You are on page 1of 456

From Sign to Signing

AUTHOR "Wolfgang G. Mller, Olga Fischer and"
TITLE "From Sign to Signing: Iconicity in Language andLiterature 3"
WIDTH "150"
VOffSET "4">
From Sign to Signing
Iconicity in language andliterature 3
Edited by
Wolfgang G. Mller
Friedrich-Schiller-Universitt Jena
Olga Fischer
University of Amsterdam
The paper used in this publication meets the minimum requirements of American
National Standard for Information Sciences Permanence of Paper for Printed
Library Materials, ANSI Z39.48-1984.
Library of Congress Cataloging-in-Publication Data
Symposium on Iconicity in Language and Literature (3rd: 2001: Jena, Germany)
From sign to signing: iconicity in language and literature 3 / edited by Wolfgang G. Mller, Olga
p. cm.
... a selection of papers that were originally given at the Third Symposium on Iconicity in Language
and Literature organized by the University of Jena in co-operation with the University of Amsterdam
and the University of Zurich and held at Jena, 29-31 March, 2001--Preface.
Includes bibliographical references and index.
1. Iconicity (Linguistics)--Congresses. 2. Philology--Congresses. 3. Sign language--Congresses.
4. Semiotics--Congresses. I. Mller, Wolfgang G. II. Fisher, Olga. III. Friedrich-Schiller-Universitt
Jena. IV. Title.
P99.4.I26 S96 2002
302.2--dc21 2002028004
ISBN 90 272 2593 1 (Eur.) / 1 58811 288 8 (US) (Hb; alk. paper)
2003 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
TITLE "Table of contents"
WIDTH "150"
Table of contents
Preface andacknowled gments
< R E F
< / R E F
List of contributors
< R E F
< / R E F
Introduction: From Signing back to Signs
< R E F
< / R E F
Olga Fischer and Wolfgang G. Mller
P:1 I
Auditory and visual signs and signing
The inuence of sign language iconicity on semantic
< R E F
conceptualization 23
< / R E F
Klaudia Grote and Erika Linz
What You See Is What You Get: Iconicity andmetaphor in the visual
< R E F
language of written andsignedpoetry: A cognitive poetic approach 41
< / R E F
William J. Herlofsky
Spatial iconicity in two English verb classes
< R E F
< / R E F
Axel Hbler
What imitates birdcalls?: Two experiments on birdcalls and their
< R E F
linguistic representations 77
< / R E F
Keiko Masuda
P:1 II
Visual iconicity and iconic mapping
Perspective in experimental shapedpoetry: A semiotic approach
< R E F
" w h
</REF "wh
John J. White
Where reading peters out: Iconic images in the entropic text
< R E F
< / R E F
Julian Moyle
Iconic representation of space andtime in Vladimir Sorokins novel
< R E F
The Queue (Ochered) 153
< / R E F
Andreas Ohme
Vision andPrayer: Dylan Thomas andthe Power of X
< R E F
< / R E F
Matthias Bauer
Diagrams in narrative: Visual strategies in contemporary ction
< R E F
< / R E F
Christina Ljungberg
Structural iconicity
The iconicity of Afrikaans reduplication
< R E F
< / R E F
C. Jac Conradie
Diagrammatic iconicity in the lexicon: Base andd erivation in the
< R E F
history of German verbal word-formation 225
< / R E F
Volker Harm
Creative syntax: Iconic principles within the symbolic
< R E F
< / R E F
Beate Hampe and Doris Schnefeld
Aspects of grammatical iconicity in English
< R E F
< / R E F
Gnter Rohdenburg
Beatrice: or The geometry of love
< R E F
< / R E F
Wilhelm Ptters
How metaphor andiconicity are entwinedin poetry: A case in Haiku
< R E F
" h
< / R E F
" h
Masako K. Hiraga
P:1 IV
Intermedial iconicity
Intermedial iconicity in ction: Tema con variazioni
< R E F
< / R E F
Werner Wolf
Iconicity andliterary translation
< R E F
< / R E F
Elzbieta Tabakowska
P:1 V
New applications of sign theory
Iconizing literature
< R E F
< / R E F
Jrgen Dines Johansen
From signal to symbol: Towards a systems typology of linguistic
< R E F
signs 411
< / R E F
Piotr Sadowski
Author index
<REF "n
< / R E F
" n
Subject index
<REF "s
< / R E F
" s
TITLE "Preface andacknowled gments"
WIDTH "150"
Preface and acknowledgments
The studies collected in this volume represent a selection of papers that were
originally given at the ThirdSymposium on Iconicity in Language and Literature,
organizedby the University of Jena in co-operation with the University of
Amsterdam andthe University of Zurich andheldat Jena, 2931 March, 2001.
The essays included here exemplify a wide range of new approaches and new
research material in iconicity studies both in language and literature. They show
that iconicity remains an important andfruitful topic for interdisciplinary work
opening up new elds for further research.
The organizers of the conference gratefully acknowledge the support of
institutional sponsors such as the Deutsche Forschungsgemeinschaft, the Univer-
sity of Jena andthe Faculteit der Geesteswetenschappen of the University of
Amsterdam. The University of Jena generously oeredits facilities to the
conference andgave logistic help. The symposium proted greatly from the
competent andenthusiastic assistance given by Dr. Eva-Maria Orth andJens
Mittelbach. Jens Mittelbachs expertise on the computer also considerably lightened
the burden of the organizers and editors. We are most grateful to Jens Mittelbach
andMarlene von Frommannshausen for their technical assistence with the index.
Very special thanks go to Max Nnny of the University of Zurich, who
helpedto organize the symposium andwhose expertise andknowled ge of the eld
was invaluable to the editors in their selection of the contributions for the present
volume. Max Nnny is the spiritus rector of literary iconicity studies and it is his
enthusiasm andcharisma which rst startedthis series of symposia on iconicity.
We wouldalso, once again, like to thank Kees Vaes of John Benjamins Ltd.
for his advice and concern as regards the volume and for helping us to see it
through the press.
A nal wordof thanks must of course go to the active participants whose
enthusiasm, detailedcase studies andinnovative approaches contributedto
making the symposium andwe hope this volume too a success.
We are indebted to the following for permission to reproduce copyright
material: in the essay by John White the illustration of Francesco Canguillo,
Milano dimostrazione 1915 is reproduced by permission of DACS (the Design
and Artists Copyright Society Ltd.); the texts from Peter Readings works in
Julian Moyles essay are reproduced by permission of Bloodaxe Books Ltd.
W. G. M. and O. F.
</TARGET "pref">
TITLE "List of contributors"
WIDTH "150"
List of contributors
Matthias Bauer
Institut fr Anglistik/Amerikanistik
Universitt des Saarlandes
PF 151150
D-66041 Saarbrcken
C. Jac Conradie
Department of Afrikaans
RandAfrikaans University
P. O. Box 29139
2109 Melville
South Africa
Olga Fischer
Engels Seminarium
Universiteit van Amsterdam
Spuistraat 210
1012 VT Amsterdam
The Netherlands
Klaudia Grote
Universitt zu Kln
Forschungskolleg Medien und kulturelle
Kommunikation SFB/FK 427
Bernhard-Feilchenfeld-Strae 11
D-50969 Kln
Beate Hampe
Institut fr Anglistik/Amerikanistik
Friedrich-Schiller-Universitt Jena
Ernst-Abbe-Platz 8
D-07743 Jena
Volker Harm
Deutsches Wrterbuch
Akademie der Wissenschaften zu
Herzberger Landstrae 2
D-37085 Gttingen
William J. Herlofsky
Faculty of Foreign Languages
Nagoya Gakuin University
1350 Kamishinano
48012 Japan
Masako K. Hiraga
School of Sociology
Rikkyo University
3341 Nishi-Ikebukuro
1718501 Japan
Axel Hbler
Institut fr Anglistik/Amerikanistik
Friedrich-Schiller-Universitt Jena
Ernst-Abbe-Platz 8
D-07743 Jena
Jrgen Dines Johansen
Department of Literature andSemiotics
Odense University
55 Campusvej
5230 Odense
Erika Linz
Universitt zu Kln
Forschungskolleg Medien und kulturelle
Kommunikation SFB/FK 427
Bernhard-Feilchenfeld-Strae 11
D-50969 Kln
Christina Ljungberg
Englisches Seminar
Universitt Zrich
Plattenstrasse 47
CH-8032 Zrich
Keiko Masuda
Department of Linguistics
University of Cambridge
Sidgwick Avenue
UnitedKingd om
Julian Hillyer Moyle
13 Churchill Road
CV22 6BT
UnitedKingd om
Wolfgang G. Mller
Institut fr Anglistik/Amerikanistik
Friedrich-Schiller-Universitt Jena
Ernst-Abbe-Platz 8
D-07743 Jena
Andreas Ohme
Institut fr Slawistik
Friedrich-Schiller-Universitt Jena
Ernst-Abbe-Platz 8
D-07743 Jena
Wilhelm Ptters
Institut fr Romanische Philologie
Universitt Wrzburg
Am Hubland
D-97074 Wrzburg
Gnter Rohdenburg
Institut fr Anglistik/Sprachwissenschaft
Universitt Paderborn
Warburger Strae 100
D-33098 Paderborn
Piotr Sadowski
American College Dublin
2 Merrion Square
Dublin 2
Doris Schnefeld
Englisches Seminar
Universittsstrae 150
D-44780 Bochum
Elzbieta Tabakowska
Institute of English
Jagiellonian University
Al. Mickiewicza 9/11
31120 Krakow
John White
Department of German
Kings College
London WC2R 2LS
UnitedKingd om
Werner Wolf
Institut fr Anglistik
Karl-Franzens-Universitt Graz
Heinrichstrasse 36
A-8010 Graz
</TARGET "loc">
AUTHOR "Olga Fischer andWolfgang G. Mller"
TITLE "Introduction"
WIDTH "150"
From Signing back to Signs
Olga Fischer Wolfgang G. Mller
University of Amsterdam University of Jena
Iconicity is one of the few elds of research in which the disciplines of linguis-
tics and literary studies both of which have regrettably drifted apart as a
consequence of specialization can fruitfully co-operate. We see iconicity (be
it of an imagic or a diagrammatic kind)
operating in everyday as well as literary
language. Andit is through the interdisciplinary nature of the iconicity symposia,
through approaching iconic phenomena from both a linguistic andliterary point
of view, that we may develop a keener perception of the pervasive presence of
iconicity in all forms of language. This will provide us with a better understand-
ing of how language is structuredand at the same time give us a deeper insight
into the tools andmethod s usedby poets andwriters, leading to a fuller appreci-
ation of the literary text itself.
In this thirdvolume in the series, we nd a number of new departures. One
of these is the interest in gestures, andmore specically, in signing: the
gestural mode of signed languages. The other is the concern with intermedial
iconicity, which we shall discuss in connection with the contributions found in
Part IV. The interest of scholars of signedlanguages in iconicity is fairly recent.
Before the 1960s andup to the 70s, signedlanguages were generally regardedas
rather primitive languages devoidof prepositions, conjunctions andabstract
words (Eisenson andBoase 1950, quotedin Battison 2001: 6); languages that
were boundto the concrete andrather limitedwith respect to abstractions,
humor andsubtleties such as gures of speech which enrich expression (Davis
andSilverman 1970, in Battison 2001: 6). Signed languages only came to be seen
as interesting in their own right after the Chomskyan revolution, that is, when
transformational-generative linguists, with their mentalistic approach to language,
focused on the machinery producing language rather than on the product itself.
This took the attention away from the still exclusive focus on aural andoral
modes of communication andopenedup possibilities for another channel, the
visual one used by signers. Indeed, new research into signed languages, which
was greatly inspiredby the seminal work of W. C. Stokoe (1960), showedthat
signedlanguages turned out to be organizedandacquired like other languages
(Newport andSupalla 2001: 107). That is, just like spoken languages, they
possess a phonological system (phonemes in sign language are the smallest
non-meaningful, visual elements), morphology andsyntax; they have anaphors
(cf. Wilbur 2001: 236) andeven null arguments (cf. S. D. Fischer 2001: 204); and
agreement in person andnumber may be signedonto the verb in connection with
both objects andsubjects (cf. Newport andSupalla 2001: 107110) etc. In other
words, the underlying abstract linguistic system is the same in both spoken and
signedlanguages, only the tools with which the signs are made, necessarily,
dier. This dierence in tools oers opportunities as well as restrictions. It is
not surprising, therefore, that linear order, tone and emphasis play an important
role in spoken languages, while space, motion andlocation come to the fore in
signedlanguages. Similarly, simultaneous activities can be quite easily expressed
in sign language while in spoken language dierent means have to be resorted
to (see also below).
It is noteworthy, however, that with the new interest in signedlanguages in
the seventies, there was a move away from their iconic nature. The reason for
this was, rst of all, that the gestural mode of communication had marked o
the language as primitive in the earlier days. Secondly, the new Chomskyan
linguistics hadelevated the arbitrariness of the sign almost to a rst principle.
Thus, iconicity hadno place within this new research paradigm. It is only very
recently that iconicity re-acquireda condent position in sign language research.
There are a number of reasons why many more signs in signedlanguages
have iconic origins andwhy the structure of signedlanguages appears to
[have] a higher degree of iconicity (S. D. Fischer 2001: 206). The rst reason is
clear and may be expressed, as Fischer has done (ibid. p. 206), succinctly:
because it can! Objects can be picturedmuch more easily andtransparently by
gestures than by sounds. It also explains why cross-linguistically, signed
languages are typologically much more alike than spoken languages. For a
similar reason it is not surprising that a higher degree of iconic types has been
foundin written language, where the spoken mode of language has been made
visible. Contributions to the rst two iconicity volumes by Matthias Bauer
(1999), Andreas Fischer (1999), Max Nnny (1999, 2001), Robbie Goh (2001)
andChristina Ljungberg (2001) have illustratediconic usages that cannot occur
in spoken varieties of language.
There may be another, equally important reason, however, why signed
languages may have preservedtheir iconicity more clearly.
S.D Fischer
(2001: 106) remarks that signedlanguages may not grammaticalize so rapidly as
spoken languages because most sign language users do not have parents who
use the same language andthat therefore sign language must be recreolizedin
every generation. It is true that, in structure, signedlanguages share features
with creoles andcreolizedlanguage, which are also more transparent (i.e. more
iconic, especially diagrammatically) than spoken languages with a longer history.
However, this does not mean that signed languages do not grammaticalize at all.
As statedabove, they do have full grammatical structure, andso they do have
signs which have become conventionalizedso that their earlier iconic origin or
motivation is no longer immediately visible. Indeed, Klima and Bellugi (1979)
have establishedthat most iconic signs are not transparent to hearing participants,
but they are often translucent, i.e. once the meaning of the sign is known, the
sign can be understood as motivated. Moreover, the degree of transparency does
not only depend on perceptual features but also on cultural features. Thus,
Pizzuto andV olterra (2001) have shown, by means of some very interesting
experiments, that signs usedin Italian sign language were more easily understood
by Italian than by non-Italian hearers, because these signs resembledculturally-
specic Italian gestures. As in all cases of iconicity, the perception of the
similarity between sign and referent depends on the interpreter and the context
that he is part of.
We welcome the attention paidto gestures and signedlanguage in this
volume because the research into gestural signs andthe structure of signed
languages can cast new light on the structure of spoken languages (in both oral
andwritten form), andmore particularly, it may tell us more about the iconic
foundations of spoken languages andthe way iconicity has evolvedin them.
There are a number of studies in this volume that are directly or indirectly
relatedto sign language or sign language research (gatheredtogether in Part II).
It must be noted, however, that the ve broad sections distinguished in this book
are meant as a loose categorization; most of the contributions cannot be limited
to the section in which they have been placed. For this reason, we will not
always follow these sections closely in this introductory chapter; rather, we will
indicate how the various topics discussed in each study refer to and are linked
with topics discussed in others.
The most direct investigation into the role played by iconicity in signed
language is provided by Klaudia Grote and Erika Linz. Their contribution pays
special attention to the fact, already remarked on above, that the interpretation of
a sign depends very heavily on its context, especially as concerns the extent to
which its occurrence is perceivedas conventional. In their study, Grote and Linz
rst remindus that only certain qualities of a referent can be iconically repre-
sentedin a sign, i.e. the similarity between a sign andits referent is neither
complete nor objective,
andthe recognition of its iconicity is always ltered by
the interpretation of the perceiver. Quite understandably, therefore, the more
conventional the immediate context of the sign is, the less likely it is that a
conventionalizedsign will be interpretediconically . The main aim of their article
is to explore the inuence of sign language iconicity on semantic conceptualiza-
tion. They test whether the iconic quality of a signedword helps to establish the
semantic concept more quickly, andwhether it inuences which quality of the
concept stands out more or is more prototypical. This, in turn, may provide
information about the semantic structuring of the lexicon. The experiments
presentedin their study involve confronting both signers and non-signers as well
as bilinguals with fully lexicalized(conventionalized ) signs (visual or oral ones)
as well as with certain qualities connectedor unconnectedwith the referent of
the sign. It turnedout that signers andbilinguals were faster in judging iconic
sign picture relations than the non-signers, who reactedto all qualities connected
with the sign in the same way. This was in fact the outcome that the researchers
hadexpectedbecause for the non-signers there was no iconic visual connection
between the sign andone of the visual qualities, their sign representing a sound
rather than a visual word. This suggests that the iconicity of the sign persists in
sign language even when the sign has become lexicalized. The experiments thus
show that the semantic organization of the mental lexicon of signers and
bilinguals is inuenced by the iconicity of the sign. These results additionally
suggest that language may inuence conceptualization, i.e. that it is not just
conceptualization that steers linguistic expression. In other words, the outcome
suggests a moderate version of linguistic relativity (p. 36).
A rather dierent experiment was conducted by William Herlofsky, who
shows how research in sign language, in this case the signing of metaphor in
Japanese sign language, may be benecial in helping to develop a comprehensive
theory of metaphor andiconicity . Working within a cognitive theory of language
where form, meaning, metaphor andiconicity are all equally relevant and fully
integratedinto the framework of the theory, Herlofsky illustrates how the
relationship between real-worldspace andmental space (which he investigated
earlier [Herlofsky 2001] from an evolutionary point of view) may be made more
visible by considering the use made of signing spaces in signed poetry. He rst
shows that the iconicity of visual language can best be approachedthrough the
analysis of metaphor. Metaphor provides the foundation for our conceptualization
of many basic abstract ideas, as the author illustrates by referring to the work of
Lako andJohnson, who show how complex metaphors arise through primary,
cross-domain associations acquired in early childhood (this is called conceptual
blending). After a brief discussion of the types of iconicity that occur in sign
language to express concrete objects (these make use of structural correspondenc-
es in both form andpath between our conception of objects in the real world
andthe form and movement of the articulators in the signing space, usedto sign
these concepts), Herlofsky moves on to the signing of abstract concepts, where it is
much more dicult to create signs that resemble their object. He shows that
here the same type of metaphorical blending takes place as is the case in regular,
i.e. non-signedlanguage. By illustrating how this blending actually occurs visibly
in the signing space in the performance of signedpoetry , Herlofsky gives us a
better idea of how metaphoric blending may take place in regular language.
At this point we will discuss a contribution by Masako Hiraga (even though
we have placedit in Part III because it is also concernedwith structural iconic-
ity) since it likewise deals with the notoriously dicult problem of metaphor
andiconicity and since it refers to Herlofskys study andd raws on the same
example, a Japanese haiku, which is also available in sign language. In Hiragas
study the vagueness of the Peircean notion of metaphor is counterbalanced by a
more precise denition derived from cognitive theory. Elaborating and rening
Turners and Fauconniers model of blending, Hiraga explains the dynamic
interplay of metaphor andiconicity from two angles: (1) iconicity manifested as
image-schema in metaphor, (2) metaphor giving an iconic interpretation to form.
The example she uses to illustrate her theoretical model is a haiku by Basho
Matsuo whose bipartite metaphorical structure is perfectly suitedto substantiate
her argument. The essay also demonstrates that in the revision of the poem, kanji
(Chinese logographs) eectively strengthen the link between form andmeaning.
The studies by Herlofsky and Hiraga, which deal with the same problem from a
linguistic point of view on the one handand a literary one on the other, exempli-
fy particularly well the advantages of the interdisciplinary approach taken in this
andthe previous volumes on iconicity.
To return to the topic of Part I, Axel Hblers contribution to this section
does not actually make use of sign language proper but of gestures used by
speakers of spoken language. He is interestedin the connection between gestures
andlinguistic signs, andhe shows how the loss of the one may leadto change
in the other, thus giving an indication of how originally iconic gestures may
emerge in a dierent form linguistically. By oering a glimpse of how such
gestures may have been translatedinto spoken language, we may acquire a
clearer idea about what links there are between spoken andsignedlanguages and
about the structure ultimately underlying both. Hbler rst develops the idea that
there is an iconic relation between the verbal and the gestural mode in the
expression of so-called redundant phrasal verbs (i.e. expressions in which the
particle does not alter the propositional content of the verb, as in swallow down
for swallow) and in the expression of pure spatial verbs (such as to up).
Thus, the accompanying gestures not only resemble but highlight a certain aspect
of the spatial meaning inherent in the verb and/or the adverbial particle. In other
words, a phrasal verb like lift up accompanied by an upward movement of the
hand highlights a part of the event itself. Similar gestures could be used in a
metaphorical spatial way in verbs such as yell out. Hbler suggests further that
this cross-modal form of iconicity is the result of the link that exists between
these verbs and gestures on the operational level; i.e. the spatial concepts,
whether expressed verbally or gesturally, are linked to the same part of the brain.
These observations are then used to explain a rather interesting historical-
linguistic development: the rise of most of these redundant and pure phrasal
verbs in the sixteenth and seventeenth centuries are linked to the great eorts
courtly society took in subjecting the body to a rigorous control (p. 74). Too
much gesturing became frowned upon. Hblers (still tentative) suggestion is that
the resulting reduction in gestures, was compensated for verbally by means of an
additional but semantically redundant particle.
Keiko Masudas investigation into birdcalls has a link with the above studies
because it likewise involves the more direct mimicking of a real-world object,
but this time by oral rather than visual means: she investigates in how far the
sounds produced by birds are actually reected in the linguistic signs that we use
to refer to those sounds. The experiments show how close the phonetic word of
a particular language, in this case English, is to the actual sound made, but the
investigation also indicates what phonetic or phonological constraints of the
language in question are in force, or indeed relaxed. It is well-known that even
in onomatopoeia, which is considered to be one of the most direct forms of
imagic iconicity (i.e. a type of icon that comes close to the real thing),
conventional phonological system of language plays a role, and it is to be
expected that the choices made to represent birdcalls are somehow constrained
by the phonological framework of the language in question.
Iconicity, then, is
a creative device that to some extent does and does not follow the conventions
of language (cf. Lecercle 1990, Fnagy 2001: 2.). Sound combinations that do
not occur in the phonemic inventory of English such as ts, pf, ps, may still crop
up in onomatopoeic expressions, as in tse tse, phft, phsst (the last two are
examples from comic books given in Crystal 1995: 250). However, the more con-
ventional the onomatopoeic words become, the more likely it is that they follow
the normal phonological rules. In other words, exclamations or interjections in
comic books may disregard these rules more easily than verbs which are next
formedfrom those exclamations, simply because verbs are more rule-governed
than interjections.
Similarly, the description of birds sounds may stretch the system of the
language more than the actual names given to birds on the basis of these sounds
because again these names as nouns fullling a regular part in the lexicon and
grammar will tendto conform to the regular pattern. Still, they remain
distinctly iconic, and therefore preserve some of the exceptional behaviour.
Marian Klamer has shown in a number of studies (e.g. 2001, 2002) that expres-
sives in a language (viz. lexicalization[s] of vividsense impressions, which
include names, and morphemes with negative connotations or referring to
undesirable states [2001: 166]) are usually distinguished formally from non-
expressives. This formal distinction could involve some direct or imagic iconic-
ity, but in establishedor conventional lexicalizations, according to Klamer, it is
the diagrammatic iconicity that has been best preserved: the formal complexity
of these expressives resembles their semantic complexity, i.e. the marked
phonetic form is iconic of the markedsemantic content.
By means of an acoustic analysis of the actual sounds of birds, Masuda
shows in her experiments which parts of the sounds are used in the linguistic
realization of it. She nds that the front cavity resonance (in which both the
secondandthirdformants in the spectogram play a role) is most crucial in the
imitation of a call, i.e. the selection of the vowels for a linguistic representation
seems determined by the frequency of the front cavity resonance. As far as
consonant selection is concerned, it appears that especially plosives are selected,
which are very well suitedto express both the abrupt onset and the extremely
short duration of most calls. There is also evidence that further acoustic factors
may inuence both the place and voicing of the plosive. In addition, Masuda
conducts some experiments the other way around, i.e. from the perceptual point
of view, by confronting human subjects with other linguistic, but deviant, signs,
which likewise imitate the birdcall. These latter experiments show conclusively
that indeed the front cavity resonance pattern is crucial for the type of vowel
used. As far as the use of deviant initial and nal consonants is concerned, the
matter was less clear. It seems that in the selection of consonants, other factors,
apart from pure perception, play a role as well. It is quite possible that the
language-specic phonological system is involvedhere. Although this was not
part of the present study, it would be interesting to nd out by an investigation
of the same birdcalls in other languages in how far the signs used there dier
from the ones foundin English, and thus in how far the phonological system of
the language itself determines the shape of the linguistic sign.
In the next section (part II), there are a number of studies that all concen-
trate on visual iconicity, a type, as we have mentionedabove, that is more
common in literary texts, andof course in signedlanguages, as we have just
seen. The study by John J. White is innovative in several respects. With its
exploration of instances of perspective in shapedpoetry on the page as a two-
dimensional surface and its semiotic eects, it opens up a new eld to
iconicity research. Albeit rootedin the Peircean tradition, it is post-Peircean in
that it utilizes theoretical approaches derived from the visual arts (E. H. Gom-
brich, Umberto Eco, Nelson Goodman), approaches which are concerned with the
role playedby perceptual conventions andcultural codes in processes of iconic
signication. The central part of his essay deals with Italian Futurists experi-
ments with typographical iconicity. Starting with minimalist Futurist examples of
shapedpoems by Bruno Sanzin, which iconize perspective in a rather simple,
albeit instructive way, he passes on to more sophisticatedexamples such as
Francesco Cangiullos free-wordcollage poem Milan-Demonstration (1915),
which is shown to evince both visual andacoustic depth. In the papers last
section, White moves from avant-garde poetry of the early twentieth century to
one of the most signicant new forms of shapedpoetry at the endof the century,
Eduardo Kacs holopoems. These poems manifest forms of signication that
have emancipated from the static renderings of perspective in earlier avant-garde
poetry using the protean eects of perspective in a holograph.
An equally complex form of visual iconicity is exploredby Julian Moyle,
namely the iconic use of the corrupt andblurredand partly illegible typing of
poems in Peter Readings Last Poems. His essay attempts to enrich the iconic
image by showing that it may express much more than just a one-to-one
relationship between wordand thing. Moyle rejects a simplistic interpretation of
Readings poem, which wouldund erstandits fragmented andillegible typograph-
ical form as iconically reecting the idea of a text under erosion. By relating the
poem to other poems in the volume the untitled nal two pages (presenting
a text even more corrupt than Erosive) and[Untitled ] (which looks like an
uncorruptedversion of Erosive) andto the wider context of Readings
work, Moyle calls in question its conception as a totalising Entropicon
(p. 142). Referring to Michel Serres complex model of entropy, which includes
negentropy, i.e. negative entropy, he proposes an interpretation of the contem-
porary British poet, which sheds light both on individual poems and his whole
vision as an artist.
A much less intricate instance of iconicity or so it seems at least at rst
sight is the topic of Andreas Ohmes analysis of Vladimir Sorokins The
Queue (1985), a generically unclassiable text, which represents the incessant
polylogue of people in an endless queue an everyday Soviet experience
waiting to buy consumer goods. While the texts dialogue is radically desemanti-
cized, its iconicity is foregrounded, the typography miming the length of the
queue, the similarity andeven identity of the replies andthe roll-calls as well as
the monotony andrepetitiveness of the whole process. A similar technique is to
be found, at least in the English translation of the text, in the use of blanks and
blank pages to iconize pauses andphases of sleep, andin the representation of
sexual encounters which is typographically reduced to moaning sounds followed
by dots. What may on the whole seem to be a rather straightforward satiric
treatment of an unpleasant part of Soviet reality is shown to be an extremely
eective subversion of the ideology and practice of Soviet Realism. The very
iconicity of the text challenges andsubverts the authority of the ideologized
doctrine of art.
The subject of Matthias Bauers essay also concerns a deceptively simple
iconic representation, i.e. Dylan Thomas poem Vision andPrayer, which, with
the rhomboidandtriangular arrangement of its lines, seems to refer explicitly to
the classical tradition of shaped poetry. But Bauer shows that there is no
straightforwardrelationship between the poems topic as expressedin the title
andits pictorial form. He makes it plausible that the poems pattern is based
on the form of the letter X, an elementary geometrical form which, as he argues,
evokes shapes profoundly signicant in mystical religious thought such as the
double pyramid, the legs of a stork, the cross, a kiss, and the gure of a naked
man with the arms andlegs spanning the globe in what is a combination of the
letters X andO. In the poems mystical geometry the X (andd erivative forms
such as triangle, diamond, double pyramid, and cross) emerges as an icon of the
poets aiming for the creative word. The iconic use of mystical geometry is here
orientedto forms or shapes without numerological implications as expounded in
Ptters essay on Dantes geometry of love in this volume (see further below).
The iconic signicance of visual elements in postmodernist ction is the
topic of Christina Ljungbergs study, which succeeds in exemplifying the
heuristic value of Peircean thought even with regardto recent experimental
literature. Ljungberg is concernedwith diagrammatic iconicity, focussing on the
interaction between visual artifacts such as photographs andmaps and the
verbal level of expression. In texts such as Ondaatjes prose poem The Collected
Works of Billy the Kid andAtwood s The Blind Assassin, photographs are used
not only to reect the structure of the narrative itself andintra-textual relation-
ships (e.g. relationships within a group or those of characters to their surround-
ings), but they are also employedas devices for self-reexive comment on
representation in general, andon the art of writing in particular. Maps are shown
to be usedby a postcolonial writer such as Merlene Nourbese Philip (Looking for
Livingstone: An Odyssey in Silence) as visual correlates of their fragmented
cultures. In self-reexive ction such as Paul Austers City of Glass, maps
iconize the problem of representation.
Moving on from visual iconicity, which is closer to the imagic type, we
now turn to Part III andto more diagrammatic forms of iconicity. These play an
important role in the way natural languages are structured. But here too the
iconic patterns range from the more concrete to the more abstract. Jac Conradies
contribution to this section concerns repetition, a very common device in
language, which is often used in a concrete iconic way. Indeed, it is likely that
all conventional repetitive patterns in language were originally iconic. Conradie
considers the device of reduplication in Afrikaans, where it is a much more
frequent phenomenon than in Dutch, which is the base language from which
Afrikaans is derived. It is very likely that the higher number of reduplicative
forms in Afrikaans is due to Malay inuence, one of the languages that play a
role in the creolization of Afrikaans. It is well-known that reduplication is a
common feature in Malay (for instance the plural is formedby repeating the
stem), as it is in many languages derived from a pidgin (cf. Tabakowska, this
volume, andKouwenber g andLaCharit 2001).
Reduplication is a versatile and multifarious device in that it can be used in
many functions. The same is true for repetition as a literary device. In literature,
it is commonly usedto express similarity, continuity, regularity, monotony,
emphasis (Nnny 1986: 205). Because the reduplication in Afrikaans is fairly
it is still more clearly iconic, i.e. there has been little grammaticalization.
Not surprisingly therefore, the functional categories distinguished for Afrikaans
by Conradie, closely resemble the categories found to be relevant for literary
language, where the repetition is not usually conventional. The most concrete
function of reduplication in Afrikaans is indeed the suggestion of repeated
action. The mechanism is especially encounteredin the names of games, which,
as Conradie states, may be regarded as essentially repetitive (p. 207). It is also
usedto suggest repeatede orts, which comes close to another concrete function
of reduplication, i.e. intensication (similar to Nnnys emphasis). Another
function Conradie illustrates, again close to the pure notion of repetitiveness, is
its use to suggest intermittent activities, which in turn is relatedto the idea of
interruption or discontinuity. Discontinuity, of course, can only exist in a
situation in which continuity is foregrounded, so again we have a link here
with one of Nnnys functions of repetition in literary language, anda link with
the notions of continuation or extension, which Conradie has also established
for Afrikaans. What makes the categorization of the functions of reduplication in
Afrikaans so convincing is the fact that all functions clearly hang together, as
well as the fact that they are usedcreatively in precisely the same way as in
literary language. As with all types of iconicity, the key to repetition is similari-
ty. In reduplication, the repeated form is literally similar to the rst form, which
probably accounts for its simple but strong iconic quality, andfor its natural
presence in both everyday language and literary language.
The other contributions placedin this section show structural iconicity of a
more abstract type. When the iconic relation is more abstract, it is more dicult
to observe it directly. At the same time, however, it is also more frequently
present in language. It is a well-known fact that iconicity increases with an
increase in structure. Thus, compounds and derivations are more iconic than
simple stems, andsyntactic structures are more iconic than words. Volker Harm
shows that there may be another iconic tendency in the lexicon which not so
much concerns a diagrammatic relationship between morphemes within words
andthe way they share morphemes with other words, but which concerns an
increase in isomorphism between the meaning andform of a whole (simple or
complex) word. He investigates a development in the history of German whereby
in a class of verbs derivationally related to one another (e.g. hren, erhren,
verhren, gehren), which were once all more or less synonymous (andtherefore
non-isomorphic between their form andtheir meaning), the one-to-many relations
between form andmeaning change slowly into a one-to-one (i.e. isomorphic)
relation, so that each individual verbal form ends up with only one of the
meanings that they previously shared. What makes this isomorphic tendency even
more interesting from an iconic point of view, is the fact that the central, most
prototypical meaning gets attachedto the most central sign or form, i.e. the
unprexed stem, while the more peripheral meanings fall to the prexed forms.
Thus we see an (iconic) equivalence between a morphologically markedform
anda semantically markedmeaning. This is rather similar to the type of iconicity
that Klamer (2002) discovered in two unrelated languages between phonological-
ly markedforms (of the phonaesthetic type, such as immer, icker) andtheir
semantically markedmeanings.
The study by Beate Hampe andDoris Schnefeldd eals with structures
larger than the word. What Hampe andSchnefeldinvestigate is verb phrases
which are combinedwith arguments that they do not normally subcategorize for.
These new argument frames, however, can be understood by virtue of similar
frames usedwith other, more general verbs. The reason why they can be
understood is that the verbs now sharing the same pattern begin to share
a number of conceptual properties. This type of iconicity works a little bit like
metaphor in that the verbal concept, normally expressedby these rather general
andfrequently occurring verbs, is now expressedmore concretely andvivid ly by
another verb using the same subcategorization frame as the oldverb, a frame that
it did not possess before. In all cases the new verbal construction highlights a
number of expressive qualities that the general verb did not possess. Thus, the
general verb put is associatedwith the following argument frame:
put NP
into NP
as in, She put the child into his high chair. Next, this same frame comes to be
usedwith the verb wrestle, even though wrestle normally occurs only in the
following frame:
wrestle with NP
When a language user produces an utterance such as, She wrestled a screaming
Dudley into his high chair, he is using the frame creatively, xing the pattern of
put onto that of wrestle, thereby not only changing the subcategorization frame
of wrestle, but also giving it a new meaning, i.e. to put somebody/something
into a place with great diculty. This new construction can only be understood
by reference to the old one. In other words, it is the diagrammatic iconic link
between the oldand the new, that gives the new construction its meaning. It
works like metaphor, but the basis of comparison is structural andnot conceptu-
al. What is transferredfrom tenor to vehicle is what Hampe and Schnefeld
call a schematic icon (p. 244), andnot a particular conceptual quality. In both
cases, however, we have the same result, i.e. the expression of the familiar by
the unfamiliar, making the utterance new andfresh. In both cases, too, it is the
context that helps to make the correct interpretation. So this study is another
illustration of the fact that the recognition of iconicity very much depends on
clues provided by the context, as was also emphasizedby Grote andLinz.
Language users may creatively manipulate the argument or complementation
patterns in which verbs appear provided that enough clues are present to interpret
the manipulatedcomplementation structure correctly. It is important to note too
that the iconic diagram in this study is dierent from metaphor in that it is
motivatedintra-linguistically , it is a form of endophoric iconicity (cf. Nth
2001). Metaphor is motivatedby qualities of the external world(exophoric) and
is therefore more likely to occur cross-linguistically (more universal). In other
words, the use of the same metaphor in dierent languages depends on cultural
phenomena, whereas the use of the same iconic schemes or diagrams depends on
the grammatical system of a particular language.
Another study that is interestedin the kindof (diagrammatic) iconic
principles that play a role in grammatical structuring is the contribution by
Gnter Rohdenburg. Rohdenburg especially addresses the question of what
determines grammatical form in cases of variation. One of the most salient
aspects of iconicity in this respect is isochrony, i.e. the phenomenon that the
order of linguistic elements referring to events in the real world mimics the real-
world order (cf. Tai 1985). This indeed constitutes Rohdenburgs rst principle
of linear order. The other two major principles which he distinguishes are the
quantity principle and the distance principle (the distance principle is similar
to Givns [1985] proximity principle andalso incorporates Bybees [1985]
relevance principle).
In his essay for this volume Rohdenburg is concerned
only with the latter two, but it is interesting to note in connection with the rst
principle that, although the temporal order of events is indeed frequently
mimickedin language, it is dicult, if not impossible to represent the simultane-
ity of events iconically in this way (cf. Haiman 1985). We have briey noted
above that this is not the case in signedlanguages where the iconic expression
of simultaneity plays a most important role. Clearly the possibilities for iconic
forms are constrainedby the dierent modes of communication (gestural/visual
vs. oral/aural). Nevertheless it is clear that in spite of the problem of representing
simultaneity iconically in terms of order, spoken languages still make use of
iconic means to convey the occurrence of simultaneous events. Jansen andLentz
(2001), for instance, have shown that simultaneity can be iconically suggestedby
intertwining the two simultaneous events, or by embedding one in the other. In
this way the linear distance between the sub-event and the main event is reduced
(andso the proximity principle is calledupon). Another possibility is using the
principle of quantity. Discussing two simultaneous events, Jansen andLentz
show that the one that is of minor importance, is given minor linguistic form
minor in the sense that explicit verbal andnominal markers are lacking
while semantically too the form is less specic than the form which represents
the major event.
The principle of quantity, as Rohdenburg briey notes, is also at work in
repetition and reduplication, as is indeed shown in detail in Conradies contribu-
tion to this volume. In his own investigation, Rohdenburg shows how both the
principle of quantity and of distance play a role in the determination of a number
of grammatical variants in Present-day English. He looks at the role of quantity
in the expression of verbs which may function as auxiliaries as well as full verbs,
and nds that only auxiliaries can be shortenedandphonetically reduced.
Similarly contractedforms with not occur only with auxiliaries, andd ialect
evidence shows that uninected forms of do or be (so forms with less quantity)
are far more frequent in the case of auxiliaries.
The quantity principle also plays a role in areas of greater syntactic
complexity. For instance, in cases of variation, such as between the comparative
form with -er andthe periphrastic form with more, the longer form is often used
in syntactically more complex constructions. In such cases, extra quantity is
used, as it were, to indicate the markedness of the total structure. In this same
case of variation, the distance principle is also at work in that in adjective-noun
phrases where adjective and noun belong together conceptually, the use of more
is foundto be more frequent. Using -er in such cases wouldcreate distance
between adjective and noun; so we nd more high-minded rather than higher-
minded. Other examples where the choice of inexion is inuenced by the
distance principle concerns the variation in voicing present in the plural forma-
tion of words like hoof (hoofs vs. hooves) and in the adjectival derivation of
phrases like loud-mouthed, where the choice is between [t] and[d ]. In each
case reduced quantity is iconic of reduced referential meaning.
A literary example of structural iconicity which assumes the form of
mathematical or geometrical analogies with profoundpoetological implications
is treatedby Wilhelm Ptters in a large-scale exploration of the intertextual
relation between Dantes two main poetical works, Vita Nova and Divina
Commedia. The connection between these texts is shown to consist in a specic
form of iconicity which relates dierent levels of textual organization by
mathematical strategies andgeometrical designs. Central to the texts mathemati-
cal designs is Beatrice, whose name is interpreted as a motivated sign. The essay
demonstrates an extraordinary knowledge and use of mathematics in Dantes
construction of his poetic universe. Beatrice is given a numerical identity, and
the numbers relating to her inform the chronological structure of the romance in
Vita Nova andthe geometrical conception of the Commedia. The mathematical
design is revealed to be a key to the hidden meaning of the two works. The
study is amply provided with gures, clarifying the intricate design. It closes
with a scheme which summarizes the whole mathematical conception of Dantes
poetic cosmos, whose spiritual corner-stones are love of Beatrice, love of God,
andlove of philosophy. Ptters elucidation of the mystical and poetological
signicance of numerological andgeometrical correspondences in Dante can be
relatedto Matthias Bauers contribution in this volume which examines the
mystical use of geometry in Dylan Thomas poem Vision andPrayer.
Hiragas paper on metaphor andiconicity , which we discussed above, is
placedin this section because her approach to metaphor focuses on the interplay
of structure andmeaning in the form of an analogical mapping andbecause her
analysis of a Japanese haiku examines all the texts formal resources (lexicon,
syntax, orthography, form of letters, soundpatterns etc.), which together enrich
the poems meaning.
An entirely new departure in iconicity research is to be found in the studies
contained in part IV which deal with the interdependence and interaction between
dierent art media on the level of form, a phenomenon called intermedial
iconicity. This type of iconicity emerges (1) when a work of art is transferred
from one medium to another one and in this process retains formal features
inherent in the source medium, and (2) when a work of art adopts or imitates
formal features characteristic of another medium. The latter type of intermedial
iconicity is systematically exploredin Werner Wolfs contribution, which is
devoted to the phenomenon that form in literature can mime other arts and
media. It discusses three types of intermedial iconicity in which literary form
imitates other media, without ever actually incorporating these media in the text:
(1) pictorialization of ction, illustratedby Thomas Hardys Under the Green-
wood Tree, a novel which is subtitled A Rural Painting in the Dutch School, (2)
lmicization of ction, exemplied by the last chapter of DavidLod ges
Changing Places, and(3) musicalization of ction, representedby Nancy
Hustons The Goldberg Variations. This paper, which opens new perspectives for
further research, also addresses the problem of reception, i.e. the question of how
the reader may be induced to identify intermedial references. It is, however, not
as radically reader-response oriented as Johansens contribution to this volume,
which is focussedon the construction of iconicity in the readers mindd uring the
process of reception.
Although Elzbieta Tabakowskas study seems to have a narrower scope than
Wolfs, it examines a highly fruitful topic, namely the problem of how to
preserve or recreate iconicity in translation. This also involves interaction
between one linguistic medium and another. In her theoretical approach Taba-
kowska proceeds from the axiom that similarity is basic to iconicity, but that in
natural language use, in the processes of lexicalization andgrammaticalization,
the earlier transparency may disappear, causing the iconic features to change into
conventional or symbolic ones. It is interesting to explore the boundaries between
these expressive andconventional stages, and Tabakowska shows that a know-
ledge of iconic practices which are conventional in one language but may still be
expressive in another is important in translation. Only if iconic devices are
indeed expressively used, i.e. used in order to achieve a particular communica-
tive purpose (p. 366), shouldthey also be translated in such a way that the
special purpose comes out in the other language. Obviously, one cannot always
use the same devices in the other language because the same device may be part
of that other languages conventional system. In other words, the choice made by
the translator needs to be new and expressive for that language in order to be
eective. Tabakowska shows how this can be done in practice by comparing a
Polish poem by Wisawa Szymborska andan English translation of it, and by
showing how even in a more conventional prose text, ad hoc forms of iconicity
may be spotted, which therefore deserve to be preserved in the translation.
Especially in the latter case, Tabakowska makes clear how dicult it is to make
a decision about whether the iconicity is intentional or incidental. Although it is
true that the intentionality of it could be checkedagainst the backgroundof the
use of language in the rest of the book, or in the language in question as a whole
(as for instance Shapiro [1998] has done concerning the use of particular sounds
in the sonnets of Shakespeare), ultimately the decision of whether something is
intentional or not depends on the knowledge of the translator of the two langua-
ges involved, as well as on his/her imagination, intuition and acumen.
The volume closes with two theoretically-orientedcontributions (Part V),
one looking at iconicity in terms of reader-response theory, and one relating it to
systems theory. Jrgen Dines Johansens contribution is rather exceptional in this
volume andin iconicity research in general in that its approach is entirely
reader-oriented. (To some extent reader response is also considered in the essays
of White andW olf.) It shifts attention from iconic signs as constituents of the
text to the readers iconization of the text in the process of reception. The study
in fact views the literary text as a set of instructions for dierent ways of iconi-
zation. On the one handbeing innovative, post-Peircean in fact, in the application
of reader-response aesthetics to iconicity, Johansens essay is on the other hand
rmly grounded in Peircean semiotics. In accordance with Peirces categories of
image, diagram, and metaphor, it deals with imaginative, diagrammatic, and
allegorical iconization occurring in the reading process. Each of these categories
is amply illustratedby examples from literary texts. The Peircean axiom of the
interdependence of the iconic, indexical, and symbolic dimensions of signs,
which has been neglectedin recent research, informs the whole discussion. (This
trendalso emerges in the linguistic contribution by Grote andLinz in the present
volume). Johansen notes that even though there is a constant interplay of the
three modes of iconization in literature, texts may receive their iconic prole by
the predominance of one mode.
Piotr Sadowskis study intends to provide a classication of linguistic signs
that diverges from Peirces, using the framework of the so-called systems theory
of information. As in Peirce and in other semiotic theories, a distinction is made
between information as purely physical facts, andpara-information, the inter-
pretation or processing of these facts by animate beings, which turns the purely
physical signals into (meaningful) signs. An important part of the theory is
that systems interact with one another by exchanging information andener gy,
andin this way the systems undergo change. In the course of this interaction,
language as a system of communication continually evolves. Sadowski considers
the dierent signs that exist in language in terms of information: emotive,
indexical, iconic and arbitrary signs. It is clear that emotive signs represent the
most simple type of para-information in that the interpretation that makes them
emotive signs does not so much take place by means of association, but by
means of instinct, and this is almost direct, subject to physiological conditions
only, andnot to cultural ones. When such emotive signs are displacedin time or
space, they become indexical, and involve para-information of a somewhat
higher order. Here the associations performed are not instinctive but acquired
behaviour. The next step in the order of signs, are signs which are no longer
physically co-present with their signals, but which resemble them only in their
structure: iconic signs. The imitations of sounds are probably the earliest iconic
signs, as well as the use of gestures. The association here works by means of
analogy andis acquiredthrough observation and experience. In all these cases,
then, the association retains a physical connection between signal andsign, but
one that becomes less and less direct. In their development towards arbitrariness,
signs have become so conventionalizedthat this direct link is lost. Such signs, in
other words, have to be acquired purely by learning. From an evolutionary point
of view, this development was highly eective in that it enabledspeed of
communication andcultural group cohesion; it liberatedhumans from the
constraints laid down by nature. Sadowski adds one further sign in this develop-
ment, a sign which is no longer promptedby the perception of the signal (as the
others are), but which derives from the para-informational level of the sign; he
calls this meta-informationally derived sign, a symbol (i.e. a symbol in the
usual poetic sense, not in the Peircean sense). The meta-informational level
serves to distinguish between literal and metaphorical meaning, between
denotation andconnotation, between truth and ction. Meta-information is
a specically human development, and a later stage in the evolution of language.
Sadowski applies the notion of meta-information especially to literary language,
but the distinction would be equally useful for an understanding of the develop-
ment that takes place in grammaticalization processes (the development of
grammatical systems), which according to Sweetser (1990), proceeds from the
socio-physical domain through the epistemic domain into the speech-act domain,
i.e. from para-informational to meta-informational.
While the last two contributions are essentially theoretical in their approach,
the analytical studies containedin this thirdvolume are also often basedon new
theoretical departures. This is particularly evident in the investigations into
gestural iconicity, sign language, sign poetry, intermedial iconicity and the
iconicity of metaphor. We believe that this new collection will add a further
stimulus to research into the iconic nature of language as usedin spoken and
signedlanguages andin literature.
1. For an elaborate description of the various types of iconicity that can be distinguished, see the
introduction to the rst volume in this series, edited by Nnny and Fischer (1999).
2. Its position is not too assuredyet, witness the chapter on sign languages in the recent Handbook
of Linguistics (Sandler and Lillo-Martin 2001), in which the helpful eect of iconicity on the
process of language acquisition is calledinto doubt, andin which many characteristics of signed
languages that couldquite easily be relatedto their stronger iconic character, are not linkedto
iconicity as yet; e.g. the type of verb agreement discussed on p. 544, the morphology of verbs
of motion andlocation discussedon p. 545, the noun classiers on p. 546, etc.
3. Starting from the reasonable assumption that at the very beginning all language, both spoken
andsigned , was more strongly iconic, and that languages have lost many iconic features
through conventionalization in the course of their evolution, cf. Bolinger andSears 1981: 129,
Fnagy 2001: 668 .
4. See also the reference to Nth in note 5.
5. Nth (2001: 19), referring to Peirce, shows, that no icon is ever purely iconic, because a
hundred percent fully iconic sign can only be the object itself. So all icons, however direct, are
only hypoicons. On the scale of iconicity however, from hypoicon to pure icon, onomatopoeia
is close to the pure one.
6. Newman (2001: 251) indeed writes: Ideophones are somewhat dierent from prosaic words,
but they are not outre-systme, i.e. they usually stretch the system of some language a bit, but
they do not totally disregard it.
7. That it is recent, is also clear from its form. Most reduplicated forms discussed by Conradie
concern full repetition rather than the repetition of just one syllable or part of the stem, which
is the more common form once reduplication has been grammaticalized, cf. the examples in
Kouwenberg andLaCharit (2001).
8. This principle states that elements that belong together semantically tendto be placed together
syntactically. The proximity principle usually applies on the basis of content or meaning and
then considers form, as is the case with most types of iconicity, but Fischer (1994) has shown
that it may also work the other way around, from form to content, i.e. giving a new semantic
content to a phrase, whose parts were placedtogether syntactically for non-semantic reasons
(e.g. because of syntactic changes elsewhere in the grammar).
Battison, R. M. 2000. American Sign Language linguistics 19701980: Memoir of a
renaissance. In K. Emmorey andH. Lane (eds), 516.
Bolinger, D. andD. A. Sears. 1981 (3rded .). Aspects of Language. New York: Harcourt,
Brace, Jovanovich.
Bybee, J. L. 1985. Morphology. A Study of the Relation between Meaning and Form.
Amsterdam: Benjamins.
Crystal, D. 1995. The Cambridge Encyclopedia of the English Language. Cambridge:
Cambridge University Press.
Davis, H. andS. R. Silverman. 1970. Hearing and Deafness. New York: Holt, Rinehart
andW inston.
Eisenson, J. andP . Boase. 1950. Basic Speech. New York: MacMillan.
Emmorey, K. andH. Lane (eds) 2000. The Signs of Language Revisited. An Anthology to
Honor Ursula Bellugi and Edward Klima. Mahwah, NJ: Lawrence Erlbaum Associates.
Fischer, A. 1999. Graphological iconicity in print advertising. In M. Nnny and O.
Fischer (eds), 251283.
Fischer, O. 1994. The development of quasi-auxiliaries in English and changes in word
order. Neophilologus 78: 137- 164.
Fischer, O andM. Nnny (eds). 2001. The Motivated Sign. Iconicity in Language and
Literature 2. Amsterdam: Benjamins.
Fischer, S. D. 2000. More than just handwaving: The mutual contributions of sign
language andlinguistics. In K. Emmorey andH. Lane (eds), 195213.
Fnagy, I. 2001. Languages within Language. An Evolutive Approach. Amsterdam:
Givn, T. 1985. Iconicity, isomorphism, andnon-arbitrary coding in syntax. In Iconicity
in Syntax, J. Haiman (ed.), 187220. Amsterdam: Benjamins.
Goh, R. 2001. Iconicity in advertising signs. Motive and method in miming the Body.
In O. Fischer andM. Nnny (eds), 189210.
Haiman, J. 1983. Iconic andeconomic motivation, Language 59: 781819.
Herlofsky, W. 2001. Icons, anaphors, andthe evolution of language. In O. Fischer and
M. Nnny (eds), 5566.
Jansen, F. andL. Lentz. 2001. Present participles as iconic expressions. In O. Fischer
andM. Nnny (eds), 277288.
Klamer, M. 2001. Expressives andiconicity in the lexicon. In Ideophones, F. K. E.
Voeltz and C. Kilian-Hatz (eds), 165181. Amsterdam: Benjamins.
Klamer, M. 2002. Semantically motivatedlexical patterns: A study of Dutch and
Kambera expressives, Language 78: 258286.
Klima, E. andU. Bellugi. 1979. The Signs of Language. Cambridge Mass.: Harvard
University Press.
Kouwenberg, S. andC. LaCharit. 2001. The iconic interpretations of reduplication:
Issues in the study of reduplication in Caribbean Creole languages. In Iconicity, O.
Fischer andM. Nnny (eds), Special Number of the European Journal of English
Studies (EJES) 5: 5980.
Lecercle, J. J. 1990. The Violence of Language. London: Routledge.
Ljungberg, C. 2001. Iconic dimensions in Margaret Atwoods poetry and prose. In O.
Fischer andM. Nnny (eds), 351366.
Nnny, M. 1986. Iconicity in literature. Word & Image 2, 3: 199208.
Nnny, M. 1999. Alphabetic letters as icons in literary texts. In M. Nnny andO.
Fischer (eds), 173198.
Nnny, M. 2001. Iconic functions of long andshort lines. In O. Fischer andM. Nnny
(eds), 157188.
Nnny, M. andO. Fischer (eds). 1999. Form Miming Meaning: Iconicity in Language and
Literature. Amsterdam: Benjamins.
Newman, P. 2001. Are ideophones really as weird and extra-systematic as linguists make
them to be?. In F. K. H. Voeltz andC. Kilian-Hatz (eds), pp. 251258.
Newport, E. L. andT . Supalla. 2000. Sign language research at the millennium. In K.
Emmorey andH. Lane (eds), 103114.
Nth, W. 2001. Semiotic foundations of iconicity in language and literature. In O.
Fischer andM. Nnny (eds), 1728.
Pizzuto, E. andV. Volterra. 2000. Iconicity andtransparency in sign languages: A cross-
linguistic cross-cultural view. In K. Emmorey andH. Lane (eds), 261286.
Sandler, W. and D. Lillo-Martin. 2001. Natural sign languages. In The Handbook of
Linguistics, M. Arono and J. Rees-Miller (eds), pp. 533561. London: Blackwell.
Shapiro, M. 1998. Sounds and meaning in Shakespeares sonnets. In M. Shapiro, The
Sense of Form in Literature and Language. New York: St Martins Press, 6696.
Stokoe, W. C. 1960. Sign Language Structure: An Outline of the Visual Communication
Systems of the American Deaf. Bualo, NY: University of Bualo.
Wilbur, R. B. 2000. Phonological andprosod ic layering of nonmanuals in American Sign
Language. In K. Emmorey andH. Lane (eds), 215244.
</TARGET "intro">
TITLE "Auditory andvisual signs andsigning"
WIDTH "150"
P:1 I
Auditory and visual signs and signing
</TARGET "p1">
AUTHOR "Klaudia Grote and Erika Linz"
TITLE "The inuence of sign language iconicity on semantic conceptualization"
WIDTH "150"
The inuence of sign language iconicity
on semantic conceptualization
Klaudia Grote Erika Linz
University of Cologne University of Cologne
1. Introduction
The medium of visual-gestural communication which sign languages employ has
characteristics which are very dierent from the auditive-vocal modality of
speech. Whereas spoken words are formedfrom vowel andconsonant sounds,
signedlanguages are articulatedwith the hands, upper body andface. Simulta-
neous combinations of specic congurations of the handwith non-manual
expressions of the body andface are usedto create signs.
These dierences in
modality have an impact on the articulation process, i.e. spoken words are
articulatedone after another in a linear sequence
andstay always in a temporal
relation to each other, whereas signs are not only expressedin a sequential order
(as in spoken languages) but simultaneously in space as well. Space is directly
usedto linguistically express properties of a referent like shape, location, motion,
manner, direction, features or qualities by movement of the hands through syn-
tactic space (see e.g. Schick 1990; Collins-Ahlgren 1990; Liddell 1990; Supalla
1986, 1990; Engberg-Pedersen 1993).
Sign languages oer a very interesting
eld for the study of iconicity because the visually basedlinguistic system shows
a much greater disposition to iconic signs than the auditory system (cf. Arm-
strong et al. 1995: 192). Due to the predominance of visual over auditory per-
ception in the interaction with external objects, there are many more possibilities
to depict visual similarities than there are to produce acoustic ones in the process
of sign-creation. It is easier to create a visual correspondence between an
external referent andlinguistic properties of visual-gestural signs than an acoustic
correspondence between a referent and vocal signs. As a direct consequence, all
sign languages possess many more signs of high imagic-iconicity (cf. Fischer and
Nnny 1999: xxif.) in comparison to the limitednumber of onomatopoeic words
foundin spoken languages. But in the history of sign language research the high
degree of iconic signs has scarcely been attributed to the visual-gestural modali-
ty. In fact, it has long been denied that sign languages are fully natural languag-
es. Before Stokoe (1960) discredited this attitude, it was widely believed that
sign languages were a kindof pantomime, purely iconic with no formal linguistic
structure. According to the classical hierarchical sign-typology, iconic signs were
consideredto be primitive andnot able to fulll the linguistic functions of
arbitrary words. In the 1980s new research began to appear on the topic of
iconicity in spoken languages. It is now widely accepted that iconicity works on
all levels of language andtherefore cannot serve as counter-evidence against the
language status of sign languages.
Nonetheless the relationship andthe interplay between the apparent imagic
iconicity of many signs in all sign languages andtheir status as linguistic signs
remains for the most part unexplained. Particularly, modern semiotic theory has
not been fully taken into account in sign language research, although it is
possible to nally overcome the long-lasting prejudice of sign languages as
primitive communication systems. Thus, before presenting two empirical studies
which explore the inuence of sign language iconicity on semantic conceptual-
ization, two theoretical aspects concerning the relation between iconic and
symbolic characteristics of linguistic signs will be discussed on the basis of
Peircean andSaussurean semiotics.
2. Iconization and symbolization of signs
The rst point to be emphasized regards the referential underspecication of
iconic signs. Peirce denes icons as likenesses, i.e. An icon is a sign t to be
usedas such because it possesses the quality signied (EP 2: 307; see also
p. 5f., CP 2.276). In contrast, symbols are dened as conventional signs (CP
2.295; EP 2: 9). According to Peirce, an icon is extremely indeterminate in its
signicance (Ransdell 1986: 62) because it signies qualities only. Since the
relation of likeness can only refer to single qualities, never to distinct concepts,
it is impossible to identify the object which an icon stands for without additional
information. This leads Peirce in his later writings to the conclusion that an icon
can only be a fragment of a completer sign (EP 2: 306). Icons therefore can
only be interpretedprecisely when supplementedby indexical and symbolic
specications. The Peircean view correlates with the thesis, as expressedby
Fischer andNnny , that iconicity is not self-explanatory but that the perception
of iconic features in language and literature depends on an interpreter who is
capable of connecting meaning with its formal expression.
With respect to sign
languages, the thesis has been conrmed by various international studies. It has
been shown that non-signers interpret at most ve to ten percent of even highly
iconic signs adequately (cf. Bellugi andFischer 1972; Bellugi andKlima 1976;
Klima andBellugi 1979; Hoemann 1975). Moreover Pizzuto andVolterra (2000)
found that the ability to guess the meanings of signs depends on (1) the compe-
tence in identifying iconic-transparent features of signs, (2) experience with a
sign language system, and(3) cultural-specic factors. In all cases the perceived
iconicity is not an objective likeness between a referent anda linguistic form but
a mentally constructed correspondence between two cognitive products.
The second point to discuss concerns the widespread misunderstanding of
iconicity andarbitrariness as contradictory characteristics of signs. On the one
hand, this misleading view results from a restricted understanding of the
Saussurean notion of arbitrariness. Larbitraire dusigne does not refer to the
common idea that the signiant is independent of the signi andthat the
relationship between sign form andsign meaning is a conventionalized one.
Rather, both the signier andthe signied are arbitrary in the sense that they do
not exist independently from the sign creation process (SECI: 232f., 253255f.;
SCIII/KH: 138142f.; see also Fehrmann andLinz 2002; Jger 1978; 1986).
Saussure himself applies the concept of arbitrariness not to the signier alone but
also to the sign as an inseparable whole. As Engler (1995: 40) explains, Saussure
refers with this notion to the arbitrariness of sign nexus which implies a reciprocal
determination of the signier and the signied, both of which were indistinct until
the relationship was set up. Thus though in contrast to the common equation
of arbitrary andunmotivatedsigns (cf. Engler 1995: 44) the Saussurean notion
in its original version does not contradict motivated sign formation.
On the other hand, the Peircean subdivision of signs into icons, indices and
symbols is, up until now, often misinterpretedas an exclusive classication, so
that a sign is thought to be either iconic or symbolic. Mainly in his later work
Peirce himself emphasizes that symbols may be in part iconic andin part
indexical (cf. EP 2: 10, 274f., 320; CP 2.302, 5.119; see also Elgin 1996; Nth
2001). Symbol and icon do not designate mutually exclusive classes and likewise
iconicity and symbolism do not refer to disjunctive properties of a sign, but
rather are functionally guided and context-dependent characterizations of signs.
