphorical extensions
stem from a common word which no longer exists in English. Therefore, speak-
ers are not likely to find connections among the contemporary senses. This may
justify our view of considering the aforementioned senses as independent from
each other.
Stage
Macmillan
1. a particular point in time during a process or set of events
2. the part of a theatre where the actors or musicians perform
3. the place or situation in which something happens, especially in politics
OED: a raised floor, platform, building

The case of stage is similar, since its first sense of ‘raised platform’ fell out of use
and all contemporary usages are simply metaphorical derivations. However, un-
like with issue, speakers may be able to take the notion of ‘part of a theatre’ as the
basic one, and therefore draw possible metaphorical connections with the other
meanings. Through folk etymology, then, today’s English speakers may make
sense of ‘a particular point in time’ and ‘the place or situation in which something
happens’ as stemming from a spatial conceptualization of time and, therefore,
locate a part of a theatre in a time-frame and project the idea of theatrical perfor-
mances to real-life events, respectively.
Judgments about the metaphoricity of a lexical item may consequently vary
because of historical variation. Determining what counts as the basic meaning
of a lexical unit will depend on whether a diachronic or synchronic approach
is adopted. If language is seen as an evolving semiotic system, the oldest sense
might take primacy in deciding what counts as the basic meaning of a lexical unit.
However, because our aim is to mark as metaphorical all those words that have a
potentially active metaphorical basis, we are in need of a synchronic description
of the language. Therefore we are basing our decisions on a contemporary users’
dictionary that is based on corpus research.
6. More about distinctness and similarity
Once both the contextual and the basic meaning of the lexical unit have been es-
tablished, the next step in the MIP is to determine whether the contextual mean-
ing is sufficiently distinct from the basic meaning and whether the contextual
meaning can be understood in comparison with the basic meaning. In order for
the lexical unit to be used metaphorically, the contextual and basic meaning have
to be related by some form of non-literal similarity. If we take our example sen-
tence (4), we can see that the contextual meaning of attack, ‘to strongly criticize
someone or something for their ideas or actions’, is sufficiently distinct from the
basic meaning ‘to use violence to harm a person, animal, or place’ since the for-
mer is verbal and abstract whereas the latter is physical and concrete. The con-
textual meaning also needs to be understood in relation to the basic meaning,
because criticizing someone is like launching a verbal battle against an opponent.
Complications with both requirements will be discussed in more detail below.
6.1 Distinctness
Essential to MIP is the presence of a contrast between the contextual and the
basic meaning which allows the analyst to treat them as independent entities. In
our operational definition, different senses of a specific lexical unit are consid-
ered to be sufficiently distinct if they have been given separate, numbered sense
descriptions in the Macmillan dictionary. This entails that the lexical unit has
to be polysemous. If different senses have been included under the same sense
descriptions (indicated by letters instead of numbers) they are considered to be
different manifestations or applications of the same core meaning and they can-
not therefore be compared.
However, different senses may be conflated under the same sense description
due to Macmillan’s status as an advanced learners’ dictionary, for reasons of econ-
omy or space, or to help learners see patterns and connections between senses. For
instance, Macmillan sometimes conflates concrete and abstract senses, human and
animal senses, or animate and inanimate senses. When we are confronted with such
entries in Macmillan and we suspect that the dictionary may have been compro-
mising, we refer to a second dictionary, the Longman Dictionary, to check whether
there is reason to believe that we are dealing with distinct senses after all. If Longman
demonstrates a similar pattern of conflation, then the senses may indeed be differ-
ent manifestations of the same meaning; but if Longman separates the senses, then
we overrule Macmillan and treat the different senses as sufficiently distinct, making
note of this decision by adding a separate code in the annotations.
This practice may be worrying to some, as it may suggest that it is our pre-
conceptions rather than the data and the tools that lead us to a decision. But it
simply goes to show that we use our tool critically instead of blindly. Moreover,
our strategy of overruling Macmillan by Longman is only applied if it can support
the decision to include potential cases of metaphorical word use: these will always
have to be further analysed at a later stage from a conceptual point of view, so
that posthoc checking against other criteria is assumed to take place. Finally, our
strategy has turned out to be necessary in a small minority of cases.
One example of such conflation in Macmillan is the verb groom. In Macmillan
this verb has received the following sense description:
groom
Macmillan
1. to clean and brush an animal, especially a horse or a dog.
1a. if an animal grooms itself or another animal, it cleans itself or another ani-
mal.
1b. to look after your appearance by keeping your hair, body, and clothes clean
and tidy.
2. to prepare someone for a particular job or activity by giving them special
training and attention.
Yet in Longman the division of senses is as follows:
groom
Longman
1. to clean and brush an animal, especially a horse.
2. to prepare someone for an important job or position in society by training
them over a long period.
3. to take care of your own appearance by keeping your hair and clothes clean
and tidy.
4. if an animal grooms itself or another animal, it cleans its own fur and skin or
that of the other animal.
Macmillan conflates human and animal senses here, with only two separate sense
descriptions that can be compared for metaphorical usage. Longman, however,
makes four distinctions, allowing for several different mappings. The abstract
sense ‘to prepare someone’ would be a candidate for metaphorical usage in both
dictionaries, but only in the case of Longman can humans grooming themselves
and an animal grooming an animal be seen as potentially metaphorical. In this
case we would therefore overrule Macmillan and take the human+animal sense
(Longman sense 1) as basic.
6.2. Similarity
If the contextual and the basic meanings are sufficiently distinct, the next step is
to determine whether they are related by similarity without identity. Identity nec-
essarily implies similarity because the entities involved are in essence instances
of the same thing. Metaphorical mappings, on the other hand, require entities
that are not essentially the same thing to still be similar. This similarity typically
homes in on external or functional resemblances, such as attributes and relations,
between the concepts designated by the contextual and basic meaning.
It may well happen, though, that the contextual meaning and the basic mean-
ing have distinct sense descriptions in the dictionary but are still within the same
conceptual domain, and are therefore not similar but related by means of another
kind of relationship, such as synecdoche or metonymy. One example is the noun
appeal, in a sentence such as
(7) An impassioned appeal for Labour to back electoral reform was made by a
leading Shadow Cabinet member yesterday. (BNC-Baby: A1J)
Here we have a general and vague contextual meaning, ‘an urgent request for
people to give you something you need such as help, money, or information’, that
is sufficiently distinct from the more specific basic meaning, ‘a formal request
for a court of law or similar authority to change its decision’. However, the senses
move from abstract to abstract and stay within the domain of ‘requests to some-
one’, with only the ‘someone’ changing from originally God to the king to the law
to anyone in authority to anyone at all. We would therefore consider this to be a
case of generalization without metaphor.
In other cases an abstract contextual meaning can be contrasted with a suffi-
ciently distinct, concrete meaning but the relation between the two does not seem
to work via similarity. If it is unclear how they are related, we must check the OED
to see if they were both derived from a common source which has disappeared
from the language. This original meaning may support the decision that both
sense descriptions are in fact equally basic in the current state of the language.
One example is the adjective right, for which Macmillan gives inter alia the fol-
lowing senses:
1. if something is right, it is correct according to the facts

2. morally correct
3. in the position, state, or situation that you would normally expect someone or
something to be
4. on or relating to the side of your body that is towards the east when you are
facing north
These senses are all related, but there is not one clear basic meaning onto which
the others can be mapped via similarity. The original meaning ‘straight; not bent,
curved, or crooked in any way’ has disappeared; these senses would therefore all
be treated as equally basic.
Finally, an abstract contextual meaning can sometimes be contrasted with
more than one concrete meaning. For instance, consider position in:
(8) Kahlo’s adoption of Tehuana dress […] asserted both a feminist and an anti-
colonialist position. (BNC-Baby: A6U)
The contextual meaning of the noun position – ‘an opinion about an important
issue’ – can be compared with both ‘the way that someone’s body is placed’ and
‘where something is in relation to other things’. We would therefore mark po-
sition as metaphorically used, making note of the different possible mappings.
These mappings will be determined in the next stage of our project, when con-
ceptual structures are analysed. This example therefore also demonstrates that
the different phases of metaphor analysis can be performed independently, and
that linguistic metaphor can be found without the need to establish a precise
conceptual mapping.
7. Concluding remarks
Finding metaphor in natural discourse is not an easy task. The Pragglejaz meth-
od has offered some help to metaphor researchers who aim to be more explicit,
systematic, and reliable than is customary in much metaphor research. It has
provided a useful starting point for the corpus-linguistic work which we have
done in our two research programmes at the VU University Amsterdam. Its ap-
plicability to large samples of British English and Dutch are promising indica-
tions of its more general use.
Our practical experience has also suggested one or two issues which we have
had to solve in different ways than those proposed by the Pragglejaz Group. One
issue has to do with the limitation of lexical units to their relevant grammatical
category. We do not treat entire lemmas as lexical units, but only consider as units
those grammatical categories and subcategories which can be used to express the
same type of referent in discourse (verbs to indicate actions or processes, nouns
to express entities, and so on).
Another issue has to do with the historical dimension of language. In their
definition of basic meanings, the Pragglejaz Group have listed concrete human-
oriented experience in one breath with historically older meanings. Although this
is a frequent combination, not all historically older meanings are also the more
concrete ones, so that the two criteria may yield different results for the analyst
who needs to decide about what counts as the basic meaning of a word. In our
work we have emphasized the synchronic approach without denying that the role
of the history of language needs to be verified at a later stage.
We have also hinted at other issues that need to be resolved. This chapter
is not intended as a technical manual, however, but rather as a short practi-
cally-oriented guide which focuses on what we see as the key issues. We hope,
though, that our aim of promoting the use of the method has been achieved
without losing sight of the critical if constructive spirit in which such an ap-
plication has to be pursued.
Acknowledgements
This research was sponsored by NWO-Vici Grant 277-30-001, ‘Metaphor in dis-

course: Linguistic forms, conceptual structures, cognitive representations’, award-
ed to the first author.
References
Cameron, Lynne (1999). Identifying and describing metaphor in spoken discourse data. In
L. Cameron & G. Low (Eds.), Researching and applying metaphor (105–132). Cambridge:
Cambridge University Press.
Press.
Lakoff, George & Mark Johnson (1999). Philosophy in the flesh: The embodied mind and its chal-
lenge to western thought. New York: Basic Books.
Longman dictionary of contemporary English online (2008). Essex: Pearson Education Limited.
Low, Graham. (1999). ‘This paper thinks ...’: Investigating the acceptability of the metaphor AN
ESSAY IS A PERSON. In L. Cameron & G. Low (Eds.), Researching and applying metaphor
(221–248). Cambridge: Cambridge University Press.
Pragglejaz Group (2007). A practical and flexible method for identifying metaphorically-used
words in discourse. Metaphor and Symbol, 22, 1–39.
Rundell, Michael (Ed.) (2002). Macmillan English dictionary for advanced learners. Oxford:
Macmillan publishers.
Steen, Gerard J. (2001). A reliable procedure for metaphor identification. In J. Barnden, M. Lee
& K. Markert (Eds), Proceedings of the workshop on corpus-based and processing approaches
to figurative language – Held in conjunction with Corpus Linguistics 2001 (67–75). Lan-
caster: University Centre for Computer Corpus Research on Language.
Steen, Gerard J. (2002). Metaphor identification: A cognitive approach. Style, 36(3): 386–407.
Steen, Gerard. J. (2005). What counts as a metaphorically used word? The Pragglejaz experi-
ence. In S. Coulson & B. Lewandowska-Tomaszczyk (Eds.), The literal-nonliteral distinc-
tion (299–322). Frankfurt am Main: Peter Lang.
Steen, Gerard J. (2007). Finding metaphor in grammar and usage: A methodological analysis of
theory and research. Amsterdam & Philadelphia: John Benjamins.
Steen, Gerard J., Aletta G. Dorst, J. Berenike Herrmann, Anna A. Kaal, Tina Krennmayr &
Trijntje Pasma (2010). A method for linguistic metaphor identification: From MIP to
MIPVU. Amsterdam & Philadelphia: John Benjamins.
chapter 10
Mapping principles for conceptual metaphors
Kathleen Ahrens
Hong Kong Baptist University
Issues surrounding novel metaphor comprehension are not well understood.

In order to address this problem, this paper proposes the Conceptual Map-
ping Model, which puts forward the idea that examining the linguistic map-
pings found in a particular source-target domain pairing allows hypotheses
to be formulated regarding the underlying reason for this conceptual pairing.
These hypotheses are formulated as Mapping Principles and allow for predic-
tions regarding novel metaphor usage to be made. Four experimental studies
on native Mandarin Chinese speakers demonstrate that novel metaphors that
follow Mapping Principles receive significantly lower acceptability and inter-
pretability ratings than conventional metaphors, as well as longer listening
times when making these judgments. In addition, novel metaphors that do not
follow Mapping Principles receive significantly lower acceptability and inter-
pretability ratings than novel metaphors that do follow Mapping Principles, as
well as longer listening times. In sum, the Conceptual Mapping Model proposes
that a contrastive linguistic analysis of conventional conceptual metaphors can
increase our understanding as to why conceptual domains are linked, which in
turn will facilitate our ability to predict and model how conceptual metaphors
are processed in both conventional and novel usages.
Keywords: conceptual metaphors, Mapping Principles, metaphor processing,

ratings tasks, reaction time studies, Conceptual Mapping Model
1. Introduction
Increasing interest in real-world application and understanding of conceptual

metaphors has heightened the need for a model of metaphor usage that can predict
varying comprehension levels of metaphor among native speakers. Although Con-
ceptual Metaphor Theory (Lakoff & Johnson, 1980, 1999; Lakoff, 1993) has deep-
ened our understanding of the pervasiveness of metaphor in our language and
cognitive system, the theory has focused on explaining the underlying conceptual
186 Kathleen Ahrens
scenarios involved in conventional metaphors (cf. Lakoff, 1993: 206). While doing

so has greatly expanded our understanding of the cognitive basis of conceptual
metaphors, it has also left the theory open to criticism that there is an ‘anything
goes’ aspect to metaphor understanding and interpretation (Tsur, 1999). Research-
ers have proposed various models to account for the linguistic data more precisely
(Clausner & Croft, 1997; Grady, 1997), yet these models fall short of predicting
how conventional and novel metaphors will be comprehended by speakers.
Clausner and Croft’s (1997) approach is to narrowly constrain the source do-
main so as to limit the mappings that may take place between two domains. They
propose that specifying a metaphor at its appropriate level of schematicity, and
carefully describing the semantic structure, can constrain the concepts that can
correspond between a source and a target domain. For example, they argue that
the metaphor an argument is a buliding is too broad and should instead be
an argument is the structural integrity of a building. The revised con-
ceptual metaphor predicts that concepts such as ‘foundation’ would be used in the
metaphor, but concepts such as ‘chimney’ and ‘plumbing’ would not, because they
are not part of the structural integrity of a building. It is not clear, however, how
to determine the source domain in this approach, nor how to avoid postulating
too narrow a source domain.
Grady (1997) takes a different approach and suggests that theories are
buildings is instead a compound metaphor made up of two primary metaphors:
organization is a physical structure and persisting is remaining erect.
When the compound metaphor is decomposed into primary metaphors, the gaps
in the examples of theories are buildings (i.e. possible instances of the meta-
phor that do not occur) are considered to be “predictable consequences of the
underlying structure” (Grady, 1997: 280). Furthermore, novel extensions are no
longer considered to be capricious or defined in terms of token frequency, but
instead can be “defined in a principled manner” (Grady, 1997: 280).
While I will expand upon the idea that novel extensions of conventional
conceptual metaphors can be defined in a principled manner, doing so through
primary metaphors is problematic for the purposes of examining how speakers
process conceptual metaphors on-line. That is, while there may be valid theoreti-
cal reasons to break compound metaphors into primary metaphors, there are em-
pirical advantages to using basic-level conceptual categories in terms of analyzing
native speaker intuitions, running corpora-based analyses, and running psycho-
linguistic experiments. In addition, I depart from Grady’s (1997) assumptions
that native speaker evidence is ‘inconclusive’ and that corpora-based evidence is
‘merely’ statistical and not essential to theoretical analyses (Grady, 1997: 281). Al-
though Grady (1997) makes a valid point about how a decompositional account is
an improvement to Conceptual Metaphor Theory, and can capture relationships
Chapter 10. Mapping principles for conceptual metaphors 187
between different metaphors, I would like to suggest here that in order to un-
derstand how metaphors are processed, it is also necessary to have a variety of
empirical methods available to determine what metaphors can be considered
novel and what metaphors can be considered conventional, in order to be able
to either falsify or replicate one’s findings. This is especially important, since such
distinctions are fundamental to understanding language processing issues related
to conceptual metaphors.1
One reason that distinctions concerning conventionality and novelty are im-
portant is because work within the broad framework of Conceptual Metaphor
Theory has come under attack by researchers who argue that it cannot explain the
psycholinguistic data (i.e. Glucksberg & Keysar, 1993; Glucksberg, McGlone &
Manfredi, 1997; Glucksberg & McGlone, 1999; Keysar, Shen, Glucksberg, &
Horton, 2000; McGlone, 1996, 2007; Murphy, 1996). McGlone (1996), for exam-
ple, tested the validity of conceptual metaphors in a series of off-line psycholin-
guistic studies and argued that there was no evidence that native speakers use the
knowledge mapping that conceptual metaphors provide. Instead, he proposes that
the attributive categorization hypothesis better accounts for the data. This model
says that metaphors are category-inclusion statements of the form (X is a Y).2
The attributive categorization view explicitly argues against a set of pre-existing
correspondences between two domains, instead arguing that when a metaphor
arises, the source domain provides properties that are then attributed to the target
domain. One flaw in studies that show evidence for the attributive categorization
view is that they do not distinguish between novel metaphors and conventional
metaphors in their linguistic stimuli. It is quite possible that these researchers did
not find evidence to support Conceptual Metaphor Theory because they were not
testing conventional conceptual metaphors.3 Thus, it becomes incumbent on re-
searchers supporting Conceptual Metaphor Theory to give criteria to distinguish
between novel and conventional conceptual metaphors, as well as to distinguish
between metaphors that syntactically constrain a categorization interpretation (of
the type X is a Y) and those that do not.
In what follows, I will present the Conceptual Mapping Model and examine its
limitations. Next, I will discuss the processing predictions suggested by this model
and present off-line and on-line experimental data that support these predictions.
2. The Conceptual Mapping Model
The Conceptual Mapping Model is designed to operationally define a method

to determine the underlying reasons for the source-target domain pairings of a
conceptual metaphor. These reasons, called Mapping Principles, can be tested
188 Kathleen Ahrens
experimentally in both off-line and on-line studies. The main idea is that the
lexemes involved in the conceptual metaphor must be identified and the associ-
ated groupings analyzed.4 Once the lexemes that map for a certain conceptual
metaphor have been analyzed, the underlying reason why a particular target has
selected a particular source domain will be able to be postulated. This underly-
ing reason will be stated in terms of a Mapping Principle. A six-step paradigm is
used to collect and analyze the data, and is demonstrated for idea is building
in Mandarin Chinese below, in order to facilitate comparison with previous work
done on this metaphor in English (Clausner & Croft, 1997; Grady, 1997). These
steps were used to aid in determining the Mapping Principles for all conceptual
metaphors used in the experimental studies that follow. All steps were completed
by a group of five linguistically-trained native speakers who then met to discuss
their answers. Step four required unanimous agreement (regarding convention-
ality); all other steps required four out of five members to agree. Ten additional
linguistically-trained native speakers later verified these analyses.5
In the first step, native speakers generate as many metaphorical examples as
they can think of within a proposed target domain, in this case idea. Second,
each speaker evaluates each example generated and groups it according to its
source domains, such as building, food, commodity or infant. The group-
ings are then discussed and modified if necessary (i.e., should the source do-
main be referred to as ‘baby’ or ‘infant’?). At this point, conceptual metaphors,
such as idea is a building are postulated. Third, three questions are asked
concerning what is known about each source domain in terms of real world
knowledge. This real world knowledge is conceptual. However, this conceptual
knowledge is able to be expressed through the (meta) linguistic expressions as
used below in (1).
(1) Real world knowledge [about buildings]
Q1. What entities does the source domain have?
[foundation, structure, base, model, layout, cement, brick, steel bar,
sandstone, (bamboo) scaffolding, roof, wall, worker, window, door,
plumbing, decoration]
Q2. What qualities does the source domain (or the entities in the source
domain) have?
[shaky, high, short, strong, weak, flimsy]
Q3a. What does the source domain do?
[to protect, to shield, to shelter]
Q3b. What can someone do to (or in) the source domain?
[to live in, to build, to construct, to tear down]
The fourth step involves determining whether the linguistic expressions that have
been generated are conventional or novel through group discussion. If any mem-
ber of the group feels the metaphor is unusual, it is not included in the group of
conventional conceptual metaphors. The examples in (2) below are all considered
conventional. The entities that correspond between the source and target domains
are jiagou ‘frame’ (2a), genji ‘base’ (2b), moxing ‘model’ (2c), cuxing ‘miniature’
(2c), and ge ju ‘layout’ (2d), while the quality that corresponds between the source
and target domains is songsan ‘loose’ as in (2a).
(2a) zhege lilun jiagou hen songsan
this theory frame very loose
‘The framework of this theory is loose.’
(2b) nide lundian genji shi sheme
your argument base be what
‘What is the foundation of your argument?’
(2c) wode zhege xiangfa zhi shao you moxing/cuxing
my this idea just slightly have model/miniature
‘My idea is only beginning to take shape.’
(2d) zhege xiangfa de geju tai xiao le
this idea de layout too small aspect-particle
‘The layout of this idea is too small.’
The functions that correspond between the source and target domains are jiangou
‘build’ (2e), chengxing ‘take shape’ (2f), and dongyao ‘shake’ (2g).
(2e) zhe qun wenren zheng nuli jiangou i tao lilun
This group scholar at the present work hard build one set theory
‘This group of scholars is working hard at constructing a theory.’
(2f) tade sixiang jiagou kuan cheng xing le
his thought framework soon taking shape aspect-particle
‘His thought’s framework is taking shape.’
(2g) tade xiangfa kaishi dongyao
his idea begin shake
‘His idea has begun to waver.’
The fifth step involves analyzing the examples for the linguistic expressions that
occur in these conceptual metaphors (3). These actual mappings are a subset of
the correspondences that exist in the real world.
190 Kathleen Ahrens
(3) Actual mappings/correspondences that exist between idea and building

Q1. What entities does the source domain have that are mapped to the
target domain?
[framework, foundation, model, layout, e.g. (2a), (2b), (2c) & (2d)]
Q2. What qualities does the source domain or the entity in the source
domain have that are mapped to the target domain?
[loose, shaky, e.g. (2a)]
Q3a. What does the source domain do that is mapped to the target
domain?
[shake, e.g. (2g)]
b. What can someone do to (or in) the source domain that is mapped
to the target domain?
[to construct, to take shape, e.g. (2e), (2f)]
Sixth, once the actual mappings have been analyzed, the correspondences that
have been mapped are compared with what could have been mapped (i.e. in the
case of real world knowledge). From the analysis in (3) above, it can be seen that
expressions relating to the concepts of foundation, stability and construction were
mapped. Concepts relating to the position of the building, internal wiring and
plumbing, the exterior of the building, windows and doors were not mapped.
Thus, the target domain of idea uses the source domain of building in order
to emphasize the concept of structure. Buildings can stand because of a founda-
tion and a well-built stable structure. Ideas are deemed worthy if they also have
a solid basis and structure. Thus, when people talk about ideas and want to ex-
press positive notions concerning organization, they may use the source domain
of building to express this idea. The Mapping Principle in this case is therefore
the following:
Mapping Principle for idea is building

Ideas are understood as buildings, in that buildings involve a (physical) structure
and ideas involve an (abstract) organization.
The result for idea is building in Mandarin Chinese is similar to the results
found in English, since the fundamental mapping relates to structure (Clausner &
Croft, 1997; Grady, 1997).
The steps as formulated above are not uncontroversial. They rely a good deal
upon consensus being reached by native speakers at each point. However, recent
advances have been made since these steps were first postulated (in early 2000) in
terms of analyzing metaphors in large-scale corpora (Ahrens, Chung, & Huang,
2003, 2004; Chung, Huang, & Ahrens, 2003; Chung, Ahrens, & Huang, 2004,
2005; Huang, Chung, & Ahrens, 2006). For example, when running step one on a
corpus, a particular lexeme, such as ‘idea’ is chosen, and all metaphorical instances
are manually selected. This narrowing of the target domain is advantageous when
dealing with large amounts of data and allows for additional related lexical items,
such as ‘theory’, to be independently analyzed. The results for both lexemes may
then be compared in terms of their respective mappings and postulated map-
ping principles. The second step, grouping into source domains, can be handled
computationally through either top-down or bottom-up methods (Chung, 2007).
This step can also be verified by using categorization tasks with native speakers.
For example, in two pre-tests discussed below, participants are asked (a) to de-
cide if a lexical item postulated to be from a particular source domain in fact be-
longs in that source domain, or (b) to rate how strongly it belongs in a particular
source domain. Step three involves making decisions about core versus peripheral
knowledge, while step four involves decisions about which expressions are novel
and which ones are conventional. Both steps require native speaker intuitions. In
step five, actual mappings need to be compared to what could be mapped – again,
this judgment is made by a group of native speakers. Step six is currently based
on a comparison between steps five and three. While all these steps currently rely
on consensus among native speakers, recent work on corpora (Huang et al., 2006)
has demonstrated that mapping principles derived in this way are supported by
frequency data. That is, the most frequently occurring lexical item in a source-
target domain pairing is postulated to correspond to the mapping principle. In
addition, I will show below that the linguistic stimuli that are created based on
mapping principles derived in this way also show statistically significant differ-
ences between experimental conditions – differences that we would not expect to
see if this method was substantially flawed.
The Conceptual Mapping Model predicts that there will be two different kinds
of novel metaphors: novel metaphors that follow mapping principles and novel
metaphors that do not follow mapping principles. Thus, under the Conceptual
Mapping Model, metaphors can be grouped into four kinds ranging from most
conventional to most novel: (1) conventional metaphors, (2) novel metaphors that
follow the mapping principle, (3) novel metaphors that do not follow the mapping
principle, and (4) anomalous metaphors, i.e. novel metaphors that use a source-
target domain pairing that rarely, if ever, occur in the language, such as beliefs
are machines in Mandarin Chinese.6 Clausner and Croft’s (1997) approach, on
the other hand, only distinguishes two kinds: (1) metaphors that fall inside the
scope of the source domain, and (2) those that fall outside the scope of the source
domain. The attributive categorization hypothesis, on the other hand, predicts
that there will be no discernible differences between the comprehension of con-
ventional and novel uses of metaphor. Lastly, hypothetical predictions cannot be
formulated for Grady’s (1997) approach since, when the conceptual metaphors
192 Kathleen Ahrens
are broken down into primary metaphors, there are multiple conceptual map-
pings involved for each primary metaphor. Since these are conceptual mappings,
there is no way to ascertain through either native speaker intuitions, or through
frequency-based data from corpora, which particular mapping is most salient. In
what follows, I will present experimental evidence for the Mapping Principles that
are formulated within the Conceptual Mapping Model.
3. Psycholinguistic experiments
The predictions of the Conceptual Mapping Model for the first three kinds of
metaphors mentioned above are as follows: conventional conceptual metaphors
(i.e. exemplars that follow the mapping principle and are common in the lan-
guage) will be treated on par with literal language (Stewart & Heredia, 2002).
Next, metaphors that follow the mapping principle but are novel usages will re-
ceive slightly lower acceptability and interpretability ratings and slightly higher
processing times when making acceptability and interpretability decisions. Lastly,
novel metaphors that do not follow the mapping principle will involve even lower
acceptability and interpretability ratings and even higher processing times.
Using both on-line and off-line experimental methods has the advantage
of testing one aspect of what Gibbs, this volume, calls a real-world approach to
metaphor. In particular, he points out in this chapter and in Gibbs (1994) that re-
searchers need to differentiate between four different types of metaphoric under-
standing: metaphor processing, metaphor interpretation, metaphor recognition,
and metaphor appreciation. Of these four, the first two types are examined here
as the two on-line studies in Sections 6 and 7 look at metaphor processing, which
Gibbs defines as “the fast, mostly unconscious process that lead to metaphor com-
prehension in real-time listening and reading”. Sections 4 and 5 look at metaphor
interpretation, which involves the reflective processes that can be found in off-
line acceptability and interpretability judgment tasks.
4. Off-line acceptability ratings of metaphors
a. Participants
132 National Taiwan University (NTU) undergraduate student volunteers from

four different classes participated in this study. Participants were native speakers
of Mandarin. They had to rate their general proficiency of Mandarin as being 5 or
above (on a scale of 1 to 7) to qualify for the experiment.
b. Materials
Eighteen metaphors were used in the experiment. Each metaphor was first ana-
lyzed and its mapping principle determined by five linguists trained in the Con-
ceptual Mapping Model. Then each analysis and mapping principle was presented
to a group of ten linguists from the Chinese Knowledge Information Processing
Group. Conceptual metaphors found in Mandarin Chinese (i.e. love is a plant)
were used in this experiment only when the analyses and mapping principles
were agreed upon by this latter group.
Each metaphor contains six types of sentence, as given in (4) below. The ex-
amples are derived from the conceptual metaphor love is baggage.
Example (4a) is a conventional metaphor, while (4b) is a ‘literal match’ to
the conventional metaphor example, as it replaces the word ‘love’ with ‘baggage’,
which then allows for a straightforward literal interpretation of the sentence. In
the next set, (4c) is an example of a novel metaphor that follows the mapping
principle, while (4d) is its literal match (i.e. the only difference between these two
examples is that in (4c) ‘love’ is the subject of the sentence and in (4d) ‘baggage’ is
the subject. In the last set, (4e) is a novel metaphor that does not follow the map-
ping principle, while (4f) is a literal match to (4e). In each metaphorical sentence
the source domain is shaded, and the target domain is underlined.
Conventional conceptual metaphor
(4a) zhe aiqing zhen jiaoren wu fa fuhe
the love really make people no way carry
‘People have no way to carry this love.’
Literal match to the conventional conceptual metaphor
(4b) zhe baofu zhen jiaoren wu fa fuhe
the baggage really make people no way carry
‘People have no way to carry this baggage.’
Novel metaphor that follows the Mapping Principle
(4c) zhe aiqing zhen jiaoren wu fa beifu
the love really make people no way bear
‘People have no way to bear this love.’
Literal match to the novel metaphor that follows the Mapping Principle
(4d) zhe baofu zhen jiaoren wu fa beifu
the baggage really make people no way bear
‘People have no way to bear this baggage.’
194 Kathleen Ahrens
Novel metaphor that does not follow the Mapping Principle

(4e) zhe aiqing zhen jiaoren wu fa dakai
the love really make people no way open
‘People have no way to open this love.’
Literal match to the novel metaphor that does not follow the Mapping Principle
(4f) zhe baofu zhen jiaoren wu fa dakai
the baggage really make people no way open
‘People have no way to open this baggage.’
Based on an a priori analysis of the metaphor love is baggage, the proposed

Mapping Principle is: Love is understood as baggage, in that baggage involves car-
rying a physical weight and love involves carrying an emotional burden. (4a) is an
example of a conventional usage of this metaphor, while (4c) is a novel usage,
since ‘beifu’ is not used to talk about love conventionally. However, the novel
example in (4c) does follow the mapping principle for this source-target domain
pairing, since the meaning of ‘beifu’ still has to do with carrying a weight. (4f) is
an example of a novel metaphor that does not follow the mapping principle for
this particular source-target domain pairing, since opening a suitcase is not usu-
ally related to how much it weighs.
In addition, before the acceptability experiment was run, several conditions
had to be met. First, all the metaphorical terms (such as fuhe ‘carry’, beifu ‘bear’
and dakai ‘to open’) had to be considered as being in the source domain (in
this case baggage) by a group of native speakers. Thus, the following two pre-
tests were run: a source domain rating test, and a yes-no source domain test.
31 National Taiwan University (NTU) undergraduate students participated in
the source domain rating tests and 20 NTU undergraduates participated in the
source domain yes/no test. All participants had to meet the language require-
ments described previously. The participants were given the source domain and
were asked if the word next to it was related to it. On the rating task, they were
asked to rate the strength of the relationship from 1 to 7, with 7 being highly
related. In the yes-no task, they were simply asked to circle yes or no. Items were
only included in the acceptability experiment that both had (a) a mean over all
participants of 4.9 or above on the ratings test, and (b) a mean of .7 or above for
the yes/no test.
Moreover, it was also ascertained that key words in (4a) and (4b) (i.e. fuehe
‘carry’), (4c) and (4d) (i.e. beifu ‘bear’), and (4e)–(4f) (i.e dakai ‘open’) did not dif-
fer in terms of frequency of lexical items. The frequency norms are based on CKIP
(1993). The means of the groups are 11667, 11767, and 9996 with an SD of 9354,
8936, 6276 respectively; F(2,47) = .231, p = .795).
The prediction for the Conceptual Mapping model is that the literal sentences
should be equally acceptable (i.e., b = d = f). In addition, the conventional concep-
tual metaphors (a) should be ranked as equally acceptable to the literal sentences.
However, both kinds of novel metaphor should be ranked lower in acceptability
than their literal counterparts (i.e., (c) < (d), (e) < (f)), and there also should be
a gradation in the ranking of all metaphorical sentences, such that (a) > (c) >
(e), since conventional conceptual metaphors should be processed automatically,
and novel metaphors that follow mapping principles should involve additional
processing resources when compared with the conventional conceptual meta-
phors. Furthermore, these resources should be less when compared with novel
metaphors that do not follow mapping principles, since this last kind requires the
listener to make a new connection between the source and target domains, while
the novel metaphors that follow mapping principles require only the activation of
an underlying connection.
Clausner and Croft’s model, on the other hand, predicts that (a) and (c) should
not differ in acceptability because (c) in their model would still fall under the con-
strained source domain. Furthermore, the attributive categorization hypothesis
predicts no difference between (a) and (c) and (e), since in this model, each meta-
phorical mapping is a unique event that is dealt with individually.
c. Procedure
The 108 experimental sentences were divided into six booklets with 18 examples
each using a counter-balanced design, so that no participant saw a sentence from
any one of the conceptual metaphors more than once. The participants were ran-
domly given one booklet containing a set of instructions. They were instructed
to rate the sentences according their acceptability. If they felt the sentence to be
not strange at all and acceptable, then they were instructed to give a rating of 7. If
the sentence was strange and unacceptable, then they were instructed to give the
sentence a rating of 1. If they felt that the acceptability was in between acceptable
and unacceptable, then they were to choose from 2 to 6 depending on the level
of acceptability. Two practice examples were given before they started to rate the
acceptability of experimental sentences.
d. Results
The data from 132 participants were tallied and the means were calculated across
all participants for the six sentence conditions. The means and standard deviation
for each sentential condition are given in Table 1.
196 Kathleen Ahrens
Table 1. Means for acceptability ratings of literal and metaphorical sentences
Sentence type Mean SD
A (conventional conceptual metaphors) 6.0 1.7
B (literal matches to A) 6.0 1.7
C (novel metaphors that followed mapping principle) 5.0 2.1
D (literal matches to C) 6.0 1.5
E (novel metaphors that did not follow mapping principles) 4.0 2.2
F (literal matches to E) 5.4 1.9
For the overall analysis of variance, which consisted of the between-participant

variables of booklets (6) and within-participants variable of Sentence type (meta-
phorical (A, C, E) versus literal (B,D,F)), a significant effect main effect of Sen-
tence Type was found over participants (F(5,630) = 112.5, MSe = .782, p < .05)
and over items (F(5,72) = 11.4, MSe = 1.052, p < .05). Of major importance to the
hypothesis under investigation, a priori planned comparisons were performed on
the metaphorical sentences and their associated literal matches. The conventional
metaphorical sentences (A) did not differ significantly from their literal matches
(B), t = .28, p = 1.0. The novel metaphors that followed mapping principles (C) did
however differ significantly from their literal sentence matches (D), t = 8.9, p < .05.
In addition, the novel metaphors that did not follow mapping principles (E) also
differed significantly from their literal sentence matches (F), t = 13.12, p < .05.
Moreover, further planned comparisons of the three metaphorical sentences
show that the difference in ratings between the conventional metaphorical sen-
tences (A) and the novel metaphorical sentences that followed mapping principles
(C) was significant, t = 9.6, p < .05. Furthermore, the difference in ratings between
the novel metaphorical sentences that followed mapping principles (C) and the
novel metaphorical sentences that did not follow mapping principles (E) was also
significant, t = 9.5, p < .05.
5. Off-line interpretability ratings on metaphors
Interpretability ratings for these sets of sentences were also examined, since the-
ories that treat grammatical judgments and semantic interpretation as distinct
modules would postulate that there might be different patterns involved. For ex-
ample, if interpretability is the question under study, it might be the case that the
novel forms are equally uninterpretable, even though there were gradations in
the differences of their acceptability. However, if the critical issue is the mapping
principles involved, and not the semanticality or grammaticality decision, then
there should be a gradation in interpretability ratings between the three sets of
metaphorical stimuli similar to those found for the acceptability ratings.
a. Participants
An additional 132 participants from NTU were tested in the off-line interpretabil-
ity experiment. The materials and design were exactly the same as in the previous
experiment, except that participants were instructed to rate the sentences accord-
ing to their interpretability, following similar instructions to those given above.
b. Results
The data from 132 participants were tallied and the means were calculated across
all participants for the six sentence conditions. The means for each sentential con-
dition and its related standard deviation are given in Table 2.
Table 2. Means for interpretability ratings of literal & metaphorical sentences
Sentence type Mean SD
A (conventional conceptual metaphors) 6.5 1.1
B (literal matches to A) 6.5 1.2
C (novel metaphors that followed mapping principle) 5.6 1.8
D (literal matches to C) 6.6 1.1
E (novel metaphors that did not follow mapping principles) 4.6 2.2
F (literal match to E) 6.2 1.5
For the overall analysis of variance, which consisted of the between-participant

variables of booklets (6) and within-participants variable of Sentence Type (meta-
phorical (A, C, E) versus literal (B, D, F)), a significant effect main effect of Sen-
tence type was found over participants (F(5,630) = 108.7, MSe = .748, p < .05)
and over items (F(5,72) = 24, MSe = .462, p < .05). Of major importance to the
hypothesis under investigation, a priori planned comparisons were performed on
the metaphorical sentences and their associated literal matches. The conventional
metaphorical sentences (A) did not differ significantly from their literal matches
(B), t = .07, p = 1.0. The novel metaphors that followed mapping principles (C)
did differ significantly from their literal sentence matches (D), t = 9.6, p < .05. The
novel metaphors that did not follow mapping principles (E) also differed signifi-
cantly from their literal sentence matches (F), t = 15.27, p < .05.
In addition, further planned comparisons of the three metaphorical sentenc-
es showed that the difference in ratings between the conventional metaphorical
sentences (A) and the novel metaphorical sentences that followed mapping prin-
ciples (C) was significant, t = 8.6, p < .05. Furthermore, the difference in ratings
between the novel metaphorical sentences that followed mapping principles (C)
and the novel metaphorical sentences that did not follow mapping principles (E)
was also significant, t = 9.4, p < .05.
198 Kathleen Ahrens
6. On-line acceptability decision experiment
The two previous experiments involved off-line (untimed) ratings, and demon-
strated that participants ranked metaphors differently depending on what degree
of novelty was involved: conventional conceptual metaphors (i.e., no novelty was
involved), novel metaphors that followed mapping principles (i.e., some degree
of novelty was involved, as the mapping principle had to be activated to interpret
the new lexical usage), or novel metaphors that did not follow mapping principles
(i.e., a large degree of novelty was involved as the metaphor did not fit in with the
mapping principle associated with this source-target domain pairing and a new
connection and interpretation had to be created.) McGlone (1996) and Glucksberg
et al. (1993) have pointed out, however, that conceptual metaphors are not neces-
sarily accessed in on-line reaction time studies. Under this assumption, one would
not expect to see a gradation in reaction time between conventional conceptual
metaphors, novel metaphors that follow mapping principles, and novel metaphors
that do not follow mapping principles, as predicted by the Conceptual Mapping
Model. Such a finding would falsify the mapping principle hypothesis and instead
lend support to the attributive categorization view. The following two on-line ex-
periments for acceptability and interpretability test this possibility.
a. Participants and materials
92 National Taiwan University undergraduate students participated in this study

for NT$100. All were native speakers of Mandarin as determined by the previously
discussed criteria. They had not participated previously in any related experiments.
The materials included the original 18 sets of stimuli and six additional sets that
underwent similar pre-testing and frequency controls to those described above.
b. Apparatus
The 144 sentences (24 sets of stimuli multiplied by six conditions in each set)
were recorded by a female speaker to the hard drive of an IBM-compatible Pen-
tium computer with the aid of the Creative Wave sound card using the Creative
Wave program. They were counter-balanced over six lists. An internal dedicated
CPU in the button box measured the time from the auditory presentation of the
sentence until a response was made on the button box or 3 seconds had passed,
whichever was earlier. The measurements were made to the nearest millisecond.
The sentences occurred in random order, and there was a three-second delay be-
tween sentences.
c. Procedure
Participants sat in front of a computer monitor in a soundproof room and listened

to instructions read aloud by a native speaker. These instructions explained that
the participants were taking part in an experiment about how language is com-
prehended and that there were no tricks involved. They were then asked to listen
to the sentences on the headphones and to decide if the sentence was acceptable
or unacceptable. The participants made this binary decision with a two-button
button box specifically designed for this purpose. Participants pressed one button
with their index finger if the sentence was unacceptable and pressed the other
button with their other pointer finger if it was acceptable. Half the participants
were assigned to press the right-hand button for ‘acceptable’ and the left hand
button for ‘unacceptable’ and vice versa. The participants were asked to keep their
fingers on the buttons at all times and to respond as fast and accurately as pos-
sible. The computer screen was covered with a piece of dark gray cardboard so
that participants’ attention would be focused on listening to the auditory stimuli.
Although participants were not given a definition of what was acceptable and
what was not acceptable, they were given the opportunity to practice making this
decision before the experiment began as a practice experiment with eight sen-
tences was run to familiarize participants with the procedure, and if they inquired
as to what ‘acceptable’ meant (and only a few did), they were told that they should
use their own judgment.
d. Results
After screening for participants who did not respond more than 15% of the time
(2 participants), there were 90 participants with reliable data, 15 participants for
each list. The outliers from the RT data were trimmed by excluding outliers above
and below two standard deviations from the mean, which resulted in excluding
5.82% of the data.
Table 3 below presents the mean reaction times of all responses in each condi-
tion, as well as the percent considered acceptable for each condition.
A one-way ANOVA was run on individual participants’ data employing
Lists (4) as a between-participants factor (materials counterbalancing factors).
An overall significant main effect of Sentence Type (metaphor versus literal)
was found for participants (F1 (5,420) = 27.40, MSe = 25067, p < .05) and for
items (F2(5, 138) = 6.842, MSe = 28650, p < .05). There was also no effect of List
(F(5, 84) = .378, MSe = 407652, p = .862). Of major importance to the hypothesis
under investigation, a priori one-tailed planned comparisons were performed on
200 Kathleen Ahrens
Table 3. RT means & percent considered acceptable for literal

and metaphorical sentences
Sentence types RT mean Percent considered
(in ms) acceptable
A (conventional conceptual metaphors) 544 85%
B (literal matches to A) 580 82%
C (novel metaphors that followed mapping principle) 691 59%
D (literal matches to C) 616 83%
E (novel metaphors that did not follow mapping principles) 788 40%
F (literal matches to E) 640 77%
the metaphorical sentences and their associated literal matches. As predicted, the
conventional metaphorical sentences (A) did not differ significantly from their
literal matches (B) in terms of reaction time (t(178) = −1.0, p = .159) or accept-
ability level (t(178) = 1.032, p = .152). However, the novel metaphors that fol-
lowed mapping principles (C) did differ significantly from their literal sentence
matches (D) in terms of reaction time (t(178) = 1.657, p < .05) and acceptability
percentage (t(178) = 6.667, p < .05). Novel metaphors that did not follow map-
ping principles (E) also differed significantly from their literal sentence matches
(F) in terms of reaction time (t(178) = 2.907, p < .05) and acceptability percentage
(t(178) = 10.000, p < .05).
In addition, further planned comparisons of the three metaphorical sen-
tences show that the difference in reaction times and acceptability percentages
between the conventional metaphorical sentences (A) and the novel metaphorical
sentences that followed mapping principles (C) was significant, t(178) = 3.555,
p < .05, and t(178) = 7.565, p < .05, respectively. Furthermore, the difference in
ratings between novel metaphorical sentences that followed mapping principles
(C) and novel metaphorical sentences that didn’t follow mapping principles (E)
was also significant in terms of mean reaction time, t(178) = 1.949, p < .05 as well
as in terms of acceptability percentage t(178) = 4.773, p < .05.
7. On-line interpretability ratings of metaphors
a. Participants and materials
96 National Taiwan University undergraduate students who fitted our language

criteria participated in this study for NT$100. They had not participated previ-
ously in any related experiments. The materials and apparatus were the same as
those used in the previous on-line experiment.
b. Procedure
The procedure was the same as in the previous on-line experiment except that in
this experiment they were asked to make a decision as to whether the sentences
they heard were interpretable or not.
c. Results
After screening for participants who were interrupted during the experiment (2
participants), or who had 15% or more of ‘no’ responses (4 participants), there
were 90 participants with reliable data and 15 participants for each list. The outliers
from the RT data were trimmed by excluding outliers above and below two stan-
dard deviations from the mean, which resulted in excluding 6.34% of the data.
Table 4 below presents the mean reaction times of all responses in each condi-
tion, as well as the percent considered interpretable for each condition.
Table 4. RT means and percent considered interpretable for literal

and metaphorical sentences
Sentence types RT mean Percent considered
(in ms) interpretable
A (conventional conceptual metaphors) 382 95%
B (literal matches to A) 402 94%
C (novel metaphors that followed mapping principles) 489 81%
D (literal matches to C) 399 95%
E (novel metaphors that did not follow mapping principles) 575 63%
F (literal matches to E) 416 91%
A one-way ANOVA was run on individual participants’ data employing Lists (4)
as a between-participants factor (materials counterbalancing factors). An overall
significant main effect of Sentence Type (metaphor versus literal) was found for
participants (F1 (5,420) = 30.84, MSe = 16162, p < .05) and for items (F2(5, 138) =
6.767, MSe = 22030, p < .05). There was also no effect of List (F(5, 84) = 1.432,
MSe = 163299, p = .221). Of major importance to the hypothesis under investiga-
tion, a priori one-tailed planned comparisons were performed on the metaphorical
sentences and their associated literal matches. As predicted, the conventional meta-
phorical sentences (A) did not differ significantly from their literal matches (B) in
terms of reaction time (t(178) = −.735, p = .232) or interpretability level (t(178) =
.515, p = .304). However, the novel metaphors that followed mapping principles
(C) did differ significantly from their literal sentence matches (D) in terms of reac-
tion time (t(178) = 2.825, p < .05) and interpretability percentage (t(178) = 5.217,
202 Kathleen Ahrens
p < .05). Novel metaphors that did not follow mapping principles (E) also dif-
fered significantly from their literal sentence matches (F) in terms of reaction time
(t(178) = 4.849, p < .05) and interpretability percentage (t(178) = 7.298, p < .05).
In addition, further planned comparisons of the three metaphorical sentenc-
es showed that the difference in reaction times and interpretability percentages
between the conventional metaphorical sentences (A) and the novel metaphorical
sentences that followed mapping principles (C) was significant, t(178) = 3.330,
p < .05, and t(178) = 5.390, p < .05, respectively. Furthermore, the difference in
ratings between the novel metaphorical sentences that followed mapping prin-
ciples (C) and the novel metaphorical sentences that did not follow mapping
principles (E) was also significant in terms of mean reaction time, t(178) = 2.319,
p < .05 and in terms of interpretability percentage t(178) = 4.475, p < .05.
8. Discussion and conclusion
The findings in the off-line and on-line experiments for acceptability and inter-
pretability verified the predictions of the Conceptual Mapping Model. In particu-
lar, conventional conceptual metaphors were rated as being equally acceptable and
interpretable to literal expressions, and were rated higher than novel metaphors.
Furthermore, conventional conceptual metaphors were rated more highly than
novel metaphors that followed the postulated mapping principles. In addition,
novel metaphors that followed the mapping principles were rated more highly
than novel metaphors that involved the same source domain, but did not follow
the mapping principle.7 In addition, contrary to views that conceptual metaphors
might not be accessed on-line, listening times were faster for conventional meta-
phors as compared with novel metaphors that followed mapping principles when
participants were making acceptability and interpretability judgments. Moreover,
listening times were faster for novel metaphors that followed mapping principles
as compared with novel metaphors that didn’t follow mapping principles when
participants were making these same judgments.8
These findings, along with the fact that the stimuli were tightly controlled in
terms of syntactic structure and lexical frequency, suggest that speakers do access
mapping principles when making interpretability and acceptability judgments
about sentences that contain conceptual metaphors. Sentences with conventional
conceptual metaphors require no additional processing time when compared
with literal sentences, but novel metaphors that follow mapping principles re-
quire more processing time as compared to literal sentences, and novel metaphors
that do not follow mapping principles require even more time. Although similar
processing times do not necessarily indicate that similar processing resources
were used, as in the case of the conventional versus literal examples (cf. Ahrens
et al., 2007), the longer processing times do indicate that there was an increased
processing load depending upon whether or not the novel metaphors followed
the mapping principles postulated.
While Clausner and Croft’s (1997) model might predict a difference between
conventional and novel metaphors based simply on frequency (or perhaps on the
related concept of familiarity), they would not predict a difference between con-
ventional metaphors and novel metaphors that follow mapping principles, since
in their theory novel metaphors that follow mapping principles and the conven-
tional metaphors are both part of the same constrained source domain. The At-
tributive Category Hypothesis, moreover, predicted that there should be no dif-
ference in acceptability between any of the three kinds of metaphor (unless they
also invoke either frequency or familiarity for the conventional metaphors) since
under these models, there is no coherent system of conceptual correspondences
and each metaphoric mapping is unique. In addition, even if frequency or famil-
iarity is invoked to account for the conventional metaphors, their model runs up
against the same constraints as Clausner and Croft’s model since it cannot explain
the differences between the two kinds of novel metaphors, as both these kinds
never occur in the language and are equally unfamiliar.
However, what Clausner and Croft’s (1997) model, Grady’s (1997) model, and
the Conceptual Mapping Model have in common is the assumption that concep-
tual metaphors reflect systematic mappings between source and target domains.
While each model chooses to represent this systematicity in different ways, all
three differ fundamentally from the attributive categorization view, which pro-
poses that a source domain provides properties that are attributed to the target
domain on the fly. The main contribution of the Conceptual Mapping Model is
that it sets up criteria for evaluating, based on native speaker intuition, corpus
data, and psycholinguistic experiments, how this systematicity is represented in
the language and dealt with from a processing point of view.
This is not to say, however, that any of the above models should be discarded.
As discussed in endnote 1, metaphor as a linguistic phenomenon is so complex
that it may be the case that more than one theory is needed to account for how it is
processed (Gibbs, 2006). Gibbs also mentions this in this volume where he notes
that one of the key challenges for the future is to examine whether researchers
who embrace different theoretical positions are perhaps looking at different types
of metaphorical language. That is, it is plausible to think that in cases where there
are no pre-existing conceptual mappings, statements of the type ‘X is a Y’ may
in fact be dealt with as class-inclusion statements as suggested by the attributive
categorization view. These class-inclusion statement metaphors may be processed
in a different manner from conceptual metaphors which occur in more varied
204 Kathleen Ahrens
types of syntactic construction (cf. Grady’s (1999) discussion of resemblance ver-

sus correlational metaphors). Thus, it is suggested that many factors need to be
taken into consideration when testing a model’s processing claims, including the
syntactic frame involved, the metaphor type, the degree of conventionality of the
metaphor, and the task type.
In sum, the Conceptual Mapping Model proposes that by delimiting the source
and target domain to conceptually coherent categories (such that participants can
make judgments about whether an item belongs in, or is related to, that category
or not), and by examining the linguistic evidence (i.e. the entities, qualities, and
functions that map between the source and target domains) or the corpus evidence
for the frequency of mappings, a mapping principle can be formulated which will
explicate the underlying reason for the existing correspondences between these two
structured knowledge domains. The model postulates that it is the mapping prin-
ciple that constrains the possibilities for novel metaphor extensions, and that it may
be the mapping principle that is activated during judgment tasks. Recent research
has also demonstrated mapping principles are activated in on-going integration of
sentences into discourse (Gong & Ahrens, 2007), and further research will examine
if the two kinds of novel metaphor proposed here (novel metaphors that follow
mapping principles and novel metaphors that do not follow mapping principles)
show differing areas of brain activation, in order to further understand how novel
language is comprehended and integrated in on-going language processing.
Acknowledgements
This paper has benefited from discussions with Biq Yung-O, I-wen Su, Chu-Ren
Huang, Mei-chun Liu, Chung-Ping Cheng, Chia-Ying Lee, Adele Goldberg, John
Barnden, Valentina Bambini, Nicoletta Calzolari, Alan Wallington, Petr Simon,
Katarzyna Horszowska and many others, including the participants of RAAM-6.
I would also like to thank the anonymous reviewers of this paper for their help-
ful suggestions. The two off-line studies discussed in this paper were presented
in preliminary form at the First Cognitive Linguistics Conference at National
Cheng-chi University in 2002. I would also like to thank my project assistants:
Hui-Ru Hsiung, Tzu-yin Lai, Dora Lu, Ya-hui Sung, Shu-ping Gong, and Siaw-
Fong Chung for the time and energy they have put into this project. I am also
grateful to the Institute of Computational Linguistics at the Center for National
Research in Pisa, and the Institute of Linguistics at Academia Sinica in Taipei
for providing office space and facilities during the writing of this paper. Lastly, I
would like to thank the National Science Council of the R.O.C. for its financial
support (NSC grant #94-2411-H-002-038 and #95-2918-I-002-007).
Notes
1. In fact, as Gibbs (2006) points out, “the complexity of metaphoric language makes it un-
likely…that any single theory will be capable of explaining how verbal metaphors come into
being, and how they are ordinarily produced and interpreted” (p. 435). Thus, the model be-
ing proposed in this paper should be viewed as one complementary account to both Grady’s
decompositional account and Clausner and Croft’s domain-based account. All three accounts
have as their starting point the idea that Conceptual Metaphor Theory needs to be better con-
strained in order to make more relevant linguistic predictions about conceptual metaphor use.
2. Grady (1999) suggests that primary metaphors are understood in terms of a correlation
between concepts, while other metaphors, such as those of the type ‘X is a Y’, can be dealt with
in terms of a resemblance model. In this paper, I will focus on what Grady calls correlation-
based metaphors, although the conceptual metaphors I discuss are not necessarily primary
metaphors as Grady defines them.
3. Other reasons might involve whether the experimental tasks are conducive to accessing the
processes involved in understanding conceptual metaphors. For example, Gong and Ahrens
(2007) have noted that line-by-line presentation does not allow context to influence decisions
about the appropriateness of conceptual metaphors, while paragraph presentation does. This
would explain, for example, why Glucksberg et al. (1993) did not find evidence of conceptual
metaphor activation in on-line processing, since their presentation method was line-by-line.
4. This notion of a mapping principle is different from the general principles that Lakoff
(1993) postulates, since he hypothesizes that these general principles are part of the conceptual
system, but not part of the grammar or lexicon. However, I am suggesting here that linguis-
tic correspondences can and should be identified in order to identify mapping principles (cf.
Ahrens, 2002; Huang et al. 2006).
5. As the native speakers who generate the stimuli are all studying for their master’s degrees
at National Taiwan University, it is reasonable to assume that the decision as to what consti-
tutes a conventional or novel metaphorical usage is valid for educated Mandarin speakers in
Taiwan in their 20’s.
6. Ahrens et al. (2007) ran a functional Magnetic Resonance Imaging (fMRI) study to look at
the haemodynamic response when contrasting literal sentences with conventional metaphori-
cal sentences and anomalous metaphorical sentences. They found that anomalous metaphori-
cal sentences use extensive additional right and left hemisphere resources, as compared with
conventional metaphorical sentences or literal sentences, indicating that this kind of metaphor
was using additional processing mechanisms to understand and interpret this never-before-
seen source-target domain pairing.
7. One reviewer notes that participants might be baffled by an unconventional formulation of