As Ransdell (1986: 57) points out:
Thus when we identify some sign as being iconic, for example, this only
means that the iconicity of that sign happens to be of peculiar importance to
us for some reason or other implicit in the situation andpurpose of that
analysis, but there is no implication to the eect that it is therefore non-
symbolic or non-indexical.
Appliedto the example of sign language it follows that the potential iconic
dimension of signs does not interfere with their linguistic function. An iconic
sign can work both as an imagic-icon anda linguistic sign. Whether it acts
primarily as an icon or as a symbol is determined by its actual use. The onset of
either one of these sign transcription processes (cf. Jger 2002) is to some
extent dependent on the linguistic context. We call the process in which the
focus of attention shifts from the symbolic meaning of the linguistic sign to the
iconic meaning of the linguistic properties of the sign iconization of a symbol.
The reverse process, when the iconicity of the sign is blankedout andthe
recipient sees through the material form of the sign directly onto the semantic
meaning (cf. Frishberg 1975; Klima andBellugi 1979), we call symbolization
of an icon.
We claim that in discourse the interpretation of linguistic signs
even those which are highly iconic is predominately guided by the systematic
structures, i.e. by the paradigmatic and syntagmatic relations of a language
system. This symbolization process of icons has already been described by Peirce
with respect to vocal language: If the sounds were originally in part iconic, in
part indexical, those characters have long lost their importance. The words only
standfor the objects they do, andsignify the qualities they do, because they will
determine, in the mind of the auditor, corresponding signs (CP 2.92). Thus,
iconic signs behave mainly as symbols after being integratedinto the language
system. Nevertheless the linguistically governedsymbolization of an iconic sign
does not generally exclude its renewed use as an icon. In fact, the iconic
dimension of a sign can be regeneratedwhen the sign is unhingedfrom the
general linguistic system. In this case the semantic aspects recede into the
backgroundandthe focus of attention draws mainly on the material form of the
sign. Such a re-iconization process of linguistic signs can often be foundin
aesthetical or advertising contexts as well as in disturbed communication.
However, in the reverse process of linguistically driven symbolization of iconic
signs the question of whether the iconic aspect of the sign still retains an eect
on conceptual structures remains unsettled. More precisely, is the inuence of
imagic iconicity limitedto a motivational aspect in the process of sign creation
or does the iconic dimension maintain an impact on the semantics of the sign
after being lexicalized? If we holdwith Pelc (1986: 12) that iconicity andthe
symbolicity are degree-properties of signs, i.e. relative and not classicational
in character, the question is raisedwhether the iconic dimension becomes
completely inert in cases of regular linguistic use or whether both semiotic
functions, the iconic andthe symbolic, can operate to a certain extent simulta-
neously in the act of language use.
3. Empirical investigation
Based on these theoretical considerations, two experiments were developed to
explore the potential cognitive eect of sign language iconicity on semantic
conceptualization processes. In particular, the empirical studies presented here are
aimedat investigating whether the iconicity of a sign aects the structure of the
semantic network of deaf and hearing signers in the way that certain characteris-
tics, which are highlightedby the iconic aspect of a sign, play a central role in
the corresponding semantic concept.
In other words, when a sign in German
Sign Language (DGS) like house provides an image of a prototypical house in
the way that the articulators (the hands) sign the trace of the shape of a house
with a pointedroof, does the roof then play a special role in the semantic
concept of the house?
Study 1: Verication task with signs/words and pictures
In the rst study we examined whether deaf and hearing signers of German Sign
Language (DGS)
andhearing speakers of German Spoken Language (DLS)
showedd ierent response times in a verication task where they hadto judge
the presence or absence of a semantic relation between a sign/wordand a picture.
The experiment was basedon the presumption, that the characteristics or aspects
of a referent which are accentuatedby the iconic feature of the sign play a
central role in the corresponding semantic network. Thus, the hypothesis tested
in the verication task was that the Response Times (RTs) of the deaf and
bilingual participants for the pictures which correspondto the iconic features of
the signs are shorter than the response times for the pictures which do not
correspondto the iconic features. Since the translatedequivalents of the signs in
spoken language were all non-iconic, the response times of the hearing partici-
pants were expectednot to dier for both types of pictures.
For deaf and bilingual subjects the sign-video and the picture were displayed one
at a time on a computer monitor (Figure 1). Timing andrecord ing responses
were under program control. In order to prevent visual masking, the sign-video
was centeredat the top of the monitor, the picture at the bottom.
A sign was considered to begin when the hand(s) entered signing space and
Figure 1.Test-procedure sign-picture
to endwhen the hand(s) began to move out of the sign conguration andback
down to resting position.
The words were presented via two loudspeakers which were attached to the
Figure 2.Test-procedure word-picture
computer (Figure 2). In addition to the acoustic presentation a visual marker
(icon of a loudspeaker) appeared on the screen at the same position where the video
was presented for the deaf subjects. At the bottomof the screen the picture appeared.
The deaf, hearing and bilingual subjects were seated directly in front of the
monitor at a distance of about 60 cm with the two index ngers resting on two
response buttons. Response times were measuredfrom the onset of the target
item. The subjects were informedthat they would see several pairs of items. The
inter-stimulus interval between the rst andthe second item was 1000 ms. The
signs andspoken words were grouped in three series of 40 + 40 + 40 trials (120
trials), with 2 rest periods. Each pair of items (sign-picture vs. word-picture) was
followed by a blank screen for 2 seconds before the next pair appeared. The
right-handedsubject was instructedto press the right button if there was a
semantic relation between the two presenteditems andthe left button if there
was none. Left-handed subjects used the response buttons the other way around.
Equal numbers of relatedandnon-related items were assignedto the test list
randomly. The subject was toldto perform the task as accurately andquickly as
possible. The subject couldrespondanytime from the start of the secondtest
item, but fast responses did not alter the timing of stimulus presentation. Each
subject was instructedby a standardizedexplanation (signed andwritten vs.
spoken andwritten) andwas given 10 practice trials.
The experimental material consistedof 120 pairs composedof 10 repeatedly
presentedsigns/word s and60 semantically relatedand60 unrelated pictures. The
sign/wordrepresenteda specic semantic concept, for example Adler (eagle)
(Figure 3). In this stimulus set, the semantically relateditems were Schnabel
(beak), Flgel (wing) and Kralle (claw). Not semantically relatedwere
Bagger (digger), Kette (necklace) and Koer (suitcase). One of the related
pictures correspondedto the iconic aspect of the presentedsign (iconic sign-
picture correspondence), the other pictures did not correspond (no correspon-
dence). Referring to the example, the corresponding picture was the picture with
the crookedbeak (markedblack in Figure 3), which is typical for eagles. In the
word condition no picture corresponded to any of the words because none of
them were iconic. In order to control the outcomes in the sign condition, i.e. to
make sure that faster response times for iconic items were causedby a general
importance of the iconic feature for both the sign andthe wordconcept, we
divided the spoken words into two groups: 1. translated words of the iconic
signs, 2. translatedword s of non-iconic signs.
The German Sign for eagle highlights the beak of an eagle by (1) the
shape of the articulator (bended index-nger), (2) location of the hand (in the
area of the signers nose), and(3) the movement of the hand (tracing the shape
of the beak). To conrm our hypothesis that the iconic features of a sign have
an impact on the relatedsemantic concept, the deaf participants andthe bilingual
participants hadto react fastest to the black-markedbeak-picture.
The selectedword s of German Spoken Language (DLS) were requiredto
correspondto the signs of German Sign Language (DGS), thus to be easily
translatable. It was essential that signs and words were well-known to deaf and
hearing subjects respectively. Subjective frequency of occurrence of signs and
words was estimated by eight native DGS-signers and eight native DLS-speakers
with the computerizedVEIP [Verfahren zur Einschtzung von Itemparametern]
(Grote 1999), a computer program for easy registration of ratings andjud gements
which we hadd evelopedfor these purposes.
An attempt was made to keep
frequency andnumber of syllables between sign items andtheir spoken transla-
Figure 3.Example Stimulus-Set Adler with related pictures (test-items) and unrelated
pictures (distractors). One stimulus set out of twenty.
tions as comparable as possible.
For analysing the data of the experiment, the mean response time of the correct
responses per relation andsubject were computed. The average response time
data of deaf (n =20), hearing (n =20) andbilingual subjects (n =20) are shown
in Figure 4.
To evaluate the dierences in response time for correct judgments of
semantically relateditems, statistical analysis was performed on the mean latency
per relation category (iconic sign-picture correspondence vs. no correspon-
For the deaf subjects the dierences between mean response times of
the categories iconic sign-picture-correspondence and no iconic correspon-
dence was signicant. Mean response times for items with iconic sign-picture-
correspondence were 160.45 msec faster than the ones for items without such
correspondence. The responses of the bilingual participants in the sign-picture
task were similar to the deaf participants. They also showed a signicant
dierence between mean response times of both item-categories. Mean response
times for items with iconic sign-picture-correspondence were 128.80 msec faster
than the ones for items without such correspondence.
For the hearing subjects no dierences in judging the two categories of
Verification Task with Signs/Words and Pictures
Deaf Bilinguals Hearing Bilinguals
Signs Words
Iconic Sign-Picture-Correspondence



No Iconic-Correspondence
Figure 4.Mean response times for the iconic sign-picture andword -picture combinations.
The rst columns represent the mean responses to the pictures which correspondto iconic
signs andto their translations into spoken words, respectively, whereas the second
columns show mean response times to pictures which correspondto non-iconic signs and
to their spoken-wordtranslations.
word-picture combinations (wordtranslations of iconic vs. wordtranslations of
non-iconic signs) were found. This result was expected since none of the words
showedany iconic correspondence to the pictures. Yet, more astonishing were
the response times of the bilinguals in the word-condition of the task. The
bilinguals showed in the word-picture-task signicantly dierent response times
comparedto the hearing subjects, but similar results to their responses in the
sign-picture-task. The response times for the words which corresponded to iconic
signs were signicantly faster (727.60 ms) than the response times for the words
which corresponded to the non-iconic signs (832.80 ms).
We had expectedthat the response times for deaf, hearing andbilingual subjects
wouldd ier depending on the category of item-pair used in the experiment. The
results are compatible with this prediction. The deaf subjects were signicantly
Figure 5.Test-procedure of the triad-comparison task with signs
faster in judging iconic sign-picture relations, such as eagle combined with
beak, than in judging non iconic sign-picture relations, like eagle combined
with wing. The hearing subjects performedd ierently from the deaf, especially
with respect to the relevant item-category containing spoken-wordtranslations of
iconic signs. Thus, the iconicity of the sign seems to accelerate the response
times of the deaf and bilinguals to the corresponding picture, whereas in the case
of non-iconic signs we didnot nd such an eect. Similar to the latter, we
foundfor the hearing group no signicant dierence in judging words which
show no resemblance to the pictures. In contrast to the hearing subjects, the
bilinguals showeda somewhat surprising result in the word-picture task. Even in
the vocal mode we obtained for bilinguals signicantly faster response times for
word-picture pairs when the preceding word was a translation of an iconic sign.
These results can be interpretedas a rst evidence that sign language iconicity
has an impact on the semantic relations in the lexicon. The properties of a
concrete referent which are highlightedby the iconic aspect of a linguistic sign
seem to play a central role in the corresponding semantic concept in the mental
lexicon. This suggests that the architecture of the mental lexicon is shapedby the
iconic properties of a language. Furthermore this inuence of iconicity on
semantic conceptualization seems to exceedthe language boundaries, i.e. it seems
to work cross-linguistically.
Study 2: Triad-comparison-task with signs/words and pictures
The second study examined whether deaf, hearing and bilingual subjects show
dierent choices in a task where they were asked to decide which of two
presentedpictures had a stronger semantic relation to a presentedsign or word,
respectively. The number of choices in favour of a specic semantic relation was
measuredfor each group of participants. It was expectedthat in the triad-
comparison task the relative number of choices of the deaf and bilingual
Figure 6.Test-procedure of the triad-comparison task with words
participants for the pictures which correspondto the iconic features of the signs
wouldbe greater than the relative number of choices for the pictures which do
not depict the features. The relative number of choices of the hearing participants
were assumednot to dier for both types of pictures.
For the deaf subjects the signs and pictures were displayed one at a time on a
computer monitor. Again, in order to provide visual masking the sign-video was
centeredat the top of the monitor, the rst picture at the left bottom and the
secondone at the right bottom. For the hearing subjects the spoken words were
connected with a visual marker followedby two pictures presentedone by one.
The subject was instructedto press the right button if he or she thought the
picture on the right bottom of the screen hadthe strongest relation to the sign/word
andthe left button if he/she thought the picture on the left bottom of the screen was
the strongest related. The subject was told to do so as accurately and as quickly
as possible at any time from the start of the thirdtest item, but fast responses did
not alter the timing. Each subject was instructed by a standardized explanation
(signed/written vs. spoken/written) and was given a practice of 10 runs.
The experimental materials overlappedthose of Experiment 1. By removing the
unrelateditems the test list contained20 target items combinedrand omly with 3
relateditems. Every possible combination of the items within a taxonomy was
realized. Thus, there were 60 combinations of a target item (sign/word) with two
relateditems (pictures). The presentation position was completely counterbal-
anced. The items were grouped in triads of three series of 20 runs.
Triad-Comparison Task with Signs/Words and Pictures
Deaf Bilinguals Hearing Bilinguals
Sign-Picture Word-Picture
Iconic Sign-Picture-Correspondence No Iconic-Correspondence




Figure 7.Relative Number of Choices for the Sign-Picture andW ord-Picture Combina-
tions. The rst columns represent the mean number of choices to the pictures which
correspondto iconic signs andto their translations into spoken words, respectively,
whereas the secondcolumns show mean response times to pictures which correspondto
non-iconic signs andto their spoken-wordtranslations
The results of study two support the outcomes of study one (Figure 7).
participants chose signicantly more pictures which depict the iconic aspects of
signs than pictures which correspondto semantically related, but not to the iconic
aspects of a sign. For the deaf subjects the dierence between the mean number
of choices of the categories iconic sign-picture-correspondence and no iconic-
correspondence was signicant. The responses of the bilingual participants in the
sign-picture task were similar to the deaf participants. They also showed a signi-
cant dierence in the number of choices for sign-picture combinations with iconic-
correspondences compared to combinations with no iconic-correspondences.
For the hearing subjects no dierences in the number of choices for the
words corresponding to the iconic signs vs. the non-iconic signs were found. As
in study 1 this result was expected. The responses of the bilinguals in the word-
picture triad-comparison task was analogous to study 1 somewhat dier-
ent from the results of the hearings andsimilar to the results in the sign-picture
task. The bilinguals showedin the word-picture-task signicantly more choices
for the pictures which correspondto the iconic features.
The results verify the hypothesis that deaf and bilinguals choose signicantly
more pictures which depict the iconic feature. E.g. with respect to the sign for
cow which highlights the horn of a cow, the deaf and bilinguals chose
signicantly more often the sign-picture combination cow andhorn than the
combination cow andud der or cow andspottedhid e. Consequently the
attribute has horns for the concept cow seems to be a central one for deaf and
bilingual signers, one that is more important, for example, than the attribute has
an udder or has a spotted hide. In contrast, the hearing participants did not
choose the picture with the horn as often, which indicates that horns are not
that central for the concept of a cow, rather the participants chose more
frequently the picture presenting a spottedblack and white hide. At least for the
participants from the region of North Rhine-Westphalia, the test-location, a
black andwhite hide seems to be the most important attribute of a cow. To sum
up, comparedto the hearing participants, the deaf as well as the bilingual
participants made signicantly more choices for pictures which correspond to the
iconic features of the matchedsigns. For the bilingual participants this eect is
foundalso in the task with words. Thus, the overall pattern of results suggests
that the semantic organization of a mental lexicon is inuenced by the iconicity
of a language.
4. The cognitive impact of sign language iconicity on conceptualization
In the present study one possible cognitive function of iconicity was explored.
The results provide empirical evidence for the inuence of sign language
iconicity on the strength of the semantic relations in the conceptual system. The
hypothesis that the characteristics or aspects of an entity, which are accentuated
by the iconic feature of the sign, play a central role in the corresponding
semantic network is supported. Thus, the iconic form of the sign not only reects
conceptual andperceptual structures but seems to have an impact on semantic
conceptualization. Hence the empirical ndings not only serve to conrm the
Peircean conception of iconicity but furthermore to specify it with respect to
language in a double way: Firstly, they demonstrate that linguistic signs as a
main class of symbols in fact can include iconic parts. When an iconic sign
becomes integratedinto a language system, it behaves mainly as a symbol, i.e.
as Peirce says, its symbolic, living character is the prevailing one (EP 2: 10).
Nevertheless the predominance of its symbolic character does not imply that the
iconic dimension of a linguistic sign becomes completely blanked out or deleted.
It seems that even in the case of primary imagic iconicity, the persistence of the
iconic function is not restrictedto specic contexts of poetics or advertising etc.,
i.e. to processes of re-iconization in which the sign is unhingedfrom its regular
usage. Rather, the iconicity appears to maintain a semiotic function in ordinary
language usage too, albeit in a reducedandmod ied manner.
Secondly, regarding the cognitive impact of signs, the results support the
repistemological key assumption of Peirce that our way of conceptualizing the
worldis mediated by signs. Following Peirce we argue that the iconic relation
between sign and object does not apply to a sign-independent real object but to
the Object as cognizedin the sign (EP 2: 495; CP 8.183; see Nth 2001: 20f.)
the immediate object in Peircean terms. With an iconic sign its object of
reference is conceptualizedin a specic way, namely as an object with just that
property that denes the iconic relation. This means that through the relation of
similarity which constitutes the iconic dimension of a sign, a specic feature is
singledout andtherewith becomes essential for the assignedobject. Exactly this
process of choosing a single feature andgiving it a prominent status in the
conceptualization of the object is what we observedin the experiments. One
main function of iconic signs might therefore consist in marking specic features
as particularly relevant for a concept.
Considering psycholinguistic models of the semantic lexicon, e.g. spreading
activation network models (e.g. McClellandandRumelhart 1981; Anderson
1983; Dell andOSeaghd ha 1991, 1992; Bock andLevelt 1994), it can be
assumed that semantically relatedconcepts which are reinforcedby the form of
a specic sign have a very close or strong connection to the relatedreferent.
Whenever specic characteristics andattributes of a referent are highlighted in
an iconic sign, the relatedconcepts become reactivatedandthe semantic relation
is reinforcedrepeated ly. Moreover, the semantic network seems not to be
bounded by a single linguistic system. The results of the bilinguals in the word-
task suggest that the sign language system somehow inuences the vocal
language system. Comparedto the hearing group the bilinguals showedfaster
responses anda greater number of choices for word-picture combinations where
the words corresponded to iconic signs. Thus, the semantic categories of
bilinguals seem to be inuenced by both language systems.
In sum, the empirical ndings provide evidence for a moderate version of
linguistic relativity. Whereas most linguists who concede the existence of
iconicity in language explain the relation between semantic structure and
iconic aspects of linguistic form in terms of a linguistic reection of conceptual
or experiential structures, the results also indicate a reverse eect from language
to semantic conceptualization. Most notably the results of the bilinguals ascribe
the possible semantic impact of iconicity not only to situational meaning but also
to the structure of the mental lexicon. The fact, that the bilinguals show an
acceleratedresponse even in the vocal modality where no iconic signs were
presented, demonstrates an eect of iconicity not restrictedto the generation of
situational meaning in the actual use of an iconic sign. Moreover, the data found
in the presentedstud y lendsupport to the thesis that linguistic iconicity has an
inuence on the formation of semantic network-relations. One major challenge
for the future of iconicity research is thus to investigate andd escribe the
underlying cognitive processes which lead to the eects discussed in this paper.
1. Sign Languages, like German Sign Language (DGS), are fully autonomous languages, which
possess all the properties common to all natural languages it is a rule-governed, grammatical
symbol system that changes over time andthat members of a community share. Signed
languages are distinct from spoken languages in both grammar and lexicon. The phonological
units to construct signs are (1) manual parameters such as handshapes, hand-orientations, hand-
movements, hand-locations, and (2) non-manual signals such as facial expressions (brow raises,
eye blinks, mouth gesture andmouth picture), headnod s, headtilts, shifts of the body and
shoulders. All parameters have grammatical functions in sign communication.
2. Tone languages in which words have an additional meaning by tone modication are an
3. The merging of multiple morphemes denoting salient attributes of the noun-referent is an
important characteristic of sign languages. It is inherently economical andan advantage of the
visual-gestural mode.
4. See their denition of iconicity on the web
5. Wilcox (2000) calls the relationship between mental conceptions of linguistic forms and
meanings cognitive iconicity.
6. When a newly introduced high-iconic sign violates linguistic regularities within the sign
language system it will lose some of its iconicity in the process of symbolization because the
symbolic form moves towards a sign which conforms more closely to linguistic rules.
7. The study is part of an interdisciplinary research project at the Center for Cultural Research
(University of Cologne) on the Mediality of linguistic signs. Our theoretically and empirically
orientedresearch is basedon the assumption that language has a constitutive impact on
conceptualization andthat the material format of a sign has a structuring eect on its content.
Two of the projects at the Center deal with medial dierences between signedand spoken
languages andtheir inuence on cognitive structures. Within this wider framework we are
interestedin language iconicity andits possible cognitive functions.
8. Hearing signers are bilinguals, who are also calledcod as, which is an abbreviation of
Children of deaf adults. The bilinguals learned German Sign Language and German Spoken
Language at the same time anduse both languages regularly. The three groups of participants
were matchedwith respect to schooling, gender andage.
9. At the moment we are still collecting data with the VEIP regarding the frequency, iconicity,
imagine ability of the signs/spoken words and pictures we are using as test-items in dierent
10. T-Tests for paired samples for 1. iconic sign-picture correspondence vs. no correspondence:
Deaf subjects t(19) =19,657; p <.0001; bilingual subjects t(19) =6,831; p <.0001, 2. word
translations of iconic signs vs. wordtranslations of non-iconic signs: bilingual subjects
t(19) =4,981; p <.0001.
11. T-Tests for paired samples for 1. iconic sign-picture correspondence vs. no correspondence:
Deaf subjects t(19) =11,741; p <.0001; bilingual subjects t(19) =4,720; p <.0001, 2. word
translations of iconic signs vs. wordtranslations of non-iconic signs: bilingual subjects
t(19) =6,545; p <.0001.
Anderson, J. R. 1983. The Architecture of Cognition. Cambridge, Mass.: Harvard Universi-
ty Press.
Armstrong, D. F., W. C. Stokoe andS. E. Wilcox. 1995: Gesture and the Nature of
Language. Cambridge, Mass.: Cambridge University Press
Bellugi, U. andS. D. Fischer. 1972. A comparison of sign language andspoken
language. Cognition 1: 173200.
Bellugi, U. andE. S. Klima. 1976. Two faces of a sign: Iconic andabstract. In Origins
and Evolution of Language and Speech, S. R. Harnad, H. D. Steklis and J. Lancaster
(eds), 514538. New York: Annals of the New York Academy of Sciences, 280.
Bock, K. and W. Levelt. 1994. Language production. Grammatical encoding. In
Handbook of Psycholinguistics, M. A. Gernsbacher (ed.), 945984. San Diego:
Academic Press.
Bouissac, P., M. Herzfeldand R. Posner. (eds) 1986. Iconicity. Essays on the Nature of
Culture. Festschrift for T. A. Sebeok. Tbingen: Stauenburg.
Dell, G. S. and P. G. OSeaghdha. 1991. Mediated and convergent lexical priming in
language production. A comment on Levelt et al. (1991). Psychological Review 98:
Dell, G. S. and P. G. OSeaghdha. 1992. Stages of lexical access in language produc-
tion. Cognition 42: 287314.
Elgin, C. Z. 1996. Index and icon revisited. In Peirces Doctrine of Signs. Theory,
Applications, and Connections, V. M. Colapietro andT . M. Olshewsky (eds),
181189. Berlin, New York: Mouton de Gruyter.
Engberg-Pedersen, E. 1993. Space in Danish Sign Language. The Semantics and Morpho-
syntax of the Use of Space in a Visual Language. Hamburg: Signum.
Engler, R. 1995. Iconicity and/or arbitrariness. In Iconicity in Language, R. Simone
(ed.), 3945. Amsterdam, Philadelphia: Benjamins.
Fehrmann, G. andE. Linz. 2002. Zeichen als Medium. In Die Kommunikation der
Medien, J. Fohrmann andE. Schttpelz (eds), (in press). Tbingen: Niemeyer
(Special Issue of IASL).
Fischer, O. andM. Nnny. 1999. Introduction. Iconicity as a creative force in language
use. In FormMiming Meaning. Iconicity in Language and Literature, M. Nnny and
O. Fischer (eds), xvxxxvi. Amsterdam, Philadelphia: Benjamins.
Frishberg, N. 1975. Arbitrariness andiconicity . Historical change in American Sign
Language. Language 51: 696719.
Grote, K. 1999. VEIP-Verfahren zur Einschtzung von Itemparametern. Software,
University of Cologne.
Hoemann, H. W. 1975. The transparency of meaning or sign language gestures. Sign
Language Studies 4: 151161.
Klima, E. S. andU. Bellugi. 1979. The Signs of Language. Cambridge, Mass.: Harvard
University Press.
Jger, L. 1978. F. de Saussures semiologische Begrndung der Sprachtheorie. Zeit-
schrift fr Germanistische Linguistik 6: 1830.
Jger, L. 1986. Der saussuresche Begri des Aposeme als Grundlagenbegri einer
hermeneutischen Semiologie. In Zeichen und Verstehen, L. Jger andC. Stetter
(eds), 733. Aachen: Rader.
Jger, L. 2002. Transkriptivitt. Zur medialen Logik der kulturellen Semantik. In
Transkribieren Medien/Lektre, L. Jger andG. Stanitzek. (eds), 1941. Mn-
chen: Fink.
Liddell, S. 1990. Four functions of a locus: Re-examining the structure of space in
ASL. In Sign Language Research: Theoretical Issues, C. Lucas (ed.), 176198.
Washington, DC: Gallaudet College Press.
McClelland, J. L. and D. E. Rumelhart. 1981. An interactive activation model of context
eects in letter perception. Part 1: An account of basic ndings. Psychological
Review 88: 375407.
Nth, W. 2001: Semiotic foundations of iconicity in language and literature. In The
Motivated Sign. Iconicity in Language and Literature 2, O. Fischer andM. Nnny
(eds), 1728. Amsterdam, Philadelphia: Benjamins.
Peirce, C. S. 199298. The Essential Peirce. Selected Philosophical Writings. Vol. 1
(1992), ed. by N. Houser and Ch. Kloesel; Vol. 2 (1998), ed. by the Peirce Edition
Project. Bloomington, Indianapolis: Indiana University Press. (Quoted as EP)
Peirce, C. S. 193158. Collected Papers. Vols. 16, ed. by C. Hartshorne and P. Weiss;
vols. 78 ed. by A. W. Burks. Cambridge, Mass.: Harvard University Press. (Quoted
as CP. Reference is made to vols. and paragraphs)
Pelc, J. Iconicity. Iconic signs or Iconic uses of signs? In P. Bouissac, M. Herzfeldand
R. Posner (eds), 715.
Pizzuto, E. andV. Volterra. 2000. Iconicity andtransparency in sign languages: A cross-
linguistic cross-cultural view. In The Signs of Language Revisited. An Anthology to
Honor Ursula Bellugi and Edward Klima, K. Emmorey andH. Lane (eds), 261286.
Mahwah, NJ: Erlbaum.
Ransdell, J. 1986. On Peirces conception of the iconic sign. In P. Bouissac, M.
HerzfeldandR. Posner (eds), 5174.
Saussure, F. de. 1968. Cours de linguistique gnrale. dition critique par R. Engler,
Tome I. Wiesbaden: Harrassowitz. (Quoted as SECI)
Saussure, F. de. 1993 (191011). Troisime cours de linguistique gnrale dapres les
cahiers dE. Constantin. Saussures third course of lectures on general linguistics from
the notebooks of E. Constantin. French text ed. by E. Komatsu, engl. transl. R.
Harris. Oxford, New York: Pergamon. (Quoted as SCIII/KH)
Schick, B. S. 1990. Classier predicates in American Sign Language. International
Journal of Sign Linguistics 1: 1540.
Stokoe, W. C. 1960. Sign Language Structure: An Outline of the Visual Communication
Systems of the American Deaf. Bualo, NY: University of Bualo.
Supalla, T. 1986. The classier system in American Sign Language. In Noun Classes
and Categorization. Proceedings of a Symposium on Categorization and Noun
Classication, Eugene, Oregon, Oct. 1983, C. Craig (ed.), 181214. Amsterdam:
Supalla, T. 1990. Serial verbs of motion in ASL. In Theoretical Issues in Sign Language
Research. Vol. 1: Linguistics, S. Fischer andP . Siple (eds), 127152. Chicago:
Chicago University Press.
Wilcox, P. P. 2000. Metaphor in American Sign Language. Washington, DC: Gallaudet
University Press.
</TARGET "gro">
AUTHOR "William J. Herlofsky"
TITLE "What You See Is What You Get"
WIDTH "150"
What You See Is What You Get
Iconicity andmetaphor in the visual language of
written andsignedpoetry: A cognitive poetic approach
William J. Herlofsky
Nagoya Gakuin University
The transformation of the world of objects into the world of signs is founded
on the ontological presupposition that it is possible to make replicas.
Yuri M. Lotman
1. Introduction
In earlier research (Herlofsky 2001) speculating on the role of iconicity in the
evolution of language, I discussed a possible scenario for a perceptual/conceptual
shift from real-world-space to mental syntactic-space. In footnotes to that
research (Herlofsky 2001: 6465) I briey suggestedplausible intermediate steps
in the evolutionary transition, including reference to the notions of mental space
(Fauconnier 1985, 1997, 2001) andmetaphor (Keller 1998), andthe possibility
of sign languages providing further insights into the evolutionary emergence of
language (see also Armstrong et al., 1995, andStokoe 2000). Although an
extended discussion of the notions of mental spaces, metaphor, and sign languag-
es was not within the scope of the previous research project, in the present paper,
I wouldnow like to consider (though not from an evolutionary perspective) the
relationships among these notions, andhow the signing spaces of sign language
might make some of these mental spaces visible.
This paper attempts to illustrate the appropriateness ande ectiveness of the
mental space/blending framework (Fauconnier 1997, Fauconnier and Turner
1996, Fauconnier 2001) of cognitive linguistics for the investigation of iconicity
andmetaphor in visual language, especially the analysis of visual iconicity in
written andsignedpoetry . Sections 2 and3 provide basic backgroundinforma-
tion on previous relevant cognitive linguistic andsign language research that will
be combinedin a cognitive poetic analysis of visual iconicity in written and
signedpoetry in Section 4. Section 2 presents a brief summary of Lako and
Johnsons (1999) integratedtheory of primary metaphors, where it is claimedthat
metaphors are an integral part of our conceptualization of the worldaroundus,
andthat conceptual metaphor is one of our central cognitive andlinguistic tools.
Section 3 provides various examples of iconicity in sign languages, examples
basedon Taubs (1997, 2001) treatment of iconicity andmetaphor in American
Sign Language (ASL), andmy own analysis of metaphor andiconicity in Japan
Sign Language (JSL). Section 4 begins by illustrating how the cognitive ap-
proach in Section 2 can be appliedto literary analyses by examining the relevant
portions of Hiragas (1999a,b, 2000, this volume) research on metaphor,
iconicity andblend ing in written poetry, andthen combines this andthe analyses
from previous sections to illustrate the overall eectiveness of the cognitive
poetic framework for the analysis of visual iconicity in JSL poetry. The paper
ends by concluding that the generalizations made concerning metaphors and
iconicity in spoken/written languages are applicable, without major modality
eects, to sign languages.
2. An integrated theory of metaphor
This paper approaches iconicity in visual language through the analysis of
metaphors within the cognitive linguistic framework. There are two main reasons
for using the cognitive linguistic framework in this paper. The rst reason is that
cognitive linguistics does not consider language to be autonomous from other
cognitive functions, nor does it deal with form and meaning separately, points of
view that contrast sharply with, for example, Chomskys formalist framework, in
which competence andperformance are consideredseparate domains, andwithin
which it wouldbe next to impossible to explain how the semantic component
could inuence the form that a language takes. The second reason is that
metaphors are vital to the analysis of iconicity in sign languages in that they
allow for the scope of iconic signs to be extended beyond the concrete to
abstract concepts (see, for example, Taub 1997, 2001, andW ilcox 2000), and in
cognitive linguistics, conceptual metaphors are a central component. Within the
cognitive linguistic framework, then, form, meaning, metaphor andiconicity can
all be treatedtogether , andthis holistic, embodiedapproach is most suitable to
the present papers objective of a clear andaccurate description of iconicity in
the visual language of signedpoetry .
As a brief introduction to cognitive linguistics and metaphor, it will be
benecial to utilize passages from Lako andJohnson (1999, hereafter L&J),
since their work is the most recent andsuccinct summary of the cognitive linguistic
theories relevant to the analysis of metaphors in this paper. First, let us start with
L&Js (1999: 3) summary of three major ndings in cognitive science research:
The mind is inherently embodied.
Thought is mostly unconscious.
Abstract concepts are largely metaphorical.
According to L&J, then, the body and mind are not separate entities, as much of
Western philosophy andscience (andChomsky s Cartesian approach) seem to
assume, but are instead intimately connected, in fact, united, in the embodied
mind. All of our senses, our movements, and the sensorimotor system in general,
in this conceptualization, interact directly with the embodied mind, and our
sensorimotor bodily experiences of early life become part of the unconscious
embodied mind, which later in life, become part of our abstract conceptual
system that is largely metaphorical.
What is especially important for the present analysis is the cognitive
linguistic notion that metaphors are not just a supercial way of using language
in some poetic language performance (although there will be a discussion of
metaphor in poetry later), but that they are in fact integral to our conceptualiza-
tion of many basic abstract ideas. L&J contend that the way we think about
many abstract ideas is to a large extent metaphorical, and that although these
abstract ideas may have non-metaphorical conceptual skeletons, each is eshed
out by conceptual metaphors.
It is L&Js hypothesis that the early experiences of all normal human children
result in the embodiment of various primary metaphors. L&J utilize the work of
other cognitive scientists to formulate their integratedtheory of metaphors. First,
consider Johnsons (1997) theory of conation, as summarized in L&J (1999: 46):
For young children, subjective (nonsensorimotor) experiences and judgments,
on the one hand, and sensorimotor experiences, on the other, are so regularly con-
ated undierentiated in experience that for a time children do not dis-
tinguish between the two when they occur together. For example, for an infant, the
subjective experience of aection is typically correlatedwith the sensory experi-
ence of warmth, the warmth of being held. During the period of conation,
associations are automatically built up between the two domains. Later, during
a periodof dierentiation, children are then able to separate out the domains,
but the cross-domain associations persist. These persisting associations are the
mappings of conceptual metaphor that will leadthe same infant, later in life,
to speak of a warm smile, a big problem, anda close friend.
The cross-domain associations can be illustrated as in Figure 1, where the
concrete, physical sensorimotor experience of warmth can be locatedin the
source domain on the left, and the subjective (non-sensorimotor) experience of
aection in the target domain on the right.
These cross-domain associations then develop into conceptual blending as in
Figure 1.
Gradys (1997) description of primary and complex metaphors (L&J 1999: 46):
All complex metaphors are molecular, made up of atomic metaphorical
parts called primary metaphors. Each primary metaphor has a minimal struc-
ture andarises naturally, automatically, and unconsciously through everyday
experience by means of conation, during which cross-domain associations are
formed. Complex metaphors are formed by conceptual blending.
Blending can be diagrammed as in Figure 2 below, where again warmth is on the
left, anda ection on the right. In the blend, some extension of the simple
association of warmth anda ection emerges, anda more complex metaphor like
a warm smile, or perhaps, a warm atmosphere is created.
Combining these and other ideas, L&J (1999: 128) have developed what
Source Target
Figure 2.
they refer to as an integratedtheory of primary metaphors, simple descriptions
of which appear below:
Correlations in our everyday experience inevitably lead us to acquire
primary metaphors, which link our subjective experiences andjud gments to
our sensorimotor experience. These primary metaphors supply the logic, the
imagery, andthe qualitative feel of sensorimotor experience to abstract
concepts. We all acquire these metaphorical modes of thought automatically
andunconsciously andhave no choice as to whether to use them.
Many, if not all, of our abstract concepts are dened in signicant part by
conceptual metaphor. Abstract concepts have two parts: (1) an inherent,
literal, nonmetaphorical skeleton, which is simply not rich enough to serve
as a full-edged concept; and (2) a collection of stable, conventional
metaphorical extensions that esh out the conceptual skeleton in a variety
of ways.
The fundamental role of metaphor is to project inference patterns from the
source domain to the target domain. Much of our reasoning is therefore
This, then, has been a very brief and simplied introduction andillustration of
the role of metaphor in conceptualization as theorizedin cognitive linguistics.
The following sections will examine how the integratedtheory of metaphor and
the above model of blending are relevant to a description of iconicity in sign
language, as well as iconicity in written andsignedpoetry . Section 3 will rst
briey consider iconicity in sign language.
3. Iconicity in sign language
Let us begin this section with Volterra andErting s (1994: 2) statement about the
existence of iconicity in sign language:
An interesting dierence between the signedand the spoken modality is that
iconicity is present in sign language to a greater degree than it is in spoken
languages (Klima andBellugi 1979; Mandel 1977). While iconic characteris-
tics of sign languages have undergone signicant historical modication and
grammaticalization, it is often possible to perceive a relationship between a
sign andits referent. At the same time, each sign language can choose to
conventionalize a particular iconic relationship in an arbitrary way. In sign
languages, then, a certain degree of iconicity coexists with one of the funda-
mental features of languages: arbitrariness.
Now let us consider the denition of iconicity that appears on the home page
( of the Iconicity in Language and
Literature society:
Iconicity as a semiotic notion refers to a natural resemblance or analogy between
the form of a sign (the signier, be it a letter or sound, a word, a structure
of words, or even the absence of a sign) and the object or concept (the
signied) it refers to in the world or rather in our perception of the world.
The references to concepts and our perception of the world open the door to
considering iconicity to be a phenomenon of the perceptual/conceptual domain of
cognitive mental spaces, or the mental models from Taub (1997) discussed below.
Taub (2001, 1997: 36), in her study of iconicity in ASL, begins by claiming
that for an expression to be iconic, some aspect of the expressions physical
form resembles a physical referent. Importantly, however, she adds that this
resemblance is not simply an objective fact about two entities, but a product of
our cognitive processing. Andnot only are schematization and conceptual
mapping involved, but mental models of the entities as well.
Specically, when we compare two entities (for similarity), we attempt to set
up structure-preserving correspondences between our mental models of the two
entities. This means that for each entity, we gure out its relevant parts and
the relations between the parts: this is the perceivedstructure of the entity.
Then, given the structure of one entity, we look for corresponding structure in
the other entity. The more correspondences we can nd, the more we believe
the two entities resemble each other.
Taub illustrates the structure-preserving correspondences of iconicity in an ASL
sign by comparing a pair of human legs with the middle and index nger of a
human hand extended downwards from a st (like an inverted V-sign). The two
ngers can be seen as resembling a pair of human legs because of the structure-
preserving correspondences (pairs of two long, thin moveable objects, with joints
in the middle, and joined to a base (the human body and the st)). These
structure-preserving correspondences make it easy to use these two extended
ngers in iconic signs for verbs involving the legs, such as WALKING and
In this example, the FORM (Taub uses the term SHAPE here, but to
avoidconfusion with the term handshapes, I will use the term FORM) of the
legs corresponds to the FORM of the ngers, and so, this is a FORM=FORM
correspondence or resemblance. When the ngers are wiggled in a certain way
andmovedalong a certain path for the signs for WALKING or RUNNING, the
structure-preserving ngers-for-legs correspondence is still a FORM=FORM
relationship, but the moving of the ngers along a path to imitate the walking
or running along a path, is a PATH=PATH relationship (see Taub 1997, 2001
for further discussion).
Before preceding any further, it may be helpful to note that the signs of sign
languages are usually articulated with the upper body (the hands and arms for
manual signs, and the upper body (from waist to head) and face for non-
manual signs), involving movements and handshapes at dierent locations
andwith dierent orientations within the signing space, which is usually
somewhere between the top of the headandthe waist, extending horizontally a
bit wider than shoulder space to the left and right, and to about arms-length in
front of the body. And since the signing space is usually in front of the upper
body, and the articulators are also usually above the waist, having the two-
ngered handshape (like an invertedV -sign) with sucient structure-preserving
correspondences to represent two legs is not only convenient but ecient. It is
much more convenient andener gy/space-ecient to let your ngers do the
walking, than to require that your legs and whole body do all the work.
The above FORM/PATH description illustrates how the components or
building blocks of iconic signs can be identied according to their structure-
preserving correspondences both in form and movement. In sign languages, then,
dierent from spoken (sound-based) languages, where iconic expressions
generally imitate sounds, it is form (FORM) and movement (PATH) that can be
readily imitatedby the forms andmovements of signs. For concrete forms/
movements the physical forms/movements of the signs are sucient, andthe
general types of associations of form andmovement can be explainedand
diagrammed as follows.
The FORM=FORM correspondences, as in the ngers-for-legs example
discussed above, are perhaps the most transparent type of iconicity in sign
languages. This type of FORM=FORM iconicity also exists in the Chinese
characters or kanji that are usedin the Japanese writing system, as illustrated
below, in the kanji for tree, where the horizontal line in the kanji on the left
indicates the branches/leaves, and the diagonal lines on both sides of the vertical
line indicate trunk/roots.
In American Sign Language, the sign for TREE is also a FORM=FORM
relationship, where the arm andhandcorrespondto the tree trunk and branches,
For JSL, however, the TREE sign is a PATH=FORM relationship, where the
hands draw an outline of a tree trunk in the signing space.
There is also a PATH=PATH relationship, as in the JSL sign for GO below.
And, a less common FORM=PATH relationship, as when the two ngers moving
out from the eyes indicate the path of vision in the JSL signs for LOOK and SEE.
As one might expect, in dynamic living languages, these FORM-PATH
dichotomies of actual signs are not always so clear-cut, since signs are often
made up of combinations of the above relationships, as seen in the WALKING/
RUNNING example above. In addition, the involvement of metonymy as well as
iconicity complicates matters, as in the JSL sign for TREE above, in which only
the trunk of the tree represents the tree, andthis sometimes makes the iconicity
less transparent. These complications andothers that wouldhave to be included
in a complete description of iconicity in JSL are beyond the scope of the present
paper (see Taub 2001 andW ilcox 2000 for more complete discussions of
iconicity in ASL). The descriptions above andbelow , however, shouldbe
sucient to leadus into the discussion of metaphor andiconicity in signed
poetry in the next section.
For concrete objects that can be representedby handshapes or other upper
body parts that preserve the structural correspondences, then, it is not dicult
to see how iconicity is important in sign production. For abstract concepts,
however, such as understanding or remembering, for example, it is not so
obvious how the signs for these concepts, if they are to be iconic, can resemble
the referents. Let us begin the discussion of abstract signs by rst considering the
concepts understanding and remembering, and the words that are often
associatedwith them.
In spoken English, for example, to describe not understanding or remem-
bering something, we might say things like, I couldnt quite grasp what you
were saying, or I cant quite holdon to so much information. In other words,
the concrete sensorimotor experiences of grasping andhold ing are usedas
metaphors for the abstract concepts of understanding or remembering. And since
these cognitive activities are thought to be taking place in the brain/head(when
we dont understand something too dicult, we say, That was way over my
head), we have a situation like that discussed in Section 2, where there emerges
a metaphorical blendof a concrete, sensorimotor experience (grasping, holding)
and a more abstract, subjective experience (understanding, remembering) into a
metaphor where our brain can reach out andgrasp things, andthen holdon
to them. This is illustratedin Figure 4 below.
We can now consider how this type of metaphoric blend can be seen in
grasping, holding
understanding, remembering
Figure 4.
the iconic nature of the signs of JSL. First, let us describe the JSL sign for
RECEIVE. For this sign, an arm is rst extended out in front of the body, with
the handopen, palm facing upward, andthen the handis closedas the arm is
withdrawn toward the body. This RECEIVE sign is thus a metonymic and iconic
imitation of getting something in the handandpulling it towardthe body.
A metaphorical extension of this sign is createdfor the JSL sign for
REMEMBER. The location of many signs associatedwith cognitive functions
occur close to the head(since, as mentionedabove, cognitive activities are
thought to take place in the brain/head, also see Sweeter 1990 for a similar
discussion for spoken languages), and the sign for REMEMBER is no exception,
made by extending an arm upwards, palm open and facing toward the rear, and
then closing the hand andbringing the closed hand (st) down close to the head.
This, then, is the sign for RECEIVE performednear the head, andso, metaphori-
cally, it enacts the receiving of information into the head.
This sign is of course quite similar to the representation of the blendthat
Figure 5.
appears in Figure 4. In this way, mental-space blends become somewhat visible
in the signing space of sign languages. This possibility will be exploredin more
detail in the following section.
Although the iconic andmetaphorical aspects of these andsimilar signs
have been recognizedfor quite some time, an integratedd iagrammatic model for
describing this iconic and metaphoric mapping between the cognitive domains of
form andmeaning has been lacking (see, however, Lako andT urner 1989,
Taub 1997 and2001, Wilcox, 2000). Hiraga (1999a,b, 2000, this volume),
however, in a series of insightful papers, has woven the relationships between
metaphor andiconicity into a cohesive fabric, andapplied this fabric to the
analysis of written poetry. A brief summary of the relevant portions of Hiragas
framework will be presentedat the beginning of the next section, andthen there
will follow andextension of Hiragas analysis to signedpoetry .
4. Iconicity and metaphor in written and signed poetry
Since Hiragas (this volume) analysis is readily available to anyone reading the
present chapter, this section will begin with only a very brief summary of the
relevant portions of Hiragas chapter. In her analysis of a Japanese haiku poem,
Hiraga treats metaphor andiconicity together in a framework similar to that
described in previous sections. Haiku are short seventeen syllable (or mora)
poems dealing with some aspect of nature and/or the seasons (see Hiraga (1999,
2000, this volume) for further discussion). Hiraga analyzes a haiku written by
Matsuo Basho more than three hundredyears ago, andthis haiku, with a
translation, appears below:
araumi ya sado ni yokotau ama-no-gawa
rough sea: Sado on lie down heavens river
Rough sea: Lying down on Sado Island, the Milky Way
The metaphorical juxtapositions of the rough sea, Sado Island, and the
Milky Way in this poem are one of the focal points of Hiragas analysis. The
poem describes a natural scene in which a rough sea surrounds the island of
Sado, and the Milky Way, arching above, appears to lie down across the island.
This short poem contains much complicated embedded cultural information and
allusions, andHiraga (this volume) summarizes this haiku as follows:
In sum, this haiku text can be seen as a global metaphor in which several
images are comparedandpreserved to produce multi-layeredmeanings in the
blend. Most prototypically, such images include a rough sea and the river of
heaven (Milky Way) as an obstacle, a separating ow of water, andthe
legendary couple andthe prisoners as people preventedfrom reuniting with
their lovedones []. Furthermore, there are other implicit meanings in the
text. The text may evoke, for example, a feeling of elegy or a realisation of the
helplessness or nothingness of human beings in the face of powerful nature, in
this case, representedby terrifying rough waves andvast starry skies. It may
also imply varying kinds of contrast: a contrast of motion between the violent
waves andthe peaceful skies; a contrast of colour andlight between the black
andd ark sea and the silvery andbright skies; anda contrast of the real and
legendary between life stories of people and the love story of stars.
As can be seen in this summary, there are many complicatedid eas involvedin
this poem, andsome of the complicatedinterwoven blends are diagrammedin
Hiragas Figure 3, on p. 327 (this volume). Hiragas gure illustrates that the
possibilities for metaphors and cross-domain blends in this short seventeen-
syllable poem are quite amazing. The terms rough sea, Sado Island and
Milky Way activate numerous cultural associations andimage schemata, and
conventional metaphors, such as life is a (boat) journey andwater (waves) is
an obstacle to that journey. There is an additional allusion in the Chinese legend
about the Milky Way separating two lovers (stars). The Milky Way, Heavens
River in Japanese, is thus also an obstacle, andthe Japanese term for it involves
water. These, then, are some of the metaphors and cross-domain correspondences
representedin Hiragas Figure 3. (Again, for further details on these aspects of
the poem, please refer to Hiragas chapter. The intention here is just to indicate
the possibilities, andthen move on to how these metaphorical blends can be
mappedonto iconic form.)
In addition to the metaphors indicated above, the Chinese characters, or
kanji compounds that are used to write the three nouns in the poem, rough sea
( ), Sado ( ), and Milky Way ( ), all contain the radical (the three
dots on the left side of , and ) for water in the second kanji in the com-
pound. In other words, the metaphors of water as a path and water as an
obstacle are iconically mappedin the visual representations of the kanji. These
relationships andmappings are illustratedin Hiragas Figure 4, p. 332.
This brief description of Hiragas analysis should be a sucient foundation
for the following discussion of the same poem signed in JSL. To begin the
discussion of the JSL Rough Sea haiku, let us rst look at a literal JSL
translation of the poem in Figure 6.
This literal translation was suppliedby two deaf acquaintances of mine who
were not familiar with the original poem. The rst two signs are relatively
straightforward, with crookedly bent ngers making rough jagged movements
to the side. The sign for SEA used here is also the sign for WATER, but since
all the signs in the poem are accompaniedby mouth movements of the Japanese
equivalent word, in this case umi, a bilingual deaf person would have little or no
trouble interpreting the phrase.
The sign for SADO is also accompaniedby mouthing of the Japanese word
sado, but in this case the sign is the same as the sign for NIIGATA, the prefec-
ture that the islandof Sado belongs to. Therefore, to avoidconfusion with the
prefecture, ngerspelling of the sounds SA and DO can be added. The
HEAVENS RIVER signs are also accompaniedby mouthing, and this is especially
helpful for the sign for RIVER, which it shouldbe pointedout, is the same as for the
Figure 6.
SEA sign in the beginning of the poem. Though the signs are the same, the
mouthing of the words, and the context, should reduce ambiguity to a minimum.
The endof the poem consists of the sign for LIE DOWN, which, it should
be noted, comes at the end of the signed poem, but before Milky Way in the
original. It wouldbe technically possible to follow the original wordord er in the
signedpoem, but both signers/translators agreedthat the verb-nal form of the
poem was more natural in the JSL version.
Although the above JSL translation of the Rough Sea haiku is an accurate
translation, both deaf translators, after they had become more familiar with the
poem, were not completely satised with the artistic level of the translation. In
other words, though the translation conveyed the necessary information of the
original, it did not really convey in an artistic way the imagery of the original. After
some discussion and deliberation, the following more artistic, imagic version of the
poem was created. It is interesting that the changes correspond to the characteris-
tics of signedpoetry identied in Sutton-Spence andW oll (1999) for British
Sign Language poetry (see also Klima andBellugi 1979 who provide a similar
analysis for ASL poetry). These characteristics can be summarizedas follows:
more balance (symmetry) in use of (both) hands
smoother andmore graceful movements (uidity)
slower more fastidious blending of one sign into another
expansion of signs out of normal signing space
creating signs to t the poetic situation
There is at least a slight suggestion of all of these characteristics in the more
poetic-performance-like version of Rough Sea below (for space considerations,
the rst two signs for rough sea are omittedbecause they are the same as in
the previous version).
The increase in balance andsymmetry is only slightly suggestedin the sign
for HEAVENS RIVER (for a much clearer example of increasedsymmetry , see
Klima andBellugi 1979). In the rst version, only the right handwas used, but
in the second version, both hands are utilized, as the stationary left hand remains
in front of the stomach as a remnant of the SADO sign. A dierence in uidity
is impossible to detect in these drawings, but in the actual performance of the
poem, an increase in uidity was detectable. For reduced speed and more
fastidious blending of signs and movements as well, the drawings provide little
to make a judgment on, but what can be seen is the tendency to leave one hand
in place while the next sign begins. This can be seen in the sign for SADO,
when the left handis left in place as the right handsigns WATER/SEA. The left
handstill remains in place for the HEAVENS RIVER sign, andstays until being
combinedinto the LIE DOWN sign. This, then, gives an indication of the
blending of one sign into another in sign poetry, and a beautiful imagic
Figure 7.
representation of the Milky Way coming to lie down on Sado Island.
The sign for HEAVENS RIVER provides a slight indication of expansion
out of the normal signing space. Although the sign for RIVER is usually given
aroundstomach level, in this version it is performedabove the head, so it really
is in fact a heavens river. An example of the creation of a sign can be seen in
the new sign for SADO in the artistic version. To the rst part of the sign, of
hands moving back and forth in front of the body, there is added a motion like
WATER aroundthe stationary left hand, which remains in place to represent an
island, that is, the island of Sado.
This sign is not a standard sign, but a sign created for this artistic version
of the poem. As mentionedin the discussion of the rst version of the poem, the
sign for SADO is ambiguous, since it couldmean either Sado or Niigata. The
createdsign combines this rst sign with a variation of the sign for ISLAND.
The sign for ISLAND is made by making one hand(left) into a st andhold ing
it out in front of the body, palm facing down, and bring the other hand (right)
aroundit in a variation of the WATER sign, much like as in the created SADO
sign, the only dierence being the stationary hand. In the created sign, the
stationary left handis one half of the standardsign for SADO/NIIGATA (open
hand, palm facing upward), that remains in place while it is surrounded by the
(right-hand) WATER sign. In both the standard sign for ISLAND, and the created
sign for SADO, the basic iconic imagery is an island(the stationary hand)
surroundedby water (the WATER-sign handmoving around the stationary hand).
In this way, the artistic performances of signedpoetry can be analyzed, and
generalizations can be made that holdfor sign language poetry aroundthe world
(again, see Klima andBellugi 1979, andSutton-Spence andW oll 1999 for
supporting evidence from ASL and BSL, respectively). But if we ask why this
is so, then it will be helpful to return to our original cognitive linguistic frame-
work of mental spaces and blending. Recall Hiragas discussion of the poem, and
her diagram of iconic and metaphorical mapping in Figure 4 on p. 332. It is now
quite a simple task to produce a similar diagram for the signed version of the
same poem, andthat diagram appears below in Figure 8.
We can assume that the metaphorical mapping is all but identical for the
written version of the poem andthe signedversion. The dierence comes in the
iconic mapping. There is visual-iconic mapping, not for written kanji, but for the
visual signs. But note that, although there is clearly a dierence between the
kanji andthe JSL signs, the aspects of the visual shapes that are being imitated
are quite similar. That is, the path of water metaphor is involvedin both the
metaphorical mapping andthe iconic mapping in both the written andsigned
versions of the poem. For the written version, it is the three dot water radical on
the left of the kanji for , andthat iconically represents water, and for
the JSL version, it is three repetitions of the WATER sign (in Figure 8, the
second, fourth andsixth signs, where the right handmoves in a wiggling, wave-
like motion) in the same three nouns that iconically represent water.
Figure 8.
It is now time to try to summarize what has been discussed in this paper.
First of all, metaphors are an integral part of our conceptualization process. The
automatic and unconscious mapping of cross-domain associations builds from
atomic metaphorical parts into primary metaphors, and then more complex meta-
phors in conceptual blending. This blending of concepts in metaphorical mapping
can also extend to the iconic mapping of visual shapes. This iconic mapping can
occur as the result of a local blend, as, for example, for the sign for REMEM-
BER in Figure 5, that extends to the iconic mapping illustrated in Figure 9.
It is also possible to expand to more global metaphors, and more global
iconicity, such as that illustrated for written poetry in Hiragas Figure 4 p. 332,
and for signed poetry, in Figure 8. In this global metaphor/iconicity, the whole
text of the poem becomes a metaphoric/iconic blendof form andmeaning. And
Metaphorical Mapping
Generic Space
cognitive activity is in the head
hands can grasp, hold things
understanding, remembering
Input I
Input I
(abstract, subjective)
(concrete, sensorimotor)
grasping, holding
Iconic Mapping
Visual Shape
Figure 9.
nally, these metaphoric/iconic blends of the mental spaces become somewhat
physical andvisible in the signing spaces of signedpoetry . Although this may be
just a beginning stage for a cognitive approach to JSL research, I am convinced
that sign language research along the lines exemplied in this paper will be
important in the development of a comprehensive theory of metaphor and
iconicity, in which major modality eects shouldbe nonexistent.
I wouldlike to thank Olga Fischer, Masako Hiraga, Sumiko Saito andthe participants at the Third
Symposium on Iconicity in Language andLiterature for their valuable comments on earlier versions
of this paper. I wouldalso like to thank Sumiko Saito andother members of the Friday Sign
Language Salon for their comments and advice concerning the JSL data. I am, however, solely
responsible for any misinterpretations of this valuable advice, and any other errors that may have
occurred in the discussion of the data.
JSL is one of the languages indigenous to Japan (along with the Japanese language, the Ainu
language of Hokkaido, andthe Ryukyu language of Okinawa), andin one form or another is used by
most of the estimated half million deaf andhearing impairedin Japan. Though it is linguistically
distinct from the Japanese language, it has been greatly inuenced by the Japanese language andthe
dominant Japanese culture. One important reason for the strong inuence of the Japanese language
on JSL is the fact that most Japanese deaf andhearing-impairedlearn the national language,
Japanese, rst, in the oralist approach which dominates in Japanese schools for the deaf and hearing-
impaired, where JSL, although not usually ocially bannedany longer, is still rarely used andoften
resistedby (the mostly hearing) teachers. JSL, then, for most Japanese deaf andhearing impaired, is
usually a secondlanguage, acquiredafter a rst (often not successful) attempt to acquire the Japanese
language, and, apparently, many Japanese deaf and hearing-impaired do not consider themselves to
be a linguistic community united by JSL. Despite this reality, JSL is nonetheless an independent
language, andcan andshouldbe studiedby linguists as a rich source of linguistic data. (See, for
example, Kanazawa 2001, for discussions concerning JSL and deaf education in Japan.)
Also, as is conventional for transcriptions of the signs of sign language, all signs in the present
paper are written in capital letters with the nearest English equivalent of the meaning of the sign.
Armstrong, D., W. Stokoe, andW . Sherman (eds.). 1995. Gesture and the Nature of
Language. Cambridge: Cambridge University Press.
Fauconnier, G. 1985. Mental Spaces, Cambridge Mass.: MIT Press.
Fauconnier, G. 1997. Mappings in Thought and Language, Cambridge: Cambridge
University Press.
Fauconnier, G. 2001. Conceptual blending and analogy. In The Analogical Mind:
Perspectives from cognitive Science, D. Genter et al. (eds), 255285. Cambridge,
Mass.: MIT Press.
Fauconnier, G. andM. Turner. 1996. Blending as a central process of grammar. In
Conceptual Structure, Discourse and Language, A. Goldberg (ed.), 113130.
Stanford: CSLI.
Grady, J. 1997. Foundations of Meaning: Primary Metaphors and Primary Scenes.
Ph.D. Dissertation, University of California, Berkeley.
Herlofsky, W. 2001. Goodprobes: Icons, anaphors, andthe evolution of language. In
The Motivated Sign: Iconicity in language and literature 2, O. Fischer andM. Nnny
(eds), 5568. Amsterdam: Benjamins.
Hiraga, M. 1999a. Blending and interpretation of haiku: A cognitive approach. Poetics
Today 20: 128.
Hiraga, M. 1999b. Rough sea andthe Milky Way: Blending in a haiku text. In Computa-
tions for Metaphors, Analogy, and Agents, C. Nehaniv (ed.), 2736. Berlin: Springer.
Hiraga, M. 2000. The interplay of metaphor andiconicity in a haiku text: A cognitive
approach. Paper presented at PALA 2000, Goldsmiths, University of London.
Johnson, C. 1997. Metaphor vs. conation in the acquisition of polysemy: The case of
SEE. In Cultural, Typological and Psychological Issues in Cognitive Linguistics. M.
Hiraga, C. Sinha, and S. Wilcox (eds), 152171. Amsterdam: Benjamins.
Kanazawa, T. (ed.). 2001 Rookyooikuno Dakkoochiku (The Deconstruction of Deaf
Education). Tokyo: Akashi Shoten.
Keller, R. 1998. A Theory of Linguistic Signs. Oxford: Oxford University Press.
Klima, E. andU. Bellugi. 1979. The Signs of Language. Cambridge Mass.: Harvard
University Press.
Lako, G. andM. Turner. 1989. More Than Cool Reason: A Field Guide to Poetic
Metaphor. Chicago: University of Chicago Press.
Lako, G. andM. Johnson. 1999. Philosophy in the Flesh: The Embodied Mind and Its
Challenge to Western Thought. New York: Basic Books.
Mandel, M. 1977. Iconic devices in American Sign Language. In On the Other Hand:
New Perspectives on American Sign Language, L.A. Friedman (ed.), 57107. New
York: Academic Press.
Stokoe, W. 2000. Gesture to sign (language). In Language and Gesture, D. McNeill
(ed.), 388399. Cambrdige: Cambridge University Press.
Sutton-Spence, R. andB. Woll. 1999. The Linguistics of British Sign Language. Cam-
bridge: Cambridge University Press.
Sweetser, E. 1990. From Etymology to Pragmatics: Metaphorical and Cultural Aspect of
Semantic Structure. Cambridge: Cambridge University Press.
Taub, S. 1997. Language andthe Body: Iconicity andMetaphor in American Sign
Language. Ph.D. Dissertation, University of California, Berkeley.
Taub, S. 2001. Language in the Body: Iconicity and Conceptual Metaphor in America Sign
Language. Cambridge: Cambridge University Press.
Turner, M. andG. Fauconnier. 1995. Conceptual integration andformal expression,
Metaphor and Symbolic Activity 10.3: 183204.
Turner, M. 1998. Figure. In Figurative Language and Thought, A. Katz, C. Cacciari, R.
Gibbs, and M. Turner (eds), 4487. Oxford: Oxford University Press.
Volterra, V. andC. Erting, 1994. Introduction. In From Gesture to Language in Hearing
and Deaf Children, Volterra, V. andC. Erting (eds), 14. Washington D. C.:
Gallaudet University Press.
Wilcox, P. 2000. Metaphor in American Sign Language, Washington, D. C.: Gallaudet
University Press.
</TARGET "her">
AUTHOR "Axel Hbler"
TITLE "Spatial iconicity in two English verb classes"
WIDTH "150"
Spatial iconicity in two English verb classes
Axel Hbler
University of Jena
1. Aims
Of all English verb classes, two subgroups standout for being highly typical of
the English verb vocabulary and, at the same time, for having posed major
challenges to linguists. While the one has receivedquite a lot of attention, the
other has been largely neglected. The approach put forward here oers an
equally satisfying explanation of both of them. The verb classes I am referring
to are (1) the so-called redundant phrasal verbs and (2) the class that basically
consists of zero-derivations from adverbial particles; for lack of a better term I
will call them pure spatial verbs. Both are unitedby the particular prominence
they give to spatial adverbial particles. In the following, I will rst give a brief
denition and then discuss them as to their iconicity, more specically with
respect to gestures. In addition, I will justify their existence by relating them to
a specic socio-cultural phenomenon of the sixteenth andseventeenth centuries.
These analyses reect my intention to radicalize (on a trial basis) for these two
verb classes the cognitive approach with its central concept of embodiment.
2. Redundant phrasal verbs
2.1 Denition
A phrasal verb consists of a verb andan adverbial particle which together behave
as an integral unit. Its meaning is to a higher or lower degree idiomatized (cf. the
examples in block [1]). It is characterized as redundant when the phrasal verb is
propositionally equivalent to its simple counterpart (cf. examples in block [2]).
(1) put down put (2) swallow down = swallow
go o go mail o = mail
give up give connect up = connect
That does not amount, however, to claiming that these phrasal verbs are entirely
synonymous with their corresponding simple verbs. As Hampe (2002: 246247)
aptly points out, in terms of cognitive concepts,
redundancy in verb-particle constructions must be regarded as a form of
conceptual overlap [] Rather than inserting completely new conceptual
material, the particles prole material already (to some degree) activated [ ]
What exactly is being proled is the resultant state of the respective particles
trajectory [] The proling achievedby the particles can therefore [] be
dened as end-point-proling. [] The change in the construal of a scene
achievedby proling is a change in the relative salience of the subparts of a
conceptualization [].
Adverbial particles in redundant phrasal verbs have spatial meanings. When their
proling eect does not activate the spatial dimension inherent in the verb itself,
it can be interpretedin a metaphorical way. For example, the out in stretch out
proles the extension andgoal state of the movement itself, in yell out it results
from the metaphor : soccroiciN is : coN1:iNr; lend out rests on the
conceptual metaphor iossrssioN is coN1:iNxrN1 and proles the resultant state
of its trajector in the act of giving away. The up in lift up proles the trajectors
upwardmovement and nal position, whereas in phrasal verbs such as shut up
or fry up it proles the goal state, i.e., the completion of the action, andis
motivatedby the conceptual metaphor coxiir1ioN is ci (cf. Hampe
2002: 181248).
Going beyondthis explanation, I want to suggest that, when the adverbial
particles (such as up or out) do not aect the corresponding propositional
content of the verb, they represent the common spatial denominator of gestures
that in speech couldco-occur with these verbs. The idea as such is certainly not
entirely new (cf. Lindner 1983: 231), but it has never been more than an
interesting aside andnever been tackledseriously .
2.2 Aspects of iconicity
As I am applying it, the term iconicity refers to similarities that holdbetween
verbal expressions andgestures. Assuming that communication consists of three
modes the verbal, the prosodic and the gestural/kinesic mode (cf. Arndt
andJanney 1987), the iconic relation may be specied as cross-modal. In
redundant phrasal verbs, it can be identied at three levels. Firstly, we nd
iconicity on the compositional level. Here, we understand the interaction between
verb andparticle (in phrasal verbs) as being similar to the interaction recogniz-
able in the co-occurrence of a simple verb anda gesture. According to McNeill
(1992), gestures take an integral part in the development of ideas, they are not
simply a performative embellishment. Where speech is spontaneous andthe
speaker (naturally) involved, a gesture occurs together with that linguistic
element of an idea unit that forms its center. It usually coincides with the nucleus
of the corresponding tone unit. McNeill interprets it as manifestation of what he
takes to be the growth point of the idea unit. It is a neuro-cognitive construct
andis dened as follows:
The concept of a growth point unites image, word, and pragmatic content into
a single unit. In this way it is a unit with the properties of the whole []. It
is also a unit that encompasses properties of an opposite character both
imagistic andlinguistic, idiosyncratic and social, global andsegmented , holistic
andanalytic [] (McNeill 1992: 220).
Both parts separate subsequently, during the productive cycle,
andreunite at the
moment of utterance, manifesting themselves bi-modally, i.e., as gesture and
language. Due to its spatial meaning, the adverbial particle of the redundant
phrasal verb recalls andpreserves some of the imagistic andglobal potentials of
the gesture. As such it may thus be saidto have a similar share (together with
the abstract andparticularistic verbal constituent) in the realization of the growth
point of the idea unit as the gesture has.
Secondly, we nd iconicity between the particles of redundant phrasal verbs
andgestures on the operational level. According to Lako andJohnson
(1999: 580), there is no fundamental dierence in neural activity between the
characterization of a motor concept (the meaning of a motion verb, for example)
andthe characterization of a motor schema, which organizes and controls the
corresponding movement. Such a stance is compatible with the view that spatial
concepts, no matter whether they are expressedverbally (as adverbial particles)
or gesturally, are global andimagistic in nature andare linkedto the same
hemisphere of the brain, i.e., with right-handers normally to the right hemisphere.
Thirdly, redundant phrasal verbs show an iconic relation to gestures on the
functional level. Just as gestures co-occur with verbal (propositional) concepts,
of which they highlight only certain aspects, adverbial particles highlight some
aspects of the co-occurring verbs. Thus, the gesture depicted in illustration (a),
for example, highlights, when accompanying the word box, its three-dimensional-
ity andsquareness. In the same way, the particle away, when interacting with the
verb hide, highlights a (perceptual) eect of the act.
Furthermore, like gestures, the particles may be saidat times to take the
lead. The gesture described in example (3) anticipates the utterance of the
corresponding linguistic concept, in (4) it does not nd a verbal correspondence
at all, in this case due to Broca-aphasia.
(3) this gives [a complete duality]
The right hand presents the idea of a dual by looping upward
(McNeill 1992: 166)
(4) Experimenter: What was the cat doing?
ah ah []
Left hand with index nger pointing up, rises up over head and then
straight down to lap the cat plunging to earth
(McNeill 1992: 338)
Correspondingly we may say following Bolinger (1971) that with the
redundant phrasal verb in
(5) a. They burnt down the house
it is the particle down that ultimately carries the central meaning, while the verb
burn merely species it, as the paraphrase given in (5b) illustrates:
(5) b. They downed the house by burning
The existence of pure spatial verbs, the secondclass to be discussedlater in this
paper, provides support for this view.
2.2.1 A case study
In order to further substantiate my approach, I will have a closer look at the
particle up in redundant phrasal verbs. For reasons mentioned earlier and to be
explainedlater , I will restrict myself to examples from the sixteenth and
seventeenth centuries.
What cognitive linguists usually do is describe the proling eect of the
adverbial particle in spatial terms, literally or metaphorically; I have given a brief
illustration of such an approach in Section 2.1. In the approach I am advocating,
the spatiality of the adverbial particle refers to gestures; it, consequently, does
not prole parts of the content of the main verb, but highlights a certain aspect
of it just as gesture (a) highlights the squareness of the concept box. In line
with this literally body-centered approach, the adverbial particle up in redundant
phrasal verbs is considered a verbal correspondent to a non-verbal, gestural
element, andthus preserves some of the gestures global, holistic, andimagistic
root. Thus all verbs that form a redundant phrasal verb together with up as well
as these redundant phrasal verbs themselves combine typically with gestures for
which up is a characteristic feature.
The evidence I will present for supporting the claimis certainly not very strong,
but may suce for a pilot study. I have collected the gestures that a few informants
and myself spontaneously produced in connection with a number of redundant
The gestures elicitedcan be lumpedtogether into ve groups.
Characteristic for the rst group is a vertical movement of one handeither
straight upward, or circumscribing a curve (cf. illustrations [b] and [c]).
b. c.
The rst variant highlights a part of the event itself (upwardmovement), the
seconda characteristic of the eect (having an upwardd imension). The rst
variant goes with verbs such as rise (up), lift (up), pile (up); the secondvariant
with store (up), clutter (up), ll (up) andthe like. As the following two examples
show, it does not make any dierence whether the meaning of the verb is literal
or gurative.
(6) He rose up on his fete quyckly (1530, Palsgrave)
(7) Andif my hope sometimes ryse vp, by some redresse: It stumbleth
straite, for feble faint (1542, Wyatt)
The gestures, which form a secondgroup, go with verbs like wake (up), call (up),
or summon (up). Characteristic is an upwardmovement of one hand(as in
illustration [b]) or simply a movement of the four ngers of the open hand, palm
turnedup (cf. illustration [d]).
The gesture conveys andhighlights the eect of the action referredto, i.e., being
present in person andmind . Example (8) illustrates the literal use, (9) a gura-
tive one.
(8) That thy power Might call up him who left untoldThe story of
Cambuscan bold(1632, Milton Penser.)
(9) Now Madam summon vp your dearest spirits (1588, Shakespeare
L. L. L. ii.i.)
The gestures belonging to the thirdgroup are executedwith two hands. They
consist of an opening movement with the palms turnedup as in illustration (e).
This group has two readings, i.e., creating space and disintegrating. The rst
reading would apply to verbs like clear (up) and light (up), the secondread ing to
break (up) and tear (up). As the following examples illustrate, the space involved
can be concrete or abstract, gurative:
(10) What brightnesse is this I see? Have you light up any candles? (1649,
Roberts Clavis Bibl.)
(11) A savoury dish, a homely treat, Where all is plain, where all is neat,
Clear up the cloudy foreheads of the great. (1700, Dryden)
The gestures of group four show a reverse movement. One or both hands,
preferably spread, go, with the palms facing down, downwards, as if put on a
(at or slightly rounded) surface. It is from the perspective of the entity below
the hands that the movement can be conceptualized as up. The putting of the
hand(s), on a at or slightly rounded surface is articulated as a short or extended
movement (cf. illustrations [f] and[g]).
In the former case, the gesture collocates with verbs like shut (up), lock (up),
block (up) andhighlights the forced nature of the action concept. In the latter, it
occurs with verbs like heal (up), freeze (up), seal (up), highlighting the non-forced
nature. Gesture andparticle interacting with the verb are open to literal as well
as gurative use, as is illustratedby the following example.
(12) He hath shytte up his treasour in a wall (1530, Palsgrave)
(13) Though all the Conduits of my blood [be] froze vp (1590, Shake-
speare Com. Err.v.i.)
The fth (and nal) group to pay attention to is certainly somewhat precarious,
mainly because it may appear to be a collective group for all the gestures that do
not t into any of the preceding groups. They have, however, one feature in
common that not only unites them as a group but also distinguishes them from
the gestures tackledso far. The gestures of this last group do not incorporate a
spatial element for which the adverbial particle up wouldbe an obvious and
immediate verbal correlate. They fall into two groups. Either they accompany
phrasal verbs like chop (up), mash (up), and polish (up), where they depict some
aspect of the single verbal concept, as in illustrations (h)(k).
h. i.
The following example shows that the meaning of the verb andthe correspond-
ing gesture can also be gurative.
(14) With your ngers you handle the reall, corporall, substanciall, identi-
call presence of Christ, beholdthe same with your eyes, andchoppe
him uppe at a morsell (1581, J. Bell Haddons Answ. Osor)
Or the gestures simply consist in beats made out of the wrist (cf. illustration [l]).
According to McNeill (1992: 15), a beat is typically
a simple ick of the handor ngers up andd own or back andforth; the
movement is short andquick [] The semiotic value of a beat lies in the fact
that it indexes the word or phrase it accompanies as being signicant, not for
its own semantic content, but for its discourse-pragmatic content.
Such beats may accompany verbs like nish (up), fry (up), or sweeten (up).
(15) Goe, sweeten up thy labours andthy life With fresh delights (1644,
Quarles Sol. Recant)
In all the cases of group ve, the (redundant) particle up is not motivatedby the
specic spatial dimension of upness inherent in the gestures, but by a more
general feature, namely handactive, and, subsequently, the application of the
widely profused metaphorical concept :c1ivr is ci. It is therefore not by chance
that up seems to be the most productive particle among redundant phrasal verbs.
A nal wordof precaution. We might be temptedto assume that redundant
phrasal verbs going with the same or a similar gesture wouldform a special class
and, consequently, to compare the resultant classication with those found in the
cognitive literature, in Lindner (1983) or Hampe (2002), for example. This would
lead us astray, though. All I have done was demonstrate that gestures which are
naturally co-expressive with up-phrasals typically show some characteristic that
motivates up as an adequate adverbial particle. This undertaking, however, is
embedded in a line of reasoning that deserves particular attention because it is
fundamentally dierent from current treatments. The central point relates to the
signicance of up (or some other redundant particle). In my view, redundant
particles such as up do not exert an impact on the meaning potential of the corre-
sponding verbs in the rst place, but function rather as a complement to it: They
add to the propositional meaning of these verbs their non-propositional counter-
part. That is the global, holistic, andimagistic share in speaking and thinking,
which normally manifests itself in gesture.
3. Pure spatial verbs
3.1 Denition
Remarkably enough, for all spatial adverbs which can occur as part of redundant
phrasal verbs we nd instances of their verbal or verb-like use. In other words,
we have verbs such as to up, to out, to down, to through; andwe have in
imperative contexts elliptic uses of the spatial adverbs, i.e., uses in which the
(imperative) verb itself does not show up, whereby the adverb acquires a verb-
like appearance. The semantic specication is provided by the context; the OED,
therefore, can list for most of them several readings, extending from the physical
to the mental. Again, up can serve as a goodillustration in point. Examples (16)
(18) illustrate the verbal conversion of these spatial particles, (19) and(20)
their elliptic use.
(16) The Swan-heard shall vp no Swan nor make any sale of them,
without the Maister of the Swannes be present (= to drive up and
catch [swans, etc.] so as to provide with the mark of ownership)
(15841585 Order for Swans)
(17) An Animal together blowdandmad e, Andupdof all the shreds of
every Trade (= to make up, form, or compose of something) (1685,
Cleveland London Lady 102)
(18) The true-bredGamester ups a fresh, andthen, Falls tot agen (= to
rise to ones feet) (1643, Quarles Embl. ii.xiv)
(19) Debbora sayde vnto Barak: Vp, this is the daie wherin [etc.] (= com-
mandor exhortation to action, activity, rising from bed, movement,
etc.) (1535, Coverdale Judges iv)
(20) Vp with my Tent, heere wil I lye to night! (= denoting erecting, rais-
ing) (1594, Shakespeare Richard III, v.iii)
This is only a selection from the meanings to be foundin the sixteenth and
seventeenth centuries, some of which go back to earlier times. Other meanings
pop up in later times.
3.2 Iconicity
It is again my claim that these spatial verbal expressions relate to gestures and
that the relationship can be described as iconic. They show iconicity at the
operational level which can be described in similar terms as above. But we also
nd a functional type. It may be more obvious with the elliptic use of such
particles in imperative contexts (cf. examples [19] and[20]) than with other cases.
A gesture consisting of an upwardmovement of the hand usually is co-expressive,
i.e., is accompanying a verb. It may, however, acquire the status of a language-
independent gesture with a xed meaning (Ekman and Friesen 1969 call such
gestures emblems). Using the particle up elliptically with its imperative reading
is similar to performing the (corresponding) gesture in its emblematic function.
Where up does not adopt an imperative force, the iconic relationship to the
gesture is less easy to establish. We have to distinguish between the (action)
concept expressedas up andits characterization in temporal andpersonal terms.
As far as the action concept is concerned, we could argue along the following
line: where in the production process both gesture and word surface (which is
usually the case), the gesture often precedes the corresponding word by some
milliseconds (cf. McNeill 1992). If the word (for the denite concept) is lacking,
the adequate gesture may come up nonetheless; this holds true for pathological
as well as normal circumstances. Thus the gesture, consisting in an upward
movement, could realize the central idea of upwardness alone.
(21) but then, of course, his spirits
This gesture, in turn, couldbe thought to become subject to verbalization andthe
temporal andpersonal specication in forms like ups (18) or upd (17) couldbe
considered a post hoc strategy for adjusting the verbalized gesture to the gram-
matical requirements of a clause. The result is a hybridandas such reects the
mixedverbal non-verbal constituency, typical for the expression of a units
central idea in spontaneous speech.
4. Signicance
Detecting a gesture-relatediconicity in the two verb groups tackled wouldbe a
rather vain undertaking, if it were not for some reason of higher order. What I
wouldlike to oer is an argument that rests upon a combination of cultural-
historical andneuro-cognitive considerations. The illustrations given for both
groups of verbs date back as will be remembered to the sixteenth and
seventeenth centuries. It is indeedthe periodwhich, diachronically speaking, has
a dominant share in these formations. In fact, almost half of the phrasal verbs
that Hampe (2002) discusses as potentially redundant can be said to originate in
these two centuries, if we take the rst quotations in the OED as indicative; a
few are older, but the rest are of more recent origin. An even more denite picture
emerges for the pure spatial verbs, if we take the seventeenth century as the dividing
The same holds true for the elliptical use of these adverbial particles.
The sixteenth andseventeenth century is the time for which there is rich
evidence that the English (courtly) society took great eorts in subjecting the
body to a rigorous control as part of what Elias (1939) calls the civilizing
process. This control also extended to the use of gestures in conversation, as
emerges clearly from the study of the courtesy books of the time (cf. Hbler
2000 and2001: 176202). While these books hardly ever oer descriptions of
appropriate kinesic behaviour in positive terms, they provide negative models, as
illustratedby the following quotation.
Others there are, who can never enter into any set or serious Discourse, but
they must play with a button [] but this mimicke andapish action keepes
small concurrence with the Postures of a Gentleman, whose Speech as it
shouldbe free, native andgenerous; so should the action of his bodie admit of
no phantasticke imitation or servile aectation, which expresseth little, save a
degenerate qualitie or disposition (Braithwait 1630/1994: 87).
Or they explicitly formulate bans on certain forms of kinesic behavior; a most
glaring case is Hawkins (1646/1980: 34, 39), from whom the following extracts
are taken:
15. Thou oughtst not to make a face, or use any other action of undecency
with thy mouth, eyes, or thy hands, to express what thou wouldest deliver;
neither oughtst thou to holdthy handbehindthy back, either clasped or across
[] When thou talkest, be circumspect how thou carriest thy body. Shake not
thy head, nor move thy hands much, and hold thy feet still.
34. If any one hadbegun to rehearse a History [] and if he relate it not a-right,
andfully; shake not thy head, twinkle not thine eyes, andsnigger not thereat.
In view of such a situation, we may well draw the conclusion that the develop-
ment of redundant phrasal verbs as well as of pure spatial verbs served the
purpose of substituting gestural behaviour. Findings from neuro-cognitive
research, according to which gestures form an integral part of speech and thought
production (cf. McNeill 1992) and thus call for substitution in case of inhibition
(cf. Rim andSchiaratura 1991), warrant such an assumption.
Both verb classes were not ideal substitutes, though, mainly for two reasons:
(a) they can do their business only where action concepts form the centre of an
idea unit and(b) their application is subjectedto certain pragmatic constraints. In
other words, they are not iconic in respect to one feature standing out for
gestures, this is their ubiquity, i.e., their potential to occur repeatedly/with very
high frequency. The English courtly society, therefore, hadto develop a better
substitute eventually. This was found, I suppose, in their prosody. But thats
another story
1. In Hbler (1998a), I have pursueda similar argument for idioms such as to raise ones eyebrows
or to tap ones forehead, i.e. idioms that verbalize gestures.
2. Cf. McNeills (1992: 236) self-organization model.
3. There is one drawback, however; while the gesture usually anticipates slightly the correspond-
ing verbal item andnever follows it, the particle always follows the verb. It is counterbalanced,
however, by the fact that, as far as intonation is concerned, the particle usually attracts the main
4. The gesture (described in italics) is co-occurrent with the spoken text set in square brackets.
5. The question that I askedmyself and others was simply What gesture do you think wouldgo
well with the following expression x.
6. Of the redundant particles nowadays in use, only o must be considered a post-seventeenth
century verb; the verbal use of through, up, and down originates during the period under
consideration; for out and in, nally, the verbal use is testied in Middle English times.
7. Except for in, whose elliptical use is evidenced for the 18th century, all others date back to
earlier times, through to the 16th century, the others still earlier.
Arndt, H. and R. W. Janney 1987. InterGrammar. Toward an Integrative Model of Verbal,
Prosodic and Kinesic Choices in Speech. Berlin, New York: Mouton de Gruyter.
Bolinger, D. 1971. The Phrasal Verb in English. Cambridge, Mass.: Harvard University
Elias, N. 1939. ber den Prozess der Zivilisation. Frankfurt: Suhrkamp [1997].
Ekman, P. andW . V. Friesen 1969. The repertoire of nonverbal behavior: Categories,
origins, usage, andcod ing. Semiotica 1: 4998.
Hampe, B. 2002. Superlative Verbs. A Corpus-Based Study of Semantic Redundancy in
English Verb-Particle Constructions.Tbingen: Narr.
Hbler, A. 2001. Das Konzept Krper in den Sprach- und Kommunikationswissen-
schaften. Tbingen, Basel: Francke.
Hbler, A. 2000. From body language to embodied language: Changes in the expressive-
ness of conversation during the 16th and 17th centuries. In English Diachronic
Pragmatics, G. di Martino and M. Lima (eds), 5778. Napels: CUEN.
Hbler, A. 1998. Worte statt Gesten? Zu Mglichkeiten undGrenzen der Verbalisierung
des Nonverbalen am Beispiel des Englischen. Anglistik 9: 6579.
Hbler, A. 1998a. The Expressivity of Grammar. Grammatical Devices Expressing Emotion
across Time. Berlin, New York: Mouton.
Lako, G. andM. Johnson 1999. Philosophy in the Flesh: The Embodied Mind and its
Challenge to Western Thought. New York: Basic Books.
Lindner, S. J. 1983. The Lexico-Semantic Analysis of English Verb-Particle Constructions.
Trier: L. A. U. T. (Series A, No. 101)
McNeill, D. 1992. Hand and Mind. What Gestures Reveal about Thought. Chicago,
London: The University of Chicago Press.
Rim, B. andL. Schiaratura 1991. Gesture and speech. In Fundamentals of Nonverbal
Behavior, R. S. Feldman, and B. Rim (eds), 239281. Cambridge: Cambridge
University Press.
</TARGET "hub">
AUTHOR "Keiko Masuda"
TITLE "What imitates birdcalls?"
WIDTH "150"
What imitates birdcalls?
Two experiments on birdcalls
andtheir linguistic representations
Keiko Masuda
University of Cambridge
1. Introduction
Many linguists have insistedthat the associations of meanings and sounds
dened by the words of a language are arbitrary (Sapir 1929). However, some
non-arbitrary relations between certain sounds and meanings or senses have been
claimedas far back as Socrates. This type of non-arbitrary relation between the
soundof a wordandits meaning is calledsoundsymbolism.
One of the most famous forms of soundsymbolism is onomatopoeia: the
coining or use of a wordwhich attempts to represent a non-linguistic soundby
a combination of appropriate segments selectedfrom the ordinary phoneme
inventory of the language (Trask 1996: 247). Onomatopoeia lies partially outside
the phonological inventory of the language. To some extent, language-specic
constraints are relaxed, such as the number and combination of consonants
allowedin a cluster, and the distribution of a given phoneme. The best-known
example may be animal sounds such as bow-wow and tweet-tweet. Some linguists
have pointedout that the secondformant pattern (F2 henceforth) of a linguistic
representation of a birdcall or even some bird names mimics the birdcall (e.g.
Rhodes 1984: 279). Bladon (1977) attempted a cross-linguistic study of onomato-
poeia and an acoustic analysis of spectrograms of a few sounds including animal
sounds and external noises. Although his study was original and novel, the
analysis was very limited.
This paper attempts to examine the claim that F2 mimics the birdcalls from
an acoustic point of view, which has not been extensively explored. It seeks
correlation between birdcalls andtheir linguistic representations andthe eect of
phonological or phonotactic constraints in forming the representations. The results
of an acoustic analysis will be presented, followed by the results of a perception
experiment which tests the hypothesis drawn from the acoustic analysis.
2. Experiment 1
This experiment investigates the correlation between acoustic properties of
birdcalls and their linguistic representations. Throughout this paper, a linguistic
representation refers to a phonetic realisation of a birdcall by human voices,
which can be expressedin the orthography of the language (English, in this case)
as foundin the voice section in books on birds (e.g. a reference book for
birdwatchers). For instance, a linguistic representation of a Blackbird call is
tsink-tsink-tsink. Such linguistic representations will be comparedacoustically
with the original birdcalls to nd out the correlation between them.
2.1 Method
2.1.1 Experimental representations
Ten English birdnames of onomatopoeic origin were selected for the purpose of
the experiment according to the availability of the source of their calls (Table 1).
Due to a shortage of candidate names for the experiment, dialectal bird names of
onomatopoeic origin were also included (Gooders 1982, Heinzel et al. 1990,
Hume 1990, etc.). Clear tokens of their calls were digitised at a sampling rate of
16 kHz from a pair of CDs (Sample 1996) into a Silicon Graphics Unix computer
in the Phonetics Laboratory, University of Cambridge.
Linguistic representations of the birdcalls can dier slightly from reference
to reference. In such cases, the representation that is listedmost frequently in
references was adopted. When more than one representation is listed in an equal
number of references, all of the representations were adopted. When one bird has
several distinctive calls (e.g. breeding call, ying call, etc.), the one that clearly
is the origin of the name was chosen. It may be possible to pronounce one
linguistic representation in several ways. A simple survey asking ve near-RP
speakers of English how they pronounce the linguistic representations was
carried out to avoid possible individual dierences in pronunciation. The major
pronunciations were adopted, and in some cases, more than one pronunciation
was usedas a representation. A phonemic transcription was put next to each
representation to show the speakers who took part in the experiment how to
pronounce it to avoid individual dierences in pronunciation.
2.1.2 Speakers
The speakers were four near-RP speakers of British English in their twenties
(two males and two females), who were all Cambridge University students at the
time of recording.
Table 1. Birds used in the experiment and their linguistic representations together with the
possible pronunciations
Bird Name Linguistic representation(s) Pronunciation(s)
Chicha chi-cha chi-cha /twIf twf twIf twf/
Coot teuk
Crake (Corncrake) crex-crex
/kr7ks kr7ks/
[kp kp ]
[kr kr ]
Cuckoo cuc-koo /k~ku/
Curlew cur-lee
/k8 li/
/kf li/
Keelie (Kestrel) kee-kee-kee /ki ki ki/
Kittiwake kitti-wark /kItI w"k/
Pewit (Lapwing) pee-wit /pi wIt/
Twite chweet /twwit/
Whew (Wigeon) whee-oo /wi u/
/&i u/ (/hwi u/)
2.1.3 Recording
Sixteen pronunciations of thirteen linguistic representations for ten bird names
were put in a random order in a list, together with some llers inserted in the
beginning and at the end of the list to avoid the marked reading which tends to
occur in the beginning or at the end of the whole production. The speakers were
asked to read the complete list of representations twice in citation form with the
help of phonemic transcriptions. All the productions were recorded on a DAT tape
in the sound-treated booth in the Phonetics Laboratory. The recorded data were then
digitised at a sampling rate of 16 kHz into a Silicon Graphics Unix computer.
2.1.4 Analysis
The digitised data of both birdcalls and their recorded linguistic representations were
analysed in terms of formant frequencies, fundamental frequency, duration, and
their overall dynamic patterns in proportional timing. Proportional timing means
duration expressed as a proportion of the whole call (e.g. 25%). The choice of
phonemes used to imitate particular parts of the components was also considered.
As an analysis of the recordings of the four speakers revealed that there was
no major individual dierences among the speakers, one production of one
speaker, whose productions were the clearest of all, was used for a detailed
analysis andwill be presentedin this paper.
Some birdcalls have clear harmonic-like concentrations of energy at certain
frequency levels (Keelie, Kittiwake, Pewit, Twite andWhew) showing up as dark
bars on spectrograms. In this case, these dark bars are referred to as compo-
nents, the darkest of which was taken for analysis if they dieredin density.
Otherwise the component around2,000 Hz was usedto compare with linguistic
representations, as this frequency level is around the middle point of the range of the
perceptually most important four formants of vowels of human speech. Other
calls have a single pure-tone-like concentration of energy (Chicha, Cuckoo
andCurlew), which will be called the component. The Coot andthe Crake, on
the other hand, have a more scattered pattern rather like a fricative, in which
energy is nevertheless concentratedrather more at some frequencies than others.
2.2 Results
In this paper only two cases, the Whew andthe Coot, will be presentedd ue to
space limitations.
2.2.1 Whew
The call of the Whew is expressedas whee-oo, /wi u/ or /hwi u/ in most
references. The spectrogram pattern of the birdcall looks like two harmonics
rising steeply towardthe peak point andthen falling gradually (Figure 1).
The spectrogram of the linguistic representation whee-oo /wi u/ is shown
in Figure 2. Figure 3 shows both the main component of the Whew call andthe
four formants of the linguistic representation in proportional timing.
It may look as if F2 intersects F3 during the transitions /w/ /i/ and/i /
/u/, although in the latter case the intersection is not very straightforward. It
seems that the entity representedby F2 of /w/ rises, crossing the next higher
resonance, to become itself F3 of /i/, andthen falling again. The most striking
observation of these two spectrograms may be the similarity of the trajectory of
this resonance andthat of the components of the birdcall. Formant frequencies
4.4 4.5 4.6 4.7 4.8 4.9 5.0 5.1 5.2
Frequency (Hz)
Time (ms)
Figure 1. Spectrogram of a Whew call.
18.8 18.9 19 19.1 19.2 19.3 19.4 19.5 19.6 19.7
Frequency (Hz)
Time (ms)
Figure 2. Spectrogram of a linguistic representation of a Whew call, whee-oo /wi u/.
may be determinedby the size of front andback cavities (Fant 1960). It is
normally F2 that is associatedwith the front cavity. For a high front vowel,
however, it is F3 that is associatedwith the front cavity. The Front Cavity
Resonance (FCR) changes smoothly andcontinuously , andtherefore it is more
appropriate to regardthe FCR (rather than a sequentially numberedformant) as
responsible for imitating the birdcall.
It is assumedthat /wi/ may be usedto imitate the steep rise of the harmon-
ics andthe transition of /i u/ for the gradual fall because of their FCR levels.
2.2.2 Coot
0% 25% 50% 75%
Time (proportional)
Component of a Whew call
Figure 3. Dynamics of the component of a Whew call and F1, FCR, Back Cavity Resonance
(BCR), and F4 of the linguistic representation, whee-oo in proportional timing.
A Coot call is very sharp, high-pitched, short, and about 100 ms in duration. It
begins very abruptly and its components slightly decline towardthe end(Figure 4).
The linguistic representations vary quite considerably from reference to
reference, but some common ones are teuk (/tjuk/) and kowk (/ka~k/ or /k6~k/).
It is observedin all representations that the FCR (F2 itself in this case) seems to
track the trajectory of the component(s) of the call. The transition from the
higher FCR of /j/ of teuk /tjuk/, /a/ and/ 6/ of kowk /ka~k/ and/k 6~k/ to the
lower FCR of /u/ of /tjuk/, /~/ of kowk /ka~k/ and/k 6~k/ may indicate the
declining signal of the call.
Figure 5 shows the main component of a Coot call andthe four formants of
one of the representations, kowk /ka~k/. The grey broken line indicates the
energy concentration of /k/. The initial voiceless plosives /t/ and/k/ convey the
very abrupt onset of the call very well. In this sense, the syllable onset couldbe
/p/ or a voicedplosive, or even an aricate. However, the concentration of the
energy of /p/ is lower than that of /t/ and/k/, and therefore a labial may not be
ideal in this case as it would not track the high frequency onset of the call. The
voicedplosives are probably avoidedbecause the call is very high-pitchedand
the formant transitions wouldbe more audible.
The voiceless plosive may serve best as a word-nal sound. Because of the
4.8 4.85 4.9 4.95 5 5.05 5.1 5.15 5.2 5.25 5.3 5.35 5.4 5.45
Frequency (Hz)
Time (ms)
Figure 4. Spectrogram of a Coot call.
so-calledpre-fortis clipping eect, a vowel before a voiceless plosive is shorter
than the one before a voicedplosive, and therefore a voiceless plosive is the best
candidate as a word-nal sound to express the very short call. Although there is
no element in the call corresponding to the burst of the plosive, this may be
toleratedbecause word-nal plosives are often not fully releasedor they are
often even no more than glottal stops for many English speakers.
2.3 Summary
From the results of the experiment, the following observations were made. First,
there is a strong tendency that it is the front cavity resonance (FCR) of the
linguistic representations, rather than F2, that tracks a component of birdcalls.
This was clearly observedin representations where there is a soundsequence that
involves a rapidmovement in the oral cavity from the high back vowel to the
high front vowel. In such a sequence, there is a change in cavity aliation,
which makes F2 andF3 look as if they intersect each other. As their continuity
is interrupted, unlike the dynamic pattern of the birdcall component, it may be
more appropriate to consider the continually changing FCR as a responsible
factor. Segmentally speaking, the selection of vowels for a representation seems
determined by the frequency of their FCR, except in some cases where other
factors such as phonological or phonotactic ones seem to have stronger inuenc-
es. In the selection of consonants, the majority of the consonants usedin the
representations are voiceless plosives, the voiceless velar plosive /k/ in particular.
It may be because the location of the energy concentration of the voiceless velar
plosive matches itself to the secondor thirdformant of the adjacent vowels,
which may least aect the auditory impression.
3. Experiment 2
0% 25% 50% 75%
Time (proportional)
Component of a Coot call
Figure 5. Dynamics of the component of a Coot call and F1, FCR, BCR, and F4 of the
linguistic representation, kowk (/ka~k/) in proportional timing.
The secondexperiment tests if the observations made in Experiment 1 can be
supportedfrom a perceptual point of view. In this experiment, an occurring
(received) linguistic representation used in the rst experiment is presented,
with a representation which violates the hypothesis, for subjects to choose which
better imitates the birdcall. That is, this experiment tests whether linguistic
representations with the same FCR dynamic pattern as the main component
dynamic pattern of the birdcalls are preferred by subjects, and whether the
selection of consonants for linguistic representations of birdcalls conforms to the
predictions above.
3.1 Method
3.1.1 Experimental materials
Five birdcalls andone linguistic representation for each were selectedas
experimental materials: Coot (kowk /ka~k/), Curlew (cur-lee /k8 li/), Keelie
(Kestrel, kee-kee-kee /ki ki ki/), Twite (chweet /twwit/), andWhew (whee-oo /wi
u/). These were selectedfor various reasons including the simplicity of their
forms andtheir relative unfamiliarity to people.
For each birdcall, the received linguistic representation mentioned above
was taken as the expected one on the assumption that its conventional use is
well motivated. For each expected representation, deviant representations were
createdwhich deviate from the expectedrepresentations in ways which test
predictions of the FCR and consonant models. Of the birdcalls mentioned above,
some birdcalls have more than one linguistic representation and some linguistic
representations can be pronouncedin more than one way. In spite of the
possibilities of multiple receivedlinguistic representations, a single expected
representation was chosen for each birdcall for the sake of simplicity of the
experiment. Instead, the excluded receivedrepresentations were usedto compare
with the selected ones to see more detailed preference by subjects. For instance,
both kowk /ka~k/ and kowk /k6~k/ couldin principle be the expectedrepresenta-
tion, both of them having the same FCR dynamic pattern (falling) and diering
only in the absolute frequency in the beginning of the FCR dynamics. It is
hypothesisedfrom the rst experiment that the FCR trajectory is the essential
factor in forming a linguistic representation. This hypothesis does not, however,
dene the ner details of the FCR trajectory, e.g. the absolute frequency of the
FCR. Therefore, it wouldbe interesting to see which linguistic representation
wouldbe chosen between these two with the same FCR dynamic pattern and
only diering in the absolute frequency. The excluded receivedrepresentations
will be referredto below for convenience as deviant counterparts of the expected
ones (i.e. selected receivedrepresentations).
For each expectedrepresentation, twelve deviant representations were
created: four diering in the FCR dynamic pattern, four diering in the initial
consonant, andfour diering in the secondor word-nal consonant.
Of the four deviant linguistic representations in the FCR dynamic pattern
group, one hadan invertedpattern (i.e. if the expectedone had a rising FCR
dynamic pattern, the deviant one had a falling pattern), two had a level pattern,
each at dierent frequency levels, andthe fourth had the same pattern with
dierent absolute frequency. Taking the Coot for example, the expectedrepre-
sentation kowk /ka~k/ has a falling FCR pattern. One of the four deviant
representations is /kaIk/, which has a rising pattern, /kik/ and/k uk/ have a level
pattern with dierent frequency level, and/k 6~k/, as discussed above, has the
same pattern but with dierent absolute frequency.
In creating the deviant linguistic representations diering in the initial
consonant, dierent procedures were used depending on the initial consonant of
the expectedrepresentations. If the expected representation begins with a
voiceless velar plosive, as in kowk /ka~k/, two voiceless plosives of a dierent
place of articulation (/p/ and/t/), andthe voiced counterpart of /k/, the voiced
velar plosive /g/ were selectedas an initial consonant of the deviant representa-
tions. A representation without any initial consonant was also created. This
selection procedure is based on the hypothesis that the voiceless velar plosive is
dominantly preferred. Therefore voiceless plosives of a dierent place of
articulation were selectedto test if or how much the location of the energy
concentration matters. Similarly, /g/ anda null consonant were selected to test
whether the disturbance to the auditory transition is important. It may be argued
that other consonants shouldalso be tested andthat the hypothesis cannot be
otherwise supportedin a strict sense. This is not done here mainly because it is
presumedthat other consonants, such as fricatives or nasals, are unlikely to be
preferredas a consonant to express an abrupt onset of these three birdcalls,
which can normally be best representedby a plosive.
In the case of whee-oo, four consonants were selectedfor deviant represen-
tations. Although /w/ of /wi u/ is normally considered as a consonant in English
phonology, it is treatedwith the vowels here since it clearly contributes to
forming the FCR dynamic pattern. It therefore seems appropriate to test whether
a null obstruent is preferredto an obstruent existing before the vowels. Together
with the voiceless labial-velar fricative /&/ of the other expectedrepresentation
/&i u/ (or /hwi u/), three voiceless consonants /k/, /t/, and/s/ were selected
because they are voiceless and less disturbing to the overall auditory impression.
Besides, of all the voiceless consonants in English, only these three consonants
are able to form phonotactically correct sequences when followedby /wi/
(Gimson 1980: 242243).
Creating deviant linguistic representations diering in the second/word-nal
consonant followed the same procedure as with dierent initial consonants. For
the expectedrepresentations with a second/word-nal consonant such as kowk
/ka~k/, consonants with a dierent place andmanner of articulation anda null
consonant were selected. For the expected representation without a second/word-
nal consonant such as whee-oo /wi u/, three voiceless plosives /p,t,k/ anda
voiceless labiodental fricative /f/ were chosen among other voiceless consonants.
As there is no word-nal consonant in the expected linguistic representations, it
is assumedthat the best consonant, if one were to be inserted, should be as
inaudible as possible. The voiceless plosives were selected because they are less
audible when in word-nal position than other English voiceless consonants such
as /s,w,tw/, as they are not fully released. The voiceless labiodental fricative was
chosen because it has much weaker energy than other fricatives.
Altogether 60 pairs of expectedand deviant linguistic representations were
prepared. In order to distract subjects from guessing the expected representation by
repeatedly hearing the same one, 40 llers were created using dummy represen-
tations. Five practice pairs of linguistic representations were also prepared.
3.1.2 Recording
In total 123 linguistic representations (5 expected, 60 deviant, 48 dummy, and 10
practice) were readout by a female near-RP speaker of British English in her
twenties with the help of a phonemic transcription of each representation, and
recorded on a DAT tape in the sound-treated booth in the Phonetics Laboratory.
The speaker was one of the four speakers who participatedin recording for the
rst experiment. All the representations were pronouncedin normal citation
prosody. The recorded data were then digitised at a sampling rate of 16 kHz into
a Silicon Graphics Unix computer.
3.1.3 Subjects
Thirty students at the University of Cambridge participated as subjects. They were
all native speakers of British English in their late teens or twenties. None of them
was a birdwatcher or an ornithologist, and therefore they had presumably little or no
familiarity with the birdcalls, and no familiarity with their linguistic representations
in reference books, ensuring their performance is not aectedby prior knowledge.
3.1.4 Procedure
The task for the subjects was to listen to a birdcall followed by two possible
linguistic representations of it each time andto choose which one of the repre-
sentations better imitates the call. The stimuli were presentedusing a Unix
computer and headphones in a sound-treated room. For this purpose, one hundred
pairs of the linguistic representations including the dummy ones were pro-
grammedso that they were presentedaud itorily with their birdcalls to subjects in
a sequential order. A birdcall was presented rst, followed by two representa-
tions (the expectedandthe deviant, or the deviant andthe expected), andthe
birdcall was playedagain at the endof the sequence. After a subject heardall of
these four, s/he was promptedfor an answer by a screen message, and entered
a choice by pressing either 1 or 2.
The pairs were randomisedand preparedin four ways to avoid two possible
order eects, one by the order of presentation of the two representations within
a pair andthe other by the order of presentation of pairs. The four lists are as
follows. A list of 60 pairs with the expectedrepresentations placed 30 times as
Choice 1 and30 times as Choice 2 was randomised together with 40 llers (List
1A). For List 1B, the order of the 100 pairs of List 1A was reversed. For List
2A, the order of the two representations within each pair in List 1A was reversed
for all the 100 pairs. For List 2B, the order of the pairs of List 2A was reversed.
For each list, ve practice questions were insertedat the beginning. At least
seven subjects heardeach list.
3.1.5 Analysis
The raw data of each subject contained 105 answers of either 1 or 2. Of these,
answers for 40 dummy questions and 5 practice questions were excluded.
Therefore, the number of answers to be analysedwas 60 for each subject, i.e.
1,800 altogether.
3.2 Results
First, the data was analysed to see whether there were any signicant order eects
in the performance of subjects. Then the data was divided into three groups, one
with the pairs with deviant FCR, one with deviant initial consonant, and one with
deviant second/word-nal consonant. Each of these was analysed separately.
3.2.1 Order Eects
To analyse whether there were any order eects in performance, the Fisher
Exact Probability Test was used. This test may be the most suitable since in this
experiment the scores from two independent random samples (i.e. List 1A&1B
andList 2A&2B) all fall into one or the other of two mutually exclusive
classes (i.e. Choice 1 or 2) and the two independent samples are small (Siegel
andCastellan 1988: 103).
A statistical analysis revealedthat there was no eect of order of presenta-
tion within a pair (i.e. whether the expectedlinguistic representation is presented
rst or second) in general (p =0.231). On the individual level, however, the
results of three pairs seem to have been aectedby the order of presentation as
shown in Figure 6. In all of these three pairs, the representation that was
presented rst (shown as a dark column) was preferredby each set of subjects.
This eect seems to have occurredwhen both members of a pair sounded to the
subjects almost equally suitable as a linguistic representation of the birdcall.
As for the eects of where a pair occurs in the presentation list, no
signicant dierence was foundin performance on average between the forward
(1A, 2A) andreversed(1B, 2B) lists (p =0.252). Looking at individual cases, the
results of two pairs seemedto be signicantly aectedby position: /ka~k/]
/ga~k/, and/wi u//swi u/. The pair /ka~k//ga~k/ receiveda higher expected-
answer rate when presentedlater , while the other pair /wi u//swi u/ received
a higher expected-answer rate when presented earlier in the experiment. In the
latter pair, the eect may be explicable, as the higher expected-answer rates may
/ka k/ ~
/a k/ ~
/ka k/ ~
/a t/ ~
\t wi t\ w
\t wi k\ w
Number of
Presented first
Presented second
Figure 6. Responses for pairs with signicant order eects within a pair.
be attributedto the presumably better concentration of subjects at an earlier stage
of the experiment. However, it is hardto explain why there should be an order
eect in the rst pair /ka~k//ga~k/. The pairs presentedright before the pair
/ka~k//ga~k/ do not seem to have anything to do with the dierence, either.
Thus, even though it is useful to consider factors which might explain a
position eect, no watertight explanation can be found. However, since there
were only two pairs out of 60 that showeda signicant dierence, the absence
of a full explanation is relatively unimportant.
3.2.2 FCR Dynamic Pattern
To analyse the data individually, the binomial test was used in each section
below since it is a single-sample test andall of the possible observations from
the population fall into one of two discrete categories, 1 or 2 in this case (Siegel
andCastellan 1988: Ch. 4).
In the FCR dynamic pattern group, the overall average expected-answer rate
was 82%, which is statistically signicant (p <0.001), supporting the FCR theory.
Further dividing the data into subgroups according to the shape of the FCR
trajectory, the theory was signicantly supportedin all the subgroups.
In the subgroup of pairs in which the deviant representation had an inverted
FCR dynamic pattern, the result gave the strongest support for the hypothesis,
i.e. the invertedFCR dynamic pattern was seen as the most inappropriate, with
an average expected-answer rate of 88% (p <0.001), as shown in Figure 7. An
asterisk next to the phonemic transcriptions indicates statistically signicant (in
what follows).
In the subgroups of pairs where deviant representations had a level FCR
0 5 10 15 20 25 30
/ju i /
/wi u /
/t t/ I~ w
/t wi t/ w
/ki l / 8
/k li / 8
/ka k/ I
/ka k/ ~
Number of responses
Figure 7. Responses for pairs consisting of representations with expected FCR dynamics
versus deviant linguistic representations with inverted FCR dynamics.
dynamic pattern, the average expected-answer rate was 80% (p =0.001), as
shown in Figure 8. Of the eight individual cases, the pair /twwit//twit/ for the
Twite showedthe lowest rate of 63%, which is not signicant. The dierence
couldbe understoodif there was lip-rounding for the aricate in the production
of /twit/, causing the lowering of formants, andtherefore the deviant representa-
tions sounded similar to the expected representation /twwit/. However, this was
not the case as can be seen in Figure 9.
As Figure 9 shows, the speakers production of /twit/ does not indicate any
trace of lip-rounding, which wouldbe shown in particular by loweredformant
frequencies, as seen in the expectedrepresentation /twwit/ (Figure 10).
Therefore, the possibility of formant frequencies lowered by lip rounding does
not explain this result. Another possibility is that the complexity of the birdcall
itself might have made it relatively dicult for subjects to make a decision. The
fact that the other pair for the Twite, /twwit//twut/ hadalso a slightly lower
expected-answer rate (73%) than other pairs might indicate that it was the case.
0 5 10 15 20 25 30
/ki li /
/k li / 8
/k l / 8 8
/k li / 8
/ku k/
/ka k/ ~
/ki k/
/ka k/ ~
Number of responses
0 5 10 15 20 25 30
/u wu /
/wi u /
/i ji /
/wi u /
/t u t/ w
/t wi t/ w
/t i t/ w
/t wi t/ w
Number of responses
Figure 8. Responses for pairs consisting of representations with expected FCR dynamics
versus deviant representations with level FCR dynamics.
36.15 36.2 36.25 36.3 36.35 36.4 36.45 36.5 36.55 36.6 36.65 36.7 36.75 36.8 36.85 36.9
Frequency (Hz)
Time (ms)
Figure 9. Spectrogram of a deviant linguistic representation /twit/.
Although the result for the pair /twwit//twit/, which was the only non-signi-
cant pair, is hardto explain, it may be concludedthat the overall result for this
subgroup indicates that the linguistic representations with an FCR dynamic
pattern tracking the dynamic pattern of the main components of birdcalls are
predominantly preferred to the representations with a level FCR dynamic pattern.
In the next subgroup of pairs, the deviant linguistic representations have the
same FCR dynamic pattern as the expected linguistic representations, but at a
dierent frequency level.
The average expected-answer rate was 76%, which is statistically signicant
(p =0.003). Although the rate was lower than the former two subgroups (88% for
the inverted FCR dynamic pattern group and 80% for the level FCR dynamic
28.3 28.35 28.4 28.45 28.5 28.55 28.6 28.65 28.7 28.75 28.8 28.85 28.9 28.95 29 29.05
Time (ms)
Frequency (Hz)
Figure 10. Spectrogram of an expected linguistic representation /twwit/.
0 5 10 15 20 25 30
/w / 6 8 ~
/wi u /
/t w t/ 8 w
/t wi t/ w
/ku ku ku /
/ki ki ki /
/k li / f
/k li / 8
/k k/ 6~
/ka k/ ~
Number of responses
Figure 11. Responses for pairs consisting of representations with expected FCR dynamics
versus deviant representations with the same FCR dynamics at a dierent frequency level.
pattern group) as expected, the overall result supports the FCR hypothesis, as
shown in Figure 11.
On closer inspection, the deviant representation was chosen approximately
as often as the expectedone in the pair /twwit//tww8t/ for the Twite. The
complexity of the birdcall might have made it dicult for subjects to choose a
representation to match it. Another possible explanation may be that the slope of
the FCR dynamic pattern of /tww8t/ is more similar to the slope of the main
component of the birdcall in terms of angle than that of /twwit/. In fact, for the
latter two-thirds of the Twite call, it has its lowest component although it is
very weak in energy at a frequency level quite similar to the FCR of the
deviant representation /tww8t/ (Figure 12). Although the expectedrepresentation
/twwit/ has its FCR at a frequency similar to the main component of the birdcall,
the angle of slope might have been a more determining factor in this pair.
The question then arises why /tww8t/ is not usedin references. As far as
0% 25% 50% 75%
Time (proportional)
Frequency (Hz)
Lowest component of a Twite call
Main component of a Twite call
FCR of
FCR of
\t wi t\ w
\t w t\ 8 w
Figure 12. Main and lowest components of a Twite call and FCRs of representations.
phonotactics is concerned, /tww8t/ and/t wwit/ are in the same situation, both
phonotactically non-standard in having the /tww/ sequence. The phonotactic possibili-
ties of a sequence /Cwi/ and/Cw 8/ seem to be the same judging from the possibili-
ties of /twV/, /kwV/, /dwV/, /gwV/, /wV/, and/swV/ (Gimson 1980: 242243). A
possible explanation might be that /i/ occurs three times more frequently than /8/
(Cruttenden 2001: 148) and that /twwit/ might soundlexically more stable than
/tww8t/ to native speakers of English. Another explanation that is more plausible
may be historical in nature; /8/ is relatively new, arising from loss of post-
vocalic /r/, comparedto /i/. The actual reason, however, remains uncertain.
Looking at the other four pairs, the results were expectedly not very distinct
in the pairs /ka~k//k6~k/ and/k 8 li//kf li/, as the deviant representations of
these pairs can also be foundin some references as a received linguistic repre-
In the pair /ka~k//k6~k/, these two representations are presumedto be
phonetic realisations of one orthographic representation kowk. There is therefore
no way to predict a preference for one or the other by the frequency of listing in
references, where only orthographic representations are available. In terms of
absolute frequency, the FCR of /ka~k/ has a more steeply declining slope than
that of /k6~k/ in RP, with the FCR of /a/ starting at higher frequency than /6/ and
ending at the same frequency level. The actual production of /~/ of the diphthong
/6~/ by the speaker, however, seems to reect the fronting of the secondelement
of /6~/ andwas centralised as [ ~], making the FCR higher than [~]. Therefore,
the declining slope of the FCR of /6~/ in her production is less steep than it
shouldbe in RP. However, this should not have too much eect, since both the
slope of [6 ~] andthat of [6~] are less steep than that of /a~/, andvery similar to
that of the main component of the Coot call (Figure 13). The angle of the slope
does not seem to be the factor of preference here. It might be the absolute
frequency in the beginning of the slope that inuences the preference. The Coot
call has, however, multiple harmonic-like components with their beginning
ranging from 800 Hz upward. Consequently, it cannot be said that the absolute
frequency level at the beginning is the determining factor solely because it
matches one of many harmonic-like components. The question why /ka~k/ was
preferredto /k6~k/ remains unansweredhere.
3.2.3 Initial Consonant
The results of the pairs with deviant initial consonants were not clear-cut, and
seem to vary from pair to pair. Detailedresults will be shown only for the Coot
call andthe Whew call due to space limitations.
For the Coot call, the initial consonant for the expectedlinguistic representa-
tion is /k/. As Figure 14 shows, the representation with the expectedconsonant
/k/, /ka~k/, was signicantly preferredto the deviant representations in two pairs,
/ka~k//pa~k/ and/ka ~k//ta~k/ (p<0.05).
In the other two pairs, however, both
expectedandd eviant representations receiveda similar number of responses.
In the /ka~k//a~k/ pair, the result that the deviant representation /a~k/
received a considerable number of responses may be explicable as the birdcall
has a very abrupt beginning and/a ~k/ couldexpress it no less than /ka~k/. In
fact, /a~k/ may even better express it since it does not contain any aspiration in
the beginning, which the birdcall itself does not have, either. The result for this
pair seems to have been greatly inuenced by the order of presentation within
the pair as discussed earlier. Of 15 subjects testedwith List 2A andList 2B,
where the deviant representation /a~k/ was presented rst, 13 of them (73%)
0% 25% 50% 75%
Proportional timing
Main component of a Coot call
FCR of \ka k\ ~
FCR of of the speaker
FCR of 'RP' (estimated)
\k k\ 6~
\k k\ 6~
Figure 13. Main component of a Coot call and FCRs of its linguistic representations.
chose /a~k/ as an answer, while only two out of 15 subjects (13%) testedwith
List 1A andList 1B, where /a~k/ was presented second, did so. This may
indicate that both the expected representation /ka~k/ andthe deviant representa-
tion /a~k/ may be able to express the birdcall almost equally well and can be
consideredas an expectedlinguistic representation.
In the other pair, /ka~k//ga~k/, the result might seem odd at rst glance, as
the voicedconsonant is approximately as frequently chosen as the voiceless
consonant. The voiceless velar plosive is always usedin the receivedrepresenta-
tions for the Coot call including the ones that this and the rst experiments did
not use. This may be explainedby the characteristics of the English voiced
consonants; the English voicedconsonants are not fully voiced unlike the French
voicedconsonants (Ladefoged1993: 144).
It is /k/, not /g/, however, that is usedin the receivedrepresentations. One
reason may be that the English voicedstops are voicedparticularly between
vowels, while they are not fully voicedin non-intervocalic position. An extensive
survey of linguistic representations for other birdcalls has shown that the voiced
obstruents are normally usedfor rather low-pitched andharsh birdcalls including
Fulmar (ag-ag-ag-arr). It may therefore be saidthat the voiceless consonants are
in general preferredfor expressing high-pitchedsound s.
The Whew subgroup is dierent from other subgroups in that the expected
representation does not have any initial obstruent. Of the four Whew pairs, two
0 5 10 15 20 25 30
/ga k/ ~
/ka k/ ~
/ta k/ ~
/ka k/ ~
/pa k/ ~
/ka k/ ~
/a k/ ~
/ka k/ ~
Number of responses
Figure 14. Responses for Coot pairs consisting of a representation with an expected initial
consonant /k/ versus deviant representations with a deviant initial consonant.
pairs (/wi u//kwi u/ and/wi u//swi u/) showedthe result that the expected
representation without any initial consonant was signicantly preferredas shown
in Figure 15, the responses being 70% and80% respectively.
This may indicate that the birdcall was perceived as containing no such
elements corresponding to the onset consonant of a linguistic representation. The
higher expected-answer rate of 87% for the pair /wi u//swi u/ may also
support this suggestion, since the hissing noise of the sibilant can cause more
auditory disturbance than the stop burst.
In the /wi u//twi u/ pair, although this deviant representation may behave
similarly to /kwi u/ andcouldactually disturb the formant transition much more
than that, /twi u/ receivedmore responses than the previous two deviant
representations andthe preference for the expectedrepresentation was not
signicant. One possible explanation might be the subjects familiarity with the
sequence /twi/. It is part of tweet-tweet /twit twit/, the common linguistic
representation for a generic birdcall, and therefore /twi u/ might have sounded
familiar andmore acceptable to subjects.
The result for the pair with the deviant representation (/&i u/) which can be
a more historical pronunciation of the receivedlinguistic representation whee-oo
showedthat the modern pronunciation /wi u/ was preferred, although it was not
statistically signicant (p =0.1). This might suggest that even the weak fricative
0 5 10 15 20 25 30
/twi u /
/wi u /
/swi u /
/wi u /
/kwi u /
/wi u /
/ i u / &
/wi u /
Number of responses
Figure 15. Responses for Whew pairs consisting of an expected representation with no
initial obstruent versus deviant representations with a deviant initial obstruent.
energy of /&/ is not preferred when the birdcall does not seem to have corre-
sponding signals.
3.2.4 Second/word-nal Consonant
The results for the pairs with deviant linguistic representations containing deviant
second/word-nal consonants seem to have been even less clear-cut. Two cases
of birdcalls, the Coot andthe Whew, are to be presentedhere as well.
For the Coot call, the word-nal consonant for the expected linguistic
representation is /k/. The results for the Coot pairs show that in two pairs the
expectedlinguistic representation was signicantly preferred, as shown in
Figure 16, although the overall results did not support the hypothesis that the
voiceless velar plosive /k/ is preferred. In the pair /ka~k//ka~t/, 67% of the
responses signicantly chose the expectedrepresentation /ka~k/ (p =0.049), while
in the other pair with a deviant representation with another voiceless plosive /p/,
/ka~k//ka~p/, the expectedrepresentation receivedfewer responses (57%),
though still a majority. These three voiceless plosives, /p,t,k/, all condition the
pre-fortis clipping eect, andthe duration of /ka~/ (from the beginning of the
burst of /k/ until the endof voicing of /~/) of the three representations is not very
dierent (233 ms for /ka~k/, 264 ms for /ka~p/, and256 ms for /ka~t/). Therefore,
the dierence in the expected-answer rate of the pairs /ka~k//ka~p/ and
0 5 10 15 20 25 30
/ka g/ ~
/ka k/ ~
/ka t/ ~
/ka k/ ~
/ka p/ ~
/ka k/ ~
/ka / ~
/ka k/ ~
Number of responses
Figure 16. Responses for Coot pairs consisting of a representation with an expected word-
nal consonant /k/ versus deviant representations with a deviant word-nal consonant.
/ka~k//ka~t/ might be attributedto the place of articulation, or more concretely,
the location of the energy concentration of the voiceless plosives. The FCR
dynamics of /a~/ having the falling pattern, the deviant representation /ka~t/
receivedfewer responses comparedto the other deviant representation /ka~p/,
probably because the location of the energy concentration of /t/ is relatively high
while that of /p/ is low, which might make the FCR transition more natural
although it may do so less when the consonant is after rather than before a
vowel. It must be mentionedhere, however, that the result for the pair /ka~k//ka~t/
was signicantly aectedby order of presentation within the pair. That is, there
was a tendency for subjects to choose the representation that was presented rst.
This could mean, as already discussed above, that both representations for the
birdcall sounded equally appropriate to the subjects. Therefore it might be
suggestedthat the choice for the word-nal consonant between /k/ and/t/, andin
fact among all voiceless plosives, is rather nely balanced. This ne balance
may also be supportedby the fact that the name of the bird, Coot /kut/ is
onomatopoeic, anduses /t/ as a word-nal consonant (Lockwood1984: 4647).
In the pair /ka~k//ka~g/, the expectedlinguistic representation /ka~k/ was
predominantly preferred. This may be because of the voicing of /g/ that makes the
diphthong /a~/ in /ka~g/ longer than that in /ka~k/. The duration of /ka~/ in /ka~g/
was 515 ms, whilst that in /ka~k/ was only 233 ms. Since the duration of the Coot
call is very short, lasting only about 100 ms, /ka~k/ may have been preferredfor
its shorter duration. It may also be because /g/ rarely occurs after a diphthong in
English; the sequence /-a~g/ indeed is unphonotactic (Gimson 1980: 247).
As for the pair /ka~k//ka~/, however, the result seems a little complicated.
In this pair, both representations receivedan equal number of responses. The pre-
fortis clipping eect may have workedin favour of /ka~k/, while the absence of
signal in the Coot call corresponding to the word-nal consonant may have acted
in favour of /ka~/. In terms of duration, the duration of /ka~/ in /ka~k/ (233 ms)
is much shorter than that of /ka~/ (487 ms), presumably because of the pre-fortis
clipping eect, which may make /ka~k/ a more appropriate representation. On
the other hand, lack of a word-nal consonant in /ka~/ may make /ka~/ more
suitable than /ka~k/, which has an unnecessary acoustic element at the endas the
birdcall does not have a corresponding signal.