the metaphor rather than by an unconventional mapping. However, if participants were simply
baffled by an unconventional formulation of a metaphor, they would have produced similar
results for both kinds of novel metaphor, which was not the case.
8. Note that visual iconicity as discussed in Hiraga (2005) is not apt to influence the four
experiments discussed here, as the latter two are listening experiments in which no characters
206 Kathleen Ahrens
are involved, and the former two are reading experiments in which participants are reading at
a natural speed and not analyzing the semantic structure of the characters they are reading.
Furthermore, all four experiments show the same patterns with regard to the degree of nov-
elty of the metaphors, which also suggests that the participants were not spending additional
resources trying to analyze iconicity of the characters when they were reading the sentences in
the first two experiments.
References
Ahrens, Kathleen (2002). When love is not digested: Underlying reasons for source to target
domain pairing in the contemporary theory of metaphor. In Y.-C. Hsiao (Ed.), Proceedings
of the first cognitive linguistics conference (273–302). Taipei: Cheng-Chi University.
Ahrens, Kathleen, Siaw-Fong Chung, & Chu-Ren Huang (2003). Conceptual metaphors:
Ontology-based representation and corpora driven mapping principles. In the Proceed-
ings of the ACL workshop on the lexicon and figurative language (35–41). Association of
Computational Linguistics.
Ahrens, Kathleen, Chung Siaw-Fong, & Chu-Ren Huang (2004). From lexical semantics to con-
ceptual metaphors: Mapping principle verification with WordNet and SUMO. In D. H. Ji,
K. T. Lua Kim, & H. Wang (Eds.), Recent advances in Chinese lexical semantics: Proceedings
of 5th Chinese lexical semantics workshop (CLSW-5) (99–106). Singapore: COLIPS.
Ahrens, Kathleen, Ho-Ling Liu, Chia-Ying Lee, Shu-Ping Gong, Shin-Yi Fang, & Yuan-Yu Hsu
(2007). Functional MRI of conventional and anomalous metaphors in Mandarin Chinese.
Brain and Language, 100, 163–171.
Chinese Knowledge Information Processing Group (CKIP) (1993). Corpus-based frequency
count of words in journal Chinese. Taipei: Academia Sinica.
Chung, Siaw-Fong (2007). A corpus-driven approach to source domain determination. Unpub-
lished Doctoral Dissertation. National Taiwan University.
Chung, Siaw-Fong, Chu-Ren Huang, & Kathleen Ahrens (2003). economy is a transpor-
tation device: Contrasting representation of source domain knowledge in English and
Chinese. In the Proceedings of the special session for the international conference on natural
language processing and knowledge engineering (NLP-KE): Upper ontology and natural lan-
guage processing (UONLP) (790–796). Beijing: IEEE Press.
Chung, Siaw-Fong, Kathleen Ahrens, & Chu-Ren Huang (2004). Using WordNet and SUMO to
determine source domains of conceptual metaphors. In D. H. Ji, K. T. Lua Kim, & H. Wang
(Eds.), Recent advances in Chinese lexical semantics: Proceedings of 5th Chinese lexical se-
mantics workshop (CLSW-5) (91–98). Singapore: COLIPS.
Chung, Siaw-Fong, Kathleen Ahrens, & Chu-Ren Huang (2005). Source domains as concept
domains in metaphorical expressions. International Journal of Computational Linguistics
and Chinese Language Processing, 10, 553 – 570.
Clausner, Timothy & William Croft (1997). Productivity and schematicity in metaphors. Cog-
nitive Science, 21, 247–282.
Gibbs, Raymond (1994). The Poetics of the Mind. Cambridge: Cambridge University Press.
Gibbs, Raymond W (2006). Metaphor interpretation as embodied simulation. Mind & Lan-
guage, 21, 434–458.
Gong, Shu-Ping & Kathleen Ahrens (2007). Processing conceptual metaphors in on-line dis-
course. Metaphor & Symbol, 22, 313–330.
Glucksberg, Samuel, Mary Brown, & Matthew McGlone (1993).Conceptual metaphors are not
automatically accessed during idiom comprehension. Memory & Cognition, 21, 711–719.
Glucksberg, Sam & Matthew McGlone (1999). When love is not a journey: What metaphors
mean. Journal of Pragmatics, 31, 1541–1558.
Glucksberg, Sam, Matthew McGlone, & Deanna Manfredi (1997). Property attribution in met-
aphor comprehension. Journal of Memory and Language, 36, 50–67.
Grady, Joseph (1997). theories are buildings revisited. Cognitive Linguistics, 8, 267–290.
Grady, Joseph (1999). A Typology of motivation for conceptual metaphor: Correlation vs. re-
semblance. In R. W. Gibbs & G. Steen (Eds.), Metaphor in cognitive linguistics (79–100).
Amsterdam: John Benjamins.
Hiraga, Masako (2005). Metaphor and iconicity: A cognitive approach to analyzing texts. New
York: Palgrave-Macmillan.
Huang, Chu-Ren, Siaw Fong Chung, & Kathleen Ahrens (2006). An ontology-based explora-
tion of knowledge systems for metaphor. In K. Rajiv, R. Ramesh & R. Sharman (Eds.),
Ontologies: A handbook of principles, concepts and applications in information systems. Vol-
ume 14. New York: Springer.
Keysar, Boaz, Yeshayahu Shen, Sam Glucksberg, & William Horton (2000). Conventional lan-
guage: How metaphorical is it? Journal of Memory and Language, 43, 576–593.
Lakoff, George (1993). The conceptual metaphor theory. In A. Ortony (Ed.), Metaphor and
thought (2nd ed.) (202–251). Cambridge: Cambridge University Press.
Lakoff, George & Mark Johnson (1980). Metaphors we live by. Chicago and London: University
of Chicago Press.
Lakoff, George & Mark Johnson (1999). Philosophy in the flesh. New York: Basic Books.
McGlone, Matthew (1996). Conceptual metaphors and figurative language interpretation: Food
for thought? Journal of Memory and Language, 35, 544–565.
McGlone, Matthew (2007). What is the explanatory value of a conceptual metaphor? Language
and Communication, 27, 109–126.
Murphy, Gregory (1996). On metaphoric representation. Cognition, 60, 173–186.
Stewart, Mark & Roberto Heredia (2002). Comprehending spoken metaphoric reference:
A real-time analysis. Experimental Psychology, 49, 1–11.
Tsur, Reuven (1999). Lakoff ’s roads not taken. Pragmatics and Cognition, 7, 339–359.
chapter 11
Systematicity in metaphor
and the use of invariant mappings
Alan M. Wallington
I argue that the origin of the apparent systematicity found in families of related
metaphors lies not in ontologically-rich skeletal source domain schemas with
slots that map to target domain correspondents. This suggests a rigidity which
is inappropriate, and misses significant cross-metaphor generalisations. Instead,
I claim that metaphors utilise just a few core source-target correspondences.
Users can extend and elaborate upon these, and by doing so give the impression
of a systematic exploitation of a domain, by incorporating into the utterance
any aspect of the user’s encyclopaedic knowledge that can be linked to the core
correspondences. However, this linkage is subject to the constraint that any
conclusions then drawn about the source domain meaning must ultimately be
grounded in specific types of information or meaning that transfer invariantly
between source and target as adjuncts to the core correspondences. It is in these
invariant mappings that systematicity is to be found.
1. Introduction
In this chapter, I will attempt to give an account of why metaphorical utterances

appear to abide by a principle of systematicity: that is, a tendency to cluster into
families of metaphors sharing common source and target domains. For example,
metaphorical utterances pertaining to love are often viewed or described as con-
cerning a journey, arguments are viewed as war, theories as buildings, and so on.
One explanation for this systematicity is that the metaphor families consist of a
fixed and stable set of mappings or correspondences between the linked source
and target domains. For example, love as a journey assumes correspondences
between: lovers and travellers; the love relationship and a vehicle; problems in
the relationship and obstacles in the path; and so on.1 Expressions such as “We
are at a crossroads in our relationship”, and “We may have to go our separate
210 Alan M. Wallington
ways” are consistent with these correspondences. These related source-target

correspondences would be termed, if we were to adopt a cognitive linguistic ap-
proach, “conceptual metaphors” and many hold them to be “a fixed part of our
conceptual system” (Lakoff, 1993: 208).
I will argue that we need to reject such a static approach, in which the cor-
respondences involved in the individual conceptual metaphors are a fixed part of
our conceptual system. I believe there are problems both with the assumption of
a fixed and stable set of detailed correspondences and with the proper definition
of a domain and these will be discussed in sections two and three. The main claim
however of the chapter is that significant generalisations concerning what is com-
mon to metaphor are being missed.
I argue, instead, that the production and interpretation of metaphorical ut-
terances must be seen as part of dynamic process that takes into account meta-
phor producers’ (encyclopaedic) knowledge, the knowledge they assume their co-
locutors to have, the linguistic, social and cultural context of the discussion and
the situation in which the conversation takes place. All factors which chime with
issues raised by Gibbs (this volume). Along with recent work on Relevance Theory
and metaphor (see Carston, 2002; Vega-Moreno, 2007; Sperber & Wilson, 2008)
and work within the ATT-Meta framework (e.g., Barnden et al., 2004; Wallington
et al., 2006), (see also Hobbs, 1990 for similar proposals), I assume that meta-
phor interpretation involves (often extensive) inferencing, or searching through
knowledge associated with the source words for specific types of information that
can answer queries thrown up by the surrounding discourse.2 Metaphor interpre-
tation is highly context dependent, as also noted by Stern, 2000; Steen, 2007; Steen
et al., this volume; Brouwer, this volume; Gibbs, this volume).
But if this is the case, then what accounts for the apparent systematicity? The
Relevance-Theoretic accounts do not give an account of systematicity. Indeed,
McGlone (2007), working within the somewhat similar approach of Glucksberg
and his associates, explicitly argues against systematicity. To attempt an answer,
I shall first propose a metaphor for the process of metaphor use: that of a jazz
musician improvising around a theme. Importantly, there are principles and con-
straints governing the process of improvisation, and what I shall attempt to do in
the first part of this chapter is to suggest general means, not tied to specific con-
ceptual metaphors, by which very basic, central, or core mappings between source
and target (of a similar nature to Grady’s “Primary Metaphors” (Grady, 1997a, b))
can be extended and elaborated upon in order to allow specific types of informa-
tion to be inferred. This ability to extend and elaborate a core mapping between
a source and target – to improvise around a theme – will give the impression of a
set of detailed mappings between different aspects of a source domain to equally
detailed targets.
Chapter 11. Systematicity in metaphor 211
Another metaphor or analogy might be made with the development over time
of Chomskyan transformational generative grammar.3 In the standard theory of
the 1960s, there were numerous different transformations involving construction-
specific rules detailing how a (structural) description of a source sentence, such as
an active sentence, could be (structurally) changed via a specific transformation
to become a related type of sentence, such as a passive. However, in the following
decades, it was claimed that there were structural commonalities underlying the
different constructions and transformations. Hence, an analysis in terms of many
different constructions and transformations was gradually abandoned in favour
of fewer, more general, principles and constraints, and by transformations that
were reduced in number firstly to two: wh-movement and np-movement, and
then to just ‘move anything anywhere’. The specific constructions and transfor-
mations of the standard model were claimed to be epiphenomena, derived from
more basic principles.
Whatever one might think of the success or otherwise of the Chomskyan en-
terprise, I will argue that a somewhat similar move is possible with the specific
conceptual metaphors and their detailed source-target correspondences. Signifi-
cant generalisations can be made by factoring out from the various conceptual
metaphors, any correspondences that have the following two properties.
Firstly, they are involved in transferring information in an essentially invari-
ant or unchanging manner between source and target, i.e. without undergoing
any metaphorical transformation. For example, suppose that there are two enti-
ties that have been metaphorically cast as A and B in a metaphorical utterance
and some sort of causal or enabling relationship holds between them. Then we
can assume that whatever A and B stand for in reality, say C and D, the enabling
relation will also hold between them.
Secondly, the invariant information the correspondences convey belong to a
privileged set of information types that recur across widely different conceptual
metaphors. Hence, the metaphor of “factoring out” correspondences.
What this “factoring out” process will leave is that which is particular to the
different conceptual metaphors: the core mapping or theme of the preceding jazz
metaphor, around which we improvise, so to satisfy our communicative needs, by
elaborating and extending the core mapping using the meaning conveyed by the
invariant correspondences.
Thus, for a successful metaphor, the speaker/writer must assume that (s)he
and the hearer/reader both share knowledge of the core source-target mapping,
the types of invariant information that can accompany such mappings, the body
of knowledge that allows the relevant inferences to be made, which will, of course,
differ across cultures, and an expectation that the reader will make necessary in-
ferences and utilise the core mapping. Given these assumptions, many different
metaphorical utterances will, as with standard Conceptual Metaphor Theory, ap-

pear to be systematically related. Different utterances may exploit different in-
ference chains in order to convey different types of information. However, since
the invariant mappings conveying this information must accompany, or act as an
adjunct to, core source-target mappings, the entities referred to by the different
utterances must all ultimately relate to the source. Furthermore, since the pro-
ducer of the metaphor must assume that the listener will easily draw the correct
inferences if communication is to succeed – at least for mundane, non-poetic,
discourse – then the different entities, attributes, relations and so on that are used
to extend or elaborate on the core mapping are likely to be quite closely and con-
ventionally related to the source part of the core mapping.
If they are quite closely related, then they might loosely and/or conveniently be
described as belonging to the same domain, but by focussing on the identification
of domains we ignore the significant generalizations that can be found by examin-
ing the rather heterogeneous types of invariant information that cut across tradi-
tional conceptual metaphors. These can be empirically investigated, at least in part,
by re-examining existing conceptual metaphors, especially where detailed (onto-
logical) correspondences have been proposed, and looking for commonalities.
This search for commonalities amongst traditional conceptual metaphors can
be described as a meta-analysis of existing domain-to-domain mappings similar
and complementary to that performed by Grady (1997a, b). Grady also aban-
doned the centrality of complex domains in favour of “primary”, concept-to-con-
cept mappings. These cut across the traditional domain mappings and interact,
blend or unify with each other to motivate some, but crucially not all, of the map-
pings underlying the complex domain-to-domain mappings. The work reported
here differs from Grady’s in that the mappings sought are invariant ones.4
In short, I wish to argue that looking for significant generalisations concern-
ing metaphor interpretation by aligning ontologically complex domains may be
mistaken. Instead, the correct place to look for generalisations is in the meaning
being conveyed; in the type of information that can transfer in an invariant man-
ner between source and target and conversely in the type of information that is
typically ignored. In Section 5, I re-examine from the perspective of a putative
set of types of information that partake in invariant transfers, many of the map-
pings motivating Jäkel’s (1995) mental activity is manipulation conceptual
metaphor. However, I wish to do more in this chapter than merely propose an
alternative to Conceptual Metaphor Theory or a reorientation towards an ex-
amination of metaphor in terms of the dimensions of similarity between source
and target.5 I would argue that there are a number of problems with the tradi-
tional domain-to-domain approach which seem to favour the approach outlined
in this chapter. Thus, I first look at some of the problems with using domains
to circumscribe the space of possible mappings. I then discuss the importance

of what have been described in the ATT-Meta approach to metaphor under-
standing (see Wallington et al., 2006; Barnden et al., 2003) as Map-Transcending
Entities or Aspects, namely source domain entities, relations, attributes and so
on for which it is hard, if not impossible, to give a corresponding target domain
interpretation, yet which play a crucial role in allowing the interpretation of the
metaphorical utterance to be inferred.
One more point about the approach needs to be made. By emphasising the
role of inference in metaphor interpretation it should not be assumed that I dis-
miss the notion of “conventional metaphors”. Firstly, the same sequence of infer-
ences repeated often enough may gradually become an almost automatic process,
triggered upon hearing the metaphor in the appropriate context or upon deciding
how to express a particular concept.. Secondly, a child hearing what an analyst
might classify as a metaphor may attempt a process of inferencing to derive the
meaning, but it is also possible that she makes a direct sound-meaning pairing,
especially if the same metaphor has been repeated frequently. Thus, a final stage
in the process of conventionalisation would be if the sequence of inference steps
were replaced by a single step and the metaphor becomes lexicalised. In short, a
metaphor may become conventionalized either over the history of the individual
user or over the history of the language, with a child, and future generations, lexi-
calising what had previously involved a sequence of conventional and automatic
inferences. However, at any particular time, there may be people who are dynami-
cally searching their knowledge for a sequence of inference steps that can ground
an utterance in one or more of the relevant types of information; there will be
others who, familiar with the utterance, have conventionalised the sequence of
inferences; and possibly yet others who have only ever had a direct lexical inter-
pretation. Furthermore, factors influencing which metaphoric phrases become
conventionalised and which do not, presumably include the type of lexical and
collocational conventions that are best explored using corpus examinations of the
type undertaken by Deignan (2005). Nonetheless, I would argue that even with
highly conventional examples, we can see evidence of the role once played by
inferencing and invariant transfers.
The above points are relevant to some of the issues raised by Gibbs (this
volume). Thus Gibb’s presents a number of dichotomies such as synchronic-dia-
chronic, and system-use. I have just suggested that underlying many conven-
tional metaphors is a sequence of inference steps to a type of information that
will transfer invariantly to the target. This can be viewed as a systems view and
a diachronic analysis; but there may be others for whom the sequence of steps is
new, or who find/convey the same types of information using slight variations on
the utterance and the inference steps. For example, in the next section I compare
the back of one’s mind with the edge of one’s mind, found in a corpus search. In
short, I would argue that the situation on the ground is messier than suggested
by the simple dichotomies.
It might be helpful to employ yet another metaphor, this time from computer
science, and talk of a “space of possible metaphors.” The work described in this
chapter might be understood as circumscribing this space, helping to determine
what is a possible metaphor and what is not. However, the work does not deal
with how actual metaphors in real discourse cluster, or are distributed, within this
space; I present practically no quantitative analyses. Such work is important, but
complementary to the approach described here.
However, this emphasis on the space of possible metaphors must not be con-
fused with approaches, criticised by Gibbs, that treat metaphor in isolation from
context. It is the discourse context that often initiates the search through one’s
knowledge for a sequence of inferences; inferences that may end ultimately in
some of the types of information that will transfer to the target.
Finally, I am in complete agreement with Gibbs about the need for metaphor
research to pay more attention to meaning, although I have used the term infor-
mation in preference to meaning in this chapter. The search for an inventory of
information or meaning types that universally transfer to the target is one of this
chapter’s central concerns.
2. Domains
The overarching argument in this section is that we need a means of deciding

whether or not an entity is in a domain or not if the notion of a productive, sys-
tematic mapping between a source domain and a target domain is to do any use-
ful work.
The notion of a conceptual domain has traditionally been central to the cog-
nitive linguistic endeavour and what distinguishes the “Contemporary Theory of
Metaphor” from theories that look on metaphor as essentially a lexical matter.
The argument is that significant generalizations about metaphor can be stated by
assuming that related metaphors share common source and target domains
There may be constraints on particular correspondences, such as Lakoff ’s
Invariance Principle (Lakoff, 1993: 207), but essentially it is the existence of do-
main-to-domain mappings that constitutes the productive or generative aspect of
metaphor. This is because a domain is not just a single concept, but a number of
related concepts and a source-target mapping is not between a single source con-
cept and a single target concept, but between a number of related concepts and a
number of related target concepts.
An instance of this apparent productivity can be seen with the domain of

journeys being used to describe (the progress of) love affairs as a consequence
of the conceptual metaphor love as a journey. Lakoff (1993) lists the following
highly conventional phrases:
Look how far we’ve come. It's been a long, bumpy road. We can’t turn back now.
We’re at a crossroads. The relationship isn’t going anywhere. We may have to go
our separate ways. We’re spinning our wheels. Our relationship is off the track.
The marriage is on the rocks. We may have to bail out of this relationship.
And a similar exercise could be done with numerous other conceptual metaphors.
Furthermore, novel variants of many conventional metaphors can be generated
by extending the conventional mappings into what Lakoff and Turner (1989: 53)
describe as “unused parts of the source domain”. For example, by virtue of the
conceptual metaphor cognizing is seeing, almost anything having to do with
vision can be used to describe the process of understanding. Grady (1997a) sug-
gests, and I agree, that the following sentence is perfectly understandable, al-
though there is no conventional link involving an “electron microscope.”
(1) You would need an electron microscope to see the point in that.
I shall return to these examples later. However, I first want to discuss a number
of related problems having to do either with the definition of a domain, or with
the assumption that domains circumscribe what can be mapped in a conceptual
metaphor. The issue is important for metaphor because without an operational
definition of what constitutes a domain there must be a suspicion that with suf-
ficient ingenuity many different entities could be claimed as members of the same
putative domain. Such a lack of constraint and detail would considerably weaken
the assumption that it is domain-to-domain mappings that constitute the genera-
tive power of metaphor in accounting for the observed systematicity.
I shall start with the question of definition, and the first problem to note is that
the term domain is highly polysemous (as can be seen from http://en.wikipedia.
org/wiki/Domain). Thus, one common usage is for domain to refer to a universe
of discourse, that is to say the set of entities – which might be reified events or re-
lations – that are specific to any one area of interest. Another common usage, par-
ticularly within Conceptual Metaphor Theory, is to equate domain with Fillmore’s
(1977) notion of a frame and use domain to refer to the background (knowledge)
necessary to understand a concept. For example, terms such as buying and selling
are to be understood against a commercial transaction frame. A concept is said to
be profiled against a domain (see Langacker, 1987; Lakoff, 1987).
In what follows, I shall largely confine the discussion to a view of domains
as being similar to frames rather than as a set of entities specific to an area of
interest.6 And an example of an issue needing consideration is whether the do-

main for one concept can itself be profiled against another domain. In a careful
study, Clausner and Croft (1999) discuss the meaning of arc and circle from this
perspective. The meaning of arc (a continuous portion of a circle) requires the
notion of a circle in order for it to be understood. But likewise, the meaning of
circle needs to be profiled against the domain of geometric shapes. Consequent-
ly, the domain for one concept can itself be profiled against another domain.
But this now raises the question of the relation of arcs to geometric shapes, or
more generally of the relation between an upper domain, such as a geometric
shape, and a lower domain or concept, such as an arc. For the purposes of do-
main-to-domain mappings in metaphor, does a concept belong transitively to
(all) upper domains even though the upper domains do not directly provide the
background necessary to understand the concept? If this is the case, it makes
the definition of a domain particularly inclusive and makes it correspondingly
harder to utilise the concept of a domain to circumscribe the space of possible
domain-to-domain mappings.
Given this problem, it might be best to choose the most restrictive definition
and limit a domain to that against which a concept can be profiled. But does this
apply in practice to proposed domain-to-domain mappings? To give a concrete
example, one of the phrases Lakoff lists as illustrating love as a journey is: “It’s
been a long bumpy road.” According to Lakoff ’s mapping, “difficulties in the rela-
tionship” correspond to “impediments to travel.” It would seem reasonable to in-
fer impediments from bumps, so we might ask whether specific bumps, as opposed
to the more general obstacles, are part of the domain of journeys. Now, bumps
may be profiled against the domain of roads (as well as against numerous other
surface domains), and roads may be profiled against journeys. We therefore have
an analogous case to that of the arc, circle and geometric shape. So somewhat
indirectly, bumps can be linked to a domain of journeys. But does the existence of
this inferential link entail that bumps is in the domain of journeys? And is it this
domain membership that allows bumpy roads to be used to talk about love affairs
via the love as a journey conceptual metaphor? One might claim instead that
bumps can be inferentially linked to the notion of journey, so allowing informa-
tion about the lack of progress to be conveyed. Is the notion of a domain some-
thing of a red herring here?
However, consideration of the bumps examples raises a more fundamental
problem. Much of the information conveyed by a metaphor does not seem to be
specific to any one domain, at least at a degree of abstraction, and Conceptual
Metaphor Theory has always agreed that mappings should be stated at the more
abstract, superordinate level, rather than a basic level (see Lakoff, 1993: 211–212).
For example, it can be inferred that bumps/obstacles prevent or disenable the
smooth running of the journey.7 But an enablement/disenablement relationship

is not restricted to the domain of journeys. For example, I would argue that in the
following sentence, which utilises the conceptual metaphor theories as build-
ings, or perhaps the more general, viability is erectness (see Grady, 1997a, b),
it can be inferred that the purpose of the buttressing is to prevent or disenable the
theory-construction from being refuted or toppling.
(2) Later, Freud did locate a couple of key historical documents which buttressed
his argument. (www.pep-web.org/document.php?id=jaa.016.0399a)
And the notion of a general mapping is not restricted to disenablement. To give

another example, Lakoff (1993) argues that in the following song lyric the emo-
tion of excitement is inferred and transferred to the target.
(3) We’re riding in the fast lane on the freeway of love.
However, emotional states such as excitement are not restricted to the domain
of journeys. Finally, we might note that in the bumpy road example, bumps
might be expected to cause physical discomfort. This physical discomfort might
itself lead to emotional discomfort, or the physical discomfort might stand
metonymically for emotional discomfort; consider for example, “heartache”. In
either case, emotional discomfort in the source corresponds to emotional dis-
comfort in the target.
The arguments so far against the utility of domain in metaphor theory for
anything other than superficial cataloguing have been, firstly, that having do-
mains does not seem to constrain possible inference chains in any natural way
and, secondly, that much of importance in the interpretation of a metaphor seems
to apply to numerous metaphors and is not profiled against any specific domain. I
now wish to consider concepts which receive a standard non-metaphorical inter-
pretation by being profiled against a domain and so should clearly be thought part
of the domain, yet which seem to lead to uninterpretable metaphors when used
in what appears otherwise to be a conventional and productive source-domain
to target-domain mapping. Thus both Grady (1997a, b) and Clausner and Croft
(1997) have discussed the conceptual metaphor theories as buildings and not-
ed the non-occurrence of such metaphors as the following:
(4) That theory lacks French windows.
The concept of windows, French or otherwise, seems to make sense in the context
of a building. If there exists a domain of buildings, windows would be a part. Yet
the sentence is difficult to understand.
In Grady’s account, not only does the theories as buildings conceptual
metaphor include many source domain gaps, such as French windows, but he also
notes that those aspects of the source domain that do appear productive are not
restricted to the specific target domain of theories. He concludes that what appear
to be productive are two mappings: organization is physical structure and
viability is erectness, and these are not just restricted to talking about theo-
ries.8,9 Clausner and Croft, too, argue that theories as buildings is too specific
and they suggest replacing the source domain with the more general, the inter-
nal structure of buildings.
However, even with very general conceptual metaphors such as mind as
physical space (see Barnden’s Metaphors of Mind database http://www.cs.bham.
ac.uk/~jab/ATT-Meta/Databank/), gaps can still be found, as sentences 5, 6 and
7 illustrate.
(5) I know he will probably do a good job but in the back of my mind I can’t
help thinking about the problems he had last year.
(www.business-english.com/backidioms/menu.php)
(6) Yes, I’m young and naive, but I think there’s something just on the edge of my
mind that I can’t quite grasp.
(www.online-literature.com/forums/showthread.php?t=19984&page=3)
(7) ?There’s something just on the right of my mind.
Sentence 5 employs a very conventional metaphor. Indeed, the Shorter Oxford

Dictionary includes, under the entry for back, the phrase at the back of one’s
mind, which it glosses as in the memory but not consciously thought of or imme-
diately recalled.
The phrase the edge of my mind is far less conventional and an anonymous re-
viewer suggested that the phrase would not occur. Nonetheless, an internet search
turned up many examples, one of which is given as 6. However, an internet search
for the right of my mind found no examples.10 Given our knowledge of physical
space, this result might appear surprising, since according to Kilgarriff ’s BNC
frequency list (http://www.kilgarriff.co.uk/bnc-readme.html.), right is the most
common spatial term in the domain of physical space.
The existence of the right of the mind gap would lead Grady to conclude that
mind as physical space is not a primary metaphor and it may be that further re-
finement can eventually come up with a set of primary metaphors for which there
exist no gaps. However an issue that bears on the nature of gaps, and which is
often overlooked, is that, even for entities which can be used felicitously in a meta-
phorical utterance, not every inference that one might make about the entity using
our normal common sense knowledge of the putative source domain is accept-
able. For example, Musolff (2004) gives examples in which the European Union
is described as a marriage. We might assume that such metaphors are licensed by
a union as marriage conceptual metaphor. But this would be a marriage with,

currently, 27 partners, which would be illegitimate in all the member countries.
But the illegitimacy of the European Union is presumably not an inference that
the user would expect to be drawn. Consider also the utterance this policy reflects
the views of the majority. Now, a reflection inverts the image and so the utterance
perhaps ought to be taken counter-factually to mean that the policy is opposed
by the majority. But of course, it does not. It is not as if it violates our common
sense knowledge of mirrors; we know that mirrors invert and even utilise the fact
in such metaphorical utterances as this is the mirror-image of what was wanted. In
short, there appear to be inference gaps as well as the more normal entity gaps.
What this suggests is that it might not be certain entities, such as windows or
right sides, per se that lead to illegitimate metaphors. Instead, it is the drawing of
certain inferences from these entities that connect to the core correspondence.
It is these that are impossible, or perverse, given the context. With acceptable
entities, on the other hand, inferences, which are likely to be made by the pro-
ducer of the metaphor and followed by its interpreter, are possible. To return to
Example (1), in which the felicity of the novel phrase: You would need an electron
microscope to see the point in that was noted, it seems easy to infer, given our
knowledge of what electron microscopes are used for, and assuming a conceptual
metaphor cognizing is seeing, that the degree to which something can be seen
and hence understood is very low. Turning next to the back of the mind and, the
less conventional, edge of the mind examples – sentences 5 and 6 – it again seems
easy to infer, given our knowledge, in this case, of the relation between the centre
of a space and the peripheral areas, that the degree to which someone at the centre
can interact with objects at the periphery is very limited. And, it is this limited
interaction with, or awareness of, the idea that represents the intended message
of these latter sentences.
Conversely, it is unclear what inferences one could draw within the context of
theories as buildings about the relation of French windows to the rest of the
building. A similar point can be made about why an interpretation of utterances
such as it lay on the right of her mind in the context of mind as physical space
is absent. Its absence makes sense if one tries to imagine what inferences could be
drawn from its use, and, in particular, whether the conclusions of such inferences
represent the type of information that typically transfers in metaphor.
To conclude, although it may be possible to give precise definitions of do-
mains, there are reasons to doubt that the notion of domain used in discussions of
cross-domain correspondences can be given a precise enough definition to ground
the notion of source-target transfers. Furthermore, much of what is transferred is
of an abstract nature that would transcend any domain boundary. Consequently,
I take a much more open-ended approach, with potentially anything that can be
i nferentially linked to a central source-target mapping being possible to incor-

porate into a metaphorical utterance.11 But, on the face of it this would fail to
account for the apparent systematicity unless there is some means of controlling
the possible inferences. The control cannot be exerted by domain boundaries. The
proposal I wish to advance in this chapter is that the notion of a correspondence
or mapping between two ontologically complex domains should be replaced by
that of a core mapping and the assumption that only information of a certain type
typically transfers. Such information will be inferred as holding of the situation
described by the metaphorical utterance if the producer and interpreter of the ut-
terance in its context can follow the necessary chain of inferences to the relevant
abstract information types.
3. Correspondence approaches and map-transcending entities
In the previous section I argued that it is a mistake to account for systematicity

by appealing to mappings or correspondences between well-defined domains. In
this section I shall aim towards the same conclusion. However I shall do this not
by querying the nature of domains, but by pursuing one of the principle theoreti-
cal claims of ATT-Meta theory and arguing that many of the entities, relations,
attributes and so on used metaphorically in an utterance to talk about a particular
target need have no actual target correspondents.
This is not the view taken by Lakoff, who states that evidence for the existence
of a system of conventional conceptual metaphors includes: “Generalizations gov-
erning inference patterns, that is, cases where a pattern of inferences from one
conceptual domain is used in another domain.” He further adds that mappings:
project source domain inference patterns onto target domain inference patterns.
(Lakoff, 1993: 245, emphasis added).
In other words, the pattern made by tracing out each step in the inference pro-
cess from premises to conclusion in the source are claimed to be projected onto
the target.
Given this assumption, I shall re-examine in more detail the conventional
expressions concerning the progress of a love affair that I cited in the previous
section, and which Lakoff claims show love conceptualized as a journey.
Lakoff (1993: 207) argues that such expressions are evidence for the love-as-
journey mapping, being “a set of ontological correspondences that characterize
epistemic correspondences by mapping knowledge about journeys onto knowl-
edge about love.” Specifically, the following set of correspondences is assumed:
THE LOVE-AS-JOURNEY MAPPING

– The lovers correspond to travellers.
– The love relationship corresponds to the vehicle.
– The lovers’ common goals correspond to their common destinations
on the journey.
– Difficulties in the relationship correspond to impediments to travel.
Lakoff does not describe the reasoning process in any detail, so I shall take the
expression we’re spinning our wheels and consider how one might infer from this
that the love relationship is not progressing as it might. Plausibly, the following
chain of inferences is entertained. If wheels are referred to, then, it is likely that
a vehicle is involved. The spinning wheels are causing the vehicle not to move as
it should. If the vehicle is not moving as it should, then it is not moving towards
its destination. A further conclusion is that this lack of progress, either literally or
metaphorically, causes frustration.
What can be made of this pattern of inferences that allows us to conclude
that the vehicle is not moving towards its destination? Both the vehicle and the
destination have correspondents in the target domain, namely the love relation-
ship itself and the lovers’ common goals. With these correspondences, it might be
assumed that the source domain conclusion can be transferred to become a target
domain conclusion. However, this is the transfer of the conclusion. The conclu-
sion would never have been reached without the premise that something – the
spinning wheels – was causing the vehicle not to proceed. So what is the tar-
get domain correspondent of the spinning wheels? None of the four ontological
correspondences would seem appropriate. Moreover, this conclusion involves a
negation: not moving towards a target. The negation transfers to the target, but
there is no formal discussion of why this should be so and whether negation in the
source always corresponds to negation in the target.
A look at the other examples listed finds similar cases where the lack of a
target domain correspondent would prevent crucial aspects of the source domain
inference pattern from mapping to an inference pattern about the target. What,
for example, is the target domain correspondent of the rocks in the utterance our
marriage is on the rocks? A similar point can be made about the location off the
tracks in our relationship is off the track. But, as before, these statements can be
easily interpreted. Both of them would seem to permit the same conclusion that
was reached about the spinning wheels, namely that the relationship/vehicle is
not progressing towards the lovers’ common goal or destination, and that conse-
quently the relationship is not functioning as it ought. But without target domain
correspondents for these items, and if it is inference patterns that map between
source and target, the conclusion cannot transfer.
There is also the problem that it does not seem quite right to assume that
the spinning wheels, rocks, or lack of tracks are impediments in the sense of the
third of Lakoff ’s correspondences. Whilst one might assume that the bumps in
the statement it’s been a long bumpy road are impediments and point to specific
difficulties that have occurred over the course of the relationship, there does not
seem to be an assumption with the other cases we have discussed that there are
specific causes to the relationship not progressing and hence functioning. What-
ever wheels might refer to, it is a vague and very unspecific target domain entity.
Furthermore, there is also the point made in the previous section that a relation-
ship described as a bumpy road suggests that there was some emotional discom-
fort, an inference that would not appear to follow from the assumption that there
was an impediment.
In short, a careful study of these examples supports the view of ATT-Meta
theory that there can be entities referred to in a metaphorical utterance that need
have no correspondent, or mapping, in any target or literal interpretation of the
utterance. Instead, their role is in allowing certain conclusions concerning the
source to be drawn and it is these conclusions that transfer, not necessarily all the
items that were involved in the drawing of the conclusions.12
Another sentence often used to illustrate love as a journey is: We’re at a
crossroads. Now, stating that we are at some location on a road might be taken to
imply that we are going somewhere along the road and hence have a destination.
This would correspond to a target domain common goal. However, there is no spe-
cific correspondent given for crossroads. The source domain inference, that if one
is at a crossroads, then there is a choice of possible destinations, cannot transfer to
the target, becoming the conclusion that there are alternative relationship goals,
if, as is stated, inference patterns are mapped from one domain to another.
Now a possible solution to the problem might be to assume a lacuna in the
correspondences listed above and add a fifth correspondent, something along the
lines of alternative destinations correspond to alternative life-goals. However, the
mapping already lists as correspondence 3, life goals corresponding to destina-
tions. It might be thought more parsimonious to factor out from this putative cor-
respondence the commonality that could be dealt with by correspondence 3. This
leaves alternatives mapping to alternatives. But, this is just the type of invariant
transfer I suggested in the introduction underlies much metaphor. If further evi-
dence shows that alternatives inferred in the source always map across to become
alternatives in the target, regardless of which conceptual metaphor is in play, then
it would be redundant and missing a generalization to include a specific corre-
spondence as part of the love as a journey conceptual metaphor.
A similar factoring out would seem possible with the fourth correspondence:
difficulties in the relationship correspond to impediments to travel, if it is first
r ephrased as: factors that prevent the relationship from progressing are factors that
prevent the vehicle from travelling. Since correspondence 2 (and 3) already lists a
relationship (progressing towards a goal) corresponding to a vehicle (travelling
towards a destination), what is particular to the rewritten correspondence 4 is
factors that prevent corresponding to factors that prevent.
I will follow ATT-Meta theory and assume that ability, inability, enablement,
disenablement, letting and prevention (more broadly, causation) tend to map in-
variantly. I will assume that an inference can be made that the journey is being
prevented from proceeding as it should. Thus, a reader will understand a bumpy
ride or spinning wheels in reality as concerning a love affair rather than a journey,
because of the core mappings of love as a journey, and she will further under-
stand in reality that the affair isn’t progressing, because of the disenablement ad-
junct to love as a journey that allows disenablement in the source to be viewed
as disenablement in the target.
In conclusion, and in agreement with the view of the ATT-Meta approach,
there often appears to be a lack of parallelism or correspondence between source
and target. Furthermore, where correspondences are stated, as in Lakoff ’s love
as a journey correspondences, they are often little more than rewordings of
the source terminology in order to transfer an invariant property, as with dif-
ficulties in the relationship corresponding to impediments to travel. All this cre-
ates problems for the assumption of domain-to-domain mappings. However, if
instead what transfers are conclusions of a certain type, and if the purpose of
the source entities in the utterance for which there are no correspondents is to
allow additional information about the source to be inferred, but information of
a type that transfers invariantly to the target, then the lack of parallelism would
be expected.13
4. Invariant transfers
The two preceding sections have suggested that it is a mistake to look for the
systematicity of metaphor in correspondences between conceptual domains.
Thus, in Section 2, I argued that the notion of domains is problematic, at least
with respect to circumscribing possible metaphor mappings, being both diffi-
cult to delimit and needing to contain elements such as excitement, degree and
so on, which presumably could be inferred from numerous unrelated meta-
phorical utterances. In Section 3, I criticized the claim that inference patterns
must transfer from source to target by casting doubt on the notion that entities,
relations, attributes and so on referred to in a metaphorical utterance must have
target correspondents.
So what universal types of invariant information can be inferred? It does not

seem to be the case that information of any type can be inferred, even if there
is strong contextual support for the drawing of such an inference. For example,
Hobbs (1990: 59, 60) noted the word elephant can metaphorically convey different
information depending upon context.
(8) Mary is graceful, but John is an elephant.
Patricia is small, but James is an elephant.
Susan forgets everything, but Paul is an elephant.
Jenifer is subtle, but Roger is an elephant.
However, the colour red is surely as salient an aspect of tomato or pillar-box as

clumsiness is of an elephant, yet the following metaphorical utterance does not
seem to work.
(9) Mary’s pencil is blue, but John’s pencil is a tomato/pillar-box.
Similarly, the well known Juliet is the sun leads to no inference that Juliet was a
bright yellow. It might be concluded that colour is not the type of information that
typically accompanies a core source-target mapping.14
I have already made some suggestions as to the types of information that may
transfer invariantly: information concerning enablement and disenablement, the
degree to which something holds, and emotional states. However, I have not yet
shown that the examples in which such information is inferred are representative
of a more general phenomenon. It might be the case that the types of inference
that transfer invariantly differ according to the particular conceptual metaphor
chosen. Indeed, the class of generic is specific metaphors (see Lakoff & Turner,
1989, or Grady, 1999), such as: long term purposeful activities are jour-
neys, where a journey is a specific instance of a long term purposeful activity,
constitute such a case.15
In this section, I shall try to motivate a set of invariant mappings which do
have a more universal application – at least by default. In the ATT-Meta approach
this set of invariant mappings have been called View-Neutral Mapping Adjuncts
(VNMAs); adjuncts because they accompany, or are an adjunct to, traditional
source-domain to target-domain mappings, and view-neutral because they can
apply regardless of which metaphorical view one is taking.16 In what follows, I
shall list a number of different types of information which seem to get transferred
between source and target by these VNMAs. In each case, I also attempt briefly to
motivate the particular VNMA by giving a few examples of the invariant trans-
fer of the type of information being discussed, using a range of what might be
thought of as different source and target domains. The list is not intended to be
complete (see Wallington and Barnden, 2006 for a fuller list with more discussion.
See also Wallington et al., 2006 and Barnden et al., 2003). In addition, since any
metaphorical utterance typically conveys a mixture of information, the examples
given do not pretend to consist of all, or even the most prominent type, of the
information being conveyed.
causation/ability/function
Causation, ability, function, prevention, helping and (dis) enablement relation-

ships between events or other entities in the source map to causation, prevention,
etc. relationships between their mappees (if they have any).
A couple of examples that have utilised the invariant transfer of enablement
(or disenablement) have already been discussed. In sentence 2, repeated below, I
argued that the buttressing documents enabled Freud’s arguments to stay erect.
This enabling relation between the documents and the arguments transfers to the
target as an adjunct to the core mapping between theories and buildings.
(2) Later, Freud did locate a couple of key historical documents which
buttressed his argument.
Similarly, I argued that the relevance of bumps in Lakoff ’s love as a journey

conceptual metaphor was that they prevented the love affair-journey from pro-
gressing as it should.
qualitative degree
If the holding of a graded property or relationship in the source maps to the hold-
ing of a graded property or relationship in the target, or vice versa, then the quali-
tative absolute and relative degrees map over identically.
The following two sentences form a minimal pair:
(10) In the far reaches of her mind, Anne knew …..17
(11) Anne had an idea in mind.
Both sentences concern some thought or idea of Anne’s. However, sentence 10,
but not 11, seems to suggest that Anne was not engaging with the idea very ac-
tively. See also sentences 5 and 6 discussed earlier. The respective interpretations
can be explained if the phrase the far reaches in sentence 10 allows an inference
that is unavailable in sentence 11. Both sentences make use of mappings between
ideas and physical objects, mind and physical space, and mental activity
and the physical interaction with solid objects.18 However, with sentence
10, the fact that the idea-object is far away suggests that Anne would only be able
to interact with the idea to a very limited degree. Then, because mental activity
is viewed as physical interaction, the low degree of physical interaction would
be viewed as a low degree of mental activity if qualitative degree transfers
invariantly.19
Sentence 12 conveys a somewhat similar idea:
(12) We can dimly make out four distinct ideas mingled in the current notion.
(Some Principles of Maritime Strategy By Julian Stafford Corbett. Page 23.
Published by BiblioBazaar, LLC, 2008. ISBN 0554251981.)
I would argue that the analysis of this sentence also involves a mapping between
ideas and objects, but in this case involves cognizing as seeing. It can be in-
ferred that the viewing, or making out, of the idea-object can only be done to a
limited degree. Then, since viewing maps to cognizing, the low degree will map
over as an adjunct.
Consider sentence 13:
(13) She was not fully cognizant of the idea.
I shall assume a mapping between a state (of being cognizant or aware) and a
full container: the same mapping that underlies idioms such as “an empty ves-
sel” or “empty headed”. In this case, the container is full to only a partial degree
and this partial degree maps over, allowing the inference that the subject’s state of
cognizance or awareness is also only partial.
I shall consider one more example.
(14) Now you’ve almost reached your destination: an informed and thoughtful
opinion.
This utilises the mapping between purposeful action and journeys and be-
tween states, such as having an informed and thoughtful opinion, and locations.
However, the main import of the utterance is that the degree to which the subject
has an informed and thoughtful opinion is high but not complete. This follows if
we assume that the inferred near completeness of the journey and thus arriving at
the destination location corresponds to being near to having the opinion.
To conclude, if the degree to which something holds can be inferred in the
source, then typically such information will also be relevant to the target. Further-
more, the degree is qualitatively the same.
(un)certainty
The level of certainty with which situations hold in the source maps at least ap-
proximately to the level of certainty with which their mappee situations hold.
Consider the following utterances:
(15) It’s a long shot but ….
(16) It was a complete shot in the dark.
These examples assume the correspondence: mental activity is physical in-

teraction with solid objects together with an accompanying reification of
ideas problems etc as objects. The causation vnma is also required to trans-
fer the ability to manipulate the objects. But it can also be inferred that because
the idea-object was distant, or there was little light to see the idea-object, then
there will be a high degree of uncertainty that it will be possible to interact with,
or ‘hit’, the idea-object.20
To consider uncertainty at play in another domain, the following sentence,
about the then future digital movement in photography, was found using an in-
ternet search:
(17) Where are you both regarding the digital movement coming up behind us?
The sentence is somewhat unusual because future events are often conceptualised
in the majority of languages as being in positions in front of the temporal experi-
encer. Drawing upon an analysis of the future in the Aymara language by Núñez
and Sweetser (2006), I would argue that when a future event is conceptualised as
being behind the temporal experiencer, it is because she has considerable uncer-
tainty about the nature of the future event and how it will impact upon her. This
can be inferred if we assume core mappings between cognizing and seeing and
between events and objects, plus the inference that an event metaphorically
cast as an object which is not co-located with the temporal experiencer, is not
now. If some coming event-object is behind, then it cannot be seen and hence one
only has uncertain awareness or cognizance of it.
alternatives21
If a number of choices/alternatives are presented concerning a mappee or a num-

ber of alternative instances of a mappee are available, then the existence of the
alternatives will also transfer to the target.
I have already discussed the sentence:

(18) We’re at a crossroads.
I assume this rest upon purposeful activity as a journey. It can be inferred

that if one is at a crossroads, then alternative continuations of the journey/pur-
poseful activity are possible.
The interpretation of the following sentence utilises different core mappings:
(19) There’s another side to this issue.
This involves correspondences between issues and physical objects and be-
tween mental activity and physical interaction.22 The fact that the object
can be manipulated/viewed in more than one way, as can be inferred by the claim
that there are different sides, will be transferred to the target accompanying the
mental activity is physical interaction mapping.
event-shape
Aspectual features of events/situations/processes, such as whether they have a

start or end, or are intermittent, map identically to mappee events/ situations/
processes.
All episodes, including source and target, need an event-shape. By default the
event-shape of the target will be inherited from the source.23 Compare for example:
(20) He searched for an answer to the question.
(21) He wanted an answer to the question.
The first example denotes an activity, the second a state, and such information
concerning the event will be mapped by the event-shape vnma. No specific
target correspondent is required for the searching or wanting.
mental/emotional states
If some agents in the source domain map to some agents in the target domain,
then mental and emotional states of the agents map identically, except that their
objects or propositional contents (if any) are modified suitably by any mapping
relationships that apply, and provided that this modification can be done. This
vnma also transfers any emotional state induced in the users of the metaphor
by the source domain description to the target.24
(22) Your room is a cess-pit.
A cess-pit is a pit for the disposal of sewage. Therefore, this utterance is using a
type of location, namely a cess-pit, to describe a state, namely the untidy, dirty or
disgusting state of the addressee’s room. However, there is no need to assume as
part of states are locations any specific mapping between dirty and disgusting
locations and dirty disgusting states. The disgust evoked by the use of the term
cess-pit is conveyed by the mental/emotional states vnma.
One could also cite here the much discussed example of the grim, unpleasant,
look of personifications of death, such as the medieval Grim Reaper. Any fear or
anxiety induced would be conveyed by this mental/emotional states vnma.
value-judgment
Levels of goodness, importance or other types of value assigned by the under-

stander to states of affairs in the source domain map identically to levels of good-
ness, etc. of their mappee states of affairs, if any.
This vnma can be very similar to and hard to tell apart from the former.
Consider the following utterance:
(23) That’s a gem of an idea.
I would argue that the only core mapping needed to understand this example
is ideas as objects. In particular, there need be no correspondence associated
with the concept of gemness. Instead, it can be inferred that gems are valuable and
this value-judgement about those objects that are gems transfers to become a
value-judgement about ideas that are metaphorically precious.
(24) A late medieval flowering ….
This metaphor presumably depends upon the correspondence, development as

growth (of plants). However, rather than having a correspondence detailing
all aspects of plant growth, the positive value-judgment concerning entities that
flower would be conveyed.
Other vnmas
I will briefly mention a number of other vnmas that I will refer to in the final sec-
tion. For more details see Wallington and Barnden (2006). The negation vnma
deals with instances where a property or relationship in the source corresponds
to some property or relationship in the target. By virtue of the negation vnma,
the non-possession of the source property/relationship maps to non-possession
of the target property/relationship. For example, I have discussed enablement and
disenablement. Strictly, disenablement follows from the negation vnma applied

to the causation/ability vnma.
Another vnma I shall refer to is rate. Something metaphorically cast that
can be inferred to take a long time will be interpreted as taking a long time in the
context of the target.25 A final vnma I shall mention is time-order, which ac-
counts for the fact that if there is a sequence of events in the source which map to
events in the target, then the time-order of the events will be preserved.
5. Invariant mappings and systematicity
In the final part of this chapter I shall briefly look at how the postulation of VNMAs
motivates a number of systematically related metaphorical utterances. These utter-
ances have been used by Jäkel (1993) to illustrate a series of sub-metaphors and cor-
respondences, what Jäkel calls: model-integrated components or MICs, that group
together to form the conceptual metaphor: mental activity is manipulation.26
I chose Jäkel’s work because it represents a particularly good and detailed use of
conceptual metaphor theory to analyse a large set of what appear to be systemati-
cally related metaphors. Like Jäkel, I too shall assume as a core correspondence the
link between physical interaction, i.e. making oneself aware of the physical
properties of the object, and mental activity. I also assume, as does Jäkel, that
ideas, thoughts, problems or issues are reified as physical objects by virtue of the
conceptual metaphor ideas are objects. I also assume mind as physical space
together with the important addition that the physical space is often conceived
as relative to a homunculus or ego representing the conscious possessor of the
mind and occupying a central position in the physical space. Since these mappings
will be used in most of the examples I analyse, I will not always explicitly men-
tion them. Also, unless the event-shape plays an important role in the analysis,
I shall usually omit to mention it. There are two final points to be made before I
begin the analysis. Firstly, it is often the case that source entities in the metaphori-
cal utterance can be inferentially linked to a core mapping using more than one
vnma, and so can be used to convey more than one type of information. Indeed
I would argue this economy of resources is an important aspect of metaphor. Sec-
ondly, Jäkel’s illustrative examples are presented without context, or rather the only
context provided is Jäkel’s claim that the examples illustrate particular MICs. One
of ATT-Meta’s central claims (matching that of Stern, 2000 and Steen et al. (this
volume) and Gibbs (this volume)) is that the context of a metaphorical utterance
helps determine the nature of the queries that the utterance is expected to answer.
Thus other readings may be possible given a more explicit context.27
Jäkel’s first model-integrated component (or MIC) is ideas are solid ob-
jects. This is a conceptual metaphor I also recognise, and have used it in a num-
ber of examples in this chapter. Jäkel also argues for its interaction with the con-
ceptual metaphors: complexity is dimensionality, difficulty is hardness,
and importance is weight. These motivate in turn the following utterances:
(25) There’s another side to the question.
(26) That’s a very hard question to answer.
(27) Let’s turn to less weighty matters.
Part of the import of utterance 25 may be explained by assuming that if there is
more than one side, then the question-object may be examined from a number
of alternative perspectives, which implies it can be manipulated (or viewed) in a
number of alternative ways. If there are alternative manipulations of the question-
object, then there are alternative mental activities concerning the question-object
(or alternative understandings). This inference will transfer to the target via the
alternatives vnma acting as an adjunct to mental activity is physical in-
teraction (or possibly cognizing is seeing).
I shall discuss utterances similar to 26 when I consider sentence 48. The anal-
ysis of 27 may make use of a conceptual metaphor importance is weight but
alternatively it may rely on the common inference that heavy (and large) objects
are often important. Importance is then transferred to the target via the value-
judgement vnma.28
Jäkel’s second MIC is understanding an idea is establishing physical
closeness. The earlier discussion of sentences 10 and 11, and the degree to which
an idea-object can be manipulated, or interacted with physically, is relevant here.
However, Jäkel subdivides this MIC into four subMICs representing different
stages in reducing the distance to the idea-object.
The first subMIC, MIC2a is: understanding starts with attempts at
finding and hunting idea-objects. This is claimed to motivate utterances
such as the following:
(28) Sally searched for an idea all day.
A plausible sequence of source domain inferences would be that Sally does not
possess the idea-object. Thus, she has no ability to manipulate it. This information
is passed to the target by the negation and the causation-ability-function
vnmas. It might also be inferred that if she were searching for an idea-object then
she desires to possess it. This desire would be passed to the target by the mental
and emotional state vnma. One might also note the specific event structure,
which was that of an activity involving Sally and the idea-object. This activity
event structure will transfer to the target via the event-shape vnma so shaping
the aspectual properties of the target. Note that the verb search contributes to
the meaning of the utterance via its contribution to the event structure and via
the mental or emotional state it suggests. However, there is no need for a specific
target domain correspondent specifically for physically searching for something.
Finally a reasonable, although not definite, inference, given her desire and the na-
ture of the target’s event structure, is that Sally undertook some action to remedy
the situation: namely to have the ability to mentally interact with an idea.
The inference of a wish or desire is absent from the following utterances, al-
though further context may suggest it, but degree and ability, within the con-
text of the core metaphor mental activity is manipulation, can be inferred.
(29) The solution is now close at hand.
(30) Nothing was further from my mind.
(31) I think the problem is within my grasp.
These inferences are likely to be made with the following examples, but further
conclusions can also be drawn.
(32) He took aim at the problem.
(33) Have a shot at solving the problem.
(34) It’s a long shot but ….
(35) It was a complete shot in the dark.
These examples all suggest that there is some uncertainty as to whether it will
be possible to shoot, hit or physically interact with the idea/answer-object. This
uncertainty will be transferred to the target by the uncertainty vnma. Thus, the
existence of the uncertainty vnma can be said to motivate the form of these
metaphorical utterances, at least partially.
Jäkel’s second subMIC, MIC 2b is: understanding is seizing idea-object
firmly. What this adds to the conceptual metaphor mental activity is physi-
cal interaction is the notion that the idea-object is manipulated firmly. Clearly
if one is holding something, then one can manipulate it, to at least some extent.
However, the degree of certainty that one will be able to continue to possess the
ability to manipulate the object may be high or low. By using the word firmly, it
suggests that the ability is high. The continuance is transferred by the event-
shape vnma; the uncertainty, ability and degree are transferred as in the
other examples.
(36) He lost his grip and things got away from him.
(37) I’m all thumbs at algebra.