The results for the Whew pairs were less complicatedcomparedto the Coot
pairs. The expected representation for the Whew call does not have any word-
nal consonant. In three pairs out of all the four results, the expectedlinguistic
representation without any word-nal consonant was signicantly preferred
(Figure 17). Even in the pair where the result was not signicant, the expected
representation needed only one extra response to be a signicant preference. It
therefore might be saidthat it was quite clear to the subjects that there should
not be any consonant at the word-nal position of the representation for this call.
3.3 Summary
In the secondexperiment, a hypothesis and observations from the rst experi-
ment were tested from the viewpoint of perception. Five kinds of birdcalls were
chosen from the ones usedfor the rst experiment, andone linguistic representa-
tion for each birdcall was taken as the expected representation. Sixty pairs of
linguistic representations were preparedfor the experiment along with forty
llers, each pair containing the expectedrepresentation anda deviant representa-
tion with a deviant FCR dynamic pattern, initial consonant, or second/word-nal
consonant. All the pairs were presented, preceded and followed by the relevant
birdcall, for thirty subjects to choose the representation that better describes or
expresses the birdcall.
In the FCR dynamic pattern group, the overall result overwhelmingly
supported the FCR hypothesis. In this group, the 20 pairs were further divided
into subgroups according to the type of FCR dynamics an inverted pattern
(i.e. inverse of the expectedpattern), a level pattern, andthe same pattern at a
dierent frequency. In the subgroup with deviant representations with an
0 5 10 15 20 25 30
/wi u f/
/wi u /
/wi u k/
/wi u /
/wi u t/
/wi u /
/wi u p/
/wi u /
Number of responses
Figure 17. Responses for Whew pairs consisting of the expected representation with no
word-nal consonant versus deviant representations with a deviant word-nal consonant.
invertedFCR dynamic pattern, the expectedrepresentations were signicantly
preferredto the deviant ones in all the pairs, andthe expected-answer rate was
the highest of all the subgroups. In the subgroup with deviant representations
with a level pattern, the expectedrepresentation was signicantly preferred
except in one pair. In the subgroup with deviant representations with the same
FCR dynamic pattern at a dierent frequency, the results were also in favour of
the expectedrepresentations although the average expected-answer rate was
lower than the two subgroups previously mentioned. To conclude the FCR
dynamic pattern group, the results supportedthe hypothesis, andthe resemblance
of the FCR trajectory with the dynamic pattern of the main component of the
birdcall may be said to be the key factor for the vowel selection for a linguistic
In the groups where the deviant representation of a pair is dierent from
the expected representation in terms of initial consonants or second/word-nal
consonants, the results were not as straightforwardas the previous group.
Although most unexpectedresults couldbe explained, selecting consonants may
be more nely balancedand aectedby many competing factors including
duration of the call, absence of signals in the birdcall, and specic eects of
certain consonants such as pre-fortis clipping.
4. Conclusion
In some aspects of language, sound andmeaning are closely linkedto each other.
The most direct link may be found in onomatopoeia, where language is used to
represent non-linguistic sounds. This paper attempted to further develop the claim
that F2 of a linguistic representation mimics the birdcall. Two experiments
revealedthat the front cavity resonance (FCR) of a linguistic representation, rather
than the secondformant, was the most important factor in creating a linguistic
representation of a birdcall. While the FCRdynamic pattern determines the selection
of vowels, consonant selection seemedmore complicated and nely balanced.
1. FCR may not be a perfectly appropriate term since there is no front cavity for open vowels in
a strict sense. The term is nevertheless usedhere because the continuity of the resonance
frequency is considered to be important. For discussion relevant to this issue, see Kuhn (1975).
2. Apart from the fact that /ka~k/ is the receivedlinguistic representation, a signicantly large
number of responses for /ka~k/ over /pa~k/ and/ta ~k/ might partly be attributable to an
articulatory reason. Since /a~/ is a [+back] diphthong, it might be easier to pronounce with a
consonant that is also [+back]; note that both /t/ and/p/ are [back].
Bladon, R. A. W. 1977. Approaching onomatopoeia. ArchivumLinguisticum 8: 158166.
Cruttenden, A. 2001. Gimsons Pronunciation of English: 6th Edition. London: Arnold.
Fant, G. 1960. Acoustic Theory of Speech Production. The Hague: Mouton.
Gimson, A. C. 1980. An Introduction to the Pronunciation of English: Third Edition.
London: Edward Arnold.
Gooders, J. 1982. Collins British Birds. London: Harper Collins.
Heinzel, H., R. Fitter andJ. Parslow. 1972. The Birds of Britain and Europe. London:
Hume, R. 1990. Birds by Character. London: MacMillan.
Kuhn, G. M. 1975. On the front cavity resonance andits possible role in speech percep-
tion. Journal of Acoustical Society of America 58.2: 428433.
Ladefoged, P. 1993. A Course in Phonetics: Third Edition. Fort Worth; London: Harcourt
Brace College Publishers.
Lockwood, W. B. 1984. The Oxford Dictionary of British Bird Names. Oxford: Oxford
University Press.
Plato. Cratylus. (H. N. Fowler (Trans.) 1926. Cratylus. Cambridge, MA: Harvard
University Press.)
Sample, G. 1996. Bird Songs & Calls. (with CDs) London: Harper Collins.
Sapir, E. 1929. A study in phonetic symbolism. Journal of Experimental Psychology 12:
Siegel, S. andCastellan, Jr., N. J. 1988. Nonparametric Statistics for the Behavioral
Sciences: Second Edition. New York: McGraw-Hill.
Trask, R. L. 1996. A Dictionary of Phonetics and Phonology. London: Routledge.
</TARGET "mas">
TITLE "Visual iconicity andiconic mapping"
WIDTH "150"
P:1 II
Visual iconicity and iconic mapping
</TARGET "p2">
AUTHOR "John J. White"
TITLE "Perspective in experimental shapedpoetry"
WIDTH "150"
Perspective in experimental shaped poetry
A semiotic approach
John J. White
Kings College, London
1. Early attempts at suggesting perspective and depth
in twentieth-century poetry
In 1925 Filippo Tommaso Marinetti attemptedto rally his troops by publishing
Italian Futurisms secondmajor group-anthology of poetry: I nuovi poeti futuristi.
The typographical work of these self-proclaimednew Futurists of the 1920s
was on the whole decidedly more adventurous than that in the founding antholo-
gy which Marinetti haded itedin 1912 under the title I poeti futuristi or that in
the pages of their house-journal Lacerba. Figure 1, for example, belonging to a
series of ve so-calledpanoramic word-tables by Bruno G. Sanzin, is more
radical in its break with horizontality in places even easy legibility and in
its diverse forms of iconicity than anything in the movements 1912 dbut
anthology. Sanzins poem was createdin 1924 as the secondin the planned
panoramic free-wordseries andappearedin I nuovi poeti futuristi the following
year. As a latecomer to the Futurist movement, Sanzin was to remain very much
on the periphery of the Italian avant-garde. Indeed, his Panoramic table no. 3 was
not publishedfor almost half a century (Sanzin 1972), even though it is the
boldest and most haunting of the panoramic works which Sanzin produced
prior to moving on to fresh pastures. Before addressing this work in any detail,
however, I shouldlike to use the earlier Panoramic table no. 2 as a basis for
considering certain features of the semblance of depth and perspective created in
shapedpoems on two-dimensional surfaces.
As is obvious from rst inspection, Panoramic table no. 2 (Figure 1) is a
sparse diagrammatic construct, combining a few drawn lines and iconically
shapedword s in contrasting typographical styles; it thereby achieves a result that
couldjust as well have borne the title schematic landscape with train entering
Figure 1. Bruno G. Sanzin, Panoramic table no. 2, 1924.
tunnel. The words running diagonally across the top left-hand portion stand out
as being narrative rather than baldly referential, as most of the works other
verbal components are. Even though they have not been identied as such, they
might be a quotation from, or possibly a pastiche of, a poem in a more passist
lyrical style andhence part of the kindof post-Symbolist/avant-garde contrast
common to many Futurist works of the previous decade. In English translation
they read: The mountainous coast rises up in leaps and bounds and reveals its
bare bones. (Sanzin hailedfrom the Trieste region and as his autobiography, Io
e il futurismo, makes clear, in the 1920s he came to be regarded by Marinetti as
Futurisms main representative in that politically sensitive outpost.) Perhaps
thanks to its prominent location or because the top of the page is where most
literary traditionalists still expect to nd the title of a poem, this sentence has
even been mistaken for the works title (Blumenkranz-Onimus 1984: 131), one
which wouldhave restricted attention to a single peripheral feature of Sanzins
poetic landscape and would have been longer than all the other parts of the work
put together. By contrast, the actual generic caption, Panoramic table no. 2,
establishes its position within a remarkably heterogeneous experimental sequence,
reproductions from which can be foundin Caruso andMartini 1974: 209. and
1977: 268f. While that top line may not supply the works title or focal point, it
is important because its iconicity involves more than the mere outline shape of
a mountainous coastal region. For while it suggests the contours of a hill by
means of a line of words, as one reads the undulating sentence running at an
upward gradient of some 40 degrees, the eye also guratively rises up in
bounds. This part of the landscape is dynamically visual, not merely iconic in
the static tradition derided as an anachronism by the Futurists.
Dropping below to the centre of the work, we nd a further key dynamic
element: the carriages (or freight-cars, Italian allows for both meanings) of what
looks like a childs wooden toy-train made up of the individual bold upper-case
letters of the plural vagoni. It is only possible to readthe word with ease either
by tilting ones headprogressively to the left or gradually rotating the page
clockwise. But then, since the nal alphabetical item of rolling stock is only
discernible as a letter i by virtue of its place in a horizontal matrix of letters,
what might have otherwise lookedlike a simple oblong box-shape becomes
interpretable as a letter of the alphabet iconically representing the trains guards
van or, to use an American term which more accurately captures the shape of the
letter: the nal box-car. The individual letters that collectively make up the
train are also iconic in their suggestion of three-dimensionality. That is to say,
we can see three of the vehicles six notional sides, which is the most any
vantage-point is likely to oer in real life. By contrast, the letters (lower left)
that make up the word campi (elds) have the appearance of being no more than
two-dimensional. Typographically, thanks to the spideriness of the writing, they
create no illusion at all of rising up above the surface of the landon either side
of the railway-line, as train, hill and tunnel-entrance do. Moreover, despite (or
perhaps because of)
the impression they give of atness, a rudimentary sense of
depth and perspective has been introduced in this segment and certainly more
successfully so than in the rest of the conguration. While the elds disposition
is accommodated to the bending railway-track, an illusion of perspectival
convergence is also implied by the angled, slightly diminishing letters; after all,
the typographical space involvedis in this case primarily transversal rather than
receding towards some distant vanishing-point and the constituent letters are at
the same time also intended to suggest the diagonal-running rectangular shape of
elds, possibly as seen from a moving perspective that is following the passage
of the train. Clearly, Sanzins operative methodhere is little more than a
simplied version of the orthogonal lines associatedwith linear perspective in
painting and drawing, a technique still impressively in evidence in his own
lifetime in Antonio SantElia andMario Chiattones majestic blue-prints for
Futurist buildings (Tisdall and Bozzolla 1977: 12035).
The archway of the tunnel-entrance is constructedout of a letter G
rotatedclockwise by 90 degrees, but at rst glance looking more like a lower-
case n and- alleria (one meaning of galleria in Italian is tunnel). In
function it is indexical as well as iconically referential, if one allows for the fact
that a tunnel-entrance is the only convenient way of visually representing a
subterranean tunnel. Apart from the train andthe tunnel it is entering, Sanzins
poem boasts one further example of motivatedlanguage. It comes in the form of
the word fumo (smoke), drawn as if emerging from the tunnel-entrance. This
element consists of an ingenious amalgam of sign-functions: not only the familiar
mixture of symbolic word andmotivatedshape (andpossibly even iconicity of
colour by exploiting the whiteness of the page), but, since the smokes position
indicates that the railway-engine has already entered the tunnel (i.e. there is a
suggestion of tunnel that takes the viewers imagination beyondthe visible
tunnel-entrance), it also reveals the direction of movement. The result is a
cartoonish variation on Charles Peirces classic example of smoke qua index (CP
2:305), but this time not so much of the re in the locomotives rebox as of the
presence of the unseen locomotive itself. Yet whereas the poems other shaped
words are on the whole easily discernible, fumo is arguably characterizedby a
more concealed form of visual iconicity. Indeed, some readers might initially see
no more than a purely non-verbal, drawn shape, rather than an iconically
motivatedreferential noun.
Sanzins ephemeral doodling can hardly be said to mark a highpoint even
in Futurist shapedpoetry (although the poets autobiography makes it clear that
he regarded this as the most important experiment in his sequence of panoramic
tables). Nevertheless, it does exhibit a surprisingly broad spectrum of semiotic
eects. For even if much of Figure 1s expressiveneness derives from the somewhat
mechanical deployment of single referential nouns shaped in various ways to
represent their objects, others communicate movement by permitting the left to right
sweep of the readers eye to cause a sensation of dynamism; and in at least one
instance, the indexical aspect of the typography is as important as any iconicity
of shape. I wouldnot want to make any extravagant claims for the semiotic
signicance of what is really little more than a Futurists ve-nger-exercise, but
its sheer minimalism can help identify certain features of importance to the
iconicity of perspective andd epth in early twentieth-century shapedpoetry .
One salient aspect of the work is the way in which andwhat is more, the
sheer fact that an illusion of depth, three-dimensionality and even contouring
is created. Such features are relatively rare occurrences in the history of shaped
poetry. The majority of carmina gurata from Greek antiquity, the Renaissance
andthe German Baroque right through to the calligrammes andpicture-poems of
the twentieth century (see Pozzi 1981, Adler and Ernst 1987 and Higgins 1987
for a conspectus of the principal examples) remain resolutely two-dimensional,
not just in the sense that they are constrainedby the context of print on at page,
but on account of their refusal to avail themselves of the techniques andcod es
eventually available for creating the illusion of spatial depth and a sense of
perspective. Although by the fteenth century the architect Filippo Brunelleschi
had inventedRenaissance linear perspective andLeone Battista Albertis treatise
Della Pittura of 1436 hadsystematized its implications for painting, gured
poetry remainedlar gely conned to two-dimensional eects for over four more
centuries. There may well have been goodreasons for this restriction, ranging
from a strict observation of the division of genres (even a gured poem remained
a poem andhence consistedof words on a page) andthe limits of early print
technology to more durable factors such as the need for readability. In the
twentieth century, collages juxtaposition of word-fragments and fragmentary
images on a at surface re-enforced the dominant paradigm of the two-dimen-
sionality of the printedsurface, andin this instance the word-pieces didnot
usually possess an iconic sign-function in any case. Thus, for whatever technical
or cultural reasons, any suggestion of three-dimensionality remained a rarity in
shapedpoetry until certain periods in the twentieth century.
2. Theoretical approaches derived from other visual arts
The Italian Futurists were obsessedwith making jejune conventional language
more expressive, which involvedthem in experimentation with forms of poetic
iconization for which they lackedan adequate analytical terminology. Now we
are closer to the appropriate semiotic frameworks, although they have seldom
been appliedto the work of the Futurists or to shapedpoetry in general, for
that matter. To appreciate just how a post-Peircean approach can illuminate not
only the illusion of perspective in Futurist shapedpoetry , but also the nature and
limits of genuine iconicity in such a genre, it will be necessary at this juncture
briey to re-visit some of the arguments in E. H. Gombrichs Art and Illusion,
Umberto Ecos A Theory of Semiotics and Kant and the Platypus, andNelson
Goodmans Languages of Art, all of which robustly, andat times contentiously,
address the role of convention in what might otherwise be thought of as an
unproblematically iconic signication process.
In a passage very relevant to the material under examination, E. H. Gombrichs
Art and Illusion convincingly shows that modern advertisings techniques of
suggesting perspective by means of foreshortenedletters often rely on our
expectation of the normal letter form to give us the impression of letters or
words arranged in depth and coming towards us with aggressive force (Gom-
brich 1960: 221). His example of the post-Futurist typographical eects in a
cinema advertisement shows that a sense of concave curvature and recession to
the lettering in the words WINDJAMMER CINEMIRACLE (set out one above
the other and appearing to arc inwards towards the centre and grow threateningly
larger as they approach) is a matter of the dierence between, on the one hand,
the radically unconventional shape and size of the lettering used, changing
perceptibly from letter to letter as did Sanzins campi, andour horizon of
expectation concerning conventional regular-sizedletters, on the other. Gom-
brichs point with this illustration is that such devices are part of an eect
which wouldbe lost on someone who did not know the conventions of [normal]
lettering (Gombrich 1960: 221). Gombrichs example is of a dynamic, Cinema-
Scope-inuenced play with word-shapes; Sanzins elds on a plain, by contrast,
are the static setting for a train in motion. Nevertheless, the underlying principle
remains the same. A typographically deviant layout only becomes meaningfully
iconic when experienced alongside our knowledge of how regularly normal
letters andword s are shapedand set out in a horizontal, linear, two-dimensional
fashion. Under such circumstances, the experimental poet has the choice of
actually juxtaposing relatively conventional typography with deviant forms or
tacitly working with the viewers expectations about conventional typography. In
his panoramic tables Sanzin works with both the conventional andthe innova-
tive, as did Gombrichs Windjammer lm advertisement.
Of course, other conventions andcod es may be involved apart from those
relating to our awareness of the regularity andtwo-d imensionality of convention-
al typography. In his polemical Critique of iconism, Umberto Eco
(1976: 191217) attempts to show that forms of graphic signication, usually
assumedto be unproblematically iconic, in fact depend for their eect on
substantial elements of cultural conventionalization. Thus, according to Eco, the
outline drawing of an animal signies [its object] by means of a cultural mode
of correlation. In the ensuing exploration of Iconism andshared properties
(Eco 1976: 19395), Eco concentrates specically on the example of a drawn
outline of a horse, no doubt because Gestalt psychology had temptingly oered
it as an illustration of a schematic perceptual convention (Khler 1947: 106). He
repeats here in English verbatim the objection he hadalread y raisedin La
struttura assente (Eco 1968), that the sharedproperty in the case of this
particular hypoicon can only be the outline that ostensibly links the drawing with
the animal thus signied:
Suppose now that I draw the outline of a horse on a sheet of paper by one conti-
nuous andelementary line. The sole property that my horse possesses (one
continuous black line) is precisely the property that a real horse does not possess.
My drawing has dened by that line the space inside the horse separating it
from the space outside the horse, whereas the actual horse is in fact a body
within or against a space. Admittedly, if I see the prole of a horse against the
background of the sky, the contrast between the boundaries of that body and
the backgroundcan appear under some circumstances as a continuous line at
whose limits the light is absorbed into the dark body. But [] the boundaries
are not so clear, andtherefore the black line iconically rendering this perceptu-
al experience is decidedly a simplifying and selective one. (Eco 1976: 193f.)
This argument, if accepted, would have substantial repercussions for the method-
ology neededto analyse certain types of shapedpoetry . What is known as lled-
form shapedpoetry (see Adler andErnst 1987: 83101 for examples) does
involve a dense body of outline-shaped horizontal black typography within or
against the space of the white page. In this respect, this particular sub-genre is
closer to the object/non-object distinction between horse and background we
intuitively make in real life. By contrast, most calligrammes in the Apollinaire
tradition, while not even depending on the suggestion of continuous lines or
outlines, nevertheless rely on the sentence-lines ability to signify the linear
continuity of an impliedoutline (of a heart, a crown or a mirror) or diagonal
lines (of falling rain). In this sense, the meandering sentence in the top left-hand
part of Sanzins Panoramic table no. 2 is, Eco wouldhave claimedbetween 1968
andthe mid-1990s, not iconic in respect of suggestedoutline as shared property.
In both La struttura assente and A Theory of Semiotics, the conclusion Eco
at the time drew from his simple horse-illustration (admittedly, one not that much
simpler than the eect in Sanzins poem, except that a continuous drawn line
anda line of words are not the same) is that a graphic convention allows us
to transform, on paper, the elements of a schematic conceptual or perceptual
convention which has motivatedthe sign (Eco 1976: 194). In other words, this
particular hypoicons iconicity is more a matter of convention than any shared
characteristic. If this is accepted, one can either join the early Eco in dismissing
the allegediconism of outline as a naive notion although it has to be noted
that the Eco of Kant and the Platypus (Eco 2000) has himself re-assessedhis
earlier counter-intuitive, Gestaltist argument
or one can modify ones model
of visual iconicity to make room for the role playedby perceptual conventions
andcultural codes in processes of iconic signication. After all, Peirce himself
hadtalked about the icon arousing analogous sensations as well as possessing
sharedcharacteristics (CP 2: 299), even if he probably assumed that the one
was a corollary of the other.
Ecos account of the visually iconic in A Theory of Semiotics is seldom
concernedwith iconicity in artistic contexts. Even his substantially revised
position in Kant and the Platypus concentrates on general philosophical issues of
perception andcognitive semantics andhence oers little help with the semiotics
of culturally coded material from painting and literature, including the illusion of
perspective within largely iconic visual sign-systems. On the other hand, Nelson
Goodman, theoretically following in this respect in Gombrichs footsteps, is
anxious to emphasize how little in [artistic] representation is a matter of
imitation. In his brief section on the conventionality of perspective (Goodman
1968: 1019), what is clearly intended to be a devastating contrast is made
between the highly articial conditions of observation in the case of a perspec-
tivizedpainting andthose which govern our perception of an object in real life.
Pictures in perspective, Goodman also claims, in a passage that comes close to
the assumptions of Ecos La struttura assente and A Theory of Semiotics, like
any others, have to be read; andthe ability to readhas to be acquired. The eye
accustomed solely to Oriental painting does not immediately understand a picture
in perspective (Goodman 1968: 14f.). For as Goodman reminds us, according to
the oldad age, there is more to vision than meets the eye.
Even where their views on perspective appear to diverge, the concern of
both Gombrich andGood man is with the role played in certain visual arts by
cultural conditioning and a familiarity with convention in the reception of an
assumedly iconic feature. However, unlike Eco, they are not working within an
explicitly semiotic framework. Goodmans concern with a series of deceptive
criteria of delity, including the so-called laws of perspective, as well as his
questioning of just how many or few features of a visually representational work
are sharedwith their object, might appear susceptible to being mapped on to
semiotics interest in determining the degree of iconicity; except that, once
discrepancies are located, Goodman tends to rest his case, a lack of imitation
having been duly demonstrated. Yet it could be argued that, from a semiotic
point of view, some questions of interest really only arise at the point of
interaction between the genuinely iconic sharedattributes andelements of
cultural conventionalization. To highlight the conventional element in perspective
does not rule out the importance of other (iconic) components, for the symbolic
andiconic elements work together to enhance the suggestion of perspective.
Ecos subsequent Kant and the Platypus paints a picture of the polemics in the 1960s
and 1970s between iconoclasts, seeking to predicate their analyses on the overrid-
ing importance of convention to the point where any iconicity in many hypoicons
couldbe dismissedas a naive assumption, andiconists stressing genuinely
shared features at the expense of acknowledging the role of conventionalized,
symbolic factors (Eco 2000: 339). Andit is clear from Ecos account that each
side still hadmuch to learn from the other. Lookedat from the outside, with
reference to the poetics of shapedpoetry , The Debate on Iconism (Eco
2000: 338.) provides cautionary lessons for both those who overstress the
importance of convention andthose who virtually ignore it in the interests of
establishing iconicity, but for the present purposes those lessons have to be
modied to accommodate them to a very dierent visual art-form. Whether it is
a matter of the hypoicons assumediconicity or of degrees of conventionalization
andtherefore Peircean symbolicity, perspective in shaped poetry can only to a
limitedextent benet from the theoretical consideration of the parallel phenome-
non in painting andschematic drawing.
3. The suggestion of depth and perspective in Bruno G. Sanzins
Panoramic table no. 3 and related works
In the case of Bruno G. Sanzins more assured deployment of perspectival codes
in his third Panoramic table (Figure 2), it is possible to isolate certain features
which will be worth bearing in mind when later considering the dierences
between techniques, both iconic and conventionalized, for suggesting depth and
perspective in painting (the medium which had most impact on literatures
typographical experimentation) andtheir equivalents in shaped poetry. The
lettering technique involvedhere is known technically as inline display face or
shaded display face and, as the terms suggest, it is a style very much connected
with commerce. Sanzins constructive principle is to take the Italian word case
(houses) andiconize the four constituent letters accordingly, placing them in a
conguration intended to suggest modern apartment blocks. Given Futurisms
involvement in architecture and the movements desire to interface various media
(as in Fortunato Deperos 1927 design for a Book Pavillion made up of various
sections in the shape of gigantic letters of the alphabet (Hulten 1986: 419), it
wouldbe possible to readSanzin s poem as a blue-print for some trendy
contemporary Alphabet Estate (although there wouldcertainly be a few less
attractive locations in some of the alphabet-blocks than in others). The method
here looks like an even more radically minimalist variation on the railway-train
component of Panoramic table no. 2. Again, in a form of Lettrism avant la lettre,
the focus is on the iconic potential of the individual letters of a single referential
noun, not motivation at word- or sentence-level. It may be pertinent to recall that
within a few years Sanzin wouldbelong to the rather amorphous sub-group of
SecondFuturist aero-poets and aero-painters, for the four tenement blocks are
clearly presentedfrom an aerial perspective. Only from such an elevatedvantage-
point are we able to recognize, not just the letters in isolation, but the entire
word, even if there is something perverse about making us start reading at the top
(with the c of case) andthus forcing us to engage rst with the at-roof of the
most distant building before reverse-zooming to the more immediate ones. Two
conventions top-down and nearest is more important appear to be at odds
with one another in this instance. The works reception is as a consequence
likely to consist of three phases: (1) synoptic registering of the overall congura-
tion, (2) decoding of the word and (3) correlating iconic sign with its object. The
fact that the nouns foreshortenedcontours give the illusion of sharing properties
of volume and receding shape with the signs object is what distinguishes this
iconic word-poem from the method of Pasquilinos 1917 moonlit street work
Chiaror di luna (Papini 1977: 227). In Pasquilinos more surreally eerie poem (a
precursor of the Metaphysical Art movement to which a number of ex-Futurists
were later attracted), the overall eect is one of similarity in exclusively graphic
terms: none of the language-elements involvedbeing, as far as I can see,
motivated. The individual letters used at various places in the picture do not
make up words and hence they have no obvious object or motivation. There is
no semiotic or semantic reason why Pasquilino shouldhave suggested that his
street is receding into the distance by superimposing the letters A and E on
the road-surface or shouldhave usedan E on its side to suggest a roof-area or
Fs to signify windows and/or shutters. Such motivation as there is remains
indirect: it is between the works title andthe geometric use of shapedletters,
not between iconically shapedword s andtheir objects.
By contrast, in Sanzins
poem the wordis quite simply its object.
The striking perspectival quality of Panoramic table no. 3 derives from
Sanzins use of orthogonal lines to create an impression of foreshortening and,
by including both the ends of the apartment-blocks nearest to the reader and the
inward-facing walls in the central area, a sense of observing position. This, like
most of the other examples citedin this paper, gives the impression that the
repertoire of devices at the disposal of the experimental poet or verbal collagist
is drastically curtailed by comparison with that in painting. From the Renaissance
onwards, painters had been able to work with a combination of linear perspec-
tive, separation of planes, layering, as well as colour- andatmospheric
perspective in order to create on a two-dimensional canvas a mathematically
correct illusion of volume, depth and accurate placement of gures and objects
in space (Dunning 1991). Inasmuch as even avant-garde typography has tended
to remain monochrome (viz. black ink on a white page), one of the above
options is de facto ruledout. Classic colour theory, according to which warm
colours appear to advance and cooler ones to retreat, remains unavailable as a
code in modern shaped poetry, although it does have a role to play in multi-
Figure 2. Bruno G. Sanzin, Panoramic table no. 3, 1924.
coloured avant-garde book-cover designs. Yet despite the overall impression of
impoverishment, typography does have other compensatory ways of suggesting
gradatedcontrasts between close andmore distant verbal elements. One can nd
two of these, albeit in a weakly motivatedcontext, in the Futuristically named
Jamar 14s untitledpicture of a party of letter-Is seemingly out for a walk
(Papini 1977: 302). Here, with the aidof some converging lines suggesting
streets disappearing into the distance, the principle according to which diminish-
ing size of letters and their relative blackness can signify degrees of distance or
proximity can be appreciatedin an isolatedform, although we shall shortly
encounter the device in a more complex conguration. In similarly paradigmatic
form, the typography of Ivo Pannagis cover for the printedversion of Ruggero
Vasaris Futurist play Langoscia delle macchine (Caruso andMartini 1974: 204)
succeeds in suggesting depth, not only by layered typography, reminiscent of the
techniques of modern gratti-artists balloon writing, but also in the basic
contrast between the atness of the word langoscia andthe iconic disposition of
letters making up the foregrounded word macchine. In this case, the inconsistent,
fracturedperspective of the foregrounded wordpresents the machines as being
simultaneously torn in various directions in their anguish.
Another technique sometimes employedin early twentieth-century shaped
poetry to create a sense of depth is to work with a relatively balanced combina-
tion of graphic andverbal elements. This was Sanzins methodin his various
Panoramic tables. A wittier, andmore inventive, example can be foundin De
Nardiss Curiosa (Papini 1977: 365). This displays the same indirect iconicity
often to be foundin Futurist typographical experiments where, very often, a title
in conventional non-motivatedlanguage supplies concepts crucial for our
understanding of the referential object of the motivated letters of the alphabet in
the work itself. In De Nardiss case, the letters in the bottom right of the work
do not spell curiosa although if added to the title-noun they take it preciously
close to curiosit! but insteadsuggest a curious woman (feminine because of
the works title), or perhaps even an animal. The letters T andA are so
intertwinedas to signify inquisitiveness and with goodreason, for the concave
drawn elements in the picture conjure up an impression of the inside corner of
a room, while the detail (including part of another letter of the alphabet) about
to come aroundthe corner implies a convex outside corner. Such a three-
dimensional destabilizing eect couldonly be achievedby a combination of
graphic, symbolic and iconic elements in what Eduardo Kac calls impossible
spaces (Kac 1996: 195).
4. The role of perspective in Francesco Cangiullos interventionist collage
My nal illustration from shapedpoetry (Figure 3) technically speaking, the
most sophisticatedof all three-dimensional Futurist works of which I am aware
is particularly instructive in the way in which it uses a contrast between
readability and illegibility to suggest depth and degrees of distance. The work,
Francesco Cangiullos 1915 free-wordcollage poem Milan-Demonstration,
involves an elaborate juxtaposition of verbal, numerical andgraphic elements.
The use of decorative numbers, letters and punctuation-marks in the rendering of
the building top-left is of course not iconic; there is no motivation to the use of
sign for denotatum in this instance, any more than there was in Cangiullos
Numerical Landscape of the same year (Northern Arts and Scottish Arts
Council 197273: 52); isolatednumbers in modern literature are not automatical-
ly referential in the same way as individual words can be. Cangiullos rst work
involving humanized numbers and letters of the alphabet, 4 Carabinieri, 5
Hunchbacks, 2 Wetnurses andthe Dancer of 1913 (Northern Arts and Scottish Arts
Council 197273: 84), where the number ves are so positionedas to look like
hunchbacks, the fours resemble Italian carabinieri in their capes andthe generous
proportions of the two letter Bs suggest the wetnurses, still depends for its
iconicity on the works title. It is unlikely that the ve ves wouldbe interpreted
as hunchbacks without that assistance. By contrast, the process of signication
in the lower half of Milan-Demonstration is far more distinctly motivated.
As date and title suggest, Cangiullos subject is an Italian interventionist
Figure 3. F. Cangiullo, Milan-Demonstration, 1915. DACS 2002
demonstration, either calling for Italys abandonment of neutrality and belated
entry into the First WorldW ar on the Entente side or possibly even marking the
countrys May 1915 declaration of intent to do so. (Probably the former, since
the event is described as a demonstration rather than a celebration.) The
sketched-in iconic frame indicates a setting which includes both the Piazza del
Duomo andthe Galleria Victor Emanuel, with the buildings festoonedwith the
same national ags as some of the demonstrators carry. The crowd itself is
depicted by means of a combination of symbolic and iconic eects: from
symbols (the tricolour, a heroic statue, even the patriotically signicant venue)
to various forms of motivatedlanguage. The lower halfs hand-written upper-case
words standboth for the various slogans being chantedby the assembledpatriots
andfor the dense crowdfrom which they emanate. Slogans such as long live
war, down with Austria andGermany andT urkey, long live Salandra (the
prime minister who eventually implementedthe interventionist policy) and long
live the King, as well as the patriotic song LItalia s desta. About a quarter
of the way up the left-hand side one nds a politically intriguing juxtaposition:
the line starting with Viva Mussolini is just in front of one beginning Viva
Marinetti. If position is iconic (both in the sense that we start reading each line
from the left andare in this instance obliged to readthem from the front working
backwards), then the distributional code of Cangiullos collage-poem rhetorically
gures Marinetti as immediately behind Mussolini in importance.
This is evidently a perspectivized literary collage, unlike the two-dimension-
al ones on relatedtopics that were more typical of Futurist work in general, for
example Deperos New York urban landscape poem or Marinettis Tumultuous
Assembly (Caruso andMartini 1977: 127, 1974: 43). The rst dozen or so lines
of text that make up the crowdin Cangiullos work are still legible. The
subsequent pattern of diminution paired with decreasing clarity has a dual
function: it suggests the visual image of a crowdwhere those positioned at the
front standout distinctly, while those behindprogressively blur into an amor-
phous political mass. But the changing quality of the written part also possesses
an acoustically iconic aspect. The fact that the shouts of those near the front
wouldbe more likely to be audible is signied using the codes of legibility, size
and boldness of lettering. Moreover, the way the collective noise of the demon-
strators merges into an indistinguishable patriotic cacophony is signied by the
reduction of what had begun as symbolic signs (shouted slogans) to a merely
iconic representation of indecipherable sound. And the fact that the written words
swing around middle-right so that Abbasso is being shouted in a counter-
direction to Viva implies competing focuses of vocal energy. In these respects,
the collage has both visual andacoustic depth.
Leaving aside the overall gure/ground aspect, one can see that legibility is
more than just a code for proximity or distance. The angle and positioning of the
majority of words reects the overriding need for propaganda-collage to be
readable. But the position of the scenes observer or the reader is also crucial.
The crowdis seen from above; as a result, more of the layeredslogans can be
read than from an observation-point down at ground-level. Suggestively, the
readers vicarious observing position would appear to be that aorded by
standing on a balcony, as if from the location and with the eyes of the organizers
orchestrating the demonstration. It is as if we are also present among the great
andthe goodon the balcony alongside Mussolini (andperhaps Salandra and
Marinetti), as their words incite the crowd to renewed patriotic fervour. Indeed,
the words coming up from the crowd might be no more than echoes of what had
been said to them in their leaders speeches. In such an ideological collage poem,
a well conceivedvantage-point which gives meaning to the various perspectiv-
izing devices is at the same time also a politically indexical factor. From such a
point of observation the reader receives the sum total of the elements of the
political demonstration, itself presented as indexical of the mood of Italy at that
moment in history. The observing position thus spatializes an awareness of the
will of the people andacts as a contrast to the deafness of Italys then political
leaders to the mood of the time.
Without recourse to further examples, we can already distinguish between
certain conventions in modern shapedpoetry andliterary collage for suggesting
depth and perspective which have their analogues in painting and others peculiar
to the verbal medium. Converging lines andforeshortening can be foundin both
contexts, as can devices for suggesting dierent planes by a process of layering.
Monochrome shaped poems may have to substitute gradations of darkness or
degrees of legibility for the colour-contrasts of painting. But the point to stress
is that in avant-garde experimental contexts the solutions are more likely to be
sui generis than part of an established tradition of codes, as they soon became in
painting. In any case, such solutions are only requiredin those kinds of shaped
poetry which are intended to signify a particular form of iconic congurational
relationship: Sanzins objects in a three-dimensional landscape or tenement
blocks set in communal grounds, Cangiullos gures demonstrating for a well-
known cause in a recognizable Milan setting or his students eagerly mounting the
steps to their university. This is why the often still referential experiments of the
Italian Futurists provide such fertile ground for those interested in typographical
iconicity. It is less easy to nd comparable eects in Dada, Surrealism and post-
War concrete poetry. A sense of depth may still be suggested, for example by
the dierent sizes and degrees of boldness of the one word that makes up
GerhardRhm s poem Jetzt (where the German word for now appears to
hang in three-dimensional space), but in this case depth is an abstract dimension
lacking any iconic semiotic function.
5. Iconic elements in Eduardo Kacs holopoetry
As a general matter of principle, it makes sense to explore forms of perspectival
iconicity in shapedpoetry with reference to the technology available at the time
works were produced. Even as late as the Futurist period, the problem of print
technology still often hadto be overcome by resorting to hand-engraved letters,
deformed words and shapes. Coming at this topic from a deliberately dierent
angle in the nal part of the present paper, I shouldlike to break away from the
previous relatively circumscribed, homogeneous corpus of early twentieth-century
illustrations andmove forwardto the nal decade of that century. For what I
now want to do is to conclude with a brief consideration of one of the most
signicant new forms of shapedpoetry of the last century, one where the illusion
of depth and perspective is created, not by the use of conventional printing or
handwritten words, but by means of a combination of holography and digital
technology. Such experiments transcend dependence on the printed word, the
drawn shape, the page or the canvas. They are the result of work in, but
emphatically not transferredinto, a relatively new heuristic medium. In a way at
best only pregured by the poetic mobiles of the 1950s and 1960s, what are
calledholopoems are designed to break free from what their inventor rejects as
the rigidity of the [immutable] page (Kac 1996: 192). According to Kac, a
holopoem must be readin a broken fashion, in an irregular and discontinuous
movement, andit will change as it is viewed from dierent perspectives (Kac
1996: 189). Such a reception has implications for the forms of signication
involved, for it represents an emancipation from static, unequivocal forms of
iconic eect, in favour of the textual instability (Kac 1996: 193) of the multi-
layeredand multi-facettedholograph. The result is a work which continually
oscillates between morphing text andprotean images in a more fundamental
break with the semiotics of a xed sign-object iconicity than any other work of
the twentieth century hadaccomplished .
What Eduardo Kac
has inventedwith his holopoems is a form of poetry
that seeks to work within a form of four-dimensionality. In a situation where the
perception of the texts changes with viewpoint, time, embodied in the viewers
shifting perspective, also becomes a constitutive factor. As Kac puts it:
the perception of a holopoem takes place neither linearly nor simultaneously,
but rather through fragments seen by the observer, according to decisions he
or she makes, depending on [] position relative to the poem. Holopoems are
[] quadri-dimensional because they integrate dynamically the three dimen-
sions of space with the added dimension of time []. A holopoem is a spatio-
temporal event; it evokes thought processes andnot their result.
(Kac 1996: 186f.)
Inevitably, by virtue of their protean quality, such programmedmulti-sequences
resist adequate replication on the page and hence cannot be reproduced as
illustrations, as my previous examples could.
Somewhere during shaped poetrys
evolution from the creation of an illusion of three-dimensionality on a two-
dimensional surface andthe combination of virtual space andholography , a point
has been reached where conventional methods of reproduction are no longer
adequate because they would limit the experimental eld.
Here is not the place, nor is the present writer qualied, to consider the
technical means by which recent holographic experiments have createda digitally
enhanced high-denition illusion of spatial three-dimensionality; the poet has
himself given a general account of his creative strategies in Writing holopoems
(Kac 1996: 195.). However, it is clear that we have come a long way from the
codes and conventions of my earlier examples. And as was the case with the 1960s
experimental radio-plays espousal of stereophony and quadrophonic eects, the
tendency has not been to use the new sophistication for narrowly mimetic purposes,
but to develop new kinds of abstraction and syntheses of abstraction and
referentiality. This is the case with many of Kacs so-calledmobile signifying
systems. Here is his account of the methodof his early 1980s holopoetry,
work largely carriedout in Rio de Janeiro before his move to Chicago:
Holographic poetry tries to exhibit the impossibility of an absolute textual
structure, it attempts to create verbal patterns with disturbances that magnify
small changes in meaning according to the perceptual inquiry of the reader.
For example, a syntactical system can be createdin which one could see
twenty or more words occupying the same space without overlapping; a word
couldalso transform itself into another word/shape or vanish momentarily.
Letters can collapse andreconstruct themselves or move to form other words
in a time-reversal transition. These andall other latent expressive possibilities
of holopoetry are unique to its grammar andthey are only possible in part
because its space [] is an oscillatory eld of diracting light as opposedto
the tangible surfaces of pages andobjects. (Kac 1996: 193)
As this suggests, the mediums dominant pull is in the direction of abstraction,
as if what couldbe oeredis a more adventurous virtual reality version of
Rhms multiple wordhanging in impliedspace. But, andit is more than just a
latent possibility, a surprising number of Kacs works are actually basedon a
central tension between abstraction andreferences to the experiential worldfrom
which its words, shapes and choreographies have been abstracted. Consider, for
example, the following account of a poem entitledPhoenix, where the two
letters of the alphabet employed, w andi, in places suggest a bird anda
vertical ame respectively, andyet refer iconically beyondthe phoenix-legend to
other aspects of contemporary reality:
My rst piece in Chicago was Phoenix (1989), a poem composedof only one
letter that draws attention to its visual properties instead of representing a
particular sound. Designed with ambiguity, the letter w might be perceived
as a stylizedbirdwith open wings. It oats in front of the holographic lm
plane andis transxed by a vertical open ame that can be readas the letter
i and which moves randomly according to air currents. The laser transmis-
sion letter-image produces a curious harmony with the actual ame, suggesting
that we are as fascinatedby laser images today as primeval man was by re.
Where the laser redmeets the blue ame, a hybridmagenta is perceived.
(Kac 1996: 200)
Despite all the three- or four-dimensional modalities of abstraction, there is still
a strong, albeit permutational, form of iconicity in evidence in Phoenix. And
the same interplay between abstraction andiconic referentiality within a perspect-
ivizedand continually modulating poetic space can also be foundin the follow-
ing abbreviatedaccount (drawn from his catalogue raisonn in Visible Language)
of Kacs 1992 holopoem Astray in Deimos:
[The] natural subject [of Astray in Deimos] is the landscape of Deimos, one of
the two moons of the redplanet. This holopoem is imaginarily written by
someone who has visitedDeimos, which so far is only known to us through
photographs taken by the Mariner andV iking probes. (Kac 1996: 207)
[the work] explores metamorphosis as its main syntactical agent. Deimos
(terror) is the outer, smaller satellite of Mars. The piece is comprisedof two
words [] which are seen through a circle of predominantly yellow light.
Surrounding this scene is a web-like landscape made of shattered glass, which
partially invades the yellow light circle. The circle may represent Deimos as
seen in the sky from the earth, or a crater on the surface or even a spacecraft
window through which one may look down at the spacescape. (Kac 1996: 205)
There can be no doubt here about the iconicity of the non-verbal images; even
the works title suggests some equivocal form of referentiality. But what about
the words themselves? There are at root two mist and eerie (at times
suggestively eliding to create the further word mystery) that oscillate and
move in andout of the various perspectival frames as the virtual reality event
proceeds in a sequence unique to the individual beholders position and move-
ments. Here is Kac again on the mobile aspect of the words reception:
As the viewer moves relative to the piece, he or she perceives that each line
that renders the graphic conguration of each letter starts to actually move in
three-dimensional space. The viewer then perceives that as the lines and points
undergo an actual typographical transformation, they slowly start to recong-
ure a d i erent [] letter. If the viewer happens to move in the opposite
direction, the noun is transformed back into the adjective. (Kac 1996: 206)
Each viewer is in this sense astray in Deimos, his or her visual experiences
will involve dierent oscillations between the noun, the adjective and their
protean context. Since there is no overall shape to the work but simply a series
position-governedpatterns, it is like being on an alien poetic planet without a
map. The instability of the adjective (eerie) and the noun (mist) iconically
reects such a predicament. But as is usually the case in holopoetry, the iconicity
is no longer the product of an immutable, single iconic relationship between sign
andobject, but a matter of seemingly innite momentary permutations. As Kac
says, holopoems dont rest quietly on the surface: each viewer writes his or
her own texts as dierent vantage-points make words shift, blur into one
another, metamorphose into objects, vanish into mist or reverse the process
(Kac 1996: 190). Terms like turbulent syntax, textual instability, time-
reversible uid signs, luminous dissolution, impossible space, animated
fragmentation, morphing, time-smear andthe branching of holographic
space proliferate in Kacs verbal evocations of his various eects. But the point
that needs emphasizing in the present context is that in certain of them, with
titles like Chaos, Adrift, Omen, Havoc and of course Astray in
Deimos, the abstraction is by no means devoid of referentiality and there are
moments of short-livedor potential iconicity, which then characteristically morph
into some other mode of signication.
6. Conclusions
At almost the same time as Eduardo Kac was experimenting with various forms
of holographic poetry, the Eco of Kant and the Platypus was edging towards
delineating some of the implications of holography for our conception of iconicity.
Admittedly, the discussion seldom centres on the hologram, not least because
Ecos volume is very much governedby the need to reassess positions previously
heldin A Theory of Semiotics. Having written elsewhere at length on the mirrors
importance for an understanding of visual iconicity, Eco now turns his attention
to closed-circuit-television images and, more pertinently, those of an enhanced
form of television, in other words, forms of hypoicon unknown to Peirce:
Let us suppose [] that the television has been perfectedto the point that we
can have three-dimensional images large enough to correspond with the
dimensions of my eld of vision, andeven [] that the screen has been
eliminatedandthere is some apparatus that transmits the stimuli directly to the
optic nerve. In such a case, we wouldreally nd ourselves in the same
circumstances as someone looking into a telescope or standing in front of a
mirror, and this would do away with most of the dierences between what
[Ransdell 1979] calls a self-representing iconic sign (as happens in the
perception of objects in mirror images) or an other-representing iconic sign
(as in photographs or hypoicons in general). (Eco 2000: 373)
Such a television wouldin certain respects border on virtual reality, another of
the challenges to a twenty-rst century conception of iconicity. But rather
tellingly, Ecos musings about the nature of such a super-TV endwith the words
there are no theoretical limits to high denition (Eco 2000: 373). Later, Eco
cites Maldonado (1992: 40) and concurs with him on the fact that a new
typology of iconic constructs, all the way to virtual reality andtherefore not
static but dynamic and interactive iconic constructs [he could be describing Kacs
holographic experiments] sets new problems that require new conceptual
instruments. Eco concludes: I think that a general semiotics must explain the
fact that these phenomena exist (andquestion us), andnot how they work in a
cognitive sense (Eco 2000: 43). That his thinking is pushing towards the
thresholdwhen confronted with the modalities of iconicity in virtual reality, the
hologram andother forms of surrogate stimulus is hinted at when Eco refers to
a standard test for the iconicity brought about by surrogate stimuli: a good rule
for detecting surrogate stimuli [trompe lil eects like the lm of icy vapour
on the outside of the proverbial advertisements beer glass] would seem to be the
following: if I change my point of view, do I see something new? If the answer
is no, the stimulus is surrogate (Eco 2000: 356f.). But a cautious endnote adds
a rider: If the answer is yes, it is not sure that the stimulus is natural; we could
be facedwith a hologram. I suspect that the question of holograms shouldbe
approachedfrom the point of view of my further discussion on mirrors andTV
images (Eco 2000: 427). Andthere the reader is left. Reference to the hologram
andvirtual reality in Kant and the Platypus is largely in the context of sophisti-
catedhigh-d enition iconicity. Andthis may not be inappropriate, given that this
is where most of us rst encounteredholography: in the eerily accurate three-
dimensional replicas that suddenly started turning up in our shopping malls a few
decades ago. Yet as we have seen, in experimental poetry the medium has
movedin a variety of less mimetic, but nevertheless in some cases still partially
iconic directions.
When I began this paper with a piece by Sanzin, it was in order to look at
a work of minimalist referentiality andhence some rather schematic iconic codes
for suggesting perspective. Minimal, andalso ephemeral, iconicity is of course
also a feature of Kacs work, though in a very dierent andmore profoundly
innovative sense. Kacs own personal view of the early twentieth-century avant-
gardes experiments with iconicity tends to make them look like little more than
an insignicant blip in the annals of genuine experimental poetry. In a historical
preamble to his account of holopoetry, he reminds his readers:
Some poets triedto give a new direction to the ancient gurative poem (i.e.
a poem in the shape of an object), but this tendency is a minor part of modern
andcontemporary literary experiments. Even in Apollinaires uvre, shaped
words do not always signify straightforwardly the subjects of the shapes they
were molded into, creating an ideogrammatic tension between the symbolic
(verbal) andthe iconic (visual). (Kac 1996: 192)
Having commentedat length on the connement of the printedword to the at
page, Kac makes no reference to experimental works which seek to overcome
or at least compensate for that limitation by perspectivizing their shaped
words. Perspectivizediconicity is the missing link between the grandtrad ition of
two-dimensional visual poetry which Kac pays homage to (Kac 1996: 186f.) and
his own holographic poetry.
Perspective is still a major factor in Astray in Deimos: not just as a
feature of reception, but also in the various referential scenarios suggestedin
Kacs commentary. But now, insteadof the relatively xed referentiality of
perspective that we encounteredwith the Futurist material, we have protean play
with referential possibilities andnon-referential virtual spaces. Iconicity of depth
andperspective have not been jettisonedin favour of absolute abstraction, but
they have lost their xity. As a result, we are arguably as far out into the outer
space of one particular form of protean shapedpoetry as the moon Deimos is
from the planet Earth, our normal framework for visually iconic eects.
1. Diminution is easier to suggest using regular or equal at surfaces as a point of reference,
hence Renaissance painters liked to suggest depth through the rendering of tiled pavements.
Assuming as we must that the pavements are at andthe tiles identical units, we are compelled
to read their progressive diminution as recession (Gombrich 1960: 221). Sanzins elds are the
Futurist poets equivalent of the square segments of the Renaissance artists tiled ooring.
2. See Iconism andHypoicon (Eco 2000: 33792), especially Section 6.6. where the polemic
about outlines is historicizedas an example of iconoclast violence typical of the periodin
which it was conducted. From his later vantage-point, Eco argues that the hypoicon in some
ways transcribes [] conditions of observation, which in earlier chapters of Kant and the
Platypus have been shown to involve cognitive types (the equivalent of Kants schemata) and
a nuclear content, but also, returning to a concept from Eco 1968: 110., surrogate stimuli.
These are dened as stimuli that surrogate dierent phenomena, which depend partly on the
form of the object and partly on the way in which we decide to look at it (Eco 2000: 353). The
emphasis on surrogate stimuli does not negate the role played by graphic conventions and other
forms of cultural codication in shaping this surrogacy. In general, Eco appears to concede
what a hot potato the missing outline illustration was, especially in the 1960s, and, rather than
seeking to defend the choice, sidesteps the polarizing terms of the earlier debate by focussing
more on conditions of perception.
3. We are in a grey area where post-Peircean literary semiotics, because it argues in terms of the
word(as sign) andits object, oers insucient help with the role playedby the isolated letter
of the alphabet in many avant-garde works. A test-case would be Francesco Cangiullos
Universit (Hulten 1986: 147), which depicts some steps leading up to a large colonnaded
building. The works title is written on the back wall between the columns and below that
stands, as a possible sub-title, the phrase studenti in lettere. Hence, we immediately read the
various letters of the alphabet climbing the steps or promenading beneath the columned
archways as representing students in lettere. Yet for all the framing, the iconicity remains
weak. Some of the letters (A, K, H) display one leg shorter than the other as they climb up
the stairs. On the other hand, the main intended link between letters of the alphabet and
studenti in lettere would appear to depend more on the works title than on any primary
iconicity between sign andobject at the humanizedalphabet level, to use Cangiullos term.
4. For such a reading, see Adler 1987: 10519.
5. Eduardo Kac, a Brazilian poet working at the time in Rio de Janeiro, invented the holopoem in
1983, thus freeing words from the page. In 1989 he moved from Brazil to Chicago where, as
he puts it, I was able to work andexperiment on an ongoing basis. In 1995 he receivedthe
highest international accolade in the eld of holography, the Shearwater Foundation Award, for
his invention andd evelopment of holopoetry. He has works in several international public and
private collections andhas written extensively on innovative poetry andthe visual arts. He is
also on the editorial board of the journal Leonardo. For further information, see Kacs website:
6. Because of their irreducibility as holographic texts, holopoems resist vocalization and paper
print reproduction. Since the reception of the texts changes with viewpoint, they do not possess
a single structure that can be transposedor transportedto andfrom another medium (Kac
1996: 201).
Adler, J. and U. Ernst. 1987. Text als Figur: Visuelle Poesie von der Antike bis zur
Moderne. (Ausstellungskataloge der Herzog August Bibliothek, 56) Wolfenbttel:
Herzog August Bibliothek.
Adler, J. 1987. Progressive Universalpoesie. Die Kunst Gerhard Rhms, Protokolle, 87,
2: 10519.
Blumenkranz-Onimus, N. 1984. La posie futuriste italienne: Essai danalyse esthtique.
Paris: Klincksieck.
Caruso, L. andS. M. Martini (eds). 1975, 1977. Tavole parolibere futuriste (19121944).
2 vols. Naples: Liguori.
Dunning, W. V. 1991. Changing Images of Pictorial Space: A History of Spatial Illusion in
Painting. Syracuse, New York: Syracuse University Press.
Eco, U. 1968. La struttura assente. Milan: Bompiani.
Eco, U. 1976. A Theory of Semiotics. Bloomington, Indiana: Indiana University Press.
Eco, U. 2000. Kant and the Platypus: Essays on Language and Cognition. Tr. A. McEwen.
London: Vintage.
Edgerton, S. Y. 1975. The Rediscovery of Renaissance Perspective. New York: Basic
Gombrich, E. H. 1960. Art and Illusion: A Study of the Psychology of Pictorial Representa-
tion. London: Phaidon.
Gombrich, E. H. 1982. The Image &the Eye: Further Studies in the Psychology of Pictorial
Representation. London: Phaidon.
Goodman, N. 1968. Languages of Art: An Approach to a Theory of Symbols. Indianapolis:
Higgins, D. 1987. Pattern Poetry: Guide to an Unknown Literature. Ithaca, New York:
Cornell University Press.
Hulten, P. 1986. Futurismo & Futurismi. Milan: Bompiani.
Ivins, W. M. 1946. Art and Geometry: A Study in Space Intuitions. New York: Dover.
Kac, E. 1989. Holopoetry andFractal Holopoetry: Digital Holography as an Art
Medium. Leonardo 22: 397402.
Kac, E. 1996. Holopoetry. Visible Language 30: 184212.
Khler, W. 1947. Gestalt Psychology. New York: Loveright.
Maldonado, T. 1992. Appunti sulliconicit. In Reale e virtuale. Milan, Feltrinelli,
Northern Arts andScottish Arts Council. 197273. Futurismo 19091919: Exhibition of
Italian Futurism. Newcastle upon Tyne: Hudson.
Papini, M. C. (ed.). 1977. LItalia futurista (19161918). Rome: Edizioni dellAteneo &
Peirce, C. S. 193159. Collected Papers. Vols. 16 ed. Charles Hartshorne and Paul
Weiss; vols. 78 ed. Arthur W. Burks, Cambridge, Mass.: Harvard University Press.
(Quotedas CP. Reference is made to vols. andparagraphs.)
Pozzi, G. 1981. La parola dipinta. Milan: Adelphi.
Ransdell, J. 1979. The Epistemic Function of Iconicity in Perception. Peirce Studies
1: 5166.
Sanzin, B. G. 1972. Documenti. Trieste: Il Timavo.
Sanzin, B. G. 1976. Io e il Futurismo (condenze in libert). Milan: Istituto Propaganda
Sonesson, G. 1993. Pictorial Semiotics, Gestalt Theory andthe Ecology of Perception.
Semiotica 99: 319399.
Tisdall, C. and A. Bozzolla. 1977. Futurism. London: Thames and Hudson.
Wertheim, M. 2000. The Pearly Gates of Cyberspace: A History of Space from Dante to the
Internet. London: Virago.
White, J. 1967. The Birth and Rebirth of Pictorial Space. London: Faber and Faber.
</TARGET "whi">
AUTHOR "Julian Moyle"
TITLE "Where reading peters out"
WIDTH "150"
Where reading peters out
Iconic images in the entropic text
Julian Moyle
University of Nottingham
1. Introduction
In his essay Iconicity in literature (1996), Jrgen Dines Johansen outlines some
dierences between the iconic diagram and the iconic image. These dierences
help account for the widely held perception amongst theorists of iconicity that
the diagram functions as a versatile andcomplex type of iconicity, based on
relational similarities between the structuring of the sign andits real world
equivalent, whilst the image, in contrast, appears as a rather static anduncompli-
cated reection of an object or concept in the real world.
Firstly, a diagram can express generality, and may enjoy a rather large
freedom in the way in which it renders the relations of the object (Johansen
1996: 40). The image, we can infer, is hindered by the specic requirement that
it must share qualities with its object, a requirement that serves to tie it to the
thing signied, delimiting its possible modes of iconic representation. Related to
its dependent tie to the object, the image depicts an original, whilst diagrams
may as well precede their object as follow it (40), such as a blueprint for the
building of a house. This sense of the late and derivative aspect of what might
be more accurately described as an after-image contrasts with the creative and
constructive qualities of the diagram which enable it to represent intelligible
relations as well as the sensible (40). In this way, the image is made to appear
like a passive emulation of its objects physical presence. A further point of
comparison accounts for the opposing fortunes befalling the image andthe
diagram in the recent history of studies on iconicity. Diagrammatical features,
we are told, are not only eminent in gures and drawings, but, according to
Peirce, are present, andnecessarily present, in the syntax of language (40).
Whilst the diagram can boast such an integrated role in the workings of lan-
guage, the image is restrictedto the linguistic loop-hole of onomatopoeia or to
creative shapings of language such as those observedin modern poetry and
advertising. As linguistics has been the engine room behind recent studies of
iconicity, the image has receivedless attention than the diagram.
Tightening the restriction of the image to the marginal spaces of language
expression, is its complicity in an undesirable history of absolutist attempts to
show, or perhaps force, the existence of a one to one relation between language
and the world. The tradition, dating back to Platos Cratylus, is sketchedout in
GrardGenette s Mimologics (1995), which describes attempts made by mimo-
logists to give individual letters, or groups of letters, xed values that correspond
in a directly imagic way with properties of the world. These attempts often have
an imperialist tone, resulting in a valorization of the elements of ones own
language against those of others.
Facedwith the crisis of the arbitrariness of the sign, the image went
underground, only to re-appear on the pages of modernist poetry in a dierent
Such modernist images were often not as pure as their historical
antecedents, exhibiting less the elemental aspects of language and more a
synthesis of textual features into a total form that was supposedto correspond in
an imagic way with some other object or form. Generally, these were hybrid
forms comprising imagic and diagrammatic modes of iconicity. The staggered
lineation in the opening section of Ezra Pounds Canto III (Figure 1), which
iconically suggests the outline of steps, has, for example, been included by Max
Nnny within a selection of modernist diagrams (Nnny 1986: 216), whereas
Olga Fischer has re-usedthe steps as an example of the sort of iconic feature
that is not to be her concern in an account of the relation between diagrammatic
iconicity andlanguage change, precisely because the steps represent a device
that is much closer to imagic iconicity (Fischer 1997: 71).
Whilst Nnny adheres to the strictness of Wendy Steiners statement (which
Figure 1. An icon of steps at the opening of Ezra Pounds Canto III.
he cites) that verbal art achieves whatever iconicity andpresence it can claim
through relational, diagrammatic means (Steiner 1982: 22), I prefer Fischers
use of the adjective imagic as it acknowledges the tendency of the lines to assume
the character of the image (in their attempt to assume a total and unied shape
that is supposedto correspondd irectly with a total impression of a set of steps).
The problem for modernist images, such as Pounds steps, is that they
perpetuate the aura of absolutism andimperialism which, rightly, contributedto
the denigration of mimological practice in the nineteenth century. The perception
of this politicising hereditary aw has informed the charge levelled against
modernist writers by critics, who have seen something fascistic in the way an
image halts andtotalises the temporal progression of a text within a singular
spatial or imagic array.
For example, when we see an image, such as Pounds
steps, we have to suspend the process of reading that ideally manifests itself as
a continuous anduninterruptedtranslation or activation of a spatial layout
(printedword s on the page) into an internally processed conception of the
temporal andsequential dimensions of the text. We then cast a spatialising gaze
back upon the printedtext, transforming the sequence of language into a singular
iconic image that is seen to reside in the print on the surface of the page.
Against this backdrop of the image viewed as absolute, totalising and
derivative, images have been more recently analysed, but have tended to be read
as features operating on a small-scale, in isolatedimitative spots within texts.
Here, I shall examine a case where the iconic image can be seen to disperse its
eects throughout the wider context of a literary work, one that produces results
that are less absolutist, andmore socially responsible, and which have implica-
tions relating to wider issues of reading. Michel Serress work on the subject of
entropy will inform this analysis.
In order to consider how Erosive, a poem by the English poet, Peter
Reading, may enable us to re-evaluate the category of the iconic image, it is rst
necessary to discover how the poem is situated both within the immediate
boundaries of the book where it is found, and in the wider eld of Readings
oeuvre. We will see how this situating of the poem plays with the readers
preconceptions concerning the presence of the iconic image in literary texts.
2. Image or draft?
Peter Readings Last Poems (1994) opens with a foreword(Read ing 1996: 246),
which, in true postmodernist fashion, is presented as having been written by a
ctitious and pedantic academic named John Bilston, who claims to have the
task of bringing the nal work of the poet Peter Reading to publication. As
Bilston writes of his discovery from Melaleuca, the tip of Southern Australia, it
wouldappear that the poet, Peter Reading, or rather his ctitious doppelgnger,
has disappeared down under in more than one sense. Bilston informs the reader
that the poems are printed in the order in which they are found and that they
were discovered in an envelope bearing the superscription, in the authors hand,
Last Poems. Two paragraphs follow in which Bilston displays his extensive
knowledge of classical literature and metrical form, where he labels various
poems as versions of specic Ovidian, Homeric and Anglo-Saxon sources. It
is, though, the fourth paragraph of the forewordwhich is particularly likely to
rouse the readers curiosity. Here, Bilston writes that Erosive and the untitled
nal two pages are reproduced as found, noting that it is unclear whether the
author intended them to appear in their present form, or whether they represent
drafts towards an unrealised work in progress.
The reader, then, on the verge of turning the page to discover the rst of the
poems that lies beyondthe foreword, is facedwith an option. It is possible to
proceed sequentially, condent in the order of the book. Alternatively, the reader
can respondto Bilstons uncertainty as to whether or not Erosive and the pages
to which Bilston refers represent completedwork. Shouldthese pieces have been
included in the volume facing the reader? What is it about their appearance that
makes them stand out? To go directly to these end-pieces from the foreword of
Last Poems is a move prompted by Bilstons hesitancy; a hesitancy which challenges
the reader, on the threshold of the opening poem, to resolve the question of
whether the nal pieces are drafts, or have reached completion. Such a move is,
also, licensed by the statement that the poems were discovered, unbound, in an
envelope. The reader of Last Poems has an indirect invitation to shue the
pack, andmay , rst, turn to the written pieces lying at the end of the volume.
Touching down at Erosive (Reading 1996: 278), we immediately grasp
why a scholar who excels in displaying the rigours of traditional literary research
methodology could have such a problem in classifying the poem. Whereas
Bilston is quick to seize upon and dene the underlying metrical patterns in
classical poetry, which might evade the untrained ordinary reader, he is stumped
when facedwith the more literal impediment of a poem presentedin a condition
where semantic legibility anda given path for reading (through a sequence of
easily identiable lines) are no longer assured (Figure 2).
Some elements of the poem are well markedout, including various connect-
ed words andphrases, such as crackle of brittle andrustle of old. Other
phrases force us to stop andthink, but prove easy to negotiate; anhydrous
laurel can be made out clear enough, allowing us, with little semantic
diculty, to construct, as leaves, the cluster of letters which follows laurel.
However, it is hardto work out the rst wordin lous holograph.
Figure 2. Erosive (1994).
Faced with these interruptions to reading, Bilston can only conclude that
either the visual dilapidation of Erosive is intentional, in which case he has no
idea as to what the intentions are, or that Erosive is a draft, an unrealised
work in progress which was not complete when he foundit. This second
interpretation results from Bilstons receptivity to a set of signiers which do
suggest draft: the misalignment of letters and punctuation, the poor quality
printing of certain letters, the primitive type font, the handwritten insertion of
the wordbond , the misalignment of sections seemingly skewedat angles to
each other that make up the poetic line, andthe incompleteness of such lines.
This set of signiers is, however, quite obviously in tension with a second
set of signiers, to which Bilston fails to respond. Both semantically and
formally, this secondset of signiers builds up andenacts an image of a (the)
written text suering from processes of erosion. This signication is anchored
in the inter-action between the fragmentedsemantics of the title Erosive and
phrases such as pages of faded, eroded leaves and rotted the frail bond,
and more formal signiers indicating textual breakdown, erosion or implosion:
a group of lines, for example, begins at the left handsid e of the page, and
disappears (often in mid-word or phrase) on collision with a group of lines which
appears out of this line of division the regularity of this divide suggests a
textual ssure, or inclinedfault line, at which the text collapses in on itself; also,
we might note how the V, in the title Erosive, does not just appear to be a
mistyping that contributes to the signication of a draft, but also functions as an
alphabetical icon of a ssure or crevasse that is consciously deployed below the
line that governs the rest of the letters in the title.
The power relation between the two sets of signiers, however, would
appear to be unbalanced. The signications suggesting draft can be subsumed
within the logic of the signications of erosion, so that the image of a draft
becomes re-interpretedas an image of the state of incompleteness that character-
ises a text under erosion. However, the reverse cannot occur to give a satisfacto-
ry reading. In other words, it seems illogical that signiers of erosion be
subsumedwithin the logic of the over-arching signication draft, for the image
of erosion or regression appears to have little to oer to the sign that is sup-
posedto indicate the progressive formulation of a text.
The reader has little diculty in substituting, for Bilstons uncertainty, a
model which views, not the text as the image of a draft, but the text as contain-
ing the signs of a draft which help construct an image of a text under erosion.
The function of the foreword, so it would seem, is, rather simply, to harness the
complicity of the reader in pouring light scorn on pedantic scholarly research,
which may result in the identication of the dactylic tetrameter, though remains
blindto more visible signs, as a result of a lack of familiarity with visual codes
in operation in the external worldandin popular culture.
Whilst our principal focus is Erosive, it is important to see if this
resolution also applies to the words andlines of the untitled nal two pages
(Reading 1996: 27980). These two pages, reproduced as Figure 3, display as a
visual mess what we infer to be pages of poetry. What couldbe interpreted as
splashes or lines of ink wouldleadBilston to consider categorising the text as a
draft, belonging to a particularly careless author. In contrast with Erosive,
where lines contain the promise of legibility at the left handmar gin, only to slip
into a ssure of unreadability, the lines on these pages fail to give much hope to
the assiduous reader. It is even uncertain whether a title belongs to these pages.
Bilston failedto spot one, though a separated line in boldfont slanting upwards
at the top of the page, couldbe made out to readNothing for Anyone.
Whereas Erosive containedsemantic fragments that self-reexively referredto
the text as an image of disintegration, there is little to nothing in the verbal
language that can be discerned on these nal two pages that help describe the
text as representative of an image.
However, in the light of the evidence found in Erosive, that the poem was
not a draft, it is extremely dicult for the reader to avoid reading the nal two
pages with the same glasses. Are the ink streaks not to be better readas the
signs of water damage? The secondpage has some lines pastedup the right hand
margin of the text. It is unlikely that a draft would deliberately involve typing
vertically down this right hand margin. Instead, the reader seems to be faced
with some sort of image of spatial collapse. Of Erosive andthe nal two
pages, the application of the word draft (outside of the contribution of draft
signs to an over-arching image of erosion) seems to be wrong on all counts.
3. Contextual framing
A reading which replaces the possibility that Erosive andthe untitled nal
two pages of Last Poems are drafts, with the certainty that they constitute
imagic patterns of erosion, is buoyedby two contextual factors, the rst of which
may escape the reader who is unfamiliar with the history of the production of
Last Poems.
This rst factor involves the knowledge that the poems (if it is appropriate
to describe them as poems) were originally two pieces of visual/textual art
forming part of an exhibition, also calledErosive, which was organisedby
Peter Reading and Peter Kennard at the South Bank Centre in London in June
The pieces were included as late additions to Last Poems, conceived after
the rest of the volume hadreachedcompletion in June 1993.
The secondfactor is relatedto the purpose of this exhibition, though its
manifestation in Readings work will not have escaped the assiduous reader of
Last Poems, or indeed, of Readings oeuvre as a whole. This refers to the fact
that Erosive, the exhibition, aimed to present visually dilapidated versions, or
images, of poems that Reading had already written.
When we took up John Bilstons indirect invitation to classify Erosive and
the untitled nal two pages of Last Poems, we avoided the standard reading
practice of turning the page to readthe rst poem in the volume. On tracing a
path back to the rst poem, having encounteredErosive, we nd ourselves
confrontedwith something familiar (Figure 4).
Where Erosive appears to be a degenerate version of [Untitled]
(Reading 1996: 247), a poem that lies at the beginning of the same volume, the
Figure 3. The untitled nal two pages of Last Poems (1994).
untitled nal two pages represent doctored photocopies of pages of a poem,
which hadbeen published(without ink streaks) seventeen years earlier. This
earlier poem, the rst page of which is reproduced as Figure 5, bears the clear
title Nothing for Anyone (Reading 1995: 127), which is, also, the title of the
Figure 4. [Untitled] (1994).
volume in which it is located.
Peter Reading has described to Isabel Martin how he went about the process
of making the dilapidated texts for the exhibition: I produced 30 A4 sheets
recycled (grand euphemism) from previous stu of mine, from Final Demands,
Perduta Gente, Shitheads, and Last Poems, andthese sheets were worked on to
make them visual (enlarged by photocopier, burnt, otherwise damaged, torn
&c) (Martin 2000: 22829). In a separate interview with DavidClark, Reading
notedthat the implication is that here is a set of Museum exhibits from some
defunct civilisation (Clark 1994: 27). These production details, in materially
underpinning the suggestion of a direct correspondence between the dilapidated
texts at the endof Last Poems andearlier poems, addweight to the impression
that the eroding written pieces represent excellent examples of iconic images
functioning within a literary work. We must, however, be content with two recent
developments in the eld of iconicity studies which aect our classication.
Firstly, a commitment to the character, as opposedto the technical purity,
of the iconic image, allows us to accept that, though Erosive is not an exact
image of [Untitled] (there is clearly some creative doctoring to suggest that we
are seeing an image of [Untitled] under erosion), Erosive functions principal-
ly to imply direct correspondence. This is facilitated by resemblance between the
words, and the shape of both poems, together with certain correlations between
the ordering of lines (some re-arrangement has occurred during transition).
Erosive in fact fuses a picture of erosive processes with the image of [Unti-
tled] to create a hybrid image. Likewise, our interpretation of the resemblance
between Erosive andour conceptual idea of what an erosive pattern might look
like, involves interpreting relational similarities. For example, we buildup a
conceptual image of a text under erosion by interpreting correspondences
Figure 5. The rst page of Nothing for Anyone (1977).
between broken relations at the level of the signier andour idea of how the
broken relations in the constitution of an object in the worldmight be indexical
of the presence of erosion. It is only after a considerable piecing together of
relations that we see the image of a text (a previous text) under erosion.
Secondly, we must accept, provisionally, Winfried Nths distinction
between an exophoric iconicity, basedon the established form miming meaning
equation, andan endophoric iconicity, where form mimes form (Nth 2001: 22).
It is, of course, highly problematic, to claim that Erosive represents an object
or concept in the real world, because questions ensue with respect to whether
conceptual representations of erosion oer a stable real world referent (and
whose conceptual representations are they?), andwhether [Untitled] can count
as an object in the real world.
Endophoric Iconicity, for Nth, is more like intra-
textual miming (I wouldextend this to include inter-textual miming), though we
may still retain a valuable distinction between, on one hand, reference through
mime andresemblance and, on the other, reference as it exists through more
arbitrary semiotic channels.
In loosening some of the semiological technicalities concerning how
Erosive, as an iconic image, should plot a one to one correspondence with its
object(s), we are displaying a commitment to the role played by the context of
a literary work in determining the character of iconic relationships.
4. Append(im)ages
To determine an iconic image on the basis that an arrangement of text has an
imagic character in a particular context is a move which gives, to the critic of the
literary image, some scope, beyondthat aorded by the strict linguistic deni-
tion of the image. What we must constantly be preparedto de-construct,
however, is the notion of context, precisely because the context of the image
may be conceived selectively and in accordance with culturally dominant pre-
suppositions (pre-suspicions) regarding the images character.
We should, therefore, consider how far the identication of Erosive and
the untitled nal two pages as iconic images is motivatedby notions of
lateness and otherness which characterise the following contextual consider-
ations: the sense of the poems being post-apocalyptic debris exhibited in a
museum; the original site of the images being an art exhibition; the late addition
of the pieces to the volume Last Poems; the knowledge that the pieces are
visual versions of earlier poems in Readings work.
The notion of lateness has its basis in the assumption that Erosive and
the untitled nal two pages, being images of poetic originals (which occupy a
place of temporal precedence in Readings work), have, in eect, already been
read in their original, non-imagic forms. As such, both pieces function as, rather
simply, pictorial precipitations of the semantics of disintegration already
expressedin the originals. In Last Poems, then, Erosive becomes framedas a
visual parasite, or a hollow reproduction, feeding o an earlier site, where it is
assumedthat the meaning of the image was generated.
The sense of otherness is associatedwith problems relating to the generic
identity of the visual/textual pieces. Both pieces, originally placed in an art
exhibition, become framedas graphic images intruding upon a volume of poetry
which, through its signications of literary classicism, presents itself as a site
prioritising verbal (not visual) language as the dominant semiotic in the volume.
Any images ought, within this generic framework, to be construedas verbal
images. The generic codication of the volume extends to its expected reception,
in accordance with a model of reading (a classical one?), where priority is given
to a linear methodof text reception, involving the sequential translation of lines
of material text into concepts or images that form in the readers imagination.
This process requires the visible surface of a text to be smoothedout, or
anaesthetized, so that such a reading can proceed more easily. Erosive, in
drawing attention to its visible surface, as well as encouraging readings that
visualise this textual surface as a singular graphic image, nds itself marginalized
andframedas a threat to the generic purity of the rest of the volume.
Perhaps the poems couldbe liberatedif we were to readthem as we might
approach a concrete poem, where we wouldneed to remain attentive to the play
between visual andverbal semiotic strategies? Even though concrete poetry is
marginalizedas a minor inter-semiotic genre, the space which it occupies in
the gap between the central concerns of university departments of Literature and
Art History at least retains a certain autonomy that enables specic analytical
practices to be developed in order to deal with semiotic problems posed by the
hybridity of such verbal/visual constructions. However, the specic marginali-
sation of Erosive andthe untitled nal two pages is more acute than the way
that concrete poems are marginalizedfor the reason that, whilst identied as
other to the generic core of the volume (andby extension Readings oeuvre),
both written pieces remain held in a relationship of dependency towards origi-
nals which occur within that generic core. This Derridean paradox of inclusion/
exclusion, often appliedto post-colonial discussions of marginalisation, also
describes the situation of these late additions to Last Poems within the physical
space of the book that in/excludes them.
The implication that Erosive must be interpretedthrough a semiotic of
art (Huisman 1998: 41) that is other to a semiotic of language (41), works to
repress, more insidiously, the recognition of what Rosemary Huisman has
termedthe semiotic of seen language (41). This alternative form of semiosis
threatens to de-stabilize the notion of a generic core of verbal semiosis, not
through hybridization at its boundary (as with the concrete poem), but through an
uncovering of the other lying within. It is here, restrainedinsid e the book, that
Erosive must plot its revolution.
A contextual framework, then, which aligns Erosive andthe untitled nal
two pages with notions of lateness andotherness, reinforces Johansens set of
denitions which perpetuate a marginalisation of the iconic image. Also rein-
forced is the desire to impose order, structure and linearity upon a selection of
poems, which, we shouldremember , were discoveredunboundin an envelope.
5. De-constructing the Entropicon
A paradox exists in a reading of Last Poems which happily synthesizes the
semantic references to the text becoming undone with belief in the existence of
a xed link between Erosive and[Untitled ] or, on a wider scale, between the
untitled nal two pages andNothing for Anyone. Thinking of images and
originals, in this way, works to do up the text by threading together linear paths
in the face of the thematic emphasis on disintegration.
What is more, Isabel Martin has described how Readings strangely
attractive exhibition artefacts (do we spot the otherness of the iconic image
again here?) work at physically fortifying the strongly intratextual unity of his
oeuvre (Martin 2000: 229). It seems that Erosive must function both as a
weakened replica of an earlier, original site of reading and as a strengthening
device, casting back a visual aura onto the original, which receives some added
presence from the connection. The specic relationship between, say, Erosive
and[Untitled ] can, by extension, be imaginedas being applicable right across
Readings oeuvre, so that all poems can be conceived to exist in relation to an
iconic breakdown of themselves. In this way, the entire work can become
perceivedas a total, iconicizedand fortied structure, bearing a close similarity
with Charles Morriss conception of a total icon (Morris 1971: 422).
This paradox is retained for as long as Readings work is conceived as an
entropic structure involving linear, temporal ows between xed points. Entropy
involves a general and irreversible drift from order to disorder (described by the
secondlaw of thermodynamics). It is accompaniedby a loss of energy, a
condition which ts the way in which the iconic properties of Erosive are
rendered passive at the same time as they signal rising entropy.
It is fairly straightforwardto think of entropy as a kindof invertedteleology .
Where teleology describes a progressive path from disorder to order or an ideal
state, entropy can be thought of as a regressive path from order to disorder. Con-
ceivedas such, entropy, like teleology, takes a linear form. This concept of an
entropic line moving from order to disorder ts well the idea of a line of
continual disintegration running from, say, [Untitled] downstream to Erosive.
An assumption that Erosive is positionedon a linear temporal axis
sanctions the poems complicity in a form that has structuring andtotalising
implications for the conception of Readings work. Because the line plots out a
quantiable dierence between Erosive and[Untitled ], signs of greater
degeneration downstream reexively point to the presence of greater order
upstream, so that the visual exposure of the later poems physical constitution
can be interpretedas being indexical of a physicality which the earlier poem
embodies. Unlike in the case of the later poem, this physicality is controlled
because its surface remains concealed from the dissecting powers of the readers
gaze; the implication being that an ordered physical presence lies somewhere else
(somewhere deeper), perhaps in the regular rhythmic utterance of the poems
metrical sinews. We might also note, at this juncture, that the paradoxical
impulse of weakening/strengthening displays the characteristics of the epic
lament (a phrase Robert Potts has used to describe Readings work [Potts
1995: 5]); a form which idealises an earlier world through its broken ruins.
The resultant conception of Readings work as a totalising Entropicon
fragments shoredagainst their ruins loosens once we widen our contextual
vision to consider evidence lying at tangents to the railroad track between
Erosive and the original we assume it to desire. A reader, for example, who is
familiar with the history of the production of Last Poems, couldpoint to how the
performative gesture of adding two last poems to Last Poems, including the
presentation, in an even more last condition, of a poem that has already been
announcedas one of the last ([Untitled] becomes Erosive) involves just a
touch of irony. This deconstruction of the last, and, indeed, the lament, points
forwardto further volumes of Readings poetry, such as Work in Regress (1997)
andthe poets nal obituary Ob (1999).
Moreover, there is one other short poem bearing the name (or non-name) of
[Untitled] in Last Poems. Addto that, seven poems called[Untitled ] which
have sprung up in the volume also entitled[ Untitled](2001), appearing in print
seven years after the publication of Last Poems. Each of these untitledpoems
stakes a claim to the non-identity of the title, diluting any claims to originality
that a poem called [Untitled] may have had.
Separately, we might consider a collage from the volume [Untitled],
reproduced as Figure 6 (Reading 2001: 34). Here, amongst the grati andthe
runes, lie the phrases crackle of anhydrous and frail wisps which are
identical to phrases that appear in the poem [Untitled] from Last Poems. If we
persist in thinking of Readings work as a linear ow, then Erosive is certainly
exposedas not being a xed point at the endof a line (for surely the phrases in
the collage are also images of parts of the opening poem in Last Poems).
The evidence that the phrases in the collage are less imagic of [Untitled]
Figure 6. Collage from the volume [Untitled] (2001).
or Erosive than of the nal page of a 1988 volume Final Demands, reproduced
as Figure 7 (Reading 1996: 158), from which the phrases were photocopied,
really unxes the points of reference. Publishedsix years before [Untitled]
appearedin Last Poems, the presence of the nal page of Final Demands
completely subverts any claim the later poem still has to originality.
Another sub-version of the image is found embedded in Last Poems taking
the form of a poem entitledReiterative (Reading 1996: 263). Beginning with
the phrase [Churned out in 76, / the eroded, faded text], Reiterative
replays (andslightly distorts) sections of Nothing for Anyone a few pages
Figure 7. Final page of Final Demands (1988).
before we are presentedwith the damagedphotocopies also imaging Nothing
for Anyone (andalso, surely, Reiterative).
Reading amongst these hypertext-
ual references, it becomes increasingly dicult to be certain as to what an
image is an image of.
Readings work operates in accordance with a more complex model of
entropy than one taking a linear form. Such a model has been described by
Michel Serres, in an essay entitledThe origin of language: Biology, information
theory, & thermodynamics (1982).
Firstly, Serres abandons the nineteenth century notion of a thermodynamic
system where there occurs an irreversible drift from order to disorder or from
dierence to the dissolution or dissemination of a homogeneous mixture (Serres
1982: 72), on the basis that such a thermally insulatedsystem is dened as being
isolatedandclosed(72). Serres, instead, turns to the example of the living
system, which bears the characteristics of a thermodynamic system (operating at
high temperatures, tending towards death according to an unpredictable and
irreversible time [7374]), andyet functions as an open system in that it
receives, stores, exchanges, andgives o both energy andinformation in all
forms, from the light of the sun to the ow of matter which passes through it
(food, oxygen, heat, signals) (74).
Crucial to an understanding of this system is the conception of negentropy,
or the negation of entropy, which describes a struggle against irreversible time,
well illustratedby the Darwinian concept of mutations of selection (74) to
which Serres refers. Negentropy functions as a stabilising andord ering process
in the face of the deathly and irreversible entropic ow. If there was no negen-
tropy or stabilisation, our bodies would be unable to process the noise and chaos
of information that surrounds us.
How we understandboth time andlanguage may be conceptualisedwith
reference to negentropy. Serres writes that even though we willingly accept
the fact that the things aroundus do not all share the same temporality as in
pockets of local orders in rising entropy, and that none of these things
disturbs us, it has still seemed dicult to intuit a multitemporality (75). This
is because living bodies (perhaps unconsciously and out of necessity) work to
synchronize meanings andd irections, the continuous andd iscontinuous and
the local andthe global (75). The living organism, therefore, is a converter of
time (7576), which is why it is able to learn about systems dierentiatedby
their individual time: the world, re, and signs (76).
Similarly, living organisms can be described as apparatuses which produce
language from noise and information (79), a process likenedto a series of inter-
locking boxes which, at progressive andunconscious levels, lter out and
substract ambiguity in order to create and receive increasingly comprehensible
signals. Rather poetically, Serres describes the traditional view of the uncon-
scious as the nal black box, the clearest box for us since it has its own
language in the full sense (80). We only needto consider the allusion to an
aeroplane crash site to understand what is meant, when Serres writes that
beyondit (the black box) we plunge into the cloudof meaningless signals
(80). Finding the black box subtracts ambiguity from the wreckage.
Linearity is an unsuitable form for this model of entropy and negentropy,
for we are not talking about a greater order upstream, and a greater disorder on
the downstream side. Instead, entropy and negentropy, dissemination and
temporary stabilisation are simultaneously everywhere.
To help us picture this open system, Serres asks us to imagine a river:
It is a river that ows andyet remains stable in the continual collapse of its
banks andthe irreversible erosion of the mountains around it. One always
swims in the same river, one never sits down on the same bank. The uvial
basin is stable in its ux andthe passage of its chreodes; as a system open to
evaporation, rain, andcloud s, it always but stochastically brings back
the same water. What is slowly destroyedis the solidbasin. The uid is stable;
the solidwhich wears away is unstable (74)
As a result, Serres denes the living system as homeorrhetic (74), a term
formedfrom the Greek words homos, meaning same, and rhysis, meaning
ow. Josu V. Harari andDavidF . Bell inform us that Serres replaces the
normal term describing the equilibrium of a self-regulating system, homeo-
stasis, by homeorrhesis in order to emphasize the idea of continual movement
andexchange as opposedto the less dynamic idea of stasis (74).
Just as a rivers ow achieves relative stability through the processes of
exchange which constitute the water cycle, so too does the same language appear
to ow through Readings volumes of poetry, appearing over and over again in
new versions and expressions. Indeed, the opening of Reiterative admits as
much: [Churned out in 76, the eroded, faded text] (Reading 1996: 263). To
attempt to locate a dierence between the language of the 1976 andthe 1994
constructions is to miss the point that the repetitive assertion of entropy, the
replaying of the same tiredword s, collapses linear temporality.
At the same time, instability is to be foundin the forms which contain this
language, the banks through which ows the river of words. These banks are,
in fact, iconically alludedto by the placement of brackets around[Untitled ] in
Last Poems. I do not just refer to those surrounding the title, but also, and
pointedly, to those that enclose all nine lines of the poem. The poem is a re-
forming of the same language seen in, say, the 1988 volume Final Demands, and
in later formal congurations (some in greater or lesser appearance of instabili-
ty), such as Erosive or the collage in the 2001 volume [Untitled]. The form of
[Untitled] is, however, under imminent threat of collapse.
We can more fully complete the jump from a linear conceptualization of
Readings work to a spatial one by considering Serress description of the
organism as a barrier of braided links that leaks like a wicker basket but can
still function as a dam (Serres 1982: 75). The image of a leaking sieve (where
holes appear evenly distributed over a surface) modies the earlier image of a
river by forcing us to think of Readings work less as a sequence of upper,
middle and lower river courses, and more as a slowly eroding delta, where an
endless number of low energy channels dissipate on a relatively at surface. We
must apply the image of a delta to Readings oeuvre, which is open, constantly
dying and never completed. As Reading has said too much decays insu-
cient dies (Reading 1995: 53).
6. Conclusion
In Peter Readings body of poetry, the vertical distinction between an iconic
image andits object no longer remains. The proliferation andrepetition of
language amongst unstable forms results in images becoming objects andobjects
themselves functioning as images. This immediately brings into question the
notion of the image being derivative of its object. The slipperiness of reference
andthe collapse of linear sequence means that, on a spatial plane, images can
just as much be readto precede their object as follow it.
A consequence of this is that the disintegration of the object in an image
(e.g. [Untitled] in Erosive) points to, not just what becomes of the object, but
also what constructs it. The breakdown of Erosive, therefore, speaks of the
construction of [Untitled] from its material constituent parts and highlights its
visual presence as a nine-line icon.
John Bilstons confusion in the forewordnow has a more complex reso-
nance. Unknowingly to him, he has stumbledon the answer. Erosive is both
image and draft, de-constructing its objects materiality from both sides.
Erosive de-mythologises the conception that traditional poetic forms are
non-material andnatural expressions of orality, ideality andself-presence
(witness the traditional educational emphasis on learning such poems by heart:
a process which fuses poem and self). This illusion of naturalness depends on
the concealment of the signs that betray the material construction of the poem.
For example, the regularity of lines andthe absence of a jaggedright hand
margin draw attention away from the visual unit of the line.
Understoodas an image andd raft, Erosive works to transform our
perception of the sort of traditional poetic form which [Untitled] embodies.
The reader is encouraged to view these forms as iconic images in themselves.
For example, a reader whose eyes stumble across a form, such as a Sonnet, or
simply an image of lines arrangedin certain regular patterns (e.g. quatrains), may
recognize the poem as a Sonnet, or may otherwise experience a icker of
recognition of having seen such a visual pattern (irrespective of its metrical
properties) somewhere before. Each employment of a particular form, then,
becomes an iconic image of that form usedpreviously (whether in that text, or
others). That form, through repeateduse, becomes increasingly objectied, even
if the myth of naturalness requires such object-status to remain unacknowledged.
Objects andimages are, then, everywhere, revealedto wear an iconic value
as a result of cultural exposure even if it is assumedthat such objects are merely
arbitrarily constructedforms. Gradually, the arbitrariness of articial construction
declines through usage.
In de-mythologising more traditional poetic forms, and exposing their iconic
Figure 8. Page from a vagrants diary from Perduta Gente (1989).
status andmaterial construction, Erosive reduces such forms to an existence
locatedon the same physical plane as other collage-elements in Readings work,
which are more usually interpretedas visual, material or iconic. An example of
such an element is the piece of notepaper from a vagrants diary, reproduced as
Figure 8 (Reading 1996: 164). Erosive forces the reader to abandon any notion
that formal poetic modes mark out the central ground for the operation of
reading. The material interference aecting apparently lower forms of expres-
sion does not signal the end of reading, but the need to generate new strategies
to explain our encounters with these complicated elds.
As in a river delta, all forms in Readings work are levelled out on a writing
surface that manifests itself in the inter-dependence between concrete sites of
language on specic pages andthe spatial, open network that envelopes these
unstable andshifting contents.
The juxtaposition of all formal vehicles in Readings work, therefore, results
in a kindof iconic dialogics. Here, the Adonic jostles against newspaper cuttings
and notepaper. The experience of reading Reading becomes like an encounter
with a materially shifting world, where the eye must negotiate disparate iconic
elements, andwhere language is consistently being regressed to its minimal
This is the start of the story of the release of the iconic image from its
denitional trauma. Hereon, the image, re-inscribed with a greater freedom of
reference, works at refracting, deecting and dissipating the readers gaze across
the cruel and paradoxical social scenarios that Reading presents.
I am grateful for the support receivedfrom the Arts andHumanities Research Boardof the United
Kingdom. I would like to thank Ming-Qian Ma for his comments following a presentation of this
1. The re-appearance of the image on the modernist page represents one of the substitutive
formations (Genette 1995: 336) into which the genre of mimology wouldburst forth, when
threatenedwith its demise.
2. W. J. T.Mitchell speaks of polemical attacks against Joseph Franks 1945 essay Spatial form
in modern literature which have come from critics (Robert Weimann and Frank Kermode are
mentioned) who regard spatial form as an actual, but highly regrettable, characteristic of
modern literature and who have linkedit with anti-historical andeven fascist ideologies
(Mitchell 1980: 541). Mitchell writes that Franks basic argument is that modernist literary
works (particularly by Eliot, Pound, and Joyce) are spatial insofar as they replace history and
narrative sequence with a sense of mythic simultaneity andd isrupt the normal continuities of
English prose with disjunctive syntactic arrangements (541). Franks essay appeared in
Sewanee Review 53 (Spring, Summer, Autumn 1945) becoming revisedfor his The Widening
Gyre (New Brunswick, N. J., 1963).
3. For example, see Nnny (1999).
4. A box full of visual/textual work displayed in the exhibition entitled Erosive: Thirty-Seven
Texts, Hand-Produced by Peter Reading (1993/4) is housedat the Poetry Library, London.
5. Indeed, C. S.Peirce acknowledged as much when he noted that only an idea, in the sense of a
possibility, couldbe a true icon. Thus, it became useful to introduce the term hypoicon which
wouldaccommod ate a material image, such as a painting, in spite of it being largely
conventional in its mode of representation. Similarly, the adjective iconic became a looser, and
more successful means to express how a sign may represent its object mainly by similarity, no
matter what its mode of being (Peirce 1933: 157).
6. The notion of inter-textual miming would, I think, on further exploration, cast some doubt on
the validity of the distinction between the categories of endophoric and exophoric iconicity,
though the issue cannot be centrally addressed here.
7. This iconicizedstructure is similar to Charles Morriss conception of a total icon (Morris
1971: 422). Morris noted how part of a text or work of art sets up demands and expectations
which are met, or partially met, by other aspects of the text, which cause the interpreter to
have to perform a complex perceptual activity, passing from part to part of the art object,
responding to certain parts as signs of others, and building up a total response (and so total
object of perception) in terms of the partial responses (422). What saves Morriss total icon
from simply being another rendition of the formal structure of a work of art, or literary object,
is the emphasis on the part that iconic signs (as distinct from non-iconic signs) play in the
perceptual activity that builds up a conception of a total icon. Morris writes: In this process,
non-iconic signs play their part as in any perceptual process; what dierentiates esthetic
perception from other perceptual activities is the fact that perception is directed to value
properties which are directly embodied (though perhaps only partially embodied) in certain of
the iconic sign vehicles which form part of the total sign complex (422).
8. This linear conception of entropy, where degeneration downstreamtriggers a strengthening impulse
upstream, supports Sean OBriens interpretation that Readings political exhaustion, may
produce a desperate eloquence but leaves a huge hole where causality ought to be (OBrien
1998: 130), whilst the nger remains narrowly pointedat the perpetrators of social ills (131).
9. Which preceded Marfan (2000) and[ Untitled] (2001).
10. Nothing for Anyone was publishedin 1977. Readings reference to 76 must refer to his
writing of the poem.
Clark, D. 1994. Essential reading. Interview with Peter Reading. Whats On. 22 June
1994: 27. Citedin Martin (2000), 229.
Derrida, J. 1976. On Grammatology. Trans. Gayatri Chakravorty Spivak. Baltimore: John
Hopkins University Press.
Fischer, O. 1997. Iconicity in language andliterature. Language innovation andlanguage
change. Neuphilologische Mitteilungen 98: 6387.
Genette, G. 1995. Mimologics. Trans. T. E. Morgan. Lincoln: University of Nebraska
Huisman, R. 1998. The Written Poem. Semiotic Conventions from Old to Modern English.
London: Cassell.
Johansen, J. D. 1996. Iconicity in Literature. Semiotica 110: 3755.
Martin, I. 2000. Reading Peter Reading. Newcastle upon Tyne: Bloodaxe Books.
Mitchell, W. J. T. 1980. Spatial form in literature: Towarda general theory. Critical
Enquiry 6: 53967.
Morris, C. 1971. Writings on the General Theory of Signs. Thomas A. Sebeok (ed.). The
Hague: Mouton.
Nnny, M. 1986. Imitative form: The Modernist poem on the page. In Poetry and
Epistemology, R. Hagenbuchle and L. Skandera (eds), 213231. Regensburg: Pustet.
Nnny, M. 1999. Alphabetical letters as icons in literary texts. In Form Miming
Meaning: Iconicity in Language and Literature, M. Nnny andO. Fischer (eds),
173198. Amsterdam: John Benjamins.
Nth, W. 2001. Semiotic Foundations of Iconicity in Language and Literature. In The
Motivated Sign: Iconicity in Language and Literature 2, O. Fischer andM. Nnny
(eds), 1728. Amsterdam: John Benjamins.
OBrien, S. 1998. The Deregulated Muse. Newcastle upon Tyne: Bloodaxe Books.
Peirce, C. S. 1933. Collected Papers of Charles Sanders Peirce. Volume II: Elements of
Logic. C. Hartshorne and P. Weiss (eds). Cambridge: Harvard University Press.
Potts, R. 1995. Englands unocial laureate. Review in The Guardian, 11 August
1995: 5.
Pound, E. 1987. The Cantos. 4th Collected Edition. London: Faber and Faber.
Reading, P. 1994. Last Poems. London: Chatto and Windus.
Reading, P. 1995. Collected Poems: 1: Poems 19701984. Newcastle upon Tyne:
Bloodaxe Books.
Reading, P. 1996. Collected Poems: 2: Poems 19851996. Newcastle upon Tyne:
Bloodaxe Books.
Reading, P. 1997. Work in Regress. Newcastle upon Tyne: Bloodaxe Books.
Reading, P. 1999. Ob. Newcastle upon Tyne: Bloodaxe Books.
Reading, P. 2000. Marfan. Newcastle upon Tyne: Bloodaxe Books.
Reading, P. 2001. [Untitled]. Newcastle upon Tyne: Bloodaxe Books.
Serres, M. 1982. Hermes. Literature, Science, Philosophy. J. V. Harari andD. F. Bell
(eds). Baltimore: John Hopkins University Press.
Steiner, W. 1982. The Colors of Rhetoric. Problems in the Relation between Modern
Literature and Painting. Chicago: Chicago University Press.
</TARGET "moy">
AUTHOR "Andreas Ohme"
TITLE "Iconic representation of space andtime in Vladimir Sorokins novel The Queue (Ochered) "
WIDTH "150"
Iconic representation of space and time
in Vladimir Sorokins novel The Queue (Ochered)
Andreas Ohme
University of Jena
1. Introduction
Of the multifacetedforms of iconicity in literature, the most obvious result from
an unconventional layout.
Visually shapedpoetry is the classic example of this,
but typographic devices for generating meaning are also made use of in prose
ction. This is the case with the novel discussed here, entitled The Queue
(Ochered). It was publishedin Paris in 1985 by the Russian author Vladimir
Sorokin, who was born in 1955. Before discussing the novel itself, I would like
to make some general remarks about problems that may arise when dealing with
texts whose aesthetic eect is basedon an interplay of language on the one hand
andvisual elements on the other.
Whether verse or prose, such texts call for faithful reproduction in editions
andtranslations into other languages in order to prevent changes in meaning.
While this observation may seem rather banal from a philological point of view,
it is unfortunately not always taken into consideration in publishing practice.
Reasons for changes in the typography of a text can vary. The possibilities range
from additional re-workings by the author himself to editors or publishers
oversights or lack of access to the original. Another possibility is the intervention
by censors, which has to be taken into consideration in the twentieth century
especially in regardto texts that were published in the former socialist countries.
However, it is often dicult to ascertain why exactly such changes were made
andthe reasons remain open to speculation. It shouldbe evident however that
when analysing literary texts whose aesthetic eect is heavily dependent on
typographic structure, the version of the text shouldbe chosen very carefully.
This also applies to Vladimir Sorokins novel The Queue. With respect to the
readers of this volume, I have based my remarks on the English translation.
However, it will occasionally prove necessary to go back to the Russian original
when the translation deviates from it enough to inuence its reception.
2. The principles of Socialist Realism as literary background to The Queue
As the title already indicates, Sorokins novel depicts a situation that can be
considered indicative of the everyday Soviet experience.
Waiting for consumer
goods of all kinds, from basic foodstus to luxury items, is a phenomenon all
too familiar to the contemporary audience. In the course of the novel, however,
this phenomenon is increasingly subject to hyperbole. Contributing to this is the
length of the steadily growing queue and the subsequent waiting period of three
days, which is not even sucient for the main character to reach the sales
The length of the queue leads to absurd-seeming consequences, as
when, due to the long waiting period, goods acquired elsewhere get spoilt
when some people leave the queue for a length of time to accomplish other tasks
that also involve standing in queues, for example, the purchase of vodka (103).
In general, other queues are one of the recurring topics of conversation
among the people waiting (e.g. 7, 8, 9, 16, 103, 105). This creates the impression
that waiting is the main activity of the people in Moscow. The situation becomes
almost grotesque due to the fact that, over the course of time, the goods for
which the people wait are constantly interchanged. If at rst it is a matter of
shoes (20), later the talk is of jeans (37), andon the next day again of jackets or
coats with astrakhan collars (68), and nally of pieces of polished furniture
Thus, the waiting is no longer contingent upon a desired object, but
becomes rather an endin itself. It becomes a quasi-existential situation: The
queue turns into a place for the routines of everyday life. People eat, drink,
communicate, andform friendships. Thus, the queue becomes a surrogate for real
andthe novel itself a parody of the principles of Socialist Realism, in that
these principles are overturnedat all levels of the text.
For example, the principle calling for an artistic depiction of everyday life
results in The Queue consistently in the portrayal of the permanent shortages in
the Soviet Union. Accordingly, Vadim, the protagonist of the novel, reaches his
goal the goods not through suggestions for improved production methods,
a common theme of the so-called production novel, but rather by accidentally
meeting one of the people who organise the selling of the goods. In this regard
the parody is made even more apparent in that the traditional heroes of this type
of novel are either ridiculed like the front-line workers (37), or cause outrage like
the occupants of three buses who are allowedto cut to the front of the queue
because of certain privileges, being assumedby the waiting people to be workers
(23) or unionists (25). Another aspect of Socialist Realism made ridiculous here
is the so-calledsolid arity with the masses (massovost). This aspect, which refers
to the principle that calls for general intelligibility in literature and for its didactic
character, is satirically personied in The Queue by the character of a novelist
who has only joinedthe queue for the purpose of getting to know the masses
(100). Instead, however, he immediately leaves the queue to meet with a girl
whose acquaintance he has just made. The novelist is further discredited because
he is able to easily obtain for the girl the goods she is waiting for (100). Through
his obvious privileges, he proves himself a part of the authority responsible for
the shortage.
In addition, the literal application of other fundamentals of Socialist Realism
leads to a perversion of this ideological concept of art. An example of this is the
way the novel deals with Socialist Realisms demand for linguistic simplicity
(prostota) that originally was usedas a slogan against the art of the avant-garde.
In The Queue, this is achievedthrough a style that resembles spoken language
and through the excessive use of an especially drastic form of substandard
speech, the russkij mat, whereas Socialist Realism propagatedthe exclusive use
of standardwritten language. One last point that shouldbe mentionedin this
connection concerns the insistence of Socialist Realism on a faithful and
historically accurate presentation. This principle is also taken literally in The
Queue. In Socialist Realism, this was usually accomplishedby the use of an
omniscient narrator who not only describes the inner andouter worldof the
characters in exact detail, but also steers the reader through ideologically sound
positive and/or negative value judgements. In The Queue, on the other hand,
narrative mediacy is missing entirely, even though the work is classied as a
novel (though not in the English translation). Thus, Sorokins text proves to be
a true polylogue, whose number of characters cannot be precisely determined. In
this way, the demandfor a faithful andhistorically accurate presentation is so
literally executedin The Queue that in literary criticism the text has even been
described as hyper-realist (Vishnevskaja 1985: 176).
3. The communicative structure and the plot of The Queue
Sorokins novel lacks not only the perspective that comes with a narrator and
which acts as a point of orientation for the reader, but also the mediating system
that is crucial when it comes to dispensing information in prose ction. Hence,
the sole sources of information in The Queue are statements of the characters and
structural elements which must be attributedto the level of the impliedauthor .
In terms of typography andin terms of the way the reader receives information,
Sorokins text approximates the genre of the drama, though with the decisive
dierence that the side text is missing. This has far-reaching consequences for
the reading process. For the reader it is not only unclear when and where the
depicted events take place, but also who the individual speakers are. In fact, it is
only through the course of conversation that the reader is given clues that allow
him to gure out that the setting is Moscow in the summer of 1982.
individuals can be identied from the numerous dierent voices that take part
in the polylogue only through recurring topics of conversation, the direct address
of the interlocutor by name, or idiosyncrasies of the characters speech.
This primarily applies to the protagonist Vadim Alekseev, the female
characters he meets while waiting, Lena Troshina andLjud mila Konstantinovna
(calledLjud a for short), and to certain other characters who are, however, not
important to the development of the plot. This plot, developed through Vadims
character, can be summarizedas follows: Vadim, who is employedby a maga-
zine as a proof-reader of articles on history, joins a queue one afternoon, where
he meets Lena, a student. Along with the others they receive a number that
guarantees their position in the queue. As night slowly falls, they retire to a
nearby square, where they spendthe night on park benches. Vadim uses this
opportunity to make his rst advances to Lena. Their slumber is rudely interrupt-
edat three a.m. by the rst roll-call, during which the peoples names are called
one after another. This is repeatedin the morning. The people who are waiting
then move in groups to a self-service restaurant, where Lena makes the acquain-
tance of the aforementionednovelist. Soon after that, she leaves the queue for
good, more than likely in order to meet with this same novelist. Vadim, on the
other hand, shares a bottle of vodka with two others. After yet another roll-call
and one more bottle of vodka, Vadim lies down on a grassy patch to sleep o
his inebriation. After he has woken andhas returnedto his place in the queue, a
thunderstorm forces the waiting people to seek shelter in the nearby houses. Here
Vadim meets Ljuda, an inhabitant of one of these houses, and she invites the soaking
wet man into her apartment. They share a meal andthen also spend the night
together. When, upon waking the next morning, Vadim intends to rejoin the queue,
Ljuda reveals that she is one of the organizers of the sale and can therefore
procure the goods for him. The action ends with Vadim falling asleep again.
4. Iconic devices in The Queue
These events form the basis for the ongoing polylogue, which develops at the
end into a dialogue between Vadim and Ljuda. It is crucial that from a semantic
point of view these conversations are of little use to the reader as vehicles of
information. Except for the few passages that provide background information
about Vadim, Lena, andLjud a, the waiting periodis lled with predominantly
trivial dialogue. The conversations are for the most part situation-dependent and
concerned with the queue itself or with the immediately discernible surroundings.
People pass the time by solving crosswordpuzzles, reading aloud notices from
the newspaper, or discussing recent events from the realms of politics, sports,
andeconomy . Because of the banality of the topics, their disparity, and their
seemingly arbitrary sequence, the content of most of the remarks has no immedi-
ate function for the reader. Moreover, the individual statements again with the
exceptions of those by Vadim, Lena and Ljuda do not lend the respective
speakers any individuality. In other words, the symbolic function of the linguistic
signs is nearly reduced to insignicance. Instead the iconic quality of the signs
is foregrounded.
This iconic quality is usedprimarily to illustrate the temporal
andspatial aspects of the depictedevents.
Thus, the absence of the narrator is
partially compensatedfor in so far as the reader is given information at the level
of the impliedauthor .
By applying the term spatial, I do not refer to the setting as such, that is,
the various locations in Moscow mentionedin the characters speech itself, but
rather the spatial dimension that the queue takes within this setting. Its excessive
length, which is repeatedly referredto by the characters, is expressedgraphically
with the help of iconic devices, as the following examples will show. Because of
the intense heat, the people redirect the queue to a stand where kvas, a typical
Russian soft-drink, is being sold. This gives rise to the following polylogue
between the thirsty people andthe saleswoman:
)&, $o:\T4N. Two large.
)&,>"*P"H\ <,:@R\ *"&"6H,. Twelvehavent you got any
U"F B@4VJ " &@H &@2\<4H, Ill just have a lookah,
herethere we are
)DJ(@, *,:@. A@0":J6FH". Thats more like it. Thank you.
%"<? Next?
#@:\TJ`. One large.
G"8 R,HZD, &"T4 Therefour kopeks
?6, 8"8 $DZ2(",H J &"F Whoops, its splashing
! &Z @H@6*4H, @HF`*". %4*4H, HJH Move away from here now. Its
<@8D@ &F, all wet, look.
#@:\TJ`. One large.
),&bH\ $,D4H, &"<? Thats ninethere you are
)&, $@:\T4, 4 @*>" <":,>\8"b. Two large andone small.
AbH>"*P"H\ B@*"6H, 8DJ08J Fifteenpass me back that
@HHJ*" mug
;":,>\8J` *"6H,. A small one, please.
;":,>\8J` H"8 $,D4H, One small there you go
#@:\TJ` One large
;,:@R\ *"&"6H,, H@&"D4V4 I needchange, comrades
%@H *,FbH\. Heres ten.
Q,HZD, Four
#@:\TJ`. One large.
E@D@8 R,HZD, Forty-four (31)
(Sorokin 1985: 28f.)
This example shows that except for rare insertions, the people place their orders,
one after the other, andthe saleswoman calls out the amount to be paid or the
amount of change to be returned. This is just a small excerpt of the scene at the
kvas stand, which occupies a total of four pages (3134). Clearly, its function is
not to inform the reader about the characters or the development of the novels
plot. Instead, the constant repetition of the same types of utterances serves to
point out the length of the queue andthe repetitiveness of what happens in it.
A similar sequence with the same function occurs in the scene where the people
are in front of the cash register at the self-service restaurant (94f.).
This iconic function of the linguistic signs is even more evident in view of
the aforementionedroll-calls. The longest of these occupies seventeen pages in
the English translation (113129) andeven thirty in the Russian original (Sorokin
1985: 115144). This dierence stems from the fact that the English version is
printedin two columns at this spot, whereas the Russian version is, with two
exceptions, printedin one. During this roll-call, the people in the queue are
calledby their names, whereupon they conrm their presence with a curt yes.
The strange thing about this is that these proper names have apparently no
immediate function for the reader. With the exception of Vadim and Lena, he
does not learn anything about the identities of their bearers, nor is he able to
connect them with the various speakers in the preceding and subsequent poly-
logues. This circumstance appears especially paradoxical from a sign-theory
perspective because proper names are distinguished from other signs in that they
grant the bearer a certain degree of individuality.
In The Queue this very issue
is ironically foregrounded. In a few cases, the persons name is incorrectly listed
and is of course immediately correctedby the character concerned(124, 125,
126). Undoubtedly, it is important to the bearer of a name in the ctional world
to be calledby his correct name, not only because it belongs to his identity, but
also because it is only the correct name that guarantees his position in the queue.
For the reader, on the other hand, it is completely inconsequential whether a
character is namedBerbetov or Berberov, Trusova or Turusova, Vosk or Volk,
quite simply because they are as meaningless to the development of the plot as
the other 552 people on the list.
To repeat it once more, the recurrence of the
identical types of utterances serves primarily to iconize the ideas of length and
repetitiveness. The linear sequence of the names mimes the line of the waiting
characters. This text structure is not without consequences for the reading
process. The inverse relationship between the length of this passage andthe
amount of information containedtherein tempts the reader to leaf aheaduntil the
conversation among the waiting people begins again. Whether or not the
individual reader actually does so is irrelevant. What is important is that the
reader is in this way made conscious of the reception process itself.
To an even greater extent this is true for the iconic expression of time. In
The Queue, the length of the depicted time is, as can be deduced from the
comments of the characters, aroundforty-two hours: The action begins at about
2 p.m. (11) with the arrival of Vadim in the queue and ends at about 8 a.m. two
days later (195). As stated above, narrative mediacy in Sorokins novel is absent.
This leads to the overlapping of the internal and external communication systems
characteristic of drama (Pster 1988: 4). Thus the impression is created that in
the portrayal of the forty-two hours the depicted time and the time of the
depiction converge.
From this, the audience receives an impression of immedia-
cy as it is usually only achieved in drama (Pster 1988: 5). The reader is witness,
as it were, of a nearly uninterrupted ow of utterances.
This apparent conver-
gence of the time of the depiction and the depicted time oers the possibility of
portraying the passage of time iconically. This happens, on the one hand, in a
way analogous to the iconic depiction of the length of the queue. An example of
this is the representation of the sexual intercourse that takes place between
Vadim and Ljuda towards the end of the novel. It is depicted in a sound-
naturalistic way (Witte 1989: 163), in that only moaning noises andterms of
endearment are represented. The dots after the moans point to the physical
activities which otherwise remain unexpressed:
?6 <4:Z6 @@@N Oohdarlingooh.
M""" Haaa.
!""N Aaah.
M""" Haaa.
!""N Aaah..
M""" Haaa
!""N Aaah..
M""! Haa!
!""N 8@H, >@ R,8 Aaah.babybaby.
M""" Haaa. (179)
(Sorokin 1985: 201)
The portrayal of this sexual act takes up eighteen pages (174191) andis only
occasionally interruptedby longer passages of real dialogue. Thus once again,
the symbolic function of the linguistic signs retreats into the backgroundin
favour of the iconic, for neither the moaning noises nor the terms of endearment
have any immediate informational value for the reader. This should be qualied
by adding that the reader receives at least one other piece of information,
something that concerns the increasing intensity of feeling during the sexual
intercourse. This is once again iconically conveyedby the growing length of the
moaning noises (181 and189f.). Another conspicuous phenomenon should be
mentionedin this context. On pages 176 and180, the linear development of the
text is interrupted by the addition of a second, upside-down column of utteranc-
es. Whether or not one can interpret this as the iconic illustration of a certain
sexual practice is open to debate. The fact remains that such a layout calls for
further explanation. Nevertheless, the amount of information in these passages,
which primarily serve to illustrate the duration of the sexual act, is comparatively
low, so that again the reader might be tempted to leaf them through.
Along with this foregrounding of the iconic quality of the linguistic signs
comes another eect that is important in the context of Sorokins parodic
reference to the fundamentals of Socialist Realism. As mentioned above, one of
the characteristics of Socialist Realism is its didactic character that, amongst
other things, is achievedby the use of a special type of hero, the so-called
positive hero, that is, a character with the correct ideological consciousness that
shouldserve as a model of the new (socialist) man. In The Queue, however, we
nearly learn nothing about the characters attitude towards life. Instead, they are
shown while they are performing their daily routines. Thus, they lose their
individuality, as the example of the roll-calls makes evident. Even the main
characters are reduced to their physical activities, as is shown by the representa-
tion of the apparently endless sexual intercourse between Vadim and Ljuda.
Connectedwith this portrayal of love-making is another technique that
expresses the passage of time in an iconic manner andpersists throughout the
entire novel. As Roman Ingarden (1973: 237) observed, time in a literary text is
generally not experiencedas such, but rather through the activities with which it
is lled. The Queue represents an exception to this rule, because the convergence
of the time of the depiction and the depicted time results in the typographical
marking of even the pauses in speech. Examples of this can be foundin those
passages in which the characters are for whatever reason preventedfrom
speaking, for instance because they are eating (96f. and167f.), drinking (101,
107, 134, 167f.), kissing (174f. and186), or licking (186f.). In all of these cases,
the characters silence is indicated by the absence of speech.
Instead, the
utterances, which are always introducedwith a dash andin this way assignedto
the individual characters, are here lled with a varying number of dots, which in
turn are now andthen interruptedby single words, so that the lengths of the
pauses within the individual utterances are made clear.
One could, to borrow a
term from Jurij Lotman (1977: 95), describe this absence of speech as a minus-
device, or zero sign from a linguistic point of view (Sebeok 1976: 118), since
meaning is produced in these cases through the absence of linguistic signs. Of
course, it is not the individual events as such that are represented here eating,
drinking, kissing, etc., which are either directly referred to in the characters
speech or inferable from the context but rather the time in which these
activities are performed. In other words, what is illustrated here is the duration
of these activities.
This technique is used almost to excess to indicate the duration of sleep,
which occurs three times. First, when Vadim and Lena spend the night on park
benches, next, when Vadim sleeps o his inebriation, and nally when Vadim
falls asleep in Ljudas bed. In the rst instance, the reader is alerted to the fact
that the characters are about to fall asleep by the lengthening of pauses between
the utterances, which is made obvious by the gradually increasing empty space
in the text (5358). The complete absence of speech, which in the English
translation is extended over several blank pages, iconizes the duration of the
sleep itself (5964). In these passages the English version of the novel diers
once again from the original in that the number of these pages is much smaller
than in the Russian text (Sorokin 1985: 5666). The same is true of the two other
passages of this type. While in the translation they only take up two andthree
pages respectively (139f. and192194), in the original there are ten andtwenty
of such pages (Sorokin 1985: 154163 and215234). Here the aforementioned
reduction of the characters becomes most obvious since because of this absence
of speech they almost seem to vanish out of the text. Andhere at the latest the
reader will actually leaf ahead because of the extreme reduction of the informa-
tional content of these passages.
Speaking of empty pages is not entirely true with reference to the Russian
original. While in the English translation these pages are actually white, in the
original they are coveredwith amorphous grey-black prints. These vary in their
density and their size and serve in this way to iconize the depth of slumber.
While the previous examples are to be attributedto diagrammatical forms of
iconicity, here we are dealing with non-verbal images.
In the end, however, it
is of secondary importance whether the pages are white, as in the English transla-
tion, or coveredwith amorphous prints, as in the original, because both types of
textual arrangement serve the same function. Like the iconic devices discussed
above, they direct the attention of the reader to his own reception process.
If The Queue, with its absence of narrative mediacy, represents a massive
breach of one of the central conventions of prose ction, this is no less true for
the iconic devices which are a corollary of this absence. As the nal examples
show, they break with the tradition of the typography that exhibits no signicant
gaps andin the Russian original even with the tradition of the exclusive use of
verbal signs in literature, namely that the pages of a narrative are coveredwith
text from top to bottom andfrom left to right.
Bearing this in mind, one could
even go so far as to call Sorokins entire novel iconic. For the utterances, which
are with few exceptions left-boundandone-lined , form, typographically, a textual
queue that runs through the entire book, just as the characters in the ctional
worldform a queue in the streets of Moscow. This eect becomes most obvious
in the internet-publication of the novel (see:
texts/ochered.shtml). Here, the reader is confronted with a seemingly endless
vertical ow of text lines, which is neither interruptedby empty space nor by the
horizontal process of leang.
5. Conclusion
The starting point for the analysis of The Queue presentedhere was Sorokins
parodic reference to the fundamentals of Socialist Realism, which is found at all
levels of the text from style to theme. Moreover, this parody is extended to the
level of the layout. One of the basic principles of Socialist Realism is the
creation of a self-contained ctional worldwhich the reader is, as it were, invited
to enter. Thus, the works of Socialist Realism belong to the long tradition of
literature that creates an aesthetic illusion.
The iconic devices used in The
Queue aim at the exact opposite eect. They expose the reader to his own
reception process and, in the end, to the materiality of the art object.
On the other hand, the characteristic of the iconic sign is that the signier
is motivatedin its relation to the signied. Hence, Thomas A. Sebeok (1976: 128)
postulates a connection between iconicity andthe Platonic concept of mimesis.
In The Queue the mimetic eect arises from the analogy between the layout of
the text andthe topic of the novel. Thus, The Queue represents a paradox
concerning the means of portraying the worldin literature. The foregrounding of
the iconic aspect of the signs andthe use of non-verbal images contribute to
breaking up the conventions of illusion formation that is crucial for Socialist
Realism, but at the same time the very use of iconic devices makes Sorokins
novel appear more realist than the texts of Socialist Realism ever could.
1. A short description of the dierent forms of iconicity can be foundin Nnny andFischer
1999: xxixxvi.
2. Georg Witte (1989: 156) speaks of one of the specic chronotopes of Soviet culture.
3. The depicted time amounts to only a little less than two days, but a comment from one of the
characters indicates that she was already in the queue the day before (Sorokin 1988: 8). Further
references to the English translation (Sorokin 1988) will be citedsimply by page number.
4. For example, when meat begins to drip because of the heat (17) or Lenas carnations wilt (44),
causing her to throw them away (75).
5. It is simply not true that, as Ulrich Schmidclaims, the characters do not know what they are
waiting for (Schmid2000: 214). Rather, they are well aware of the multiple exchanges of the
various goods, as the conversations about the types of product, their quality, and their country
of origin show (20, 44, 68, 109). Furthermore, the type of product is conrmed by eye
witnesses (20, 68, 109) although their claims cannot be proved. But the grotesque eect
basically arises from the fact that the characters wait even though the goods are interchanged.
6. Typically, the sense of their own conduct proceeds not from activity, but from passivity. It is
because the queue grows longer that it seems to the people that they are getting closer to their
goal (see the relevant comment from the protagonist Vadim, 22). However, this applies only in
relation to their position in the queue comparedto that of the new arrivals, it does not aect
the distance between them and the sales counter.
7. It is true that the doctrines of Socialist Realism had undergone several changes since their
sanctioning at the 1st All-Union Congress of Soviet Writers in 1934, andalso a relaxation
during the late 50s, 60s and70 s. However, Sorokins novel representedsuch a massive
breach of them that despite the increasing liberalization in the middle of the 80s it still could
only be publishedabroad . Hans Gnther (1984: 1854) oers a survey of the principles of
Socialist Realism. A detailed documentation of the debates at the 1st All-Union Congress of
Soviet Writers was presentedby H.-J. Schmitt and G. Schramm in 1974. For a survey of the
literary styles in twentieth century Russian literature, see Veldhues 1997.
8. While it is impossible to establish the exact dates of the novels action, there are various hints
that allow the determination of a relatively small time frame. The events must take place
between July 5th andNovember 10th, 1982. On the 5th of July, Brazil playedItaly in the
WorldCup Soccer tournament in Spain, andthe game is mentionedin a conversation by the
people waiting in the queue (27). They also speak of LeonidBrezhnev as the current head of
state, andhe diedon the 10th of November, 1982. However, Brezhnev is not mentionedby
name in the English translation. Here it only refers to Gen Sec (General Secretary) (91). But
this can also be regarded as an indirect reference to Brezhnev, since this term was reintroduced
by him in 1966. Of course it also couldrefer to one of his successors andthus it is ambiguous.
In the Russian original, on the other hand, the name of Brezhnev is explicitly used (Sorokin
1985: 93). Additionally, because the unbearable heat is one of the recurring topics of conversa-
tion, it can probably be assumedthat the action occurs in July or August of 1982.
9. For instance, Vadims repeated Tara-ra-tara-ram while singing the melody of a song.
10. Thomas A. Sebeok has repeatedly pointed out that the classication of signs is not based on
their absolute function, but on sign aspects that exist in certain contexts in a hierarchical
relationship to one another (Sebeok 1976: 41 and120).
11. One iconic device in The Queue that does not play a role in this connection should be
mentioned, however. It concerns the repeated announcements of a police ocer with a
megaphone (24., 72f. and1 11f.). In contrast to the usual utterances of the characters, these
announcements are consistently printedin capital letters, by which their loudness is typographi-
cally portrayed.
12. Naturally this spatial aspect is connectedto the temporal, as the dimensions of space and time
always are, andwhich Bakhtin aptly termedthe chronotopos.
13. This should be qualied by adding that this applies primarily to the surnames and not to the
same extent to the rst names. See Sebeok 1976: 138f.
14. Here Vadim and Lena once again are exceptions. Of course, this list of names calls for further
explanation. I wouldlike to mention only two aspects here. One function is to indicate the
multitude of nations living in the Soviet Union in that the list contains, among others, Russian,
Ukrainian, Armenian, Georgian, German andJewish names. Other names serve to arouse a
comical eect. This applies to such names that in the Russian context soundexotic and refer
to famous composers or artists like Britten (116), Dvorzhak (118), andGropius (118), as well
as to names that remindof persons that were of importance for the ideology andthe politics of
the Soviet Union like Dzherzhin (115), Maks (115) andLein (117).
15. See also Witte 1989: 158. However, an absolute alignment between the depictedtime andthe time
of the depiction in literary texts is fundamentally impossible, as Roman Ingarden has pointed
out (Ingarden 1973: 237). It can only succeed in the stage-play, where events occupy the time
between the utterances andthe action is determinedalso by the use of non-verbal codes.
16. The necessity of this qualication will become apparent later in the argument.
17. A similar phenomenon can be foundd uring the roll-call, where the missing replies of those
characters who have already left the queue are rendered typographically with empty lines,
enabling the resulting silence to be iconically illustrated(65 . and1 13.).