(38) It’s a slippery concept.
Consequently grip, thumbs and slippery need have no specific target correspon-
dents. Note that Example (38) may also utilise the value-judgement vnma to
transfer negative connotations about slippery objects.
MIC 2c is understanding is picking up idea-objects and this motivates
utterances such as the following:
(39) He’s slow/quick on the uptake.
(40) Where did you pick up such ideas?
Jäkel argues that such metaphors utilise the conceptual metaphor conscious
is up. However, it might be possible instead to appeal to a default assumption
that manipulation of an object takes place when it is in an easily accessible
position, and not on the ground. Thus, there would be no need for a target cor-
respondent of up.
MIC 2d is understanding is taking idea-objects into the mind-
container. This assumes the conceptual metaphor mind is a container, which
I assume is a special case of mind is physical space: one of our central mappings
between source and target.29 I have assumed that the manipulator of idea-objects,
ego, is assumed by default to occupy a central position in the mind-space. Thus
in the following sentence a simple inference would be that ego, being absent from
the mind-space now no longer has the ability to physically interact with idea-
objects in the mind-space. This lack would be transferred by the negation and
causation-ability-function vnmas.
(41) He’s gone out of his mind.
The following utterances allow the inference that there is a high degree of cer-
tainty that idea-objects will be able to enter the mind-space. Consequently the
broadness or openness of the mind needs no target equivalent.
(42) Keep an open mind.
(43) Her parents were broadminded.
Jäkel’s next new model-integrated component, MIC 3 is: within the mental
workshop, idea-objects are stowed away according to urgency. He as-
sumes that the following utterance utilises Johnson’s (1987) conceptual metaphor,
importance is centrality.
(44) She put the idea to the back of her mind.
This metaphor can be analysed in a similar manner to the far reaches example, i.e.
sentence 10, that I examined in the previous section, and which involved a low
degree of ability. Over and above this analysis, the time course and event-shape
of the idea-object being in the centre and then being put at the back will be trans-
ferred by the time course and event-shape vnmas.
MIC 4 is thinking is working on problem-objects with the mind-
tool. As with MIC 2, Jäkel also divides this into four subMICs. I shall not exam-
ine them in detail here, nor list them all. However, I shall comment on a few of
the metaphorical utterances that motivate this MIC. Jäkel assumes the conceptual
metaphor: intelligence is sharpness of the mind-tool and supposes that
this motivates utterances such as the following:
(45) He has a razor wit.30
(46) The intrigues of the court had sharpened her wits.
(47) He used extensive reading to hone his intellect.
However, the familiar metaphor analysing is cutting can just as well be used. I
can then assume a source domain inference that the degree to which one can cut
something, and hence analyse it, is high if the tool is sharp. This will be conveyed
by the degree vnma.
A somewhat similar analysis can be given to utterances such as:
(48) This is a hard nut to crack
But in this case the idea-object or nut is being profiled, rather than the cutting or
cracking implement.
The following utterances about the mind-tool can be analysed as allowing
the inference that the mind-tool, or alternatively a person metaphorically cast as
a machine, is not functioning properly:
(49) He flies off the handle quite easily.
(50) Her mind was not on it
This lack of function will be conveyed by the causation-ability-function

vnma. There is no need for any specific correspondents for handles or flying.
Jäkel uses the following utterance to motivate his MIC 4d, solving a prob-
lem is taking the solution-contents out of the problem-container.
(51) We dredged up a load of sordid facts about her.
The notion of facts-object being brought to a position where they can be easily
operated upon or interacted with is similar to many other examples I have an-
alysed. However, note the negative value-judgment conveyed by the word being
used to describe the act of bringing the facts to this position, namely dredge. I
shall assume this negative value-judgment is conveyed to the target by the value-
judgment vnma.
MIC 5 is forming ideas is shaping raw materials. This differs from other
examples in that the idea-object is produced or manufactured, rather than be-
ing a pre-existing object. Consequently most of the examples involve the event
structure vnma.
(52) He produces new ideas at an astonishing rate.
This example may also involve the rate vnma, although arguably the astonishing
rate applies to the outcome the source-target mapping and is not part of the infer-
ence process involved in interpreting what is meant by producing new ideas.
MIC 6 is judging diverse arguments is weighing up idea-objects. In
the discussion of MIC 1, I suggested that the inference can be made that heavy
things are important and that the importance would be conveyed via the value-
judgement vnma. The examples Jäkel discuss here also involve the alterna-
tives and degree vnmas. The only other thing that is required is a conceptual
metaphor involving a mapping between considering something and weigh-
ing it, which is just a special case of analysing is measuring.
(53) He weighed the alternatives in his mind.
(54) The scale just tipped in favour of my voting yes.
Under this analysis, scales or weights need no target domain correspondents.

Jäkel’s seventh MIC is the valuables of knowledge are kept in the memory
store for further use. He assumes that this involves a conceptual metaphor,
ideas and knowledge are valuables, which motivates utterances such as:
(55) Treasure something up in one’s memory.
But all that is required is the conceptual metaphor ideas as objects and mind
as physical space, which have been assumed throughout this section, plus
the value-judgment vnma to convey the value that might be inferred from
being told that an object is to be treasured. See the discussion of Example (23)
in Section 4.
Jäkel divides this MIC into two subMICs in which memorising is putting ob-
jects into the memory store and remembering is bringing the objects back. These
motivate examples such as:
(56) You must commit them to memory.
(57) Can’t you bring this to mind?
Assuming ideas as objects, mind as physical space and mental activity

is physical interaction, together with the assumption of ego, or a homuncu-
lus representing the conscious possessor of the mind, in the centre of the mind-
space, then all that is required to analyse these examples are the time-order and
event-shape vnmas.
Jäkel’s final MIC is MIC 8: abandoning ideas is dumping discarded ob-
jects and empty issue-containers at the end of the manufacturing pro-
cess. This covers such examples as the following:
(58) You’ll have to let go of your belief in Santa Claus.
This example can be analysed by assuming that after the event of letting go, EGO
will only have a low degree of ability to manipulate the belief-object.
6. Conclusion
In this chapter I have argued that contemporary metaphor theory has paid too
much attention to describing the nature of putative domains used in source-target
mappings and to the objects, such as buttresses, windows, crossroads and so on,
that might or might not be found in these domains, but not enough attention to
describing and cataloguing the types of information or meaning that seem to be
used in metaphor. I suggested an alternative to accounts that explain the syste-
maticity found in metaphor in terms of ontologically complex, circumscribed,
source domains that map to specific target domains. Instead, I proposed a means
whereby core metaphor mappings can be extended and elaborated. I argued that
what is required is just the interaction of a few core source-concept to target-
concept correspondences, together with the assumption that anything that can
be inferentially linked to the source concepts, given the context, is a possible can-
didate for use in a metaphorical utterance by the speaker. However, this candi-
dature is subject to the constraint that the speaker must assume that the hearer
will draw the correct inferences and, most importantly, that the information so
inferred belongs to one of the types of privileged information that transfer invari-
antly between source and target as an adjunct of the core mappings. Following
the ATT-Meta approach, I proposed a set of vnmas that characterise the types
of meaning or information that appear to map in an invariant manner. It must
be stressed that determining the set of vnmas is an empirical exercise, which
requires careful examination of a wide range of metaphorical utterances in order
to factor out common invariant properties.
The claim made in this chapter, however, consists of more than that there is
an alternative means of explaining the systematicity found in metaphor. I argued
that there is no operational means of defining source domains that could usefully
be used to delimit the set of possible metaphors and I noted that much of what is
transferred would not normally be characterised as belonging to just one particu-
lar source domain. I also noted that many of the entities, relations, attributes, and
so on in particular metaphorical utterances do not have any correspondents in the
target. Thus I would claim that there are empirical problems with domain-to-do-
main mapping approaches. In general, it is not entities that are important, or even
how different entities relate or are structured, but rather the types of conclusion
that can be drawn from their use.
It could be argued that the proposal is more of a shift in emphasis in doing met-
aphor research. Others have noted the importance of invariance in metaphor. In-
deed a traditional claim is that metaphor involves searching for a similarity between
the source and the target, although ATT-Meta’s assumption that correspondents
are not required for many source entities argues against too superficial a similarity.
This lack of source-target parallelism also argues against the claim that metaphor is
a type of analogy or is strongly related to analogy (see Gentner et al., 1988). If one
turns to specific approaches to metaphor and analogy, further differences from the
approach presented here can be found. Thus, approaches based on Structure Map-
ping Theory (Gentner, 1983) argue that systematic relations and the entities involved
in them will transfer between source and target, whilst one-place relations such as
attributes will not. Such an approach may well account for the causation-ability-
function vnma. However, I also assume that value-judgment attributes that
pertain to something will also transfer by default, as will degrees.
The importance of invariant information has also been discussed within the
context of Cognitive Linguistics. Thus, Lakoff has formulated an Invariance Prin-
ciple that acts as a constraint on source-target mappings (Lakoff, 1993: 207):
Metaphorical mappings preserve the cognitive topology (that is the image sche-
ma structure) of the source domain, in a way consistent with the inherent struc-
ture of the target domain.
As with Gentner’s Structure Mapping Theory, this seems to emphasise the role
of structure over the meaning, or the type of information that is transferred in-
variantly. But even more importantly, note that the structure is image schematic
structure (i.e. imagistic or what can be experienced by the senses). If the source
and target domains are structured by image schemata, then one would expect to
find similarities between the source and target. And, in this sense, obstacles to
a journey might reflect a basic journey image schema. However, whilst this ap-
proach may explain the use of ‘bumps’ in metaphors about journeys, emotional
states such as excitement or misery are not imagistic; nor are value-judgments;
and nor, I would contend, is much that is conveyed by metaphor.
Perhaps the closest approach is that of Carbonell (1982) who presents an in-
variance hierarchy of relations that are preserved under metaphoric/analogical
transfer. However, I make no claim that the vnmas are ordered in terms of a hi-
erarchy of application. I also assume that VNMAs are adjuncts to a core mapping.
It is worth emphasising that vnmas form a rather a heterogeneous set. What
they all do is “afford” the participants in a discourse the means of describing what
might at the time of the discourse, and for whatever reason, be difficult to state us-
ing more domain-specific language. There is no obvious means of reducing them
to some more general principle. Indeed, we could speculate that the different vn-
mas have little to do with metaphor as such. One might suppose, for example,
that value-judgments, such as X is good to eat or Y is dangerous represent a very
basic categorisation principle. Similarly, recognising the emotional and men-
tal state of an individual; having a ‘theory of mind’; is a fundamental cognitive
ability, and assuming causal links between entities is if anything over-exploited
and seen as relating events where no such relation exists. Thus looking for further
vnmas, analysing metaphors in terms of vnmas and looking for motivation
for the vnmas that have been proposed represents a considerable research en-
deavour that has barely been started.
Acknowledgements
This work is located within the ATT-Meta approach to metaphor interpretation

and ATT-Meta’s influence together with that of my colleagues with whom I have
worked on ATT-Meta (John Barnden; Mark Lee; Sheila Glasbey and Rodrigo
Agerri) permeates this work, including many places where acknowledgements
should have been given, but have not. This work was largely supported by grants:
EP/C538943/1 and GR/M64208, from the Engineering and Physical Sciences Re-
search Council.
Notes
1. I shall follow the convention of capitalising source-target mappings and continue the pro-
cess both when the mappings are from concept to concept, rather than domain to domain and
also for types of information that are mapped in an unchanging or invariant manner between
source and target.
2. But unlike Relevance Theory where the interpretation of a metaphorical utterance is broad-
ened or narrowed until its relevance to the discourse is able to ground the inference process, I
will argue that what grounds the inference process is whether information of a specific type that
is relevant to the discourse can be inferred.
3. This analogy implies no commitment to Chomskyan linguistics. The importance of con-

structions has recently received a great deal of support from linguists working in a number of
different theoretical approaches (e.g. Goldberg, 2006; Sag et al., 2003).
4. I also make no assumptions concerning the embodiedness of the concepts involved.
5. Using the term Conceptual Metaphor Theory overlooks the fact that there have been major
changes and advances in the theory over the years (see Lakoff, 2008).
6. Barnden (2008) has a good discussion of some of the problems such a view of domain
would pose for metaphor theory.
7. According to the online edition of the OED (http://dictionary.oed.com/) the meaning of

“disenable” is: To render unable or incapable; to disable: the reverse of enable.
8. A phrase motivated by organization is physical structure that is not about theories

might be: “the fabric of society”. A phrase motivated by viability is erectness that is not
about theories might be: “This situation will not stand”.
9. For reasons such as these, Grady abandons the notion of a domain for what he calls “pri-
mary metaphors” and assumes instead concept-to-concept mappings, relying on a somewhat
sketchy notion of the unification and/or blending of primary metaphors to account for meta-
phorical systems such as theories are buildings.
10. Examples were found in which “off ” had been misspelled as “of ” and the words occur
within the larger phrase “I have not lost the birth-right of my mind” in Act 3, scene 1 of John
Dryden’s play Aureng Zebe.
11. Note that to define a domain as any entities, relations, attributes and so on that can be in-
ferentially linked to a core element leads to a circularity.
12. In ATT-Meta theory, entities that are not directly part of a mapping, either a pre-exist-
ing mapping such as the correspondences in the love as a journey mapping, or a mapping
that needs to be created on the fly, are termed “Map-Transcending Entities” or MTEs. One of
the principal theoretical claims made by ATT-Meta is that very little source-to-target mapping
needs to be done on the fly.
13. We might go further than stating that unmapped source entities allow inferences to be
made to stating that they initiate the inference process. In the ATT-Meta system this is en-
shrined in the principle of “Transfer-Warranting by Unmapped Structure” (see Wallington et
al., 2006; Barnden & Lee, 2001).
14. The accompaniment ought to be emphasised here. I am not claiming that a colour cannot be
the target of image metaphors. An added complexity is that the presence or absence of a deter-
miner seems to make a difference. “John’s is pillar-box” seems marginally more acceptable to me.
15. The importance and widespread use of generic as specific metaphors is often overlooked,
the generic as specific being disguised by the terminology used to describe the source and
the target. Thus, Grady decomposed theories as buildings into persisting is remaining
erect and organization is physical structure. But “persisting” is a close synonym of “re-
maining” and a “structure” is a type of “organization.” Consequently, the two metaphors could
be reworded as: persisting is persisting in an erect state and organization is physical
organization. If this is done, the generic is specific structure is brought out.
16. The term COMMAs (Conceptual Metaphor Mapping Adjuncts) was used instead of VN-
MAs in Wallington et al. (2006).
17. From Linden Gross, "Facing up to the Dreadful Dangers of Denial," Cosmopolitan, 216 (3),
USA ed., March 1994 (see Barnden, 2008, for further discussion).
18. This is essentially Jäkel’s (1995) mind as a workshop metaphor.
19. There s an alternative, but very similar analysis of these sentences and that involves re-
placing the mental activity as physical interaction (with an object) mapping with
cognizing as seeing (an object), and inferring from the phrase “the far reaches” that the
idea-object can barely be seen. In fact, we might collapse the two different mappings into the
more general mental activity as interaction with an object through the senses, with
the former involving the sense of touch and the latter, the sense of vision. This, of course, sug-
gests that the senses of hearing, smell, and taste can also be used to describe mental activity. I
will not pursue the idea further here.
20. Note the existence of the degree vnma at work as well.
21. This VNMA frequently operates in conjunction with the mental and emotional states
vnma, to be discussed shortly, allowing the inference of a choice.
22. Or possibly cognizing as seeing; it is often difficult to determine whether the tactile or
visual senses are being utilised.
23. There is some dispute in the literature as to whether or not the Event Structure of idioms
is the same as that of their literal equivalents. See Glaseby (2003, 2007) and McGinnis (2002,
2005) for opposing views.
24. I assume that mental and emotional states are frequently combined and interconnected,
making it difficult to talk of pure mental or emotional states. It may even be the case that there
is a cline between mental and emotional states of mind.
25. As with the degree vnma, rate must be interpreted qualitatively rather than in terms of
some absolute rate, for example as: slow, fast or at a medium rate, and inferred using everyday,
common sense, knowledge about the metaphorically cast object. For example, the following
sentence was found via an internet search: “Evolution occurred at a glacial pace” (www.straight.
com/print/118354). It can be inferred that glaciers move extremely slowly and this very low
degree is what is transferred to the rate of change of evolution. But of course, glaciers move at
many times the order of magnitude quicker than evolution proceeds.
26. This is broadly the same as the core metaphor I have called mental activity is physical
interaction.
27. It should be noted that space permits consideration of only a selection of his examples and
I shall usually discuss only some of the inferences that may be drawn, but see Wallington and
Barnden (2006) for a fuller analysis.
28. In fact Grady (1997a) motivates the importance is weight primary metaphor by just this
type of inference.
29. See Barnden’s Metaphors of Mind database http://www.cs.bham.ac.uk/~jab/ATT-Meta/

Databank/ for discussion of the relation between mind as a container and mind as physi-
cal space.
30. An anonymous reviewer has questioned “razor wit”. The following is from an internet
search: “Sexual perversity in Donegal. With whiskey, suicide and razor wit, the Dublin Theatre
Festival is in fine form.” www.bookrags.com/highbeam/art-sexual-perversity-in-donegal-with-
19991013-hb
References
Barnden, John A. (2008). Metaphor and artificial intelligence: Why they matter to each other.
In R. W. Gibbs, Jr. (Ed.), The Cambridge handbook of metaphor and thought (311–339).
Barnden, John A. & Mark G. Lee (2001). Understanding open-ended usages of familiar concep-
tual metaphors: An approach and artificial intelligence system. Technical Report CSRP-01-
05, School of Computer Science, The University of Birmingham.
Barnden, John A., Sheila R. Glasbey, Mark G. Lee, & Alan M. Wallington (2003). Domain-
transcending mappings in a system for metaphorical reasoning. Conference companion to
the 10th conference of the European chapter of the Association for Computational Linguistics
(EACL 2003), 57–61.
Barnden, John A., Sheila R. Glasbey, Mark G. Lee, & Alan M. Wallington (2004). Varieties and
directions of inter-domain influence in metaphor. Metaphor and Symbol, 19, 1–30.
Carbonell, Jaime G. (1982). Metaphor: an inescapable phenomenon in natural-language com-
prehension. In W. Lehnert & M. Ringle (Eds.), Strategies for natural language processing
(415–434). Hillside, NJ: Lawrence Erlbaum.
Carston, Robyn (2002). Thoughts and utterances: The pragmatics of explicit communication.
Oxford: Blackwell.
Clausner, Timothy C. & William Croft (1999). Domains and image-schemas. Cognitive Lin-
guistics, 10, 1–31.
Clausner, Timothy C. & William Croft (1997). The productivity and schematicity of metaphor.
Cognitive Science, 21, 247–82.
Benjamins.
Fillmore, Charles J. (1977). Topics in lexical semantics. In R. W. Cole (Ed.), Current issues in
linguistic theory (76–138). Bloomington, IN: Indiana University Press.
Gentner, Dedre (1983). Structure-mapping: A theoretical framework for analogy. Cognitive Sci-
ence, 7, 155–170.
Gentner, Dedre, Brian Falkenhainer & Janice Skorstad (1988). Viewing metaphor as analogy. In
D. H. Helman (Ed.), Analogical reasoning. Dordrecht: Kluwer.
Glasbey, Sheila R. (2007). Let’s paint the town red for a few hours: Composition of aspect in idi-
oms. In A. M. Wallington (Ed.), Proceedings of the ACL Workshop: The lexicon and Figura-
tive Language (42–49). East Stroudsberg, PA, Association for Computational Linguistics.
Glasbey, Sheila R. (2007). Aspectual composition in idioms. In L. de Saussure, J. Moeschler &

G. Puskas (Eds.), Recent advances in the syntax and semantics of tense, aspect and modality
(71–89). Berlin & New York: Mouton de Gruyter.
Glucksberg, Sam (2001). Understanding figurative language: From metaphors to idioms. Oxford:
Oxford University Press.
Goldberg, Adele (2006). Constructions at work: The nature of generalization in language.
Oxford: Oxford University Press.
Grady, Joseph E. (1997a). Foundations of meaning: Primary metaphors and primary scenes. PhD
Dissertation, University of California at Berkeley, Dept. of Linguistics.
Grady, Joseph E. (1997b). Theories are buildings revisited. Cognitive Linguistics, 8, 267–290.
Grady, Joseph E. (1999). A typology of motivation for conceptual metaphor: correlation vs.
resemblance. In G. Steen & R. Gibbs (Eds.), Metaphor in cognitive linguistics (79–100).
Amsterdam & Philadelphia: John Benjamins.
Hobbs, Jerry R. (1980). Literature and Cognition. Lecture Notes, Number 21, Center for the
Study of Language and Information, Stanford, California.
Jäkel, Olaf (1995). The Metaphorical concept of mind. In J. R. Taylor & R. E. MacLaury (Eds),
Language and the cognitive construal of the world (197–229). Berlin & New York: Walter
de Gruyter.
Johnson, Mark (1987). The body in the mind: The bodily basis of meaning, imagination, and
reason. Chicago: University of Chicago Press.
Kövecses, Zoltán (2002). Metaphor: A practical introduction. Oxford: Oxford University Press.
Langacker, Ronald W. (1987). Foundations of cognitive grammar. Vol. 1: Theoretical prerequi-
sites. Stanford: Stanford University Press.
Lakoff, George (2008). The neural theory of metaphor. In R. W. Gibbs, Jr. (Ed.), The Cambridge
handbook of metaphor and thought (17–39). Cambridge: Cambridge University Press.
Lakoff, George (1993). The contemporary theory of metaphor. In A. Ortony (Ed.), Metaphor
and thought, 2nd ed. Cambridge: Cambridge University Press.
Lakoff, Goeorge (1987).Women, fire, and dangerous things: What categories reveal about the
mind. Chicago & London: University of Chicago Press.
McGinnis, Martha (2002). On the systematic aspect of idioms. Linguistic Inquiry, 33, 665–672.
McGinnis, Martha (2002). Painting the wall red for a few hours: A reply to Glasbey (2003).
Snippets, 10. Retrieved from http://www.ledonline/snippets/
McGlone, Matthew S. (2007). What is the explanatory value of a conceptual metaphor? Lan-
guage & Communication, 27, 109–126.
Musolff, Andreas (2004). Metaphor and political discourse. Analogical reasoning in debates about
Europe. Basingstoke, UK: Palgrave-Macmillan.
Núñez, Raul E., & Eve Sweetser (2006). With the future behind them: Convergent evidence
from Aymara language and gesture in the crosslinguistic comparison of spatial construals
of time. Cognitive Science, 30, 1–49.
Sag, Ivan, Thomas Wasow, & Emily Bender (2003). Syntax: A formal introduction. Stanford:
CSLI Publications.
Sperber, Dan & Deidre Wilson (2008). A deflationary account of metaphor. In R. W. Gibbs, Jr.
(Ed.), The Cambridge handbook of metaphor and thought (84–109). Cambridge: Cambridge
University Press.
Steen, Gerard J. (2007). Finding metaphor in grammar and usage. Amsterdam & Philadelphia:
John Benjamins.
Stern, Joseph (2000). Metaphor in context. Cambridge, MA: MIT Press.
Vega-Moreno, Rosa E. (2007). Creativity and convention: The pragmatics of everyday figurative
speech. Amsterdam & Philadelphia: John Benjamins.
Wallington, Alan M. & John A. Barnden (2006). Similarity as a basis for metaphor: Invari-
ant transfer and the role of VNMAs. Technical Report CSRP-06-02, School of Computer
Science, The University of Birmingham.
Wallington, Alan M., John A. Barnden, Sheila R. Glasbey & Mark G. Lee (2006). Metaphorical
reasoning with an economical set of mappings. Delta, 2, Especial, 147–171.
chapter 12
Attitude, style and context

Matching cognitive and aesthetic accounts
of poetic interpretation
Elsbeth C. Brouwer
University of Amsterdam
The use of productive imagination was traditionally identified as the cognitive

attitude typifying the interpretation of art or poetry. From the point of view of
cognitive linguistics, however, all types of discourse are inherently metaphori-
cally structured, and therefore require a productive use of imagination for their
understanding. From this perspective, the traditional criterion for determin-
ing the poetic character of a given text, i.e. through the productive, cognitive
response it elicits, then fails to delimit poetic text. However, the question of
how we then recognize poetry as poetry remains unanswered within cognitive
approaches. Although some literature exists on the distinction between poetic
or literary and other metaphors, a systematic study of the recognition of genre
and related strategies of interpretation seems to be lacking within the cogni-
tive approach. In this chapter I discuss alternative aesthetic characterizations of
art/poetry, based on style and context, and explore their relevance for a cogni-
tive approach to poetic interpretation. After an initial theoretical discussion of
such alternative criteria in the first part, the second part of the chapter pursues a
more concrete analysis of how these aspects may function in interpretation. For
this, I explore how a nowadays widely accepted interpretation of works from the
historical avant-garde has come to be construed. I conclude with a brief discus-
sion of the ways in which a cognitive approach may benefit from aesthetics in
the investigation of poetic interpretation.
Keywords: aesthetics, poetic metaphor, interpretation in context
1. Introduction
In philosophy there is a long tradition of defining the poetic through the kind of
interpretation it requires. As an exemplary instance, poetic metaphors have been
246 Elsbeth C. Brouwer
held to present a new world-vision (Ricoeur), to intimate rational ideas by analogy

(Kant), to prompt new insights (Davidson), to change concepts (Black), to create
similarities (Indurkhya), and so on. The productive combination of concepts in
imagination has thus generally been considered as the hallmark of aesthetic and
hence also of poetic interpretation. Such emphatically productive aspects of inter-
pretation, historically recognized in aesthetics, now recur in theories that model
cognition more generally. The dynamics of metaphorical interpretation is therein
generalized to such an extent that it applies to all acts of conceptualization; poetic
interpretation included. From the cognitive perspective, we can then no longer
diagnose a given text as poetic on the basis of the special productive use of imagi-
nation, involving analogy, projection or a combination of concepts: These cogni-
tive resources are no longer held to be exclusive for aesthetic interpretation.
Since the characterization of the poetic in terms of imaginative processing
largely coincides with that of metaphorically structured discourse more gener-
ally, it seems that the moral from cognitive science for aesthetics is that the given
understanding of interpretation unifies the poetic with other discourse. As such,
the move toward conceptual metaphor can even be seen to “rob metaphor of its
special qualities” (Gibbs, this volume). Indeed it seems to be the explicit aim of
different authors to show that the meaning of poetic language is not substantially
different from that in other discourses. Stockwell claims that in Cognitive Poet-
ics generally the programme is to reconnect literary phenomena with ordinary
everyday experience of language (Stockwell, 2002), just as Turner claims that it
attempts to offer an alternative to the “mandarin activity” of literary theorists
(Turner, 1991). As such, Lakoff and Turner set the example with their More than
Cool Reason, in which the cognitive theory of metaphor is used to understand
some well-known poems. Leaving aside the question whether their interpreta-
tion fully does justice to these poems when they point out their connection to
everyday meanings, it is a rather different question whether such an approach
can account for the exceptional status such poems (and others) have acquired.
Throughout their book, Lakoff and Turner seem to take for granted that there
is poetry, and that poets do things like “creating new ways of understanding the
world” (p. 203). Yet they discuss neither how we establish that a text presents such
a new way of understanding, nor how we recognize a text as being poetry. From
the perspective of cognitive metaphor theory, as noted above, the characterization
of poetic text requires a definition of genre that does not appeal to a cognitive at-
titude or a specific strategy of interpretation. How do we then recognize a text as
poetic, and does this have an effect on the interpretation we engage in? And if so,
does that mean we do after all have a different cognitive approach to poetic texts?
In other words: can the cognitive theory of metaphor explain what sets poetic
text apart from other genres? Is there something specific to the interpretation of
Chapter 12. Attitude, style and context 247
poetry, such that ‘mandarins’ want to treat it differently from other types of dis-
course? How are we to explain the specific nature of poetic interpretation, while
acknowledging that other types of discourse are interpreted with the use of the
same cognitive resources? Or, to place the question within the scope of the cog-
nitive theory of metaphor: what would distinguish a poetic metaphor from any
other metaphor?
It is precisely on this topic that aesthetics may be of relevance for the cognitive
theory of metaphor. Below I explore several alternative aesthetic characterizations
of art/poetry that could be of use for the cognitive theory of metaphor. In the
first part I consider a number of definitions of the poetic (or of art in general)
that bear upon attitude of interpretation, style, and cultural practice: topics whose
relevance is only just beginning to be explored within the cognitive approach. The
definitions discussed are drawn from the philosophical field of aesthetics, from
art theory and poetics. In the second part, I focus on how these perspectives can
be united with a cognitive approach, in line with Gibbs’ call for a rapprochement
between these fields (Gibbs, this volume). To this end, I present a case study: a
discussion of the interpretation of two works from the historical avant-garde. I
conclude my discussion with a brief indication of how a cognitive approach could
benefit from the ‘mandarin’ activities of literary theorists and others: it is precisely
the analysis of this practice that helps us to understand what is poetic about un-
derstanding cognitive metaphors occurring in poetry.
2. Diagnosing the poetic
2.1 Aesthetics: Detachment
The founding father of aesthetic theory is Immanuel Kant, who gave a fully-
fledged cognitive account of aesthetic interpretation in his Critique of Judgement
(hereafter CoJ). One of the main theses in this work is that aesthetic interpreta-
tion involves subjective judgement based on a reflective use of imagination. Pure
aesthetic judgements, for Kant, are the result of reflection on conceptually unde-
termined objects; they are judgements pertaining to how the cognitive faculties
can deal with form, rather than yielding a purposeful conceptualization of given
objects. In the contemplation of an artwork the subject attempts to produce a
judgement on its aesthetic value. For this, one contemplates the object with re-
gard to its effects on the faculties of understanding. In this process, the object
itself is never fully determined; the interpreter is not interested in the objective
existence of the object, nor in its purpose, but merely in the effect upon her facul-
ties. Hence, if the work of art allows for long-lasting or even endless imaginative
contemplation, we take delight in its form. In such a case, the object will be more
likely judged beautiful than one with clear-cut form, which one cannot help but
recognize and determine immediately. Thus Kant states that aesthetic reflection
requires a disinterested attitude, meaning that in this state of mind you are not
supposed to be concerned with determinate conceptions of the objects in front of
you. The freedom of imagination in the process of interpretation is thus secured;
nothing depends on the possible outcome of aesthetic reflection. Kant calls this
attitude one of objective disinterestedness: The subject engages in a contemplation
of its own imaginative representations, and has no interest in the objective exis-
tence of the object.
To characterize the realm of art by means of disinterested reflection does of
course not determine which objects belong to that realm, since aesthetic inter-
pretation depends on the (disinterested) attitude of the interpreter and not on
(determined) properties of the object. Hence, objects of art that conventionally
belong to the public domain of art need not become the object of individual aes-
thetic reflection. Kant further remarks that some individuals are capable of see-
ing something that was previously determined conceptually as if it were new and
unfamiliar. Therefore, even the most common objects could become the object of
aesthetic reflection (CoJ: 231). This aspect of Kant’s analysis has been interpreted
in DeDuve (1998) as implying that any common object can therefore be consid-
ered as art, a view which reinforces the idea that there is indeed no separate realm
for objects of art.
The idea that aesthetic reflection involves objective disinterestedness has
pervaded deeply in aesthetics, and can be witnessed in different forms in many
theories of art or poetry. It reoccurs, for instance, in the concept of aesthetic dis-
tance, or detachment from the everyday world, traditionally described (or rather
prescribed) as the proper attitude in contemplating art in for instance Bullough
(1912), or Langer (1953).
The feature of disinterestedness in aesthetic judgement is further echoed in
Jakobson’s notion of the poetic function. Dominance of this function entails that
the text draws attention to itself through the purely formal qualities of the poetic
message, at the expense of the referential (objective) function of the text. This
aspect of interpretation can also be witnessed in the possibility of dissociating
from a given context, and of then imaginatively invoking other, new contexts of
interpretation. The possibility of dissociation is thus firmly tied to the possibil-
ity of applying concepts from one domain to objects belonging to another. All
metaphorical projection, in this respect, presupposes the possibility of detach-
ment; the latter can then in itself hardly be understood as specific for aesthetic
interpretation.
The dissociational aspect of metaphorical interpretation was taken up in

Ricoeur (1978). For Ricoeur, the suspension of reference is a crucial moment in
the interpretation of poetic metaphor, since it allows the interpreter to form a
new, poetic ‘world vision’. That is, after first undoing any objective references in
an interpretation, the interpreter imaginatively forms a ‘world vision’, which he
may only then newly apply to reality. Once again, the interpretation of poetry and
art is related to an attitude of objective disinterest as a condition for imaginative
reflection. To recognize poetry as poetry, according to this view, encourages the
interpreter to disengage from an objective interest, and to use productive imagi-
nation to metaphorically describe the world under a new perspective.
Hence productive metaphorical interpretation is identified with poetic in-
terpretation. The question we were trying to answer is how the reader would
recognize poetry as poetry, or art as art – how would a reader recognize which
text should be interpreted in this way? Since metaphorical processing is properly
identified as not merely occurring in poetic interpretation, but in virtually every
type of linguistic understanding, the question of how to determine whether a text
is poetic has become even more acute. On the one hand, it no longer suffices to
say that poetry is by definition whatever we interpret metaphorically; yet, surely,
not every other text we interpret metaphorically allows for the formation of ‘new
world-visions’ or makes use of ‘free imagination’ to the extent that is emphasized
here. The question is whether there is another definition of poeticity that could
capture those qualities of poetic text, or at least could explain why poetic interpre-
tation has been thought to stand out in this way.
Despite the dictum that “a metaphor, after all, is not a linguistic expression”
(Lakoff & Turner, 1989: 203) it has been recognized, within the cognitive theory of
metaphor that the overt recognition of whether a text contains metaphors might
influence its interpretation. Thus Gibbs (2002) suggests a distinction between, on
the one hand, a ‘literary strategy’ of interpretation, involving the explicit recogni-
tion of the metaphorical character of a given (poetic) text, and, on the other hand,
the general mode of understanding any given utterance in context, which does
not draw on the actual recognition of the use of explicit metaphorical figuration.
However, this discussion appears within the context of research regarding general
diagnostic criteria for metaphor (e.g. in relation to the research in Steen (2002)
and Steen et al., this volume). In this line of research, the criteria for recognition
are mainly sought with respect to the identification of conceptual metaphors, and
not specifically poetic ones. Still, the discussion serves to show that recogniz-
ing specific properties of a text may influence its interpretation. Furthermore, it
points to an ambiguity in the terminology we have to work with. The term meta-
phor has of course a double understanding: firstly the term (the way it appears in
the Rhetorics of Aristotle) indicates a rhetorical figure, a property of a textual or

stylistic nature. Secondly a metaphor is a cognitive construction, indicating a rela-
tion between concepts. Naturally the cognitive approach to metaphor emphasises
the second understanding. The diagnostic question asks how and whether these
conceptual structures appear in the structure of discourse; that is, it seeks to find
a relation between the two levels on which we may characterise metaphor.
The issue of recognizing poetic text may involve a similarly ambiguous ter-
minology and require a subsequent project of disentanglement; authors such as
Ricoeur thematise the cognitive aspects of poetic interpretation, and take the rela-
tion to the text more or less for granted. There is, however, no reason to assume
that what conventionally belongs to the genre of poetry actually meets such as-
pects of interpretation. The question is how the textual, stylistic properties relate
to poetic interpretation, and whether these may indeed yield a diagnosis of the
poetic. Before turning to those questions, I will first take a closer at stylistic diag-
nostics for the recognition of the poetic.
2.2 Stylistics: Form
The larger issue of recognizing art is already present in Kant’s discussion of ob-
jects of art. Although an object of art is intended for our reflection, it can only be
reflected upon freely – that is, within the attitude of objective disinterestedness –
if it has the guise of not being intended for the purpose of our pleasure. In other
words, an artwork should not betray that it is designed for our faculty of reflec-
tion, since it would then merely represent a technique, or a rule that the artist used
to make the work. Still, art should have some sort of recognizable form: “[I]n all
free arts something of a compulsory character is still required, or as it is called, a
mechanism, without which the soul, which [...] alone gives life to the work, would
be bodiless and evanescent (e.g. in the poetic art there must be correctness and
wealth of language, likewise prosody and metre)” (CoJ: 304, my italics). Hence,
Kant points to the necessity of some formal criteria that allow us to distinguish
objects of art, even if these cannot be recognized particularly obviously.
One such criterion has been inferred from the attitude of detachment, the at-
titude of interpretation discussed above. In the wake of his definition of the poetic
function mentioned earlier, Jakobson (1973) develops what he calls an empirical
criterion for poeticity: a means to establish that the poetic function in the text is
dominant over its other functions. Accordingly, in poetic text the poetic use of lan-
guage outweighs its more practical functions, such as reference or communicative
intent. According to this view, the poetic text draws the reader’s attention to its for-
mation, rather than to its reference, its communicative intentions, or the emotional
state of the writer. That is to say, in a poetic text the principles for selecting words
are constitutive for the construction of the text. The relevant principle of selection
is analysed in Jakobson (1969) as the principle of equivalence, in his terminology
metaphoricity, as opposed to the principle of combination, namely contiguity (or
metonymy). In Jakobson’s terminology, in poetry the ‘metaphorical axis’ is pro-
jected onto the ‘metonymic axis’; that is, words belonging to different conceptual
domains are combined in a text on the basis of some aspect of likeness. ‘Likeness’
here refers not only to conceptual similarity or structural analogy, but also includes
similarities of sound, source, spelling or context. In poetic text the principle of
likeness overrules principles of combination, such as familiar word order, gram-
maticality, or subject-predicate containment. In our times, some of this seems to
resound when Fabb (2002: 217) concludes “verbal art is experienced as aesthetic
because it exploits to the full every option for making verbal behaviour difficult”.
This principle of metaphoricity thus presents a stylistic, empirical criterion
for whether a given text belongs to the poetic genre. To analyse a text as belonging
to a genre, a type of text with a specific function, is sometimes indeed understood
as the ultimate purpose of stylistics, for instance according to the definition of
Widdowson (2002: 167): “[The purpose of stylistics] is to explore the linguistic
texture of texts, literary or otherwise, to converge on their discursive significance
and to typify them as kinds of social action”.
However promising this may look to some, as a general criterion to determine
whether a text is poetic or not, the exploration of ‘linguistic textures’ alone will
not do. First of all, it requires an understanding of genre – or, for that matter, of
types of social action – which in itself seems to be defined exactly through such
stylistic analysis. Stylistic analysis itself, however, is not suitable to present such a
definition, for the following two reasons.
On the one hand, stylistic analysis provides no necessary ground for the de-
termination of the genre, since no stylistic devices are strictly used only within the
genre of poetry or literature – they reoccur in speeches, commercials, songs, graphic
designs and the like. Jakobson also warns against an all too eager application of his
criterion to genres; poetry need not principally be poetic, nor is the poetic function
exclusively restricted to poetry. In other words, even in Jakobson’s stylistic unifica-
tion of poeticity and metaphoricity, the cross-domain use of predicates, signs, or
sounds yields no unambiguous criterion for the context of use in itself.
On the other hand, stylistic properties provide no sufficient condition for the
identification of genre either. If we can learn one thing from the history of lit-
erature and of art, it is that conventional forms of literary text that are dominant
at one time do not restrict what may come to be considered as literature or art
in the future. Many examples can be given to demonstrate the insufficiency of
formal criteria based on traditional genres. One by-now standard example of the
changing form of art is that of the Fountain, an upside-down urinal submitted
by Duchamp to an exhibition in 1917; it was refused as a bad joke at the time.
The same object however is now held to be the founding work of the highly re-
spectable genre of conceptual art. André Breton gave a literary variety of the objet
trouvé when he included the listing of several Bretons in the telephone directory,
followed by his name, under the title PSTT in a poetry book (Breton, 1966/1923).
One could cite many other examples. Put generally, the criterion of conventional
form as determining the realm of art allows only for conformist art forms. It con-
flicts with our view of art or literature as creative, renewing and original, since it
would, for example, exclude the admittance of unconventional text in the poetic
genre. Such observations may tempt one into the romantic thought that uncon-
ventional form itself may be an essential property of art, if indeed we expect of
art that it is original or innovative. However the property of unconventionality
does not in itself provide a criterion, since something can only be unconventional
with respect to certain conventions, leaving others intact. Furthermore, even as a
criterion for merely extending the domain of art, it denies itself. If we expect the
unconventional to be art, a truly unconventional artwork would be one that does
not count as unconventional. But, having figured this out, it would be conven-
tional to make something non-unconventional, and so on.1
Naturally, this does not entail that there are no conventional stylistic char-
acteristics of poetic discourse. Clearly stylistic properties often function as a cue
for different sorts of interpretation, and may as such deeply affect the cognitive
processing of a given text, as empirical evidence shows (cf. e.g. Hoorn, 1997).
That is, however, not to say that the recognition of stylistic properties is either
necessary or sufficient for poetic interpretation. At most, it can yield a partial and
incomplete definition of the poetic genre, and as such only reflects some of the
dominant conventions of poetic form at a given time. It certainly fails to typify the
‘social action’ that writing new, unconventional poetry would be. In the next sec-
tion, we will look closer at the social conventions, customs, and norms surround-
ing the presentation and reception of art, and see whether we may in fact begin to
understand what sorts of social act are involved in the practice of art and poetry.
2.3 Art theory: Valuation
A further aspect Kant mentions with regard to the valuation of selected art-works
is the way we are taught to recognize them. He recommends the study of dead
languages as the field par excellence for training the faculty of taste, since here
the student can become acquainted with beautiful form. Some products of art are
considered to be exemplary, and they allow the individual to properly develop

his own faculty of taste (CoJ: 232). Kant also observes that the appreciation of
the beautiful indicates a sensitivity to what is morally good. To have good taste
is not a culturally neutral characterization, but indicates a person who is more
likely to identify with the values and norms of a society. Hence, one element of
aesthetic judgement is to be trained to give aesthetic judgements within a cultural
tradition, and hence to have appropriated specific cultural values and to judge ac-
cordingly. The use of stylistic conventions in poetic text then may be understood
as a way to adhere to a canonical tradition, and as such presupposes the value
attached to poetic discourse. This points again to what I observed in the former
section, namely that stylistic analysis must necessarily follow an understanding
of the genre of poetry in terms of social action, rather than that it could yield a
definition of it.
In some form, identification with artistic values is always present in dis-
cussions on art, whether in the emphasis on being cultured, or on meeting the
standards set by the canon of great works. Such identification with the aesthetic
seemingly presents a source of pleasure in itself, quite distinct from the cognitive
dwelling on the possible meanings of the object. To read literature or to view art
apparently involves another characteristic: an attachment to the aesthetic, which
does not seem to require any specific cognitive responsivity, nor even an attitude
of aesthetic detachment. Typically, objects of aesthetic reflection are deemed wor-
thy of attention, as a matter of both cultural and personal taste. For example in a
social context, the appreciation of an artwork by a highly regarded person often
leads to a more general appreciation of that art-work in the same context. This
is also reflected in the often discussed relation between expert appraisal and the
economic value of an artwork. If a prominent museum, or a renowned collec-
tor acquires the work of some artist, her work more likely than not becomes the
object of both economic speculation and of art criticism, and is thereby situated
firmly in the realm of art (cf. Bourdieu, 1979; Hughes, 1990).
Overt recognition of the aesthetic value of a given work within society, as
expressed in its economic value or in the appearance of reviews, may provide
both a reason for and guidance in aesthetic reflection. Such exemplary valuation
of artworks reveals another type of conventionality in aesthetic reflection, namely
that of an institutionalized socio-cultural practice of interpreting and judging art.
Some authors think that it is through such conventions that we can determine
whether something is art. According to Dickie (1974), the social context that
brings an object to our attention as art determines whether we look upon it as art.
Something is a work of art if it is an artefact brought to our attention as a “candi-
date for appreciation” through the judgement of one or more agents acting in the
name of a specific social institution: the art world. Dickie’s criterion of convention
is not conservative with respect to the form of the artwork; as long as an agent of
the art world presents it as such, any object can be appreciated as art.2
The general problem with such an approach is that mere cultural or institu-
tional conventionality presupposes another criterion for deciding what art is in
an object’s first judging. That is, the expert agents of the art world are left in need
of a proper criterion for their judgement; hence the criterion of institutional rec-
ognition can only provide a secondary, rather than a constitutive definition of art.
The institutional analysis does, however, reveal two sides to the question on how
art should be recognized. On the one hand, there are conventionally accepted
works of art. One learns to appreciate these as a matter of cultural education, and
hence recognition is partly determined by cultural tradition. On the other hand,
this collection of publicly recognized works is continuously extended through the
recognition of new works of art. If Dickie is right in emphasizing the institutional
status of art, then his theory should be supplemented with first an understanding
of how the individual (an ‘agent of the art-world’) recognizes how and when this
collection may be extended, and secondly an explanation how it is possible that
even agents have differences of opinion.
Lüdeking (1988) discusses the normal use of the word ‘art’ in a Wittgenstei-
nian manner. When someone judges something to be art, we recognize that that
thing has been attributed a certain value. Thus, if an extra-terrestrial friend from
Mars were to translate our word ‘art’ into *art*, and use it for exactly the same
objects as we do, but then proceed to treat these objects as common garbage, we
would not think he had grasped the meaning of the word. Moreover, if he newly
presented some object as *art* and then encouraged us to treat it like dirt we
would not consider the object as art (Lüdeking, 1988: par. 64–69). The epithet ‘art’
to us is a title of honour; it means we treat the bearer of the name with reverence,
and social codes demand from other people that they respect our valuations, even
if they do not agree with or understand them. An object can be presented as art
before any aesthetic reflection is engaged in. Hence it is not only the outcome of
individual reflection upon its value that determines what is art, but again also
a social fact. This conception of conventions surrounding artistic objects goes
beyond Dickie’s institutionally proclaimed status of art objects, since it points to
norms and values of a more general cultural practice.
Indeed, such broad cultural practice, governed not only by norms of polite-
ness but also by traditional standards of judgement, seems to come closest to what
could be called a normal context of aesthetic interpretation. To name an object
‘art’ cannot be just a matter of personal baptism, since, in doing so, we point to
the object as something valuable, deserving special attention by others as well. We
expose the object as one that is worthy of reflection. In other words, to personally
adopt a specific attitude is not the same as calling something art. The interpretation
is a private matter between the interpreter and the object. But to call an object ‘art’
is a social deed; one points to an object as a candidate for more general appraisal.
Social factors, such as one’s reputation, intentions and credibility on the one hand,
and formal aspects, such as the more or less conventional make-up of the object
on the other hand, play a role in how such judgement is received, and whether it
is followed by others. Depending on the social role of the interpreter, as well as
on how aspects of form and presentation meet culturally formed expectations, an
object may acquire the social status of art after being named ‘art’.3
Parallel to the use of the word ‘art’, the epithet ‘poetic’ indicates an utterance or
phrase that deserves special attention. Calling an utterance a poetic metaphor is to
say that it is of specific value, and promises a rewarding moment of reflection. The
poetic nature of a text is then not just a property of the text drawing attention to
itself, but also depends on contextual presentation, and on the social convention
that given its presentation in such a context, the text should draw the interpreter’s
attention to the text in itself. Thus, to recognize a text as partaking in a poetic con-
text allows a reader to approach the text as one whose value has been established,
and thus poses a normal occasion for poetic interpretation. In such a recognizably
normal poetic context, we know very well how to go about finding relevant formal
aspects, as well as indications of rewarding imagery or figuration. The explicit rec-
ognition of metaphors in a literary text is hence prepared by our recognition of its
genre, and the same holds for the recognition of stylistic properties of the text. That
is not to say that stylistic characteristics and the recognition of overt metaphori-
cal imagery have no role in the recognition of a text as poetry. However, their role
should not be overstated; contextual factors, such as historical background knowl-
edge, presentational conventions, expectations within a given aesthetic practice, or
explicit indications, all help to establish the poetic character of a given text.
When we thus recognize the aptness of poetic interpretation, we get prepared
to bring along our own domains of interpretation, as well as comparisons to previ-
ously read texts. Equipped with all our guiding tools for interpretation, we have
prepared ourselves for a ‘poetic’ reading, maybe even before having read a single
word. In this sense, any cognitive analysis of poetic text is simultaneously bound to
an analysis of social context. Within the context of cognitive linguistics, the need
for an analysis of social context has recently been emphasized by Weber (2005),
who also points to methods of discourse analysis developed by Van Dijk (cf. e.g.
Van Dijk, 1988, 1995 and 1998). Here, however, the issue rather seems to be how a
given text may be interpreted in relation to a social practice, and, as such, presup-
poses clarity on a more principal question: what constitutes a poetic context.
With these considerations, we must now conclude that many factors play a
role in calling a given text or utterance poetic. A cultural practice of poetic inter-
pretation allows us to attach values to specific formal qualities, as well as to certain
circumstances of the presentation of a given text. A purely cognitive definition of

interpretation, I conclude, can hardly yield a diagnostic criterion for poeticity, and
neither, as we saw earlier, can a purely stylistic one. The question, then, is whether
and how interpretation can be affected by the recognition of the poetic character
of a given text. Do such factors as contextual presentation yield a specific cogni-
tive response, and if so, how do they do this? In the next part of this paper I dis-
cuss the history of the reception of Duchamp’s Fountain, by way of a case study. In
the discussion, I try to reconstrue a widely accepted contemporary interpretation,
and thereby focus especially on the interplay of contextual and cognitive aspects.
In doing so, I consider how both context and metaphorical interpretation may
play a role in the interpretation of innovative poetic works.
3. Interpretation in practice: Extending the cognitivist approach
3.1 The reception of the objet trouvé: A case study
After a troublesome initial reception, Duchamp’s Fountain is presently regarded

as the founding work for the genre of conceptual art.4 It consists of a common
urinal, hung upside down, with a signature in black paint reading “R. Mutt”. It
can easily be felt that the imaginative pondering on a urinal is in itself not quite in
agreement with the assumed sensitivity for the moral good that artistic training
was supposed to engender, according to traditional aesthetics. In line with such
considerations, the object was refused for an exhibition of the Society of Indepen-
dent Artists in New York in 1917.5 Nevertheless, mention of the Fountain and the
scandal surrounding it persisted. Photographs of it circulated, a (second) urinal
appeared in an exhibition in the 1940’s, and finally a replica signed by Duchamp
was exhibited in 1964.
At its initial reception, the discussion focused on whether the object could be
accepted as art, and consequently the epithet ‘art’ itself became the topic of discus-
sion. On these grounds, Lüdeking (1995) interprets the Fountain as a ‘semantic
experiment’ redirecting the course of art theory. That is to say that in discussions
on the value of the work itself, one can only get entangled in higher-level dis-
cussions of the meaning of ‘art’. In such a discussion, the interpreter ends up in
a self-reflective examination of attitudes of interpretation. Put like this, viewing
the urinal as art requires one not so much to detach oneself from the common
purpose of the object in the Kantian sense noted above, but rather to distance
oneself from the ‘normal’ approach to art, such as an imaginative pondering on its
representation or mood. Hence, the work’s interpretation is generally formed on
a meta-level of theoretical conceptions of art. Indeed, in the wake of the Fountain
aesthetic theories of interpretation have taken a radically different course. For

instance DeDuve (1998) concluded on similar grounds that after Duchamp, any
object can be rightly termed art by any individual, and recently Carroll (2004) and
Shelley (2004) have discussed the possibility of ‘non-perceptual’ aesthetic qualities
in relation to the objet trouvé, thereby questioning the traditional understanding
of aesthetic interpretation as contemplation and conceptualisation on the basis of
sensory presentation alone.
It is certainly the case that a different type of aesthetic interpretation is ap-
pealed to in the interpretation of the Fountain as a ‘semantic experiment’. The
object does not present new or original form to the viewer: the object is quite
familiar, not unique, and it represents a social taboo. Each of these aspects seems
in blatant contradiction to the definition of art as objects that bring ‘delight to
the senses’. Where once the term ‘reflection’ in connection with the Fountain
would have indicated the contemplation of the images and thoughts called up
by a urinal, the same term can now be used with a different sense: reflection on
the concept of art. In literature and theatre ‘experiments’ have similarly paved the
way for novel theories of interpretation. The strategy of Befremdung or estrange-
ment propagated by the Russian Formalists and later explored by Brecht, as a
means of making the theatre audience aware of the reality of the situation, simi-
larly makes presentational conventions explicit. Other examples of such experi-
mentation are abundant, and would include absurdist writings, surrealist poetry,
or DADA texts, such as the Breton poem mentioned earlier. Henceforth it is
possible to – anachronistically – classify these works under the aspect of form,
for instance as a strategy of semantic experimentation, or as conceptual art that
lacks formal concern.
In the case of the Fountain, nobody, including Duchamp, expected or even
wanted art theory to follow suit and turn the work into a defining exemplar for a
respected genre.6 In a sense, its present day respectability is precisely the failure
of the historical avant-garde: art theory simply evolved and usurped their revolu-
tionary works. What we can now recognize as ‘novel form’ was in fact an aggres-
sive attempt to undermine, ridicule and ultimately end art, as well as its embodied
ideology as it existed then, an intention clearly expressed in the many manifestos
of the time.
Lakoff and Turner (1987) see avant-garde writers as taking up a “standard
challenge” of expressing new metaphorical schemes “to create new ways of un-
derstanding the world” (p. 203). They sum up the ways in which avant-garde po-
ets may do so: by offering new modes of metaphorical thought, by employing
conventional basic metaphors in unusual ways, or by destabilising them and thus
revealing their inadequacies (pp. 51–52). However, throughout their book they
offer no analysis of how such radical innovation or destabilisation would take
place, other than by mentioning that some metaphors – the less conventional
ones – seem to be “more dispensible” (p. 56). As explained above through the
work of Duchamp, the historical avant-garde has not only led to the introduc-
tion of new formal genres within the arts, but has also, and more importantly,
introduced a different way of looking at art. From the point of view of cognitive
theory, it is then worthwhile to examine more closely how the historical avant-
garde created such “new ways of understanding the world”. In the remainder of
this section I will try to reconstrue how the more recent classifications of objets
trouvés have come about, and how form, context and theory have come into play
with the cognitive process of interpretation. For this purpose, I turn to a literary
variety of the genre: Breton’s poem PSTT.
PSTT is the title of a page in a poetry book. It presents a list of names (all
‘Breton’), occupations and telephone numbers. Its last line is “André Breton”. The
graphic presentation of the lines is that of a telephone directory. Since it does not
present any text beyond a fragment of the telephone listing, the poem poses the
reader a similar situation as with Duchamp’s urinal. What should a reader looking
for poetry do with such a prosaic listing? For this we may recall Jakobson’s broad-
er notion of ‘metaphoricity’, indicating similarity on any plane, whether semantic,
visual, intertextual or auditory. From the point of view of style, the organisation
of PSTT is similar to that of a poem; the occurrence of the author’s name seems
to be in line with a presupposition of autobiographical content, and the repetition
as well as the use of white space conform to stylistic devices in poetry. As such,
it perfectly displays what Jakobson called poeticity: ‘metaphorical’ similarity and
lack of communicative or (in this context) purposeful content.7 Thus it can be
interpreted as a highly polemic statement against classical poetry, which as a rule
presents rich metaphorical imagery and appeals to the reader by its potential rich-
ness of emotional expression through the use of the same stylistic means.
The starting point of this antagonistic interpretation is, however, not so much
the formal, stylistic analysis of the printed text, nor the search for a metaphorical
scheme of similarity. Rather, it starts with the need to achieve an interpretation of
this page in the book as a poem, that is, in relation to some understanding of what
poetry is. Hence, the text functions as poetic text through its context of presenta-
tion – that is, it becomes a metaphor of the form: telephone page is poem. The
explicitly reflective quality of the interpretation it acquires, as at once a parody
and a poetic statement on the nature of poetry, in itself involves another meta-
phorical – or rather metonymical interpretation. This interpretation is formed
through the construal of the given poem as a poem, that is as an exemplar for
the poetic genre, or, in the terminology of Nelson Goodman: the poem exempli-
fies its genre. As such, within the terminology of cognitive metaphor theory, we
may consider the poem as a source-object, which allows us to understand the
more general or abstract concept of poetry: poem stands for poetry. The text
has sufficient similarity to the target (poetry) in respect of its presentation, both
regarding formal aspects (metre, repetition, blank spaces) and its contextual pre-
sentation (collection of poems) to warrant the collapse of these two into a single
scheme: telephone page stands for poetry. Its obvious prosaity conflicts with
this understanding. And yet, it is precisely its triviality, unimaginativeness, liter-
ality and so on that allow for a polemic interpretation. The occurrence in a book
of poetry adds to the plausibility of this interpretation: the surprise or Befrem-
dung that the appearance of this alien kind of text calls up in the given context. It
prompts the question: why? and as such invites further thought on the nature of
presentation, genre and interpretative response. In the metaphorical construal of
interpretation, the raising of questions carries over to the domain of poetic text
more generally, and ultimately leads to a questioning of the dominant conven-
tions for poetic interpretation.
The contemporary view, that such avant-garde texts call into question the tra-
ditional practice of aesthetic interpretation, can thus be understood as the result
of metaphorical interpretation, with the use of metonymy (cf. the analysis of me-
tonymy as an instance of metaphor, in Lakoff and Johnson (1980: 35)). The latter
involves seeing the particular page standing for poetry in general, through its
present appearance in a poetic context. The part-for-whole reading of the page is
sustained by the visual similarities in style between poems and listings. In order to
construe this interpretation, some knowledge of stylistic conventions in poetry is
required, as well as the capacity to recognize the context as poetic. A reader who
only uses the page to conveniently look up the butcher’s telephone number will
not come to construe such an interpretation. As Davidson notoriously noted: in-
terpretations explicitly point to an insight that need not explicitly or semantically
be realized in metaphor. The mere juxtaposition of images, terms, or descriptions
suffices for creating a metaphorical tension (Davidson, 1978). Although I dis-
agree with Davidson’s understanding of semantics and his consequent rejection of
metaphorical meaning, the point he makes here is an attractive one: metaphoric-
ity is construed in interpretation, and is relative to context. Certainly in the case
of PSTT, it is largely due to the contextual presentation that we as a reader come
to consider a page of a telephone book as a model for poetry.
Furthermore, it seems to require some openness or willingness on the part
of the interpreter, to let go of one’s expectations, raised through the knowledge
that one is about to read a book of poetry. An important aspect of the interpreta-
tions discussed is the appeal made to conventions of interpretation. In both the
reading of PSTT and the Fountain, a more narrow, normative conception of art or
poetry not only figures as normal background knowledge that guides us through
the interpretation (construing a context as an aesthetic one), but is foregrounded
through its violation (when obvious everyday objects enter the domain of aesthetic
interpretation). In both cases, the interpretation involves a reflective awareness of
the process of interpretation in general: we reflect upon the work as a critical com-
ment on the act of making, exposing, and interpreting art or literature. Hence both
works under their interpretations here serve as a means to at once render the inter-
preter conscious of his or her (conventional) approach to art and to criticize it.
Just as our standards for interpretation, based on previous experience and
training, help us judge a work of art, we may use the work to reflect on those stan-
dards. Explication of this process, then, is what leads to the theory-driven inter-
pretation of the ready-made as a semantic experiment on the nature of ‘art’. Thus,
we can come to understand the ready-mades as a comment on accepted artistic
practice, thereby affecting and changing that practice. At present it is common
practice in most forms of modern art that we sometimes let ourselves be guided
by theoretical considerations on the nature of art, whether in production, percep-
tion, recognition or interpretation of an artwork. Even naive, realist or expres-
sionist art for instance are motivated and defended from a theoretical position:
the position that art should involve something other than theoretical evaluation,
and ought instead to appeal to intuitive perceptual processing, immediate recog-
nition or excelling craftsmanship.
3.2 Conclusion: Matching cognitive and aesthetic accounts

of poetic interpretation
In the introduction to this paper, I mentioned Turner’s wish to distance himself

(and the whole of Cognitive Poetics) from the ‘mandarin’ activities of literary the-
orists. As we have seen, the willingness to reflect on a theoretical background may
be brought about in some sense by a work of art itself, or by its presentation in a
specific context. As history shows, there seems no ‘right’ or ‘wrong’ theory here,
other than changing notions of what would be in good taste. Obviously in More
than Cool Reason Lakoff and Turner interpret poetry against a theoretical back-
ground of cognitive metaphor theory, that predicts conceptualisations through
more familiar or more concrete basic concepts. With the discussion of the inter-
pretation of the objets trouvés above, I think it clear that a reductive analysis in
terms of basic cognitive metaphors cannot tell the whole story.
In the analysis of the interpretation of the Fountain and PSTT we have drawn
on some understanding of reflective interpretation and of cultural values sur-
rounding contexts of presentation. It was possible to reconstrue different interpre-
tations of these works by appeal to these notions, derived broadly from aesthetic
theory, in addition to an understanding of metaphorical interpretation. It was not,
or not just through a specific, conventional metaphorical scheme that we could

construe such an interpretation, since we rather looked at how such a scheme
could come about in the first place. For this, an understanding of context, and es-
pecially the possibility that this understanding explicitly enters into reflection, was
necessary. Hence, although the role of both general metaphorical schemes and of
general cultural knowledge have already often been emphasized in interpretation,
we have here found another aspect that enters into the aesthetic interpretation of
innovative works of art: explicit reflection on the act of interpretation itself.
The moral of the story, then, is that interpretation turned out to be neither just
cognitively nor just perceptually motivated, nor could we explain it by a ‘recon-
nection to ordinary language’. In the given cases, the objects were themselves very
ordinary. Metaphorical interpretation did not emerge through perceptual proper-
ties of the objects – since we only recognised them for the objects they are. Rather,
such an interpretation only emerged in the effort of contemplating the why? of
their present context, forcing us to become aware of the nature of that context,
and forcing us to reflect explicitly on the values bestowed on it as a matter of
convention. To put it briefly: the interpretation could only emerge after making
ourselves conscious of our background theory on what art is.
When engaging in a cognitive theory of interpretation, which in addition
encompasses social and contextual aspects of the practice of interpretation, we
should not indulge in an aesthetic preference for one or another type of inter-
pretation; nor would the possibility of any such type of interpretation provide
a conclusive argument for a cognitive account. If cognitive metaphor theory is
committed to understanding the cognitive resources that are at work in linguistic
understanding, then, for all its dislike of mandarin interpretations, it must be ca-
pable of explaining even those interpretations that are based on theory and reflec-
tion. Regardless of one’s opinions and preferences here, surely one cannot deny
the cognitive possibility, nor the social reality, of such interpretations. As such, a
cognitive approach to poetry is bound to extend its scope to the investigation and
explanation of existing practices of poetic interpretation.
In order to understand what is poetic about poetic metaphor or its interpreta-
tion, we should then not only take into account cognitive processing or stylistic
conventions, but also social practice with its implicit norms for interpretation.
Indeed, Lakoff and Turner themselves conclude that “[i]t is vital that we under-
stand our worldviews” and that the study of poetry can be a means to promote
“ethical, social and personal awareness” (p. 214). To gain greater insight in how
such awareness grows and influences the practice of writing, interpreting and ap-
preciating poetry, the collective history of aesthetics, stylistics and literary theory
presents a wonderfully documented case-study.
We can safely conclude that apart from the emphasis on imaginative process-
ing, aesthetics shares other concerns regarding interpretational processes with
cognitive theory. From this perspective, aesthetic understandings of the nature of
context and cultural conventions may be of great value for cognitive theory. To
begin with, such understandings may offer greater insight into the nature of cul-
tural conventions and context-bound values. But secondarily, the dynamics of art
theory represents valuable empirical material for cognitive research, as it allows us
to study the interaction between cultural practice and theory in interpretation.
Acknowledgements
I would like to thank the following people for discussion and comments on earlier
drafts of this paper: Marianne Brouwer, Raymond Gibbs, Charles Forceville and
Josef Früchtl. I am also much obliged to Graham Low and to the anonymous ref-
erees of this volume for their suggestions. Needless to say, any remaining mistakes
or misgivings are entirely my own.
Notes
1. The construction of such an argument may seem far-fetched, but it has actually been put
forward, e.g. in Bierens (2000). Bierens gets entangled in this paradox when he argues that the
objet trouvé was an artwork when presented by Duchamp in 1917, but is no longer so when
presented by Tracy Emin in the nineties.
2. A proposal that similarly emphasizes the role of ‘experts’, inspired by formal semantics, was
offered by Matthews (1980) and Carney (1982). Here the word ‘art’ is understood as a natural
kind term with rigid designation, in the sense proposed by Kripke and Putnam. Accordingly
art works are identified as art by experts. Other people use the word with an implicit, causal
reference to such expert usage.
3. Joseph Früchtl noted that Thomä (2001) presents a further argument in his comment on
Rorty. He states that any original act of expression relies on the existent, social use of lan-
guage. Rorty replies that indeed, without such binding to social use, any novel use of an expres-
sion “leaves the individual without a clue as to whether he is a madman or a genius” (Rorty,
2001: 321, my translation). Hence to declare something to be art, even if it is entirely non-con-
formist, would still appeal to social norms for usage of the term.
4. Its general recognition is testified for instance by an auction at Sotheby’s in 1999, where a
1964 replica (one from an edition of eight) of Fountain was sold for $1,762,500. The catalogue,
with a photo of Fountain up front, states that “Duchamp’s legacy lives on, through the young
artists today who are also devoted to investigating conceptual strategies inherent in appro-
priation and replication, increasingly recurrent themes in contemporary art that [...] Duchamp
ingeniously pioneered.”
5. ”We can’t show it [...]. It is gross, offensive! There is such a thing as decency.” From an account
of the meeting of members of the Society, at which the Fountain was rejected (related by Wood,
1985). Arensberg and Duchamp withdrew from the committee immediately after its refusal.
6. ”The substitution of an anonymous article of everyday use can be understood only as an

intellectual or philosophical decision...It is simultaneously a visual demonstration, as Duchamp
has said, ‘of the futility of ever defining art’ and [it is] conclusive proof that all art ...is only to
be understood ...as the victory of consciousness over non-conscious matter, of will over taste.”
(Hamilton, 1976).
7. ”These poems express nothing, and do not want to express anything.” (Jouffroy in Breton,
1966: 16, my translation).
References
Bierens, Cornel (2000). De Beeldende Kunst Moet Onderduiken. In NRC Handelsblad.

19/5/2000.
Bourdieu, Pierre (1979). La distinction, critique sociale du jugement. Paris: Minuit.
Breton, André (1966). Clair de terre. Reprinted with an introduction by A. Jouffroy. Collection
Poésie. Paris: Gallimard. (Originally published in 1923).
Brouwer, Elsbeth (2003). Imagining metaphors. Amsterdam: University of Amsterdam, ILLC-
Dissertation Series.
Bullough, Edward (1912). ‘Psychical Distance’ as a factor in art and an aesthetic principle.
British Journal of Psychology (June), 5, 87–98.
Carney, James (1975). Defining art. The British Journal of Aesthetics, 15, 191–206.
Carroll, Noël (2004). Non-perceptual aesthetic properties: Comments for James Shelley. The
British Journal of Aesthetics, 44, 413–423.
Davidson, Donald (1984). What metaphors mean. In D. Davidson, Inquiries into truth and
interpretation (245–264). Oxford: Clarendon Press. (Originally published in 1978).
DeDuve, Thierry (1998). Kant after Duchamp. Cambridge Mass: MIT Press.
Dickie, George (1974). Art and the aesthetic: An institutional analysis. London: Cornell Uni-
versity Press.
Dijk, Teun, A. van (1988). Social cognition, social power and social discourse. Text, 8, 129–57.
Dijk, Teun, A. van (1995). Discourse semantics and ideology. Discourse and Society, 6, 243–289.
Dijk, Teun, A. van (1998). Ideology: A multidisciplinary approach. London, Sage.
Fabb, Nigel (2002). Language and literary structure: The linguistic analysis of form in verse and
narrative. Cambridge: Cambridge University Press.
Gibbs, Raymond W. (2002). Psycholinguistic comments on metaphor identification. Language
and Literature, 11, 17–23.
Gibbs, Raymond W. (1998). The fight over metaphor in thought and language. In A. N. Katz,
C. Cacciari, R. W. Gibbs & M. Turner (Eds.), Figurative language and thought (88–118).
Oxford: Oxford University Press.
Hall, Geoff (2003). The year’s work in stylistics: 2002. Language and Literature, 12, 353–370.
Hamilton, George (1976). Inside the green box. Appendix to: The bride stripped bare by her
bachelors, even. Stuttgart: Edition Hansjörg Mayer. No page numbers. (Originally pub-
lished in 1960).
Hoorn, Johan (1997). Behavioral and psychophysiological research into literary metaphor pro-
cessing. PhD Dissertation, Vrije Universiteit Amsterdam.
Hughes, Robert (1990). Art and money. In: Nothing if not Critical. Selected Essays on art and
artists (387–404). New York: Knopf.
Jakobson, Roman (1969). Linguistics and poetics. In T. A.Sebeok (Ed.), Style in language (360–
377). Cambridge Mass: MIT Press. (First published in 1960).
Jakobson, Roman (1973). Qu’est-ce que la poésie? Transl. by M. Derrida, in Questions de Poé-
tique (113–126). Paris: Seuil.
Johnson, Mark (1987). The body in the mind. Chicago: University of Chicago Press.
Kant, Immanuel (1952). The critique of judgement. Transl. of Kritik der Urteilskraft (1790) by
J. Meredith. Oxford: Oxford University Press.
Langer, Susanne (1953). Feeling and form. A theory of art developed from philosophy in a new
key. New York: Charles Scribner’s Sons.
Lüdeking, Karlheinz (1988). Analytische Philosophie der Kunst. Frankfurt am Mainz: Athenae-
um Verlag.
Lüdeking, Karlheinz (1995). A brief guide to the avoidance of style. Nordisk Estetisk Tidskrift,
13, 19–32.
Matthews, Robert (1980). Traditional aesthetics defended. The Journal of Aesthetics and Art
Criticism, 38 (1979–1980), 39–50.
Ricoeur Paul (1978). The metaphorical process as cognition, imagination and feeling. In S. Sacks
(Ed.), On metaphor (141–157). Chicago: University of Chicago Press.
Rorty, Richard (2001). Erwiderung auf Dieter Thomä. In T. Schäfer, U. Tietz, & R. Zill (Eds.),
Hinter den Spiegeln: Beiträge zur Philosophie Richard Rortys (319–324). Frankfurt am
Main: Suhrkamp.
Schäfer, Thomas, Udo Tietz, & Rüdiger Zill (Eds.) (2001). Hinter den Spiegeln: Beiträge zur
Philosophie Richard Rortys. Frankfurt am Main: Suhrkamp.
Steen, Gerard (2002). Towards a procedure for metaphor identification. Language and Litera-
ture, 11, 17–23.
Stockwell, Peter (2002). Cognitive poetics: An introduction. London: Routledge
Thomä, Dieter (2001). Zur Kritik der Selbsterfindung. Ein Beitrag zur Theorie der Individua-
lität. In T. Schäfer, U. Tietz, & R. Zill (Eds.), Hinter den Spiegeln: Beiträge zur Philosophie
Richard Rortys (292–318). Frankfurt am Main: Suhrkamp.
Turner, Mark (1991). Reading minds: The study of English in the age of cognitive science. Princ-
eton: Princeton University Press.
Weber, Jean Jacques (2005). From bad to worse: Pragmatic scales and the (de)construction of
cultural models. Language and Literature, 14, 45–63.
Widdowson, Henry (2002). Verbal art and social practice: A reply to Weber. Language and
Literature, 11, 161–167.
Wood, Beatrice (1985/2006). I shock myself: The autobiography of Beatrice Wood. San Francisco:
Chronicle Books.
chapter 13
A genre approach to imagery in winespeak

Issues and prospects
Rosario Caballero and Ernesto Suarez-Toste

Universidad de Castilla-La Mancha
This chapter is concerned with the figurative language used by wine experts in
the genre of the Tasting Note. Our main objectives are to draw attention to the
complexities found in the figurative language used in this genre, and discuss
some of the methodological issues researchers need to overcome when dealing
with specific discourses, and in particular questions concerning the identifica-
tion, classification and interpretation of figurative data in textual contexts. The
analysis is done on a corpus of 12,000 tasting notes drawn from specialized
wine publications, and follows the main guidelines in both cognitive and lin-
guistic approaches to metaphor.
Keywords: winespeak, tasting note, metaphor identification and classification,

genre, synaesthesia
1. Introduction
This paper explores the metaphors underlying winespeak, the specific jargon used
by wine experts to discuss wine. For, if as linguists claim, mastering the language
is a must in becoming a member of a given community (professional or other-
wise), in the wine realm it inevitably implies learning how to use figurative lan-
guage. This may help newcomers not only to understand how reality is construed
by the expert members of the wine community, but also to acquire the most im-
portant tool to overcome the difficulties involved in communicating to others the
experience of drinking wine.
Exploring how wine experts use figurative language involves making deci-
sions on three matters. Firstly, one needs to select a manageable and representa-
tive context from what is often a wide range of possibles. In our view, focusing
on a particular genre can be a useful way to narrow down potential contexts, as
recent approaches to genre (e.g. Swales, 1990, 2004) emphasise both the purpose
266 Rosario Caballero and Ernesto Suarez-Toste
and the patterned quality of the texts, or rather language-use events, involved.
Restricting the focus to one genre can help researchers build up hypotheses about
not just the figurative expressions likely to appear, but also the reasons that might
explain their presence. The researchers can then check their hypotheses against
the data found in the genre analyzed, and discuss the relevance of figurative lan-
guage by relating it to the intentions underlying the authors’ use of such language
and the audience’s expectations. The genre we have chosen in order to discuss
wine metaphors is the Tasting Note (henceforth TN), written by expert wine tast-
ers and critics for a presumably knowledgeable audience – even if, as discussed
later, the genre’s readership also includes non-expert wine lovers who will gradu-
ally gain expertise by regularly engaging with tasting notes.
Secondly, there is a need for a principled way of analyzing linguistic data,
which involves, in the first place, distinguishing figurative from non-figurative
language in the texts concerned – an issue often taken for granted by theoretically-
grounded, cognitive-oriented metaphor scholars, but questioned by the grow-
ing amount of research adopting a more discursive approach (e.g. Goatly, 1997;
Cameron, 1999a, 1999b; Low, 1999; Steen 1999, 2002, 2007, this volume). Finally,
after identifying the figurative language relevant to the study, the next step is to
classify the underlying metaphors into types, in order to determine which expe-
riential domains occur most often, which metaphor(s) are the most prevalent,
and how the diverse figurative schemas acknowledged in the literature (such as
metaphor and metonymy) interact.
These issues are far from simple when attempting to explore metaphor in gen-
eral discourse, but they are particularly problematic when exploring the meta-
phors used to discuss the specific topic of wine. Discussing this beverage is a high-
ly complex task which draws on our range of sensory experiences and depends
on our capacity to isolate and memorize them. Part of the complexity relates to
the fact that our tongues can detect just four flavours (sweet, sour, salt, and bit-
ter); the remaining multiplicity of distinctions is made through the nose. Given
this situation, it is hardly surprising that metonymy and synaesthesia appear to
‘contaminate’ everything in wine description. There is, for example, no such thing
as “a sweet smell”; something smells sweet when we recall past experiences where
similar aromas turned out to belong to sweet-tasting products. This is congruent
with the role of metonymy in salience construal, evoking or highlighting par-
ticular aspects of a given entity, concept, or situation (Panther & Radden, 1999;
Radden & Kövecses, 1999; Panther & Thornburg, 2003; Paradis, 2004), as well
as with synaesthetically motivated expressions, where a property typical of one
sensory domain or mode is used to talk about something perceived through a dif-
ferent sense (Day, 1996; Ramachandran & Hubbard, 2001; Yu, 2003).
Chapter 13. Imagery in winespeak 267
The result of trying to share organoleptic experiences with others is that our
verbal capacity is frequently tested and the limits of the lexical repository at our
disposal are stretched to the point that we rapidly exhaust conventional expres-
sive resources. The issue has been addressed by scholars dealing with wine dis-
course, all of whom have pointed to the need to develop a standardized termi-
nology if wine tasting is to reach a ‘scientific’ or reliable status (Lawless, 1984;
Solomon, 1997; Gawel, 1997; Gawel & Oberholster, 2001; Hughson & Boakes,
2001; Nedlinko, 2006).
One obvious way out of the dilemma is to use figurative language, though
even this solution is not entirely uncontroversial. Thus, although some scholars
see figurative language as an indispensable resource to overcome the shortage of
lexis related to sensory experiences (e.g. Amoraritei, 2002; Lehrer 1975, 1983,
1992, 2009), others, like Peynaud (1987) or Gluck (2003), regard it as some kind
of undesirable (if occasionally necessary) ‘juggling’ with words. Such contradic-
tory views on the value of figurative language in wine discourse may well depend
on what is regarded as metaphorical. By way of illustration, consider the follow-
ing two TNs:
(1) This has a sweet and savoury nose, with pears, herbs and damp straw notes.
It’s quite complex. It smells a bit sweet. The palate is complex and intense,
with an unusual texture: it is a little soft and flat in the middle but then there’s
an acid kick giving it life and definition. It almost has a cidery character.
There’s amazing length and real separation of flavours. […] It’s a wine that is
hard to pin down; it’s continually changing and showing different facets.
(2) [T]he 2001 Batard-Montrachet offers a thick, dense aromatic profile of

toasted white and yellow fruits. This rich, corpulent offering reveals lush
layers of chewy buttered popcorn flavors. Medium-bodied and extroverted,
this is a street-walker of a wine, making up for its lack of class and refinement
with its well-rounded, sexually-charged assets. Projected maturity: now-2009.
Both texts display numerous examples of non-literal uses of language (in italics),
some of which appear to be more saliently figurative than others, depending on
both the degree of acculturation to wine discourse of the person reading them,
as well as his/her approach to winespeak.1 Thus, whereas for some wine experts
and critics the metaphorical expressions in Example (1) are likely to go unnoticed,
those in Example (2) are more conspicuous. In fact, this example may well raise
some eyebrows among readers (wine experts included) even if the – positive –
assessment of a wine by equating it to a whore comes from Robert Parker, cur-
rently one of the most authoritative voices in the field.2 For linguists, however,
both TNs offer interesting figurative data, irrespective of whether they belong to
the conventional language used to talk about wine (e.g. terms like “nose”, “pal-
ate”, “medium-bodied”, or “extroverted”) or they creatively exploit some of the
customary figurative frames underlying wine discourse. Indeed, if wines are often
personified by means of adjectives such as “corpulent” or “extroverted” (see, for
instance, Amoraritei, 2002; Lehrer, 2007), why should they not also be qualified as
“street-walkers” – however far-fetched this may seem to the non-initiated?
Given the extreme difficulty of communicating comprehensibly and with
clarity about wine, the main aim of this paper is to explore the complexities found
in the figurative language used by wine professionals in TNs and in the process
discuss some of the methodological hurdles researchers need to overcome when
dealing with specific discourses. The chapter begins by examining the genre cho-
sen, moves on to a discussion of the methodology followed in the identification of
figurative expressions, and ends by considering the criteria necessary to classify
the expressions conceptually.
2. The research context: The TN genre
Our discussion draws upon the results of a research project on the discourse of
wine funded by the University of Castilla-La Mancha, Spain. The main purpose
of the project was to explore the metaphorical language used in one of the most
representative genres within wine discourse, namely the Tasting Note. In order
to do so, we built a corpus consisting of 12,000 TNs of both red and white wines
written (in English) by the most authoritative British and American wine critics,
and published in Decanter, Wine Enthusiast, Wine Spectator, The Wine Advocate,
Wine News, The Wine Pages, and The Wine Anorak. The corpus includes TNs
evaluating an equal number of red and white wines; in order to ensure repre-
sentativeness, the wines were sourced from the major wine-producing regions
in the world.
Our choice of genre was determined by three considerations. The first was the
quantitative import in the discourse practices of the community concerned.3 The
second was functional importance; TNs play a crucial role in wine acculturation,
in the sense that wine aficionados regularly learn how to describe and assess wines
by reading numerous TNs (Silverstein, 2003, 2004; Lehrer, 1983, 2009). Last, but
not least, the genre needed to be a context where figurative language occurred to
a marked degree.
TNs are short texts (from 20 to 200 words) devoted to describing and evalu-
ating (that is to say, reviewing) wines. They usually fulfil at least four functions,
namely (a) describe and evaluate the aromatic profiles of specific wines and
place them in context with their ‘peers’; (b) describe the wines’ mouthfeel (which
subsumes both the wines’ flavour(s) and texture); (c) provide an idea of their
quality/price ratio; and (d) estimate an ideal drinking ‘window’ for the wines to be
consumed at their best. This is an extremely complex set of tasks which involves
employing the senses of sight, smell, taste, and touch, with the bulk of the assess-
ment being based on a combination of smell, taste and touch, since these three are
intrinsically related in tasting procedures.
The rhetorical organization of TNs is not only determined by the four key
functions (above), but the critical parts or sections of the text in fact mirror the
three key stages in the highly ritualized event of wine tasting (see also Silverstein,
2004): namely the assessment of a wine’s colour, smell (metonymically referred to
as its nose), and ‘feel’ in the mouth (metonymically referred to as its palate). For
illustration, here are two fairly typical TNs:
(3) 2005 Domaine la Noble Merlot. Languedoc Roussillon, France. $9. 88
points. Appealing aromas of sweet red and blue fruits emerge from the nose
of the 2005 Merlot. A terrific value [sic], it is rich, suave, satin-textured
and packed with candied bilberries as well as cherries. This fruit-forward
offering’s finish reveals notes of chocolate before exhibiting some structured
tannin. Projected maturity: now-2008. Importer: Dan Kravitz – Hand
Picked Selections, Warrenton, VA; tel (540) 347-9400.
(4) 86. Finca Sophenia 2005 Altosur Malbec Rosé (Tupungato); $10. Reddish in
color, which announces that it’s Malbec. The nose is sizable but fresh, with
riper, sweeter aromas. A bit more zesty in the mouth than you might expect,
with an expansive, almost Beaujolais-like finish. Well balanced; a good pink
wine. Imported by Tastings Import Company. Best Buy.
As in (3) and (4) above, critics start their commentary by identifying the wine
under assessment, often following this with a first evaluation of the wine (“A
terrific value”), and closing the TN with a final evaluative commentary (“a good
pink wine”) plus recommendations on consumption. The global evaluation of
the wine is also usually provided numerically; most critics have adopted a 100-
point rating system (over 85 meaning that a wine is worth considering). This
score may appear at the beginning of the TN or in a separate ‘technical’ card
where details such as the winery’s location, the wine’s importers, its price, or the
number of bottles produced are provided (although this information may also
appear as part of the main text).
Naturally, both description and evaluation in figurative language are bound
to be susceptible to a greater or lesser degree of subjectivity. This is particularly
true when not only parameters but different degrees, amounts, or measurements
of the parameters involved are being discussed. In the following example is a good
illustration of the problems of wine assessment being subjectively expressed:
(5) It’s a deep coloured wine with a seductive but not showy nose of intense ripe
black fruits. The concentrated palate is bold and spicy with […] smooth but
firm fine-grained tannins.
Although some expressions could be objected to from a ‘scientific’ point of view

in spite of their high ability to communicate effectively (e.g. “sweet nose” in Ex-
ample 1 discussed earlier), there are also apparently contradictory expressions
such as “smooth but firm fine-grained tannins” in (5). The contradiction lies in
the fact that one of the first things wine novices learn is that rough tannin is not
necessarily a bad thing (similarly, one of the first vices acquired by novices is to
equate the “raspy” or “rough” mouth-feel of a wine with its structure; a wine may
have a perfectly sound structure without being raspy at all). Tannin need not be
aggressive after a certain while, so the coupling of “smooth” and “firm” may prove
confusing – though we may note that the author does connect them with the ad-
versative “but”, suggesting an awareness of the potential conflict.
Similarly, taste measurements too are necessarily subjective. Somehow, the
critic must assess the wine relatively, in context with its peers (regionally, varietal-
ly, or from a vintage point of view), as well as in absolute terms. These extremely
subtle appreciations virtually demand that the readers know the author from pre-
vious experiences. How else could one estimate what exactly a writer means by
“a seductive but not showy nose of intense ripe black fruits” in (5)? Where is one
expected to draw the line between “seductive” and “showy”? In fact, expert con-
sumers often speak of ‘calibrating’ their personal tastes against those of the critics
they follow.4 It is an interactive process most of the time, both fun and frustrating,
but it is precisely this ambitious aim for descriptive perfection that tests the critic’s
and perhaps the reader’s verbal creativity to the limit.
3. Identifying figurative schemas in TNs
The first step in exploring metaphors in the TNs was to discriminate between lit-
eral and non-literal (or figurative) expressions in the critics’ commentaries. This
proved an interesting exercise since our respective experiences with wine and,
above all, approaches to the topic were radically different. Whereas one of us is a
wine connoisseur and, as a result, highly acculturated to wine discourse, the other
is a novice in this respect, whose main interest in wine (apart from drinking it) is
the richness of winespeak for metaphor research. Although our identifications of
metaphorical language often coincided, particularly in the case of the most salient,
unusual or ‘flashiest’ expressions like labelling a wine as a “street-walker” (above),
when it came to discussing the literal or non-literal quality of some of the most
frequent technical terms in winespeak, there were marked differences of opinion.

A good example of this was the case of the adjectives hard, long or round; all are
standard in wine commentary, to the extent that they would not be normally re-
garded as metaphorical by a professional wine critic or many of his/her readers for
that matter. Yet, for someone outside the wine community, it is difficult to see how
wines can be thus qualified, unless the terms are regarded as used in a non-literal,
figurative way, however commonplace they may be in the wine realm.
Indeed, these initial problems with metaphor identification reflect two highly-
debated issues in metaphor research (see also Gibbs, this volume, in this respect).
On the one hand, they illustrate the weight of the creativeness-conventionality
opposition in determining whether a use of language is metaphorical or not.
Thus, whereas for some scholars creative, metaphorical uses of language are those
characterized by their high degree of semantic and/or pragmatic incongruity with
regard to certain conventions (Kittay, 1987; Partington, 1998), incongruity is not
an issue for cognitive researchers, given the postulate that metaphorical language
comprises a large amount of the conventional language used in everyday com-
munication, professional or otherwise. Indeed, professional jargon is one of those
cases where creativity (or metaphoricity) is not at odds with conventionality, as
shown by research on metaphor in professional or specific discourses (Henderson,
1998; Boers, 1999; Pickering, 1999; Charteris-Black, 2000; Flannery, 2001; Larson
& Johnson, 2004; Musolff, 2004; Caballero, 2006). The fact that certain linguistic
expressions may be regarded as metaphorical by some communities yet be seen
as non-metaphorical by others is also addressed in Cameron (2003) and Steen
(2007), both of whom point to the weight of context and social convention in
metaphor awareness and/or identification.
Over and above relating to the creative-versus-conventional argument, our
insider-versus-outsider approach to metaphors in winespeak reflects two differ-
ent trends in metaphor studies, namely what Low (1999) refers to as user-centred
approaches and analyst-centred ones – the former taking into account the users’
point of view, that is, whether readers/users regard their language as metaphorical
or not, and the latter describing a situation where the researcher decides unilater-
ally what is and what is not metaphorical. In our case, the combination of both
approaches proved most beneficial. Once the initial reticence had been overcome,
the close acquaintance of one researcher with wine (the target domain for the
metaphors) and the conventions of the wine community in general and TNs in
particular was crucial for spotting the aspects of wine being discussed; the ex-
pertise on metaphor of the other researcher helped us notice expressions which
would have been otherwise overlooked and to organize them according to the
source domains involved in the underlying metaphors.
In the research summarised here a linguistic expression was considered meta-

phorical if it involved the understanding of and/or reference to wine or any of its
attributes or elements, but belonged to an experiential domain other than wine.
The degree of innovation involved was irrelevant. By selecting these criteria, we
in effect adopted the viewpoint of the interested, but non-expert reader. In order
to operationalise the criteria, we took into account linguistic form (which often
involved incongruous collocation, as in, for instance, a wine “bursting” from the
glass), yet also considered the potential of the expressions to activate a cross-do-
main mapping at a conceptual level (e.g. a wine qualified as “tightly-knit”, “broad-
shouldered” or “jarring”). In this regard, we followed the Lakoff & Johnson
approach to metaphor as a cross-domain transfer, since the reduced transfer ap-
proach proposed by Wallington (this volume), although appealing in many ways,
needs further empirical research before it can be applied to specialised datasets.
Nevertheless, the fact that an expression may be accounted for via a cross-domain
mapping does not imply any claim regarding the actual activation of such map-
pings each time wine critics use figurative language. Similarly, it does not mean
that they are consciously aware that knowledge is transferred from one domain to
another whenever a wine is assessed as “a little soft and flat in the middle”.
4. Classifying the figurative language in the TNs
Beyond the mere identification of metaphoric expressions lies the task of classify-
ing them according to coherent parameters that may illuminate their conceptual
underpinning and aspects of their use. Unsurprisingly, it is in this phase where
we encountered the greatest obstacles. The first major methodological distinc-
tion we needed to make was between conceptual and synaesthetic metaphors. A
caveat is in order at this point. We start from the well-known definition of con-
ceptual metaphors as mapping knowledge from a usually concrete or structured
domain or world entity onto a more abstract or less structured domain or entity
in order to understand it (Lakoff & Johnson, 1980, 1999; Lakoff & Turner, 1989);
synaesthetic metaphor maps information across different sensory domains and, in
this regard, is more specific or ‘restricted’ since the domains involved are the five
human senses – in contrast to the wide range of domains that may be involved in
conceptual metaphors. Of course, this does not mean that synesthetic metaphors
do not have a conceptual basis (Yu, 2003), yet they cannot be treated in the same
way as conceptual metaphors. Indeed, the distinction is particularly relevant in
the case of wine since, while conceptual metaphors help describe this beverage in
all its complexity (alluding to its structural properties, kinship relationships with
other wines, etc.), synaesthetic metaphors help convey the sensory experiences
afforded by wine. In other words, the properties of wine articulated by conceptual

metaphors are put to the test and described/assessed via smell, taste and touch,
which makes both types of metaphor relevant when looking into the metaphors
in winespeak – an issue discussed in more detail below.
We accordingly established a classification of metaphors taking into account
their different (a) source domains (e.g. architecture, physiology, anatomy, or tex-
tiles), and (b) source senses or modalities (e.g. sight, touch or sound). However,
it was immediately clear that some of their instantiations could exemplify both
conceptual and synaesthetic metaphors, like, for instance, satiny, which could
relate to either (or both) a textile metaphor or a touch metaphor. Another prob-
lem arose when figurative language drawing upon diverse domains co-existed
in the same TN and, worse still, in the same phrase and/or sentence (e.g. “Give
this puppy a few years to knit together” discussed later). Apart from the interpre-
tation problems involved in these cases (addressed in the final section below),
the expressions were difficult to classify according to the metaphor apparently
motivating them.
4.1 Experiential domains in winespeak
In wine assessment, abstract properties are very often articulated in terms of

concrete, or at least more familiar, entities from diverse experiential domains.
Thus, among the figurative repertoire of wine critics, wines are frequently por-
trayed as diverse living organisms (plants, animals or human beings), manufac-
tured entities (cloth, musical pieces, or buildings), and three-dimensional, geo-
metrical bodies.
One conspicuous set of expressions draws upon the domain of physiology,
which is far from surprising since wine is a mutable entity resulting from an or-
ganic process: the juice extracted from grapes changes considerably across its life
inside both oak casks and bottles (a process referred to as breeding or ageing), as
well as during the process of opening it, letting it breathe as it sits in the glass, and
drinking it. Accordingly, wines are often described as living organisms that can
be born, flesh out, get rid of [their] puppy fat, mature and age, grow tired, thin out,
and, finally, die. The following TN is a good illustration:
(6) Lovely toasty notes on the nose here, a nice mineral and smoky maturity,
and still very good, pure fruit. […]. Fine acid balance, and really no sign of
decay or tiredness despite not being from a top vintage.
Another set of expressions draws upon the domain of anatomy and focuses on
the structural properties of wines. Among the terms instantiating this human or
animal schema we find big-bodied, robust, fleshy, backbone, sinewy, long-limbed,

fat, flabby, broad-shouldered, lean, or disjointed. Both Examples (7) and (8) reflect
the anatomical schema:
(7) The big, bold, and sultry 2001 Chevalier-Montrachet is a highly expressive,
fleshy, supple wine […] This white chocolate, truffle, cream, and spice-
scented beauty is medium-bodied, concentrated, and sexy.
(8) [This wine] has a nicely buried backbone of acidity and tannin. […] ultimate
impression of both muscle and flesh.
This anatomical conceptualisation of wines is at times applicable to either ani-

mals or humans, but at others is clearly human, to the point where, as with the
“streetwalker” in Example (2), we may talk about personification. The bound-
ary is not always clear-cut; sexy, boisterous, assertive, sensitive, demure, shy, or
expressive would seem primarily to mark human personality traits, but other
attribute terms like pretty, handsome, beautiful and possibly even curvaceous
can be variably applied.
Other figurative expressions suggest a view of wines as three-dimensional ar-
tefacts. These may involve two kinds of entity. On the one hand, wines may be ren-
dered as geometrical bodies which have edges, layers, contours, backs and fronts,
or are described as square, angular, well-delineated, wide, threadlike, long, pointed,
deep or round. On the other hand, we also find wines portrayed in architectural
terms, as suggested by reference to some of their constitutive elements (e.g. acid,
alcohol and tannins) as their building blocks, to wines themselves as ‘edifices’ built,
buttressed, backed up or fortified by all or some of those constituents, or by their
qualification as monumental, massive, monolithic, or foursquare. This can be seen
from the following five passages:
(9) [The wine is] a big, blockish, masculine style, a bit angular and still quite oaky.
(10) This is a round, generous Shiraz that’s packed with layers of flavour […].
(11) Structured and built for the long haul, the densely packed aromatics hint at
the power within.
(12) A magnificent edifice of a wine, elegant and refined in structure but dripping
with flavour […].
(13) [This wine’s] future is all but ensured, with its tight core, solid wall of tannins
and a deep, black cherry finish that lasts a good two minutes.
A final conspicuous schema instantiated in TNs draws upon the domain of

cloth-making. Textile terms are mostly used either (a) to describe the structural
roperties of wines (which have a fabric or weft, may burst at their seams, and can
p
be variously described as a tapestry, open-knit, well meshed, or tightly wound) or
(b) to evaluate their feel in the taster’s mouth by means of adjectives like silky, vel-
vety, satiny, or pillowy. Other textile terms refer to wines’ elements as their cloak,
glove, frock, or mantle, which may well again suggest a personified view of wines.
The following examples illustrate this schema:
(14) A monster in a beautiful frock. […] Full-bodied, with loads of velvety
tannins and a long, long dried cherry finish.
(15) Intricate and dry, this single-vineyard beauty feels easy and silky, but there’s an
iron fist in that velvet glove.
(16) An extraordinary wine, and not just because it wears its 16.4% alcohol so well.
Of course, this activation of different metaphors by a single linguistic expression

is far from unusual, as pointed out by metaphor scholars dealing with diverse
types of discourse (see Goatly, 1987; Forceville, 1996; Caballero, 2006), yet it
poses classification problems that need to be solved. One implication is that in
identifying the metaphors that motivate figurative uses of language, consider-
ation of the semantic and syntactic aspects of those uses is important, yet knowl-
edge of other sorts is needed as well (Steen, 2007). In the first place, as pointed
out earlier, familiarity with the community whose metaphors are under focus
may help us understand the figurative construal of their world. In turn, impor-
tant information may also be derived from the dynamics of the texts where such
metaphors occur.5
Nevertheless, contextual factors can also be tricky in metaphor classifica-
tion. Indeed, as shown in the examples seen so far, the TNs often combine lexis
drawing from diverse domains, and this diversification may lead to confusion
when the author does not stick to his/her original source domain and moves on
to a totally different one, particularly within the same sentence. This is the case
with passage (17):
(17) This dramatic wine is dark, young and closed. It opens with a flourish of
black cherry and currant aromas conjoined with the smoky vanillins of
oak, but it hasn’t all come together. […] Give this puppy a few years to knit
together.
In this example the reader misses the consistency of Example (6) discussed ear-
lier, where the physiological imagery of maturity was carried forward logically
to decay or tiredness. In (17) we find youth expressed via physiology (“puppy”)
and maturity via textiles (“knit together”) in the same sentence. While the overall
message remains perfectly understandable, it is nonetheless a sign of the difficul-

ties deriving from the wine critics’ lack of systematization in language use – an
issue pointed out by research on wine terminology, yet still waiting to be success-
fully solved (see Nedlinko (2006) for a discussion of some of the terminological
issues in the wine domain).
The problematic nature of some other instantiations derives from the fact that
they may ‘belong’ to different domains. A case in point is lexis drawing upon
anatomy and architecture – the best exponents being structure and balance (see
also Lehrer, 2009). Contextual factors may help the analyst resolve the issue, but
when this is not the case, the ambiguity remains. The following three passages
illustrate the point:
(18) On the palate the wine is clean and fruity, but underneath this rather simple
facade hides some tannic structure and complexity.
(19) Full-bodied, with admirable structure, weight, and volume, [this wine] will be
even better in another 1–2 years, and last for 10–12.
(20) [This wine] is ambitious […] with much sweet berry fruit, light vanilla and
cloves on the nose, significant density and structure on the palate […].
In Examples (18) and (19) the ambiguity of structure is somewhat pinned down
by the architectural and anatomical imagery co-occurring with the term (fa-
cade and full-bodied respectively); however, there is no hint towards either in
Example (20) and, therefore, it is impossible to isolate a single source domain
with any certainty. A plausible explanation for the difficulties in discriminat-
ing architecturally-biased language from anatomically-biased language is the
frequent use of lexis drawing from anatomy in architectural discourse and vice
versa. Consider, for instance, terms such as skeleton or spine referring to the
basic supporting system in a building, or the use of well-built to qualify a big,
hefty person. Whatever the reason, whenever the context provided no clues as
to the metaphor being instantiated, the terms were tagged as illustrating both
anatomy and architecture metaphors.
Things got even more difficult when the traits discussed involved concrete
rather than abstract aspects of wine.
4.2 Conceptual metaphor, synaesthesia, or both?
Problems appeared when the metaphors did not bridge the abstract-concrete gap,
but connected similar degrees of concretion. For although helping our under-
standing of the most abstract via the most concrete is one of the most salient
properties of metaphor, this does not rule out the potential for both the source
and target to be fairly concrete in certain metaphors. Another conspicuous trait
of metaphor is that it can involve various modalities or senses (e.g. vision, hear-
ing, touch, or taste). An extreme example of multiple modalities is synaesthetic
metaphor, which maps information across sensory domains or modes. For in-
stance, acidity of wines can be described as searing, screeching or scintillating (i.e.
by means of adjectives prototypically related to touch, sound and sight).
Indeed, as has been pointed out, wine tasting is inherently cross-modal in the
sense that while the nose complements the palate in the perception and identi-
fication of aromas, the tongue perceives much more than just flavours (namely,
extremely detailed tactile impressions). It is only through this comprehensive sen-
sory experience that we can access the intricacies of a given wine. In this regard,
synaesthesia and synaesthetic metaphors are crucial both in wine appreciation
and in the verbal communication of this experience. However, synaesthetic meta-
phors can be problematic research-wise (particularly, in metaphor classification)
because in many cases they incorporate a particular entity, the experience of in-
teracting with it via a given sense, and the sense itself. A case in point are exam-
ples using lexis from the textile domain, particularly terms such as silky, velvety,
satiny, or pillowy when used to evaluate a wine’s feel in the taster’s mouth. It is this
last aspect, the mouth, that points to the close relationship between synaesthesia
and some figurative language in winespeak. For ‘silkiness’ and the like can only be
fully appreciated via the sense of touch against the skin which, in turn, is used to
articulate the properties of wines as perceived through its feel in the mouth. The
difference between both tactile experiences is that whereas feeling with the hand
or against the cheek usually involves the nerve endings in our skin, ‘feeling’ in the
mouth is felt through our tastebuds. In short, although the metaphor involves a
touch-to-touch mapping, the different nature of the source and target touch ex-
periences makes it worthy of being seen as a case of synaesthesia – even if it can
also illustrate a textile metaphor.
Another difficulty arises because most synaesthetic instantiations can be
arranged in a gradation or cline, which is congruent with critics’ assessment of
wines’ attributes as involving different degrees, amounts, or measurements of pa-
rameters (none of which are categorical). This is particularly well illustrated by
lexis drawing upon the sense of sight, which is often used in the TNs to assess
a critical aspect of both the structure and mouthfeel of wine, namely, its acidity.
The various expressions instantiating the acidity is light metaphor easily lend
themselves to arrangement in a cline. This is schematized in Figure 1.
Figure 1. ACIDITY IS LIGHT
ACIDITY IS LIGHT: WHITE WINES
INTENSITY
lightning bolt blaze

+ spark laserlike
beam
bright
brightness
brighten
radiant/radiance
vivid
shine
light up
enlighten
glistening
luminous
shimmering
scintillating
INTENSITY
fade
+
PERSISTENCE
PERSISTENCE
Figure 1. acidity is light
Given the idiosyncrasy of wine tasting, the terms have been arranged according to
the parameters of ‘force’ (or intensity) and ‘duration’ since both are salient features
that need neither imply nor exclude each other. In fact, this reflects the many ways
in which acidity may be detected in the taster’s mouth: from a sudden impact that
disappears briskly to a pleasant lingering sensation. Because with age acidity is
integrated and harmonized, the many stages in a wine’s life will need expressions
to communicate this. Again, the terms allow a certain degree of subjectivity or
inconsistency (native speakers, especially poets and physicists, might argue end-
lessly over potential modifications to our cline). These are clearly not pH indices
and it is unrealistic to expect scientific precision, but they are in fact more useful
in context than numbers because what they express is perceived relative acidity
(i.e. measured against other parameters). That speaks to the reader about organo-
leptic balance, not just chemistry, and provides much more relevant information
than absolute acidity in grams per litre. The following extracts exemplify the use
of this schema in three whites and one red:
(21) This really lights up the palate. The flavors are deep and intense, set against
an ironlike structure that blazes with minerality on the finish.
(22) A lightning bolt of acidity jolts the palate […]
(23) A laser beam of honey, apricot and mineral accents, […]

(24) The vintage’s bright fruit shines through with stunning clarity and purity.
These are illustrative examples of the role of acidity in wine. Acidity (described
in terms of light) balances the other elements and helps them stand out (i.e.
highlights them, makes them “shine”) because the correct focus (or delineation)
is needed to appreciate them in all their splendour. Without the proper amount of
acidity, the wine would feel dominated by alcohol and/or tannin, and feel heavy
and clumsy (i.e. unfocused, blurred). In this sense, expressions incorporating the
term ‘laser’ (laser-like, laser-sharp, laser beam, laser focus) epitomize perfect acid-
ity: ‘laser’ does not merely express concentration; it also incarnates the ideal of
light with zero dispersion and, as such, provides perfect definition (i.e. focus) to a
wine. This is the explicit point of the following description of a white wine: “The
difference between this and the typical [wine from the same area] is intensity; the
fruit is more focused. It’s the difference, say, between a flashlight and a laser.”
The complex and multidimensional nature of wine metaphors made it nec-
essary to adopt a corresponding multiplicity of parameters in order to classify
them.6 Our tentative, working proposal for classifying the figurative language in
winespeak involves the following three basic dimensions:
4.2.1 The amount of information conveyed by the figurative expression

This dimension is closely related to the structuring potential of the underlying
metaphor, that is, the projection of partial or complete structure from source onto
target. The greater or lesser structural potential of a metaphor is determined by
the aspects of the mapping that may be activated in an expression in its context
of occurrence. For instance, examples like “The already assembled, luxury cuvée
is […]”, or “The wine is a fruit-filled, seamlessly constructed effort” all point to the
metaphor winemaking is building. This would recruit rich information from
several elements in the logic of the source (building) and subsume more specific
metaphors such as wines are buildings or elements in a wine are building
elements (as suggested by reference to acid, alcohol and tannins as the building
blocks of wines, or qualification of wines as monolithic or foursquare). In contrast,
metaphors such as wines are three-dimensional artefacts (instantiated in
qualification of wines as angular, round, or hollow) would involve the mapping
of a smaller set of traits, i.e. recruit poorer knowledge from their metaphorical
sources. We found that metaphors with a high structuring potential helped critics
focus in their TNs on a greater number of aspects related to wine (from winemak-
ing processes to particular attributes of wines) than did metaphors characterized
by a lower degree of structuring potential, which only highlighted a few aspects of
wine (prototypically, its texture).
4.2.2 The perceptual quality of the figurative expression, that is, its ability
to convey sensory experiences related to smell, taste and touch
Thus, whereas an expression such as “like its older siblings this wine will be deli-
cious” is unconcerned with the wines’ sensory properties (focusing, rather, on
its ‘kinship’ relationship with other wines) and was regarded as an example of
conceptual metaphor, expressions like “roaring nose”, “scintillating acidity” or
“supple and caressing finish” draw upon typical lexis from perception via hearing,
sight and touch to qualify wine’s properties. The latter cases illustrate synaesthetic
metaphor and were classified according to the sense involved in them (e.g. sight,
touch, sound metaphors).
4.2.3 The conventional quality of the figurative expressions