18. In one such situation, this type of text structure is even motivated. At Vadims request in the
self-service restaurant to tell him about Hungary, Lena answers with the maxim, Kogda ja em,
ja glukh i nem (Sorokin 1985: 98) [When I eat my lips are sealed. (96)]. This saying,
which refers to a certain manner of conduct at the table, is here typographically transposed in
the absence of speech.
19. The truth of this statement is admittedly relative, since what is discerned from the pauses in
speech is the fact that the aforementionedactivities are performed, not the activities respective
lengths. Therefore, it is highly speculative if, as Georg Witte (1989: 158) claims with regardto
these phases of silence, they take up as much space as the corresponding duration of the
utterances would demand.
20. For a dierentiation of diagram and image, compare Nnny 1986: 199. It is however unclear
whether this form of text structure goes back to Sorokin himself or to the illustrators A. andM.
Gran, who furnishedthe book with silhouettes of the queue. These amorphous prints are at any
rate once again missing in the Russian edition from 1998 (Sorokin 1998: 261406). The number
of empty pages is also extremely reduced here, making it seem logical to refer back to the
original edition from 1985, even more so because the new edition contains no editorial remarks
that couldprovid e information about which version of the text it was basedon.
21. In connection with this, see also McHale 1993: 181184.
22. More information about illusion formation andillusion breakthrough can be foundin Wolf
Gnther, H. 1984. Die Verstaatlichung der Literatur: Entstehung und Funktionsweise des
sozialistisch-realistischen Kanons in der sowjetischen Literatur der 30er Jahre.
Stuttgart: J. B. Metzler.
Ingarden, R. 1973. The Literary Work of Art: An Investigation on the Borderlines of
Ontology, Logic, and Theory of Literature. Evanston: Northwestern University Press.
Lotman, J. 1977. The Structure of the Artistic Text. Ann Arbor: University of Michigan.
McHale, B. 1993. Postmodernist Fiction. London and New York: Routledge.
Nnny, M. 1986. Iconicity in Literature. Word and Image 2: 199208.
Nnny, M. andO. Fischer (eds) 1999. Form Miming Meaning: Iconicity in Language and
Literature. Amsterdam and Philadelphia: John Benjamins.
Pster, M. 1988. The Theory and Analysis of Drama. Cambridge et al.: Cambridge
University Press.
Schmid, U. 2000. Flowers of Evil: The Poetics of Monstrosity in Contemporary Russian
Literature (Erofeev, Mamleev, Sokolov, Sorokin). Russian Literature 68: 205222.
Schmitt, H.-J. andG. Schramm(eds) 1974. Sozialistische Realismuskonzeptionen: Dokumente
zum 1. Allunionskongre der Sowjetschriftsteller. Frankfurt/M.: Suhrkamp.
Sebeok, T. A. 1976. Contributions to the Doctrine of Signs. Bloomington andLisse:
Indiana University and Peter de Ridder.
Sorokin, V. 1985. Ochered. Paris: Sintaksis.
Sorokin, V. 1988. The Queue. New York and London: Readers International.
Sorokin, V. 1998. Sobranie sochinenij v dvuch tomach (Tom 1). Moskva: AdMar ginem.
Veldhues, Ch. 1997. Zur Evolution der russischen Literatur im 20. Jahrhundert.
Zeitschrift fr slavische Philologie 56 (1): 130.
Vishnevskaja, Ju. 1985. Giperrealizm povsednevoj zhizni. Sintaksis 14: 175178.
Witte, G. 1989. Appell Spiel Ritual: Textpraktiken in der russischen Literatur der
sechziger bis achziger Jahre. Wiesbaden: Otto Harrasowitz.
Wolf, W. 1993. sthetische Illusion und Illusionsdurchbrechung in der Erzhlkunst:
Theorie und Geschichte mit Schwerpunkt auf englischemillusionsstrenden Erzhlen.
Tbingen: Max Niemeyer.
</TARGET "ohm">
AUTHOR "Matthias Bauer"
TITLE "Vision andPrayer"
WIDTH "150"
Vision and Prayer
Dylan Thomas andthe Power of X
Matthias Bauer
Universitt des Saarlandes
1. An iconic poem?
Dylan Thomass Vision andPrayer seems to be the most obviously visual and
iconic of his poems if iconic is taken in one of its most simple senses, that is,
the visual shape of a text imitating a particular meaning (see e.g. Fischer and
Nnny 1999: xvixvii, xxii).
Its clear-cut outer form is without parallel among
Thomass poetry, the only possible exception being the somewhat unusually
shapedpoem Now (Garlick 1973: 42). All commentators of Vision and
Prayer mention the baroque pattern poem andin particular George Herberts
Easter Wings, which seems to be the model at least for the second part of
Vision andPrayer. Thus Howard Seargent (1962: 65) points out that it is so
reminiscent of Herbert in its hourglass form andJames A. Davies (1998: 193)
observes that the hourglass shape of the stanzas in the secondpart stresses time
already ebbing away. But is the shape of Thomass poem merely a repetition of
Herberts? In the rst place, Herberts poem is not shapedlike an hour-glass, at
least not in the rst edition of The Temple, in which Easter-Wings, following
the model of Simias of Rhodess Pterygion,
really looks like a pair of wings.
Accordingly, if there is a relationship to Herberts poem it is, quite literally,
an oblique one; the main dierence being not just a rotation of 90 degrees but,
more importantly, the fact that whereas Herberts typographical arrangement
exactly represents what his title announces, there seems to be no such straightfor-
wardrelationship between title andpictorial form in Dylan Thomas. To put it
dierently: what do the remarkable shapes of Vision and Prayer represent if
spontaneous reaction as well as literary tradition lead us to assume that they are
not arbitrary but iconic? There are references to wings in the poem (the
wingedwall in line i.29, the Cyclone of his wing in i.56 and the wings of
the children in i.91) as well as to time (To the burn and turn of time i.11) but
all these references are made in the rst part, which is not wing-shaped or
hourglass-shapedat all but has the form of rhombuses or lozenges.
Perhaps the assumption of a pictorial signicance is wrong after all
(Garlick 1973: 43) or the shape is just vaguely suggestive of a meaning, as
Moynihan (1968: 14) seems to think, who speculates that the rst part suggests
a movement from nothingness to nothingness yet, because it is diamond-shaped,
simultaneously conveys the feeling of richness andvalue, or Korg (1992: 120),
to whom the shape of the stanzas in Part I seems to reect the idea of opening
that prevails in it, both in relation to birth andto spiritual awakening. Similar
impressions are given for the shape of Part II.
As an alternative to such negative
or mainly associative responses, however, one might consider the shape of Vision
andPrayer as the iconic image of a basic geometric structure which in itself serves
as a sign as well as a symbol pointing to a number of interrelatedmeanings.
2. Pyramidal rays
Now let us look at Vision andPrayer and this is meant quite literally. A
poem whose rst title wordis Vision andwhich has such a remarkable shape
is, I suggest, to be looked at before it is to be read. In other words, close reading
here in the rst place means close looking, assuming the role of seer for a
moment, deliberately unfocussing ones eyes so that one does not decipher but
become aware of the dark shapes against the background of white paper. We
thus realize that the one shape is the negative of the other: in the manner of
tilting images we see, in Part I, either black rhombs or white triangles (more
precisely tricquets displayed as George Puttenham has it, quoted by Thomass
editor Daniel Jones [1985a: 273]) and in Part II black triangles or white rhombs.
The pattern thus revealedis basedon the elementary geometrical form of the
letter X, which provides the outline of both shapes.
With a pattern based on the interplay of foregroundandbackground , it may
be permittedto think of a background even at this initial stage of approaching
Vision andPrayer. To a reader of seventeenth-century English literature, the
X-shapedpoem, whose title links the sense of vision andthe appeal to God,
inevitably recalls one author whom Dylan Thomas is known to have read(Maud
1968: 12 andn. 14) andwho was fascinatedby this geometrical pattern: Sir
Thomas Browne, who in The Garden of Cyrus regards what he calls the
quincunx (St. Andrews cross) as a basic pattern of creation, giving evidence
of the fact that nature is the Art of God (Religio Medici i.16; Browne
1964: 16). What is more, human perception of the worldis itself an example of
this geometrical structure. It is, as Sir Thomas Browne puts it,
gratefull to the Eye: For all things are seen Quincuncially; For at the eye the
Pyramidal rayes from the object, receive a decussation, and so strike a second
base upon the Retina or hinder coat, the proper organ of Vision; wherein the
pictures from objects are represented, answerable to the paper, or wall in the
dark chamber; (Browne 1964: 167)
This reference to the dark chamber of the camera obscura is further elaborated
by Browns reference to Bovillus, who maintained(in De intellectu) that the
inner or intellectual reection also takes place in the shape of a double pyramid
with the understanding acting as a lens or focal point between the res in mundo
andthe res in memoria:
Browne further refers to Egyptian philosophy in which the geniall spirits of the
divine and the human world do trace their way in ascending and descending
Pyramids, mystically apprehended in the letter X and the open Bill and straddling
Legges of a Stork, which was imitatedby that Character; this goes together with
the myth of creation in Platos Timaeus (37B), in which the rst creation andthe
unfolding of the spheres is described as a process to be outlined by the letter X,
as the world-soul is divided cross-wise. George Herbert refers to this X-shape of
creation in Prayer (II), when he speaks of Gods great arm, which spans the
east andwest, / And tacks the centre to the sphere! (89). The X implicitly
appears in the hour-glass image of the following line as well: By it do all things
live their measurdhoure (10).
Switching from backgroundto foregroundagain, we now begin to read (or
to readmore than just the title of Thomass poem) in order to see whether the
suggested concept of the double pyramid is borne out by the words and images
of Vision andPrayer. As the opening lines make clear, its rst subject is the
birth of an unknown person: Who/Are you / Who is born / In the next room/
When we think of the X as the shape of a stork the theme of a childs arrival
may seem quite appropriate. With Dylan Thomass quaint sense of humour and
predilection for birdimagery this is not as far fetchedas it may appear, especial-
ly since in folk legend, the stork also appears in connection with Christs cross
andas a tyrant that devours his subject (Evans 1981: 107475). The link
between the sun andthe bird in stanza 4 (In/The spin / Of the sun / In the
spuming / Cyclone of his wing) makes the Egyptian connotations of the
hieroglyphical X on which the poems shape is basedappear quite likely.
syncretism is very much like Sir Thomas Brownes, whose double pyramid not
only corresponds to the shape of Vision and Prayer but also coincides with
Dylan Thomass general interest in that Egyptian geometrical symbol, manifest-
ing itself, for example, in My Worldis Pyramid (no. 95) which, although
it is not a pattern poem, is very much concernedwith symmetrical shapes, the
fellow halves (II, l. 19). Just as My Worldis Pyramid combines references
to Egypt andthe Orient with an English valley, the land of the Bible (the
crossing Jordan), the Arctic, andthe South, there is in Vision andPrayer
a syncretistic link between the Christian belief in the divine childandthe Egyptian
belief in the divine sun-bird. The blending is of course traditional, for the proto-
typical Egyptian sun-birdis the Phoenix (being worshipped at Heliopolis, Suntown
in Egypt), which has always been regarded as a type of Christ. Like Shakespeares
The Phoenix and Turtle, Vision andPrayer includes a threnos or burial song of
birds: at the beginning of Part II, the burial song/ Of the birds of burden becomes
part of a ghostly mock-resurrection, for Thomass birds are bearing / The ghost /
From/ The ground/ Like pollen / On the black plume / And the beak of slime
(ii.713). Shakespeares poem draws attention to the iconic expressiveness of the
letter X, too, when, for example, paradoxical contraction goes together with the
noticeable use of the letter in Distance, andno space was seen / Twixt this
Turtle andhis queen (3031, my emphasis). The letter is of course also fore-
groundedby the repeatedreference to the phoenix itself (23, 35, 50 56).
3. Mystical geometry
The emphasis on seeing (Distance, andno space was seen) reinforces the
connection with the letter X as the structural principle of Vision andPrayer .
The outline of Thomass poem is both an example of the Vision or theoria to
be practised(we have to look at this poem as well as to readit) andan iconic
representation of the visual process itself with its double pyramid of rays
focussedby the lens of the physical or mental eye andleaving a picture on the
Thus, although the rst sense impression in the poem is hearing (So
loud in l. 4, referring to the mothers cry of pain in childbirth; cf. the moan /
Of the mother in stanza 2), the speaker nds himself indeed in a dark cham-
ber into which light streams only when in stanza 2 the wingedwall is torn / By
his [i.e. the new borns] torridcrown / Andthe dark thrown / From his loin / To
bright / Light (2934). It is impossible to follow up all the implications of this
image here but a particularly striking one has to do with the fact that it is the
new borns loin (the centre of his body) from which darkness disappears,
surely with biblical overtones of procreation (e.g. Gen 35:11 Andkings shall
come out of thy loins or Acts 2:30, taking up 1 Kings 8:19, in which the
resurrectedChrist is calledthe fruit of his [Davids] loins). The juxtaposition
of loin andlight alludes to the loins girdedabout andthe lights
burning of those who expect the arrival of the Lord(Lk 12:35), while loin,
together with the headof pain andthorn andcrown is evocative of the
loincloth of the crucied Christ. In Thomass vision, the birth of the childby the
mothering maiden (i.42), The adored / Infant light (ii.2829) born For / All
men (ii.2627) coincides with his death on the Cross. The beams of light which
nally, in the last stanza of the poem, are so bright that the speaker is lost in
the blinding/One enter the I or eye in a crosswise manner. For the paradoxi-
cal, mystical experience of perceiving or being lost in the deluging / Light
(ii.5556) means, to put it in the words of St. Paul, to be crucied with Christ.
The lightning that nally answers the speakers cry has indeed the eect of
an X-ray in that it makes hiding impossible (I wouldturn back andrun / To the
hidden land). The comparison is warranted by the fact that the place of hiding
for which the speaker prays is inside the body. He desires to return to the
mothers womb, To the birth bloody room (ii.21), a wish that is nally
overcome by the speakers letting himself be discovered, found by the loud
sun (ii.92).
The once hooded room (i.47) is replaced with the cauldron / Of
his / Kiss (i.4951), a vessel which corresponds to the shrine / Of his blazing /
Breast (i.7173) and the worlds wound (ii.99) as well as the high noon / Of
his wound which Blinds my / Cry (i.6568). This reference to the Crucixion
(Garlick 1973: 46; Kidder 1973: 161) synaesthetically combines the penetrating
beams of light with the re andheat of the mouth by which the speaker is
touched(a bonre in / His mouth; i.4344).
The mouth is unitedwith the eye in this excruciating experience. Thus from
the context of Thomass writings it becomes evident that to him kiss and cross
belong together. When one leafs through his Collected Letters, for example, one
soon realizes that he fanciedthe letter X as a sign for kiss at least as much as
Willy Nilly does in Under Milk Wood (Thomas 1985b: 43). This is of course a
widespread convention, but the conspicuous way in which Thomas uses the letter
emphasizes his awareness of its potential expressiveness. This is, for example, how
Dylan Thomas signedone of his rst letters to his wife Caitlin (Thomas 1987: 248).
Dylan X Caitlin
In Thomass poetry, too, the kiss is visualizedas a point of contact, the centre of
crossing lines: notice the position of the wordkiss (anagram of iks) at the
endof the third stanza right at the centre of the cross formedby the outlines of
stanzas 3 and4 together; or notice such a phrase as O see the poles are kissing
as they cross in the poem I see the Boys of Summer (no. 86, l. III.6). Here
the crossing lines of vision (O see) go together with the kissing-crossing.
Vincent Leitch (1992: 341) noticedthe similarity to Herberts poem The
Search: East andW est touch, the poles do kisse, / And parallels meet. In the
geometrical mysticism to which Herberts and Thomass lines are indebted,
parallels meet andcross when they form the innite sphere which is God himself
(Mahnke 1937: passim, e.g. 2021, 8486, 173).
This mystical notion, which for
Thomas characteristically includes the experience of sexual encounter as a
personal cross(ing) of cosmic dimensions, also appears in A Prospect of the
Sea (1937), which is a story about vision (as the title indicates) as well as about
the kissing of a boy and girl andabout procreation andgenesis (the girl will
have a baby on every hill). The geometrical hieroglyphics or double pyramids
of Vision andPrayer are not only foreshadowed by the girls sister living in
a pyramid but also by the boy, who realizes that in this encounter his own space
has become no wider (and, implicitly, no smaller) than Eden or the loving
room of the world, and[that] the two poles kissedbehindhis shoulder blades
(Thomas 2000b: 93). In one of his early letters to Pamela HansfordJohnson,
Dylan Thomas testies to the formative inuence of this geometrical mysticism
when he tells her: I lie in the dark andthink. I think of GodandDeath and
Triangles (Thomas 1987: 129). Powers of overwhelming impact are imaginedin
terms of geometrical shapes, just as, indicative of a natural mysticism that
corresponds to the geometrical one, the smallest kind of these shapes is seen to
be inhabitedby a divine person. Thus in the same letter Dylan Thomas main-
tains: The chromosomes, the colour bodies have a god in them that doesnt care
a damn for the howls of our brains. It is of course the letter X (together with
the similarly triangular Y) that, because of its shape, gave its name to the sex
In Vision andPrayer the erotic meeting markedby kiss andprocreation
and the encounter with the blinding light of the sun or son, which is a death
pregured by the death on the Cross, are shown to coalesce by the visible cipher
of the letter X.
We are reminded by this of traditional constructions of this
letter as the shape of a nakedman spanning the globe or the four corners of the
earth, whose centre is his navel. An example is the combination of O andX
(and, implicitly, the rhomb and the X) in Geofroy Torys Champ Fleury (Tory
1973: fol. 18v):
How exactly the shape of Thomass stanzas ts those X-shapedhuman letter-
bodies can be seen in an example from Agrippa of Nettesheims De occulta
philosophia (Agrippa 1987: 284):
Like the poles that kiss as they cross, this two-dimensional image on the page is
to be visualizedthree-d imensionally as well. Semantically, the spatial or spheri-
cal dimension of the lines comes to the fore in expressions like Cyclone of his
wing (i.56) and O spiral of ascension (i.79), which add, as it were, an O to
the X (or a circle to the cross, as in Donnes The Crosse, see note 9). The X
marks a globe or cell or, to return to the image of the beam of light an
interior dark chamber which opens up and unites with another in a process full
of pain andterror . The emphasis on space becomes obvious right at the begin-
ning: Who/Are you / Who is born / In the next room/ So loudto my own ?
This unknown person X is born in a room which not only blends with the space of
the womb (whose wren-bone walls open in childbirth) but also with the space of
the poem itself: the characteristically shapedstanza is the room (stanza means
room) in which the childis born, it is a room that is punningly visualizedas a
rhomb, andeven the mysterious wren ts with this dark chamber (the wall thin
as a wrens bone, i.9), for the wren is not just any birdbut one whose zoologi-
cal name is troglodytes, which means cave-dweller (cf. OED troglodyte; wren
bone is of course also an anagram of new born[e]).
The birth bloody room of the womb and the room that is the space of the
poem itself, are connectedby a chain of words and images that serve to integrate
the secondhalf of the poems title into the iconic conception of the whole. The
once hoodedroom from which the speaker runs (andto which he later wishes
to return), is the space of child-hood as well as the womb in which the embryo
is coveredby a caul (the amnion). Thomas punningly points this out by juxtapos-
ing the once hooded room with the cauldron/Of his kiss, i.e. of him who has
a bonre in / His mouth (i.4344). Similarly, in From Loves First Fever to
Her Plague (no. 74), the moment of birth is that of the scissoredcaul (3). To
be born with a caul, as we know from the beginning of David Coppereld,
means to be protectedfrom drowning, but in Vision andPrayer the caul is
scissored, too (the wingedwall is torn / By his torridcrown), andthe speaker
will inevitably drown in the worlds wound of the son. The allusion to the pair
of scissors is of course yet another reference to the X-shape of the poem itself
(andvice versa).
4. The name of X
These new spaces or rooms, the mouth, the shrine / Of his blazing / Breast
(i.7173), the exhaling tomb (i.76, which, in the context, is an X-hailing
tomb as well), then the shrine of the wound(ii.37), later the hollow of the
handin which the speakers voice burns (ii.101) form the counterpart to the
hidden land of the mothers womb. To move inexorably from the one to the
other means proceeding from the speechlessness of infancy to the nal speech-
lessness of the Infant light (ii.29; with reference to infans, speechless), in
which the sun roars at the prayers end. In between lies the space or room of
language which cannot be avoided by returning Before the lips blaze and
bloom/ To the birth bloody room (ii.2021). It is the space of the poem itself,
in which the speaker, invoking another mouth-cave-image, tolls the tongue of the
sleepers (ii.54).
Or, to compare once more From Loves First Fever: And
from the rst declension of the esh / I learnt mans tongue, to twist the shapes
of thoughts / Into the stony idiom of the brain (3133).
There are, however, dierences as well as similarities between the two
poems, for the predominantly negative view of language has been replaced in
Vision andPrayer by a more hopeful or at least ambivalent one. In From
Loves First Fever, the speaker disparagingly (and desperately) exclaims: The
root of tongues ends in a spentout cancer / That but a name, where maggots have
their X (3738). In Vision andPrayer, the X or cross as the central pattern or
cipher is not merely a sign of death and decay. It is a cipher that turns into an
eective word, not but a name but the name (i.28) that the turbulent new
born burns into the speaker, the mark by which the nding one relentlessly
claims him his own. In From Loves First Fever the X, as Mayer (1995: 41)
has remarked, recalls the signature of the illiterate (who has, as it were, no
name). By contrast, in Vision andPrayer, it points, as we think of the letter
Chi (Jones 1966: 81) for Christ (as in X-mas), to the name above every name
(Phil. 2:9), the root andendof all naming.
In a letter to his American publisher (James Laughlin), Dylan Thomas,
returning the proofs of Vision andPrayer, insisted that the shape hadto be
absolutely symmetrical with no variations in the straight diamond lines and
their complete reversal in Part II (Thomas 1987: 54243).
Thomas thus
obviously thought of the shape in terms of a diamond. In the context of linguistic
self-reection, the precious stone is the emblem of a language quite dierent
from the stony idiom of the brain; it is, moreover, to be contrasted with the
speaker lying still as stone or praying in the name of the stone / Blind
(ii.3940), who want to go on sleeping In the dark / And deep / Rock (ii.4143).
With the name of the childbeing burnt into the speaker he becomes, in the sixth
stanza of Part I, like upright Adam who Sang upon origin (i.8990). The
diamond, whose mineralogical name is adamas, is appropriate to this rst man
praising creation by giving names to it. Furthermore, it is a traditional symbol as
well as a name of Christ, whom GerardManley Hopkins in That Nature is a
Heraclitean Fire called immortal diamond (l. 24; Tindall 1996: 239).
Vision andPrayer begins by asking someone who he is. The name is
never explicitly mentionedbut burnt into the speaker who, as Wardi
(19992000: 193) recently pointedout, by saying that in the name / Of no one /
Now or / No / One to/Be I pray (ii.7580) as it were inadvertently prays to /
Be. Andwhen he says I / Am found, the speaker actually pronounces the
name that has been burnt into him, I AM (Ex. 3:1314). Similarly, in the last
lines of the poem: Now I am lost in the blinding/One which also means:
Now I am, lost in the blinding one, i.e. having given myself up and having
become one with him, I truly am. To Hopkins, diamond (24) rhymes with I
am and, meaning I am all at once what Christ is (22; Leimberg 1998: 113).
When the speaker prays In the name of the fatherless (ii.70) he evokes, in the
very negation, the name of the father. Andwhen, after amen (ii.86) he
turn[s] the corner of prayer (ii.87) the letters themselves are turnedinto
name (ii.89) andare again enclosedin damned (ii.89; readin a crosswise
fashion, this suggests that being damned means being d-name-d). The letters
themselves are thus shown to indicate the way in which Thomas pursues the
poetic purpose of redeeming the contraries with secretive images saying
two things at once in one word, four in two and one in six (letter to Charles
Fisher in 1935; Thomas 1987: 182).
Thomas both avoids the name of Christ and refers to it (ii.35: unchris-
tened; ii.9294: But the loud sun / Christens down / The sky). He emphasizes
the visible sign that comes before andafter the spoken or written word, just as
the Christ-cross-row, the alphabet as it traditionally appeared in horn-books
andprimers, was preceded(andsometimes followed) by the sign of the cross
(Tuer 1979: 64, e.g. illustration on p. 59).
But this sign, this vision or theory is
nothing without the eective word, the actual being, and this is why we are
justied in calling it iconic. What Thomas seeks to drive home is the power of
the sign,
and in order to do so he chooses a shape that, on the one hand,
participates in non-verbal signication (X is a mark, as in the expression X
marks the spot), andon the other hand, is a letter of the alphabet (and has a
meaning only when combinedwith other letters). But here again Thomas says
four things in two for the X-shape is of course also, on the one hand, a symbol,
representing, for example, Christ (or the number ten;
the shape of the poem
thus points beyondthe number nine, which is the limit of the number of
in each line) and on the other hand, a spherical diagram that makes
visible the coincidence of opposites. The X is thus an icon of the poets aiming
for ultimate evidentia or energeia, for the (almost) physiological reality and
sacramental eectiveness of the word. It is, in the words of Wallace Stevens (in
a poem written about a year before Vision andPrayer), The vital, arrogant,
fatal, dominant X (The Motive for Metaphor l. 20, Stevens 1972: 240).
two- andthree-d imensional shapes (lozenge andtriangle, diamond andd ouble
pyramid) of the poem, which are variants and elements of the letter X, are iconic
images (and diagrams) of the creative word, which draws its letters from the rays
of light, as Thomas put it in the poem calledIn the beginning (no. 87): the
word / That from the solidbases of the light / Abstractedall the letters of the
void (1921). The vision, as well as the prayer, is one of thing and sign being
originally andultimately identical (the wordis to be seen, the sign to be heard).
Nevertheless, the cipher of Vision andPrayer is but an icon of this original word,
which can never be fully grasped. It is, by denition, a sign of the unknown, X,
giving shape to an account of what the poet most deeply fears and desires.
1. This is not an attempt at a comprehensive reading of Thomass complex poem, and even the
one particular perspective chosen does not include all aspects relevant to it, such as prosody.
But at least I hope to have chosen an approach invitedby the poem itself. Thomass poems are
quotedfrom D. Joness edition (Thomas 1985a). I am grateful to Professor Inge Leimberg for
a number of suggestions.
2. See e.g. in Adler and Ernst (1988: 3031) the reproduction from a 1545 edition of Theocrituss
3. E.g. Moynihan (1968: 134): The double pyramid may conceivably be seen as springing from
an immeasurable groundof hopefulness, passing through dryness, nally coming to an
expansive point of exultation andacceptance. Moynihan himself regards his comments as a
completely subjective reaction. Korg (1992: 120) suggests that in Section II, the convergence
of forces or reversal suggestedby the stanza form corresponds with the conict of impulses that
is the subject. It also reects (more particularly by its rhythm) withholding, followed by the
yielding of assent. More specically, Korg draws attention to the emphasis given by the shapes
to individual words, e.g. I appearing as the axislike connective between the parts of the last
stanza. Cf. also the interesting observation made by McKay (1969: 79), who sees a spiral of
ascension which is exemplied, perhaps, in the poems characteristic diamond and hour-glass
shapes which employ combinations of the spiral idea. McNees (1992: 138) suggests that the
emblematic hourglass stanzas of Part 2 visually reinforce the kenotic emptying and pleromic
fulllment of the eucharistic service. She sees the two shapes as marking two contrasting times
the rst ebbing away from, andthe second fullling the rst andrefers to Davies (1977: 52)
who regards the shape of the rst part as a diamond pattern which represents, symbolically,
birth, andtherefore the womb, andthe pattern of the second part as cross-like. For the resem-
blance to the cross, see also Jones (1966: 81). Emery (1962: 257) seems to deny the expressiveness
of the shapes altogether: the self-consciousness of the poet as craftsman belies his furor poeticus.
4. The diagram from Bovilluss De intellectu (1510) f. 85 is reprintedin Martins ed. (Browne
1964: 357).
5. Cf. Thomass sonnet Among those Killed in the Dawn Raid (no. 143), which ends: a
hundred storks perch on the suns right hand (14).
6. In Max Nnnys classication, the X of Vision andPrayer would thus probably be grouped
among the translucent letter-icons (which reveal themselves as icons in a ash; Nnny
1999: 175), although its iconic function is not entirely coveredby Nnnys typology, which is
concernedwith the shapes of letters used in a text rather than the text itself being letter-shaped.
Nnnys article nevertheless provides an excellent background for my reading of Thomass
poem, since, focussing on the letter O, it shows how wide the range of letter-icons may be.
7. A close reading of the poem in the context of the traditional mystical concept of the blinding
light (or visionary darkness) seems a desideratum. It might prove fruitful to take Crashaws
Epiphany hymn as a starting point, which presents the advent of the Bright Babe (1) to a
speaker Lost in a bright / Meridian night / A Darkness made of too much day (1618) and
such iconic implications of the Cross as the chiasmus of All circling point. All centring
sphear (26). BehindCrashaw s imagery appears, of course, Dionysius Areopagita.
8. The synaesthesia is reminiscent of the beginning of the Prologue in Heaven in Goethes
Faust. Thomas seems to have been familiar with the play, as his remarks on Louis McNeices
translation indicate (Thomas 1987: 711).
9. Thomas could nd the characteristic combination of the meeting parallels andthe speakers
own body in John Donnes The Crosse (Donne 1952: 26): Who can deny me power, and
liberty / To stretch mine armes, andmine owne Crosse to be? (1718) andAll the Globes
frame, andspheares, is nothing else / But the Meridians crossing Parallels (2324). McNees
(1992: 139) is reminded by Thomass poem of Donnes insistence on personal crucixion as
a requirement for Gods grace.
10. See OED X chromosome. It shouldbe noted that both sex and six (the number of stanzas in
each section) include the letter X.
11. For an introductory historical survey of the mysticism of letters, see esp. Ch. 4 in Drucker
(1999: 7292).
12. Cf. Ackerman (1991: 216): We must always remember that his favourite prose writer was
13. The mouth-bell is another cave which is both closedandopen.
14. See also his letter to the director of Dent, A. J. Hopp (18 September 1945, Thomas 1987: 569),
in which he asks that the poem be printed exactly as it shouldbe andthe letter of 6
November 1945, in which he thanks him for having done so (Thomas 1987: 572).
15. The editors of Thomass Collected Poems (Thomas 2000a: 246) cite his letter to Vernon
Watkins (15 November 1944) to suggest that Thomas acknowledged some inuence from
Francis Thompson, but not from Hopkins, nor from George Herbert. This is in fact not quite
what the letter says. Thomas points out that he does not remember seeing any Hopkins after
the poem was nished (Thomas 1987: 532) but this does not mean, of course, that he denies
Hopkins being among the formative (andcreatively transformed) inuences. The question of
Herberts inuence does not arise here because Watkins, in his letter, obviously only mentioned
Thompson andHopkins.
16. Thomas speaks of the Christ-cross-row of death in When once the twilight locks no longer
(no. 90, l. 24).
17. Cf. OED power n.
3.b.: The soundexpressed by a character or symbol; the meaning
expressedby a wordor phrase in a particular context: = force n.
18. See Meyer andSuntrup on the number ten (1987: 591614) as a sign of perfection (591) and
of Christ (598 on the X).
19. See Maud (1963: 160), who notes a few exceptions (the last stanza of Part I andthe thirdstanza
of Part II).
20. Other references to the letter X in Stevens, who seems to regardit as a mark of the desired (and
feared) unknown origin of the poetic process, include Someone Puts a Pineapple Together
(Stevens 1972: 295): Himself, may be, the irreducible X/ At the bottom of imagined artice
(89). Cf. also The Creations of Sound (Stevens 1972: 250). My attention was drawn to
Stevenss fascination with the letter by Firmage (1993: 256).
Ackerman, J. 1991. A Dylan Thomas Companion. London: Macmillan.
Adler, J. and U. Ernst. 1988. Text als Figur: Visuelle Poesie von der Antike bis zur
Moderne. Weinheim: VCH.
Browne, T. 1964. Religio Medici and Other Works. Ed. L. C. Martin. Oxford: Clarendon
Crashaw, R. 1957. The Poems English Latin and Greek of Richard Crashaw. Ed . L. C.
Martin. 2nd ed. Oxford: Clarendon Press.
Davies, A. T. 1977. Dylan: Druid of the Broken Body. Swansea: Christopher Davies. First
Davies, J. A. 1998. A Reference Companion to Dylan Thomas. Westport, CT: Greenwood
Donne, J. 1952. The Divine Poems. Ed. H. Gardner. Oxford: Clarendon Press.
Drucker, J. 1999. The Alphabetic Labyrinth: The Letters in History and Imagination.
London: Thames and Hudson.
Emery, C. 1962. The World of Dylan Thomas. Coral Gables, FL: University of Miami
Evans, I. H. 1981. Brewers Dictionary of Phrase and Fable. London: Cassell.
Firmage, R. A. 1993. The Alphabet Abecedarium: Some Notes on Letters. Boston: Godine.
Fischer, O. andM. Nnny 1999. Introduction: Iconicity as a creative force in language
use. In FormMiming Meaning: Iconicity in Language and Literature, O. Fischer and
M. Nnny (eds), xvxxxvi. Amsterdam and Philadelphia: John Benjamins.
Garlick, R. 1973. The Shapes of Thoughts. Poetry Wales 9.2: 4048.
Herbert, G. 1941. The Works of George Herbert. Ed. F. E. Hutchinson. Oxford: Clarendon
Jones, M. R. 1966. The Wellspring of Dylan. English Journal 55: 7882.
Kidder, R. M. 1973. Dylan Thomas: The Country of the Spirit. Princeton: Princeton
University Press.
Korg, J. 1992. Dylan Thomas. Updated ed. New York: Twayne.
Leimberg, I. 1998. The name of the bow is life: Order and meaning in Hopkins That
Nature is a Heraclitean Fire andof the comfort of the Resurrection. In Zum Begri
der Imagination in Dichtung und Dichtungstheorie: Festschrift fr Rainer Lengeler
zum 65. Geburtstag, M. Beyer (ed), 95116. Trier: WVT Wissenschaftlicher Verlag.
Leitch, V. 1972. Herberts inuence in Dylan Thomass I see the boys of summer.
Notes and Queries 217: 341.
Mahnke, D. 1937. Unendliche Sphre und Allmittelpunkt: Beitrge zur Genealogie der
mathematischen Mystik. Halle: Niemeyer.
Maud, R. 1963. Entrances to Dylan Thomas Poetry. Lowestoft: Scorpion Press.
Maud, R. 1968. Poet in the Making: The Notebooks of Dylan Thomas. London: Dent.
Mayer, A. E. 1995. Artists in Dylan Thomass Prose Works: Adam Naming and Aesop
Fabling. Montreal andKingston: McGill-Queens University Press.
McKay, D. 1969. Dot, line, andcircle: A structural approach to Dylan Thomass
imagery. Anglo-Welsh Review 18: 6980.
McNees, E. J. 1992. Eucharistic Poetry: The Search for Presence in the Writings of John
Donne, Gerard Manley Hopkins, Dylan Thomas, and Georey Hill. Lewisburg:
Bucknell University Press; London andT oronto: AssociatedUniversity Presses.
Nnny, M. 1999. Alphabetic letters as icons in literary texts. In Form Miming Meaning:
Iconicity in Language and Literature, O. Fischer andM. Nnny (eds), 173198.
Amsterdam and Philadelphia: John Benjamins.
Meyer, H. andR. Suntrup. 1987. Lexikon der mittelalterlichen Zahlenbedeutungen.
Mnchen: Fink.
Moynihan, W. T. 1968. The Craft and Art of Dylan Thomas. Ithaca: Cornell University
Seargent, H. 1962. The religious development of Dylan Thomas. A Review of English
Literature 3 (2): 5967.
Shakespeare, W. 1992. The Poems [The New Cambridge Shakespeare]. Ed. J. Roe.
Cambridge: Cambridge University Press.
Stevens, W. 1972. The Palm at the End of the Mind: Selected Poems and a Play. Ed . H.
Stevens. New York: Vintage.
Thomas, D. 1985a. The Poems. Ed. D. Jones. London and Melbourne: Dent. Rev. ed. rst
Thomas, D. 1985b. Under Milk Wood. London and Melbourne: Dent. First published
Thomas, D. 1987. The Collected Letters. Ed. P. Ferris. London et al.: Paladin Grafton.
First published1985.
Thomas, D. 2000a. Collected Poems 19431953. Ed. W. Davies and R. Maud. London:
Phoenix. First published1993.
Thomas, D. 2000b. The Collected Stories. Ed. W. Davies. London: Phoenix. First
Tindall, W. Y. 1996. A Readers Guide to Dylan Thomas. Syracuse, NY: Syracuse
University Press. First published1962.
Tory, G. 1973. Champ Fleury ou lart et science de la proportion des lettres (1529). Ed .G.
Cohen. Genve: Slatkin Reprints. First published1931.
Tuer, A. W. 1979. History of the Horn Book. New York: Arno Press. First published1897.
Wardi, E. 19992000. A boy in the listening: On voice, space, and rebirth in the poetry
of Dylan Thomas. Connotations: A Journal for Critical Debate 9: 190209.
</TARGET "bau">
AUTHOR "Christina Ljungberg"
TITLE "Diagrams in narrative"
WIDTH "150"
Diagrams in narrative
Visual strategies in contemporary ction
Christina Ljungberg
University of Zurich
1. Introduction
In the light of the harsh criticism by postmodern critics that the overwhelming
ow of free-oating images in our culture has causeda loss of referents in
modern art andliterature and a cleft between representation andreality in the
digital and the mass media (cf. Foucault 1970; Lyotard 1979; Jameson 1979;
Baudrillard 1981; see also Nth in press), it is interesting to explore the extent
to which contemporary writers employ visual strategies in their ction. The
interplay between text andvisual media has become an important strategy for
writers who exploit the specic characteristics of each medium in order to bring
to the fore problematic issues like the representation andmed iation of reality and
the real world. In particular, the interaction between visual artifacts such as
photographs andmaps and writing opens up intriguing relationships between
verbal andvisual representation, emphasizing its literal meaning of bringing
something into someones presence, at the same time as it functions as a
technique of reader involvement.
The interplay between visual andverbal media is by no means a new
development. As Nancy Armstrong (1999: 4) convincingly argues, even in the
nineteenth century, the newly inventedtechnique of photography already played
a decisive part in the process of increasedinteraction between verbal andvisual
images which resultedin Victorian realism. In her view, the unprecedented
number of mass-reproduced images detached from their referents and from
reality that we see today came about because the new medium of photography
introduced new ways of categorizing the real. Hence, the belief that photogra-
phy could andd id represent reality brought with it a new way of seeing
which came to be equatedwith knowing. By verbally recreating visual informa-
tion that hadbeen previously photographed, writers were able to communicate
visual categories that were both a way of seeing anda picture of the world
(Armstrong 1999: 19).
The map is another, albeit much older, visual device that has often attracted
writers as a means not only to plot in words but also to make diagrams of their
visions andinterpretations of real or imaginary worlds. Yet, it was not until the
late nineteenth century that maps start to appear more frequently in ction, as
geographical exploration made space an issue of general interest and the
increaseduse of graphics in the printedmed ia through new techniques made
cartography accessible to a wider audience. By stimulating their readers visual
imagination in order to draw them naturally into the space-time continuum of
the narrative, maps proved to be an excellent device for achieving reader
One of the reasons for the intense criticism of andincreased attention to
both photographs andmaps may well be that both have traditionally been
considered faithful representations of reality. A photograph, for instance, was
traditionally considered a mirror of reality, or, in Charles Sanders Peirces
terms, an icon. Peirce himself, however, characterizedphotographs both as icons
and indices, since, on the one hand, they are in certain respects exactly like the
object they represent and, on the other, they have a physical connection with
their object, since photographs, by virtue of the photochemical process, were
physically forcedto correspondpoint by point to nature (CP 2.246).
Maps, too, have traditionally been considered iconic, although their indexical
properties have been arguedfor (cf. Nth 2000: 489; Sebeok andDanesi 2000: 60)
they are causally relatedto the territory they depict, andthey also orient their
users either in their immediate geographical environment or in their mental imagi-
nary space. As an icon, therefore, the map is primarily considered a diagram, an icon
or a form of relations analogous to those of the object it represents.
In the contemporary ction that will be discussed here, photographs and
maps function primarily as instances of the specic form of iconicity called
diagrammatic iconicity. Although a diagram has both indexical and symbolical
features, Peirce calls it a kindof icon particularly useful, because it suppresses
a quantity of details, andso allows the mindmore easily to think of its important
features (CP 2.282). I will argue that it is the aspect of making important
features visual that makes diagrammatic iconicity so frequent in postmodern
ction the ction createdin what Susan Sontag (1979: 153) has calledThe
Image-World because diagrams allow us to visualize something we cannot
physically see. As iconic signs of rationally relatedobjects that are createdin our
imagination, they enable us to make mental experiments or to trace imaginary
journeys on a map. This makes diagrammatic iconicity an eective visual
strategy for creatively exploiting the possibilities of interaction between various
semiotic systems.
2. The diagrammatic function of photographs in narrative
From the very time of its invention, photography has been comparedto both
writing and drawing. The photographic technique developed by William Henry
Fox Talbot, which he documented in The Pencil of Nature in 1844, startedout as
an attempt to improve a drawing device that was popular at the time. This
consistedof a piece of paper mountedinsid e a camera obscura which produced
a reversedimage that couldbe sketchedin, andwhich was further rened by
Talbots invention of the calotype, a kindof photogenic drawing that
produced a natural image that could imprint itself on paper (Armstrong
1999: 1415). As Armstrong argues, the process discovered by Talbot yielded
a picture that seemedcapable, through repetition, of representing the visible
worldas it actually was, whole and entire. The fantasy that reality couldbe
represented unmediated, by natures hand, thus implied that the only true
picture was a photograph, a claim which has been repeatedly questioned in the
present era of picture manipulation through various digital techniques such as
computer morphing andcomputer graphics.
Exploring the role of photography in our perception of reality, Susan Sontag
writes (1979: 158):
Photography has powers that no other image-system has ever enjoyedbecause,
unlike the earlier ones, it is not dependent on an image maker The mechani-
cal genesis of these images, andthe literalness of the powers they confer,
amounts to a new relationship between image andreality . Andif photography
couldalso be saidto restore the most primitive relationship the partial
identity of image and object the potency of the image is now experienced
in a very dierent way. The primitive notion of the ecacy of images
presumes that images possess the qualities of real things, but our inclination is
to attribute to real things the qualities of an image.
To Sontag (1979: 160), photographic images achieve a particular form of
perception that views reality as an endless set of situations which mirror one
another, with the result that reality has come to be understood as a kind of
writing that has to be decoded. Hence, as a method of reading, photography has
changedthe relationship between reality andimage.
That the two concepts of reality andphotographic image have thus
become complementary andimpossible to distinguish from one another has also
been discussed by Lucia Santaella (1988: 257), who argues that, in the same way
as photographs change our perception of the world, this changed perception
creates new forms for their production and interpretation. What changes the
result of this inseparable relationship is not only our vision or concept of
reality, but the nature of reality itself: even though photographs have a replicat-
ing function, like all other signs they combine with reality and, thus, increase its
complexity to what we today call hyper-reality.
Photographs are strange incarnations that conjure up both intimacy and
alienation, presence andabsence. According to RolandBarthes (1980: 176),
photography accomplishes the unparalleledid entication of reality (thats how
it was) with truth (thats how it is). Someone or something that is absent
becomes present, but only to the extent that it is allowedto by the viewers
power of imagination. This potential has, of course, often attractedwriters. In his
fascinating book on Marcel Prousts interest in photography, Brassa (1997)
shows how Prousts attempt to conjure up a lost worldandtime in Recherche
was greatly inspiredby his passion for andund erstanding of the ambiguity and
duplicity of photography, its potential to represent reality in fragments, and its
ability to make things simultaneously present andabsent. As Brassa (1997: 20)
points out, Proust not only has photographs function as eective narrative props
which give rise to comical or tragic incidents, but he also uses photographic
terms such as linstantane (snap-shot), pose, limpression (print), clich
(negative), chambre noire (darkroom), dveloppement (development), and
xage (xing). Furthermore, in Prousts work, photography functions as a
metaphor for his whole quest to recapture time andspace in which life is
constitutedby strange andcontrasting images that, like negatives, lie buriedin
our minds. Hence, Prousts Recherche is not just a romantic or realistic narrative
that spans a certain period, but it consists of a myriadof recollectedmoments,
viewpoints andfoci that only gain coherence and meaning once they are
developedandlaidone on top of the other (ibid.: 173174).
To my mind, this last point is particularly relevant to the frequent occur-
rence of photographs in postmodern ction, in which they are often used
diagrammatically. They are employed not only to reect the organization of the
narrative itself, as well as relationships within a group, or a characters relation-
ship to her or his surroundings, but also to problematize the representation of
reality andof appearances, and the relation between the two, making photogra-
phy a device for self-reexive comment on representation in general, and on the
art of writing in particular.
That is how Michael Ondaatje uses photography in his long prose poem The
Collected Works of Billy the Kid, which is about William Bonnie, the American
outlaw who became a symbol of violence andwas given his nickname because
he killeda man at the age of twelve. The poem starts andend s with a photo-
graphic frame, an empty one at the beginning (Figure 1), andthen one with a
small portrait of a boy dressedup as Billy at the end(Figure 2). The poem is
made up of contemporary accounts, periodphotographs andd ime novels,
interspersedby contrasting poems, prose sections andphotographs. The contrasts
between static photographs, lyrical andviolent poems and prose which
function more as a slide show than as anything else seem to question the
ability of both literature andphotography to represent life.
Figure 1. The beginning of The CollectedW orks
of Billy the Kid (1996: 5).
Figure 2. The end of The Collect-
edW orks of Billy the Kid
(1996: 107)
The static quality of the photographs is also reected in the discursive scenes,
which do not leadanywhere near a traditional plot andwhich can be readin
almost any order from within. Starting with Billy the Kid announcing his own
death, the non-linearity of the text, which matches the non-sequential order of
reading a photograph, subverts the linearity of traditional narrative. Hence,
photography functions here as the organizing principle andframe in Ondaatjes
work, at the same time as it becomes an icon of the creative act itself. The traces
left by light on a blank lm are comparedto the poems birth from an empty
page one is again reminded of the pencil of nature, except that, this time,
the truth of representation is calledinto question: the commentary on the rst
empty photographic frame (Figure 1) concerns the diculty of photographing a
moving object; another ctive photograph of Billy is saidnot to have been
included, since the picture made him rough and uncouth (Ondaatje 1996: 19),
which would be unt for a legend. Furthermore, it is due to a reversed negative
that Billy who was right-handed was reputed to be left-handed. The notion
caught on, since it tted his image as an outsider, one who is not right; it also
alludes to the frequent occurrence of the theme of the artist as an outlaw in
Ondaatjes writing (cf. Solecki 1985: 193; Barbour 1993: 37). This theme is also
suggestedon the front cover, on which Ondaatjes name is written on top of the
photograph presenting Billy with his left handon his gun, with The Collected Works
of Billy the Kid underneath, and again by Ondaatje inserting himself as a child
dressedas a cowboy into Billys textual worldin the nal photograph (Figure 2).
The result is a blurring of identities, making Ondaatje, as part of the process,
as dicult to grasp as the I-narrator Billy, who begins his story by reporting
his own death, and who repeatedly talks to us from the other side. As Doug
Barbour (1993: 43) points out, this results in a continuous slippage in which
I will slide from implied author to narrator to one of the characters as
narrator; you will slide from implied reader to narratee to particular charac-
ters; andthe name, the site of our various interest, will simply multiply into an
assortment of possibilities.
It is this process that is enactedby the photographs, miming the evasiveness of
a reality that refuses to be xed: Billy the Kid, as a historical person, is long
dead, and the boy in the photograph who was once Ondaatje does not exist any
more. By trying to describe reality as seen in a photograph or through a camera
lens, the world becomes detached, framed and alienated, as an antithesis to life.
Instead, this results in what Barbour (1993: 66) calls a carefully chaotic
collage. Collectedby an ambiguous I who leaves the story after a bad night
(Ondaatje 1996: 105), it leaves us, its readers, to bring the collage poem to life
through the act of reading.
Writing andphotography are also key topics in Margaret Atwoods recent
novel, The Blind Assassin (2000). Atwoodemploys a manipulatedphotograph as
the main icon andor ganizer of the novel, andshe has it appear as dierent
versions of reality. The novel itself is a complex web of layerednarratives,
dierent text types a wildmix of pulp ction, detective mystery, newspaper
clippings andnarrative devices ashbacks, multiple time schemes,
ambiguous and indeterminate plots in which dierent voices andd ierent
viewpoints intersect to create a wide panorama of English Canada in the
twentieth century.
Referring to the novels wordplay, bits of arcane information, historical
artifacts, gossip, social and political commentary that leads the reader into an
labyrinthine search for the secret chamber at the heart of the novel, a reviewer
has proposed that reading Atwood is like playing an interactive videogame
(Martin 2000). I would suggest that this is, in fact, mainly due to Atwoods
methodof spatial plotting andher use of chiastic structures in the narrative itself
(cf. Nnny 1997: 59; see also Ljungberg 2001: 358) that gives her novels their
unique visual ability, which Figure 3 shows.
The novels inner story frame (BA: Prologue BA: Epilogue) is formedby
the verbal description of two almost identical photographs that function as the
key elements of the narrative, constituting both the prologue andthe epilogue of
the novel within the novel:
She has a single photograph of him Shes preservedthis photo carefully,
because its almost all she has left of him The photo is of the two of them
together, her andthis man, on a picnic [].
Over to one side you wouldnt see it at rst theres a hand, cut by the
margin, scissoredo at the wrist, resting on the grass as if discarded. Left to
its own devices. (Atwood: 2000: 45)
This is repeatedalmost verbatim in the epilogue, but with an important dier-
ence in the last three lines:
In the lower left corner theres a hand, scissored o at the wrist, resting on
the grass. Its the handof the other one, the one who is always in the picture
whether seen or not. The handthat will set things down. (Atwood2000: 517)
Stemming from a single negative, these two photographs present two versions of
a shot taken on a picnic, showing Iris, the narrator, andher sister Laura anking
the man they both were in love with: him in the middle with the two of us, like
bookends me to the left of him, Laura to the right (2000: 192). The three
people in the photograph are also the novels three main storytellers: Laura,
whose single, posthumously publishednovel The Blind Assassin made her a cult
author; the male lover of her story, who is also a writer of pulp science ction;
andIris, who is painstakingly writing her memoirs.
The importance of the handin the photographs is emphasizedby its reappear-
ance in the narrative, as the bodiless hand scrawling across the wall (2000: 512);
Figure 3. Narrative diagramof The Blind Assassin. Iris refers to Iriss narration (in which
each section is composed of multiple time schemes of even greater complexity than is shown
here); news to newspaper clippings; and BA to the interlacing of Lauras cult novel within
the novel, The Blind Assassin, and to the pulp science ction story about the blind assassin
that her lover is telling her.
I. The bridge
Newspaper clipping: Laura s death
BA: Prologue: Photograph
II. BA/news/BA/news/BA news/BA/news /
III. Iris 1998
IV. BA BA BA BA BA /news/ /news/ /news/ /news/
V. Iris 1998
VI. BA BA BA BA BA / /news/ / /news/
VII. Iris 1998
VIII. BA BA BA /news/ /news/
IX. Iris 1999
X. BA BA BA /news/letter/ /news/
XI. Iris 1938/1999
XII. News/news/ / / / BA BA BA BA
XIII. Iris 1999
XIV. Iris 1999
XV. BA: Epilogue: Photograph
Newspaper clipping: Iris s death
The threshold
this, interestingly, is also how Atwoodrefers to her own writing in a recent inter-
view (Martin 2000). Hence, the hand, or hands, would seem to refer to the urge
andto the Proustian necessity to break open the past, in order to conjure it up
andto write it; in this case, to Iriss desire both to give her past coherence and
meaning and to do justice to her sister whom she let down when she needed her.
Only too late did she realize how mutually dependent they were on one another,
andthat Laura was my left hand, andI was hers (2000: 513).
By naming the novels rst chapter about Lauras death The bridge and
the last one, which is Iriss posthumous comment, The threshold, Atwood
seems to use these two traditionally liminal spaces of passage (cf. van Gennep
1960: 21) not only to refer to the deaths of the two sisters but also to locate
creative activity between life and death. This seems to be further underlined by
the ability of both photography andwriting to freeze a moment in time, to both
petrify it and have it allude to the latent eternity of the indestructible
(Santaella 1998: 162).
In addition, the theme of immortalization is inherent in the conventions of
autobiography andmemoir , conventions to which Atwood has repeatedly
returnedto examine andquestion (Ljungberg 1999: 37113). In The Blind
Assassin, too, the question of truth andreliability pervades the novel, which is
also suggestedby the manipulatednegative. Hence, the two versions of the
photograph reect Atwoods playing with issues of remembering and forgetting,
of repressing uncomfortable truths andfabricating others: Iriss rambling memoir
is populatedby characters that are too stereotypical to be believable, which
makes the reader suspect that her story is as much a product of imagination as
of memory, andof presenting her own actions in a more forgiving light. This is
also reected in the structure of Iriss narration, which starts out in the present,
inevitably moves back into the past, andthen back to the present again (see
Figure 3), a technique which Atwoodhas usedpreviously (cf. Ljungberg
1999: 3941, 151160; 2001: 357360). In this context, these narrative triads
also allude to photographys ability to make things both past and present, at the
same time as they recall Barthess (1980: 176) contention of photography being
a cross-over between reality andtruth, which I mentionedabove. The interesting
twist in this novel is, to my mind, the fact that it is driven forward, not by Iriss
self-preoccupiedand often boring narration, which is like that of most people,
but by the ctional story within the story; that is, by the mixture of Lauras once
scandalous love story which, as we nd out at the end, was written by Iris
and published under her dead sisters name in anticipation of the scandal and
her lovers pulp science ction promptedby the photographs. As with photogra-
phy, it is the selection andframing, as well as the manipulation of point of view
that makes a story interesting, not the myriad of details embraced. Thus, in
contrast to the role of photographs playedin Atwoods earlier ction (e.g., The
Edible Woman 1969), where they functionedprimarily as metaphors of death and
stasis, of acquisition andof a consumer-like relationship to the world, in The
Blind Assassin, her use of photography celebrates ctions potential to create a
fascinating andcomplex narrative out of a rather limitedandd isappointing life
and, hence, to recreate life in art.
3. The diagrammatic function of maps in narrative
Although maps, like photographs, represent our worldthrough a two-dimensional
medium, they are more similar to language in so far as they use an elaborate
system of arbitrary symbols in order to locate or describe geographical places.
Writers seem to have discovered the maps representative potential at an early
stage: even Sir Thomas More (1516; 1518) already employed the famous maps
of Utopia as instruments of seeing, in order to help his readers to see the social
wrongs in their own worldby comparing it with an ideal worldand , thus, to
function as potential instruments of change (cf. Wooden and Wall 1985: 245).
Another famous early literary cartographer, Jonathan Swift, usedhis maps in
Gullivers Travels (1726) to ridicule his contemporaries limited geographical
knowledge, whereas the maps in John Bunyans (1678, 1684) The Pilgrims
Progress were designed to show readers the way to a better life. To Robert
Louis Stevenson, sketching the imaginary shorelines of Treasure Island (1892/
1924), a map is an indispensable tool for creative thinking: even when a map is
not all the plot, as it was in Treasure Island, it will be foundto be a mine of
suggestion (Stevenson 1892/1924: xxxi).
Stevensons assertion about maps coincides with what a contemporary,
namely Charles Sanders Peirce, said about diagrams, which I quoted at the
beginning, calling the diagram a kind of icon particularly useful (CP 2.282).
Because of their abstract nature, maps can clarify andshow relationships and
patterns in new andstimulating ways; in Franco Morettis (1999: 5)
maps bring to light the internal logic of the narrative: the semiotic domain
aroundwhich a plot coalesces andself-or ganizes. Hence, maps make visible the
only real issue of literary history: society, rhetoric andtheir interaction (ibid.).
In his view, in nineteenth-century ction, literary spaces reect less a social
reality than a complex network of power andauthority , as well as the readership
it addresses.
I wouldar gue that this also accounts for the frequent occurrence of maps in
contemporary ction, which fully exploits the insight that spaces are socially
constitutedthrough language andother sign systems. In particular, investigations
into spatial practices andspatial meaning have convincingly shown how space is
both producedandconstructed . This is, I wouldsuggest, one of the reasons for
the abundant use of maps both by postcolonial writers, who employ them as
metaphors of their fragmentednational cultures, andby postmodern writers,
whose self-reexive ction uses maps in order to reect the writers mapping of
her or his own ctional landscape and/or in order to problematize the referential
function of language.
Marlene Nourbese Philips Looking for Livingstone: An Odyssey in Silence
(1991) is a literal expression of the intersection of language, space, time and
identity. An experimental novel about a woman travelling alone through time,
Africa andunnamedspace, in search of Dr. Livingstone, Africas famous
discoverer, Philips text, which is a mixture of prose and prose poems populat-
edby both imaginary andhistorical characters, represents the disjointedand
female black body as partly mappedand colonizedcultural andsocial space. The
exploratory character of her odyssey is announced both by the book cover, which
presents an oldandbarely legible map of the interior of Central Africa adorned
with a redbeacon with a ashing lighthouse placedslightly o-center (Fig-
ure 4), andby the opening lines, which continue the map theme by describing
the speakers own map as a primitive one, scratchedon animal skin (1991: 1).
Philips text, which is dedicated to her ancestors who have been silent
for too long literally enacts the binary opposition of silence andword by
personifying the two concepts andopening up a dialogue between them. The
word silence is made into six dierent anagrams, presentedas the names of six
dierent African peoples whose encounters are dramatizedandrend ered as a
dialogue reminiscent of the Bakhtinian notion of the internal dialogicity of the
novel (Bakhtin 1994: 293; cf. Heim 1999: 13). This dialogue is reinforced in the
frontispiece map, whose fragmentary character questions its claim to truthfulness,
at the same time as it opens up its silencedandd iscovered geography by
suggesting that an awareness of the interdependency of word and silence is a
necessary prerequisite for the enlargement of ones own experiental space.
In a very dierent vein, Paul Austers intriguing maps in his highly
metactional City of Glass (1988) represent the problem of the writers encounter
with language andthe creation of meaning. The mystery in which the New York
writer Daniel Quinn becomes involvedstarts with bizarre phone calls asking for
a private detective named Paul Auster. A paperback detective writer whose
ction features a private eye calledMax Work operating under the nom de plume
William Wilson, Quinn gets himself involved (under the identity of Auster) in
the strange case of Peter Stillman, a young man of defective behaviour. Stillman
Figure 4. The map cover of Looking for Livingstone (Philip 1991).
spent his childhoodisolatedin a dark room by his father, Peter Stillman Sr., a
well-known linguistics professor, who wantedto teach him the prelapsarian
Gods language. Quinn, who has no real detective experience, is hired by
Stillman Jr. andhis wife, to protect Stillman Jr. from Stillman Sr., who has been
releasedafter 13 years in prison andwho has threatenedto nd and kill his son;
intrigued, Quinn accepts the job andgets lost in the gap formedbetween his
metactional worldandreality.
Consciously leaving many worthwhile aspects of this outstanding text aside,
I wouldlike to focus on the problems of meaning in this very self-reexive text,
and how they are communicatedby the maps in Quinns rednotebook (Figures
58). Since one of the books major intertexts is Lewis Carrolls Through the
Looking Glass, which describes a region where the laws of space, time, language
and logic are partially suspended, I would like to examine the anomalies of
semiosis in City of Glass, where Quinn, like Alice, is constantly searching for
signs. Thrown from his initial position as a perfect metactional character, a
detective writer with no experience of the real underworld of crime except of
that in other crime ction, Quinn gets lost in a Wonderland of signs. Taking my
cue from Quinns own analysis of his writing self as a triadof selves [in
which] Wilson servedas a kindof ventriloquist, Quinn himself was the dummy,
andW ork was the animatedvoice that gave purpose to the enterprise (1988: 6),
I will approach the text from a Peircean perspective. Semiosis, according to
Peirce, is basedon a semiotic triad, a triple connection of sign, thing signied,
[and] cognition produced in the mind (CP 1.372). However, as Quinn steps out
into reality, he nds that the process of interpreting signs becomes both
incomprehensible andd isorienting.
At the outset, Quinn is perfectly content with his theoretical andmeta-
ctional worldof crime: What interestedhim about the stories he wrote was not
their relation to the worldbut their relation to other stories (1988: 8). His
problematic relationship to his writing begins when he gets tangledup in a real
story: by becoming a real detective, he can no longer withdraw into his
constructedtriadof selves but is forcedto interpret signs in real life. His
meeting with Stillmans theories of representation pushes his disorientation even
further: like Humpty-Dumpty the purest embodiment of the human condition
a philosopher of language, as Stillman calls him (1988: 98) Stillman
assigns ad hoc meaning to his signs, andto the objects that he collects and
names, inventing new words that will correspond to the things (1988: 94),
reviving dead metaphors and creating new deadpan ones such as For all men
are eggs (1988: 98).
What lies behindStillman s quest to create a natural, prelapsarian language
is his horror at the fragmentation of language: even the worldis in fragments
(1988: 91); signs have lost their function to represent anything andonly refer to
themselves. His project is to invent [a] language that will at last say what we
have to say. For our words no longer correspond to the world (1988: 92).