Adjectives big and full-bodied are conventional jargon terms applied to wines rich
in extract, alcohol and glycerine (and, therefore, providing a feeling of weight
in the mouth), whereas adjectives muscular or broad-shouldered are more inno-
vative – not to mention the qualifications of wines full of sexual connotations
through terms such as harlot, whore or street-walker.
Finally, it should be noted that, since metaphorical cases may illustrate diverse
degrees of information richness (plus structuring potential), perceptual quality, and
conventionality, each dimension needs to be treated in terms of a cline rather than
in absolute terms. Moreover, as in Cameron (1999b), although the dimensions were
initially used one by one, they were not regarded as mutually exclusive when trying
to understand and explain our figurative data, since many cases illustrated more
than one dimension; for instance, a term like tightly-knit would be classified as illus-
trating a conceptual metaphor focusing on product even if it also suggests the pro-
cess to achieve this quality of wine, and is a fairly conventional term in winespeak.7
5. Interpreting the figurative language in TNs
Although the use of figurative language and the application of the clines are rela-
tively consistent in professional writers (as far as our 12,000 text corpus can be
considered representative), there are still many grey areas where three or four
instantiations are virtually synonymous and the only way to explain their coex-
istence is lexical variation. This is precisely what happens with textile metaphors
realised interchangeably by velvety, silky and satiny, where what seems to be get-
ting mapped is merely the soft texture and the luxury connotation, but not other,
rather different, aspects such as the potentially important temperature distinction
between silky and velvety. Indeed, the fresher feel of silk versus the warm feel of
velvet could be easily mapped onto the relative temperatures of white and red
wines. However, these terms are often used interchangeably and hence are ap-
plicable to red and white wines. That is why silky is not exclusive to white wines,
nor velvety to reds, as can be observed in Examples (25) and (26), describing a red
and a white wine respectively:
(25) Most tasters […] praised its massive cherry and cocoa flavors that flirt with
over ripeness, and the smooth, silky texture. […] Likely to soften and knit
together with a few hours of decanting […]. (Red wine)
(26) Medium-bodied and velvety-textured, this wine has loads of depth, concentra-
tion, and fruit. (White wine)
There is also a fun and pleasure aspect to the metaphors in the TNs (in line with
the claims in Ritchie & Dyhouse, 2008), though the fun may well be connected to
attempts to persuade readers to actually part with their money. Commercialism
notwithstanding, many of these metaphorical expressions reflect the creative efforts
of specific writers. After describing several hundred similar wines, or perhaps en-
raptured by particularly sublime examples, the critic grows understandably tired of
using the same language, and introduces here and there some innovation that he
or she deems fit. It is wine that is being described, not nails, and TNs, after all, are
literature. People read them, often with an ear for the poetic potential of winespeak,
and critics are aware of this aspect as an area where they may introduce an artsy
difference from their competitors. In fact, a brief look at the numerous wine blogs
and forums in the Internet illustrates that innovation in wine language is constant.
Sometimes innovation is a synthetic attempt to convey more meaning using fewer
words, a joke to poke fun at a poor wine, or just to sound witty for its own sake.
Whatever the intention, many TNs reveal an author and a style, and pride of au-
thorship occasionally intrudes and hinders effective communication. That is why
every now and then we have found expressions where there can be little doubt that
we are facing figurative language and yet nothing meaningful emerges from the
analysis: we miss the point and suspect few, if any, readers would get it A failure in
communication alone is little more than that for the reader; often the expression
appears downright nonsensical and that is all. But for researchers sometimes those
expressions can expose major inconsistencies within the system. For example, by
extrapolating wines are human beings to hyperbolic lengths, some critics have
produced expressions like “contemplative” that simply do not seem to correspond
to any objectively identifiable trait of a wine:
(27) Sweet, expressive, soft porty nose. The palate is ripe and sweetly fruited.
Distinctive wine that is expressive and contemplative.
(28) A real fruit bomb, loads of peppery fruit, rich mocha notes, fragrant and
expansive. Wonderfully lucid – combines explosive fruit with firm new oak.
Again, character is one of the most abused and ambiguous terms in this regard.
Thus, whereas in Example (29), “character” genuinely endows the wine with an
individual personality, in (30) “has a […] character” could easily be replaced
by the terser “is” and in (31) the word is almost completely empty of meaning,
implying simply that there is a predominance of certain fruits in the wine’s aro-
matic spectrum:
(29) An inky red/purple colour, this is a very individual wine. […] Very
concentrated, this is quite an unusual wine with real character.
(30) This medium-bodied wine has a satiny-textured, refined character that reveals
red currants, cherries, lily, and apricot flavors.
(31) Full-bodied, it’s so harmonious, with plum, red berry and blackberry charac-
ter, a touch of mocha and grilled nuts from the oak.
A final necessary distinction involves the categorization of descriptive versus

evaluative expressions. These often overlap and create further grey areas, where
culture intrudes and complicates matters unnecessarily. This is the case with terms
such as masculine/feminine, which, are merely descriptive despite the expectations
that readers’ cultural backgrounds may generate. Because wines are described and
evaluated within a peer context (i.e. in terms of geographical proximity, grape
variety, style, and vintage), a wine may or may not comply with expectations de-
riving from what is typical of its peers. A perfect example is Pinot Noir. To speak
about a “feminine Pinot” in absolute terms would be a pleonasm or a marker of
emphasis, adding little more information than the expression “a textbook Pinot.”
However, describing a Burgundian Pinot Noir as “feminine” is a different matter.
Because of Burgundy’s northerly location and its relatively cold weather, Pinot
produces a fragrant, delicately floral, low-alcohol, high-acid, light coloured wine.
Against a global context of fairly powerful red wines, one could thus say in a rela-
tive way that Burgundian Pinot Noir is a feminine wine. Within the confines of
Burgundy, though, many further subtle distinctions can be made and, as a result,
some communes or sub-zones produce wines that are fuller bodied than what is
considered average in the context of the entire appellation. Therefore it is gener-
ally accepted that Pommard is a masculine red Burgundy. But against a Cabernet-
based Bordeaux, the same Pommard would be a feminine red.
On a global scale, however, Pinot Noir is grown virtually everywhere. Thus, if
we compare the delicate French examples with their far more robust, alcoholic and
glyceric counterparts from California, the only possible conclusion is that most of
the Pinot-based wines emerging from much warmer California are not describ-
able as feminine in the Pinot paradigm as this has been classically defined by Bur-
gundy. A traditionalist would claim that these American examples are oversized
caricatures of “true” Pinot; a modernist might deem Burgundy thin and acidic
(rather than just feminine). A ‘balanced’ connoisseur should enjoy Burgundy for
its fragrance and complexity, and when faced with the Californian Pinot derive
pleasure from both its upfront fruit and its non-canonical appeal.
In sum, there is no evaluation inherent to the masculine/feminine distinction:
there is nothing per se good or bad in being one or the other; while it may be true
that the current market trends favour power over elegance (i.e. masculine over
feminine), that is essentially transitory (i.e. the market may eventually change as
it has done in the past). The terms simply mean that the drinker’s natural expec-
tations from a specific wine will not be met, because the wine is unusual for its
group in terms of mouthfeel and structure. And as pointed out earlier, this is rela-
tive. Either quality could make the wine in question more apt for a specific meal,
easier to drink at a young age, or the other way round. The important thing is to
be aware of such exceptionality and act accordingly.
6. Conclusions
In this paper we have attempted to outline some of the issues derived from the
exploration of the metaphors motivating a large amount of the language used to
discuss wine. We started from the basic assumption that any use of language –
metaphorical or otherwise – responds to contextual factors and, therefore, should
be approached accordingly. Thus, we narrowed our analysis to a particular – and
popular – genre within wine discourse, namely the Tasting Note, and used a
12,000-text corpus representative of the critical practices in the wine realm. Fi-
nally, our different disciplinary and background knowledge (one of us is a meta-
phor scholar and the other is a wine expert and critic himself) was regarded as an
asset that might prove advantageous for overcoming the – expected – problems
derived from researching metaphor in such a specialised discourse.
Indeed, most problems arose from the very experience that figurative lan-
guage helps articulate (tasting wine), as well as from the shortage of terms avail-
able in English for providing a precise account of that experience. Metaphor over-
comes these difficulties by providing wine critics with a set of conceptual frames
and the corresponding lexis to construe and share the difficult task of describing
what wines smell, taste and feel like. Moreover, given the complexity of param-
eters involved in wine assessment, the convergence of language motivated by di-
verse figurative schemas is far from surprising.
However, regardless of its heuristic role in winespeak, metaphor is not exempt
from problems, particularly those concerning the identification, classification,
and interpretation of its verbal manifestation in textual contexts. A number of
problems derived from (a) the close relationship between the source domains in
some metaphors (e.g. architecture and anatomy), (b) the co-evocation of various
metaphors by a single expression, and (c) the fuzzy boundaries between concep-
tual and synaesthetic metaphor (as these types have been defined in the cognitive
literature) in wine discourse. Some other problems were more rhetorically biased,
and derived from the critics’ lack of systematicity in creatively expanding such
entrenched schemas in the wine realm as wines are human beings or wines
are textiles, due to personal preferences or, simply, verbosity.
In short, the joint – expert plus non-expert – venture into winespeak meta-
phors here summarised proved most useful in some respects, and brought to the
foreground some of the issues involved in looking into metaphor in real life and
discourse. Thus, it confirmed the claim that metaphoricity (including the cre-
ative-conventional opposition among other things) is relative rather than absolute
and needs to be addressed from a situated perspective, that is, taking into account
the community using the metaphors at issue (as also pointed out by Gibbs, this
volume). Moreover, our combined approach helped determine the target(s) of the
various metaphors underpinning wine commentary as well as the diverse modes
involved in some such metaphors (i.e. synaesthetic metaphor). The problems de-
rived from identifying and classifying some of the metaphors in our corpus not
only underlined the complexities involved in using metaphorical language, but
also the concomitant challenges presented to those researchers attempting to un-
derstand and explain that use. Metaphors’ sensitivity to the context(s) where they
are used plays a pivotal role in this respect, which suggests that knowledge of the
idiosyncrasy of the topic at issue (wine, i.e. the target domain in the metaphors),
the community under scrutiny (wine lovers and experts), and the discourse con-
text chosen (the TN) is needed in order to gain insight into why, how and when
metaphors are used in real communication.
Notes
1. The figurative expressions under discussion are italicized here and in later examples.
2. See, for instance Goode (2008: 149) where we can find “So we have wine as a living creature;
wine as a piece of cloth; wine as a building; even, in a recent note by Robert Parker, wine as a
whore. It is easy to make fun of this sort of description, but such metaphors are born of neces-
sity. While we would like to have a more exact way of sharing our experience of wine in words,
such precision does not exist, and those who restrict themselves merely to naming aromas and
flavors end up missing out on some of the more important aspects of the character of wines that
cannot be described in this way, such as texture, structure, balance, and elegance.”
3. As it is, there are publications that consist exclusively of TNs. This is the case of The Wine
Advocate, for example, which is one of the most authoritative publications in the field.
4. This is a recurrent topic in the numerous online wine bulletin boards like, for instance,
Mark Squires’s Bulletin Board on eRobertParker.com, which featured 13,862 members and
162,168 different discussion threads in November 2008.
5. In this regard, we agree with Cameron (2008, this volume) when she argues for the im-
portance of context based on the shifting dynamics of oral text. Thus, although our texts are
very short and written, they also illustrate diverse degrees of discourse shifts, as discussed in
the final section.
6. A similar claim can be found in Cameron (1999b). In fact, two of the three dimensions
taken into account in our research point to some of the graded metaphor descriptors in her
framework. Thus, whereas dimension (a) brings to mind a combination of Cameron’s “Cogni-
tive demand of Topic and Vehicle terms and domains” plus “Connotative power of Vehicle term”,
dimension (c) is related to her “Novelty-Conventionality of Topic-Vehicle link” descriptor.
7. The use of clines in metaphor research has deserved some commentary by metaphor schol-
ars. Thus, in Steen (2007) we find it regarded as costly for those approaches aiming at reliability
or scientific status (i.e. quantitative approaches). However, although the use of clines may be
problematic in metaphor quantification, they may be usefully applied in qualitative approaches
like the one summarised in this chapter.
References
Amoraritei, Loredana (2002). La métaphore en Oenologie. Metaphorik.de, 3, 1–12. Online

document. Accessed at [www.metaphorik.de/03/amoraritei.htm]
Boers, Frank (1999). When a bodily source domain becomes prominent: The joy of counting
metaphors in the socio-economic domain. In R. W. Gibbs & G. J. Steen (Eds.), Metaphor in
cognitive linguistics (47–56). Amsterdam & Philadelphia: John Benjamins.
Caballero, Rosario (2006). Re-viewing space. Figurative language in architects’ assessment of built
space. Berlin & New York, Mouton de Gruyter.
Cameron, Lynne (1999a). Operationalising ‘metaphor’ for applied linguistic research. In
L. Cameron & G. D. Low (Eds), Researching and applying metaphor (3–28). Cambridge:
Cameron, Lynne (1999b). Identifying and describing metaphor in spoken discourse data. In
L. Cameron & G. D. Low (Eds), Researching and applying metaphor (105–132). Cambridge:
Charteris-Black, Jonathan (2000). Metaphor and vocabulary teaching in ESP economics. Eng-
lish for Specific Purposes, 19, 149–165.
Day, Sean (1996). Synaesthesia and synaesthetic metaphors. Psyche, 2 (32). Retrieved from
http://psyche.cs.monash.edu.au/v2/psyche-2-32-day.html.
Flannery, Michael (2001). The laboratory as library metaphor and some reflections on profes-
sional practice. Bulletin of the Medical Library Association, 89, 309–310.
Forceville, Charles (1996). Pictorial metaphor in advertising. London & New York: Routledge.
Gawel, Richard (1997). The use of language by trained and untrained wine tasters. Journal of
Sensory Studies, 12, 267–284.
Gawel, Richard & Anita Oberholster (2001). A mouth-feel wheel: Terminology for communicat-
ing the mouth-feel characteristics of red wine. Adelaide: Department of Horticulture, Viti-
culture and Oenology, University of Adelaide.
Gluck, Malcolm (2003). Wine language. Useful idiom or idiot-speak? In J. Aitchison &
D. M. Lewis (Eds), New media language (107–115). London & New York: Routledge.
Goatly, Andrew (1997). The language of metaphors. London & New York: Routledge.
Goode, Jamie (2008). Experiencing wine: Why critics mess up (some of the time). In F. Allhoff
(Ed.), Wine and philosophy (A symposium on thinking and drinking) (137–153). Oxford:
Blackwell.
Henderson, Willie (1998). Metaphor. In J. B. Davis, D. W. Hands & U. Mäki (Eds.), The hand-
book of economic methodology (289–294). Cheltenham: Edward Elgar.
Hughson, Angus & Robert Boakes (2001). Perceptual and cognitive aspects of wine tasting
expertise. Australian Journal of Psychology, 53, 103–108.
Kittay, Eva Feder (1987). Metaphor: Its cognitive force and linguistic structure. Oxford & New
York: Oxford University Press.
Larson, Steve & Mark Johnson (2004). Architectural metaphors in music discourse and music
experience. Yearbook of Comparative and General Literature, 50, 141–154.
Lawless, Harry (1984). Flavour description of white wine by expert and non-expert wine con-
sumers. Journal of Food Science, 49, 120–123.
Lehrer, Adrienne (1975). Talking about wine. Language, 51 (4), 901–923.
Lehrer, Adrienne (1983). Wine and conversation. Bloomington: Indiana University Press.
Lehrer, Adrienne (1992). Wine vocabulary and wine description. Verbatim: The Language
Quarterly, 18, 13–15.
Lehrer, Adrienne (2007). Can wines be brawny? Reflections on wine vocabulary. In B. Smith
(Ed.), Questions of taste. The philosophy of wine (41–78). New York: Oxford University
Press.
Lehrer, Adrienne (2009). Wine and Conversation, Second Edition. New York: Oxford Univer-
sity Press.
Low, Graham (1999). Validating metaphor research projects. In L. Cameron & G. D. Low (Eds),
Researching and applying metaphor (48–65). Cambridge: Cambridge University Press.
Musolff, Andreas (2004). Metaphor and political discourse: Analogical reasoning in debates about
Europe. Basingstoke: Palgrave Macmillan.
Nedlinko, Anatoli (2006). Viticulture and winemaking terminology and terminography. Ter-
minology, 12, 137–164.
Panther, Klaus-Uwe & Gunther Radden (Eds.) (1999). Metonymy in language and thought.
Panther, Klaus-Uwe & Linda Thornburg (Eds.) (2003). Pragmatic inferencing in metonymy.
Paradis, Carita (2004). Where does metonymy stop? Senses, facets, and active zones. Metaphor
and Symbol, 19, 245–264.
Partington, Alan (1998). Patterns and meanings. Using corpora for English language research and
teaching. Amsterdam & Philadelphia: John Benjamins.
Peynaud, Emile (1987). The taste of wine. London: Macdonald Orbis.
Pickering, Neil (1999). Metaphors and models in medicine. Theoretical Medicine and Bioethics,
20, 361–375.
Radden, Gunther & Zoltán Kövecses (1999). Towards a theory of metonymy. In K.-U. Panther
& G. Radden (Eds.), Metonymy in language and thought (17–59). Amsterdam & Philadel-
phia, John Benjamins.
Ramachandran, Vilayanur & Edward M. Hubbard (2001). Synaesthesia – A window into per-
ception, thought and language. Journal of Consciousness Studies, 8, 3–34.
Ritchie, David & Valrie Dyhouse (2008). Hair of the frog and other empty metaphors: The play
element in figurative language. Metaphor and Symbol, 23, 85–107.
Silverstein, Michael (2003). Indexical order and the dialectics of sociolinguistic life. Language
and Communication, 23, 193–229.
Silverstein, Michael (2004). “Cultural” concepts and the language-culture nexus. Current An-
thropology, 45, 621–652.
Solomon, Gregg E. A. (1997). Conceptual change and wine expertise. The Journal of the Learn-
ing Sciences, 6, 41–60.
Steen, Gerard J. (1999). From linguistic to conceptual metaphor in five steps. In R. W. Gibbs
& G. J. Steen (Eds.), Metaphor in cognitive linguistics (57–77). Amsterdam & Philadelphia:
John Benjamins.
Steen, Gerard J. (2002). Towards a procedure for metaphor identification. Language and Litera-
ture, 11, 17–33.
Steen, Gerard J. (2007). Finding metaphor in grammar and usage. Amsterdam & Philadelphia:
John Benjamins.
Swales, John (1990). Genre analysis: English in research and academic settings. Cambridge:
Swales, John (2004). Research genres: Exploration and applications. Cambridge: Cambridge
University Press.
Yu, Ning (2003). Synesthetic metaphor: A cognitive perspective. Journal of Literary Semantics,
32, 19–34.
section 3
The function of metaphor in discourse

chapter 14
Wot no similes?
The curious absence of simile in university lectures
Graham Low
University of York
Simile is known to be frequent in prose literature (Sayce, 1954; Goatly, 1997)

and Carter (2004) found many examples in conversation. Given the long-
traditional use of analogy by educationalists and the explicitness of similes,
facilitating access to the concepts involved, we might expect to find frequent
simile-type comparisons spread through educational discourse of all sorts,
especially in the genres that are less formal in style than, say, research reports
in academic journals. However, such appears not to be the case, to judge from
previous studies of written text (Low, 1997, 2008). The present paper focuses
on simile in spoken university lectures and builds on a study of metaphor in
three lectures (Low, Littlemore & Koester, 2008), by adding a fourth, to give two
conversational-style lectures and two more formal or rhetorical-style ones. The
findings are that there are almost no similes at all in the data and where they
do occur, it is only in the more conversational-style lectures. They rarely if ever
form part of long rhetorical sequences, where the writer or speaker is offering
an extended explanation. Rather they are one-off items serving immediate and
short-term rhetorical purposes, what Conversation Analysts call ‘local control’,
often associated with control of saliency and foregrounding at discourse level.
Keywords: simile, university lectures, genre
1. Introduction
It is hardly controversial to note that, to use Paprotté and Dirven’s (1985) term,
metaphor, like metonymy, is “ubiquitous” across topic areas, levels of language
and individual texts. Precise estimates of its frequency in various types of dis-
course have been produced by various researchers, though the figure is highly de-
pendent on whether stretches of discourse (in the form of metaphor ‘vehicles’) or
metaphorically-used words are the unit of identification. Cameron (2003: 86), for
292 Graham Low
example, marking vehicle terms in her primary school classroom data, reported
14–27 vehicles per 1000 words, while the oral and written texts across several
genres examined jointly by members of the Pragglejaz Group, using the proce-
dure described in the 2007 paper, regularly found that around 10–15% of words
(or more strictly lexical units), were metaphorically used.
Similes are, however, a different matter. There is, unsurprisingly perhaps,
evidence of their use in literary texts (e.g. Goatly, 1997), in conversation (e.g.
Carter, 2004) and in primary classroom discourse (e.g. Cameron, 2003, where
the teacher used them to introduce and explain difficult concepts). However,
the starting point for this paper is the virtual absence of similes in two corpora
of relatively informal academic discourse: the first comprising four book re-
views and three editorials in academic journals, plus two journalistic reports
of scientific/technological advances (Low, 1997: c. 8,300 words): the second
comprising twenty book reviews in high status academic journals (Low, 2008:
c. 18,900 words). The entire 27,000 words included just one clear (and another
possible) metaphoric simile.
If similes are rare in this sort of written text, one might hypothesise that they
would be rather more common in university lectures, which are oral, given face to
face and generally, one might suppose, aimed at helping students to conceptualise
and understand ideas. The purpose of this paper is to build on a short exploratory
study of metaphor in three UK university lectures (Low, Littlemore & Koester,
2008) and extend the dataset slightly by adding a fourth.
The study can be characterised as real-world research in that it addresses a
relatively common social context which can be problematic. The key problem is
that, despite there being numerous listeners and speech that is designed to facili-
tate the listeners’ learning, active negotiation of meaning can be hard to achieve,
either at points of incomprehension, or indeed ever, depending on the structure
of the lecture. The research is also real-world as it uses language that was actually
spoken in lectures, with the result that the data is frequently “messy” (as Gibbs,
this volume, Section 4.1, puts it) and identification is at times far from obvi-
ous. Finally, in line with the suggestions that Gibbs (again this volume) makes
for real-world research, an attempt is made to provide quantitative support for
qualitative analyses.
2. Simile and metaphorical simile
A simile is a comparison involving two entities which (a) focuses on similarity

rather than difference and (b) is explicitly flagged by a marker. The marker and
indeed the comparison itself may be visual or possibly even aural (Teng & Sun,
Chapter 14. Wot no similes? 293
2003), but in this chapter the focus is on linguistically expressed similes, so the
marker will be a word such as like, or a phrase such as as if. An example, using
looked like is Monica Ali’s,
The sun was shining and it was raining…. It looked like someone had thrown a
stripy sock across the sky Alentejo Blue, 2006: 203
Carter (2004) found a range of about a dozen such linguistic markers in his con-
versation data (see also Darian, 1973). The two entities compared are usually
present in the utterance (or in consecutive utterances, as in the Alentejo Blue ex-
ample), and the basis, or ground1, for the comparison (the stripy above) may or
may not be mentioned.
The comparison may or may not also be metaphoric. In John physically resem-
bles Fred it is not, but in the stripy sock example it is; a stripy sock and the colour
of the sky are very different domains. The interest here lies in metaphoric, rather
than non-metaphoric similes.
3. Identifying similes in discourse
Similes are not quite as easy to identify as is sometimes thought. Firstly, several
marker expressions, including the prototypical like, can have different meanings,
so it is not always clear whether (or how far) a comparison is intended. Thus
the expression It feels like this could describe an example stimulus which causes
physical pain (i.e. ‘This is what I felt’), or it could be a intended to suggest a com-
parison. Similarly, kind of can mark, amongst other things, subcategorisation or
exemplification (a lion is a kind of animal), partial identification (He’s kind of cute),
hesitation to do or admit something (I’m just kind of, stuck here! Text PS087 BNC,
concerning a card game) or resemblance (it’s kind of a lion, or Over by the door
of the church I could see Meriel Viney, wearing a kind of white sack and a pair of
what looked like tennis shoes Text HR9, BNC). For kind of, the syntax can to some
degree be used to help clarify the sense involved (a kind of X versus kind of Y).
Because the aim is to identify a metaphor within the comparison, identifying
metaphoric similes inherits many of the problems related to identifying metaphor
in general. Just as it is often unclear whether to treat two domains as separate or
discrepant, so it will often be unclear whether a simile really should be treated
as metaphoric.2 A humorous example of where the like could flag a metaphor, a
literal comparison or both at the same time, comes in the sardonic exchange be-
tween bride-to-be Katie and the jeweller, in Mark Haddon’s (2007) novel A Spot of
Bother, about the need for his and hers rings (i.e. similar items for both the man
and the woman) – where we find the following (p. 65):
294 Graham Low
The jeweller asked if they wanted inscriptions and Katie was temporarily foxed.
Did wedding rings have inscriptions?
‘On the inside, usually,’ said the man. ‘The date of the wedding. Or perhaps
some kind of endearment.’ …
‘Or a return address,’ said Katie. ‘Like on a dog.’
In this case, we might reasonably postulate an underlying implied metaphoric

proposition A HUSBAND IS A DOG and the accompanying implication of the
need to control his errant behaviour. However, if the simile had come earlier in
the text (Did wedding rings have inscriptions, like dog collars?), there would be
the same discrepancy of subject matter, but far less certainty about the need to
construct a metaphor.
Hanks (2007), in an analysis of P. G. Wodehouse’s (1917/8) Piccadilly Jim,
further suggests that we should extend the concept of simile to cover cases where
there is no explicit marker of comparison at all, but a comparison can nevertheless
be inferred. He cites examples such as millionaires … with gimlet eyes and sharp
tongues (Chapter 12) or a man in a grey uniform was beginning a Salome dance
(Chapter 2). This would generate a new subcategory of ‘implicit simile’, but even
here, particularly in the Salome example, it remains arguable whether categorisa-
tion (‘of the sort danced by Salome’) might be involved, rather than (or in addition
to) comparison. For the purposes of this study, I allowed a broad range of mark-
ers, such as apparently or seemingly, but drew the line at no marker. However, the
multiple potential senses of even markers such as like mean that the analyst needs
to make a context-based inference for each simile identification; few if any similes
are 100% formally explicit.
If one has chosen to use the so-called MIP procedure described by Steen (2004)
and the Pragglejaz Group (2007), or the later MIPVU version (Steen, this volume),
based on identifying metaphorically-used words in discourse, a further difficulty
arises. The procedure requires a word or lexical unit to show a difference between a
‘basic’ meaning and its meaning in the context of the discourse. However, when like
(or a similar marker) is used, this is not the case; the stripy sock in the Alentejo Blue
example means just that, a stripy sock. That is not to say the components of a simile
can never be used metaphorically – cool in cool as a cucumber (to take one of Gibbs’s
1994: 270 examples) clearly is – but simply to note that the MIP procedure cannot
be used as a general means of identifying metaphoric similes.3
For the purposes of this study, the procedure described by Cameron (2003)
was used. This requires the components of the comparison to be semantically
discrepant and (like MIP) source domain meaning to be used to make contextual
sense of the vehicle; this has the advantage that it allows the researcher to isolate
multiword stretches of utterance, not just single words (or lexical units).
4. What similes do
Similes, whether metaphoric or literal, clearly establish a connection based on

similarity between two or more things, but the marker word indicates to the read-
er that the connection should be (or may be) treated as partial, unreal or imagi-
nary. Thus even in a comparison such as John is like his father, John is still not the
same as his father in all ways. Moreover, while metaphor temporarily identifies
A with B, simile, following the characterisation in the last section, pulls back and
refrains from making the identification. The result can have different impacts.
Firstly, the fact that the marker indicates prevarication or hedging can imply
tentativeness or prevarication in the discourse. Indeed, Goatly (1997) found that
authors often began a series of metaphoric expressions with a simile (i.e. once the
reader had accepted the connection, it could be firmed up as a metaphor).
Secondly, the hedging and the syntax of making the noun or noun phrase de-
pendent on like can create a backgrounding effect at discourse level (see Low, 1999;
Khalil, 2005), allowing the speaker to introduce comparisons into the discourse,
but not in such a way that the focus on the main entities being discussed is lost.
One key role of similes may thus be to help control saliency and foregrounding.
Thirdly, markers can do more than hedge a metaphor in the way that Lakoffian
tradition suggests (Lakoff & Johnson, 1980). Littlemore and Low (2006: 42–43)
noted that adding or removing the marker could change the interpretation of
several sentences. They suggested, for example that while life is a joke convention-
ally ignores the prototypical function of a joke to be amusing or funny to the par-
ticipants, life is like a joke is not so restricted. Similes may thus be more suited to
focusing on central or core characteristics, while metaphor may be better where
peripheral characteristics are important.4
The somewhat paradoxical combination of (a) retaining the focus on the
main message and (b) nevertheless having an explicit comparison makes similes
ideally suited to educational explanations – since the learners ought to be clear
firstly, that an explanation is being given and secondly, that the entity being used
to explain is not to be confused with the entity being explained. One might also,
following Gentner (1982) and Gentner, Bowdle, Wolff, & Boronat (2001), perhaps
expect many educational or academic analogies (whether standardly part of the
discipline/theory concerned or coined by a teacher for a particular situation) to
be detailed and for the detail to relate to structural relationships more than to
physical characteristics or emotional overtones. In short, it seems reasonable to
expect that university lecturers, one of whose jobs is to interpret, evaluate, clarify
and explain academic ideas to students, would make reasonably frequent use of
similes in their lectures.
296 Graham Low
5. Four university lectures
Low et al. (2008) analysed the role of metaphor in three UK university lectures,
from the BASE corpus.5 One concerned race relationships, one looked at the his-
tory of the European Monetary Union (EMU) and the third, which was much
more conversational in style, related to international business strategy. Metaphor
was initially identified following the MIP procedure; metaphoric simile and direct
analogy were identified separately.
The results showed that the two more ‘rhetorical style’ lectures had just one
metaphorical simile each, while the conversational one had seven. In order to see
if this was simply a quirk of sampling, a second conversational style lecture was
selected for the present analysis: on creative writing for first year computer sci-
entists. Although this lecture included a lot of interactive and informal language
(e.g. / well (.) / # / that’s / bollocks (2) / ) and had a fairly short mean tone-group
length of 3.5 words, it was very fluently delivered, with just 48 filled pauses and 13
broken tone groups in 45 minutes.6 Stylistically it was therefore not quite as con-
versational as the Strategy lecture. The overall corpus was just over 33, 400 words
and the baseline details are given in Table 1.
Table 1. Baseline information on the four lectures

Lecture Race EMU Strategy Writing
Topic UK race European Company Effective writing
relations currency strategy
BASE code Ss022 Ss024 Ss032 Ps041
Audience Undergraduate Presessional Graduate Undergraduate
(mixed)
Position in course Guest lecture Mid-course Initial lecture Initial lecture
Multimedia – – OHT and OHT
whiteboard
Duration 55 min 61 min 80 min 45 min
Words 5,015 9,499 13,109 5,796
Intonation units 1,225 2,190 3,712 1,576
Interaction with Just ‘I/you/we’ 3 questions Throughout ‘I/you’, 2 tasks
audience at end and audience
laughter
Total word count: 33,419; total tone group count: 8,703

Adapted from Low et al. (2008: 433)
6. Findings
6.1 The more rhetorical-style lectures
The Race lecture had just a hammer-like implementation of policy used for em-
phasis and alliterating with an earlier hinge in the same utterance (each line rep-
resents one tone group):
(1) 209 the state’s attitude in general (1)
→ 210 was to hinge (1)
211 policy
212 attempts at
213 resolving those issues around the one (2)
→ 214 basically hammer-like implementation of policy around immigration
The EMU lecture simply had the hedged

(2) 385 so there’s a
→ 386 there’s a kind of optimistic myth
387 underlying (.)
388 much talk about the EU
389 that says
Assuming this is a simile (as it is arguable how far comparison, rather than
purely exemplification, is actually involved here), it introduces a summary of a
preceding discussion (in tg 262–4, not cited here) about a popular myth and im-
mediately prefigures the climax about the focus of the myth. The kind of hedge
in the simile (above) also parallels in summary form the complex hedging when
myth is first introduced:
(3) 262 and a popular (.)
→ 263 well I think it’s a popular myth anyway
264 a popular myth
265 that surrounds (.) the
266 motives
267 underlying the creation
268 of #
269 the
270 European Union
One might argue (with Hanks, p.c.) that a kind of optimistic myth is more of a
‘type of ’ statement than a comparison (making it a hedged metaphor), but as one
could substitute something similar to for kind of, it is here treated as a simile.
298 Graham Low
6.2 The more conversational-style lectures
The Strategy lecture had seven overt metaphorical similes and one case where
a simile was apparently changed, via complex hedging in mid-utterance plus a
change of verb (underlined below), into a metaphor involving buffering (reducing
the effect of an impact or shock):
(4) → 1721 the I mean organizations are like sort of
→ 1722 # you know
→ 1723 have huge buffering power
1724 to actually stop any change
Three similes simply formed part of a group of thirteen Hell expletives and com-
prised two like Hells and a sure as Hell. These similes may perhaps be seen as
adding variety and as emphasising the conversational style being adopted.7 Two
others were one-off examples; one was a conventional idiom (/ this this area / that
was growing like Topsy / tg 2878–9), the other was a comparison using a common
hyperbolic saying:
(5) 1954 being the sort of macroeconomic contribution
→ 1955 which is a bit like saying
→ 1956 eating babies is wrong
1957 #
1958 so
1959 it
1960 you don’t get any brownie points for that
More interestingly, the lecture contained three analogies, two of which involved
overt metaphorical similes. Analogy one compared the business environment to
an onion. The lecturer began with complex hedging and then hedged the grounds
(you sort of peel). The simile comes at the start (tg 1098) of the analogy, with the
climax several tone groups away at the end (tg 1104).
(6) 1098 because it’s rather like
→ 1099 like an onion
1100 you know
1101 you
1102 you sort of peel this thing
1103 through various layers
1104 in order to understand the macroenvironment
A similar situation arose a few minutes later, when business guru Charles Handy
was metaphorically identified with St Paul, then compared twice to God. This
time, the focus was more on physical characteristics and the ground of the com-
parison was given in considerable detail. Each simile functions as a minor climax,
with the second one immediately prefiguring the summary climax (tg 1089):
(7) 1074 this is
1075 Charles on the road to Damascus
1076 I mean
1077 this guy
1078 I always think has an enormous advantage
1079 because if you listen to Thought for the Day
1 → 1080 he sounds like God
1081 and then when you meet him
1082 you know
1083 elderly
1084 white hair
1085 white Anglo-Saxon WASP
2 → 1086 he looks like God
1087 so
1088 at least in Michelangelo’s sense
1089 so he’s got a lot going for him
The Creative Writing lecture contained four clearly metaphorical similes, one
possible one and a group of five feel likes which seemed to move in and out of
metaphoricity. In this lecture the speaker was greatly concerned to forge a con-
nection between ‘mundane’ science reporting and the successful communication
of writers like Hawking and Gould, which were held to be just one step from
creative writing. The aim was to give the students a familiar ideal to work towards
and to make them feel both excited at the unknown territory in front of them, yet
unthreatened because of the familiarity of the famous science writers.
The first simile followed an introduction to the scientists; it seems potentially
metaphoric, as the speaker expects the audience to perceive scientists’ and their
own work as worlds apart and of a different order.
(8) 164 # the result is
165 is that when a few scientists (.)
166 actually succeed in communicating (.)
167 to the public (.)
168 like Hawking here
169 they get called geniuses
170 see the word up there
171 genius (.)
172 well (.)
300 Graham Low
173 scientists like Stephen Hawking

174 or Bill Gates
175 or Stephen J Gould (.)
176 hh
1 → 177 what they actually do (.)
1 → 178 the content of their message
1 → 179 is fascinating in its own right
2 → 180 it’s just like the work that you do (.)
2 → 181 what you do
2 → 182 is fascinating in its own right (.)
Textually, it follows a three-part list of names (tg 173–175) and forms the precur-
sor to the climax of a parallel structure involving two parts, each of three tone
groups (tg 177–182, labelled 1 and 2). The need to reach a climax involving you
not them presumably determines why it is the students’ work, and not that of the
already famous scientists, that receives the hedging (and very temporarily back-
grounding) effect of the like.
The second simile was an explanation of stanza, which the speaker suddenly
realised might not be understood by the audience:
(9) 908 # sorry (.)
909 your first set of reading (.)
910 is on your handouts (.)
911 I’d like you to read this material (.)
912 before your first seminar (.)
913 it includes an analysis of Stephen Hawking’s (.)
914 style (.)
915 in A Brief History of Time (.)
916 an excerpt
917 from an essay by George Orwell
918 the author of Nineteen Eighty Four (.)
919 hh
920 there is also a stanza
921 of a poem
922 by the poet
923 John Keats (1)
924 a stanza in a poem
925 if you don’t w-,
926 know what that is
→ 927 it’s like a paragraph in a
928 in
929 # a piece of prose
It is again arguable how far like a paragraph is metaphorical, but in that prose is
in general distinct from poetry (if we ignore such things as Baudelaire’s poèmes en
prose) and the familiar is being used to explain the less familiar, I treat it here as at
least potentially metaphorical. The simile is, as before, associated with hesitation
and functions as the immediate precursor to the climactic tone group # a piece of
prose (tg 929) – whose effect is in the event somewhat diminished by the realisa-
tion that the students might not understand prose either! Yet again, the simile
forms part of a short but complex rhetorical structure, involving parallelism in
the form of a three-part list.
Example three (like a foreign language, tg 1127) is again part of a small rhe-
torical group, but this time there is a metaphor cluster involving interlaced in-
stances of language, tongue, talk, an intertextual protesteth8 and literal speak and
write (the sequence is: protesteth > talk > tongues > language > tongue > lan-
guage > speak > write > language). It comes at a boundary, after comments about
George Orwell:
(10) 1112 now
1113 I think George Orwell had a strong point for nineteen forty six
1114 but I think that text
1115 protesteth
1116 too much
1117 for nineteen ninety eight (.)
1118 the truth is
1119 that we all have to talk in many tongues
1120 and we need to (.)
1121 there are many professions (.)
1122 the language of science
1123 itself (.)
1124 is an international tongue
1125 it’s
→ 1126 like
→ 1127 a foreign language (.)
1128 to most people (.)
1129 the trick is to speak
1130 or write
1131 in more than one
1132 language (.)
The simile once more acts as the immediate precursor to a climax (tg 1129–32),
though again it is a subclimax, as the main climax occurs several tone groups later,
after a list of three I would argue:
302 Graham Low
(11) 1163 what I’d like you to do

1164 all
1165 today (1)
1166 is
1167 turn
1168 that
1169 poem (.)
1170 turn that
1171 piece of writing (2)
1172 into a piece of science (1)
Simile four (like crazy, tg 1484) is a one-off example that is part of a three-part list
of things the speaker is going to do, but seems purely there for emphasis and to
create an adverbial:9
(12) 1482 the difference between this session of writing and the last one is (.)
1483 I’m going to try to distract you
→ 1484 like crazy (1)
1485 I will attempt to
1486 overwhelm you (.)
1487 from this simple task (.)
1488 of writing (.)
1489 I’ll be using music to do it (.)
Finally, the five feel likes occur in two groups and all are designed to establish the
feeling that creative writing engenders and the sense of being overwhelmed that
it can bring.
(13) → 1 1432 what does it feel like (.)
1433 well
1434 it’s a feeling of distraction (.)
1435 it’s a feeling of
1436 helplessness
1437 before the written word (.)
1438 and
1439 often turns into
1440 #
1441 a mistaken belief by yourselves that in fact it’s a condition
of hatred (.)
1442 mixed with fear (.)
1443 of the written word (.)
→ 2 1444 writing (.)
1445 it feels like ignorance (.)
1446 but it is not ignorance (.)

→ 3 1447 it feels like this (.)
1448 hh
1449 please pick up your pens again
* * * *
1505 and I promise you (.)
1506 the sensations that you’ll feel (.)
1507 during the next two minutes (.)
1508 that feeling of overwhelming (.)
1509 that
1510 is the feeling
1511 that can be avoided (.)
1512 if you don’t (.)
1513 sorry
1514 that’s the feeling that you can actually
1515 just have to get this right (.)
→ 4 1516 that’s what it will feel like
1517 if you don’t acquire the tools (.)
1518 to write essays
1519 and technical reports (.)
1520 articles (.)
1521 poems (.)
1522 and novels (.)
→ 5 1523 that
1524 is what it will feel like
1525 if you don’t acquire the tools now (.)
1526 you’ll be overwhelmed (.)
1527 when you’re faced with information overload (.)
The first point to note is that the feel likes are all involved in short structures
involving parallelism or listing. In both extracts, the last feel like acts as the im-
mediate precursor to the climax. In both cases too, the feel likes form part of a
wider lexical cluster of feeling(s). How far the five examples could be said to be
metaphorical is arguable: perhaps only (2) is clearly so. However, the speaker ap-
pears to be playing on the edge of metaphor in both extracts.
7. Discussion
It has been found in some lectures that one use of a metaphor triggers another,
to form a sequence of clusters through the lecture (Corts, 1999; Corts & Pollio,
304 Graham Low
1999; Corts & Meyers, 2002), at times forming a single overarching metaphor
(Ponterotto, 2003). The Strategy lecture did end by bringing together briefly sev-
eral metaphoric themes used at earlier points (see Low et al., 2008), but there was
no evidence anywhere (apart from the feel likes) of similes being repeated, clus-
tered or turned into full metaphors. In short, metaphoric similes were rare in the
lectures and, where they were used, functioned primarily as one-off expressions,
answering locally relevant, not global needs.
In recent years, Chiappe, Kennedy and colleagues have suggested (e.g.
Chiappe, Kennedy & Chiappe, 2003; Chiappe, Kennedy & Smykowski, 2003)
that writers prefer simile over metaphor where a comparison is felt (a) not to be
very apt, (b) to be hard to understand, or (c) to require focusing or narrowing.
Roncero, Kennedy & Smyth (2006) hypothesised from this that similes should be
more likely than metaphors to be accompanied by because …-type explanations,
and conducted two experiments using similes and ‘A is B’ format metaphors from
the Internet which supported the hypothesis. Nevertheless, it has to be said that
the data they cite show relatively few cases of similes or metaphors with explana-
tions and just 4 of the 31 ‘A is (like) B’ pairs show the simile to be accompanied
more frequently than not by an explanation. However, the suggestion is interest-
ing and it is worth checking for in the lecture data. Roncero et al. only categorised
something as an explanation if a because could be inserted; I relaxed this to allow
paraphrases of because. Even using the broader criteria, however, the result is that
none of the similes was accompanied by explanation, though three were given
the grounds for the comparison and one was followed by an implication. It would
appear that, in the sort of education discourse examined here, similes tend to be
treated as understandable or self-evident.
8. Conclusion
The conclusions are fairly clear and stark. Overall, the four lectures contained
very few metaphoric similes, and those that did occur are only found in the two
more conversational lectures. Few, apart from the onion analogy were imagina-
tive, non-conventional or attention grabbing – the speakers presumably felt they
had the audience’s attention, or if not, they used metaphor rather than simile to
get it (Low et al., 2008). In neither lecture were the similes used in rhetorical net-
works that spread, or ‘arched’, over large areas of discourse; in this they resembled
the metaphors, that rarely had a global function either (Low et al., 2008). Nor
were they standardly accompanied by explanations, as Roncero et al.’s (2006) find-
ings might have led one to expect. Moreover, although one might have expected
several of the similes to be used to explain complex concepts or theories, none
could be treated (following Boyd, 1979/1993) as “theory constitutive”; even the

onion analogy was, to use Boyd’s labels, more “exegetical” or “pedagogical”. Turn-
ing to the structural roles of the similes, only occasionally did one have a tentative
or introductory function and precede a full metaphoric identification later in the
lecture: a somewhat different situation from what Goatly (1997) found in his data.
The metaphoric similes were, however, regularly bound up with short rhetorical
sequences involving three- (or more) part lists and used to resolve local problems
of emphasis or saliency. In several cases, the simile was used to build towards a
later climax, but did not itself function as one; the speaker thereby kept the focus
of the argument on the main topic, rather than shifting it to the entity involved in
the comparison.
If simile is envisaged by lecturers, albeit probably unconsciously, as a device
to solve local problems, or to exercise local control of either saliency or potential
misunderstanding, it is perhaps unsurprising that such an explicit form of com-
parison occurs more in the two more conversational-style lectures. This would
support Carter’s (2004: 125) general finding that simile seems to be associated with
everyday speech, as it frequently does not require too much interpretive effort on
the part of the listener. However, the data do not dovetail with Carter’s suggestion
that “simile is more frequent than metaphor in everyday speech” (p. 125). The
scarcity of similes in the lectures contrasts starkly with the proportion of meta-
phorically-used words (i.e. excluding similes), as reported in Low et al. (2008):
Race lecture, 13%; EMU lecture, 11%; Strategy lecture, 10%. Precise figures were
not given for deliberately-used or highly active metaphor, but even the examples
discussed in the 2008 paper far exceed the total simile count here.
The sample used in this study was clearly very small, and it would be useful
to know how far similar patterns of use occur in other lectures, especially ones
involving multimedia presentations or more interaction by the students. It would
also be of interest to discover how far the use of simile in lectures parallels its use
in other university genres, such as seminars, workshops, or supervision sessions.
Notes
1. The term ‘ground’ is used in a range of different ways in education, psychology and cogni-
tive science. It is used in this paper purely in the sense that Richards (1936) used it: to refer to
the features or correspondences which are the basis of the comparison.
2. It is sometimes suggested that the marker like is itself metaphoric in metaphoric similes
(Kittay, 1987: 19) and that this fact makes the marker a useful identification tool (see Fishelov,
1993: 14). However, deciding whether like flags a relationship that is somehow ‘real’ or ‘literal’
seems to be as difficult as establishing domain separateness or distance, and so the technique
has not been adopted here.
306 Graham Low
3. Steen (2007) and Steen, Dorst, Herrmann, Kaal, Krennmayr, & Pasma (2010) discuss how
a separate simile identification procedure was developed for their project, but I did not have
precise details of this when the lecture data were analysed.
4. Fishlov’s (2003/7) elicited results from context-free similes appear to support this conclu-
sion: “people tend to rely as much as possible on existing semantic categories when they inter-
pret metaphorical expressions, including novel ones” (2007: 74) “and [they] shun the possibility
of opening and reinterpreting existing categories” (2003: 76).
5. The transcriptions and recordings used in this study come from the British Academic Spo-
ken English (BASE) corpus project. The corpus was developed at the Universities of Warwick
and Reading under the directorship of Hilary Nesi and Paul Thompson. Corpus development
was assisted by funding from BALEAP, EURALEX, the British Academy and the Arts and Hu-
manities Research Council.
6. The transcripts follow the BASE convention of marking a filled pause by a hash sign #.
Unfilled pauses are marked by brackets containing the time to the nearest whole second, e.g.
(4). A (.) means a noticeable pause but one lasting less than a second. The times were mea-
sured by myself.
The system for categorising tone groups was the same as used in Low et al. (2008). A tone
group is classed as broken where the speaker breaks off in the middle of a word, as in / if you
don’t w- / (Creative writing, tg 925). When tone groups are cited in the middle of a sentence,
they are marked by single slashes (as with 925 above).
7. Moon (2008) argues that because as Hell, in as X as Hell, can be applied to a huge range of
adjectives, the formula should be treated as a template for generating emphatic phrases, rather
than as simile. However, since as Hell seems unlikely to intensify adjectives denoting heav-
enly attributes, such as saintly, unless one is trying to evaluate them or the holder negatively,
I consider it to have some residual links with comparison, and so treat it and like Hell here as
conventionalised simile.
8. The expression harks back, I assume, to Hamlet Act 3, Scene 2 and the irony of Queen
Gertrude’s comment on the Player Queen’s hyperbolic assertions that she would never marry
again, were the (Player) King to die: “The lady doth protest too much, methinks”.
9. Like crazy (tg 1484) poses the same identification problem as like Hell; the phrase’s key
function is clearly to intensify and/or to emphasise, but there nevertheless seems some residual
sense of distracting the listener so actively and wildly that one appears to be crazy.
Acknowledgements
My thanks to Hilary Nesi for permission to use the BASE data and to David
Fishelov, Patrick Hanks, Berenike Herrmann and Lynne Cameron for comment-
ing on drafts of the paper.
References
Ali, Monica (2006/7). Alentejo blue. London: Black Swan.

BASE (2005). The British academic spoken English corpus. University of Reading and Universi-
ty of Warwick. Home page: http://www2.warwick.ac.uk/fac/soc/al/research/collect/base/
Boyd, Richard (1979/1993). Metaphor and theory change: What is “metaphor” a metaphor for?
In A. Ortony (Ed.), Metaphor and thought (356–408). Cambridge, UK: Cambridge Uni-
versity Press.
Cameron, Lynne (2003). Metaphor in educational discourse. London: Continuum Press.
Cameron, Lynne & Juurd Stelma (2004). Metaphor clusters in discourse: Methodological is-
sues. Journal of Applied Linguistics, 1, 107–136.
Carter, Ronald (2004). Language and creativity. London: Routledge.
Chiappe, Dan L., John M. Kennedy, & Penny Chiappe (2003). Aptness is more important than
comprehensibility in preference for metaphors and similes. Poetics, 31, 51–68.
Chiappe, Dan, John M. Kennedy, & Tim Smykowski (2003). Reversibility, aptness, and the con-
ventionality of metaphors and similes. Metaphor and Symbol, 18, 85–105.
Corts, Daniel P. (1999). Spontaneous production of figurative language and gestures in college
lectures: A comparison across disciplines. Unpublished doctoral dissertation, University of
Tennessee, Knoxville.
Corts, Daniel P. & Karen Meyers (2002). Conceptual clusters in figurative language production.
Journal of Psycholinguistic Research, 31, 391–408.
Corts, Daniel P. & Howard R. Pollio (1999). Spontaneous production of figurative language and
gesture in college lectures. Metaphor and Symbol, 14, 81–100.
Darian, Stephen (1973). Similes and the creative process. Language and Style, 6, 48–57.
Fishelov, David (2003). Shall I compare thee? Semantic categories and the interpretation of
similes. Paper presented at the 5th International Conference of Researching and Applying
Metaphor, Paris September 3–5. Reference is to the Program and Abstracts.
Fishelov, David (2007). Shall I compare thee? Semantic categories and the interpretation of
similes. Journal of Literary Semantics, 36, 71–87.
Fishelov, David (1993). Poetic and non-poetic simile: Structure, semantics and rhetoric. Poetics
Today, 14, 1–23.
Gentner, Dedre (1982). Are scientific analogies metaphors? In D. Miall (Ed.), Metaphor: Prob-
lems and perspectives (106–132). Brighton: Harvester.
Gentner, Dedre, Brian F. Bowdle, Phillip Wolff & Consuelo Boronat (2001). Metaphor is like
analogy. In D. Gentner, K. J. Holyoak & B. N. Kokinov (Eds.), The analogical mind: Perspec-
tives from cognitive science (199–253). Cambridge, MA: MIT Press.
Gibbs, Raymond W. (1994). The poetics of mind. Cambridge, UK: Cambridge University Press.
Goatly, Andrew (1997). The language of metaphors. London: Routledge.
Haddon, Mark (2007). A spot of bother. London: Vintage.
Hanks, Patrick (2007). The function of similes. Paper presented at the 2007 Metaphor Festival,
University of Stockholm, 20–21 September.
Khalil, Esam N. (2005). Grounding between figure–ground and foregrounding–background-
ing. In F. J. Ruiz de Mendoza (Ed.), Annual Review of Cognitive Linguistics 3 (1–21).
Kittay, Eva F. (1987). Metaphor: Its cognitive force and linguistic structure. Oxford: Oxford Uni-
versity Press.
308 Graham Low
Press.
Littlemore, Jeannette & Graham Low (2006). Figurative thinking and foreign language learning.
London: Macmillan Palgrave.
Low, Graham (1997). Celebrations and SQUID sandwiches: Figurative language and the manipu-
lation of academic writing. Unpublished project report, University of York, UK. Retrieved
from http://www.york.ac.uk/depts/educ/people/Celebration.pdf
Low, Graham (1999). Teaching foregrounding skills in academic text. In L. Kasanga (Ed.),
Language(s) and the new democracy: Participation, rights and responsibilities. Proceedings
of the 18th Annual Conference of the Southern African Applied Linguistics Association
(SAALA), University of the North, 7–9 July 1998 (103–123). Sovenga, SA: University of
the North Press.
Low, Graham (2005). Explaining evolution: The use of animacy in an example of semi-formal
science writing. Language and Literature, 14, 129–148.
Low, Graham (2008). Metaphor and positioning in academic book reviews. In M. Zanotto,
L. Cameron & M. Cavalcanti (Eds.), Confronting metaphor in use: An applied linguistic ap-
proach (79–100). Amsterdam & Philadelphia: John Benjamins.
Low, Graham, Jeannette Littlemore & Almut Koester (2008). Metaphor use in three UK univer-
sity lectures. Applied Linguistics, 29, 428–455.
Moon, Rosamund (2008). Conventionalised as-similes in English: A problem case. Interna-
tional Journal of Corpus Linguistics, 13, 3–37.
Paprotté, Wolf & René Dirven (Eds.) (1985). The ubiquity of metaphor: Metaphor in language
and thought. Amsterdam & Philadelphia: John Benjamins.
Ponterotto, Diane (2003). The cohesive role of cognitive metaphor in discourse and conversa-
tion. In A. Barcelona (Ed.), Metaphor and metonymy at the crossroads. A cognitive perspec-
tive (283–298). Berlin: Mouton de Gruyter.
Pragglejaz Group (2007). A practical and flexible method for identifying metaphorically-used
words in discourse. Metaphor and Symbol, 22, 1–40.
Richards, Ivor A. (1936). The philosophy of rhetoric. New York: Oxford University Press.
Roncero, Carlos, John M. Kennedy & Ron Smyth (2006). Similes on the Internet have explana-
tions. Psychonomic Bulletin & Review, 13, 74–77.
Sayce, Richard A. (1953). Style in French prose. Oxford: Clarendon Press.
Steen, Gerard (2004). What counts as a metaphorically used word? The Pragglejaz experience.
In S. Coulson & B. Lewandowska (Eds.), The literal–nonliteral distinction (299–322). Ber-
lin: Peter Lang.
Steen, Gerard (2007). Finding metaphor in grammar and usage. Amsterdam & Philadelphia:
John Benjamins.
Steen, Gerard J., Aletta G. Dorst, J. Berenike Herrmann, Anna A. Kaal, Tina Krennmayr &
Trijntje Pasma (2010). A method for linguistic metaphor identification: From MIP to
MIPVU. Amsterdam & Philadelphia: John Benjamins.
Teng, Norman Y. & Sewen Sun (2003). Grouping, simile, and oxymoron in pictures. Metaphor
and Symbol, 17, 295–325.
Wodehouse, Pelham Grenville (1917/8). Piccadilly Jim. Project Gutenberg. Retrieved from
http://onlinebooks.library.upenn.edu/webbin/gutbook/lookup?num=2005
chapter 15
Metaphor marking and metaphor typological

and functional ranges in business periodicals
Hanna Skorczynska Sznajder

Universidad Politécnica de Valencia
This chapter analyses the distribution of marked and unmarked metaphors

(Goatly, 1997) within the typological and functional ranges identified in a
corpus of business periodicals (Skorczynska & Deignan, 2006). The corpus
contained approximately 400,000 words and was built from journalistic articles
published in The Economist, Business Week, and Fortune. The metaphors ana-
lysed functionally in a previous study (Skorczynska & Deignan, 2006), and their
co-text were searched for metaphorical markers (Goatly, 1997) with a concor-
dancer. The results obtained show that there is no significant variation with re-
gard to metaphor marking and metaphor types identified (active and inactive).
However, there are notable correlations between metaphor marking and their
functions (generic, filling terminological gaps, illustrating). This confirms an
earlier finding that metaphor anticipation on the text level by means of linguis-
tic expressions does not depend on the degree of metaphor conventionalisation,
but rather on the context in which a particular text is produced (Cameron &
Deignan, 2003).
Keywords: metaphor marking, metaphor types, metaphor functions, business

discourse
1. Introduction
Real-world metaphor research answers questions about how metaphors are used in
real-life situations by providing empirical data found in samples of all possible dis-
course types (Gibbs, this volume). The requisite of ‘ecological validity’ for this type
of research entails examining the context in which metaphors are used, so that ap-
propriate adjustments can be made to the existing theoretical perspectives, most of
which were formulated in the pre-corpus period of metaphor studies. In this sense,
the present analysis attempts to contribute to the corpus-based metaphor research
310 Hanna Skorczynska Sznajder
by looking into ways in which metaphorical uses are signalled to the reader in the
context of business journalism. This, in turn, would provide additional evidence
on how metaphors are used in that particular communicative situation.
The linguistic expressions surrounding a metaphor on the text level may offer
clues to understanding how metaphors are intended and interpreted in differ-
ent communicative contexts (Cameron & Deignan, 2003; Glucksberg & Keysar,
1993; Goatly, 1997). The effect exerted by those expressions on linguistic meta-
phors generally consists of reducing or increasing their metaphoricity, so that
the listener or reader can be conveniently cued into the intended identification
and interpretation of a metaphor. The specific functions of the so-called hedges
(Glucksberg & Keysar, 1993), metaphorical markers (Goatly, 1997), or tuning de-
vices (Cameron & Deignan, 2003) may range from preventing a metaphor from
being understood literally, avoiding a metaphor’s misinterpretation, specifying
the type of mapping between the topic and the vehicle, to enhancing or toning
down a metaphor’s force.
A number of parameters of metaphorical marker use (Goatly, 1997) could
vary depending on the type of text and its social context. A broader collection
of marker categories, a larger number of individual markers, and their higher
frequencies were identified in a corpus of business periodicals, as compared to
a related corpus of business research articles (Skorczynska & Piqué Angordans,
2005). Apart from the text type and its context, the differences in the co-text of
linguistic metaphors could be directly related to the type of metaphor used and its
functions (Goatly, 1997). A diverging view confines the correlation of metaphor
with its co-text to the degree of metaphor expectedness in a particular discourse
context, regardless of the type of metaphor used (Cameron & Deignan, 2003).
This chapter discusses the correlations mentioned using the corpus of busi-
ness periodicals and the description of metaphor typological and functional
ranges identified in a previous study (Skorczynska & Deignan, 2006). The re-
search presented herein confirms that in this particular text type and its context,
metaphorical marking is related to metaphor functions, rather than to the level of
metaphor conventionalisation.
2. The framework of study
2.1 Metaphor functional and typological ranges
Metaphor choice and its functions in a particular text seem to be determined by

the intended readership, and thus more generally by social context (Skorczynska
& Deignan, 2006). Within the framework established and based on previous
Chapter 15. Metaphor marking, functions and types 311
work by Henderson (1986), Lindstromberg (1991), and Goatly (1997), the anal-
ysis of metaphor linguistic realizations in a corpus of business periodicals by
Skorczynska and Deignan (2006) revealed that metaphors used there fulfilled
three main functions: generic, filling terminological gaps, and illustrating busi-
ness phenomena.
One of the corpora used for that study was compiled by randomly selecting
articles from 3 business periodicals: The Economist, Business Week, and Fortune,
and contained 404,251 words. The business periodicals used are widely known
sources of information about world business and current affairs. They basically
target professionals involved in different areas of management, although the read-
ership also includes the general public. The texts were written by journalists spe-
cialised in the area of economy and business between 1997 and 2003.
Regarding metaphor functional ranges described in Skorczynska and
Deignan (2006), the generic function was fulfilled by metaphors belonging to
the conventional stock of the language, that is, by those metaphors whose mean-
ing was covered in a general dictionary (Rundell, M. & Fox, G. Macmillan Eng-
lish Dictionary for Advanced Learners), but not in a specialist one (e.g. Collin, P.
Dictionary of Business). Example (1) shows the use of “pocket” with its generic
metaphorical sense.
(1) In exchange, Zurich was granted the right to the first $225 million from a
sale, leaving Ritt and his team to pocket 60% of anything left over.
(Behar, 2003: 138)
Other metaphors identified in the corpus that were included in the specialist dic-
tionary, but not covered or mentioned as genre-specific in the general dictionary,
performed the function of filling terminological gaps. Such metaphors seemed to
have emerged originally because of the absence of a term for a particular entity,
quality or action. The example that follows includes the metaphorically-originat-
ed term “bull market”.
(2) But, as industry after industry was streamlined and deregulated, the seeds of
the 15-year bull market in America were sown.
(Bremner & Capell, 1997: 111)
Finally, the function of illustrating business phenomena was carried out
by metaphors best understood as an adaptation of Henderson’s (1986) and
Lindstromberg’s (1991) view of metaphor as textual decoration or illustration
to the genre-specific corpus used in the study. In this sense, “drunks” in Ex-
ample (3) illustrates the unpredictability of economies.
(3) Yet economies, like drunks, continue to move in wavy lines.
(Woodall, 2002: 5)
The study found that the most frequent function fulfilled by metaphors in the
corpus was the generic one (60%). It was followed by filling terminological gaps
(35.8%), and illustrating business phenomena (4.2%).
With reference to metaphor types, the study focused only on active and in-
active metaphors, a typology taken from Goatly (1997), which was originally
extended to sleeping, dead, and buried metaphors. Metaphors that are well-es-
tablished in the language and covered by the dictionary (Rundell & Fox, 2002)
were considered as inactive (Example (4) with “blossom” and “wither”). On the
contrary, the metaphors not included in the dictionary were regarded as active
(Example (5) with “shoal” and “shipwreck”). Goatly’s dead and buried metaphors
were excluded, as they show no semantic connection with their metaphorical ori-
gin for current speakers of English.
(4) In the 1980s Tokyo blossomed, then withered, along with Japan’s economic
boom and bust; but for most other financial centres national economic
strength has been playing a diminishing part in success or failure.
(Edwards, 1998: 4)
(5) To ensure that markets remain open and that globalization fulfils its promise,
a new consensus will have to develop for managing the process – steering
vulnerable economies past the most dangerous shoals and coping better with
shipwrecks when they occur. (Pennar, 1998: 116)
The research showed that nearly all of the metaphors identified were inactive,
with just 4.1% of active metaphors.
2.2 Metaphorical markers
As noted previously, metaphorical markers are words and phrases occurring in

the environment of a metaphor’s vehicle term (Goatly, 1997). An inventory of 20
types of metaphorical marker was used in this study (see Figure 1).
Goatly’s inventory is possibly the most readily available list of metaphorical
markers, although it shows certain inconsistency in the use of categorisation cri-
teria. For instance, explicit markers, intensifiers, hedges and downtoners, or sym-
bolism terms have probably been distinguished on the functional basis. Other
categories, such as semantic metalanguage, mimetic terms, perceptual processes,
misperception terms, or cognitive processes have more in common with a se-
mantic rather than with a functional classification. Finally, modals, conditionals
or copular similes are clearly grammatical categories. In spite of these inconsis-
tencies, the inventory turned out to be especially useful for a concordance-based
query of an electronic corpus.
Figure 1. Metaphorical markers (Goatly, 1997: 174–5)