By having Quinn trail Stillmans enigmatic wandering through the streets of
New York, Auster in fact has Quinn iconically enact Stillmans project, the
process of inventing a new language (1988: 92). Space becomes text; New
York, the most forlorn of places, the most abject, [where t]he brokenness is
everywhere [and] the streets an inexhaustible storehouse of shattered
things (1988: 94) becomes physically imbuedwith meaning. By selecting a map
gridof New York streets (Figure 5) andwalking the letters of TOWER OF BABEL
(Figures 68)
signs which Quinn reproduces in his red notebook as maps

into it, Stillman represents the text as physical space within which he is constantly
on the move, forever on [his] feet, going from one place to the next (1988: 91).
Yet, as Quinns maps in his rednotebook show, representation can never be
a 1:1 correspondence. Icons, too, have their limits, as Carroll shows in Sylvie
and Bruno Concluded (1893), where Mein Herr, a German professor, has devel-
Figure 5. Map grid of Manhattans Upper West Side, reaching from the Hudson river to
Amsterdam avenue which is the space of Stillmans meandering (Auster 1988: 81).
Figures 68. The letters O, W and E formed by Stillmans walks, interpreted by Quinn
(Auster 1988: 8184).
opeda map of 1:1 relations. This upset the farmers: They saidit wouldcover
the whole country, andshut out the sunlight! So now we use the country itself,
as its own map, andI assure you it does nearly as well! (1893: 169). Sign
economy demands a practical function: as Winfried Nth (1994: 21) points out,
sign-vehicles in natural semiosis must necessarily be adapted to the practical
conditions of sign usage Sign-vehicles, whether iconic or not, are normally
dierent from their object.
By having Stillman violate these kinds of pragmatic rules, Auster has
Stillman use the same wonderfully zany logic as that pervading Carrolls books,
playing on incomplete semiosis andthe transformation of signs into unusual and
unexpectedpotentialities of meaning. This is why Stillmans project of a new
conception of language appeals to Quinn, as a writer: Stillmans fear that
[u]nless we can embody the notion of change in the words we use, we will
continue to be lost (1988: 103) is closely tiedto the writers struggle for
meaning. The prospect that words may be able to open up a potential access to
reality fascinates him; however, as he becomes drawn into the storys time and
place, he becomes more and more disoriented. Finding himself in a junk heap
consisting of broken people broken things broken thoughts (1988: 94),
in which there is no reassuring story ending or revelation in sight, he loses his
bearings and nally disappears out of the story, leaving his rednotebook behind.
Noting the correspondence between the red notebooks maps of Stillmans
walking signication and Michel de Certeaus description of New York urban
space, I wouldlike to argue that the maps in City of Glass form the key element
of the story. Describing urban space as perambulatory rhetorics andgestures
play[ing] with spatial organization, de Certeau (1985: 136137) assigns a mythical
structure to these perambulatory gures which form a narrative cookedup out
of [sic] elements drawn from sharedsites, an allusive, fragmentedtale whose
gaps fall into line with the social practices it symbolizes. This is, to my mind,
exactly what Auster explores andrecreates ctionally in this fascinating novel in
which the walkedmaps iconically enact the problem of representation.
To conclude, by using and integrating essential characteristics of various
media, contemporary writers experiment to develop new ways of articulation.
Although the frequent occurrence of diagrammatic icons such as photographs and
maps in contemporary ction have made these important ingredients in contem-
porary texts, this is still a surprisingly uncultivated eld which calls for a more
thorough investigation into the interplay between visual andverbal representation
than the one I have been able to oer here. By using form to add meaning,
diagrammatic icons not only serve as aids for reader involvement, but also help
us to see relationships andpatterns in fresh andunexpectedways, thereby
creating new and interesting dimensions of reading.
1. In his interesting study, Atlas of the European Novel 18001900, Franco Moretti (1999) investi-
gates how geography shapedthe narrative structure of the European novel. His maps, drawn in
collaboration with Serge Bonin, demonstrate how space acquires a plot function by locating the
spatial relationships both within the novel and in relation to other novels. His discussion does,
however, not include maps published in ction that relate directly to the narrative.
2. However, the loss of the referent that Stillman deplores does not necessarily mean a crisis of
representation, despite the signs traditional denition as aliquid pro aliquo, something standing
for something else. According to Peirce, on the one hand, semiosis is a never-ending process
of signs; on the other, a sign without a reference can still be a sign. As Nth (in press) points
out, every sign, even if it refers only to itself, has the potential to produce an eect in a
possible interpreter, which means that it can function as a sign, andthat semiosis is possible.
In addition, according to Nth (ibid.), re(-)presentation does not mean the mere repetition of
a previous sign. Instead, it always involves a dierence from what precedes (as a sign or
referent), andit is the dynamics of the eect of this dierence which results in what Peirce
calls the growth of signs (cf. CP 2.302).
3. Babel refers, of course, both to New York as the extravagant, morally lax and decadent
Babylon, andto Stillmans Miltonian project of a single language of meaning.
4. Cf. Foucaults (1970: 64) discussion of the theoretical development of representation in Western
civilization, in which he argues that even the seventeenth century philosophical andsemiotic
school of Port-Royal (ArnauldandNicole 1662/1965: 45) already regardedtopographical maps
andspatial representation as tools necessary for what they calledthe way of getting to know
something by abstraction or precision (la manire de connatre par abstraction ou precision).
As Foucault points out, the rst example of a sign given by the Logique de Port-Royal is not
the word, not the cry, not the symbol, but the spatial and graphic representation the drawing
as map or picture (ibid.).
Armstrong, N. 1999. Fiction in the Age of Photography. Cambridge, Mass.: Harvard
University Press.
Arnaud, A. and P. Nicole. 1662/1965. La Logique de Port-Royal. B. Baron von Freytag
Lringho andH. E. Brekke, (eds). Stuttgart - BadCannstatt: Friedrich Frommann
Verlag (Gnther Holzboog).
Atwood, M. 1969. The Edible Woman. Toronto: McClelland& Stewart.
Atwood, M. 2000. The Blind Assassin. Toronto: McClelland& Stewart.
Auster, P. 1988. City of Glass. The New York Trilogy. New York: Penguin.
Bakhtin, M. 1994. The Dialogic Imagination: Four Essays by M. M. Bakhtin. M. Holquist,
(ed.). Austin: University of Texas.
Barbour, D. 1993. Michael Ondaatje. New York: Twayne.
Barthes, R. 1980. La chambre claire. Note sur la photographie. Paris: Gallimard.
Baudrillard, J. 1981/2000. Simulacra and Simulation. Ann Arbor: University of Michigan
Brassa 1997. Marcel Proust sous lemprise de la photographie. Paris: Gallimard.
Carroll, L. 1865/1971. Through the Looking Glass. New York: Norton.
Carroll, L. 1893. Sylvie and Bruno Concluded. London: Macmillan.
de Certeau, M. 1985. Practices of Space. In On Signs, M. Blonsky (ed.), 122145.
Baltimore: Johns Hopkins University Press.
Foucault, M. 1968/1970. The Order of Things. An Archaeology of the Human Sciences.
New York: Random House.
Gennep, A. van. 1960. The Rites of Passage. Chicago: University of Chicago Press.
Heim, O. 1999. Reading o the map, or: What, andwhere, is experience in the writing
of Aritha van Herk andM. Nourbese Philip? In Theory and Literary Creation, J.-P.
Dunix, (ed.), 129144. Dijon: Editions Universitaires de Dijon.
Jameson, F. 1979. Foreword and introduction. In J-F. Lyotard, The Postmodern Condition:
A Report on Knowledge. Minneapolis: University of Minnesota Press, viixxv.
Ljungberg, C. 1999. To Join, to Fit, and to Make. Bern: Peter Lang.
Ljungberg, C. 2001 Iconic dimensions in Margaret Atwoods poetry and prose. In The
Motivated Sign. Iconicity in Language and Literature 2, M. Nnny andO. Fischer,
(eds), 351366. Amsterdam and Philadelphia: Benjamins.
Martin, S. 2000. Making the business of life easier. Interview with Margaret Atwood.
The Globe and Mail,, August 26, 2000.
More, Sir T. 1516/1518/1992. Utopia. Trans. R. M. Adams. New York: W. W. Norton.
Moretti, F. 1999. Atlas of the European Novel 18001900. London: Verso.
Nnny, M. 1997. Hemingways architecture of prose: Chiastic patterns andtheir
narrative functions. North Dakota Quarterly 64.3: 157176.
Nth, W. 1994. Alices adventures in semiosis. In Semiotics and Linguistics in Alices
World, R. Fordyce and C. Marello, (eds), 1125. Berlin & New York: Walter de
Nth, W. 2000. Handbuch der Semiotik. 2nded ition. Stuttgart: Metzler.
Nth, W. (in press). Crisis of representation? Semiotica. Special Issue: Crisis of
Representation, W. Nth andC. Ljungberg, (eds).
Ondaatje, M. 1970/1996. The Collected Works of Billy the Kid. New York: Vintage.
Peirce, C. S. 19311958. Collected Papers. Vols. 16 C. Harthorne & P. Weiss (eds),
vols. 78 A. W. Burks (ed.). Cambridge, Mass.: Harvard University Press. (Quoted
as CP)
Philip, M. N. 1991. Looking for Livingstone: An Odyssey in Silence. Toronto: Mercury
Santaella, L. 1998. Die Fotograe zwischen Todund Ewigkeit. Zeitschrift fr Semiotik
20: 243268.
Sebeok, T., andM. Danesi. 2000. The Forms of Meaning. Modeling Systems Theory and
Semiotic Analysis. Berlin andNew York: Mouton de Gruyter.
Solecki, S., ed. 1985. Spider Blues. Essays on Michael Ondaatje. Montral: Vhicule.
Sontag, S. 1979. On Photography. New York: Penguin.
Stevenson, R. L. 1882/1924. Treasure Island. The Collected Works of Robert Louis
Stevenson. Tusitala Edition (35 vols.) Vol. II. London: Heineman.
Swift, J. 1726/1968. Gullivers Travels. Ed. and with a foreword by C. Lloyd. London:
Wooden, W. W. andJ. N. Wall. 1985. Thomas More andthe painters eye. Journal of
Medieval and Renaissance Studies 15: 231263.
</TARGET "lju">
TITLE "Structural iconicity"
WIDTH "150"
Structural iconicity
</TARGET "p3">
AUTHOR "C. Jac Conradie"
TITLE "The iconicity of Afrikaans reduplication"
WIDTH "150"
The iconicity of Afrikaans reduplication
C. Jac Conradie
Rand Afrikaans University
1. Introduction
The repetition of a lexeme in the soundstream of natural language, constrained
as it is by Saussures principe linaire, is a rich source of iconic possibilities.
What immediately comes to mindis repeatedlexemes referring to repeated
actions or occurrences, repeatedlexemes signalling duration of action rather than
a point-in-time occurrence, repetition signalling the extension of an action or
state, andrepetition signalling an intensied state or emphasis. Less obvious,
perhaps, is the fact that the morphological break between the two components
may signal the interruption of an action andthat in some cases the rst lexeme
may be a preposed copy of the second, rather than the other way round. These
possibilities will be considered with reference mainly to examples encountered
in a number of recent Afrikaans novels which, as a corpus, illustrates the
productivity of the phenomenon in the modern language. The term reduplicat-
ion will refer to Afrikaans reduplication, unless otherwise indicated.
While the existence of a link between reduplication and iconicity is often
acknowledged, the full force of iconity as a creative force in language is seldom
recognised. Thus Botha (1988: 149, 174), while referring to a quantitative
resemblance between form andmeaning, views increase as the core meaning of
reduplication: an increase in form corresponds with an increase in the projected
referent(s) of the form. Kouwenberg andLaCharit (2001: 60, 65), conceding
that some ner distinctions need to be made within the concept of iconicity, do
take a somewhat wider view but nevertheless restrict the relationship to more of
form/more of content. Though iconicity as an explanatory principle admittedly
shouldnot be allowedfree rein, we cannot gain insight into its possibilities by
imposing a priori andunmotivated restrictions on it. The present study is an attempt
to explore some of these possibilities in the eld of Afrikaans reduplication, but will
not concentrate on the daunting task of determining where iconity ends.
Reduplication, as a morphological system consisting of the unchanged
repetition or duplication of a one-word lexeme or onomatopoeic element in its
entirety, is commonly usedin Afrikaans andnot restricted to informal speech
registers. The tendency to reduplicate, sporadically attested in the Dutch vernacu-
lar (cf. Kempen 1969: 184185), is thought to have been considerably strength-
enedand extendedin Cape Dutch mainly through the inuence of Malay slaves
(Raidt 1994: 159). Malay not only has a great variety of reduplication types, but
makes frequent use of reduplications. Scholtz (1963: 161) notes a correlation
between reduplication and the tendency towards excessive descriptiveness,
expressiveness and vividness (oormatige aanduiding, nadruklikheid en aanskou-
likheid) which he considers to be a salient characteristic of Afrikaans. On the
basis of data from Dutch and Kambera, an Austronesian language, Klamer (2001,
2002) points out a correlation between formal andsemantic markedness:
expressive words, i.e. those showing emotion, negative feelings and personal
names or epithets, tendto have less usual phonological/morphological structures.
A typical reduplication is characterised by phonological unity, viz. the
absence of a phonological break between the repeatedelements, andeither equal
stress on the two components or contrastive stress on the rst.
The morphology involved in reduplication is mainly a plural and/or
diminutive sux attachedto both elements, as in rye-rye tente rows androws
of tents and voetjie-voetjie speel playing footie-footie, the p.p. prex ge- as in
geklang-klang clanged and the ge- prex signalling as does reduplication
continuous action, as in
(1) die geklop-klop van karre en hoefslae
the Nox.tap-tap of carts andhoof-thumps
the clatter of carts andhoofs (Van Rooyen 2000: 188)
while in ys-yskoud ice-cold the rst element of a compound is reduplicated.
Reduplication is, however, never employed for plural or past participle formation
in Afrikaans (in contrast to creole languages like Jamaican, e.g. brok-brok
broken, cf. Kouwenberg andLaCharit 2001: 75), presumably because of the
retention of -s or -e plural inection andthe prex ge-, respectively. What
Kouwenberg and LaCharit (2001: 73) describe as a denite cross-linguistic
trendto what might be characterizedas a diminutive reading, e.g. Jamaican
yala-yala yellowish, from yala yellow, is also conspicuously absent in
Afrikaans which has a productive sux -erig, as in gelerig yellowish.
As a morphosyntactic andonomatopoeic process, reduplication is relatedto
phenomena such as
(a) the mere repetition of a lexeme after a phonological break what Wierz-
bicka (1986: 291) wouldrefer to as clausal as against syntactic reduplication:
(2) Gou, gou! s Addie. Hulle is n hrde, en nby.
quick quick say Addie they are a horde and near
Quick, quick! says Addie. They are a horde, and close. (Van
Rooyen 2000: 606)
(3) Aag, almal s vir my praat, praat
oh everyone say to me talk talk
Oh, everyone says to me, Talk, talk! (Vermeulen 2000: 70)
(b) onomatopoeic andother names or expressions with repetitive elements, in
which the principal components are not exact copies, such as ginnegaap snig-
ger, hinkepink hobble along, holderstebolder helter-skelter, roesemoes
tumult, hadeda hadeda ibis (bird), tinktinkie wagtail (bird), toktokkie
tapping-beetle andin
(4) Prinses kom met n vreemde kloppete-klop aangedraf
princess come with a strange kloppete-klop on.ii-trot
Princess comes trotting with a strange tapping of hoofs. (Van Rooy-
en 2000: 330)
and(c) tripartite or repeatedred uplications:
(5) die tjoef-tjoef-tjoef van die lokomotief wat al hoe verder
the choof-choof-choof of the locomotive which all the further
the pung of the locomotive moving further into the distance
(Steytler 2000: 134)
(6) die gereelde klik-klik, klik-klik van die wiele op die spoorstawe
the regular click-click click-click of the wheels on the rails
the regular clicking of the wheels on the rails (Steytler 2000: 110)
Reduplication is, furthermore, diachronically present in forms such as rrig/
regtig really from reg-reg right-right, andperhaps the nal component of
kammalielies pretending, viz. lie(g)-lie(g)s by way of lying (cf. Bosho and
Nienaber 1967). M. D. Teenstra, a Dutch traveller who visitedthe Cape in 1825
andlater commentedon the language by way of a ctitious dialogue, lets a slave
reply in regardto a tame ape (Nienaber 1971: 8):
(7) Ja sueur! die aap al te danig listig, rech! rech! hij niet spreek,
yes sir the ape all too much cunning right right he not speak
om hij geen boodschap doen wil nie.
for he no errandd o want not
Yes, Sir, the ape is quite cunning. Really! He doesnt speak, because
he doesnt want to run errands.
Various parts of speech are involvedas components, viz. adverbs, as in net-
net (lit.) just-just, only just, adjectives, as in die ou-ou storie (lit.) the old-old
story, the same oldstory, substantives, as in bietjies-bietjies (lit.) small-amounts
small-amounts, in small quantities, verbs, as in vat-vat aan die tagtig (lit.)
touch-touch on the eighty, closely approach eighty, numerals, as in twee-twee
loop (lit.) two-two walk, walk two by two andinterjections, as in eina-eina
(lit.) ouch-ouch, scanty. That reduplications are often involved in category
changes, e.g. verb + verb or numeral + numeral to adverb, will be demonstrated
by subsequent examples.
The types of iconicity related to reduplication to be considered below, are:
repetition: A lexeme is repeated: an action or occurrence is repeated.
interruption or discontinuity: A lexeme is repeatedafter a break: an action
is being interrupted.
duration or continuation: Two acts of referring suggest the beginning and
endof a time-span, andthus duration.
the extension of an action or state: The occurrence of a second, and similar,
lexeme suggests a re-mention or renewedreference, andthus a continuation
of the basic state or action.
intensication: Re-mention indicates emphasis.
Of these, repetition is probably the most obvious iconic link between form and
function. What is, however, crucial, is what is repeated. After looking at
reduplication as a direct rendering of repeated sounds and actions, we pass on to
more specialisedways in which repetition can be iconic.
2. Repetition
2.1 Sounds and actions repeated
The repetition of a lexeme is directly correlated with the repetition of an
independent, cyclically complete or rhythmical action, graphically as in
(8) die Skotte in hul rokkies wat wip-wip soos kwikstertjies
the Scots in their kilt.bix.ii that hop-hop like wagtail.ii
se stertjies
crN tail.bix.ii
the Scots in their kilts that jump up like wagtails tails
(Steytler 2000: 25)
or acoustically as in
(9) die blaasbalk se gesjoe-sjoe
the bellows crN Nox.shoo-shoo
the shoo-shooing of the bellows (Matthee 2000: 582)
A combination of the graphic andacoustic are employedmetaphorically in
(10) n maalkolk woer-woer haar soos n ballerina op ys
a whirlpool whirligig her like a ballerina on ice
A whirlpool confuses her like a ballerina on ice.
(Vermeulen 2000: 113)
The woer-woer, here, is a whirligig, i.e. a childs toy. Onomatopoeic reduplica-
tion merges naturally into types such as tjoef-tjoef-tjoef, klop-klop klop-klop and
kloppete-klop mentionedabove, in which the restriction of only one, or an exact
copy of the rst component, is absent.
This kindof reduplication is lexicalisedin a bird-name such as hoephoep
Upupa epops, namedafter its repetitive call.
2.2 The naming of games
All games may be regarded as essentially repetitive. Though conventionalised
repetitive utterances play a part in Dutch childrens games (cf. Kempen
1969: 181), the games themselves are namedby means of diminutives, rather
than reduplications as is the case in Afrikaans. Reduplication is only barred when
a game already has another conventionalised name. Thus *wegkruip-wegkruip
hide away-hide away is not available as a name for wegkruipertjie hide-and-
Repetitive reduplication, routinely contextualised by the verb speel play,
has a long-standing and productive relationship with the playing of games in
Afrikaans, anything from childrens games to adult activities looked upon
perhaps ironically as games. (Also cf. Malay main-main play aroundin fun,
kidding < main play, Dodds 1977: 109.) Examples of well-established child-
rens games are aan-aan, (lit.) on-on (a touch game), klip-klip, (lit.) pebble-
pebble (a game involving the repeatedhand ling of a klip pebble) and huis-huis
or huisie-huisie, (lit.) little house-little house (in which domestic life is
enacted). These games (as perhaps all games) clearly involve the repetition of
similar actions, the repeated handling of the same objects, etc. The directive
Speel-speel (play-play) itself a reduplication followed by a proposition
(e.g. we are frogs) functions as a game initiator or speech act facilitating entry
into the worldof make-believe. Repetition functions in an incantatory way in
formulas such as Wolf-wolf, hoe laat is dit? (Wolf, wolf, what time is it?) and
Bok-bok staan styf, hoeveel vingers om joulyf? (Goat, goat, stand rm, how
many ngers on your body?). Riddles are routinely terminated by demanding,
Raai-raai wat is dit? (Guess, guess, what is this?). The following contains an
ironic reference to a story-ending formula:
(11) Ook vir hulle is dit uit-uit my storie is uit.
also for them it is whistle-whistle my story is out
They have also reachedthe endof the road. (Philander 2000: 116)
The expression wolf-wolf skree, (lit.) to shout wolf-wolf, again basedon a
repetitive speech-act, means to misleadby raising a false alarm. In
(12) hy was te besig om spioen-spioen te speel
he was too busy for spy-spy to play
He was too involvedin playing the spy. (Meyer 2000: 346)
a caricature is createdof someones incompetent attempt at spying. The associa-
tion between reduplication and playing games is also attested by a complaint in
the press directed at the police as a result of their incompetent handling of a
mock-hijack exercise involving an airliner:
(13) Ons pleit by julle, red ware lewens, los die speel-speel vir my
we pleadwith you save real lives leave the play-play to my
driejarige kleinkinders.
three-year.:b grandchildren
We pleadwith you, save real lives; leave the games to my three year
old grandchildren. (Beeld 5/3/01)
2.3 Repetition and speech acts
While utterances such as Speel-speel, Raai-raai, Wolf-wolf and Bok-bok, as were
mentionedabove, may be lookedupon as repetitive speech-acts, certain common-
ly usedad verbs of time, such as nou-nou (lit.) now-now, presently, or a short
while ago, and gou-gou (lit.) soon-soon, quickly, may well have originatedas
speech acts andstill occur in contexts reminiscent of their origin, cf.,
(14) Gaan maak nou vir die kinders iets te ete Gou-gou,
go make now for the children something to eat quickly-quickly
Do go andprepare something for the children to eat But quickly
(Van Rooyen 2000: 294)
Gou-gou derives from Dutch gauw! gauw!, described by Van Dale as an
exclamation serving to urge on or order someone to make haste. Repetitive
speech acts may be widespread among the languages of the world. The Ancient
Egyptian reduplication snsn to be friendly with, from sn brother (Loprieno
1998: 54) suggests the speech act of repeatedly calling someone Brother!,
reminiscent of the way in which Du./Afr. soebat(ten) implore, entreat, etc.
though not a reduplicated form may have derived from Malay sobat friend,
implying the repeated addressing of someone as Friend!.
Afr. nou-nou and gou-gou dier from each other in that nou-nou, in the
sense of presently, projects further into the future than nou now, while gou-
gou, as quickly, is expectedto materialise before gou. The dierence may be
ascribedto the fact that the two reduplications are basedon dierent kinds of
speech acts. Thus gou-gou, with its connotation of ecient action, retains the
sense of urgency inherent in a repeatedd irective, Quickly! Quickly!, while
nou-nou suggests a promise repeatedin order to postpone its fullment. The
intentional vagueness of the promise is highlightedby the reassuring modal
particle sommer:
(15) Die wiel is sommer nou-nou reg.
the wheel is simply now-now ready
The wheel will be ready in a short while. (Philander 2000: 35)
It is common practice to the repeat the entire reduplication, as a radio commenta-
tor was observed to do in a recent broadcast:
(16) Nou-nou nou-nou gaan ons hoor hoe
now-now now-now go we hear how
In a short while well hear how
while the following perhaps more idiolectal utterance was heard twice in
a cooking demonstration given by a woman on TV:
(17) Jy gaan die eindresultaat nou-nou-nou sien.
you go the end.result now-now-now see
Youll see the nal result in just a short while.
Scholtz (1963: 152) points out that the rapidrate of utterance characterising
repetition in reduplicated adverbs of time denoting without delay, such as nou-
nou and gou-gou (andcf. kort-kort below), may have a natural symbolic value.
Though he didnot use the term iconic, Scholtz may in fact have envisagedan
important aspect of the iconicity of reduplications, worthy of further investiga-
tion, viz. the iconic value of speedof utterance.
2.4 Repeated eorts
One reason for repeating an action is that the initial action hadbeen perceivedas
unsuccessful or not leading to any result. A report of repeated actions, as so
many eorts or attempts at reaching a goal, may therefore signal inconclu-
sive, indecisive action, or even rudimentary, symbolic or vicarious action.
Reduplication in the following corresponds to the carefulness and even the
aversion of the person confrontedby a buriedface in a thriller:
(18) Hy skop-skop die grond van die dooie gesig af.
he kick-kick the soil from the dead face o
He removes the soil from the dead face by kicking movements.
(Dreyer 2000: 102)
Botha (1988: 112115), noting the connotations of tentatively/hesitantly/non-
intensely of these reduplicated verbs, describes them as attenuated, and
denes attenuation as such as a derived unit of meaning associated with
reduplications whose verb bases have the aspectual meanings [nocNbrb] and
[NoNicNc1c:i]. Though this predicts which kinds of verbs are likely to occur
in this type of reduplication, semantic values such as hesitancy and tentative-
ness are not suciently accountedfor . It is believedthat the iconic model of
repeatede orts/attempts not only naturally constrains the class of verbs
occurring in these reduplications, but also adequately conveys senses such
tentativeness, hesitancy andthe like. As a case in point, a few more examples
will be considered.
In (19), the blinking of the eyes is a last sign of life before death sets in,
and in (20) we note the symbolic action of an aggressive dog towards the local
(19) Hy knip-knip sy o, kiel dan vooroor
he blink-blink his eyes fall then forwards
He blinks a few times, then falls forwards. (Dreyer 2000: 186)
(20) Die hond loer na Joep en lig-lig sy lip.
the dog glance at Joep and lift-lift its lip
The dog glances at Joep and lifts its lip a few times.
(Dreyer 2000: 67)
The verb voel feel, which has the connotation of careful, tentative inspection,
is frequently reduplicated. The examples (21)(23) below furthermore contain
prepositional phrases indicating an approach to rather than the mastering of an
object a prepositional rather than a direct object, in grammatical terms i.e.
once again an unsuccessful eort or subconscious action:
(21) Desperaat voel-voel sy weer aan die diamantring.
desperately feel-feel she again on the diamond.ring
Once again she desperately ngers the diamond ring.
(Van Rooyen 2000: 62)
(22) Hulle peusel-peusel aan die maalvleis en noedels
they nibble-nibble at the mince andnood les
They nibble at the mince andnood les. (Vermeulen 2000: 28)
(23) Moersleutel skreeu en hap-hap na sy aanvaller.
Moersleutel shouts andsnap-snap at his attacker
Moersleutel (a man) shouts andsnaps at his attacker.
(Dreyer 2000: 168)
Thus, the reduplicated verb expressing inconclusiveness and the prepositionally
hedged object are seen to complement each other semantically.
Wil will andits preterite, as in wou-wou is not only the only modal
verb to be reduplicated, but is typically non-volitional and often has a non-human
agent or theme, threatening to enter into a certain state, e.g.
(24) Die kas wil-wil omval.
the cupboardwill-will fall-over
The cupboardis on the verge of toppling over. (own example)
(25) Sy wil-wil ou word, maar sy sukkel vorentoe.
she will-will pass out but she struggle forwards
She is on the verge of losing her consciousness, but struggles for-
wards. (Vermeulen 2000: 107)
Even more than the verb voel feel, the reduplicated adverbs byna and amper
almost correspondlexically to the iconic sense of goal not reached or being
on the verge of, viz.
(26) hy het haar byna-byna laat val
he have her almost-almost let fall
He almost dropped her. (Van Rooyen 2000: 186)
Italian quasi quasi seems to be an equivalent (cf. Wierzbicka 1986: 289). In the
case of net-net (lit.) just-just the goal is reached, but only just, cf.
(27) Die geluid is net-net verstaanbaar.
the sound is just-just comprehensible
The soundis barely comprehensible. (Meyer 2000: 303)
What these adverbial examples have in common, is the reapplication and
consequent renement of an evaluation.
2.5 Repeated acts of perception
Reduplications are often employed in a distributive, disjointed or intermittent
fashion, as in
(28) sien hy plek-plek skurwe rotse
see he place-place ruggedrocks
He sees ruggedrocks in places/here andthere. (Steytler 2000: 6)
(29) die sandgruis laat hom plek-plek gly
the sandy-gravel let him place-place slip
The sandy gravel causes him to slip in places/from time to time.
(Steytler 2000: 3)
(30) ek vergeet plek-plek dis Paasnaweek.
I forget place-place its Easter-Weekend
At times I forgot that it was Easter Weekend. (Philander 2000: 147)
Though stuk-stuk (lit.) piece-piece may be distributive, it has also been
lexicalisedas consecutive, cf.
(31) Ek het haar al so stuk-stuk begin terugbetaal.
I have her already so piece-piece begin.ii repay
I have already begun to repay her piecemeal. (Beeld, 8/12/01)
In examples (28) to (30) we may notice a progression from the spatial to the
temporal or, respectively, co-ordinates on a surface (28), points on a route (29),
andsubsequent moments in time (30). This represents a fragmentedrather than
holistic mode of perception and reects a non-simultaneous, distributive or
consecutive rather than static or encompassing perspective, even when no
movement is expressed, as in (28). This and probably other kinds of redupli-
cation as well are to be understoodas basedon repeatedacts of perception
or observation andstrongly corroborates the iconic nature of repetition. This
wouldbe a subtype of what Tabakowska (1999: 410) following Enkvist
refers to as experiential iconicity, which involves a more or less direct relation
between linguistic expression andperceptual relationships.
A relationship between distributed numbers and reduplication is in evidence
in some quite disparate languages. For instance, Loprieno (1998: 72) states that
in Ancient Egyptian distributive numbers, e.g. one each, were formed through
a reduplication of the cardinal number. Moravcsik (1978: 310, 318) points out
that one by one is derived from one, two by two from two, etc. through
partial reduplication in Aztec, and four by four from reduplicated four in
Turkish. Malay satu-satu one by one derives from satu one, and dua-dua in
twos from dua two (Dodds 1977: 103, 109).
To make the notion of repeatedacts of perception more precise: The distribu-
tion of a group of objects or the texture of a surface which might otherwise be
described as in groups of two, groupwise, patchy, etc., might also be described
by way of spatially shiftedtakes, viz. I see two here, then again the same number
over there, then the same number yonder, and so forth. An important characteris-
tic of the hypothesis of repeatedacts of perception, occurring of necessity one
after the other, is that no distinction is made between spatial and temporal
distribution. The interpretation if it is at all necessary to choose between the
two will depend on contextual factors. Thus, of the next three examples, the
rst two are clearly more temporal, the thirdmore spatial in keeping with the
verbs doof put out, beweeg move and staan stand, respectively:
(32) So word die ligte een-een gedoof.
become the lights one-one put-out
Thus the lights are put out one by one. (Van Rooyen 2000: 287)
(33) waar die kommando teen n skuinste af beweeg, soms
where the commando against a slope down move sometimes
twee-twee, soms drie-drie
two-two sometimes three-three
where the commando move down a slope, sometimes in twos, then
again in threes (Steytler 2000: 137)
(34) hulle staan klompies-klompies en braai
they standlittle.groups-little.groups andbarbecue
They are barbecuing in little groups. (own example)
While the reduplication of diminutives, as in *klompie-klompie, is morphologi-
cally excluded, except in game names or onomatopoeic forms, other distributives
may be in the singular or plural, e.g.
(35) die sout slaan kol-kol / kolle-kolle teen die muur uit
the salt show patch-patch/ patches-patches on the wall out
Salt exudes in patches on the wall. (own example)
A reduplication commonly used is kort-kort (lit.) short-short, denoting both
frequent occurrence andshort interval, as in
(36) Die telefoon lui kort-kort.
the telephone ring short-short
The telephone is ringing very frequently. (Vermeulen 2000: 29)
It is perhaps the only reduplication not derivable from its component kort short,
denoting for a short while as adverb of time (cf. Kempen 1969: 184). In the
reduplication, kort may well be an ellipsis of an expression such as na n kort rukkie
after a short while. Though it is a moot point whether lexical iconicity (short:
abbreviated) might have played a part here, lank-lank (lit.)long-long conspicu-
ously lacks an interpretation such as infrequently or after long intervals.
3. Interruption or discontinuity
According to Kouwenberg and LaCharit (2001: 67) the distinction between
continuous andd iscontinuous action not only characterises certain languages but
is also encounteredcross-linguistically . In this section and the next, examples
will be adduced for Afrikaans.
Even though a reduplication, unlike the repeated words of (2) and (3) above,
is a phonological unity without a pause between the two components, it is
nevertheless characterisedby the end of one lexeme andthe beginning of another
in the middle of the unit, i.e. a clear morphological break. Iconically, this break
can be employedto signal an interruption in durative action, as in
(37) Frans vorder rus-rus.
Frans progress rest-rest
Frans progresses while resting from time to time.
(Van Rooyen 2000: 623)
(38) Tant Dora en oom Ben dra drie opvou-veldstoeltjies vir die
aunt Dora anduncle Ben carry three folding-campstool.bix.ii for the
sit-sit oorkant toe.
sit-sit other-side to
Aunt Dora andUncle Ben carry three folding chairs for sitting on
while crossing over to the other side. (Steytler 2000: 114)
a forcednight-time walk by two immobile, oldpeople is described. Also cf.,
(39) Die stoet beweeg staan-staan.
the procession move stand-stand
The procession is frequently halted. (own example)
In contrast, the sense of interruption is absent in
(40) Die kleintjies word staan-staan gewas.
the little-ones become stand-stand ii.wash
The little ones are being washedstand ing up.
(Van Rooyen 2000: 161)
Interruption or discontinuation may also play a part in types such as plek-plek,
kol-kol, stuk-stuk and bietjies-bietjies mentionedabove. Thus a plek is a
restricted place, spot or point on a line or surface, depending graphically
on its intervening discontinuation or its interruption by non-plek. In the following
the areas of interruption appear to surpass the focus areas in size:
(41) lyk die omgewing nounog net plek-plek vir haar bekend
look the surroundings now only just place-place her to known
the surroundings now only appear familiar to her in places (Ver-
meulen 2000: 85)
4. Duration or continuation
Reduplication may be iconic of duration or continuation either because the two
points of reference impliedby the two constituents may signal the beginning and
endof a time-span, or because the repetition of a lexeme may indicate an
ongoing situation or a sustainedstate. Duration may refer to an action or state
persisting during the entire length of an otherwise delimited time-span. Accord-
ing to Moravcsik (1978: 328, note 13) continuedaction is the semantic category
most likely to be universally expressible by reduplication. Three functional types
will be considered here as possible components in durational reduplication: verbs,
adjectives and numerals. Though the verbal reduplications in question may have
repetitive or interruptive aspects, the durational aspect seems to predominate.
As far as verbs are concerned, it is possible to distinguish between the
(a) The principal action is expressedby means of the reduplication itself, as in
(42) Onvermydelik hoe-hoe n uil mistroostig in die kloof
inevitably hoot-hoot an owl melancholically in the ravine
The inevitable owl hoots melancholically in the ravine.
(Van Rooyen 2000: 619)
(b) The reduplication is a typical adverb of manner, i.e. it describes the way in
which the main action takes place, e.g.
(43) Die klomp kinders is toe blr-blr agter hom aan
the group of-children is then bleat-bleat after him on
The group of children then followed him bleating.
(Philander 2000: 96)
Adverbs are also derived from verbs through reduplication be it partial in
Thai, as in krasb-krasab in whispers from krasb to whisper (Moravcsik
1978: 325). The actions of main verb andred uplication may be inextricably
boundup with each other, as diving andtumbling in
(44) tot een van die rollerduiwe valvlieg-valvlieg hok toe
until one of the roller-pigeons dive.y-dive-y pigeon-house to
until one of the roller pigeons tumbledtoward s the pigeon house,
alternately diving and ying (Philander 2000: 157)
This type of reduplication is lexicalisedin the stereotypedmetaphorical expres-
sions lag-lag laugh-laugh and speel-speel play-play, both meaning easily,
without encountering any problems, as in
(45) met julle as borge sal ek speel-speel n lening kry by enige
with you.ii as guarantors shall I play-play a loan get at any
With you as guarantors Ill easily raise a loan at any bank. (Philan-
der 2000: 148)
(c) Though the actions expressedby the main verb and by the reduplication are
simultaneous, it is possible that neither of them may be the principal action, cf.
the near synonymity of:
(46) Hulle sing swem-swem.
they sing swim-swim
They sing while swimming.
(47) Hulle swem sing-sing.
they swim sing-sing
They swim while singing.
In the following there seems to be no intrinsic connection between Kallies
actions andthe conversation he is keeping up all the time:
(48) Kallie raak praat-praat haastiger tot hy later t ver weg is en
Kallie get talk-talk more-hurrieduntil he later too far away is and
ons gesels vanself doodloop
our conversation of-itself stop
Kallie became more andmore hurried, talking all the while, until he
was too far away eventually andour conversation broke o of its
own accord. (Philander 2000: 98)
When reduplicated verbs function as adverbs, reduplication may be said to
have the function of signalling category change. When, however, it is possible to
substitute a non-reduplicated adjective or adverb for a given reduplication, as is
often the case, the question arises as to what the function or added value
of the reduplication is. That reduplications are synonymous with the basic form
they derive from is, according to Moravcsik (1978: 317, note 11), the exception
rather than the rule. Thus Botha (1988: 100) claims though he retracts the
claim later on (p. 122) that reduplications such as vaak-vaak (lit.) sleepy-
sleepy, skaam-skaam (lit.) shy-shy or traag-traag (lit.) tardy-tardy convey
a sense of empathy not expressedby the nonreduplicatedforms. Scholtz
(1963: 154) remarks that reduplication seems restricted to lexemes with a less
favourable or negative meaning, such as dirty or ill, and attributes the
reduplication to the emotivity relatedto actions andsituations not perceived as
normal. In connection with emotivity andred uplication as a marked structure, cf.
once again Klamer (2001, 2002). Thus (to use andad apt some of Scholtzs
examples) one wouldgo to work but not to hospital siek-siek, (lit.) ill-ill,
while being ill, one might pack away the dishes vuil-vuil (lit.) dirty-dirty, while
still dirty, one would wear the shirt stukkend-stukkend tattered rather than heel-
heel intact, etc.
The negativity andemotivity is much less conspicuous when a person is
saidto have disappeared stil-stil (lit.) quietly-quietly, though a furtive element
is hintedat:
(49) Dis beter dat sulke mense liewer stil-stil verdwyn.
its better that such people rather quietly-quietly disappear
Its better that such folk rather disappear quietly.
(Vermeulen 2000: 70)
Malay diam-diam from diam to be silent is glossednot only as silently but
also as secretly (Dodds 1977: 109). A distinction of even greater subtlety is
between doing things saam-saam (lit.) together-together rather than just saam
together. This may hint at a greater sense of co-operation (i.e. a certain
measure of empathy), but may simply have the emotively neutral implication that
certain individuals had opted to approach a task collectively rather than as
individuals. It is after all possible to do something alleen-alleen (lit.) alone-
alone, completely on ones own as well. This is still in keeping with what
Haiman (1983: 807) refers to as the generally acceptedaxiom that what is
predictable receives less coding than what is not.
A more general claim, which wouldbe compatible with the suggestions
mentionedabove, would be that these reduplications carry the focus or a sub-
focus on new or unexpectedinformation, express a contrast with or are non-
compatible with the main proposition, or even express irony. Irony is, for
example, evident in a letter to the press (50), and a mothers complaint about the
dangerous places visited by her son (51):
(50) Dink hy eerlik ons kan vroom-vroom hier aan die suidpunt
think he honestly we can pious-pious here at the southern-point
van Afrika sit en kyk watter kant die wind gaan waai?
of Africa sit andwatch which side the windwill blow
Does he honestly think that we can wait piously here at the southern
tip of Africa to see how things will develop?
(Letter to Beeld, 18/9/01)
(51) Netnoukom Krols dood-dood hier aan, kla Mmeme by come Krols dead-dead here on complain Mmeme with
Karools may arrive here dead, for all we know, Mmeme complain-
edto Betta. (Van Rooyen 2000: 202)
Even where numerals are involved, this contrast may be observed:
(52) Een-een tel hulle onder sy beste vriende. Maar dis die nsies
one-one count they under his best friends but its the nations
Looked at in isolation, i.e. as individuals, they rank amongst his best
friends. But its the nations (Van Rooyen 2000: 304)
The contrast here even though een one cannot standin isolation in this
context is between their status as individuals and their status as members of
a nation.
5. Extension
The occurrence of a second, and similar, lexeme suggests a re-mention or
renewedreference and therefore a continuation or an extension of the basic
action or state. The verbs reduplicated in the following examples already express
repetition or duration; reduplication seems to signal further extension of the
action or state, in the rst case of a long trip on the back of an ox, in the second
of a menacing situation:
(53) Thoms gemaklik dwars op sy rug, skommel-skommel
Thoms comfortably athwart on its back rocking-rocking
Thoms [perched] comfortably athwart its back, rocking from side to
side. (Matthee 2000: 301)
(54) Sy skud-skud die koeltjie in haar hand.
she shake-shake the bullet.bix in her hand
She keeps on shaking the bullet in her hand. (Dreyer 2000: 97)
6. Intensication
There is, on the one hand, a direct and obvious relationship between repetition
andemphasis, a form of intensication, and some less obvious relationships, on
the other. Re-mention directly enhances the importance of what is mentioned.
Intensifying repetition is encounteredeven in languages not generally character-
isedby reduplication, cf. Eng. very, very good, French trs, trs cher. The
emphasis may, however, also be a derivative of what Wierzbicka (1986: 297)
refers to as the accuracy or validity of the word used or quotes (from Lepschy
andLepschy) as the identication of an authentic quality. Thus a mns-mens
(lit.) human-human wouldbe a human being displaying genuine human values.
Intensication borders on emotive language, the language of poetry,
nostalgia, hyperbole. A poem by N. P. van Wyk Louw on a tiny chisel which
eventually split the universe in half, begins with this line:
(55) Ek kry n klein klein beiteltjie
I get a small small chisel
I take up a tiny chisel
A folk-song with Malay roots starts as follows:
(56) Lekka lekka ywe, lat die ghantang nader skywe
tasty tasty onions let the lover nearer move
Tasty, tasty onions make the lover move closer to you
A quiz item, Flinkdink (fast-think), in a magazine is advertised as
(57) die vinnige speletjie vir slim-slim mense
the fast game for clever-clever people
the fast game for very clever people
Further examples that suggest intensication are:
(58) die teertoue van ou-ou skepe
the tarred-ropes of old-old ships
the tarredropes of ancient ships (Van Rooyen 2000: 95)
(59) baie-baie jare gelede
many-many years ago
very many years ago (Van Rooyen 2000: 24)
(60) sy wange is ys-yskoud, asof die dood op hom is
his cheeks are ice-ice-coldas-if the death on him is
His cheeks are icy, as if death is imminent. (Van Rooyen 2000: 127)
In some of the above examples, such as (56)(58), emphasis is enhanced
through diagrammatic iconicity. Scholtz (1963: 154155) points out the
vividness (aanskoulikheid) with which one thick pu of smoke after the other
is delineated in dik-dik skuiwe rook (lit.) to smoke thick-thick pus where
the serial character of the action, implicit in the nominal plural, is brought out
and made more spectacular by the reduplicated adjective. Similarly the ou-ou
skepe ancient ships of (58), when individualised, seem to retain more of their
former glory. Reduplication of the adjective serves a grammatical purpose by
indicating plurality in Yoruba, where custom bad-bad is understood to mean
badcustoms (Moravcsik 1978: 307).
But the examples of intensication adduced above may be iconic in yet
another way. This will be claried by two observations. Firstly, when the two
constituents of a reduplication do not receive equal stress (barring unimportant
phonetic dierences between them), the rst andnever the last constituent
may receive the principal stress. That phonological stress in such cases is indeed
on the rst constituent in English, is illustratedby Wierzbicka (1986: 300301).
Secondly, the reduplications shown above appear exclusively in attributive or
pre-headcontexts, i.e. in the SpecierHead conguration. Thus SpecierHead
seems to be the preferredcontext for intensifying reduplication. If one assumes
that the reduplication itself is also an instance of SpecierHead, i.e. an exten-
sion of the higher order construction, this would imply (a) that intensifying
reduplication is also supported by syntactic iconicity, and (b) that the copy is
not the secondconstituent, as is generally assumed, but in actual fact the rst, in
what one couldterm anticipatory reduplication. This wouldexplain why only
the rst constituent, but not the second, may receive the principal often
contrastive stress. Leftwardred uplication, but only partial, weakly stressed
andwith a grammatical function, is also exemplied by Gothic reduplicated
strong preterites, such as hahit, lalot and tatok.
A similar explanation holds for reduplicated plurals expressing emphasis,
which also prefer the attributive context. Even an isolatedplural, when emphasis-
ed, may express a great number of rather than one or two. In rye-rye tente
rows androws of tents reduplication ties in with the inherent capability of the
plural to express great number of. *ry-ry tente row-row of tents is unaccept-
able precisely because the singular refers to a structure, container, etc. but does
not possess this inherent gradability.
7. In conclusion
An analysis of a linguistic phenomenon such as reduplication from the point of
view of underlying iconic principles, reveals that iconicity as an explanatory
principle, rather than partaking of a formal system, intersects with grammar,
context andtext at any point or level, andthat various kinds of iconicity may
interact at any given moment. Iconicity may vary from a general principle such
as repeated perception to the individual case of a lexical item such as kort-kort
(lit.) short-short, every so often, where kort short might be elliptical
shortened for a phrase such as oor n kort rukkie after a short while. Iconic
principles are furthermore anti-systemic in that they may appear to be in conict
with each other and yet individually feasible, e.g. are able to signal continuation
andextension but also interruption. Attempts to harmonise all interpretations with
a single semantic value, such as increased in the case of Botha (1988: 143), or
to incorporate them in one complex semantic network, taking repeat (herhaal)
as point of departure, as in the case of Van Huyssteen (2000: 227), at all costs,
tendto be contrived.
Iconic principles may cut through grammatical categories. For example, on
the one hand, lexical verbs such as voel-voel feel-feel, modal verbs such as wil-
wil will-will andad verbs such as byna-byna almost-almost as reduplications
all have in common a sense of thwartede ort or incomplete action. On the
other hand, voel-voel feel-feel as a whole may function as a verb or as an
adverb. Reduplications thus resemble ideophones in showing a relative absence
of syntax (Klamer 2002:260). Iconic models are, moreover, productive in that
a principle such as repetition, for instance, is not only directly iconic, but
combines with various discourse and perceptive principles to create new ways of
linguistic expression.
Allen, R. E. (ed.) 1992. The Concise OxfordDictionary of Current English. London, etc.: BCA.
Bosho, S. P. E. andG. S. Nienaber, 1967. Afrikaanse etimologie. Pretoria: Die Suid-
Afrikaanse Akademie vir Wetenskap en Kuns.
Botha, R. P. 1988. Form and Meaning. A Study of Afrikaans Reduplication. Cambridge:
Cambridge University Press.
COD, see Allen 1992.
Dodds, R. W. 1977. Malay. (Series: Teach Yourself Books). Sevenoaks: Hodder &
Dreyer, T. 2000. Stinkafrikaners. Kaapstad: Tafelberg.
Haiman, J. 1983. Iconic andeconomic motivation. Language 59: 781819.
HAT, see Odendal 1994.
Kempen, W. 1969
. Samestelling, aeiding en woordsoortelike meerfunksionaliteit in Afri-
kaans. Kaapstad: Nasou.
Klamer, M. 2001. Expressives andiconicity in the lexicon. In Voeltz andKilian-Hatz
2001, 165181.
Klamer, M. 2002. Semantically motivatedlexical patterns: A study of Dutch and
Kambera expressives. Language 78: 258286.
Kouwenberg, S. andD. LaCharit. 2001. The iconic interpretations of reduplication:
Issues in the study of reduplication in Caribbean Creole languages. European
Journal of English Studies (Vol. on Iconicity, O. Fischer andM. Nnny, eds.) 5:
Loprieno, A. 1998. Ancient Egyptian. A Linguistic Introduction. Cambridge: Cambridge
University Press.
Matthee, D. 2000. Pieternella van die Kaap. Kaapstad: Tafelberg.
Meyer, D. 2000. Orion. Kaapstad: Human & Rousseau.
Moravcsik, E. A. 1978. Reduplicative constructions. In Universals of Human Language,
J. H. Greenberg (ed.), 297334. Stanford: Stanford University Press.
Nienaber, G. S. 1971
. Afrikaans uit die vroer jare. Johannesburg: Voortrekkerpers.
Odendal, F. F. et al. (eds.) 1994
. Verklarende handwoordeboek van die Afrikaanse taal.
Midrand: Perskor.
Philander, P. J. 2000. Rebunie. Kaapstad: Human & Rousseau.
Raidt, E. H. 1994. Oor die herkoms van die Afrikaanse reduplikasie. In Historiese
taalkunde; studies oor die geskiedenis van Afrikaans, E. H. Raidt (ed.), 148160.
Johannesburg: WitwatersrandUniversity Press.
Scholtz, J. du P. 1963. Reduplikasieverskynsels in Afrikaans. In Taalhistoriese opstelle,
J. du P. Scholtz (ed.), 146161. Pretoria: J. L. van Schaik.
Steytler, Klaas. 2000. Ons oorlog. Kaapstad: Tafelberg.
Tabakowska, E. 1999. Linguistic expression of perceptual relationships: Iconicity as a
principle of text organization (A case study). In Form Miming Meaning. Iconicity
in Language and Literature, M. Nnny and O. Fischer (eds), 409422. Amsterdam:
John Benjamins.
Van Dale Groot Woordenboek der Nederlandse Taal. 1984. Utrecht/Antwerpen: Van Dale
Van Huyssteen, G. B. 2000. Die reduplikasiekonstruksie in Afrikaans: Enkele aspekte
van n kognitiewe gebruiksgebaseerde beskrywingsmodel vir Afrikaans. Unpub-
lishedPh.D. thesis, Potchefstroom University for C. H. E.
Van Rooyen, E. 2000. Vuur op die horison. Kaapstad: Tafelberg.
Vermeulen, J. 2000. Geraamtes dra nie klere nie. Kaapstad: Tafelberg.
Voeltz, F. K. E. andC. Kilian-Hatz (eds) Ideophones. (Series: Typological Studies in
Language.) Amsterdam: John Benjamins.
Wierzbicka, A. 1986. Italian reduplication: cross-cultural pragmatics and illocutionary
semantics. Linguistics 24: 287315.
</TARGET "conr">
AUTHOR "Volker Harm"
TITLE "Diagrammatic iconicity in the lexicon"
WIDTH "150"
Diagrammatic iconicity in the lexicon
Base andd erivation in the history
of German verbal word-formation
Volker Harm
Deutsches Wrterbuch von Jacob und Wilhelm Grimm, Gttingen
1. Iconicity in the lexicon
While it has become more or less acceptedamong cognitive linguists that
iconicity plays an important role in grammar andgrammatical change, the
iconicity of the lexicon itself andits relevance to lexical change is still in
dispute. Generally, iconic lexical structures are considered as rather marginal:
Iconic words [] are infrequent. Iconic syntactic constructions are the rule as
John Haiman puts it (1998: 171; cf. 1985: 10). If we look at the present state of
research, this scepticism regarding the iconicity of the lexicon seems to be
correct: the phenomena studied so far under this heading are mostly restricted to
rather peripheral areas, especially to so-calledonomatopoetic or phonsthetic
expressions like cuckoo, cock-a-doodle-do, or immer, icker etc. which rely on
the sound-symbolic eects of certain vowels, consonants andconsonant clusters.
Only very recently, e.g. in an article by Friedrich Ungerer (1999), has an attempt
been made to go beyondthis image-centered level of lexical iconicity andto take
Peirces second type of iconicity, the diagram, into account. According to Peirce,
a diagram is a complex sign, representing a complex concept. The essence of a
diagram is that the relationship between the parts of a complex sign resembles
the relationship between the parts of the concept which it represents. In my paper
I will focus on two major kinds of diagrammatic iconicity. The rst is what is
often referredto as the isomorphic principle, that is the one-to-one correlation
between form andcontent, or, to borrow Dwight Bolingers classical claim: the
natural condition of a language is to preserve one form for one meaning, and one
meaning for one form (1977: x). It goes without saying that this natural
condition of language remains an ideal which is never reached in linguistic
reality. As a matter of fact, most words of a language are polysemous or at least
polyfunctional, andevery language of the world seems to have synonyms and
homonyms, andthis clearly contradicts the one-form-one-meaning hypothesis.
The notion of isomorphism between one meaning andone form becomes more
realistic if we accept that one meaning may standfor one prototypically clustered
set of meanings connectedthrough family resemblances (cf. Geeraerts
1997: 124). This reinterpretation of the isomorphic principle on the basis of
prototype theory makes it more useful for lexicological research. From the point
of view of prototype theory, polysemy within a prototypically organizedcatego-
ry, then, will not be regarded as a violation of the isomorphic principle. Only
polysemy among distinct, prototypically organized categories constitutes a non-
isomorphic conguration (cf. Geeraerts 1997: 132).
The second kind of diagrammatic iconicity I want to discuss consists of a
correspondence between morphological and semantic markedness. The notion
that morphological markedness corresponds to semantic markedness is crucial to
what Haiman calls motivational iconicity. According to Haiman, categories
that are markedmorphologically [] are also markedsemantically (Haiman
1980: 528; similar concepts are constructional iconicity in Natural Morphology
[Mayerthaler 1987: 4849] andthe quantity principle [Givn 1995: 4950]). A
simple example of such a markedness correspondence between signier and
signied is plural formation. In English (as in the majority of languages) a
morphologically complex plural, e.g. dog-s, is usually opposedto a morphologi-
cally simple singular dog. Here it is the formal contrast between a substantially
realizedmorpheme andzero which is an icon of a conceptual contrast between
less andmore, between semantic simplicity andcomplexity . In the lexicon,
markedness correlations between form and content are not so easily established.
As a matter of fact, polysemous words in most cases have more than two
meanings. The traditional Jakobsonian notion that all oppositions are privative
andbinary thus does not holdfor such words. What we needis a notion of
markedness which allows for gradual oppositions between a greater number of
elements. Here, again, prototype theory oers a number of useful insights.
Prototype theory postulates that most, if not all, linguistic information is orga-
nizedin cognitive categories with a so-calledrad ial structure of increasingly
peripheral members relatedto a central prototype. In other words, a cognitive
category consists of a number of gradual oppositions between each member of
the category andits prototypical centre. As Janda (1996) has convincingly
shown, it is this relative distance from the prototypical centre which allows us to
assign markedness values to the elements of a radial category. Thus, semantic
markedness can mainly be seen as a function of distance from the prototype of
a conceptual category (cf. also Van Langendonck 1999: 567569).
As convincing as this denition of conceptual markedness is, there remain
some practical problems with the exact determination of conceptual distance,
especially when we are dealing with historical data. Needless to say, the normal
way of dening the central and peripheral members of a cognitive category
goodness-of-example ratings, priming tests andso on is ruledout here. A
theoretically feasible way of determining the prototype structure of historical
word-meanings is the measurement of textual frequency. However, such a
quantitative approach poses signicant practical problems. The main problem for
German is that no text corpus covering major parts of the language history is
Even if we hadsuch a corpus at our disposal, it is far from certain
that a quantitative analysis wouldad mit conclusions about prototype structures.
The extreme heterogeneity of such a text collection, containing translations of the
bible, philosophical texts, heroic epics, legal texts, andso on, makes it dicult
to draw an inference from quantitative ndings to cognitive structures. Therefore,
a large-scale investigation of certain tendencies of language development
reaching from OldHigh German to present-day German, as I am going to
present here, cannot exclusively rely on frequency analyses. But there is another
possibility of determining degrees of prototypicality or semantic markedness for
historical data. Geeraerts (1997: 6061) draws attention to an important character-
istic of lexical structure andlexical change which allows veriable predictions:
Lexical innovations normally take the form of an expansion of the prototypical
centre of a category, whereas peripheral instances of a lexical item do not give
rise to new meanings. This has been shown in various synchronic andd iachronic
case studies on the structure of cognitive categories (cf. Geeraerts, Grondelaers
andBakema 1994 on words for womens clothing in contemporary Dutch, and
Geeraerts 1997: 2325; 3247 on semantic changes).
On the basis of this general characteristic of cognitive organization, we can
conclude that meanings which can be identied as the starting point of metaphor-
ical, metonymical or other innovations are likely to be situatedclose to the centre
within the semasiological range of a lexical category. By contrast, meanings
which are not or only incidentally the source of semantic extensions are likely to
be peripheral. This observation can be usedas a basis for the determination of
markedness or prototypicality of word meanings because, as a large number of
studies on historical semantics has shown, we are normally able to determine
whether a meaning is the source or the target of a lexical expansion, even when
dealing with historical data.
2. Pleonastic prexes in the history of German word-formation
Let me now turn to an application of this theoretical outline by examining some
developments in the history of German word-formation which, as I intend to
demonstrate, show a tendency towards iconic coding. Since these developments
concern the relationship between a verbal base andits prex forms it might be
helpful to give a brief outline of some general characteristics of verbal prex-
ation in the history of German. From OldHigh German to present-day German,
the main function of prexation is the semantic (or syntactic) modication of a
base. However, there have always been some prexes which do not change the
meaning of the stem. They just addsome kindof intensication to the stem, or,
to be more precise, to one meaning of the stem if the stem is polysemous. In
such cases, stem and prexed verb may be consideredto be synonymous or
partly synonymous. An example of such a partial synonymy is the modern
German verbleiben to remain, which is interchangeable with bleiben to stay,
to remain in many contexts.
This intensifying perhaps a better term would
be pleonastic function is, of course, never the only function of a prex;
most prexes of German, especially OldandMid dle High German, are multi-
functional (just as English prexes). It is with regardto such an intensifying or
pleonastic function of prexes that modern German word-formation diers
signicantly from word-formation of the older language periods. This function
is very rare in the present-day language, but, as Khnhold (1973: 356), Solms
(1989: 2830) and others have demonstrated, it was much more common in the
older periods. Those pleonastic prex forms that did survive have been lexical-
izedto a greater or lesser extent.
Hans-Joachim Solms (1989: 21) postulates that pleonastic prexations play
a major role in a process of language change which he refers to as the restruc-
turing of the lexicon (Umstrukturierung des Lexikons) and which took place
from late Middle to New High German. According to Solms, this restructuring
proceeds as follows: In Old and Middle High German there were considerable
semantic overlappings between the prexed verb andits base. In the develop-
ment of late Middle and Early New High German, these synonymies were
reduced continuously. This reduction of synonymy implies a reduction of
polysemy both of the base andits derivation. Meanings which were formerly
expressedby the simple verb as well as by its prexed counterparts are now
attachedto only one of them. Thus, the restructuring of the lexicon can be seen
as an increasing division of labour, as it were, between base and prexed verb.
Solms suggests that this process might be interpreted as a tendency towards a
higher degree of isomorphism between form and content.
3. A case study: hren, erhren, gehren, verhren
Whereas Solms gives only a sketchy account of increasing isomorphism, I will
now rst make an attempt to present a more detailedandprecise picture of its
mechanism. For this purpose I will focus on one example, the development of
hren andits derivations erhren, gehren, verhren.
Second, I will demonstrate
that this development not only involves a tendency towards a higher degree of
isomorphism, but can also be seen as expressing a tendency towards what
Haiman terms motivational iconicity. Third, I will try to substantiate my
ndings by presenting a list of analogous cases.
The verb hren is documentedas early as the endof the eighth century, that
is, the very beginning of the history of the German language. Under the entry
hren, the Althochdeutsches Wrterbuch (KFAW 4, 124045) lists the following
meanings: (a) to hear, (b) to listen to, (c) to comply with a request, (d) to
obey, to follow, (e) to belong to, to be part of, (f) to stop. Most of the
citations relate to the meanings (a) to hear and(b) to listen to. Meanings (c)
to (f) are less frequent andrather unusual from the point of view of contempo-
rary German. They can, however, be illustratedby the following Old High
German citations:
(1) pibotum [] in alleem horran (praeceptis abbatis in
the orders in everything to-follow
omnibus oboedire)
to follow the orders [of the abbot] in every respect (Benediktiner-
regel, KFAW 4, 1245)
(2) horint alliudiu teil ze einero hant
belong all these parts to one hand
All these parts belong together (Notker, KFAW 4, 1245)
(3) got miner, ih haren dir be tage, daz ne horest tu
Godmy I wait you at day this not hear you
My God, I have been waiting for you all day, but you do not hear
me (Notker, KFAW 4, 1244)
(4) diu ougen dero scalcho die man llet. uuartent ze iro
the eyes of-the servants which one smack look at of-their
herron handen [] uuieo halto sie horren uuellen
masters hands how soon they stop want
The eyes of the servants which are smackedlook at the hands of
their masters, when they will stop
(Notker, KFAW 4, 1245)
Meanings (c) and (d) may be regarded as metonymic extensions of either (a) or
(b): within the frame of oral communication between socially unequal interlocut-
ors obeying andcomplying both presuppose having heardthe instruction or the
prayer. (d), to belong to, may also be regarded as a metonymy, but its source
is to obey, to follow rather than to hear/to listen: In the context of a feudal
society, to belong to somebody may be seen as a typical implication of to
obey, to follow somebody who is in a superior position. In a second step, the
meaning to belong to somebody who is in a higher position is generalized to
to belong to, to be part of (see Figure 1 for an overview of these semantic
The origin of (f) OHG hren to stop is not entirely clear. To stop seems
to have developed from to listen to, to prick up ones ears. This becomes more
plausible if we reconstruct a frame like to listen attentively to something andto
suspendall other activities in order to concentrate on the sense perception.