Marker category Metaphorical markers
1. Explicit markers metaphor/-ically, figurative/-ly, trope
2. Intensifiers literally, really, actually, in fact, simply, fairly, just,
absolutely, fully, completely, quite, thoroughly, utterly,
veritable, regular
3. Hedges and downtoners in a/one way, a bit of, half-…, practically, almost, not
exactly, not so much … as…, … if not
4. Semantic metalanguage in both/more than one sense/s, mean(-ing), import
5. Mimetic terms image, likeness, picture, parody, caricature, model, plan,
effigy, imitation, artificial, mock
6. Symbolism terms symbol(-ic /-ically), sign, type, token, instance, example
7. Superordinate terms (some) (curious, strange, odd, peculiar, special) sort of,
kind of
8. Copular similes like, as
9. Precision similes and other material verb + like x, the y of a x, y’s x;
comparisons noun-adj., the x equivalent of
10. Clausal similes as if, as though
11. Perceptual processes seemed, sounded, looked, felt, tasted,
+ like/as though/as if
12. Misperception terms delusion, illusion, hallucination, mirage, phantom, fan-
tasy, unreal
13. Cognitive processes believe, think, regard, unbelievable, incredible
14. Verbal processes say, call, refer to, swear
15. So to speak
16. Orthography ” “ . ! white space
17. Modals + Verbal Processes could say, might say
18. Modals must, certainly, surely, would, probable/-ly, may, might,
could, possible/-ly, perhaps,
impossible/-bility
19. Conditionals if … could, would, might, imagine, suppose
20. As it were
3. Corpus and methodology
The research reported in this chapter used the corpus and the metaphorical mate-
rial from the study by Skorczynska and Deignan (2006). The linguistic metaphors
examined therein had been identified in a two-step procedure (Skorczynska &
Deignan, 2006: 89–90). It consisted first in analysing by hand a sample of the main
corpus for key vehicle terms, and after that, in searching for them in the main cor-
pus using a concordancer. The application of this procedure had resulted in the
identification of 83 different vehicle terms and their 1,659 tokens. The mentioned
number of vehicle terms also included the vehicles found in a parallel scientific
business corpus (Skorczynska & Deignan, 2006), which were then searched for in
the periodicals corpus.
In the present study, the linguistic metaphors identified were analysed in
their co-text with a concordancer (Scott, 2004), in search of metaphorical mark-
ers from Goatly’s inventory.
4. Results
4.1 Metaphor typological ranges and their marking
Most of the metaphors identified in the corpus were unmarked (94.2%), and
only 5.2% were marked by the linguistic elements from Goatly’s list. The data ob-
tained parallel the proportions of inactive (95.8%) and active (4.2%) metaphors
revealed in the previous study of this corpus, which might suggest that all active
metaphors are marked.
However, a further study of the active and inactive metaphors from the cor-
pus showed that the correlation between the type of metaphor used and its mark-
ing is insignificant. Only 19% of the active metaphors and 0.7% of the inactive
metaphors were marked. This finding provides further evidence for the view that
metaphor marking may not follow the pattern of metaphor type use (Cameron &
Deignan, 2003).
The examples that follow are active metaphors with the vehicles “eat” (Exam-
ple 6) and “cook” (Example 7). In Example (6), the marker “equivalent of ” from
the category of precision similes and other comparisons is used. The reader here
is provided with a clear indication of how “eating your seed corn” should be inter-
preted in the context of the student enrolment in graduate programmes over the
past years. Example (7) with two metaphors, one constructed through the vehicle
“eating what you cook”, and the other using “Asian cooking”, are not signalled to
the reader by means of a linguistic expression as a way of guiding her/him to their
intended interpretation.
(6) Total enrolment in graduate science and engineering programs has dropped
for three years in a row, as high pay for high-tech workers makes it more
appealing to get a job instead of going to graduate school. In the short run,
that’s the right thing to do, but it’s the equivalent of eating your seed corn:
It reduces the number of people working on the basic research needed for
years hence. (Mandel, 1998: 139)
(7) Three billion people in Asia cannot be producing and hoping that 500 mil-
lion people in the rest of the world will absorb it forever. At some point,
you have to start eating what you cook. The customers in the developed
world are getting quite full on Asian cooking. There is no reason why future
growth cannot be based on more balanced production and consumption.
(NaLamlieng, 1999, online)
Examples (8) and (9) are similar to the previous examples in the sense that two
active metaphors with the vehicle “fire” are treated differently as far as their antici-
pation on the text level is concerned. The metaphor of firing up the stock market
is intensified by “really” in Example (8), while a similarly operating metaphor in
Example (9), focused on companies, is not signalled at all.
(8) After all, the stock market was really fired up last year, with the total return
on the Standard & Poor’s 500-stock index coming in at a blistering 28.6%.
(Laderman & Smith, 1999: 72)
(9) Plenty of other companies will, in the next decade, undergo similar upheav-
als, fired by a change even more far-reaching than the harnessing of electrical
power a century ago. (“A connected world”, 1997: 3)
With regard to inactive metaphors, Examples (10) and (11) use the vehicle term
“casualty”. The first of them is anticipated by “look like” from the category of mark-
ers expressing perceptual processes. The inclusion of “casualty” in Example (11) is
more straightforward.
(10) On the surface Gruntal looks like another casualty of Sept. 11 and the post-
bubble carnage on Wall Street. (Behar, 2003: 137)
(11) The men and women still on the floor barking out orders to buy and sell occa-
sionally steal a wistful glance at the tombstones, but the line of empty swivel
chairs is enough to keep them focused on their jobs – lest they become the
next casualties. (Schwarts & Kahn, 2002: 68)
4.2 Metaphor functional ranges and their marking
The metaphors identified in the corpus fulfilled three main functions: generic,
filling terminological gaps, and illustrating (Skorczynska & Deignan, 2006). Met-
aphors used for generic and terminological purposes were not expected to be
anticipated on the text level by any of Goatly’s (1997) markers, since they are
included in dictionaries as members of the conventional stock of the general lan-
guage on the one hand, and business terminology, on the other. They should,
therefore, be easily interpreted by readers. However, metaphors illustrating busi-
ness phenomena might be signalled, as this type of anticipation would enhance
the pragmatic function in question and make it more explicit to the reader.
Graph 1. Distribution of marked and unmarked metaphors within their

The distribution
different of ranges
functional marked and unmarked metaphors within their different
functional ranges is shown in Graph 1.
0,0%
Terminological 100,0%
48,0%
Illu stratin g 52,0% Marked

Unmarke d
5,4%
Generic 94,6%
0 0,2 0,4 0,6 0,8 1
Graph 1. Distribution of marked and unmarked metaphors within their different
functional ranges
The analysis of the corpus confirmed the absence of metaphorical marking for
terminological metaphors. The metaphorically-originated business terms that are
well-established in the language certainly do not need to be signalled as non-
literal expressions to the fairly specialised reader of business news. It should be
noted here that the level of content specialization, and so of the lexis used in this
type of texts is relatively high even if we compare them to business research pa-
pers, which are written for and by specialists.
Regarding the illustrating and generic functions, the proportions of marked
metaphors differ importantly: nearly half of the illustrating metaphors (48%) were
marked, as compared to a minor percentage (5.4%) of the generic metaphors. The
results obtained support the idea that the reader may need to be tuned into the
metaphor being used if it unexpectedly illustrates a business phenomenon. By
contrast, the metaphors that are commonly used to construct everyday discourse
do not require linguistic anticipation.
Despite the fact that the proportion of marked illustrating metaphors is sig-
nificant (48%), it is less relevant in general terms: the function mentioned was
only fulfilled by 4.4% of all the metaphors studied. Nevertheless, these data sug-
gest that illustrating metaphors, though occasional in the discourse studied, are
fairly frequently anticipated by metaphorical markers. Being highly informative,
the illustrating metaphors can conveniently depict different business situations
and the use of signalling expressions can reinforce the pedagogical function of
journalist texts, as well as render business information more reader-friendly.
The metaphors in Examples (12) and (13) fulfil the illustrating function and
both contain the vehicle term “game”. The communicative purpose, for which the
first metaphor (Example 12) was used, is easily identifiable thanks to the meta-
phorical marker, “seem like”, from the category of markers expressing perceptual
processes. By contrast, Example (13) is an unmarked metaphor. Its function can

nonetheless be recognised on the basis of its syntactic and semantic configura-
tion, rather than as the result of the use of an anticipating expression.
(12) During those decades, technological progress apparently added little to
economic growth. That’s why a savings-oriented growth policy – no matter
how slow – seemed like the only game in town. (Mandel, 1998: 137)
(13) By buying stocks on the upswing (and immediately dumping any that are fall-
ing), he has scored time and time again. It’s a dangerous game, but McKinney’s
hunches paid off with Yahoo, for example. (Clifford, 2000: 161)
Similarly, Examples (14) and (15) are illustrating metaphors, and both draw on
the machine as the source domain to establish different mappings with the target
domains of world and business. The marker “as”, from the category of copular
similes signals the metaphor in Example (14). Example (15), in turn, lacks any an-
ticipating device for the metaphors of “roadkill” and “profit machine”, yet the met-
aphor’s illustrating function is performed by means of other discourse resources.
(14) The world as a single machine. (Carson, 1998: 3)
(15) Since the hyperkinetic Lebanese-Brazilian-Frenchman took over a near-
bankrupt Nissan in 1999, the carmaker has gone from roadkill to profit
machine in almost record time. (Dawson, 2003: 52)
Regarding the metaphors performing the generic function, only a small propor-
tion of them was marked (Graph 1). Examples (16) and (17) use the vehicle term
“animal”. The first metaphor (Example 16) is anticipated by two markers: “like”
from the category of copular similes, and “metaphor” from the category of ex-
plicit markers. The metaphor in Example (17), in turn, describing multinational
companies as different animals, is not signalled or anticipated by any of Goatly’s
(1997) markers.
(16) As that generation passed like a large animal through the digestive tract of
the American economy (Mr Wood’s herpetological metaphor), its numbers
and sheer exuberance did much to drive America’s consumption patterns
and even its equity markets – at least until the bust at the end of the 1990s.
(Ziegler, 2003: 3)
(17) Statistically, multinational companies play much the same part in the world
economy as they did in 1913. But they have become very different animals.
Multinationals in 1913 were domestic firms with subsidiaries abroad, each
of them self-contained, in charge of a politically defined territory, and highly
autonomous. Multinationals now tend to be organised globally along product
or service lines. (Drucker, 2001: 3)
Finally, Examples (18) and (19) are two unmarked metaphors that fill termi-
nological gaps. The first one (18) with the vehicle “bear” in “bear market”, is a
business term for the market losing value. The second metaphor (19) includes
“grow”, an economic term that refers to changes measurable according to spe-
cific economic parameters.
(18) Finally, the relentless bear market has hurt Suez because it still carries
plenty of equity holdings on its books, including big blocks of insurer Axa
and oil giant Total Fina Elf. (Rossant, 2003: 54)
(19) And he’s working to persuade them that they can make money in Russia,
whose economy has grown strongly in the last two years after a decade-long
depression. (Starobin, Belton & Crock, 2001: 68)
The quantitative data presented in this study show that contrary to previous ex-
pectations, not all active metaphors were marked, and only a few inactive meta-
phors were anticipated on the text level. This implies that metaphorical marking
in the corpus examined could have contextual motivations. The findings from
the metaphor functional analysis and the study of metaphor co-text provide evi-
dence for the correlations between the metaphor functions and its signalling in
the corpus examined.
5. Conclusions
The results obtained in this study suggest that metaphor marking could be cor-
related with the range of purposes for which metaphors are used in a particular
discourse, and so in a specific context. The analysis of metaphor functional types
and of the presence of metaphorical marking (Goatly, 1997) in a corpus of busi-
ness periodicals showed that when metaphors are used for illustrating purposes,
there is a significant tendency for the business article authors to explicitly signal
such language uses. A correlation like this was notably less pronounced for the
metaphors used with more generic functions, not specifically related to the genre
of business periodical article and its communicative context. Finally, no mark-
ing was found for the metaphorically originated business terms. This absence can
also be contextually motivated given the level of specialist knowledge required
to communicate and interpret current business information. The data presented
confirm Cameron and Deignan’s (2003) view on the use of metaphor tuning de-
vices in oral discourse. However, no sufficient evidence was found in this study to
argue that metaphorical marking depends on the type of metaphor used (conven-
tional or unconventional).
The figures presented here should, nonetheless, be treated with caution. The
process of metaphor identification is always subjective to a certain degree, and
this may depend on a number of factors, such as the metaphor definition used
or the methodology applied. Clearly, the metaphorical material examined here,
which was taken from the previous study, is limited just to the key vehicle terms
identified basically in a sample corpus, and then searched for in the main cor-
pus. The figures mentioned are representative for the corpus analysed, but should
be carefully extrapolated to other types of discourse. Finally, the investigation of
the use of metaphorical markers was limited to the words and expressions from
Goatly’s (1997) list, which, however, is convenient for concordance-based queries
of electronic corpora.
Further studies of metaphor surrounding on the text level might adopt a dif-
ferent focus from that used in the present study. Instead of a top-down approach,
where the anticipating words and phrases are searched for with a concordancer, a
combination of both software and then a hand-searched examination of a corpus
could be considered. Any account of possible words and phrases signalling linguistic
metaphors ought to be based on an in-depth analysis of a corpus and its context.
As metaphorical marking seems to be an extremely volatile linguistic phe-
nomenon, with changeable behaviour in different types of text and context, it is
suggested that studies of metaphor linguistic anticipation should be conducted
on an ad hoc basis rather than in terms of definite and universal categorisations
and inventories.
References
A connected world. (1997). Economist, 344 (8034), 3–4.

Behar, Richard (2003). The shabby side of the street. Fortune, 147 (4), 137–140.
Bremner, Brian & Kerry Capell (1997). Getting set for Japan’s rebound. Business Week, 3543,
110–113.
Cameron, Lynne & Alice Deignan (2003). Combining large and small corpora to investigate
tuning devices around metaphor in spoken discourse. Metaphor and Symbol, 18 (3),
149–160.
Carson, Iain (1998). The world as a single machine. The Economist, 347 (8073), 3–5.
Clifford, Lee (2000). My stocks are up 10,000%. Fortune, 142 (14), 156–164.
Collin, Peter (2001). Dictionary of business. London: Bloomsbury.
Dawson, Chester (2003). On your marks… Business Week, 3824, 52–53.
Drucker, Peter (2001). The next society. The Economist, 361 (8246), 3–5.
Edwards, Ben (1998). Capitals of capital. The Economist, 347 (8067), 3–5.
Glucksberg, Sam & Boaz Keysar (1993). How metaphors work. In A. Ortony (Ed.), Metaphor
and thought (2nd ed.) (401–424). Cambridge, UK: Cambridge University Press.

Henderson, Willie (1986). Metaphor in economics. In M. Coulthard (Ed.), Talking about text
(109–127). Birmingham, UK: University of Birmingham.
Laderman, Jeffrey, M. & Geoffrey Smith (1999). The best mutual funds. Business Week, 3614,
72–78.
Lindstromberg, Seth (1991). Metaphor and ESP: A ghost in the machine. English for Specific
Purposes, 10, 207–225.
Mandel, Michael J. (1998). Why the pace has to pick up. Business Week, 3593, 136–139.
NaLamlieng, Chumal (29/11/1999). The problems have to be solved one at a time. Business Week.
Retrieved July 8, 2010, from http://www.businessweek.com/archives/1999/b3657038.arc.
htm?campaign_id=search
Pennar, Karen (1998). Two steps forward, one step back. Business Week, 3593, 116–119.
Rossant, John (2003). That sinking feeling at Suez. Business Week, 3824, 54–54.
Rundell, Michael & Gwyneth Fox (Eds.) (2002). Macmillan English dictionary for advanced
learners. Oxford: Macmillan Publishers Ltd.
Schwarts, Nelson D. & Jeremy Kahn (2002). Can this bull run again? Fortune, 146 (13), 68–82.
Scott, Michael (2004). WordSmith Tools, version 4. Oxford: Oxford University Press.
Skorczynska, Hanna & Alice Deignan (2006). Readership and purpose in the choice of eco-
nomics metaphors. Metaphor and Symbol, 21 (2), 87–104.
Skorczynska, Hanna & Jordi Piqué Angordans (2005). A corpus-based description of meta-
phorical marking patterns in scientific and popular business discourse. Metaphorik.de, 9,
112–129.
Starobin, Paul, Catherine Belton, & Stan Crock (2001). Putin’s Russia. Business Week, 3757,
66–73.
Woodall, Pam (2002). The unfinished recession. The Economist, 364 (8292), 3–5.
Ziegler, Dominic (2003). The weakest link. The Economist, 366 (8310), 3–5.
chapter 16
Critical analysis of creative

metaphors in political speeches
Ralph Mueller
University of Fribourg
Creative metaphors play, in many respects, an important role in political dis-

course; however, there is still a lack of categories to identify and analyse creative
metaphors on a textual level. One precondition for analysing creativity in a non-
literary discourse such as politics is to understand that creativity exists rather
along clines and, therefore, needs to be redefined according to its context. As a
consequence, this study understands creative metaphors as expressions which
draw the attention to their metaphoricity by deviating from conventional ways
of expressing things within a particular discourse or genre (e.g. political speech-
es). In addition, an analysis of the metaphor “centre of gravity” in a speech of
the German foreign minister Joschka Fischer will illustrate how corpora can
help us to evaluate metaphorical expressions within this framework.
Keywords: metaphor, creativity, political discourse, European discourse, critical

discourse analysis
1. Introduction
What is a “creative metaphor”? So far, there is no definitive answer to this question

as it demands to some extent an act of (subjective) evaluation on the basis of a not-
clearly-defined concept. Nevertheless, individual metaphor analysts should expli-
cate criteria according to which they judge a metaphor as “creative”, even if there is
no generally accepted definition of the attribute “creative”. As a consequence, it is
necessary to develop – apart from the issue of reliable metaphor identification (cf.
Gibbs, in this volume) – additional categories of metaphor classification.
The ‘chaotic’ category “creative metaphor” includes typically literary exam-
ples such as “novel metaphors” and “poetic metaphors”. The crucial question is,
however, whether it is possible to extend this category beyond literary examples.
If this is possible – and an examination of the terms “creative” and “creativity”
322 Ralph Mueller
suggests that it is – then it will be possible to find many creative metaphors

even in political discourse which is, actually, not very famous for being creative.
Moreover, as a discussion of the metaphor “gravitational centre” will illustrate
(this metaphor had been used prominently in a political speech by the German
foreign minister Joschka Fischer in 2000), “creative” is a useful term to assess the
effect of certain metaphors.
2. “Creative” and “creativity”
The term “creative” can be used to describe both a product, such as a painting
or a text, or an individual, like Leonardo da Vinci. At the same time, it can illus-
trate a process, as it is expressed in the concept of “creative thinking” (Groeben,
2001). Moreover, creativity can be discerned according to different levels of qual-
ity, since it has been attributed to geniuses, but also to the particular behaviour
of children and animals.
The extension of the use of “creative” beyond artistic work is a terminological
development of the 20th century, when “creativity” became a buzzword to explain
how our civilisation can meet the challenge of accelerating changes (cf. Pope,
2005: 19). Such challenges are not only restricted to scientific and artistic prog-
ress; they also form part of one’s everyday experience. In this respect, “creative”
can refer to a general human faculty of producing something new. For instance,
“creative” does not only include the competence of producing ever new sentences
or metaphors (cf. Pope, 2005: 55), but also the ability to deviate creatively in ev-
eryday discourse (Carter, 2004).
This extension of the meaning of the term “creative” can be justified by the
similarity of processes that occur within the creative individuals. Nevertheless,
the difference between everyday and exceptional creativity, for instance, between
everyday talk and a Shakespearean sonnet, demands terminological choices or
decisions. As a consequence, metaphor analysts need to clarify by what standards
they evaluate creativity.
3. Creativity and political speeches
The standards of creativity are often determined in relation to venerable terms of

aesthetics such as “originality” and “invention” (cf. Carter, 2004: 27), which have
been associated with the creations of geniuses in the sciences and arts of the past.
This may seem at odds with an analysis of creative metaphors in political speeches,
since political texts are less typical examples of creativity than literary texts. There
Chapter 16. Creative metaphors in political speeches 323
are first of all differences in the presentation, as political speeches are typically
a spoken genre with one single performance, whereas literature is usually writ-
ten and printed. Moreover, literature, as an aesthetic artefact, is often expected
to be free of particular real-world interests (cf. e.g. the Kantian understanding of
“aesthetic”, Kant, 2000: 96, §6). Political speeches violate this expectation, as they
(typically) strive for immediate practical or social impact. In fact, creativity in poli-
tics has often been seen on a social dimension, as creativity may help to adapt to
changes in social practice (cf. Fairclough, 1989: 170–196). In this respect, creative
metaphors may induce new conceptualisations of problems (e.g. Chilton, 1996)
and play, therefore, a role in the history of political ideas (Drulák, 2005).
However, structuralist approaches have developed a more differentiated
understanding of the relationship between practical and aesthetic functions by
pointing out that aesthetic (or poetic) functions and practical functions are not
mutually exclusive (Jakobson, 1960; Mukařovský, 1977). In this view, literature is
rarely devoid of practical functions (Mukařovský, 1977: 69). In the same way, non-
literary discourses may contain aesthetical functions (e.g. everyday conversation,
cf. Carter, 2004).
Mukařovský’s account of the aesthetic function in terms of a theory of de-
viation is very useful in defining a creative metaphor in different discourses.
Mukařovský understood the aesthetical function as the function by which the
attention of the addressee is drawn away from the practical function to the sign
itself (cf. Mukařovský, 1977). Typically, this happens if certain phrasing devi-
ates from conventional linguistic patterns. If we understand conventionality in
terms of expectations within a given discourse or genre, we can define “creative
metaphors” as expressions which draw attention to their metaphoricity because
they deviate creatively from conventional ways of expressing things and thoughts
within a particular discourse or genre.
However, the expectations from which an utterance may deviate in politi-
cal discourse are quite different from expectations in literature. In fact, analysing
creativity in political discourse presupposes a broader understanding of creativ-
ity than in literary discourse. As mentioned above, stylistic deviation is defined
with respect to different standards, and these standards should be formulated ac-
cording to the language used in a particular genre. This has to take into account
that creative metaphors serve a practical function at the same time. For instance,
politicians often use creative metaphors to persuade, maybe even for doubtful
purposes. A “good” creative metaphor is, therefore, not necessarily a “good” meta-
phor by moral standards.
It is in accordance with research on everyday creativity (cf. Pope, 2005: 52–59)
if we define creativity according to conventions that are relevant to a non-literary
discourse or non-literary genre. In addition, it also reflects the importance of the
324 Ralph Mueller
context of production, since even exceptionally creative individuals, processes

and products should be examined in their context, as creativity has to be accepted
by the social environment. It does not suffice if an individual considers his act as
creative; it is the social environment – and in particular what Csikszentmihalyi
(1996) calls the “experts in the field” – that distinguish between madness, incom-
petence and creativity.
Although Booth (1978) proposed a programme for a metaphor critique by
metaphor scholars, we do not necessarily function as experts in the sense of
Csikszentmihalyi. This kind of evaluation is usually provided by the discourse
participants themselves. Nevertheless, we can try to investigate the use and evalu-
ation of creative metaphors. One possibility is to compare the use of a particular
metaphor with its use in large text-corpora.
First of all, corpora can help us to determine what is “not-creative” within a
particular discourse or genre. By doing this, we can find the backdrop of stereo-
typical conceptualisations and linguistic routine against which creative deviation
may be evaluated. The analysis of large corpora allows here further insights, since
it can provide an impression of both the most frequent used metaphors and their
most recurring phrasings. For instance, the German phrase (word-cluster) auf
den Weg zu bringen (‘to put something on track’) occurs verbatim 46 times in the
protocols of the German Bundestag in the year 2001. The use of such an expres-
sion is not creative. In fact, this word-cluster provides an example of the use of
conventional journey metaphors in German politics.
Additionally, we can aim for a “positive list” of criteria that allows the detec-
tion of creative metaphors (cf. Müller, 2005). In this respect, it is important to
keep in mind that deviation can be established in many ways. For instance, a
metaphor may deviate from a pattern which is established within a discourse or
a transient pattern within a single text. Therefore, it is necessary to take several
levels of analysis into account:
– co-textual: the immediate text surroundings may highlight the metaphor; e.g.
metaphor interplays with other metaphors (Goatly, 1997: 255–282), or a met-
aphor is integrated within a figure of speech (parallelism, alliteration etc.);
– textual: a metaphor is highlighted because of its role in an overall text-
structure;
– context: contextual information foregrounds a metaphor; e.g. topic-trig-
gered metaphors, or the use of metaphor in a contradictory context (cf.
Gibbs, this volume);
– conceptual: elaboration, extension and composing of mappings (cf. Lakoff &
Turner, 1989: 67–72); novel and contradictory mappings;
– discourse: novel metaphor; intertextuality, extension, elaboration.
These levels of analysis do not actually describe an exact procedure of identifica-

tion of creative metaphors, nor are they clearly differentiated from each other.
They represent, rather, a way of grouping information, as information from all
levels may be important to assess the creativity of a metaphor. Therefore, the lev-
els will be illustrated by an example from a speech by Joschka Fischer, the former
German minister of foreign affairs, on the future of the European integration.
4. The gravitational centre of the European Integration
Fischer gave his speech at Humboldt University in Berlin in May 2000, and it
had the provoking title Vom Staatenverbund zur Föderation – Gedanken über
die Finalität der europäischen Integration; the official English translation: From
Confederacy to Confederation – Thoughts on the Finality of European Integration
(Fischer, 2000).
Fischer’s thoughts on the finality of the European integration – although he
emphasised that his vision of Europe’s future was a very personal one – triggered
many reactions across Europe (cf. Winkler, 2005). For instance, French president
Jacques Chirac responded to it in his speech “Our Europe” in Berlin one month
later (Chirac, 2000). Tony Blair made comments on Fischer’s issues in a speech
made in Poland five months later (Blair, 2000).
Concerning metaphors, Fischer’s speech contains mostly metaphors from the
source domain “journey”. This starts most prominently with Fischer’s rhetorical
question “Quo vadis Europa?” and continues to the answer “onwards to the com-
pletion of European integration”.
Whereas the first part of Fischer’s speech was mostly devoted to relatively
uncontroversial statements about the importance of the enlargement of the EU
and a parallel reform of its institutions, it was the second and last part that dealt
with the so-called “finality of Europe”. It was, in particular, Fischer’s metaphor of
the “centre of gravity” that triggered interest:
(1) Ein solches Gravitationszentrum müsste die Avantgarde, die Lokomotive für
die Vollendung der politischen Integration sein und bereits alle Elemente der
späteren Föderation umfassen.
‘Such a centre of gravity would have to be the avant-garde, the driving force
[Lokomotive] for the completion of political integration and should from the
start comprise all the elements of the future federation.’ (Fischer, 2000)
The metaphor “gravitational centre” occurs in variation twelve times in Fischer’s

speech (including Gravitation, Gravitationskern and Integrationsmagnet).
326 Ralph Mueller
There are only a few particular elements on both the textual and contextual level
which may have helped to foreground this metaphor. However, the gravity meta-
phors are conspicuously concentrated in the last part of his speech. This arrangement
may provide the impression of a metaphorical summary of Fischer’s visions.
More revealing is the analysis from a conceptual point of view. A first rather
intuitive hypothesis could attribute a certain novelty to this metaphor. As there
were many reactions to Fischer’s speech across Europe, it could be inferred that
Fischer used this metaphor in a way that was new to the European discourse.
However, a more reliable assessment of the novelty of the metaphor requires a
comparison of Fischer’s metaphor with instances within a larger corpus.
A first analysis was carried out in the German COSMAS corpus, which is com-
posed of a variety of texts, but with a majority in newspapers of the last 15 years, and
which approached a billion words at the time when the analysis was carried out. The
concordances of this corpus established that Fischer’s metaphor of the “centre of
gravity” was not new. However, the dominance of newspaper articles in the corpus
had an immediate influence on the concordances: all 111 instances of Gravitations-
zentrum were metaphorical. They occurred in a variety of contexts, the majority
of which were in politics (49 instances), followed by art (32). Most uses (62) were
related to geographical entities such as countries or cities, 19 described social rela-
tionships and 15 a work of art or an artistic performance.
Of the 49 hits in politics there were 37 that referred to international relations,
23 of which are directly linked to Fischer’s speech, which mirrors a certain echo
in the media, even if the number of hits has to be relativised by the fact that the
COSMAS corpus is unevenly distributed over the years. However, there were also
three hits that used the metaphor in the context of European policy in the early
1990s. Thus, the metaphorical expression “centre of gravity” had been used previ-
ously in the context of European integration.
Nevertheless, the metaphor’s reference to a group of countries within the EU is
new, which indicates a high relevance of contextual reference for the perceived nov-
elty of a metaphor. Moreover, the combination of the centre-of-gravity metaphor
with journey metaphors (avant-garde, locomotive) had not been found elsewhere.
5. Metaphor and genres
The COSMAS corpus contains mostly texts from journalistic genres. Therefore,
it should rather be used for contrastive comparisons between journalistic genres
and political speeches. In fact, the list of (newspaper) concordances helps to find
further characteristics of Fischer’s use of the metaphor. It is, in particular, note-
worthy that Fischer avoids designating the target domain of his metaphor with
adjectives, genitive-constructions or appositions (e.g. ‘European centre of gravity’

or ‘gravitational core of Europe’), whereas such uses can be found in three quar-
ters of all journalistic concordances.
As the concordances from the COSMAS corpus turn out to be quite different
from Fischer’s use of the metaphor, a second corpus analysis was carried out with
another corpus containing the complete protocols form the German Bundestag
(1996–2006), the Austrian National Council (1997–2006) and the Swiss National
Council (1995–2006), which offer a large collection of contemporary political
speeches and provide a better assessment of the creativity of Fischer’s speech.
Fischer’s speech had a strong effect in this corpus of parliamentary debates,
as most instances occurred during a Question Time in the German Bundestag on
Fischer’s speech (19/05/2000). This importance in a large corpus of political speech-
es suggests that Fischer’s metaphor of the centre of gravity was indeed creative.
By analysing the reaction during the question time we can also confirm the
stylistic observations above: There are – as in Fischer’s speech – less explicit refer-
ences to the target domain (by an adjective or a genitive) in the political speeches
than in the newspapers. One possible explanation is that journalistic texts may
tend to be more explicit about the target domains of uncommon metaphors,
whereas speakers could assume more common ground in the unmediated com-
munication of a public speech.
A comparison of concordances from political speeches also shows that avoid-
ing explicit markers of the target domain is often paralleled by another stylistic
feature, the tendency to use a metaphor along with other metaphors. This kind
of listing of metaphors can be observed in Fischer’s speech where the “centre of
gravity” is explained as an “avant-garde”:
(2) […] wird eine kleinere Gruppe von Mitgliedstaaten als Avantgarde diesen
Weg vorausgehen, d.h. ein Gravitationszentrum aus einigen Staaten bilden,
die aus tiefer europäischer Überzeugung heraus bereit und in der Lage sind,
mit der politischen Integration voranzuschreiten?
‘[…] will a smaller group of member states take this route as an avant-garde,
i.e. will a centre of gravity emerge comprising a few member states which
are staunchly committed to the European ideal and are in a position to push
ahead with political integration?’
A few sentences further on we encounter even more metaphors (cf. Example 1

above): Fischer describes the centre of gravity as a nucleus, an avant-garde and a
driving force (Lokomotive in the German original).
The same feature can be found in the parliamentary debates about Fischer’s
speech, where alternative metaphors such as “core” and “nucleus” are repeated
several times:
328 Ralph Mueller
(3) “Gravitationszentrum”, “Kompetenzabgrenzung” und “Verfassungsvertrag”

sind Begriffe, die in der Union entwickelt worden sind und erfreulicherweise
bei Ihnen auftauchen. (Hintze, 2000: 14106)
‘‘Centre of gravity’, ‘delimitation of competences’ and ‘constitutional treaty’ are
terms that have been developed by the Conservative Party and it is fortunate
that they appear in your speech.’
(4) Schließlich besteht Konsens darüber, dass wir in der einen oder anderen Form
so etwas wie eine Avantgarde, ein Gravitationszentrum oder einen Kern
brauchen (Pflüger, 2000: 14106)
‘Finally, there is consensus that we need in some form an avant-garde, a centre
of gravity or a core.’
Fischer’s metaphor is aligned here with metaphors and other terms that originate
from an earlier Conservative white paper which had proposed a core Europe in the
mid 1990s (Schäuble & Lamers, 1994). Citing a metaphor this way could be seen as
a means of integrating a metaphor affirmatively in one’s own argument. However,
if we analyse these citations of Fischer’s metaphor in terms of Musolff ’s four basic
types of metaphor modification and negotiation (Musolff, 2004: 156), it becomes
clear that these re-uses of Fischer’s metaphor imply only a partial agreement.
Musolff points out that there are cases in which a metaphor yields conclu-
sions that another speaker wishes to reach as well. In this first category (1) the
original scenario may be enriched. However, with a growing disagreement the
metaphor is increasingly challenged and altered. There are also (2) partial agree-
ment and modification, (3) disagreement and modification, and (4) disagree-
ment and rejection.
Although Fischer’s metaphor found a lot of agreement in the question time,
these two concordances demonstrate nicely how it is possible to accept a certain
metaphor and, at the same time, suggest that Fischer’s conceptualisation was actu-
ally identical with the idea of a ‘core Europe’ as it had been expressed in the Con-
servative white paper by Schäuble and Lamers. Thus, these citations of Fischer’s
metaphor might be located somewhere between agreement and partial agreement
on Musolff ’s metaphor negotiations chart.
Of course, the evaluation of the metaphor by members of the political opposi-
tion is not really a reliable authority of evaluation. In fact, the Conservative Party’s
interpretation of Fischer’s metaphor is in contrast with the fact that Fischer’s speech
and the metaphor of the “centre of gravity” in connection with the metaphor of an
“avant-garde” triggered a broad debate, whereas the Conservative vision of a Euro-
pean core met a lot of criticism at this time. Moreover, the negotiation of Fischer’s
metaphor was, after all, a conspicuous feature of this question time.
6. Creative mappings
The fact that politicians seem to like to use metaphors along with other metaphors
and buzzwords reminds us that creativity can also emerge from the interaction of
metaphorical mappings within one text.
Lakoff and Turner, in their influential book on poetic metaphor, offer the
term of “composing mappings” – the “perhaps most powerful of all ways in which
poetic thought goes beyond the ordinary way we use conventional metaphoric
thought” – to analyse this use of metaphor (Lakoff & Turner, 1989). Lakoff and
Turner give this type a strong evaluative slant which may be appropriate for po-
etry. Nevertheless, it is an interesting type if we adapt it to the conventions of
political speeches, since it seems that combining different mappings is a recurrent
(and in the case of Fischer, also powerful) feature of political speeches.
However, the identification of mappings and the interpretation of mappings
is one of the more difficult aspects of a linguistic metaphor analysis. For instance,
Steen calls the step of “non-literal mapping identification” the “weakest of all”
(Steen, 1999: 68) and demands that this step be carried out under strict control.
Indeed, recent investigations of corpora suggest that is very difficult to establish
mappings when looking at realistic discourse material (Deignan, 2005; Semino,
Heywood, & Short, 2004).
If we look at the concordances from the COSMAS corpus and the corpus of
parliamentary speeches, there is not enough evidence to postulate a mapping of
the kind a political system is a gravitational system. First of all, the sys-
tematicity of the mapping around this metaphor remains rather vague. We can
find in the corpora references to magnets, force fields and attraction – which may
be considered as parts of the same mapping or a mapping extension – however,
there is only one instance (in the journalistic COSMAS corpus) with reference
to things such as planets, suns and galaxies (R99/Jun.44531) and one instance
with reference to a non-astronomical “big bang” (P94/MÄR.8581). Moreover,
the conceptual knowledge underlying such a mapping did not seem to be very
relevant in the discussion of Fischer’s metaphor. The two examples (3) and (4)
that tried to assimilate Fischer’s metaphor illustrate this, as they simply mix dif-
ferent metaphors alongside the gravity-metaphor without further elaboration or
questioning. Thus, if we look in the debate for clues on how to interpret these
uses of the same metaphor, we are only confronted with even more metaphors.
As a consequence, we have to rely mostly on other metaphors if we want to detect
what the users may have understood by a “centre of gravity”. However, compar-
ing a metaphor with earlier metaphors leaves the content of the conceptual sche-
mata rather vague.
330 Ralph Mueller
This particular use of this metaphor makes it difficult to assess to what ex-
tent Fischer’s metaphor may have a particularly striking effect, e.g. the effect of a
schema-refreshment that challenges and potentially refreshes the readers’ beliefs
or schemata (Semino, 1997: 197). However, it may indicate as well that analogies
between different mappings are as important as the analogies between the source
and target domains. This would suggest that new compositions of mappings could
develop many analogies between metaphors, and thereby have a strong effect of
schema-refreshment.
A comparison of these mappings is to some extent based on speculation, since
it involves at least two mapping interpretations. Therefore, only some preliminary
observations: On the one hand, “centre of gravity” seems to be less fixed and more
dynamic than a core. The combination of a journey metaphor with a gravitational
metaphor, although it is a mixed metaphor, works very well in ascribing a propel-
ling power to an abstract physical power. Fischer’s metaphor is in this respect
creative; one is almost tempted to call it a “good metaphor”.
7. A critical analysis of creative metaphors
However, looking only at the creativity aspect of a metaphor does not sufficiently as-
sess it such that it may be called a ‘good metaphor’. If an analysis of political speeches
puts too much emphasis on the aesthetic function, it may neglect important non-
aesthetical functions such as persuasive effects, and this could reduce the analysis
to a superficial examination of stylistic devices. Stylistic observations of particular
patterns and strategies may be a first step, and if they are based on a comparison of
additional concordances from a larger corpus, they may inform about particular
forms of deviation from conventional ways of expressing something. Moreover, in
accordance with Mukařovský (1977), we can expect that such creative deviations
will attract the attention of the audience (although they are not necessarily more
persuasive). However, in order to ensure a critical perspective, it is also necessary to
go beyond text- and corpus-analysis and include ethical considerations.
Such a combination of ethical and aesthetical evaluation is not uncontested.
On the one hand, stylisticians are sometimes reluctant to relate the text to its
(real-world) context. On the other hand, ethical and aesthetical comments are
often conflated. The ambiguity in the evaluative comment “This is a bad/good
metaphor!” is telling, since it may refer simultaneously to an aesthetic and a moral
judgement. Therefore, it is necessary to supplement the stylistic analysis of cre-
ativity – which per se does not and should not account for moral judgements – by
a critical analysis of metaphor (at least in the case of text types which aim at im-
mediate social or practical impact).
A good way of implementing an ethical dimension to a stylistic analysis is

achieved by looking at the implicit presuppositions of a metaphorical expres-
sion (the “vehicle”) in its source domain and their import on “social actors” (van
Leeuwen, 1995) in the target domain. For instance, Fischer’s linguistic realisation
of the mapping conceals that a gravitational centre implies a periphery on which
the gravitational centre exerts its power. Moreover, a centre of gravity is propelled
by the laws of nature and not by political subjects; consequently, it presents Euro-
pean integration as inevitable as an apple falling to the ground.
In summary, Joschka Fischer has provided us with an interesting metaphor to
illustrate the ambiguity of the evaluative statement “X is a good/bad metaphor”.
A stylistic analysis (assisted by corpus-stylistics) can show that Fischer has in-
troduced a new way of conceptualizing the process of the European integration
by applying an established metaphor to a new problem. On a discursive level,
his metaphor has triggered a broad debate, which indicates that he was also rhe-
torically successful. However, the comparison with other metaphors also shows
some aspects that have to be analysed critically, and we need a critical analysis of
metaphor in order to establish this. In this respect, Joschka Fischer’s metaphor
and its reception in politics may show that stylistic analysis and critical metaphor
analysis necessarily complement one another.
References
Blair, Tony (2000). Prime Minister’s speech to the Polish Stock Exchange [6/10/2000]. The of-
ficial site of the Prime Minster’s Office. Retrieved April 4, 2006, from http://www.number10.
gov.uk/Page3384.
Booth, Wayne C. (1978). Metaphor as rhetoric. The problem of evaluation. In S. Sacks (Ed.), On
metaphor (47–70). Chicago: University of Chicago Press.
Carter, Ronald (2004). Language and creativity. The art of common talk. London: Routledge.
Chilton, Paul (1996). Security metaphors. Cold War discourse from containment to common
house. New York: Lang.
Chirac, Jacques (2000). Notre Europe. Allocution prononcée par M. Jacques Chirac Président de
la République devant le Bundestag, Berlin 27/6/2000. Deutscher Bundestag, Retrieved March
29, 2006, from http://www.bundestag.de/geschichte/gastredner/chirac/chirac2.html.
COSMAS II (1999ff.). COSMAS II. Corpus search, management and analysis system (Institut für
Deutsche Sprache, Ed.). Mannheim. http://www.ids-mannheim.de/cosmas2/.
Csikszentmihalyi, Mihalyi (1996). Creativity: Flow and the psychology of discovery and inven-
tion. New York: Harper Collins.
Deignan, Alice (2005). Metaphor and corpus linguistics. Amsterdam: Benjamins.
Drulák, Peter (2005). Metaphors and creativity in international politics. Research Cluster ‘Dis-
course, Politics, Identity’ Working Paper No. 3. Retrieved January 6, 2006, from http://www.
lancs.ac.uk/ias/researchgroups/dpi/dpi.htm.
Fairclough, Norman (1989). Language and power. London: Longman.
332 Ralph Mueller
Fischer, Joschka (2000). Vom Staatenverbund zur Föderation – Gedanken über die Finalität
der europäischen Integration. Speech held at Humboldt University, Berlin 12/5/2000.
Retrieved August 3, 2005, from http://www.auswaertiges-amt.de/www/de/infoservice/
download/pdf/reden/2000/r000512a.pdf [English Translation: http://www.auswaertiges-
amt.de/www/de/infoservice/download/pdf/reden/redene/r000512b-r1008e.pdf].
Groeben, Norbert (2001). Literary creativity. In D. Schram & G. J. Steen (Eds), The psychology
and sociology of literature. In honor of Elrud Ibsch (17–34). Amsterdam: Benjamins.
Hintze, Peter (2000). Speech at the German Bundestag on May 19, 2000, Verhandlungen des
deutschen Bundestages. XIV. Wahlperiode, 106. Sitzung, 9973–9974.
Jakobson, Roman (1960). Closing statement. Linguistics and poetics. In T. A. Sebeok (Ed.),
Style in language (350–377). Cambridge, Mass.: Wiley & M.I.T.-Press.
Kant, Immanuel (2000). Critique of the power of judgment. In P. Guyer & A. W. Wood (Eds),
The Cambridge edition of the works of Immanuel Kant (Vol. 5). Cambridge: Cambridge
Univ. Press.
Mukařovský, Jan (1977). Two studies of poetic designation: Poetic designation and the aesthetic
function of language [1938]. In J. Burbank & P. Steiner (Eds), The word and verbal art.
Selected essays by Jan Mukařovský (65–73). New Haven: Yale University Press.
Müller, Ralph (2005). Creative metaphors in political discourse. An analysis of Swiss speeches.
metaphorik.de, 9, 53–73.
Musolff, Andreas (2004). Metaphor and political discourse. Analogical reasoning in debates about
Europe. Basingstoke: Palgrave Macmillan.
Pflüger, Friedbert (2000). Speech in the German Bundestag Mai 19, 2000, Verhandlungen des
deutschen Bundestages. XIV. Wahlperiode, 106. Sitzung, 9986–9987.
Pope, Rob (2005). Creativity: Theory, history and practice. London: Routledge.
Schäuble, Wolfgang & Karl Lamers (1994). Überlegungen zur europäischen Politik (Schäuble-
Lamers-Papier I). Retrieved April 2, 2006, from http://www.wolfgang-schaeuble.de/posi-
tionspapiere/schaeublelamers94.pdf
Semino, Elena (1997). Language and world creation in poems and other texts. London: Longman.
Semino, Elena, John Heywood, & Mick Short (2004). Methodological problems in the analysis of
metaphors in a corpus of conversations about cancer. Journal of Pragmatics, 36, 1271–1294.
Steen, Gerard (1999). From linguistic to conceptual metaphor in five steps. In R. W. Gibbs &
G. J. Steen (Eds.), Metaphor in cognitive linguistics. Selected papers from the Fifth Interna-
tional Cognitive Linguistics Conference, Amsterdam 1997 (57–77). Amsterdam & Philadel-
phia: John Benjamins.
van Leeuwen, Theo (1995). Representing social action. Discourse & Society, 6, 81–10.
Winkler, Heinrich A. (2005). Integration oder Erosion. Joschka Fischers “Humboldt-Rede”.
Absicht und Wirkung. In R. Hohls, I. Schröder & H. Siegrist (Eds.), Europa und die Eu-
ropäer. Quellen und Essays zur modernen europäischen Geschichte. Festschrift Hartmut
Kaelble (469–473). Stuttgart: Steiner.
chapter 17
Metaphor in physical-and-speech action expressions
Lynne Cameron
Open University, UK
In this chapter, I discuss a phenomenon observed on the shifting borders of

metaphor that I have called ‘physical-and-speech action expressions’. These
expressions combine a physical action verb with a prepositional particle and a
speech action verb: stand up and declare, turn round and say, sit down and talk,
and the physical action verb metaphorically carries attitude and value in respect
of the speech action. I present empirical findings around two such expres-
sions, using a methodology that combines small and large corpora, and offer
a theoretical explanation, using complex dynamic systems theory and recent
developments in grounded cognition to integrate the social, cognitive, linguistic
and affective.
Keywords: metaphor, physical action, verbs, dynamic systems, affect
Background
In my work, I explore how people use metaphor in spontaneous talk in real-world

contexts. Consider this piece of talk:
that is why it’s so important to build this bridge with you
(Jo Berry, in conversation with Pat Magee, 2001)
The speaker here is not talking about literal bridge building, but uses the phrase
build this bridge to talk metaphorically about the process of meeting a former
enemy and reaching some kind of reconciliation with him. The metaphorically-
used language is produced in the stream of talk, with the vehicle or source domain
term (build this bridge)1 embedded in and surrounded by talk about the topic
of the metaphor. Through a series of empirical studies (Cameron, 2003, 2007a,
2007b, 2008; Cameron & Stelma, 2004; Cameron & Deignan, 2006), I have come
to understand something of the nature of metaphor in spontaneous conversa-
tion – how the vehicle term usually comes ‘on its own’, rather than linked to an
334 Lynne Cameron
explicit topic term (in contrast to the nominal form of literary metaphors such
as Juliet is the sun, or conceptual metaphors in the form of a is b), how it often
comes as a verb rather than as a noun, how it dialogically shifts and changes shape
as talk proceeds, how difficult sections of interaction produce clusters of meta-
phors – both little tight clusters and longer, more spread out clusters, and how
most metaphorical talk makes use of more or less conventionalised phrases.
In constructing a theoretical framework for metaphor in talk, such features
are in need of explanation (Gibbs, this volume). In order to describe and explain
metaphor use, I have developed a “discourse dynamics approach” that combines
complex dynamic systems theory with aspects of cognitive linguistics and of socio-
cultural theory. Here, I will summarise relevant aspects of the theoretical frame-
work, referring readers to other sources for fuller explanation (Cameron, 2007b;
Cameron & Maslen, 2010; Cameron, Maslen, Todd, Maule, Stratton & Stanley,
2009; Gibbs & Cameron, 2008; Larsen-Freeman & Cameron, 2008; the metaphor
analysis website: http://creet.open.ac.uk/projects/metaphor-analysis/index.cfm).
The dialogic and social
I take metaphor in face-to-face talk as the most basic kind of ‘use’ and the most
appropriate on which to build theory and methodology (following Clark, 1996,
writing about language use more generally). In face-to-face talk, metaphor serves
the purposes of social interaction between people. As people talk together, they
express their ideas and feelings in words ‘designed for the other’. People take ac-
count of their interlocutors in choosing verbal formulations, and in accompany-
ing paralinguistic activity. People interpret what they hear and see in the light of
what they know about the other person. Talk is dialogic in this sense of involving
the other (Bakhtin, 1981; Linnell, 1998). It is social in a broader sense too, in that
people’s socio-cultural group memberships influence how they think and talk,
while each instance of talk by individuals can affect the nature and language of
social groups.
Metaphor theory has, to date, tended to ignore social and dialogic factors and
look through a monologic lens. Several reasons can be found for this individualist
stance: (a) classical metaphor theory (going back to Black (1962) and Richards
(1936)) was rooted in literary theory and worked with texts rather than with in-
teractional data, and (b) more recent cognitive theory came out of the individu-
alist generative model of language in US (e.g. Lakoff, 1993), and is connected
into the wider field of cognitive psychology where again a monologic model has
predominated (but see Pickering & Garrod, 2004). To understand and explain
metaphor in real-world contexts, we need to incorporate social interactional dy-
namics into metaphor theory.
Chapter 17. Metaphor in physical-and-speech action expressions 335
The dynamics of talk
Talk is seen as a dynamic system, as is the use of metaphor within talk. To describe
something as a ‘dynamic system’ is to highlight its continual changing nature and
the interconnectedness of its elements. In a paper with Gibbs on “the social cogni-
tive dynamics of metaphor performance”, we set out our view on the factors that
influence metaphor use as follows:
… metaphor performance is shaped by discourse processes that operate in a
continual dynamic interaction between individual cognition and the social and
physical environment. …
From a complex dynamic systems view, metaphor operates on multiple scales and
levels:
We argue that the complexities of metaphoric language use (i.e., how people co-
ordinate with each other through metaphor) emerge from self-organizational
processes that operate along a range of different time-scales, from the millisec-
ond to the evolutionary, and across a range of scales of social group size, from
the individual and dyad to the speech community. The phenomena of metaphor
performance are, we suggest, best studied in terms of continuously dynamic dis-
course processes. (Gibbs & Cameron, 2008: 65)
Adopting a dynamics systems framework prompts the researcher to recast ‘ob-

jects of concern’ into systems in continual flux. Data is then seen in terms of tra-
jectories of these systems across evolving state space landscapes. A conversation is
seen as a complex dynamic system, with a trajectory over time across a landscape
that represents all interactional possibilities between the speakers. Drawing on
complex dynamic systems, we look for self-organisation in the system of the talk
and the emergence of new phenomena from the micro-level of seconds of talk on
to higher social levels or timescales, as system trajectories move into stable pat-
terns, into attractors in their state space landscapes.
The verbal and linguistic
The third strand of the theoretical framework needed for our purposes here is lan-
guage use, the verbal expression of ideas and feelings in which metaphor occurs.
In contrast to strong cognitive views of metaphor as conceptual, I hold that meta-
phor is also intricately and inseparably linguistic in nature. The fact that we talk,
the way that our bodies permit us to talk, the role of talking in human lives – all
these insist on a role for language in how metaphor is and what metaphor does.
Language and cognition interact in human action, and so need to interact in the
theory we develop to explain it.
336 Lynne Cameron
Together with corpus linguist Alice Deignan (Cameron & Deignan, 2006), I
have begun to develop the notion of ‘metaphoreme’ to connect, within a complex
dynamics account, language and thinking. It also serves to connect individual lan-
guage use on the microgenetic (i.e. immediate, second-by-second) timescale with
social conventions of language use on the longer timescale of months, years, de-
cades and longer. The metaphoreme was posited as a bundle of lexico-grammati-
cal, cognitive, semantic, pragmatic and affective features around a phrase that has
metaphorical meaning, and that has emerged over time from discourse. Examples
include <emotional baggage> and <(not) walk away from>2 (Cameron & Deignan,
2006). Metaphoremes are held to reflect stabilities in metaphorical language use
that assist and constrain production and interpretation while allowing some vari-
ability in use. In this paper, I use the construct of metaphoreme as part of the ex-
planatory framework, and critically examine its adequacy for the task.
The shifting nature of metaphor in talk
When we track linguistic metaphor through conversations, we are forced to notice

the blurring or continuity of metaphor. Metaphor interacts with non-metaphorical
uses of language; the same lexical items are used with different degrees of meta-
phoricity (Cameron, 1999); metaphor overlaps and intersects with metonymy,
with hyperbole, with irony and other tropes (Kimmel, 2004). The amorphous
and shifting nature of metaphor in use seems to be the norm in talk and should
therefore be placed at the core of theory-building, in a reversal of earlier theories
that have centred on the highly unusual strong metaphors, leaving the blurred,
metaphor-like uses of language at the borders and edges. In the next sections, we
meet an example of blurred and metaphor-like language and attempt a discourse
dynamics explanation of its use.
Physical-and-speech action expressions: (1) sit down and talk
The particular type of linguistic expression to be considered here is compound

verb phrases that combine a physical action with a speech action, such as sit
down and talk and turn round and say. In the microgenetic dynamics of talk,
these physical-and-speech action expressions are used with meanings that occupy
a fuzzy space around metaphor, being sometimes metaphorical, and sometimes
literal but with a degree of figurative meaning. Their shifting metaphoricity and
the symbolising of most basic physical actions make them, I suggest, particularly
subtle and useful tools for human communication.
The first expression I want to examine is sit down and talk, firstly in a spe-
cific context of reconciliation conversations, and then in its use more broadly.
We begin with the use of the expression in a series of conversations between Jo
Berry and Pat Magee.3 Jo Berry’s father was killed in 1984 by a bomb planted
by Pat Magee on behalf of the Irish Republican Army (IRA). Fifteen years later,
after Pat Magee had served a prison sentence and been released under the peace
agreement, perpetrator and victim met up and talked about the bombing, its pre-
history and its aftermath. The first meeting between Pat Magee and Jo Berry, 15
years after the killing of Jo’s father, was highly charged and momentous for both
of them. What might seem a relatively insignificant aspect of their meeting – that
they sat down – becomes, as we will see, an important symbolic or figurative act
that carries import from the past into the future. When they come to their second
meeting a few weeks later, Pat Magee uses the verb sit twice in the first minute,4 in
lines 10 and 21 of Extract 1:
Extract 1
1-15 Pat you know it’s,
2 .. broken some sort of taboo here.
3 Jo … hmh
4 Pat and er --
5 it goes into that territory.
6 you know.
7 … of er --
8 there’s no right,
9 …(1.0) for me,
10 → … (2.0) to sit here and be forgiven.
11 … (2.0) if you understand me.
12 Jo hmh
13 Pat I mean,
14 in a sense there’s the political thing,
15 I knew what I was doing.
16 .. and I would even defend actions I’ve taken.
17 etcetera.
18 but,
19 when you come --
20 when it comes down to it,
21 → .. I am sitting with somebody who’s affected by,
22 … my actions.
23 …(1.0) and er --
24 …(2.0) that’s er --
25 …(2.0) there’s no preparation in the world for that.
338 Lynne Cameron
The physical act of sitting with someone who represents ‘the Other’ – Jo is both
a member of the societal group (UK governing class) who were previously the
‘enemy’ and individually a victim of his violent action – clearly affects Pat strong-
ly. The strength of its impact is both political, reflected in the breaking of a ta-
boo (line 2), and personal, reflected in the change of agency to I (in line 21, and
reached through several steps, from you in line 19 to it in line 20) and the argu-
ment of the verb with somebody who’s affected by my actions. A short time later,
Jo picks up the theme of sitting with a former enemy, connecting it to the journey
metaphor that she employs to talk about her process of making sense of her grief
after the bombing (Extract 2):
Extract 2
1 -88 Jo s- something I’ve wanted,
89 … almost since the moment the bomb went off.
90 ..I knew,
91 …(2.0) back in the moment,
92 wh- what I wanted to do,
93 … was bring as much --
94 …(2.0) something --
95 .. as much positive out of it as I could.
96 .. you know,
97 Pat [hmh]
98 Jo …(1.0) [and] I --
99 and I saw very clearly.
100 …(1.0) that the --
101 .. the end of that journey,
102 would be,
103 → .. sitting down and,
104 … talking to the people who did it.
The verb sitting, here collocated with down and with and talking, and used after
several metaphors relating to journeys, now takes on a metaphorical sense in ad-
dition to its physical, concrete, literal sense. Let’s examine the detail of this meta-
phorising process.
Extract 2 metaphorises the process or activity that Jo has undertaken since her
father’s death as a journey, with the bombing as the starting point. She describes
her purpose on this journey as to bring something positive out of it (lines 93–95),
turning the outcome of the process into some kind of concrete object (something)
that she is responsible for moving and bringing forth. In the last four lines, speak-
ing of the end of that journey focuses on that aspect of the journey domain and its
parallel in the topic domain, i.e. meeting with Pat. When Jo says would be sitting
down and talking with the people who did it, her talk does not move entirely into
the literal and physical (as would be the case if she had said, for example: the
end … would be meeting you in Ireland), since the phrase sitting down can apply
in both the metaphorical world (the vehicle domain) and the physical, concrete,
literal world of the topic domain. People at the end of a journey might sit down
for a rest, and people involved in reconciliation might sit down when they meet
each other (as in Extract 1). Through the use of the phrase, the metaphorical sce-
nario of the journey is collapsed into the physical, literal scenario in which Jo sits
down with Pat and talks with him. The use of sitting down blurs the distinction
between the vehicle domain and topic domain that had, up to this point, been
quite distinct.
Sitting down works as a “bridge” term between domains (Kittay, 1987: 166).
The bridging process has also been characterised as “literalization” of the meta-
phor (Goatly, 1997: 272), because the metaphorically-used expression is used in a
literal way. Literalisation is not a process of ‘making literal’, since the potential for
figurative meaning is not lost but is instead shifted, so that previously literal uses
become figurative or symbolic in a complementary process of “symbolization” of
the literal (Goatly, 1997: 278). Within fiction, when a metaphor vehicle is taken
into the literal world, it is often elaborated and extended, sometimes to the point
of exaggeration and hyperbole (Goatly, 1997). In this process, the metaphorical
meaning and the literal or more physical and concrete meaning are pushed apart,
rather than collapsing into a single sense.
In spontaneous talk, such effects cannot be deliberately constructed, but
something similar may happen nevertheless. Examination of the discourse dy-
namics of the expression sit down and talk (and its variants) reveals symbolizing
movement between metaphorical uses of language and non-metaphorical uses;
domains constructed in the talk are bridged by words or phrases that take the
language user from one to the other and back again, playing with the mind’s ten-
dency to understand through metaphor and through metonymy. We now look
more closely at this movement and play of metaphor and metonymy in sitting
down and talking, examining the microgenetic details of language use.
Across the dataset of two conversations and the radio interview, there are 28
uses of the lemma sit (extracted using Wordsmith Tools and shown as Appen-
dix 1). Analysis shows that:
– The lemma is used in the forms sitting (13 uses) and sit (15 uses).
– There is a clear tendency for Pat to use the verb (25 uses) rather than Jo (3 uses).
– There is a split between uses that refer to hypothetical situations, as in Ex-
tract 2: would be sitting down, (15 uses) and real or actual situations, as in
Extract 1: I am sitting with somebody… (13 uses).
340 Lynne Cameron
– Of the actual situations, 7 uses are marked for real world use by the collocate
here, one case by there, and one more specifically by in this wee kitchen.
– 15, i.e. more than half, of the verb phrases, include a modal verb such as
would or can.
The use of the verb becomes more interesting still when we examine its collocates,
in particular those that follow the verb:
– The verb is used as a phrasal verb sit down 11 times.
– Verbs of speaking are combined in various forms of sit (down) and
– The 10 verbs of speaking that co-occur with sit are as follows:
talk 5 uses
listen 3 uses
tell 1 use
exchange stories 1 use.
– The speaking verbs are usually joined to sit or sit down through parataxis us-
ing and.
Using the notion and notation of the metaphoreme, we can describe the preferred
form in this discourse as < sit (down) and talk>, where the round brackets indicate
a frequent but still optional part of the expression, and the triangular brackets
indicate a lemma form that is inflected in various, but limited, ways.
In positions after the verb sit (down), the concordance shows patterns of form
and meaning in which prepositional particles indicate, as well as relative orienta-
tion or position, affective relation or stance towards the people mentioned as the
prepositional complement. Three prepositional particles that follow instances of
sit (down) are shown below, together with the people who are referred to:
with range of people 8 uses
beside Jo 2 uses
across from ‘enemy’ 3 uses
The most neutral preposition in terms of attitudinal stance relative to the other per-
son or people is with, used by Jo and Pat to refer to each other, and by Pat to refer to
British soldiers. The use of with does not specify relative position but just proximity.
Across from is somewhat more oppositional, used by Pat when talking about the
British government and the Northern Ireland police force of the 1970s. Beside is
used to refer to Jo Berry and seems the least oppositional in both in locational terms
and in attitudinal stance. These prepositions work like bridging terms in their use
with sitting (down), bringing to the talk an affective perspective in which relative
position metaphorically refers to attitudinal perspective towards other people.
To explain how prepositions and prepositional particles work to symbolically

carry affect or attitude, we can make use of ideas about ‘simulation’ from ground-
ed cognition (Barsalou, 2008) and from work on somatic markers that integrates
emotions with cognitive processes (Damasio, 1994). Connections between meta-
phor and people’s physical interaction with the material world have been made
in cognitive metaphor theory (Gibbs, 2006a; Grady, 1999; Lakoff and Johnson,
1999), but we need to account for such connections in the moment of talking, in
situated discourse, as well as more universally or generally.6 Simulations are ‘as-if ’
activations in the brain of emotions, perceptions and actions that have come from
particular experiences. Damasio’s somatic marker hypothesis explains how emo-
tions come to be connected with experiences (Damasio, 1994). Through experi-
encing situations and related emotions, the brain builds memories with ‘somatic
markers’ that then simulate emotions in response to subsequent experiences of
similar situations. Because simulations are activities in the brain, they provide
a much quicker response than comes from actually experiencing the emotions
through the body. Actual bodily experiences, involving physiological action, such
as that around the production of hormones, take place much more slowly than
the ‘as if ’ brain activity.
The discovery of mirror neurons has helped to explain how we respond to the
physical actions of other people (Arbib, 2002; Barsalou, 2008). These neurons are
responsible for producing simulations in the brain of the actions of other people.
When bodily actions are observed, mirror neurons create simulations of those ac-
tions, while the brain also inhibits actual performance of the action in situations
where that is not required (Barsalou, 2008; Gibbs, 2006a, b). Observing others
produces imagined acting by the self.
The importance of simulations for discourse researchers lies in the fact that
simulations may be generated not just in response to a physical situation, percep-
tions or emotions but to mention of physical situations, perceptions and emo-
tions, in discourse. As Gibbs (2006b: 444) suggests ‘language use is closely tied to
embodied imagination’, and automatic simulation processes are likely to contrib-
ute to understanding linguistic metaphors encountered in discourse. Metaphors
that involve mention of physical action and/or affective force, may activate simu-
lations of physical action and associated emotional simulations. The specific dis-
course context serves to constrain physical response and support interpretation
(Gibbs, 2006 paper; Ritchie, 2006, 2010).
The symbolisation of <sit (down) and talk> can thus be explained as deriv-
ing from embodied, social and cultural experience in the physical world. Literal,
physical sitting down with someone
342 Lynne Cameron
a. is an intentional action,
b. is often face-to-face and unavoidably requires an acknowledgement of the
person you sit with,
c. who is a specific individual, not a stereotype or category, and
d. renders the participants vulnerable to each other, and thus requires a level of
trust.
Furthermore, sitting down physically contrasts with standing up, usually being
more comfortable and relaxed. In the concrete, literal world of northern Ireland,
sitting down contrasts strongly emotionally as well as physically with the up-
right, vertical, active stance of those engaged in conflict and seen on the streets of
northern Ireland, previously in action and still today in vivid murals painted on
the ends of terraces of houses (Rolston, 1992). Only at the end of conflict do these
men of action relax their guard and sit down. The socio-cultural and personal
experiences of sitting down and talking are likely to have built up connections in
memory between the physical action and emotions of trust and vulnerability, that
may be activated in simulations when the phrase is encountered in discourse.
Further examination of the physical-and-speech action expression sit down
and talk shows that it is often used metonymically in the reconciliation conversa-
tions, as when Pat says:
I felt obliged as a republican to sit down and talk about that
The physical action and the speech action are used metonymically to stand for the
larger activity of meeting the Other and explaining motives, listening and so on.7
In Extract 3, taken from the radio interview some two years after the first
conversations, metonymical sitting is accompanied by elaboration of detail as Pat
describes his first meeting with Jo. This extract includes the 24th and 25th uses
out of 28 of sitting verbs in the conversations over two and a half years, over which
time the figurative import of the verb in their discourse has been built up.
Extract 3
3-558 → Pat I’m sitting there
559 beside the woman
560 whose father I have killed
561 and at that time
562 → I was sitting in this wee kitchen
563 talking to this woman for the first time
564 whose father’s dead
In its first occurrence here (line 558), the mundane physical act of sitting is con-
trasted with the significance of the meeting through the direct and unmitigated
description of Jo as the woman whose father I have killed. The literalisation of sit-
ting here emphasises the unexpectedness and importance of the fact that it is Jo
Berry whom he is sitting beside. In the second occurrence (line 562), the repeated
literalisation is made still more striking by the additional detail of the homely
place where they sat, this wee kitchen. The bald whose father’s dead in the last line
again contrasts this significant personal detail with the apparently insignificant
wee kitchen. In these uses, sitting…talking comes to have symbolic meaning, part-
ly through the contrasts of the co-text and partly through the longer term shift-
ing and bridging in which <sit (down) and talk> has moved between metaphor,
metonymy and the physically literal.
In another development, <sit (down) and talk> comes to work as a potent
metonymy to describe reconciliation. The pronoun we in Extract 4 refers, not
to Jo and Pat, but to previously opposed groups, such as the IRA and British
government.
Extract 4
3-129 Pat after all these decades
130 we have arrived at a situation
131 where we can now sit down
132 across tables
133 and talk to each other
Summarising the chapter so far, to explain the complexity of an apparently simple

phrase sitting (down) and talking has required us to invoke processes at a range of
levels of language and thinking:
– bridging by terms in talk that facilitate on-line processes of literalisation and
symbolisation; at other times, the embodied simulations connecting action
and emotion serve to bridge between metaphorical vehicle or source domain
and what is being talked about.
– movement between the literal physical world, metaphor and metonymy,
which prompts processes of literalisation and symbolisation (talk is always
metonymic, sit down is sometimes metaphorical as well as metonymic).
– social dynamics that connect physical embodied experience to attitude and
affect through common memories and similar simulations, with associated
stabilised language forms or metaphoremes.
– microgenetic details of language choice connected to socio-cultural conven-
tions of language use through the metaphoreme.
This explanatory framework requires integration of the physical, social, cultural,
affective, cognitive and linguistic.8
344 Lynne Cameron
<sit (down) and talk> in a large corpus
Cameron and Deignan (2003, 2006) developed a methodology for working from
rich specific discourse data to large scale corpus investigation, thereby making a
dynamic systems connection between individual use in the moment of talk and
socio-cultural stabilised norms.
When we examine9 a large corpus of spoken English (the British informal
spoken corpus of the Bank of English, 15 million words) for the phrase sit down
and talk, we find that it has stabilised to express the same complex of meanings as
in the smaller dataset, positively valenced situations where speakers co-locate and
talk seriously and meaningfully together.
In the large corpus, hypothetical situations of use predominate. Of 65 uses of
the phrase in the corpus, only 8 refer to real world situations that actually exist;
the rest refer to actions that are hypothetical because they have not yet happened
or because they are suggestions or possibilities, such as:
Extract 5
… the best thing for you to do is sit down with your brother and talk with…
we don’t have pots of gold and we will sit down with the players and talk about
…
Here, the action sit down may be literal and/or metaphorical, whereas talk is met-
onymically used. We have a sense of a serious and well-intentioned act of commu-
nication, albeit with a degree of vagueness. The phrase provides only an outline of
the action, leaving details to be filled in by listeners, mentally or physically. It is at
least metonymical since the sitting down and talking stands for much more than
the literal physical and speech actions but, at the point of utterance, the degree
of symbolisation – how much more it is to stand for – is unknown and unknow-
able. The indeterminacy of symbolisation perhaps helps make this kind of phrase
particularly suited to hypothetical, future or indefinite contexts; the space beyond
the literal, the actual sitting and the actual talking, is left open for those who carry
out the action to determine.
When used to talk about real situations, the phrase is sometimes clearly used
metaphorically:
Extract 6
my mum cares about me so much in that way
…
But what does she talk to you every night about things like that
Yeah. well
Yeah
When I’m in bed we sit down and talk about all sorts of things
Sometimes it may be used both literally and metaphorically:

Extract 7
I can’t look round I can’t do what I want. So you think Well it’s not worth it. If
you want to go and visit friends I only go when [my disabled child’s] gone away
for respite
Mm
because I can’t sit down and talk to them you can’t have a good natter or whatever
because you’re always watching what he’s doing.
Because he likes a lot of electrical things obviously you see they’re breakable
In Extracts 6 and 7, the phrase sitting down and talking is given figurative meaning
(is ‘symbolised’) by the agents being engaged in similar physical actions. In Ex-
tract 6, being in bed precludes sitting down and so we are forced into a metaphori-
cal interpretation. In Extract 7, the speaker who visits friends can presumably en-
gage in the literal action of being seated and exchanging utterances, even though it
would be much less satisfying than the ‘real thing’. This context seems to highlight
the importance of the second part of the phrase; sitting down and talking includes a
substantial conversation, a good natter. In Extract 8, the medical context places the
emphasis on the sitting down, which is unusual for a busy specialist.
Extract 8
I went down to the hospital with my dad and a few days later he went and took
a colleague of his from this practice private practice with him who was a heart
specialist <tc text=pause> and actually sat down and talked to my dad about what
was happening and how why these diuretics to keep the blood thin and keep the
fluid down
It seems that there is a conventionalised stability about the form and affective
meaning of the phrase, but that there is also variability – in the degree of sym-
bolisation, in the relative importance of the physical sitting or the speech talking
action, in the prepositions used with the phrase.
The evidence from the larger corpus confirms the stabilised ‘shape’ of the
phrase as ‘metaphoreme’, as proposed by Cameron and Deignan (2006). We have,
however, broadened the type of figurative meaning that has stabilised around
the linguistic expression to include metonymy.10 Simulations that connect the
physical and the emotional add a cognitive dimension to the metaphoreme, sug-
gesting an explanation that connects stabilised language patterns with affect,
attitudes and values. We now move to the second physical-and-speech-action
expression to see whether further elaboration needs be added to this description
and explanation.
346 Lynne Cameron
Physical-and-speech action expressions: (2) turn round and say
The second physical-and-speech action phrase occurred in a different dataset,

a collection of focus group discussions around the risk of terrorism in contem-
porary UK society.11 In this extract from talk in a group of white, working-class
men from northern England, one of the participants uses the phrase turn round
and say in speaking about a television debate among British Muslims that he had
seen:
Extract 912
2 131 Leo .. and they’re sitting there,
2132 and the head Imam,
2133 in this country,
2134 → when someone turned round,
2135 and said,
2136 .. <Q what shall I do,
2137 if one of my friends,
2138 are involved in terrorism ?
2139 .. er,
2140 shall I go to the police Q>?
2141 and then the head Imam,
2142 and his .. other partner,
2143 → turned round and said,
2144 .. <Q [tuts]
2145 no no no,
2146 as a good Muslim,
2147 .. you should not go to the police,
2148 about anything,
The use of the phrase turn round and say highlights the reported speech that fol-
lows it, while at the same time conveying a disapproving or antagonistic attitude
on the part of the speaker towards the content of that reported speech and to the
reported speaker. The attitudinal perspective is constructed through metaphori-
cal use of the physical action turned round which may activate simulations of both
action and associated social emotions.
To see that turned round is metaphorical, we can suppose that, in the televi-
sion programme reported by the speaker, the Imam (Muslim religious leader)
would most likely have been facing his audience. So the someone of line 2134
who was in the audience would have been very unlikely to have physically turned
round and said since that would imply s/he was originally facing away from the
Imam. S/he could have literally stood up and said or looked up and said but not
literally turned round and said. Similarly, in line 2143, the Imam can be expected
to have been facing his questioner as he replied, so when the Imam is held to have
turned round and said…, we can assume that the physical action is metaphorical,
while the speech action is non-metaphorical.
The affective or attitudinal perspective carried by turned round can be ex-
plained, as with sit down, through experience of the social emotions associated
with the physical act, which has become sufficiently widespread across society to
connect socio-cultural norms of behaviour with the stabilised language pattern.
To make sense of the source or vehicle domain schema of turn round and say, the
person who is about to speak has to be physically positioned facing away from
his or her addressee. At the point of speaking, the speaker then turns round to
face the addressee. Our embodied and socio-cultural experience of this action
is not affectively neutral. We do not normally face away from people who we
like or even feel neutral towards, since this would be impolite. Facing away from
someone suggests rejection of that person for some reason: dislike, distaste, fear.
To then turn round prior to saying something suggests a threatening, hostile or
assertive act towards a negatively-evaluated addressee.
A search of the full dataset of focus group talk (approx 213K words of tran-
scription) found 9 further uses of the same or similar phrase, 3 more by the same
speaker and 5 by other speakers (Extracts 10 (a)–(i)):
Extracts 10
Extract 10(a)
1 677 Leo .. ‘cos there’s no getting round --
1678 if someone turned round and says,
1679 ..<Q only ginger-haired people,
Extract 10(b)
1 947 Ben .. the- the .. British Muslim Council,
1948 did turn round,
1949 .. and say <X look we X> knew,
1950 a lot of ‘em were coming over –
Extract 10(c)
1 270 Leo yeah.
1271 we should’ve turned round and said,
1272 <Q are you at war with us Q>?
Extract 10(d)
3 763 Leo you’re going to X --
3764 you’ll turn round and go,
3765 .. <Q oh.
348 Lynne Cameron
Extract 10(e)
1219 Finn if every member of the IRA,
1220 turned round,
1221 and said,
1222 <Q right,
Extract 10(f)
2533 Phil .. they actually said --
2534 turned and says,
2535 <Q you need to go on that bus Q>.
2536 she says,
Extract 10(g)
5325 Phil .. soon as Labour got in,
5326 .. my mum turned round and says,
5327 .. <Q country’s is going down X,
Extract 10(h)
835 Farid are all over the world,
836 .. they are now turning around and saying,
837 that they are the enemy.
Extract 10(i)
1287 Sarfraz at that time I was thinking,
1288 <QS you know now,
1289 every white guy is going to turn around,
1290 and kick--
1291 kick us ..all Q>,
The majority of these concern talk about hypothetical scenarios, with only (b), (f)
and (g) concerning scenarios that actually happened. Even in these actual scenar-
ios, it is unlikely that the turning round was literal; in most cases, the contextual
evidence suggests that was metaphorical.
<turn round and say> in the larger corpus
Examination of the phrase in the Bank of English UK informal spoken corpus

(15 million words) produced 99 instances of <turn (a)round and say> in various
grammatical forms (Table 1 below).
Table 1. Instances of <turn round and say> in UK informal spoken English corpus
of Bank of English
Form Number of instances Number of uses with
hypothetical scenarios
turn round and say 61 49
turn around and say 0 –
turns round and says 6 6
turns around and says 0 –
turned round and said 23 2
turned around and said 3 –
turning round and saying 9 1
turning around and saying 0 –
Total 99 58
The number of uses to talk about hypothetical scenarios is given in the righthand
column of Table 1, and suggests a distinction in use between the first two forms
and the second two forms. The hypothetical uses are exemplified by the corpus
lines in Extract 11, which show the characteristic pattern found in 25 of the 49
cases:
hypothetical scenario + modal verb + turn round and say
Extract 11
I mean using that argument you could turn round and say Rambo’s okay because
they one could actually turn round and say they’re self-contained
…who might turn round and say all right so in life there are …
In order to assess the metaphoricity of the physical action in such phrases, we

need to imagine the context of the hypothetical world. The actions often seem
highly likely to be metaphorical, at the very least they are metonymic; if indeed
there is (or might be) a physical turning around, it stands for much more than the
physical act since it invokes connected social emotions.
When the phrase was used in reference to real situations, the physical action
tended to be metaphorical, and to bring a sense of opposition or antagonism be-
tween the actual speaker and those reported to have spoken:
Extract 12
a lot of people turn round and say M’s arrogant but he’s not…
…
…. now they turn round and say we want Bush and we want Major to …
350 Lynne Cameron
The corpus evidence thus suggests the metaphoreme <turn round and say> with
the following stabilised characteristics:
– the physical action of turning round is metaphorical
– the speech action is likely to be hypothetical rather than real
– affectively, there is implied opposition, dislike or distrust between speaker
and those reported to say in the hypothetical scenario, and/or what is said.
– the grammatical form of the physical and speech action verbs is likely to be
turn/say or turns/says, i.e. simple present tense or root form
– a modal verb occurs in about 50% of uses.
Once the metaphoreme has stabilised at the level of the speech community
with this metaphorical sense to it, it becomes seem less likely to be used non-
figuratively.
Other physical-and-speech action verbs in the focus group data
A search of the dataset for other physical and speech action verbs with possible
metaphorical senses produced the following list:
stand up and say
have the finger pointed at them and said
came forward and said
came up to me and said
come out and say
(America, UK) go into other countries and say
go out and stand on a cardboard box and say13
bringing your own little part of Islam into this country and saying
look at and say
Each of these uses the speech action verb metonymically or literally, while the
physical action verb is metaphorically used, sometimes metonymically as well. In
only one instance in the data was such a combination non-metaphorical, express-
ing a chain of two actions:
he ran up the stairs and said
Conclusion
The physical-and-speech action expressions discussed in this paper are blurred,

hybrid forms that exist in the borderlands of metaphor, metonymy and the literal,
physical world, and that can invoke processes of literalisation and symbolisation,
both on the microgenetic timescale of spontaneous talk and on the longer times-
cale of stabilising metaphoremes. While their theoretical position is borderland
and blurred, their relatively high frequencies in the corpora suggest that they are
effective and important in human communication. Combining a physical action
verb with the speech action verb appears to have evolved as an important way of
implicitly expressing emotion, attitude and value in informal spoken English. The
physical action part of the phrase, which is used metaphorically, conveys attitude
towards the speaker or what is said via shared connections of the action with so-
cial emotions. Attitude and value come to be symbolised through metonymy or
association – people act this way when they feel like this. Through stabilisation of
language, concepts, pragmatics and affect, combined with a degree of variability,
these expressions can be readily constructed in the flow of talk to fit speakers’
communicative needs.
The discussion of these cases has raised several further empirical questions
that require examination of further examples in a large corpus: for example, how
extensive is the observed correlation between talking about hypothetical situa-
tions and these kinds of expressions? How do people talk about non-figurative
combinations of physical action and speech?
The discourse dynamics approach applied to physical-and-speech actions has
revealed a further discrepancy between conceptual metaphor theory and real-
world data – what appears to be a readily processed, frequently used and commu-
nicatively important type of expression is not easily explained by existing theory. I
have suggested that we need instead a theoretical framework that brings together
– a dynamic view of metaphor in language use, involving continually shifting
symbolising and literalising;
– connections between language, physical action and social emotions, for which
simulations and somatic markers currently provide the best explanation, and
that can serve to bridge between the metaphorical vehicle or source domain
and the discourse topic, in the flow of discourse;
– the notion of the metaphoreme to connect the microgenetic timescale of face-
to-face talk with longer timescales across social groups and speech communi-
ties, in a dynamic systems view of language and thinking.
352 Lynne Cameron
Notes
1. build this bridge is said to be metaphorical because (i) we can find another meaning – the
concrete construction of a physical bridge – that is distinct from the contextual meaning of
(something like) the process of reconciliation or developing empathy, and (ii) the concrete idea
of bridge building can contribute to understanding the process of reconciliation. Evidence of
metaphorical intention or interpretation is not required in this procedure for the identifica-
tion of ‘linguistic metaphor’; this would require different kinds of evidence. What is identified
concerns the potential for metaphor rather than making claims about active metaphorical pro-
cessing. (For more on issues in identifying linguistic metaphor, see Cameron (2003, Chapter 3);
Pragglejaz Group (2007)).
2. Triangular brackets are used in descriptions of metaphoremes. The form inside is in italics.
3. The project was funded by the UK Arts and Humanities Research Board (now Council).
Thanks are due to Jo Berry and Pat Magee for their co-operation and permission to use the data,
and to Juup Stelma, the project research assistant.
4. Each line of the transcript is an intonation unit (Chafe, 1996; Stelma & Cameron, 2007). On
average, an intonation unit is produced every 2 seconds. So, the number of the line indicates
roughly how far into the conversation an extract occurs.
5. The first number before the line number on each extract shows which part of the dataset the
extract is taken:
1 – the conversation recorded at their second meeting in Dec 2000;
2 – the conversation recorded at the fourth meeting in Feb 2001;
3 – a radio interview recorded in April 2003.
6. The suggested explanation is offered as an alternative to one depending on ‘image schemas’.

Image schemas in cognitive metaphor theory are held to be non-propositional structures that
organise embodied experience with the physical world (Gibbs, 2006). Recent work by Gibbs
(2006) has emphasized the embodied nature of image schemas, and how people’s experience of
interacting with the physical world develops and builds up their image schemas. However, the
original construct of the image schema is limited in explanatory adequacy in three directions
related to key features of the discourse dynamics approach as set out earlier: image schemas
are not dynamic; they are not social in that they do not seem to feature other people; language
detail does not play any role in the proposed use of image schemas.
7. and so on: such vagueness is likely to be a key feature enabled by metonymy or metaphor.
8. We have not mentioned the influence of phonology, but there is also probably some effect of
rhythm and stress that makes these phrases more likely to be used (Boers & Stengers, 2008).
9. Thanks to Alice Deignan for ‘harvesting’ these examples from the Bank of English.
10. We may in future work want to reconsider the name and move to a more inclusive term
such as ‘figuremes’.
11. The project was funded by ESRC under its New Security Challenges programme. Thanks
are due to participants for permission to use the data, and to Robert Maslen, researcher on
the project.
12. Metaphors are not underlined in this and other extracts from this dataset.
13. A cardboard box would of course be inadequate for standing on. It seems likely this is a
variant of the UK English expression about ‘standing on one’s soap box’. Nowadays this means
loudly expressing one’s feelings, but metonymically it derives from the historical scenario in
which people would take a sturdy wooden box such as one used for delivering soap, turn it
upside down and stand on it to orate in public.
References
Arbib, Michael (2002). The mirror system, imitation and the evolution of language. In
C. Nehaniv & K. Dautenhahn (Eds.), Imitation in animals and artifacts (229–280). Cam-
bridge, MA: The MIT Press.
Bakhtin, Mikhail M. (1981). The dialogic imagination. Austin: University of Texas Press.
Barsalou, Lawrence W. (2008). Grounded cognition. Annual Review of Psychology, 59, 617–645.
Black, Max (1962). Models and metaphors. New York: Cornell University Press.
Boers, Frank & Hélène Stengers (2008). Adding sound to the picture: Motivating the lexical
composition of metaphorical idioms in English, Dutch, and Spanish. In M. S. Zanotto,
L. Cameron & M. Cavalcanti (Eds.), Confronting metaphor in use: An applied linguistic ap-
proach (63–78). Amsterdam & Philadephia: John Benjamins.
Cameron, Lynne (1999). Identifying and describing metaphor in spoken discourse data. In
L. Cameron & G. Low (Eds.), Researching and applying metaphor (105–132). Cambridge:
Cameron, Lynne (2003). Metaphor in educational discourse. London: Continuum.
Cameron, Lynne (2007a). Patterns of metaphor use in reconciliation talk. Discourse and Society,
18, 197–222.
Cameron, Lynne (2007b). Confrontation or complementarity: Metaphor in language use and
cognitive metaphor theory. Annual Review of Cognitive Linguistics, 5, 107–135.
Cameron, Lynne (2008). Metaphor shifting in the dynamics of talk. In M. S. Zanotto, L. Cameron
& M. Cavalcanti (Eds.), Confronting metaphor in use (45–62). Amsterdam & Philadelphia:
John Benjamins.
Cameron, Lynne & Alice Deignan (2003). Using large and small corpora to investigate tuning
devices around metaphor in spoken discourse. Metaphor and Symbol, 18, 149–160.
Cameron, Lynne & Alice Deignan (2006). The emergence of metaphor in discourse. Applied
Linguistics, 27, 671–690.
Cameron, Lynne & Robert Maslen (Eds.) (2010). Metaphor analysis: Research practice in applied
linguistics, social sciences and the humanities. London Equinox.
Cameron, Lynne, Robert Maslen, Zazie Todd, John Maule, Peter Stratton & Neil Stanley (2009).
The discourse dynamics approach to metaphor and metaphor-led discourse analysis. Met-
aphor & Symbol, 24, 1–27.
354 Lynne Cameron
Cameron, Lynne & Juurd Stelma (2004). Metaphor clusters in discourse. Journal of Applied
Chafe, Wallace (1994). Discourse, consciousness and time. Chicago: University of Chicago Press.
Clark, Herbert (1996). Using language. New York: Cambridge University Press.
Damasio, Antonio (1994). Descartes’ error: Emotion, reason, and the human brain. New York:
Putnam.
Gibbs, Raymond W. (2006a). Embodiment and cognitive science. New York: Cambridge Uni-
versity Press.
Gibbs, Raymond W. (2006b). Metaphor interpretation as embodied simulation. Mind & Lan-
guage, 21, 434–458.
Gibbs, Raymond W. & Lynne Cameron (2008). The social-cognitive dynamics of metaphor
performance. Journal of Cognitive Systems Research, 9, 64–75.
Kimmel, Michael (2004). Metaphor variation in cultural context: Perspectives from anthropol-
ogy. European Journal of English Studies, 8(3), 275–294.
Kittay, Eva F. (1987). Metaphor: Its cognitive force and linguistic structure. Oxford: Oxford Uni-
versity Press.
Lakoff, George (1993). The contemporary theory of metaphor. In A. Ortony (Ed.), Metaphor
and thought (2nd ed., 202–251). New York: Cambridge University Press.
Larsen-Freeman, Diane & Lynne Cameron (2008). Complex systems and applied linguistics. Ox-
ford: Oxford University Press.
Linnell, Per (1998). Approaching dialogue. Amsterdam & Philadelphia: John Benjamins.
Pickering, Martin & Simon Garrod (2004). Towards a mechanistic psychology of dialogue. Be-
havioral and Brain Sciences, 27, 169–226.
Pragglejaz Group (2007). MIP: A method for identifying metaphorically-used words in dis-
Richards, Ivor A. (1936). The philosophy of rhetoric. London: Oxford University Press.
Rolston, William (1992). Drawing support: Murals in the North of Ireland. Belfast: Beyond the
Pale Publications.
Stelma, Juurd & Lynne Cameron (2007). Intonation units in research on spoken interaction.
Text and Talk, 27, 361–393.
Appendix
Concordance lines for sit* from reconciliation data
1 into that territory. you know. of there’s no right, for me, to sit here and be forgiven. if you
understand me. I mean, in a sense the
2 ve taken. etcetera. but, when you come when it comes down to it, I am sitting with some-
body who’s affected by, my actions. and that’s
3 u were aware that there’s a it’s going to be an individual who you’d be sitting down with.
sure yeah I think you’re a very intellectually open p
4 t have been a conflict. you know, there is always a way around. if you sit down, but I think
sometimes that powerful interests look for th
5 but you know, you y- maybe some time, who knows, you could sit back and be sitting
across the room from, some RUC man, who’s beaten
6 you y- maybe some time, who knows, you could sit back and be sitting across the room
from, some RUC man, who’s beaten me around the room
7 I saw very clearly. that the the end of that journey, would be, sitting down and, talking to
the people who did it. that just came in a
8 what what have you got out of those encounters I feel that too I am sitting here as a as a
perpetrator you know because that’s how we’re ca
9 t was done needed to be done again I’m conscious of the fact that I’m sitting here with a
victim of those actions justifying it’s one thing comin
10 ably wrong to say it’s an end but yo- yu- you knew one day you’d be sitting beside me you
couldn’t know how I’d respond to it and yet I hope
11 feel that he would understand why you’re going to Derry why you’re sitting here and I
remember recently reading that shortly after my ar
12 would have thought it would be relatives of British soldiers I could sit and exchange stories
or with former British soldiers with former combatan
13 happen and that’s a necessary job to do didn’t occur to me that I’d be sitting down with a
v- like a victim of Brighton you know and may
14 ther combatant’s perspective and I said one day I could see myself sitting down with for-
mer British soldiers former police officers RUC m
15 come to terms with our enemies but they’ll never forgive and then I’m sitting here with a
victim of Brighton and what that has taught me is
16 we can adapt to the needs of like all those combatants so they can sit and tell their story
and somehow get in touch with those feelings I think
17 ould li- I know that from one point realising one day I would have to sit down with British
soldiers that sort of because I I’d be obliged to
18 ginning after the Brighton bomb that’s like been my goal to be able to sit here with you and
it’s something I want I wanted this for so long y
19 pen which encouraged me obviously when somebody is so willing to sit and listen and be
open herself well that put’s you on the spot you
20 he British powers to be the British political establishment to actually sit down and listen to
the republican grievances Host so the things are
21 all these decades we have arrived at a situation where we can now sit down across tables
and talk to each other Host but as far as you wer
22 republican let me say I was I felt obliged as a republican to sit down and talk about that and
against the backdrop of the political reason
23 h at had been censored for decades so when offered an opportunity to sit down and talk
about what motivated you then you should avail of
24 enly in the presence of somebody who you’ve hurt you know er, I’m sitting there beside the
woman whose father I have killed and at that time
25 beside the woman whose father I have killed and at that time I was sitting in this wee
kitchen talking to this woman for the first time wh
26 on I know there was a time when I was just struck by if Pat and I can sit down to talk you
know then why did my dad have to die and I’ve a
27 bombing to happen Pat as an example you know the fact that we can sit here after we --
what brought us together that we can sit and talk
28 that we can sit here after we -- what brought us together that we can sit and talk and listen
and you know what are we talking about I’m --
chapter 18
The evaluative properties of metaphors
Alice Deignan
University of Leeds
It has been asserted by a number of writers from a range of disciplines that met-
aphors have evaluative and persuasive properties. In the Conceptual Metaphor
Theory literature, it is widely agreed that metaphor is evaluative, persuasive and
therefore potentially ideological. Outside the cognitive literature, metaphor’s
persuasive qualities have been studied by researchers in fields from Critical
Discourse Analysis to consumer research.
This chapter reviews the mechanisms through which evaluation seems
to be carried out, discussing a number of studies. Four themes are identi-
fied: metaphors as stories, with entailments, metaphors which map evaluative
connotations from source to target domain, the deliberate choice of a source
domain which resonates with particular groups of language users, and the use
of metaphorical scenarios. I investigate these mechanisms in a small number
of metaphorical expressions, using corpus techniques, with a focus on detailed
analysis of form as well as meaning. I find that the evaluations conveyed by the
expressions that I investigate seem most consistent with Musolff ’s scenarios
model (2006). However the evaluative meaning expressed is often even more
specific, contextually and pragmatically, than his theory would predict. I also
argue that the characteristic form of evaluative metaphorical expressions raises
challenges for our current categories of linguistic description.
Keywords: metaphor, evaluation, connotation, chunking
Introduction
The evaluative and persuasive properties of metaphor were discussed before the
development of the cognitive approach. For instance, Ullmann wrote that ‘meta-
phor is the supreme source of expressiveness in language’ (1962: 202), and Waldron
(1967) analysed tropes such as Man is a wolf, pointing out that the metaphor con-
veys evaluation rather than an objective similarity between topic and vehicle (but
see Gibbs, this volume, for a discussion of the limitations of such research).
358 Alice Deignan
More recently, these properties have been mentioned both within the cog-
nitive approach and by scholars working in other traditions. Within cognitive
metaphor theory, it is widely agreed that metaphor is evaluative, persuasive and
therefore potentially ideological, arguments that have been summarized recently
by Maalej (2007). In the early days of the theory, Lakoff and Johnson (1980: 156–
158) argued that politicians can exploit metaphor to persuade. Since, there have
been a number of studies on metaphor and persuasion within a broadly concep-
tual framework (for example, Charteris-Black, 2004; Sandikcioglu, 2000), some
of which are summarised below. Outside the cognitive and linguistic literature,
metaphor’s persuasive qualities have been studied by researchers in advertising;
for instance, McQuarrie and Phillips (2005) present evidence showing that adver-
tisers increasingly use rhetorical, indirect and metaphorical ways of presenting
their claims.
For applied linguists, a gap in much of this research is the details of how ex-
actly a metaphor is able to communicate an evaluation. Several linguistic mecha-
nisms are described briefly or alluded to, but the details are rarely explored, as in
most studies the focus is understandably on the evaluation itself and its signifi-
cance for speakers, hearers and society more widely. In this paper, I review the
mechanisms through which, according to existing studies, metaphorical evalua-
tion is carried out. I then use corpus data, a research resource increasingly used
‘to better test hypotheses about the frequency and form of metaphors’ (Gibbs, this
volume). I describe some corpus analyses of conventional metaphors, considering
how each metaphor seems to convey its evaluation. I find that these metaphorical
evaluations are consistent with Musolff ’s notion of scenarios (2004, 2006), in their
specificity, but that they are even more specific in their evaluation and pragmatics
than scenario theory predicts. I also show that the form of many of these meta-
phors is consistent with Sinclair’s (2004) argument that the lexical item, which is
usually a multiword expression rather than a single word, is the fundamental unit
of meaning.
Mechanisms through which metaphors evaluate
In this section, I review some studies of the use of metaphor to evaluate and per-
suade. Four different mechanisms for framing evaluations and ideology meta-
phorically have been proposed: the use of metaphor to create entailments; the
use of metaphorical scenarios; the choice of source domains that are meaningful
to particular language users; and the exploitation of the connotations of a word’s
literal meaning. Each is now discussed in turn.
Chapter 18. Evaluative properties of metaphors 359
Creating entailments
Schön (1979) argues that metaphors can harness stories, together with their en-
tailments, to present a particular interpretation of situations and events, the en-
tailments suggesting a course of action that is apparently natural, if not inevitable.
He examines the use of linguistic metaphors such as palliative, blighted, decay
and treated in a text about urban renewal, and argues that these are realizations
of a conceptual metaphor that he terms poor housing is a disease. ‘Disease’ is
a story or schema well known to all speakers, and its equally well-known entail-
ments include the notions that the disease must be treated, an expert needs to
treat it and that [some] diseases are infectious.
Schön argues that the conceptual metaphor maps these and related entail-
ments onto the domain of housing. When mapped onto the target domain, the
entailments suggest that outside intervention is needed to ‘cure’ the problem of
slum housing, that whole communities must be ‘treated’, just as a whole body
is treated medically, and that all ‘diseased’ areas must be removed to prevent a
few surviving pockets of disease re-infecting the rest of the area. Schön argues
that once we perceive a slum as a blighted area, we conclude that the blight must
be removed. In literal terms, this means that buildings must be torn down as
symptoms of disease. As a means of presenting a judgment about a situation, us-
ing a metaphor to generate entailments in this way may be particularly effective
because it is off-record. The reader or hearer does not need to be told explicitly
what conclusions to draw; knowing the story of the metaphor, he or she can sup-
ply them unaided.
The field of communication research offers support for the hypothesis that
metaphors create entailments. Sopory and Dillard (2002) conducted empirical
studies of the persuasive effects of metaphors, investigating six hypotheses as
to how metaphors achieve these effects. They claim that their findings are most
strongly consistent with the ‘superior organization’ hypothesis, that is, that a met-
aphor organizes facts into a story.
Lakoff and Johnson (1980) also claim that metaphor works in this way, and
give many examples of metaphorical entailments, such as the ‘war’ metaphor for
the energy crisis, which was widely used in the 1970s. Followed through, they ar-
gue, the war story can justify the use of military means to secure energy supplies,
because its entailments include:
there was an external, foreign, hostile enemy (pictured by cartoonists in Arab
headdress); energy needed to be given top priorities; the populace would
have to make sacrifices; if we didn’t meet the threat, we would not survive.
(1980: 156–7)
360 Alice Deignan
Lakoff analyzed entailments of the metaphors used to talk about the wars in the
Persian Gulf in 1991 and 2003, writing:
Since the US lost few [..] ‘assets’ in Gulf War I, the war was reported, just after-
ward in the NY Times Business section, as having been a ‘bargain’. Since Iraqi
civilians were not our assets, they could not be counted as among the ‘losses’ and
so there was no careful public accounting of civilian lives lost, people maimed,
and children starved or made seriously ill by the war or the sanctions that fol-
lowed it. […]
And why such enmity toward France and Germany? Via the Nation As Per-
son metaphor, they are supposed to be our ‘friends’ and friends are supposed
to be supportive and jump in and help us when we need help. Friends are sup-
posed to be loyal. That makes France and Germany fair-weather friends! Not
there when you need them. (2003: 1–2)
O’Halloran (2007) gives an account of the use in Critical Discourse Analysis of

conceptual metaphor theory to analyse the entailments of metaphors. He cites
another analysis of the disease metaphor, in this case to talk about riots rather
than housing (Fairclough, 2001). Again, talking in terms of disease creates sev-
eral entailments in the target domain that the metaphor makes seem natural, but
which are in fact open to dispute. These include the notion that the status quo of
society is its natural, healthy condition and anything which disturbs that status
quo is a threat that must be ‘cured’. O’Halloran critiques this kind of analysis,
arguing that the inferences created by the careful reading of an analyst may not
actually be created by the casual reader. Nonetheless, many researchers would
argue that even if inferences are not explicitly drawn in casual reading, they may
be noted subconsciously, and if the same ‘stories’ are used repeatedly, there will be
a cumulative persuasive effect.
It is important to note that all the texts analysed in the studies discussed
here are carefully crafted, consciously persuasive texts: advertisements, political
speeches and policy documents. This is a very important kind of persuasive text,
but not the only kind. It is possible that more informal or spontaneous texts may
use metaphor to evaluate in other ways.
Scenarios
Musolff developed the notion of scenarios as ‘an intermediate analytical category

between the level of the conceptual domain as a whole and its individual ele-
ments’ (2004: 13). He points out that conceptual metaphors can encompass more
detailed mappings that have apparently conflicting entailments. For instance,
he cites Lakoff ’s (1996) discussions of the conceptual metaphor a nation is a
family, which seems to be exploited in two different models of the state: the
strict father model and the nurturant parent model, which have very dif-
ferent entailments and underlie opposing political views. Each of these models
would be a scenario for Musolff. Scenarios are of interest here because they in-
clude an evaluative slant on a ‘story’. Musolff defines a scenario as:
a set of assumptions made by competent members of a discourse community
about “typical” aspects of a source-situation, for example, its participants and
their roles, the “dramatic” storylines and outcomes, and conventional evaluations
of whether they count as successful or unsuccessful, normal or abnormal, per-
missible or illegitimate, etc. (2006: 28)
Musolff analyzed a large corpus of texts about the European Union, and found
that the source domain of journeys was mapped metaphorically onto the process
of European countries moving towards a shared currency and some shared legisla-
tion. Journey is a conceptual domain in this metaphor, and it is realized through
several different scenarios, which take the form of journeys by different forms of
transport. A ‘walking together along a path’ scenario can be exploited to suggest
that some states are walking faster than others towards the goal, that some are not
walking on the right path, or that there are obstacles along the path which can slow
down some states. A train journey scenario cannot be exploited in this way: the
various European states were either on the train or off it. In Musolff ’s corpus, such
scenarios were used both positively and negatively, according to the intentions of
the speaker. He found instances of different speakers using the same set of meta-
phors, positively and negatively; often British politicians tried to turn around the
metaphors offered by a European politician to give a negative spin on European
integration. For instance, there was reference in the British press to the European
train traveling too fast, and the possibility of a disastrous end to its journey.
Musolff argues that any given conceptual mapping can generate unlimited
numbers of scenarios, each of which can carry a different evaluative slant. This
implies an important limitation to the analysis of metaphorical stories and their
entailments. It is possible to identify and count linguistic metaphors related to a
particular source domain, and the researcher can speculate as to the entailments
that these may have. However, to determine the evaluative force of the metaphor,
it is necessary to identify a scenario in more detail, and then to study how a par-
ticular writer or speaker uses the metaphors generated by the scenario.
Choice of source domain
Several researchers have argued that some writers or speakers choose metaphors
from a particular source domain as a strategy for aligning themselves with people
362 Alice Deignan
for whom the source domain is significant. For instance, Semino and Masci (1996)
studied the metaphors used by the Italian politician Berlusconi, and found that
he used a large number of football and religious metaphors. The football meta-
phor has many entailments, which seem to have been exploited by Berlusconi’s
party in the way described above. As well as the entailments these create, the
choice of the source domains themselves seems significant. Semino and Masci
argue that the use of football as a metaphor enabled Berlusconi’s party to draw on
‘the positive connotations of the most popular Italian sport, which attracts strong
passions across barriers of social class and regional origin’ (1996: 247), thus tacitly
positioning his political party as democratic and patriotic. Berlusconi also used
a number of biblical metaphors, alluding to the story of the Good Samaritan and
other images from the New Testament. Semino and Masci argue that this choice
‘seems especially appropriate to appeal to the Catholic electorate and particularly
the floating ex-Christian Socialist voters’ (1996: 255).
Koller (2004) also finds that some source domains favour particular groups of
language users over others. Drawing metaphors from such source domains can be
considered a persuasive and evaluative strategy in the sense that it suggests shared
values between the writer and certain groups of readers, thus presumably promot-
ing ease and intimacy, and a likelihood that the writer’s version of events will be
favoured over other texts and tellings. Koller analysed a corpus of business texts,
and found that they frequently draw on the source domains of war and sport. She
argues that these choices act to include men, and that they exclude many female
readers. The metaphors construct the domain of business not only as male, but in
an aggressive, violent and dominating stereotype of masculinity.
As was the case for the analyses of entailments and scenarios discussed above,
Semino and Masci’s, and Koller’s studies focus on consciously constructed texts,
and it is possible that this strategy may be less evident in other, more spontane-
ously created texts.
Connotation
Discussion of evaluation and metaphor before the development of the cogni-

tive approach often assumed that the persuasive effects of metaphor were due
to the mapping of connotations from the source to target domains. For instance,
Waldron (1967), cited above, claimed that animal metaphors gained their evalu-
ative force from the associations we hold with each animal. However, there is
only limited evidence from detailed linguistic studies to support this hypothesis.
Deignan (2000) investigated a group of words whose literal meanings are near
synonyms with different connotations, to see whether these semantic relationships

are also apparent in target domain uses. Perhaps surprisingly, little evidence was
found for this. She studied concordances for thin, slim, slender and skinny, and
for a group of near synonyms describing excess bodyweight. The data support
the general view that the literal uses of thin and skinny have negative connota-
tions and slim and slender have positive connotations. The metaphorical uses were
less straightforward. Metaphorical thin is sometimes negative, for instance when
used to describe an argument or idea that lacks substance. Skinny is rarely used
metaphorically, and the few uses found were negative. However, the main meta-
phorical uses of slim and slender also tend to be negative; frequent collocates are
chance(s) and hope(s), where a slim or slender chance is a very small one. An-
other frequent collocate is majority; a politician who wins by only a slim or slender
majority is not usually pleased about this. The negative orientation of these uses
is less surprising given that both slim and slender have had negative meanings
historically; the Shorter Oxford English Dictionary on Historical Principles (Little,
Fowler, Coulson & Onions, 1973) gives a metaphorical, negative meaning for slim
dating from 1674, and for slender from 1533. The positive literal meanings of the
words are more recent.
Corpus data show that the verb slim down is sometimes used as a euphemism
for the reduction of staff in an organization. This is presumably intended to make
the action sound more attractive by analogy to the literal meaning, but there is
corpus evidence that in recent years the expression has often been used ironically
(Deignan, 2000). Overall then, this study found that the connotational relation-
ships between words from the same source domain are not always found in their
metaphorical uses, especially where a word has a complex etymology. The study
also suggests that speakers’ and writers’ attempts to impose a positive spin on
actions and events, by exploiting source domain connotations, may in turn be
exploited for humorous or ironic effect. However, the study did not demonstrate
that connotations are never mapped metaphorically; there was evidence from fi-
nancial reporting texts that words such as bloated and flabby have negative con-
notations consistent with their literal meanings. It is possible that the study took
too simplistic an approach, especially in its focus on single words.
I have briefly summarized four mechanisms that speakers may use to evaluate
through metaphor: creating entailments, exploiting scenarios, choosing signifi-
cant source domains, and mapping connotational meaning. These are not mutu-
ally incompatible, and it seems likely that all are used, perhaps in different kinds
of texts. In the following section, I describe some metaphors found in concor-
dance data.
364 Alice Deignan
Analysing metaphors using corpora
Methodology
Corpus analysis is unlikely to disprove the existence of any of the mechanisms

described above, and indeed, it is likely that they all operate to some extent. How-
ever, a corpus analysis may help to suggest refinements to these models, and if suf-
ficient data are analysed, it could also help to show which mechanisms are most
significant. The data analysis reported here suggests very detailed and specific
matches between individual linguistic metaphors and evaluation, closest to the
connotation model of evaluation, with elements of scenarios.
Many of the studies cited above, especially those putting forward the ‘story with
entailment’ mechanism for evaluation, have used text analysis techniques. Research-
ers would generally look at patterns within an individual text, and would consider
how a metaphorically-used word or expression contributes to the overall evaluative
message of the text. In contrast, a corpus analyst would tend to look for properties
inherent in the word across a large number of instances. The immediate context of
each word would be considered in order to determine its evaluative characteristics,
but the overall evaluative stance of the text in which each word appears would not
be considered. The two approaches are complementary rather than conflicting.
The method used here is a version of corpus linguistic analysis, from which
it derives two central characteristics. First, it seeks to account for all citations of
a word in the corpus, including all its inflections, or a large random sample of
citations. Secondly, it pays close attention to details of linguistic form. The first
requirement means that every instance of a word needs to be classified as either
metaphorical or not. This presents analytical challenges that are not faced if one
considers sole instances of words for analysis. The second characteristic, attention
to linguistic form, means that the researcher looks for patterns such as the regu-
lar occurrence of a word in particular grammatical structures or with particular
inflected forms. Corpus research has shown that different forms of a word often
coincide with different meanings (for example, Hunston & Francis, 2000). The
analysis of form also includes identifying any words that regularly collocate with
particular uses of the word under study. The study of the form of linguistic meta-
phors in this and previous studies (for example, Deignan, 2005) suggested that
a number of metaphorically-used words occur in longer expressions, which are
semi-fixed. This point is discussed and exemplified below.
The analytic techniques used for this study are derived from corpus lexicog-
raphy, which generally treats metaphor as a type of polysemy. The concordance of
a word and of its inflections, or a large random sample for very frequent words, is
analyzed, initially to determine the different meanings of the word manifested in
the citations. When these are identified, the relationships between them are con-
sidered. A version of the Pragglejaz metaphor identification procedure (Praggle-
jaz Group, 2007) is used. The procedure was designed for the analysis of words
in a single text but one of its steps can be adapted to consider concordance data.
Step 3 is as follows:
a. For each lexical unit […], establish its meaning […].
b. For each lexical unit […], determine if it has a more basic meaning in other
contexts than the one in the given context. For our purposes, basic meanings
tend to be
– More concrete […];
– Related to bodily action;
– More precise (as opposed to vague);
– Historically older;
Basic meanings are not necessarily the most frequent meanings of the lexi-
cal unit.
c. If the lexical unit has a more basic current-contemporary meaning in other
contexts than the given context, decide whether the contextual meaning con-
trasts with the basic meaning but can be understood in comparison with it.
(Pragglejaz Group, 2007: 3)
In applying this procedure to corpus data, both the ‘contextual’ and ‘basic’ meanings
need to be represented in a concordance, and, as corpus analysis is usually concerned
with typical language use, need to appear a reasonable number of times in a large
concordance. If a ‘contextual’ meaning that has been identified by other means is
not found in the concordance, or is found only rarely, it would be decided that any
metaphor this represents is innovative. If a ‘basic’ meaning that has been identified
by other means is not found in the concordance, or is very rare indeed, it would be
decided that the former source domain meaning of the word has dropped out of use,
making the metaphorical link a historical one. Whether or not innovative and histor-
ical metaphors are of interest is determined by the individual researcher’s priorities.
In either case, concordance data can provide only very limited evidence for analysis.
Assuming reasonably frequent citations for each meaning in the concordance,
the ‘contextual’ meaning of the word is a candidate for conventional metaphor, and
the ‘basic’ meaning is potentially its literal counterpart. Part (c) of the procedure
involves deciding whether the relationship between the two meanings is one of
metaphoricity or another type of polysemy. For some words this is straightforward,
but in many cases difficulties arise. The corpus used in this study is the 56 million
word section of the Bank of English which is available online through ‘Wordbanks’.
. See http://www.collinslanguage.com/wordbanks
366 Alice Deignan
It consists of contemporary texts from written sources including a range of newspa-

pers, magazines and books, and transcripts of formal and informal speech.
Problem cases for metaphorical analysis
The concordances of many words with potentially metaphorical meanings pres-

ent problems for classification, for both semantic and formal reasons. Typically,
a concordance will contain some citations of single word’s literal and metaphori-
cal uses that are relatively straightforward to classify in the way described above.
However, once these citations are accounted for, there often remains a fairly large
group of citations that are not easy to classify. Problems in classifying such cita-
tions are now described.
Firstly, classifying citations semantically can often be difficult where there is
ambiguity between literal and non-literal meaning. This is found especially in
words that have a literal meaning connected to the human body, bodily actions or
processes. These often occur in multi-word strings that seem to have an element
of metonymy, which is where the ambiguity seems to stem from. The problem has
been documented by Moon (1998), and Charteris-Black (2003); Charteris-Black
discusses the string lick one’s lips, dividing citations into those where the literal
action is likely to occur and those where it is not. He points out ‘it is often ambigu-
ous whether the action occurs or not’ (2003: 296). Deignan makes the same point,
with examples such as keep an eye on, turn one’s back, and bite one’s lip (Deignan,
2005: 66), which are used with literal meanings, figurative meanings and in con-
texts where it is impossible for the analyst to decide which is intended, or even if
both are intended.
Expressions such as lick one’s lips and keep an eye on are also difficult to deal
with because of their form: they do not seem to consist of freestanding words, but
neither are they classical idioms. A corpus study of the meanings of heel and its
inflections (Deignan, 2007) showed a typical case. There are 1361 citations in the
56 million word section of the Bank of English, 887 of which refer to part of a foot
or shoe, including 2 literal citations of Achilles heel, the base of the hand, or some-
thing concrete which is shaped like the heel of the foot. These were all considered
to be literal in the analysis. There are a further 3 citations where heel is used meta-
phorically, to refer to a man who behaves dishonourably, a slightly archaic use.
The remaining 471 citations occur in expressions such as the following:
– Hard on the heels of
– Well heeled
– Head over heels
– Hot on the heels of
– Turn on your heel

– Dig in your heels
– Bring someone to heel
– Kick one’s heels
– Snap at someone’s heels
– Down at heel
– Drag one’s heels
– Cool one’s heels
– Kick up one’s heels
– Be under the heel of someone
These expressions have some element of figurative meaning, and a degree of
fixedness. They seem to hover on the boundaries of three established distinctions:
the distinction between literal and figurative meaning; between metaphorical and
metonymical grounding; and between word and idiom.
The concordance for thin presents similar difficulties for analysis. There are
citations of the literal meaning, and a metaphorical meaning found in citations
such as thin trading and a thin excuse. There is also a large number of multiword
expressions including the following:
– (Vanish into/ appear from) thin air
– Walk/ tread a thin line
– Thin end of the wedge
– Thin-skinned
The concordance for point also shows many multiword expressions with figura-
tive meanings, including the following:
– Starting point
– Focal point
– Point of view
– Point of departure
– High point
– Low point
Such multiword, non-literal expressions are rarely treated as a single category in
linguistic or semantic descriptions. Rather, they occupy the fuzzy boundary area
between figurative collocations and classical idioms. They seem to be more fixed
and specific in meaning than the usual understanding of a collocation, and yet
they lack the semantic opacity usually regarded as criterial to idioms. However,
despite being of uncertain status in linguistic description, corpus studies suggest
that they are very important numerically. They also appear to have interesting
evaluative qualities, which I discuss in the following section.
368 Alice Deignan
Multiword figurative expressions
In this section, I take a few of the multiword figurative expressions listed above,
and use corpus citations to show that they have evaluative qualities. I argue that
these are difficult to describe within a conceptual metaphor theory approach to
metaphor and evaluation, for several reasons. I begin with an expression that has
positive evaluative qualities, then consider two expressions with negative quali-
ties, and finally an expression which has highly specific connotations that are not
located on a good-bad axis.
Turning point
The expression turning point is frequently used with a non-literal meaning. The
adjectives occurring immediately before the expression in the corpus are, in de-
scending order of frequency: real, important, major, crucial, big, great, and his-
toric. These collocates suggest a positive evaluation, and on investigation of in-
dividual citations, this proves to be overwhelmingly the case. The following two
citations are typical:
(1) This round of peace talks marks a turning point, from squabbling over
procedure to negotiating the issues.
(2) That appearance taught me not to be afraid of TV and I think I would put it
down as the turning point in my career.
In the corpus data, a metaphorical turning point is almost always a point where a
situation changes from bad to better, or from one of inactivity and lack of progress
to positive progress. Turning point also appears in some citations as the name of
a charity which aims to help drug users, a name which would be expected to be
chosen for positive connotations. Turning point could be analyzed as a realization
of the well-documented life is a journey conceptual metaphor, an analysis sup-
ported by collocates such as road, path and crossroads found in some citations.
However, this does not explain the positive evaluation that the expression takes
on in the target domain. Literal turning points could take the traveler to either
better or worse places, and there is not therefore a transfer of connotations from
the source domain. Moreover, the evaluation does not apply to all metaphorical
uses of point; metaphorical starting points are neither strongly good nor strongly
bad. The evaluation seems specific to this expression and this meaning.
Down at heel
Moon recorded the evaluative stance of fixed expressions and idioms in her cor-
pus, noting that ‘negative assignments are roughly twice as common as positive
ones’ (1998: 247). This is consistent with findings from the studies reported here.
The expression down at heel is an instance of negative evaluation, both in literal
and non-literal use. A typical citation of the literal use is:
(3) … out of the gloom into the outside – wearing her creased alpaca apron, her
fat insteps leaning over her down-at-heel shoes.
The presence of collocates such as gloom, creased and fat support the analysis of
negative evaluation.
The concordance for down at heel illustrates the ambiguity mentioned above;
in the following citation, the strangers referred to may literally be wearing old
shoes:
(4) Watch out for down-at-heel strangers who want to share a room. (If the
hippie who nicked my traveller’s cheques from that guesthouse in Nicaragua
is reading this, I still haven’t forgiven you.)
However, here down at heel also connotes poverty and dishonesty, as shown in the
co-text of the citation, the negative evaluation of the literal meaning being appar-
ently mapped onto the non-literal meaning.
In the following two citations, the nouns modified by down at heel are an area
of a city and a building respectively, so the expression cannot literally be true. An
analysis of negative evaluation is supported by the collocates masses and trudging
in the first citation, and in the contrast with once grand in the second.
(5) … the masses trudging through the town’s down-at-heel docklands.
(6) … this grand, if somewhat down at heel hotel.
The citations suggest that the non-literal meaning of down at heel describes a
person or entity that was of better quality in the past but is now poor, shabby
and possibly dishonest. Here the connotations of the source domain seem to be
transferred to the target domain meaning, perhaps inevitably given that there is
some degree of fuzziness between the literal and metaphorical meanings for some
citations. A conceptual metaphor analysis of this use is much less straightforward
than for turning point, above. The expression could be analysed as a realisation
of good is up, down is bad, or as a general personification metaphor, in which
worn clothing stands for poor quality. However, neither of these analyses seems
370 Alice Deignan
to satisfactorily explain the very precise evaluations that the expression conveys in
its non-literal use. The expression is also used to describe a very narrow range of
entities, a fact not accounted for by a conceptual metaphor analysis.
Thin-skinned; have a thin skin
Like down at heel, thin-skinned or have a thin skin seems to evaluate negatively
for both its literal and non-literal meanings, though far less strongly for the literal
meaning, which suggests vulnerability, as shown in the following citation:
(7) Although there is a tendency to believe that the fair-skinned and porcelain-
thin-skinned woman will be most reactive to a whole range of irritants, this
is not necessarily true.
The non-literal meaning of thin-skinned is ‘tending to over-react to perceived

criticism’. A loose synonym would be sensitive, but thin-skinned carries a strongly
negative evaluation that sensitive does not. The following citations show this:
(8) The lengths to which some people will go to ensure that no thin-skinned
soul will ever be offended again.
(9) He is thin-skinned to criticism and likes to be liked.
(10) [He] must have developed a very thin skin if he is considering legal action.
Like down at heel, it is difficult to analyse thin-skinned as a realisation of a concep-

tual metaphor. It could perhaps be seen as a metonymy-based metaphor, in which
the body stands for the whole person, including mental attitudes. This would map
the skin onto the outer layer of the self, and would map qualities such as physical
sensitivity on to mental sensitivity. However, this analysis would not explain the
very strongly negative evaluations of the non-literal use.
Into thin air
The expression thin air is used hyperbolically and metaphorically in a small

group of related expressions which describe something or someone going miss-
ing: vanish or disappear into thin air; or being obtained or appearing: pop out
of, appear or be conjured from thin air. The following citation exemplifies a use
that is literal in the sense that the girl referred to is physically absent, though it
is obviously hyperbolic:
(11) She can’t just have vanished into thin air. She was a good-looking, outgoing
girl. Someone must have seen her, someone must know something.
The next citation can also be classified as literal in the sense that it refers to some-
body becoming physically present:
(12) “Ye gods, Jill, I didn’t even see you walk up,” Cadmar said with a grin. “You
didn’t just pop out of thin air, did you now?”
In both these citations, thin air connotes suddenness and surprise, but it is dif-
ficult in these or other corpus citations to attribute either positive or negative
evaluation. The non-literal meaning seems to have the same connotations, of sud-
denness and completeness. The following citation is clearly not literal, because
the co-text shows that the friends referred to are physically present; disappear into
thin air describes their withdrawing their friendship:
(13) Apart from my former business partner, who has been fabulous to me, my
friends disappeared into thin air. There was one couple who Sue and I used
to go around with a lot. After Sue died I spotted them in the street. They
noticed me – then crossed the road to look in shop windows.
It is difficult to determine what conceptual metaphor, if any, underlies the non-

literal uses of thin air, other than perhaps a general mapping of physical onto
abstract, or more specifically of emotional support on to physical presence. In this
case, the connotations from the literal use seem to be transferred very specifically
onto the non-literal use.
Multiword metaphorical expressions and evaluation
Although a large-scale corpus study to confirm the point has not yet been un-
dertaken, small corpus studies such as those reported here have suggested that a
large number of metaphorical uses appear in multiword expressions. It has been
argued here that many of these expressions carry an evaluative meaning. Moon
(1998) claims that idioms appear to encapsulate a culturally-shared fragment of
experience, complete with an evaluative perspective, an account which also fits
some of the multiword expressions discussed here. This seems to apply to down
at heel and thin air, which appear to carry evaluation from the source domain
meaning but in a very specific way. These expressions are thus consistent with a
mechanism through which source domain connotations are mapped. However, it
does not apply to turning point or thin-skinned, where the figurative uses seem to
have, respectively, strongly positive and negative evaluations not attributable to
their literal meaning.
Most of the multiword figurative expressions discussed were difficult to clas-
sify as realisations of a conceptual metaphor. In the cases where a conceptual
metaphor is detectable, such as turning point, the conceptual metaphor had no
372 Alice Deignan
predictive power for the evaluative slant of the linguistic expression. In these
cases, it seems that evaluative orientation is extremely specific to each expres-
sion. From the perspective of form, any underlying conceptual metaphor has no
predictive power in terms of linguistic realisation, and offers no account for the
linguistic fixedness of these expressions. There appears to be a meeting of highly
specific formal, semantic and evaluative properties in these expressions. Cameron
and Deignan (2006) discuss this in terms of complexity theory.
Conclusion
The corpus data that have been discussed here have not demonstrated that any
of the various evaluative mechanisms described in previous writing does not ex-
ist. It is not usually possible for corpus data to demonstrate the non-existence
of a phenomenon; that phenomenon may simply not occur in the data studied.
However, what this study does show is that multiword figurative expressions that
evaluate are very important numerically. These account for a large proportion of
corpus data, both in terms of types (different expressions) and tokens (numbers
of instances of them). I have argued that multiword figurative expressions are not
all easy to describe in terms of conceptual metaphor mappings; they seem to rep-
resent highly specific, if not one-shot mappings. Their evaluative properties seem
largely unpredictable from a mapping, and are very specific. If these linguistic
expressions can be taken as evidence of underlying mappings, it would seem that
one shot mappings are more frequent, and perhaps therefore more significant
conceptually, than current cognitive approaches suggest.
References
Cameron, Lynne & Alice Deignan (2006). The emergence of metaphor in discourse. Applied
Charteris-Black, Jonathan (2003). Speaking with forked tongue: A comparative study of metaphor
and metonymy in English and Malay phraseology. Metaphor and Symbol, 18, 289– 310.
Charteris-Black, Jonathan (2004). Corpus approaches to critical metaphor analysis. Basingstoke:
Palgrave Macmillan.
Deignan, Alice (2000). Metaphors of persuasion. In C. Heffer & H. Sauntson (Eds.), Words in
context: A tribute to John Sinclair on his retirement. University of Birmingham (CD-ROM).
Benjamins.
Deignan, Alice (2007). Image metaphors and connotations in everyday language. Annual
Review of Cognitive Linguistics, 5, 173–192.
Fairclough, Norman (2001). Language and power, 2nd edition. London: Longman.
Hunston, Susan & Gill Francis (2000). Pattern grammar: A corpus driven approach to the lexical
grammar of English. Amsterdam: John Benjamins.
Koller, Veronika (2004). Metaphor and gender in business media discourse. Basingstoke: Palgrave
Macmillan.
Lakoff, George (2003). Metaphor and war, again. Retrieved from http://www.alternet.org/
story/15414
Lakoff, George (1996). Moral politics: What conservatives know that liberals don’t. Chicago: Uni-
versity of Chicago Press.
Press.
Little, William (Ed.) (1973). The Shorter Oxford Dictionary on Historical Principles, 3rd edition.
Oxford: Clarendon Press.
Maalej, Zouhair (2007). Doing critical discourse analysis with metaphor: Towards a discourse
model of metaphor. In C. Hart & D. Lukes (Eds.), Cognitive linguistics in critical discourse
analysis (132–158) Cambridge: Cambridge Scholars Press.
McQuarrie, Edward F. & Barbara J. Phillips (2005). Indirect persuasion in advertising: How
consumers process metaphors in pictures and words. Journal of Advertising, 34, 7–21.
Moon, Rosamund (1998). Fixed expressions and idioms in English: A corpus based study.
Oxford: Clarendon Press.
Musolff, Andreas (2004). Metaphor and political discourse: Analogical reasoning in debates about
Europe. Basingstoke: Palgrave-Macmillan.
Musolff, Andreas (2006). Metaphor scenarios in public discourse. Metaphor and Symbol, 21,
23–38.
O’Halloran, Kieran (2007). Casualness vs commitment: The use in critical discourse analysis of
Lakoff and Johnson’s approach to metaphor. In C. Hart & D. Lukes (Eds.) Cognitive linguis-
tics in critical discourse analysis (159–179). Cambridge: Cambridge Scholars Press.
Sandikcoiglu, Esra (2000). More metaphorical warfare in the Gulf: Orientalist frames in news
coverage. In A. Barcelona (Ed.), Metaphor and metonymy at the crossroads: A cognitive
perspective (299–320). Amsterdam & Philadelphia: John Benjamins.
Schön, Donald (1993). Generative metaphor: A perspective on problem setting in social policy.
In A. Ortony (Ed.), Metaphor and thought (2nd edn) (137–162). Cambridge: Cambridge
University Press.
Semino, Elena & Michaela Masci (1996). Politics is football: Metaphor in the discourse of Silvio
Berlusconi in Italy. Discourse and Society, 7, 243–269.
Sinclair, John (Ed.) (2001). Collins Cobuild dictionary for advanced learners. London: Harper-
Collins.
Sinclair, John (2004). Trust the text: Language, corpus and discourse. London: Routledge.
Sopory, Pradeep & James P. Dillard (2002). The persuasive effects of metaphor: A meta-analysis.
Human Communication Research, 28, 382–419.
Ullmann, Stephen (1962). Semantics: An introduction to the science of meaning. Oxford:
Blackwell.
Waldron, Ronald A. (1967). Sense and sense development. London: André Deutsch.
Index of names
A Bourdieu, Pierre 253, 263 Chung, Siaw-Fong 190, 191, 204,

Abelson, Robert 164 Bowdle, Brian 295, 307 206, 207
Abkarian, Gene 21, 31 Boyd, Richard 305, 307 Clark, Herbert 103, 334, 354
Ahrens, Kathleen x, 185, 190, Bremner, Brian 311, 319 Clarke, David 22, 30, 32, 103, 120
203–207 Breton, André 252, 257, 258, 263 Clausner, Timothy 186, 188,
Ali, Monica 293, 307 Brouwer, Elsbeth x, 210, 245, 190, 191, 195, 203, 205, 206,
Amoraritei, Loredana 263 216–218, 241
267, 268, 285 Brown, Mary 207 Clifford, Lee 317, 319
Anderson, Richard 36, 59 Bullough, Edward 248, 263 Coady, James 36, 60
Arbib, Michael 341, 353 Coleman, Julie 139, 141
Arnberg, Leonore 37, 59 C Collin, Finn ix, xii
Arredondo, J. 104 Caballero, Rosario x, 265, 271, Collin, Peter 311, 319
Asch, Solomon 23, 30, 31 275, 285 Cometa, Michael 105, 120
Austin, John 150, 164 Cacciari, Cristina 21, 31, 32, 263 Corts, Daniel 303, 304, 307
Cameron, Lynne vii, ix, xi, xii, Cowen Orlin, Lena 140, 141
B 3, 9, 17, 24, 31, 32, 41, 59, 165, Crock, Stan 318, 320
Bakhtin, Mikhail 334, 353 166, 183, 266, 271, 280, 285, Croft, William 186, 188, 190, 191,
Barnden, John 79, 184, 210, 213, 286, 291, 292, 294, 306–310, 195, 203, 205, 206, 216–218,
218, 224, 225, 229, 239, 240, 314, 318, 319, 333–336, 344, 241
241, 243 345, 352–354, 372 Csikszentmihalyi, Mihalyi
Barsalou, Lawrence 341, 353 Capell, Kerry 311, 319 324, 331
Basso, Keith 106, 120 Caramelli, Nicoletta 105, 120
Beaton, Alan 51, 59 Carbonell, Jaime 238, 241 D
Bechhoffer, Robin 105, 120 Carey, Susan 24, 31 Damasio, Antonio 341, 354
Beck, Isabel 36, 59 Carney, James 262, 263 Danesi, Marcel 63, 67, 69, 79
Behar, Richard 311, 315, 319 Carroll, Noël 257, 263 Darian, Stephen 293, 307
Belton, Catherine 318, 320 Carson, Iain 317, 319 Davidson, Donald 246, 259, 263
Bender, Emily 242 Carston, Robyn 210, 241 Dawson, Chester 317, 319
Bennett, Heather 106, 120 Carter, Ronald 291–293, 305, Day, Sean 33, 41, 125–127, 130,
Bierens, Cornel 262, 263 307, 322, 323, 331 132, 134–136, 138, 231, 257, 266,
Billow, Richard 105, 106, 120 Chafe, Wallace 352, 354 285, 299, 355
Black, Max 63, 65, 78, 79, 246, Charteris-Black, Jonathan DeDuve, Thierry 248, 257, 263
256, 270, 271, 274, 275, 285, 63, 65, 78, 79, 271, 285, 358, Deignan, Alice ix, xi, 6, 17, 65,
307, 334, 353, 358, 366, 372 366, 372 79, 82, 85, 89, 98, 101, 102, 126,
Blair, Tony 325, 331 Chiappe, Dan 304, 307 141, 166, 213, 241, 309–311,
Boakes, Robert 267, 286 Chiappe, Penny 304, 307 313–315, 318–320, 329, 331, 333,
Boers, Frank 22, 31, 32, 57, 59, Chilton, Paul 323, 331 336, 344, 345, 352, 353, 357,
65, 78, 79, 82, 85, 96, 99, 100, Chinese Knowledge Information 362–364, 366, 372
102, 271, 285, 352, 353 Processing Group (CKIP) Demecheleer, Murielle 82, 85,
Booth, Wayne 324, 331 195, 206 96, 102
Boronat, Consuelo 295, 307 Chirac, Jacques 325, 331 Dickie, George 253, 254, 263
376 Researching and Applying Metaphor in the Real World
Dijk, Teun van 255, 263 Garnham, Alan 36, 60 Henderson, Willie 271, 286,
Dillard, James 359, 373 Garrod, Simon 334, 354 311, 320
Dirven, René 82, 96, 102, 103, Gawel, Richard 267, 286 Hene, Birgitta 39, 60
143, 291, 308 Gentner, Dedre 22, 23, 31, 39, 59, Heredia, Roberto 192, 207
Dorst, Lettie 165, 184, 306, 308 105–107, 120, 237, 241, 295, 307 Herman, Vimala 22, 32, 120
Dowker, Ann x, 105–107, 120 Gevaert, Caroline 140, 141 Herrmann, Berenike 166, 184,
Drucker, Peter 317, 319 Gibbs, Raymond W. ix, 1, 11, 13, 306, 308
Drulák, Peter 331 18, 22, 35, 39–41, 43, 48, 59, Heywood, John 329, 332
Dworkin, Ronald 138, 140, 141 64, 66–67, 79, 81, 100, 106, Hiebert, Elfrieda 36, 60
Dyhouse, Valrie 281, 287 119, 120, 125, 138, 147, 165, 166, Hintze, Peter 328, 332
192, 203, 205–207, 210, 213, Hiraga, Masako 31, 118, 120,
E 214, 230, 241, 242, 246, 247, 205, 207
Echevarría, Carmen 103 249, 262, 263, 271, 284, 285, Hobbs, Jerry 210, 224, 242
Edwards, Ben 312, 319 287, 292, 294, 307, 309, 321, Holderness, Jackie 103, 104
Ellis, Nick 51, 59 324, 332, 334, 335, 341, 352, Holme, Randall 65, 79
Emmenecker, Nelly 105, 120 354, 357, 358 Hoorn, Johan 252, 264
Engel, Matthew 114, 119, 121 Gimbel, Jørgen 38, 48, 59 Honan, Park 140, 141
Eson, Morris, 105, 120 Glasbey, Sheila 79, 241–243 Horemans, Arnout 79
Evans, Mary 23, 31, 41, 59 Glicksohn, Joseph 21, 31 Horton, William 187, 207
Evans, Vyvyan 139, 141 Glucksberg, Samuel 187, 198, Hsu, Yuan-Yu 206
Eyckmans, June 79, 102 205, 207, 210, 242, 310, 319 Huang, Chu-Ren 190, 191,
Gluck, Malcolm 267, 286 204–207
F Goatly, Andrew 126, 141, 266, Huang, Xiting 106, 121
Fabb, Nigel 251, 263 275, 286, 291, 292, 295, 305, Huckin, Thomas 36, 60
Fabiszak, Małgorzata 141 307, 309–315, 317–320, 324, Hughes, Robert 253, 264
Fairclough, Norman 323, 331, 332, 339, 354 Hughson, Angus 267, 286
360, 373 Goldberg, Adele 204, 239, 242 Hundt, Marianne 139, 141
Falkenhainer, Brian 105, 106, Golden, Anne x, 35, 37, 38, 48, Hubbard, Edward 266, 287
120, 241 58, 59 Hunston, Susan 364, 373
Fang, Shin-Yi 206 Gong, Shu-Ping 204–207 Hvenekilde, Anne 37, 38, 59
Fauconnier, Gilles 10, 17, 152, Goode, Jamie 284, 286
156–158, 160, 163–164 Grady, Joseph x, 31, 43, 59, 166, I
Fernández, Rodrigo 104 186, 188, 190, 191, 203–205, Irigaray, Luce 140, 142
Fillmore, Charles 215, 241 207, 210, 212, 215, 217, 218,
Fischer, Joschka 321, 322, 224, 239, 240, 242, 341 J
325–332 Graves, Michael 58, 59 Jäkel, Olaf 137, 142, 212,
Fishelov, David 305–307 Green, Melanie 139, 141 230–236, 240, 242
Flannery, Michael 271, 285 Gries, Stefan 91, 102, 141–143 Jakobson, Roman 248, 250, 251,
Forceville, Charles 157, 161, 163, Groeben, Norbert 322, 332 258, 264, 323, 332
164, 262, 275, 286 Gustafsson, Ylva 141–143 Johnson, Christopher 22, 31
Fourment, Marie-Claude Johnson, Mark vii, 9, 18, 43,
105, 120 H 60, 64, 79, 120, 126, 128, 134,
Fox, Gwyneth 311, 312, 320 Haastrup, Kirsten 38, 57, 60, 61 139, 140, 142, 169, 183, 185,
Francis, Gill 364, 373 Haddon, Mark 293, 307 207, 233, 242, 259, 264, 271,
Freebody, Peter 36, 59 Hall, Geoff 263 272, 286, 295, 308, 341, 358,
Hamilton, George 263 359, 373
G Hammer, Paul 140, 141 Jones, Alison 21, 31
Gabryś, Danuta 79, 102 Hanks, Patrick 74, 79, 294, 297,
Gamble, Dianna 23, 31, 41, 59 306, 307 K
Gardner, Howard 21, 31, 61, Hatfield, Elaine 140, 141 Kaal, Anna 165, 184, 306, 308
105–107, 113–114, 119–121 Haynes, Margot 36, 60 Kahn, Jeremy 315, 320
Index of names 377
Kamil, Michael 36, 60 Larson, Steve 271, 286 Menéndez, Elvira 103
Kant, Immanuel 246–248, 250, Lawless, Harry 267, 286 Meyers, Karen 304, 307
252, 253, 263, 264, 323, 332 Lawley, Jim 104 Miller, Joan 135, 138, 143
Keil, Frank 24, 32, 41, 60 Lazar, Gillian 66, 79, 82, 85, 96, Montanari, Angela 105, 120
Keller, Dan 164 100, 102 Montañés, Elvira 104
Kennedy, John 304, 307, 308 Lee, Chia-Ying 204, 206 Moon, Rosamund 66, 79, 108,
Kettunen, Tom 141, 142, 143 Lee, Mark 79, 184, 239, 241, 243 113, 121, 306, 308, 366, 369,
Keysar, Boaz 187, 207, 310, 319 Lehrer, Adrienne 267, 268, 276, 371, 373
Khalil, Esam 295, 307 286 Moore, Kevin 106, 120
Kimmel, Michael 336, 354 Levorato, Maria Chiara 21, 31, 32 Mukařovský, Jan 323, 330, 332
Kircher, Mary 31 Li, Fuyin Thomas 57, 60 Müller, Ralph 151, 153, 164, 324,
Kittay, Eva 271, 286, 305, 307, Lillis, Theresa x, xii 332
339, 354 Lindstromberg, Seth 22, 31, 32, Müller-Prove, Matthias
Koester, Almut 291, 292, 308 82, 85, 96, 100, 102, 311, 320 151, 153, 164
Koivisto-Alanko, Päivi 137, 142 Linnell, Per 334, 354 Murphy, Gregory 187, 207
Koller, Veronika 126, 139, 142, Littlemore, Jeannette ix, xii, 58, Musolff, Andreas 133, 142, 143,
362, 373 60, 291, 292, 295, 308 218, 242, 271, 286, 328, 332,
Korsisaari, Eva Maria 128, 140, Little, William 363, 373 357, 358, 360, 361, 373
142 Liu, Ho-Ling 204, 206
Kossmann, Bianca 141, 142 Lonergan, Julia 11, 18 N
Kövecses, Zoltán 57, 60, 82, Louw, Bill 68, 79 Nagy, William 58, 60
85–87, 96, 100, 102, 106, 120, Lovejoy, Arthur 133, 142 NaLamlieng, Chumal 315, 320
128, 129, 131, 132, 136, 137, 140, Low, Graham ix, xi, xii, 3, 58, Nation, Paul 97, 103
142, 166, 242, 266, 287 60, 102, 166, 174, 183, 262, 266, Nedlinko, Anatoli 267, 276, 286
Krasowicz, Grasyna 106, 120 271, 285, 286, 291, 292, 295, Neiders, Ivars 128, 143
Krennmayr, Tina 166, 184, 306, 296, 304, 305, 306, 308, 353 Nerlich, Brigitte 22, 30, 32, 96,
308 Lowie, Wander 82, 96, 103 103, 120
Kronqvist, Camilla 141, 142, 143 Lüdeking, Karlheinz 254, 256, Nerlove, Harriet 23, 30, 31
Kucan, Linda 36, 59 264 Nevala, Minna 140, 143
Kulbrandstad, Lisa 36, 38, 60 Ludolph, Frank 164 Nippold, Marilyn 21, 32
Kurtyka, Andrzej 96, 102 Núñez, Raul 227, 242
Kytö, Merja 139, 142 M Nurmi, Arja 139, 143
Maalej, Zouhair ix, xii, 358, 373
L MacArthur, Fiona 25, 31, 32, 97, O
Laderman, Jeffrey 315, 320 101, 102 Oakhill, Jane 36, 60
Lakoff, George vii, vii, 9, 17, 18, MacLennan, Carol 96, 103 Oberholster, Anita 267, 286
22, 32, 43, 60, 64, 79, 120, 126, Mandel, Michael 314, 317, 320 O’Brian, Jennifer 48
128, 129, 134, 136, 139, 140, 142, Manfredi, Deanna 187, 207 Ojanen, Markku 140, 142
143, 169, 183, 185, 186, 205, Mark, Geraldine 104 O’Halloran, Kieran 360, 373
207, 210, 214–217, 220–225, Masci, Michaela 362, 373 Ortony, Andrew 21, 23, 32, 61,
237, 239, 242, 246, 249, 257, Maslen, Robert 334, 353 121, 207, 242, 307, 319, 354, 373
259–261, 264, 272, 295, 308, Matlock, Teenie 66, 67, 79 Özçalışkan, Şeyda 39, 60
324, 329, 332, 334, 341, 354, Matthews, Robert 262, 264
358–360, 373 Maule, John 334, 353 P
Lamers, Karl 328, 332 McEachrane, Michael 141, 142, Palermo, David 41, 61
Langacker, Ronald 22, 32, 215, 143 Palmer, Gary 106, 120
242 McGinnis, Martha 240, 242 Paprotté, Wolf 291, 308
Langer, Susanne 248, 264 McGlone, Matthew 187, 198, Panther, Klaus-Uwe 266, 286,
Larsen-Freeman, Diane 354 207, 210, 242 287
Larsen, Vibeke 48, 59 McKeown, Margaret 36, 59 Pantz, Valerie 105, 120
Larson, Kate 128, 142 McQuarrie, Edward 358, 373 Paradis, Carita 266, 286
Partington, Alan 271, 286 Rundell, Michael 100, 103, 168, Steen, Gerard ix, x, xii, 44, 59,
Pasma, Trijntje 165, 184, 306, 184, 311, 312, 320 60, 165–167, 169, 184, 207, 210,
308 230, 242, 243, 249, 264, 266,
Pearson, David 36, 60 S 271, 275, 285, 287, 294, 306,
Pennar, Karen 312, 320 Sag, Ivan 239, 242 308, 329, 332
Perkins, David 31 Sand, Andrea 139, 141 Stefanowitsch, Anatol 102, 126,
Perkins, Roderick 154, 164 Sandikcoiglu, Esra 358, 373 127, 141, 142, 143
Peyawary, Ahmad 97, 103 Sayce, Richard 291, 308 Stelma, Juurd 307, 333, 352, 354
Peynaud, Emile 267, 286 Schäfer, Thomas 264 Stengers, Hélène 65, 78, 79,
Pflüger, Friedbert 328, 332 Schank, Roger 164 352, 353
Philip, Gill x, 63, 66, 68, 69, Schäuble, Wolfgang 328, 332 Stern, Joseph 210, 230, 243
73, 79 Shneiderman, Ben 153, 164 Stewart, Mark 192, 207
Phillips, Barbara 358, 373 Schön, Donald 359, 373 Stratton, Peter 334, 353
Piaget, Jean 39, 105, 120 Schwarts, Nelson 315, 320 Stockwell, Peter 246, 264
Pickering, Neil 271, 287 Scott, Michael 314, 320 Stross, Brian 106, 121
Pickering, Martin 334, 354 Searle, John 108, 121 Stuart, Patricia 22, 23, 31
Pinot, Giuliana 282, 283 Semino, Elena 166, 329, 330, 332, Sullivan, Michael 21, 32
Piqué Angordans 310, 320 362, 373 Sun, Sewen 292, 308
Piquer-Píriz, Ana x, 21, 25, 30, Shen, Yeshayahu 187, 207 Superfine, Wendy 103–104
32, 96, 97, 102, 103 Shipton, Paul 104 Svanlund, Jan 51, 60
Pires de Oliveira, Roberta Short, Mick 2, 12, 26, 28, 33, 37, Swales, John 265, 287
139, 143 50, 52, 53, 90, 92, 95, 183, 186, Sweetser, Eve 137, 143, 227, 242
Pollio, Howard 303, 307 188, 212–214, 219, 222, 268, Szabó, Peter 82, 85, 87, 96, 100,
Ponterotto, Diane 304, 308 277, 284–285, 291, 292, 295, 102
Poole, Joshua 130, 143 296, 301, 303–305, 314, 329,
Pope, Rob 322, 323, 332 332, 338 T
Pragglejaz Group 7, 9, 18, 44, Shweder, Richard 135, 138, 143 Taylor, John 59, 82, 103, 142,
60, 165–167, 171–174, 182–184, Siemund, Rainer 139, 141 242
292, 294, 308, 352, 354, 365, Silverstein, Michael 268, 269, Tendahl, Marcus 13, 18
373 287 Teng, Norman 292, 308
Sinclair, John 103, 358, 372, 373 Thomä, Dieter 262, 264
R Skandera, Paul 139, 141 Thornburg, Linda 266, 286
Radden, Gunther 266, 286, Skorczynska, Hanna Tietz, Udo 264
287 xi, 309–311, 313–315, 320 Tillyard, Eustace M. W.
Ragazzini, Giuseppe 68, 79 Skorstad, Janice 105, 106, 120, 133, 143
Ramachandran, Vilayanur 241 Tissari, Heli x, 125–127, 130, 132,
266, 287 Smith, Anita 106, 120 137, 142–143
Rapson, Richard 140, 141 Smith, Geoffrey 286, 315, 320 Todd, Zazie vii, 22, 30, 32, 103,
Reilly, Vanessa 104 Smykowski, Tim 304, 307 120, 334, 353
Reynolds, Ralph 61 Smyth, Ron 304, 308 Trelles, Emma 103
Richards, Ivor 305, 308, 334, 354 Soegaard, Mads 150, 164 Tsur, Reuven 186, 207
Ricoeur, Paul 246, 249, 250, 264 Solska, Agnieszka 79, 102 Turner, Mark 10, 17, 18, 43, 60,
Ritchie, David 281, 287, 341 Solomon, Gregg 267, 287 152, 156–158, 160, 163, 164,
Roazzi, Antonio 106, 120 Solomon, Robert 128, 143 215, 224, 242, 246, 249, 257,
Rolston, William 342, 354 Sopory, Pradeep 359, 373 260, 261, 263, 264, 272, 324,
Roncero, Carlos 304, 308 Sökmen, Anita 36, 60 329, 332
Rorty, Richard 262, 264 Sperber, Dan 11, 12, 18, 210, 242
Rosenstiel, Anne 40, 61, 113, 121 Stacey, Les 106, 120 U
Rossant, John 318, 320 Stanley, Neil 334, 353 Ullmann, Stephen 357, 373
Rudzka-Ostyn, Brygida 100, 103 Starobin, Paul 318, 320 Urquhart, Sandy 36, 60
Index of names 379
V Weber, Jean Jacques 255, 264 Wood, Beatrice 263, 264, 317,
van Leeuwen, Theo 331, 332 Weir, Cyril 36, 60 332
Vega-Moreno, Rosa 210, 243 West, Gretchen 21, 31 Woodall, Pam 311, 320
Verspoor, Marjolijn 82, 96, 103 Wetz, Ben 104
Vestbøstad, Per 50, 61 Widdowson, Henry 251, 264 Y
Viberg, Åke 36, 37, 55, 59, 61 Wilson, Deidre 11, 12, 18, 210, Yafe, Tal 21, 31
Vosniadou, Stella 21, 23, 24, 32, 242 Yu, Ning 106, 121, 140, 143, 206,
39, 41, 61 Wilson, Paul 61 266, 272, 287, 355
Winkler, Heinrich 325, 332
W Winner, Ellen 21, 23, 24, 30–32, Z
Waggoner, John 41, 61 40, 61, 105, 107, 108, 113, 114, Zanatta, Theresa 104
Waldron, Ronald 357, 362, 373 119–121 Zhou, Rong 106, 121
Wallington, Alan x, 79, 210, Wodehouse, P. G. 294, 308 Ziegler, Dominic 317, 320
213, 224, 225, 229, 239–241, Wolf, Dennie 105, 120 Zill, Rüdiger 264
243, 272 Wolff, Phillip 295, 307 Zingarelli, Nicola 68, 79
Wasow, Thomas 242 Zurer Pearson, Barbara 21, 32
Index of terms
A multi-modal blend 145–146, 166, 176, 204, 210, 214, 237,

advertising 163, 358, 360 149–151, 153, 155, 157–159, 245, 255, 334
aesthetics 17, 245–262, 322, 323, 162 cognitive linguistic analyses
330 Web principle 156, 160, 164 8–9
aesthetic distance 248 body 2, 8, 25, 40, 129, 136, 138, cognitive approach vii, 55,
see also Befremdung, 140, 176, 178, 180, 182, 359, 100, 246–247, 250, 261,
detachment, estrangement, 366, 370 357–358, 362, 372
objective disinterestedness bodily experience 8–9, 22, collocation 63–64, 68–70,
alliteration 297, 324 106, 176, 341 72–76, 78, 272, 340, 363, 364,
analogy 30, 120, 145, 211, 216, embodiment 17, 29, 119, 367–369
237, 238, 246, 251, 295, 298, 142, 175, 177, 257, 341, 343, collocation error 64, 69, 70,
289, 304–305, 330 347, 352 72, 73
art 17, 158, 247–263, 322, 326 see also domain, collocational patterning
artist 250, 253, 256 experiential 71, 72, 74, 77
artistic values, valuation physical action ix, 16, 23, commodity 66, 125–134, 138, 188
252–256 109, 128, 131, 146–148, complex dynamic systems
ATT-Meta framework 210–239 152–158, 160, 163, 164, 167, theory 334–335
Attributive Property Theory 10 171, 175, 179, 183, 226, 232, see also discourse dynamics
attributive categorization 251–253, 311, 333–353, 363, approach, dynamic systems
hypothesis 187, 191, 195 365, 366 compound 44, 87–89, 171–172,
attitudes 106, 138, 246–263, physical or social interaction 186, 336
340–351, 370 vii, xi, 170, 255–228, compression see blend
230–232, 236, 240, 305, concept-to-concept mapping
B 334, 335, 341 212, 239
Bank of English 344, 348–349, Bokmålskorpuset 50 Conceptual Mapping Model
365–366 book see textbook 187–193, 195, 198, 202–204
BASE corpus 296, 306 British National Corpus (BNC) conceptual metaphor theory
basic meaning 39, 43, 44, 165– 68, 72, 76, 78, 88, 126, vii–viii, xi, 3, 10, 43, 49, 57,
183, 294, 365 167–174, 181–182, 218, 293 64, 185–187, 205, 212, 215–216,
see also core meaning, 230, 239, 351, 360, 368
prototypical meaning C conceptual fluency 67, 69
Befremdung 257, 259 Career of Metaphor Theory 10 connotation 29, 41, 136, 161, 233,
beliefs 8, 14, 133, 139, 177, 191, cline 240, 277–278, 280, 285, 321 280, 358, 362–364, 368–369,
236, 330 cluster 371
blending metaphor cluster 209, 214, container (schema) 126–129,
blending of primary metaphors 301, 303–4, 307, 334 133–134, 136, 137, 140, 152, 226,
239 word cluster 324 233–234, 236, 241
Blending Theory, conceptual cognitive effect see Relevance context vii–viii, x–xi, 2, 5–6,
blending 10, 145–164 Theory 11–15, 22, 28, 35, 40–41, 43, 45,
compression 152, 157–158, 163 cognitive linguistics vii, 2, 8–9, 49–51, 55–58, 70, 73, 75–76,
integration 155–157, 160, 24–25, 31, 49, 95, 126, 139–140, 78, 85, 97, 99–100, 106, 108,
161, 163 119, 129, 136, 138–139, 147, 163,
165, 167–171, 173–175, 179, 205, diachronic language analysis E

210, 213–214, 217, 219–220, 127, 136, 138, 141, 213 Early English Correspondence
224, 230, 232, 236–237, 245, diachronic approaches to Sampler 127, 133, 140
248–249, 251, 253–256, meaning 175–178 ecological validity viii, 3, 165,
258–262, 265, 268, 270–271, decoration (metaphor as) 95, 311 309
275–276, 278–279, 282–285, delexicalisation 63, 69–70, EFL (or ELT) ix, 21–31, 82–101
292, 294, 306, 309–310, 75, 77 elicitation 40, 78
314, 318–319, 324, 326, 330, desktop metaphor 145–164 embodiment see body
333–334, 337, 341, 344–345, see also virtual desktop emergent structure 156, 159,
348–349, 364–365–366 detachment 139, 247–250, 253, 161–162
contextual implication 13–15 256 entailment 358–364
contextual meaning 167–170, see also aesthetic distance, equivalence
173–174, 179, 181–182, 352, Befremdung, estrangement, L1–L2 equivalence 51, 54–55,
365 objective disinterestedness 71
conversational implicature 12 discourse Principle of equivalence 251
core mapping 210–212, 220, 223, discourse approaches 141 Estrangement 257
225, 227–230, 236, 238 discourse analysis 139, 166, see also aesthetic distance,
core meaning 26, 86, 88–89, 255, 360 Befremdung, detachment,
91–92, 98, 101, 179 discourse dynamics approach objective disinterestedness
see also basic meaning, 334, 351–352 etymology 176–178, 363
prototypical meaning see also complex dynamic evaluation xi, 46–48, 51, 74,
corpus systems theory, dynamic 109, 129, 139, 260, 269, 283,
combining large and small systems 321, 324, 328, 330, 357–358,
corpora 7, 333, 344 disenablement 216–217, 223–225, 361–362, 364, 368–371
corpus analysis xi, 63–78, 230, 239 evaluation of creativity 322
125–141, 309–331, 357–372 domain evaluation of emotions 131–
COSMAS corpus 326, 327, conceptual domain, 3–4, 8–9, 132, 138
329, 331 16, 21, 23–25, 30, 40, 41, evaluation of image strength
large-scale corpus 190, 324, 43–44, 99, 126, 145, 146, 51
327, 344, 351, 361 151, 154, 156–158, 160–163, evaluative response 109, 110,
learner corpus 64 172, 185–195, 203–205, 116, 203
small-scale corpus 21–58, 209–220, 223–224, metaphor evaluation see
81–101, 145–164, 265–285, 227–229, 234–239, 248, metaphor, topics
291–307, 333–353, 371 251, 252, 255, 259, 260, 266, see also value
course book see textbook 271–277, 284, 285, 293, 294, event structure 231–232, 235,
creativity xi, 72–3, 252, 270–271, 305, 317, 325–327, 330, 331, 240
281, 284, 321–331 333, 338–339, 343, 347, 351, experiment viii, x, 2, 10, 15,
creative metaphor 321–325, 358–363, 365, 368, 369, 371 48, 57, 96, 107–108, 114–116,
330–331 domain-to-domain mapping 118–119, 147, 151, 160, 186–188,
identifying creative metaphor 212, 214–216, 223, 237 191–195, 197–203, 205–206,
324–325 experiential domain 266, 256–257, 260, 304
Critical Discourse Analysis 139, 272, 273–276 expert 57, 67, 71, 147, 150–151,
320–331, 360 psychological domain 106, 153, 163, 165, 253–254, 262,
cross-sensory interpretation 113, 119 265–267, 270–272, 283–284,
109–110, 116 dual-property adjectives 23 324, 359
(cross) cultural differences dynamic systems 333–335, 344, expert knowledge 147
x, 96, 106–119, 136, 161 351 expletive 298
see also complex dynamic extension 39, 71–72, 82, 85–86,
D systems theory, discourse 88–91, 95–97, 100–101, 146,
diachronic dynamics approach 177–178, 186, 204, 322, 324,
diachronic dimension 5, 213 329
Index of terms 383
meaning extension 65, 83, Helsinki Corpus of English journalism 310, 326
95, 97, 100–101 127, 129
semantic extension 25, 30, homograph 71 L
82, 90–91, 94, 96, 99–100 human-computer interaction language
145–162 comprehension 105–119
F hybrid interference 64, 75
factoring out 211, 222 hybrid form 351 learning see first language
figurative schemas 266, 270, 283 hybrid modality 152 acquisition, second
identification of figurative hyperbole 281, 298, 306, 336, language acquisition
schemas 270–272 339, 370 production 63, 65, 67
first language acquisition 39–58, transfer 64
65, 22–24, 30, 41, 50, 56, I see also translation
105–119, 157, 160, 246, 265, idiom 21, 67–68, 69, 71, 75, 76, learning 151, 156, 161
268, 270 86–87, 89, 90, 92, 96, 107, see also first language
force 112–113, 118, 172–173, 226, 240, acquisition, second
force as source domain 298, 366–367, 369, 371 language acquisition
128, 131, 132, 329 image lexical unit 43, 165–183, 294, 365
illocutionary force 150 = picture 130, 158, 219 linguistic minority 35–59
evaluative force of metaphor = visualisation 44, 49, 51–57, list (in discourse) 259, 300–303,
xi, 361, 362 255, 257–259, 265–284, 362 305
foreign language see second mental image 73, 98, 113, 136 literalisation 338, 339, 343, 351
language acquisition image schema 22, 237, 352
frame, framing 151, 163, 204, 215, see also container M
268, 283, 358 integration Macmillan English Dictionary
Freiburg-Brown Corpus of conceptual integration for Advanced Learners
American English 127, 134, 140 see blend 168, 169, 172, 174, 177–181, 310
Freiburg-LOB Corpus of British integration of sentences Map-Transcending Entities
English 127, 140 in discourse 204 213, 220, 239
frequency 6, 16, 37–38, 49–50, model integrated mapping principle 185–206
53, 55, 58, 84–85, 87–95, component see MIC mental image 73
100–101, 106, 110, 114, 116–119, integration of metaphor into metaphor, types
125, 130, 138, 186, 191–192, argument or figure of correlational 10, 22, 31, 204,
194, 198, 202–204, 218, 291, speech 324, 328 205, 207, 242
358, 368 integrating mental and social cross-sensory 106–107
functional range of metaphor phenomena 341–343 decontextualised 3, 108, 165
309-319 interface 8, 145–164 linguistic 13, 73, 182, 310, 313,
fuzzy boundaries 49, 284, 336, see also human-computer 314, 329, 336, 341, 352, 359,
367 interaction; WIMP 361, 364
interface marked 314, 316
G interlanguage 64, 68, 71 see also metaphorical
genre 38, 50, 136, 175, 246, intertextuality 258, 301, 324 marker
250–253, 255–259, 265–266, intuition 4–5, 7–9, 12, 67, 186, metonymy-based 101, 370
291, 305, 311, 318, 323–324, 191, 192, 203, 260 mixed 16, 147, 330
326–328 intuitiveness 12, 148–162 negative 14–15
gesture 17, 87, 152–153, 157 invariance 15, 210–241 non-sensory 23–24
Great Chain of Being 133 Invariance Principle 214, 237 novel 7, 9, 10, 73, 185–187,
irreversibility of metaphor terms 191–198, 200–205, 306,
H 157 321, 324
hedge, hedging 295, 297–298, overarching xi, 304
300, 310, 312–313 J physical-psychological
complex hedging 297–298 jargon 136, 158, 265, 271, 280 23, 107, 113
pictorial 17, 147 266, 268, 336, 339, 342–345, perceptual interpretations
poetic 245–263, 321, 329 349–353, 366, 367, 370 107, 111, 113–114, 116, 118
positive 14 metonymy-based metaphor persuasion 358
primary 8, 10, 31, 148, 152, see metaphor, types phrasal verb 85, 92–93, 96–100,
186, 192, 205, 210, 213, 218, MIP 166, 170–171, 173, 175, 179, 102–103, 171, 172, 340
239, 240 294, 296 phraseological patterning
relational 23 MIPVU 294 64, 66, 75–77
resemblance 9–10, 22, 31, minority see linguistic minority poetry viii, 1, 95, 106, 246–253,
204, 205 mirror neurons 341 255, 257–261, 301, 329
sensory 23, 106–107 mnemonic 153 poet, 246, 257, 278, 300
theory-constitutive 305 model-integrated component poetic metaphor
topic-triggered 324 (MIC) 230–236 see metaphor, types
unmarked 314, 316–318 moral 74, 127–129, 132–133, 138, poetic function 248, 251, 323
metaphor, topics 182, 246, 253, 256, 261, 323, poeticity 249–251, 256, 258
bridge term 339 330 political discourse xi, 321–331,
identification viii, x, 5, moral systems 139 360
25–26, 44, 64, 126, 165–167, morality 140 polysemy 24–25, 82–83, 86, 89,
174–175, 212, 249, 251, 268, motor-memory reinforcement 95–96, 100–101, 364–365
270–272, 277, 283, 291–295, 153 polysemous 22, 25, 30, 36,
305–306, 310, 314, 319, 321, multiword expression 50, 53, 55, 55, 70, 82, 85–86, 91–92,
325, 329, 352, 365 126, 358, 366, 367, 371 95–96, 175, 179, 225
evaluation 312–331, 357–372 see also polyword polyword 171, 172
metaphor understanding music 3, 158–160, 302 see also multiword expression
4–5, 11, 13, 22, 57, 186, 213 musical 4, 158, 273 Pragglejaz Group 7, 9, 44,
metaphorical understanding musician 73, 178, 210 165–183, 292, 294, 352, 365
14, 48 processing time 13, 192, 202, 203
metaphor comprehension N prototype 25, 28, 30, 49, 86, 89,
4–5, 40, 42, 192 negotiation 292, 328 91, 152, 157, 277, 279, 293, 295
metaphorical comprehension neuropsychology 5 prototypical meaning 25, 28,
38–39, 48, 56 novel metaphor see metaphor, 30, 86, 89, 91
in spontaneous conversation types see also basic meaning,
333, 339, 351, 360, 362 novelty xi, 10, 187, 198, 206, core meaning
social and dialogic factors 285, 326 proverb 87, 88, 97, 98, 105, 107,
334 112, 118
metaphorical marker 292–295, O psychological interpretation
305, 310, 312–314, 316, 319, 327 objective disinterestedness 107, 112–114, 118–119
see also marked metaphor, 248, 250
unmarked metaphor see also aesthetic distance, R
metaphor modification 328 Befremdung, estrangement, Relevance Theory 12–13, 15,
metaphor scenario 328, 358, detachment 210, 238
360–361, 363 objet trouvé 252, 256–258, 260, cognitive effects 11–15
metaphoric simile xi, 292–294, 262 reconciliation discourse
296, 304–305 observation (as research 332–352
see also simile method) 119, 157 reliability x, 7, 17, 165–167, 182,
Metaphors of Mind database 267, 285, 321, 326, 328
218, 241 P
metaphoreme 336, 340, 343, 345, perceptual 23, 41, 106–107, S
350–351 109–114, 116–119, 257, 260– salience, saliency 31, 41, 59, 69,
metonymy 16, 22, 24, 25, 66, 83, 261, 280, 312–313, 315–316 72, 73, 75, 77, 130, 161, 192,
86–87, 89, 95, 134, 146, 174, perceptual explanations 224, 266, 267, 270, 276, 278,
176, 181, 217, 251, 258, 259, 110–112, 117–119 295, 305
Index of terms 385
scenario 186, 328, 338, 339, T vehicle 2, 23–24, 209, 221, 223,
348–350, 353, 358, 360–364 Tasting note (wine) 266, 268, 285, 291–292, 294, 310, 312,
schema-refreshment 330 283 314–319, 331, 333, 339, 343, 347,
science 139, 163, 175, 214, 246, temperature words 21–30, 44, 351, 357
267, 270, 278, 285, 292, 296, 89–91, 280 virtual desktop 145–146, 149,
299–302, 305, 314, 322 textbook x, 35–58, 81–101 156–162
script 149, 150, 156, 161, 163, 164 top-down 38, 149, 191, 319 see also desktop metaphor
second language acquisition/ topology 156, 158, 163, 237 visual iconicity, icon 118, 205,
learning 5, 21–30, 35–58, topos 130 152–156
63–78, 81–101, 295 translation 5, 51, 67, 68, 70–74, vocabulary 22, 25, 36–41, 43,
selectional restriction rules 2 77, 97–99, 160, 254 46–50, 55–58, 65–66, 78, 82,
semantic field 36 typology 85, 96–99, 158
simile xi, 22, 291–306, 312–314, lexical typology 36–38 vocabulary learning 99
317 typological range View-Neutral Mapping Adjunct
simile identification of metaphor 309–319 (VNMA) 224, 227–238, 240
293–294, 306
non-metaphoric simile V W
292–293 value Web principle see blend
simulation 341–343, 354–356, 351 aesthetic value/valuation WIMP interface 146, 151–155,
somatic marker 341, 351 252–256, 260, 261, 330 160, 162
Structure Mapping Theory 237 communicational value 15 see also human-computer
stylistics 250–251, 261, 331 cultural value 29, 127, 139, interaction, interface
stylistic diagnostics 250 253, 260, 262 word class 171, 174–175
stylistic deviation 323 moral value 127, 133, 138 grammatical class 82, 85
symbolisation 312, 336, 337, 339, positive/negative 74, 129,
341, 343–345, 351 345, 351
synaesthesia 4, 266, 276–277 value judgement 229–238
synchronic 5, 22, 127, 174–176, value system 132, 133
178, 183, 213 see also evaluation
In the series Human Cognitive Processing the following titles have been published thus far or
are scheduled for publication:
26 LOW, Graham, Zazie TODD, Alice DEIGNAN and Lynne CAMERON (eds.): Researching and Applying
Metaphor in the Real World. 2010. xii, 385 pp.
25 PANTHER, Klaus-Uwe, Linda L. THORNBURG and Antonio BARCELONA (eds.): Metonymy and
Metaphor in Grammar. 2009. xiii, 423 pp.
24 EVANS, Vyvyan and Stéphanie POURCEL (eds.): New Directions in Cognitive Linguistics. 2009.
xi, 519 pp.
23 HOLŠÁNOVÁ, Jana: Discourse, Vision, and Cognition. 2008. xiii, 202 pp.
22 BERENDT, Erich A. (ed.): Metaphors for Learning. Cross-cultural Perspectives. 2008. ix, 249 pp.
21 AMBERBER, Mengistu (ed.): The Language of Memory in a Crosslinguistic Perspective. 2007. xii, 284 pp.
20 AURNAGUE, Michel, Maya HICKMANN and Laure VIEU (eds.): The Categorization of Spatial Entities
in Language and Cognition. 2007. viii, 371 pp.
19 BENCZES, Réka: Creative Compounding in English. The Semantics of Metaphorical and Metonymical
Noun-Noun Combinations. 2006. xvi, 206 pp.
18 GONZALEZ-MARQUEZ, Monica, Irene MITTELBERG, Seana COULSON and Michael J. SPIVEY
(eds.): Methods in Cognitive Linguistics. 2007. xxviii, 452 pp.
17 LANGLOTZ, Andreas: Idiomatic Creativity. A cognitive-linguistic model of idiom-representation and
idiom-variation in English. 2006. xii, 326 pp.
16 TSUR, Reuven: ‘Kubla Khan’ – Poetic Structure, Hypnotic Quality and Cognitive Style. A study in mental,
vocal and critical performance. 2006. xii, 252 pp.
15 LUCHJENBROERS, June (ed.): Cognitive Linguistics Investigations. Across languages, fields and
philosophical boundaries. 2006. xiii, 334 pp.
14 ITKONEN, Esa: Analogy as Structure and Process. Approaches in linguistics, cognitive psychology and
philosophy of science. 2005. xiv, 249 pp.
13 PRANDI, Michele: The Building Blocks of Meaning. Ideas for a philosophical grammar. 2004. xviii, 521 pp.
12 EVANS, Vyvyan: The Structure of Time. Language, meaning and temporal cognition. 2004. x, 286 pp.
11 SHELLEY, Cameron: Multiple Analogies in Science and Philosophy. 2003. xvi, 168 pp.
10 SKOUSEN, Royal, Deryle LONSDALE and Dilworth B. PARKINSON (eds.): Analogical Modeling. An
exemplar-based approach to language. 2002. x, 417 pp.
9 GRAUMANN, Carl Friedrich and Werner KALLMEYER (eds.): Perspective and Perspectivation in
Discourse. 2002. vi, 401 pp.
8 SANDERS, Ted, Joost SCHILPEROORD and Wilbert SPOOREN (eds.): Text Representation. Linguistic
and psycholinguistic aspects. 2001. viii, 364 pp.
7 SCHLESINGER, Izchak M., Tamar KEREN-PORTNOY and Tamar PARUSH: The Structure of
Arguments. 2001. xx, 264 pp.
6 FORTESCUE, Michael: Pattern and Process. A Whiteheadian perspective on linguistics. 2001. viii, 312 pp.
5 NUYTS, Jan: Epistemic Modality, Language, and Conceptualization. A cognitive-pragmatic perspective.
2001. xx, 429 pp.
4 PANTHER, Klaus-Uwe and Günter RADDEN (eds.): Metonymy in Language and Thought. 1999.
vii, 410 pp.
3 FUCHS, Catherine and Stéphane ROBERT (eds.): Language Diversity and Cognitive Representations.
1999. x, 229 pp.
2 COOPER, David L.: Linguistic Attractors. The cognitive dynamics of language acquisition and change.
1999. xv, 375 pp.
1 YU, Ning: The Contemporary Theory of Metaphor. A perspective from Chinese. 1998. x, 278 pp.

(Human Cognitive Processing, V. 26) Graham Low - Et Al - Researching and Applying Metaphor in The Real World-John Benjamins Pub. Co (2010)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Human Cognitive Processing, V. 26) Graham Low - Et Al - Researching and Applying Metaphor in The Real World-John Benjamins Pub. Co (2010)

Uploaded by

Copyright:

Available Formats

Researching and Applying Metaphor in the Real World

Human Cognitive Processing (HCP)

Editorial Advisory Board

John Benjamins Publishing Company

American National Standard for Information Sciences – Permanence of

Library of Congress Cataloging-in-Publication Data

© 2010 – John Benjamins B.V.

Editors’ introduction vii

Section 1. Metaphor and language learning

Section 2. Capturing and analysing metaphors

9. Pragglejaz in practice: Finding metaphorically used words

Section 3. The function of metaphor in discourse

Index of names 375

learning metaphors in a foreign language; a deeper understanding of these will

Together, the collection provides a snapshot of real-world metaphor studies that

without compromise, data of real-world phenomena, qualities that we hope char-

The wonderful, chaotic, creative,

This chapter provides a brief overview of the many intellectual accomplish-

Keywords: metaphor, real world, future research

challenges to contemporary theories of language processing and human cognition,

2. The scope of real-world metaphor research

I want to begin by suggesting a number of general propositions about what is

Lastly, real-world metaphor research needs to explore situations which are as

4. Important analytic advances

Despite the continued existence fundamental questions, real-world metaphor re-

4.1 The messy reality of metaphor use

Metaphor is evident throughout language, thought, and culture. But metaphor

4.2. Attention to discourse, not single utterances

4.3 Importance of corpus research

­ etaphors, which has resulted in greater caution in scholars relying solely on

4.4 Attention to reliability of analyses

An increasing concern with the variability of analysts’ intuitions in making judg-

4.5 Greater attention to quantitative analyses

Claims about the importance or ubiquity of particular metaphorical patterns, in

4.6 Questions about language and thought links

4.7 Complexity of body, mind, language and culture interfaces

Finally, in the interdisciplinary metaphor community, scholars have become in-

5. Some peeks into the future

5.1 How do you do your analyses?

(2) Einstein gave birth to the theory of relativity.

(3) Those ideas died off in the Middle Ages.

5.2 Different metaphors – different theories?

of your argument” is thought to depend on accessing a well-established concep-

5.3 What do metaphors mean?

One of the most remarkable aspects of much contemporary metaphor research

both a metaphorical assertion and a conversational implicature. On the surface, it

In general, the issue of how to individuate and count metaphor meanings

5.4 How are metaphors used?

My discussion about trying to understand which meanings of a metaphor are

New information context – contextual implication

5.5 Mixed and multiple metaphors

My final speculation on a future trend in metaphor research is the study of meta-

Cameron, Lynne (2003). Metaphor and educational discourse. London: Continuum.

Metaphor and language learning

Can people be cold and warm?

The understanding and production of figurative language in childhood has been

Keywords: figurative meanings, young learners, EFL, temperature terms

Children’s ability to understand and produce figurative meanings in their native

that often takes the linguistic form of a simile expressed as ‘A is like B’ or ‘A is

2. Warm and cold and domain knowledge

have mainly concentrated on adolescent or adult learners of EFL. However, EFL

Two research questions were addressed in this study:

3.2 Stimuli and procedure

Editors’ introduction vii

Index of names 375

etaphors, which has resulted in greater caution in scholars relying solely on

In Denmark, Golden and Hvenekilde’s vocabulary study was duplicated by Gimbel

e xamples in each category, with the ambiguous or inappropriate examples exclud-

advocated teaching the students various vocabulary acquisition strategies. Such