This short analysis of the conceptual interdependencies within the semasio-
(a)/(b) to hear/to listen
(c) to comply with a request, to hear a prayer
(d) to obey, to follow
(f) to stop
(e) to belong to, to be part of
Figure 1. Semantic expansions of OHG hren to hear, to listen.
logical range of OHG hren shows that meanings (c) to (f) have developed either
directly or indirectly from (a) to hear and (b) to listen, while (a) and (b) have
not been the target of semantic expansions, at least as far as we are able to trace
back the semantic history andprehistory of the word. Therefore we may safely
conclude that to hear, to listen is the prototype of OHG hren, whereas the
other meanings are more or less peripheral. This is supportedby the fact that,
according to the data presented in the KFAW-entry, to hear and to listen are
the most frequently attestedread ings of this item.
MHG hoeren still exhibits more or less the same range of meanings as
OHG hren. Frequency analyses of selected Middle High German texts clearly
show that to hear/to listen is still the most frequently attestedmeaning.
Hence, the prototype structure of the Middle High German word seems to be similar
to the OldHigh German one. The only dierence is that the readings to obey,
to follow, to hear a prayer, to comply with a request are not documented in
the dictionaries of Middle High German. (As these dictionaries, for the most part,
date back to the nineteenth century, this may merely reect lexicographical
lacunae rather than a real gap in wordhistory .)
The history of (Early) NHG hren is characterizedby two striking losses of
language period
8 9 10 11 12 13 14 15 16
(a) to hear
(b) to listen
(c) to comply
with a request,
hear a prayer
(d) to obey,
to follow
(e) to belong to
(f) to stop
17 18 19 20
Table 1. The history of hren and its principal meanings
meaning. From 1400 up to the present the meanings to stop andto belong to
simply disappear. The last occurrences of to stop date back to the sixteenth
century, to belong to became extinct during the eighteenth century. However,
the central position of to hear/to listen within the conceptual structure remains
unchangedfrom Old High German to present-day German. (Table 1 presents an
overview of the history of hren.)
The prexed OHG gihren is attestedas early as the eighth century. The
semasiological range of gihren largely corresponds to that of hren. Apart from
to belong to andto stop, all meanings of the base are documentedfor the
prexed verb as well, i.e. (a) to hear, (b) to listen to, (c) to hear a prayer, to
comply with a request, (d) to obey, to follow may be expressed by either the
stem or the derivational form.
(5) uzouh iu thie dar gihoret, quidu ih: ir birut salz erda
but to-you which there listen tell I you are salt of-earth
But you who are listening are the salt of the earth
(Tatian, KFAW 4, 1252)
(6) in herzen betot harto kurzero worto joh lutoro thare,
in hearts pray with-very short words and honourable there
thaz iz got gihore
that it Godmay-hear
Pray in your hearts with short andhonourable words so that Godmay
hear you (Otfrid, KFAW 4, 1254)
(7) stant uf, quad er, gihore mir, joh nim thin betti mit thir
standup said he follow me andtake your bed with you
Standup, he said, follow me, andtake your bedwith you (Otfrid,
KFAW 4, 1255)
As in the case of hren, the meanings to hear/to listen to are the most
frequent; hence, they may be regarded as the prototypical centre of the lexical
category. In Middle High German, to belong to/to be part of arises as a new
meaning of gehoeren:
(8) die gehrent unde treent ze einem hwiske
they belong and t to one family
They are part of a family (Windberger Psalter,
DWB 4, 2, 2509)
To stop is attestedas well, but only in a few citations. In the course of the New
High German period, the meanings to hear/to listen to, which occupied a central
position even in Middle High German, gradually disappear, and have not been
documented since the eighteenth century. Only the cluster to belong to/to be
part of/to t has been handed down to contemporary German (cf. Table 2).
The semantic equivalence between base and prexed verb may result from
the fact that the prex gi- originally hada purely grammatical function. It was
mainly usedas an indicator of a perfective or resultative aspect (cf. Eroms 1993).
Since hren hear, like its New High German counterpart hren, denotes the
succesful accomplishment of an auditory perception, it may be regarded as a resul-
tative or perfective verb itself (for a discussion of the semantics of perception verbs
like hren cf. Harm 2000: 93). Hence, the prex gi- intensies the central
meaning of hren.
Erhren emerged in the eleventh century andhas survivedup to the present day.
language period
8 9 10 11 12 13 14 15 16
(a) to hear
(b) to listen
(c) to comply
with a request,
hear a prayer
(d) to obey,
to follow
(e) to belong to
(f) to stop
17 18 19 20
Table 2. The history of gehren and its principal meanings
As in the case of the prex ge-, the main function of the prex er- is to mark a
perfective aspect. From its rst citation up till now, the central meaning of
erhren has apparently always been to hear a prayer, to comply with a request,
cf. (9). Another usage of erhren already recorded in the eleventh century is to
hear, cf. (10).
(9) ih rvfte ze dir herre vnde dv erhortost mih do ih in arbeiten
I calledto you Lord and you heard me when I in tribulations
was (ad te cum tribularer clamavi et exaudisti me)
I calledto you, my Lord, andyou listenedto me/you heardmy
prayer in my tribulations
(Notker, KFAW 4, 1255)
(10) der vil unervorhte degen / tet vor im des cruces segen, /
the very fearless hero did before him of-the cross blessing
da er die stimme erhorte
when he the voice heard
the fearless hero crossedhimself when he heardthe voice (Renne-
DWB 8, 1836)
But since the seventeenth century, this meaning has occurredonly sporadically
with the verb. In present-day German, erhren to hear has only survivedin the
adjective unerhrt incredible, fantastic. Another marginal reading of erhren
that disappeared after the seventeenth century is to obey, to follow.
Verhren rst occurred in Early Middle High German, cf. (11).
Here, as in the
language period
8 9 10 11 12 13 14 15 16
(a) to hear
(b) to comply
with a request,
hear a prayer
(c) to obey,
to follow
17 18 19 20
Table 3. The history of erhren and its principal meanings
majority of the Middle High German citations, it is largely synonymous with the
base hoeren. The meaning not to hear, to mishear, which shows the negative
function of the prex ver-, arose in Early New High German texts of the
fteenth andsixteenth century, cf. (12).
(11) Theophile, herre gt, / sprach er, verhre mnen mt, / waz ich
Theophilus master dear saidhe hear my thoughts what I
mit dir reden sal
with you speak will
Theophilus, my dear master, he said, listen to my thoughts about
which I want to talk to you (Passional,
DWB 12, 1, 581)
(12) er was gar mde und verhrde als das, das im Herna sagte
he was very tired andnot-heardall that which him Herna told
He was very tiredandd idnot hear, what Herna toldhim (Elisabeth
v. Nassau, Lexer 3, 132)
After the sixteenth-seventeenth century, the situation changes dramatically.
Verhren in the meaning of to hear, to listen falls into disuse. It only survives
in the specic context of legal discourse, where it is used in the sense of
interrogate, a reading, incidentally, which clearly illustrates the intensifying
function of ver-, cf. (13).
(13) der ocial ist kumen her / und wil verhrn man und frauen, /
the ocer is come here andwill interrogate men andwomen
ob iemant ber di schnur het gehauen
if someone over the trace has hit (proverb)
The ocer has come andwill interrogate men andwomen, to nd
out if someone has kickedover the traces (Fastnachtsspiel,
DWB 4,
1, 582)
The meaning not to hear, as illustratedby (12), continued from Early New
High German into the nineteenth century, but then disappeared. Since the nineteenth
century, verhren (apart from its legal usage) has been constructedwith the reexive
pronoun sich andhas as its only meaning to mishear, to hear wrongly, cf. (14).
(14) sie haben sich nicht verhrt, herr baron: Zempin und ich
you have rri not misheardmister baron Zempin andme
You did not mishear, my lord: Zempin and me (Spielhagen,
10, 1, 584)
4. Conclusion
If we try to sketch a general picture of the development of hren andits
derivations, there is one feature that stands out: In Old and Middle High German
there were signicant semantic overlappings between hren andits derivations:
not only couldthe prexed forms express the meaning of the base, hear, listen,
language period
8 9 10 11 12 13 14 15 16
(a) to hear
(b) to listen
(c) to examine,
to interrogate
(h) not to hear
(i) mishear
17 18 19 20
Table 4. The history of verhren and its principal meanings
but the base couldalso express the specializedmeanings of the prexed forms.
In New High German (that is, to be precise, after the seventeenth century) the
relationship between hren andthe prexed verbs is a fundamentally dierent
one: hren and erhren, gehren, verhren show clear-cut semantic distinctions.
At the same time, their semasiological range is signicantly reduced as opposed
to their more polysemous Old and Middle High German counterparts. In New
High German, the meaning to hear/to listen can only be expressedby the
base, whereas to belong to/to be part of is restrictedto gehren, andto hear
a prayer, to comply with a request to erhren. If we put the Old, Middle and
New High German paradigms of hren to the test of the isomorphic principle,
what emerges is that in the older language periods more meanings correspond to
one form, whereas New High German shows a higher degree of isomorphism
between form andcontent.
On the basis of this evidence, the history of hren
andits prexed verbs can indeedbe describedin terms of the isomorphic
principle as proposedby Solms.
In addition, I would like to stress that the word history I have outlined here
can also be regarded as expressing a tendency towards Haimans notion of
motivational iconicity. The point of departure for this hypothesis is the question
why the process described here took place just the way it did. What is striking
about this development is that all the meanings that have been transferred from
hren to its derivational forms since Middle High German, have to be considered
as peripheral category members from the point of view of prototype theory.
Hence, as I have outlinedabove, they may be considered as semantically marked
as opposedto the central member of the lexical category. When we turn our
attention from content to form, hren may be seen as morphologically unmarked
as opposedto its derivational forms. The latter are all markedfor the simple
reason that they all have a formal feature which hren does not have. It thus
becomes clear that form-meaning-correspondences are not established randomly
or accidentally in the course of the history of hren. What emerges is that the
older, non-prototypical meanings of hren (e.g. belong to) tendto be attached
to the derivational forms in New High German (e.g. gehren), while the
prototypical meaning, hear, becomes more andmore restrictedto the simple
verb. In other words, semantically marked meanings of hren tendto be attached
to a markedform, whereas the unmarkedmeaning is preferably renderedby the
unmarkedform. This can indeedbe interpretedas a tendency towards motiva-
tional iconicity.
It is still an open question whether the development sketched here really is
an adequate account of a general tendency in the history of the German lan-
guage. In order to provide an answer I have collected a number of cases that are
similar to the development of hren. I began by checking dictionaries of
contemporary German for deverbal er-, ge-, and ver-forms. As a secondstep, I
tracedthese verbs to earlier periods of language history using current historical
dictionaries like Grimms Deutsches Wrterbuch andPaul s Deutsches Wrter-
buch. As far as I was able to determine on the basis of these dictionaries, the
cases listed under items (15)(17) seem to be comparable to the development of
hren: for all the prexes mentioned the dictionaries provide strong evidence that
there were large semantic overlappings between simple verb and prexed verb
in historical periods, whereas in contemporary German more or less signicant
semantic dierences have developed.
(15) er-verbs
achten: erachten; bauen: erbauen; fassen: erfassen; nden: ernden;
fordern: erfordern; (sich) freuen: (sich) erfreuen; greifen: ergreifen;
holen: erholen; klren: erklren; langen: erlangen; nen: ernen;
scheinen: erscheinen; staunen: erstaunen; sterben: ersterben; streken:
erstrecken; suchen: ersuchen; tten: ertten; tragen: ertragen;
wecken: erwecken; wehren: erwehren; zhlen: erzhlen; zeugen:
erzeugen; ziehen: erziehen
(16) ge-verbs
bieten: gebieten; brauchen: gebrauchen; denken: gedenken; dulden:
(sich) gedulden; leiten: geleiten; loben: geloben; mahnen: gemahnen;
raten: geraten; reichen: gereichen
(17) ver-verbs
bieten: verbieten; bitten: verbitten; blenden: verblenden; ben: ver-
ben; enden: verenden; erben: vererben; fassen: verfassen; fhren:
verfhren; harren: verharren; hindern: verhindern; mahnen: ver-
mahnen; meiden: vermeiden; merken: vermerken; missen: vermissen;
mgen: vermgen; spren: verspren; warnen: verwarnen; weilen:
Although the development of hren andits derivations is certainly not an isolated
case, as the lists under items (15)(17) show, it is clear that, at the present state
of research, we are far from determining the exact extent of this tendency
towards more isomorphism or iconicity. The list of possible parallels has to be
checked in detail not only on the basis of dictionaries but also of text corpora. In
addition, other prexes, especially the frequent be-prex, have to be examined
as well. Only then will we be able to advance reliable hypotheses about a
possible general preference for isomorphic/iconic coding in the history of verbal
word-formation. Nevertheless, as I hope to have demonstrated in my case-study,
the notion of isomorphism andof diagrammatic iconicity is not only a fruitful
approach for the study of grammar and grammatical change but also provides us
with new insights into central aspects of wordhistory , which otherwise would
have remained unidentied.
I wouldlike to thank Sabine Buchholz, Olga Fischer, ManfredJahn andthe audience of the Jena-
conference for their interest in andtheir comments on a previous version of this paper. I am also
particularly grateful to Alexander J. Wood for checking my English.
1. Up to now, only smaller text collections have been compiled; there is still no equivalent to the
excellent Helsinki-corpus, which covers the older periods of English.
2. If there are any dierences between bleiben and verbleiben, they are of a pragmatic rather than
semantic nature.
3. Lexicalization here means that the meaning of the prexed verb as a whole cannot be analysed
as a mere addition of the meaning of the stem and the meaning of the prex; rather, in relation
to its constituents the prexed verb gains a semantic surplus which has to be stored in the
mental lexicon.
4. For the sake of simplicity, I will focus on genuinely prexed verbs andd isregard particle verbs
like aufhren, abhren, zuhren. As opposedto prexed verbs, particle verbs show initial
stress, are separable, correspondto prepositions, andcarry more semantic weight.
5. This is the only citation for OHG hren to stop; in Middle and Early New High German,
hren/hoeren to stop is well documented, cf. Lexer 1, 1340. Since Early New High German
times, the meaning to stop is expressedby the particle verb aufhren, cf. FNW 2, 493495.
6. For a detailed discussion of this semantic change see EWDS 37 and Harm (2000: 171173).
7. An analysis of selected Middle High German texts clearly shows that hear/listen is by far the
most frequently attestedmeaning of hren: in the Nibelungenlied 125 out of 126 citations of
hren have the meaning hear/listen, in Walther von der Vogelweide, the ratio is 21 to 24,
in Neidhart 25 to 27, in Gottfried von Strassburgs Tristan 58 to 62, in Wolfram von Eschen-
bachs Parzival 90 to 93.
8. The German prex ver- can occur with a large number of semantic values. This is due to the
fact that ver- goes back to three etymologically distinct Proto-Germanic prepositions which
mergedas early as pre-OldHigh German times (see EWDS 854). In the case of MHG
verhoeren, only the pleonastic function of the prex is relevant.
9. Solms (1999, 2423) argues that this tendency towards isomorphic coding in language history is
due to a transition from the holistic thinking of the Middle Ages to the more analytical reasoning
of the modern era. This view, however, is rather speculative. Specically, what it cannot explain
is why a similar development seems to have occurred already in the history of classical Latin
(for further details see Harm 2001). Rather, the reason for this tendency towards isomorphic
coding seems to be the constant anduniversal needof the speaker to be as explicit as possible.
For a similar preference of explicitness in grammatical change see Rohdenburg (1996).
10. In a quantitative approach to language change as advocated by Manczak (2000), it could be
arguedthat this is the result of frequency correlations rather than iconicity. In this case, the
most frequent meaning hear/listen has been attachedto the most frequent form, i.e., the
simple verb, whereas the meanings which have always been coded less frequently by the simple
verb have been attachedto the relatively rare derivational forms. But such a quantitative
approach cannot explain why an item like gehren has become restrictedto belong to/be part
of: in Middle High German, the meaning belong to/be part of is quite rare with the ge-form,
while hear/listen is predominant here as well (in the Nibelungenlied the ratio is 1 to 19, in
Tristan 3 to 24, in Parzival 4 to 14, in Walthers works 0 to 7, in Neidhart 1 to 4). Thus, in the
case of gehren the quantitative approach makes an incorrect prediction. Following Mayerthaler
(1981: 137140), token frequency has no explanatory power itself; both low andhigh frequen-
cies have to be seen as epiphenomena of markedness and unmarkedness, respectively.
DWB = Deutsches Wrterbuch von Jacob Grimm undW ilhelm Grimm. Bd. 116.
Leipzig: Hirzel 18541960. Quellenverzeichnis Leipzig: Hirzel 1971.
DWB = Deutsches Wrterbuch von Jacob Grimm undW ilhelm Grimm. Neubearbeitung.
Hg. von der Berlin-Brandenburgischen Akademie der Wissenschaften (vormals
Akademie der Wissenschaften der DDR) und der Akademie der Wissenschaften zu
Gttingen. Bd. 1. Stuttgart, Leipzig: Hirzel 1983.
EWDS = Friedrich Kluge. Etymologisches Wrterbuch der deutschen Sprache. 23. erw.
Au. bearb. von Elmar Seebold. Berlin, New York: de Gruyter 1995.
FNW = Robert Anderson/Ulrich Goebel/Oskar Reichmann. Frhneuhochdeutsches
Wrterbuch. Bd. 1. Berlin, New York: de Gruyter 1986.
KFAW = Elisabeth Karg-Gasterstdt/Theodor Frings. Althochdeutsches Wrterbuch. Auf
Grund der von Elias von Steinmeyer hinterlassenen Sammlungen im Auftrag der
Schsischen Akademie der Wissenschaften zu Leipzig. Bd. 1. Berlin: Akademie-
Verlag 1952.
Lexer = Matthias Lexer. Mittelhochdeutsches Handwrterbuch. Bd. 13. Leipzig: Hirzel
Paul = Hermann Paul. Deutsches Wrterbuch. 9., vollst. neu bearb. Au. von Helmut
Henne undGeor g Objartel unter Mitarbeit von Heidrun Kmper-Jensen. Tbingen:
Niemeyer 1992.
Secondary Literature
Bolinger, D. 1977. Meaning and Form. London: Longman.
Eroms, H.-W. 1993. Das Verbalprx ge- im Sptmittelhochdeutschen. Poetica 38:
Geeraerts, D. 1997. Diachronic Prototype Semantics. A Contribution to Historical Lexico-
logy. Oxford: Clarendon.
Geeraerts, D., S. Grondelaers and P. Bakema. 1994. The Structure of Lexical Variation:
Meaning, Naming, and Context. Berlin: Mouton de Gruyter.
Givn, T. 1995. Isomorphism in the grammatical code: cognitive and biological
considerations. In Iconicity in Language, R. Simone (ed.), 4776. Amsterdam,
Philadelphia: Benjamins.
Haiman, J. 1980. The iconicity of grammar: Isomorphism andmotivation. Language 56:
Haiman, J. 1985. Natural Syntax. Cambridge: Cambridge University Press.
Haiman, J. 1998. Talk Is Cheap. Sarcasm, Alienation, and the Evolution of Language.
Oxford: Oxford University Press.
Harm, V. 2000. Regularitten des semantischen Wandels bei Wahrnehmungsverben des
Deutschen. Stuttgart: Steiner.
Harm, V. 2001. Gibt es eine Monosemierungstendenz in der Wortgeschichte des
Neuhochdeutschen? berlegungen zur sprachhistorischen Interpretation lexikograph-
ischer Befunde. Zeitschrift fr germanistische Linguistik 29: 364380.
Janda, L. 1996. Unpacking markedness. In Cognitive Linguistics in the Redwoods. The
Expansion of a New Paradigm in Linguistics, E. Casad(ed .), 207233. Berlin, New
York: Mouton de Gruyter.
Khnhold, I. 1973. Prxverben. In Deutsche Wortbildung. Typen und Tendenzen in der
Gegenwartssprache, Vol. I, I. Khnholdand H. Wellmann (eds), 141362. Dssel-
dorf: Schwann.
Manczak, W. 2000. Criticism of naturalness: Naturalness or frequency of occurence?.
Folia Linguistica Historica 21: 149154.
Mayerthaler, W. 1981. Morphologische Natrlichkeit. Wiesbaden: Athenaion.
Mayerthaler, W. 1987. System-independent morphological naturalness. In Leitmotifs in
Natural Morphology, W. Dressler (et al.) (eds), 322. Amsterdam, Philadelphia:
Rohdenburg, G. 1996. Cognitive complexity and increased grammatical explicitness in
English. Cognitive Linguistics 7: 149182.
Solms, H.-J. 1989. Frhneuhochdeutsche prxale Wortbildung und die Umstruktur-
ierung des Lexikons. In Zur Wortbildung des Frhneuhochdeutschen. Ein Werkstatt-
bericht, H. Moser andN. R. Wolf (eds), 2131. Innsbruck: Institut fr Sprachwissen-
Solms, H.-J. 1999. Der Gebrauch uneigentlicher Substantivkomposita im Mittel- und
Frhneuhochdeutschen als Indikator kultureller Vernderung. In Sprachgeschichte
als Kulturgeschichte, A. Gardt (et al.) (eds), 225246. Berlin, New York: de
Ungerer, F. 1999. Diagrammatic iconicity in word-formation. In FormMiming Meaning.
Iconicity in Language and Literature, M. Nnny andO. Fischer (eds), 307324.
Amsterdam, Philadelphia: Benjamins.
Van Langendonck, W. 1999. Markedness and prototypical speaker attributes. In Issues
in Cognitive Linguistics. 1993 Proceedings of the International Cognitive Linguistics
Conference, L. de Stalder and C. Eyrich (eds), 567576. Berlin, New York: Mouton
de Gruyter.
</TARGET "har">
AUTHOR "Beate Hampe andDoris Schnefeld"
TITLE "Creative syntax"
WIDTH "150"
Creative syntax
Iconic principles within the symbolic
Beate Hampe andDoris Schnefeld
University of Jena / University of Bochum
1. Iconicity and derivative syntax
Previous work on iconicity has emphasizedthat it is mainly iconic principles that
enable language users to cross the borders of the syntactically admissible in
creative ways in order to satisfy their communicative needs. In the introduction
to Form Miming Meaning, Fischer andNnny (1999: xx) state that when lan-
guage users unleash their creative energies (as in situations as diverse as the
creation of literary texts, childrens acquisition of language, or the creolization of
pidgins), iconicity is strongly involved. So we nd that, whenever the need
arises, language users, in order to express themselves, try to be inventive and
tendto stretch their languages elements and rules to new usages. To promote the
successful communication of the speakers message, the creatively manipulated
language material must, however, oer clues as to how the novel conceptualiza-
tion with which it is associated is to be understood. The main goal of this paper
is to show that such clues often are in central respects iconic of what happens at
the conceptual level.
For a rst illustration of the kindof linguistic creativity we are talking
about, take the by now widely used, but once creative word formation worka-
holic as an example. Here the new, complex meaning/conceptualization of one
who is addicted to work is formally expressed by a fusion (traditionally called
blend) of the (partially truncated) forms work and alcoholic. Such forma-
tions are very economic ande ective because the short, fusedor blendedword
form calls up two concepts which also have to be fusedor blended in a special
way to nally refer to a unied, more complex novel concept. Ungerer
(1999: 314) observes that the fusion of the wordforms can be interpretedas
an iconic reection of the process of conceptual fusion.
In the following, we are going to show that iconicity is also a major factor
in the creative assembly of larger constructs belonging to the eld of syntax.
More specically, we claim that we nd in language use syntactic units,
particularly verb complementations, known by the terms of analogical or
derivative syntax (cf. Lamb 1991, Fillmore and Atkins 1992), which exhibit a
similar phenomenon. Examples (1) to (5) shall serve to illustrate this claim. The
full collection of authentic examples is provided in the Appendix, which the
reader is referred to by the numbering in square brackets.
(1) she wrestled a screaming Dudley into his high chair. [A1]
(2) the boys andgirls divided towards their separate staircases [B17]
(3) the lake froze solid [C29]
(4) She bored them all stupid [C32]
(5) An OxfordUniversity student was missing, feared drowned yesterday
What unites all of our examples is that the verb phrases present obviously
exceptional combinations andrefer to events not normally denoted by the
respective verbs; in each case, a highly unusual verbal construction prompts the
listener to look for or create a matching unusual conceptualization. The question
is how exactly this is achieved.
We have already suggested one major aspect of this process, namely that
such constructions contain iconic hints at the way they are to be understood. The
use of every verb is markedby the presence of an argument structure which is
at rst sight inconsistent with it andquasi-inherited(or perhaps borrowed)
from another verb. This represents a playful transgression of regular linguistic
rules, andtherefore belongs to what Fnagy (1995: 28586) called the Distorter or
Modier a secondary code that re-introduces motivation into symbolic
structure. Though these sketchy remarks already indicate what the crucial keys are
that enable listeners to create appropriate interpretations, the description of the
kindof iconicity involvedhere can still be made more precise in two respects.
As suggestedabove, the borrowedsyntactic structure of the verb phrase
provides an iconic clue that guides the listener to retrieving the meaning of a second
verb more typically associatedwith this structure. This type of iconicity is diagram-
which means that each case of derivative verbal syntax as presented in (1)
to (5) is a schematic icon of some other verbs most typical argument structure,
without this other verb appearing in the respective expression. Thus, one
predicate eectively combines two event descriptions, whereby the second one
is only hintedat by the quotation of its syntax. This is diagrammatic iconicity
of the secondd egree, which is not strictly semiotic in nature (Johansen
1996: 49.) because it is intra-linguistic. This means, it is no diagram of any
relations outside language, but reects relations within language itself. It is an
iconic force operating within the symbolic, more specically, within grammar.
Although it has been acknowledged before that even the symbolic is to some
extent iconic (Fischer 1999: 346), we consider derivative verbal syntax to be a
remarkably vital illustration thereof.
Having saidabove that one predicate thus eectively combines two event
descriptions, we shouldat once correct ourselves andsay , more precisely, that
it integrates two event descriptions. This reformulation opens up a better view on
the secondway in which iconicity is involved, this one exactly paralleling the
process-related type of iconicity discussed by Ungerer (1999: 309 .) concerning
the processual aspects of word-formation in compounding, lexical blends (such as
workaholic), andacronyms. Since we are dealing with cases in which a verb
appears with another verbs syntax, we think it appropriate to call them syntactic
blends to make the analogy more obvious. The very fact that the two event
descriptions calledup by a syntactic blendare integratedinto one form or presented
by one syntactically mixedform (insteadof two separate verb forms), constitutes an
iconic clue in that it mirrors the integration or creative mixture at the conceptual
level itself. It reects in other words a conceptual process of fusion.
To summarize (in reversedord er): the recipient of the message gets iconic
clues to the intended interpretation on two dierent levels: On the one hand, and
in a more general way, the integration at the formal level reects that there is inte-
gration at the conceptual level.
On the other hand, and more specically, the syntac-
tic structure quoted from another verb, serves as a guide to this verbs meaning and
is therefore the prerequisite for the creation of the intended complex meaning.
Otherwise, how wouldwe know what concepts to integrate at all?
2. Fusion: A construction-based explanation
This survey of the two ways in which iconicity is involvedin the creation and
understanding of derivative verbal expressions has not so far mentioned that
there may be an alternative to the iconic evocation of another verbal meaning,
namely the direct semantic contribution of the syntactic construction itself, as
suggested by Goldbergs research on constructional semantics (cf. Goldberg
1997). Of course, one needs to nd a construction-based approach to grammar
viable enough to consider this an alternative at all (cf., e.g., Goldberg 1995). The
issue is of some importance because if meaningful grammatical constructions are
consideredto play a role in the process, then they could just by providing the
creatively usedverb with the respective structure contribute their meaning to
the syntactic formation directly, andthere wouldbe no needto call up another
verbal meaning via an iconic relationship. In such a case, we wouldbe left with
only one level of iconicity, namely the process-relatediconic clue as to the
conceptualization of the integratedmeaning of the respective phrase.
Figure 1 presents a brief survey of the kinds of syntactic (i.e. lexically
unlled) constructions posited by Goldberg (cf. 1995/1997):
To rephrase the problem in a more concrete format, let us briey turn to
Goldbergs syntactic constructions
syntactic patterns: constructional meanings:
caused-motion construction
NP +V + NP + PP/adverb
subj obj adverbial: goal
intransitive-motion construction
NP +V + PP/adverb
subj adverbial: goal
resultative construction
NP +V + NP + Adj
subj L:dyn obj compl
di-transitive construction
NP +V + NP + NP
subj ind obj dir obj
transitive construction
NP +V + NP
subj obj
conative construction
NP +V + PP
subj at
Figure 1. Syntactic constructions posited by Goldberg (cf. 1995/1997).
examples (1) and(4). Do the forms wrestle somebody into a high chair and
bore somebody stupid iconically relate to particular verbs more typically
realizing these structures, such as force andmake, as suggested above? Or do
they just activate the constructional meanings of the caused-motion construction
andthe resultative construction respectively?
In the construction-basedapproach, the unusual syntactic structure prompts
the language user to align the verbal meaning with the respective constructional
meaning itself (rather than with the more potentially specic meaning of another
verb more typically realizing this syntactic construction). This is understood as
the semantic fusion of the constructions argument structure with the semantic
roles of the verb occurring in it. In this process, the construction, which species
in a highly abstract way the basic frame i.e. a scenario andits participants
allows the verb actually being used to acquire additional participant roles or to
change their syntactic realization (cf. Goldberg 1995: 51. andGold berg 1997).
For our two examples, the subsequent diagrams schematically depict this fusion
of verb andconstruction. In (1), the caused-motion construction licenses the
appearance of directional adverbials (here: into the high chair) when a verb like
wrestle is made to refer to a caused-motion scenario, although it does not
normally denote directed motion (and therefore does not have this syntactic slot).
Similarly, the resultative construction licenses the slot for the object-related
complement in (2), when a prototypically transitive verb, such as bore, is used
to denote the causation of change in its object.
Legendfor Figures 2/3:
Construction (Sem) CAUSE-MOVE < agt pat loc >
R:CAUSE/MEANS wrestle < agt secondary agt >
Syn V subj obj loc
Figure 2. Role fusion for example (1) wrestle somebody into a high chair(after Goldberg
1995: 51).
Construction (Sem) CAUSE BECOME - < agt rec pat >
R:CAUSE/MEANS bore < agt pat >
Syn V subj obj compl
Figure 3. Role fusion for example (4) bore somebody stupid (after Goldberg 1995: 54).
Sem = semantics; agt = agent; pat = patient; loc = location adverbial; R = relation/way in which the
verb is integratedinto the construction; Syn = syntax; subj = subject; obj = object; compl
; = object-
The resulting fusedinterpretations of both examples can be paraphrasedas move
somebody into the high chair by (means of) wrestling with him, and make
somebody stupid by (means of) boring them respectively.
This process of semantic fusion between a construction anda verb not
normally occurring in it is to be distinguished from such realizations of construc-
tions where the verbs simply elaborate the constructional meaning or, in
Langackers terms, instantiate them (as is the case with put, force, push,
etc. for the caused-motion construction). However, the distinction between
fusion andelaboration is a cline rather than a binary opposition, with verbs
(proto-)typically realizing a given construction being positionedat the elabora-
tion-endof the scale. Compare the following three examples:
(6) give somebody a tax disc.
(7) refuse somebody a tax disc.
(8) slip somebody a tax disc.
Example (7) is a less typical elaboration of the ditransitive construction than
example (6) since it negates its basic meaning, namely the transfer of the entity
denoted by the direct object. Yet it is not as atypical as example (8), which
involves the fusion of the ditransitive construction with the argument structure
andmeaning of a normally intransitive verb yielding the interpretation give
somebody a tax disc by (means of) slipping it (e.g. across the table).
We are now going to show that Goldbergs approach, promising as it seems
to be, does not suce to deal even with all of our examples, and that the verb-based
approach on iconic principles provides a better account of the problem at issue.
3. Construction-based vs. verb-based explanation
We detectedone major shortcoming of the construction-basedexplanation. Apart
from that, there are other issues, which can, however, be remedied in principle
within this theoretical framework. We will start by discussing the latter aspects.
Firstly, a construction-basedanalysis of our examples would presuppose a
suciently comprehensive inventory of constructional schemas. In order to
account for examples like (3) freeze solid and (5) fear someone drowned, we
would(hypothetically) have to introduce a subject-relatedresultative construc-
tion as well as an attributive construction.
The former simply completes the
range of possible syntactic realizations of constructions with resultative meanings
(i.e. includes subject-related complements in addition to Goldbergs complex-
transitive pattern with object-relatedcomplements); it would prototypically be
realizedby such verbs as become, or get. The latter construction exhibits an
Figure 4. Additional constructions needed to account for all of our derivative examples.
argument pattern similar to Goldbergs (object-related) resultative construction
(i.e. realizes a complex transitive pattern consisting of subject, verb, object and
object-relatedcomplement), but it involves a dierent set of copular verbs
namely stative (as opposedto dynamic) link verbs andwould usually be
elaborated, for example, by regard and consider.
Secondly, there are many examples the syntactic structures of which t the
constructions posited, but whose semantic specications do not. More specical-
ly, the target interpretation of many examples that t the syntactic patterns of the
intransitive- or caused-motion constructions respectively (cf. examples 9 and
10) do not necessarily involve literal movement. In these examples, the direc-
tional adverbials often realized by prepositional phrases with out of, into
andto can only be understoodas specifying a location, i.e. a source or
goal of a path, if they are construedmetaphorically . In these metaphorical
mappings, the source or goal of the movement corresponds to the nal or initial
state of a change. Thus, what such uses of the intransitive-motion construction
refer to is changes of state (cf. 9a,b), or time (cf. 9c) not movement. Analo-
gously, the caused-motion examples refer to caused change of state (cf. 10a) or
when movement is mappedonto action causedaction (cf. 10b).
(9) a. The aircraft began to roll out of control [B27]
b. The boiler had shuddered to a halt [B28]
c. January faded into February [B25] (see also [B24])
(10) a. the few owls hadto be nursed back to health [A15] (see also
b. frighten the committee into [making a decision] [A14]
The theory of conceptual metaphor (Lako 1993 and1999) provides the relevant
mappings for the understanding of action, reasoning, time and change in terms
of motion. Given that conceptual metaphors are xed correspondences between
source andtar get domains, constructional meanings should be accessible to the
mapping just as any other movement concept. Though this is not denied by
Goldberg (1995, 8189), who discusses both polysemy at the level of construc-
tional meanings andmetaphorical extensions, it is still an issue whether the
licence to use, for example, the intransitive motion construction [X MOVES TO Y]
in contexts not involving literal movement is actually provided by the
constructional meaning itself, or by the particular verb occurring. In the iconic-
ity-based/verb-centred approach, on the other hand, the problem does not arise
since the metaphorical use of movement concepts is already lexicalized in the
iconically triggeredverbs themselves andtheir collocation with spatial adverbs/
prepositions. In other words, what is actually called up in each case is not just
the general movement concept provided by the construction, but more specical-
ly such verbal concepts that normally andmost typically appear in the syntactic
frames borrowedby the expressions at hand. For the intransitive-motion verbs in
(9ac), these are get out of control, come to a halt, andturn/change into.
The caused-motion verbs triggered in (10a,b) are the most typical prepositional
verbs tting the contexts bring back to, andtalk into.
Finally, the major point of criticism concerns the semantic predictions made
by the construction-basedapproach. There are cases in which, although fusion
between the verb usedand the constructional meaning is feasible, the resulting
fusedmeanings are not nearly rich enough to account for the intended target
interpretations. Thus, examples (11) and(12) t in with the syntactic pattern of
the caused-motion construction. Possible paraphrases of the fused interpretations
wouldbe move in(to some place) by (means of) paying andmove garments
on(to some place) by (means of) slipping them:
(11) she paid herself in at the hotel [A9]
(12) She retrieveda pair of panties anda matching bra and slipped
them on. [A10]
These paraphrases are insucient, however, because what is referredto are not
the unspecic actions of paying before entering (any place) or moving
garments somewhere by slipping, but very specic scenarios typically associated
with entering places like hotels andairports, or putting on clothes. In other
words, (diagrammatic) iconic reference is made to the only typical verb-particle
constructions tting these contexts: these have to be check in andput on
respectively, which carry with them their rich frames containing knowledge about
the kinds of possible scenarios. The same argument holds for (9a), roll out of
control, and (9b), shudder to a halt, which exhibit the syntax of Goldbergs
intransitive-motion construction. They are not meant to be interpretedas moving
somewhere in a rolling or shuddering manner, not even if this movement is
conceivedof metaphorically as change of state, but as suggestedabove
are meant to be modelled closely (and iconically) after the specic, verbal idioms
get out of control andcome to a halt.
To put things in perspective, similar to the metaphor problem previously
discussed, the problem of the underspecication of the integrated meaning also
occurs only in the construction-basedapproach. This is to say, the iconicity-
based, verb-centered approach suggested here avoids the problems of the
construction-basedone andyield s simple and satisfactory solutions as to what the
plausible candidates for formal and conceptual integration are.
4. Creating integrated conceptualizations by blending
In order to demonstrate how, in the verb-based approach, the two verbal concepts
are nally integrated, let us turn next to the model of conceptual blending
(Turner andFauconnier 1995, Fauconnier andT urner 1996/1998), which we take
to be the most plausible model available for the creation of integrated conceptu-
alizations from two (or more) dierent sources. In the model, commonalities
between the two inputs representing dierent mental spaces (the so-called
input-spaces) are collectedin a thirdspace, the generic space, and, on the
basis of these, the respective counterparts from the (two) input spaces are
mappedonto one another in a fourth space, the blendedspace. Thereby, an
integrated conceptualization is produced, which can additionally contain material
not to be foundin either or any of the inputs.
In our example (11), pay oneself in (at the hotel), input space 1 contains
an unintegratedsequence of two events, an agent a performing some action
(pay (sb.) (sth.)) on the one hand, and an object o that moves (in this case
identical with a) on the other, together with the direction and goal of this motion
(dm, gm). In this particular sequence, the action of paying is the enabling
condition for the second action, the agents entering of the place in question (in
this case, a hotel). The iconically activatedverb check (sb.) in (at some specic
place, such as a hotel or an airport) provides as conceptual material for input
space 2 the frame which structures a corresponding, fully integrated event
sequence as well as its formal, i.e. syntactic, expression. To blendthese two
spaces, the counterparts (collectedin the generic space) needto be mappedonto
one another: a and a, o and o etc., as shown in gure 5 below. The role e
(conditionedmovement), which has no counterpart in input 1, is mappedto as
action (namely paying). This is why the blendcan inherit both this action and
the verb expressing it. In other words, in the integrated expression, pay acquires
the additional meaning of enabling (namely the second action: entering) by
being mappedto e. This fact is iconically signalledby the appearance of the
particle in within this particular structure. The remaining frame elements are
mappedto and lled with the roles providedby input 1, as you can see in the
diagram: a, o, dm, gm.
As for a blends potential to contain or rather to develop emergent structure
of its own, i.e. structure not present in any of the inputs, the following aspects
can be assumed to play a part in the understanding of the sentence: By fore-
grounding the action of paying, which is usually no or only a non-salient sub-
event of checking in (at a hotel), the blended conceptualization acquires a
connotation of situational urgency, an implication of some kindof danger, such
as an illness or some other handicap, which causes the person in question to have
to check in at all costs. Consequently, in this example, it is neither necessary nor
helpful to evoke the sentence-level construction of caused-motion, independent-
ly of possible lexical items typically instantiating it in this context.
Another strong argument in favour of the activation of a secondverb
(insteadof a constructional meaning) in the blending process, is the fact that
conceptual blending is strongly context-determined, or context-sensitive. It can
thus well provide for the interpretive procedures already mentioned: Cued by
some verbal form, a very specic verb frame containing knowledge about a par-
ticular scenario is invoked, rather than a very general meaning, such as that of
the caused-motion construction (cf. also 10b). Apart from that, it is also the
broader situational context which gives clues as to how the expression is to be
understood. A nice example of both can be found in a list of nonce formations
collectedby Mc Arthur (1989): being castledout modelled after being worn
out/tiredout in the context of having seen too many castles where the model
phrasal verb licenses both the use of castle as a verb andprovid es the integrated
5. One phenomenon, two explanations An either/or decision?
We have shown so far that the verb-basedapproach to derivative syntax, which
focussedon the central role of iconic principles, oers satisfactory andstraight-
forwardexplanations for a broad range of examples. However, it does not
completely exclude a construction-based account. Let us have a brief look at how
it is partially compatible with the latter andcan even be saidto incorporate it.
check (sb.) in
(at some place)
pay and gain access
to some place
NP agent a a agent
V conditioned e
(special conditions ENABLEMENT
- pay money
- show ticket
- show passport)
(NP) object o o object
Adv direction dm dm
location gm
at the hotel
she NP a
paid V e
herself NP o
in Adv dm
at the hotel PP gm emergent structure:
necessity/urgency of
checking in
a = a
o = o
dm = dm
gm = gm
Figure 5. The blending account of conceptual integration for example (11) pay in at the
Firstly, one may argue that, at least for such non-metaphorical cases in
which the verb that is iconically triggeredd oes, in fact, only elaborate the
respective construction (e.g., force andbring elaborate the caused-motion
construction in (1) wrestle into a high chair, or (A4) support out of the
stadium), it does presumably not make any dierence whether the desired
integratedconceptualization andits formal expression are arrived at via the
short-cut provided by the respective constructional meaning or else via conceptual
blending with typical verbs occurring in this construction. In fact, a language user
may have both options. This argument gains some plausibility from the fact that, as
Langacker (1987, 1991) claims, the existence of constructions (i.e. the syntactic
templates andtheir highly general meanings) is due to a process of schematizat-
ion from the frequent similar verbalizations of particular scenarios in the rst place.
Still, our considerations of semantic frames and context-sensitivity also demon-
stratedthat the construction-basedapproach is less viable where the semantic
contribution of a construction alone is insucient, i.e. not specic enough.
This leads directly to a second aspect of major relevance to the compatibili-
ty issue: We demonstrated above that a construction-based explanation does not
do for examples (11), pay in at the hotel, and (12) slip on garments, unless
one assumes that the process is basedon lexically lled constructions, namely
the verb-particle constructions check in andput on. Allowing for lexically
lled constructions, however, does make the construction-basedapproach
identical to the verb-basedone; the two approaches converge, in fact, in that the
diagrammatic iconic clue is needed in both cases. In eect, then, it does not
matter in these instances whether we talk about a specic, lexically lled
construction or a verb typically occurring in the respective syntactic environment.
Since the verb-basedapproach works for both very specic andvery general
cases without any further provisos andcan even incorporate the construction-
basedexplanation, we see a clear advantage in the former.
6. Outlook Questions for further research
6.1 Constraints
We have shown that the type of syntactic creativity at issue is not syntactically
constrainedin any way other than that exertedby the range of syntactic patterns
available in English. Since the creative aspect is to be foundin the use of regular
syntactic structures for unusual lexical entries, the point, especially for the
listener, is to get some assistance in the conceptualization s/he is intended to
form. Andexactly this assistance is providedby the diagrammatic iconic clues
we have isolatedandd iscussedhere.
As for semantic limitations, we can state in a very general way that
the two actions or events expressedin the inputs to all our examples needto
allow for a meaningful integration in order to be blended. What is considered as
meaningful largely depends on the context as well as on the individual person
doing the integration of the verbal meanings. It is often possible, for example, to
construe dierent semantic relations, which might even holdsimultaneously;
some of our examples therefore occur more than once in the list of relations
presentedbelow . However, not just any semantic relation between two events
qualies as meaningful andallows for a successful integration. We want to
suggest that the issue may be more complex than impliedby the list of semantic
criteria to be found in Goldberg (1997). Note that these conditions for meaning-
ful integration do not depend on the syntactic patterns and cross-cut the syntactic
classication; they include the force-dynamic relations: cause, means/
instrument, blockage andresult as well as (less typically) enabling actions/
preconditions, and (even more marginally) merely co-occurring activities. In our
collection, we foundexamples where the derivatively usedverb can be understood
to specify relations not included in Goldbergs list. We refer to the numbers in the
Appendix here and indicate with the help of the paraphrases in brackets how the
semantic relation between the two verbs involvedcouldbe construedin each
(a) the manner or way in which the action is realized, or only a salient aspect
of it, as in
[B18] nod up the path {move up the path so that a nodding can be per-
[B22] whoosh past {go or y past in such a way that a whooshing sound
is made}, see also[A5] and [C31];
(b) a (metaphorical) comparison with this other action, as in
[A7] boost sales into orbit {boost sales beyondall limit, so that it can be
comparedto launching an object into orbit},
[B18] nod up the path {move up the path as if nodding}, see also [B19],
All the remaining examples in the Appendix can be construed according to
Goldbergs semantic relations, the majority of which are cause/means/results
andenablement/instigation. For an illustration, consider the paraphrased
interpretations of the following examples:
[A11] magic me out {get out of the room by (means of) magic}
[B17] divide towards the staircases {move and part as a result of that
[B23] stand back {step back and, as a result, stand there};
[C29] freeze solid {become solid by freezing},
[C32] bore them stupid {make (sb.) stupid by (means of) boring them},
[D35] fear drowned {(almost) consider (sb.) dead, out/because of fear}.
Regarding Goldbergs criterion of co-occurrence/simultaneity, we doubt that
the simultaneous occurrence of two events perceivedor experiencedas unrelated
wouldsu ce to inspire the creation of an integratedrepresentation. Thus, if
simultaneity is the relation, the co-occurring events are actually seen as aspects
of one andthe same activity or process, with the derivatively used verb referring
to a very salient aspect of the situation, as in:
[A2] bow them through the doors {show them through the door while
[B17] divide towards the staircases {moving and simultaneously dividing}
[B27] roll out of control {get out of control while rolling}
6.2 Productivity and entrenchment
As regards the question of whether some verbs lend themselves more easily to
the process of being usedin a borrowed syntactic environment than others
i.e. the problem of the productivity of the structures involved in derivative syntax
our data, though they do not necessarily constitute a representative sample,
suggest a ranking in frequency with the types of caused-motion and intransi-
tive-motion verbs at the top. Concerning the enrichment of the structural
inventory of the language, we observedthe following. No new patterns arise,
which are not yet licensedby the rules of syntax. Yet there is an enrichment of
the verbal means of English in that frames borrowedfrom other verbs become
lexicalized with new verbs, which do then refer to dierent scenarios, provided
these scenarios are of such importance that language users feel the needto retain
the new expression. In other words, their potential entrenchment depends on
frequent usage andoccurrence.
To illustrate this, consider the OEDs report on the use of the verb bow, for
example, which documents the following steps in its historical development: In
OldEnglish, the word bow was a strong verb bugan, standing among other
things for bending ones knee/head, in token of reference, respect or
submission. In 1606 (i.e. in Early Modern English), it was found to mean to
express by bowing, in 1651 it is documented to mean to incline the body or
head( to a person) in salutation, acknowledgement of courtesy, polite assent, etc.;
to make or give a bow. Only from 1819, it has been recorded to mean: to usher
in or out with a bow, or bows; so to bow (any one) up or down (stairs etc.) (cf.
OED [on CD-ROM]). We also came upon this usage in our corpus, as example
(A2) bow them through the doors shows. Many of our examples are not novel,
but have been lexicalized andentrenched . These are asteriskedin the Appendix.
6.3 Discourse functions
More research certainly needs to be done on discourse functions, but we can
indicate some tendencies: As one can see from the respective contexts (for
illustration see A18, A26, andA33), derivatively usedverbs are more expressive
than verbs that only elaborate a syntactic construction. This is not only due to
their being exceptional (i.e. marked in the classical functional sense of the
term), but also to the fact that verbal blends are a very economical way of
expressing a complex conceptualization. Apparently, the bringing together of
dierent (even unexpected) scenarios very often has witty, or funny eects.
Derivatively usedverbs can also serve to indicate a particular perspective, as in
(A18), which suggests the view-point andthoughts of a small child. Without
having carriedout systematic research, we found that derivative structures are
relatively frequent in childrens literature, particularly of the fantasy type, and in
crime ction. High-brow literature seems to use the devise only rarely, but then
very eectively (see A18). Comparatively many examples can be foundin the
magazine parts of newspapers andin reviews, where they seem to be exploited
for the above-mentionede ects, too.
7. In retrospect
We emphasizedat the beginning that iconicity plays a major role in the creative
use of language. The analysis presentedhere has shown that, in cases of
derivative syntax, the reader will be able to produce the respective conceptualiza-
tions eortlessly to the point of not even noticing the problem on the
basis of a number of verbal cues provided by the speaker/writer.
We isolatedseveral cues as being iconic reections of the processes going on
at the level of conceptualization andshowedhow these cues help unpack the
meanings of verbs that occur with verbal complementations usually not their own.
In each case, the deviant form prompts the reader/listener to do the following:
i. retrieve the concept referredto by the verb actually occurring in the
ii. retrieve a secondverbal concept, normally and more typically occurring in
the syntactic frame of the expression at hand,
iii. meaningfully blendthe two concepts to produce a thirdevent conceptualiza-
tion appropriate in the context.
We specied the cues involvedin part (ii) of this instruction as being diagram-
matic icons of the secondd egree, whereas (iii) exposes a process-relatedtype of
In the course of our discussion, we furthermore made it clear that such a
verb-centred, iconicity-basedaccount is to be preferredto its construction-based
1. Haiman denes diagrammatic icons as iconic in that although the component parts may not
resemble what they standfor , the relationship among those components will approximate the
relationship among the ideas they represent (Haiman 1994: 1629).
2. Given that there is an establishedform-meaning isomorphism between the typical syntactic verb
frame/structure andthe scenario this frame designates, the occurrence of the unusual syntactic
frame/structure represents an irritation. This disturbance is balanced out in the creation of a new
isomorphism relating the new syntactic form to a novel conceptualization.
3. We do not aim at augmenting Goldbergs model, or generally completing the range of syntactic
constructions in English. We simply apply her approach to the problem at issue in order to
detect its shortcomings.
4. All examples that realize the respective constructional meaning metaphorically are markedout
in the appendix by the bold print of their numbering.
5. In the Appendix, we suggest for each of our examples one verb that may be iconically triggered
by the respective derivative form. These are printed in italics after each example.
Appendix: Collection of authentic examples of derivatively used verbs
The following examples are collectedfrom dierent samples of written English, both newspaper
articles and ction. Here is a list of the sources with the abbreviations used below providedin
The Independent, Dec. 30, 2000 (Ind), The Observer, March 18, 2001 (Obs), J. Fielding: The Deep
End(DE), J. K. Rowling: Harry Potter and the Philosophers Stone (HP I), J. K. Rowling: Harry
Potter andthe Chamber of Secrets (HP II), J. K. Rowling: Harry Potter and the Prisoner of Azkaban
(HP III), D. Lessing: The Summer before the Dark (SBD), K. Manseld: Sun and Moon (SM). The
examples in the appendix are ordered according to the syntactic patterns/constructions they exhibit.
A. Predicates exhibiting the syntax of caused-motion verbs
V + NP(Obj) + PP/Adverb (locational adverbial)
1. she wrestled a screaming Dudley into his high chair. (HP I: 8) force
2. A pair of goblins bowed them through the silver doors (HP I: 57) show
3.* Professor McGonagall pointed them into a classroom (HP I: 112 show
4. Hermione supported Ron out of the stadium. (HP II: 125) bring/help
5.* Joanne found a deck of playing cards and slid them from their package (DE:
6. We are trying to increase numbers of otters by encouraging them into new
areas through the provision of good, undisturbed habitat. (Ind: 10)
7. the chain reaction that can boost sales into orbit (Ind, review: 1) launch
8.* Malfoy trickedyou Filch knew someone was going to be in the trophy
room, Malfoy must have tipped him o. (Harry Potter I: 118)
send o
9. Paying o the taxi man, she paid herself in at the hotel, andwas askedif she
were well enough to sit in the foyer andwait (SBD: 131)
check in
10.* She openedthe top drawer andretrieveda pair of plain white panties anda
matching bra and slipped them on. (DE: 165)
put on
11. [Harry being lockedup in a room says to his friend] But you cant magic me
out either. (HP II: 32)
get out
12.* Hagrid walked them back up to the castle. (HP III, 298) bring
13. a delicious view of how little it takes to provoke modern celebrity-jessies
into having a nervous breakdown (Obs, magazine: 3)
drive into
14.* Malfoys dads have frightened the committee into it [a decision], saidHer -
mione, wiping her eyes. (HPIII:)
talk into
15. the few owls hadto be nursed back to health by Hagrid (HP I: 143) bring back
16. (In the context of burning books) Can I be as insouciant about smoking out
my own words as I have been about other peoples. (Obs: 2)
blot out
B. Predicates exhibiting the syntax of intransitive-motion verbs
V + PP/Adverb (directional adverbial)
17. Through the portrait hole [= an entrance] andacross the common room, the
boys andgirls divided towards their separate staircases. (HP III: 106)
18. When you staredd own from the balcony at the people carrying them the
ower pots lookedlike funny awfully nice hats nodding up the path (SM:
19. Professor Flitwickhadgold en bubbles blossoming out of his wand. (HP I:
20.* he collapsed into a chair. (HP I: 26) fall
21. Before neighbours who have gone away for the summer begin ltering back.
(DE: 362)
22. Show her your acceleration, Harry., Fredyelled , as he whooshed past in
pursuit of a Bludger [a ball in a game called Quidditch] that was aiming for
Alicia. (HP III: 282)
23.* Here we are, Paul said, pushing open the double doors of the cafeteria and
standing back to let Joanne pass through. (DE: 170)
step back
24. He lay andwatchedhis birthday tick nearer (HP I:38) come
25. January faded imperceptibly into February, with no change in the bitterly cold
weather. (HP III: 266)
26. Or they hada foolish goodnature andawful defenceless niceness, like the
weak laugh that sounds as if it is going to ebb into tears. (SBD: 92)
27. The aircraft began to roll out of control when the automatic pilot disengaged
in the struggle. (Ind: 3)
28.* The boiler in my at had shuddered to a halt (Obs, review: 2) come
C. Predicates exhibiting the syntax of resultative verbs (subject- and object-related)
+ Adj(Comp)/V
+ NP(Obj) + Adj(Comp)
29. the lake froze solid (HP I: 143) become
30. Without warning, Joanne felt an invisible presence standing beside her bend-
ing close to whisper in her ear. (DE: 59)
31.* Joanne rose slowly from her eat andunlockedthe door to slide it open. (DE:
32. She bored them all stupid with ying tips (HP I: 108) make
33. I felt like I was just fucking squeezing the toothpaste dry. (Obs Magazine,
rev. 3)
34. tall lean shafts of raggedstone exploding from grass chewed smooth by a
platoon of sheep (Ind. 2)
D. Predicates exhibiting the syntax of the attributive construction (object-related)
+ NP
+ Adj
University student was missing, feared drowned yesterday after high winds
toppledthe nine-man rowing crew into a fast-owing Spanish river (Ind: 2)
Fauconnier, G. andM. Turner 1996. Blending as a central process of grammar. In
Conceptual Structure, Discourse and Language, A. Goldberg (ed.), 113130.
Stanford: CSLI Publ.
Fauconnier, G. andM. Turner 1998. Conceptual Integration Networks. Cognitive
Science 22(2): 133187
Fillmore, C. andB. Atkins 1992. Towards a frame-basedlexicon: The semantics of
RISK andits neighbors. In Frames, Fields and Concepts, A. Lehrer andE. F. Kittay
(eds), 75102. Hillsdale, NJ: Lawrence Erlbaum.
Fillmore, C. andB. Atkins 1994. Starting where the dictionaries stop: The challenge of
corpus lexicography. In Computational Approaches to the Lexicon, B. T. S. Atkins
and A. Zampolli (eds), 349393. Oxford: Oxford University Press.
Fischer, O. 1999. On the role playedby iconicity in grammaticalisation processes. In
M. Nnny andO. Fischer (eds), 345374.
Fischer, O. andM. Nnny 1999. Iconicity as a creative force in language use. In M.
Nnny andO. Fischer (eds), xvxxxiii.
Fnagy, I. 1995. Iconicity of expressive syntactic transformations. In Syntactic Iconicity
and Linguistic Freezes, M. E. Landsberg (ed.), 285304. Berlin: Mouton de Gruyter.
Givn, T. 1995. Isomorphism in the grammatical code. In Iconicity in Language, S.
Raaele (ed.), 4776. Amsterdam: Benjamins.
Goldberg, A. 1995. Constructions. Chicago: University of Chicago Press.
Goldberg, A. 1997. The relationship between verbs and constructions. In Lexical and
Syntactical Constructions and the Construction of Meaning, M. Verspoor, K. D. Lee
and E. Sweetser (eds), 383398. Amsterdam: Benjamins.
Haiman, J. 1994. Iconicity. In The Encyclopedia of Language and Linguistics, R. E.
Asher and J. M. Y. Simpson (eds), 16291637. Oxford: Pergamon Press.
Johansen, J.-D. 1996. Iconicity in literature. Semiotica 110: 3755.
Lamb, S. 1991. Syntax: Reality or illusion?. LACUS-Forum 18: 179185.
Langacker, R. 1987. Foundations of Cognitive Grammar 1. Stanford, Ca: Stanford
University Press.
Langacker, R. 1991. Foundations of Cognitive Grammar 2. Stanford, Ca: Stanford
University Press.
McArthur, T. 1989. The long-neglectedphrasal verb. English Today 18 (April): 3844.
Nnny, M. andO. Fischer (eds). 1999. Form Miming Meaning. Iconicity in Language and
Literature. Amsterdam: Benjamins.
Turner, M. andG. Fauconnier 1995. Conceptual integration and formal expression.
Metaphor and Symbolic Activity 10: 183204.
Ungerer, F. 1999. Diagrammatic iconicity in word-formation. In M. Nnny and O.
Fischer (eds), 307324.
</TARGET "ham">
AUTHOR "Gnter Rohdenburg"
TITLE "Aspects of grammatical iconicity in English"
WIDTH "150"
Aspects of grammatical iconicity in English
Gnter Rohdenburg
University of Paderborn
1. Introduction
This paper proposes to give an overview of iconically motivatedvariation
phenomena in English syntax andmorphology .
We shall be concentrating on
phenomena which have either not been described at all or whose iconic motiva-
tions have been largely ignored. It seems to me that most if not all relevant
grammatical variation phenomena may be characterizedin terms of the three
major iconic principles recognizedby Givn (19841990, 1991a) andmany
others, (1) the principle of linear order, (2) the quantity principle and (3) the
distance principle.
The rst principle says with respect to the serialization of clauses that
their sequential ordering will tendto correspondto the temporal order of the
depicted events (Givn 1990, 1991a). This principle does not play any role in
motivating the grammatical variation phenomena to be presentedhere. Moreover,
it is probably too well known to deserve a more detailed illustration. So I only
mention it here for the sake of completeness. The remainder of this paper will
then be devotedto an exploration of the secondandthirdprinciples.
The quantity principle (which John Haiman [1983] a bit misleadingly refers
to simply as economic motivation) states that A larger chunk of information
will be given a larger chunk of code (Givn 1991a: 87). The principle may be
interpreted to cover so-called markedness correlations as discussed, for instance,
in the typological literature (e.g. Givn 1991b; Wurzel 1998). As is well-known,
there is a universal tendency to code semantically more specic or more highly
marked categories like the plural or the past tense by means of additional
morphological devices. Another and more clearly iconic manifestation of the
quantity principle may be seen in the phenomenon of repetition or reduplication
in morphology. Andcrossing the morphological boundary into phonology, we
nd that the use of geminatedor lengthenedsegments may produce largely
parallel eects. For instance, both repetition/duplication and phonological
lengthening are likely to be usedfor purposes of intensication or even pluraliza-
tion (cf. also Conradie, this volume).
According to the distance principle, the linguistic distance between
expressions corresponds to the conceptual distance between them (Haimann
1983). T. Givn (1990: 970, 1991a: 89), who uses the term proximity principle
instead, gives the following characterization:
Entities that are closer together functionally, conceptually, or cognitively will
be placedtogether at the code level, i.e. temporally or spatially.
The principle involves a large spectrum of syntactic manifestations some of
which will be discussed in later sections. But it can also be seen to cover a
number of universal morphological tendencies which Bybee (1984, 1985)
captures in terms of the so-calledrelevance principle. Its most important insight
is that the relevance of a morphological category to the semantics of a verb is
reected in its closeness to or fusion with the verb stem. In English, for instance,
the high degree of relevance of the category (past) tense is seen in the fact that
the past tense marker may be fusedwith the stem as in sent, lent and bent.
2. Some implications of the quantity principle in English
Informally, the quantity principle may be capturedby means of the slogan More
form more information/informative. Comparing grammatical elements or
function words with semantically richer or more specic lexical elements we nd
again andagain that the relatively unspecic function words tendto be very
much shorter in all languages than the rest of the vocabulary. For instance, this
is true of articles andmost pronouns, which generally constitute the shortest
words in natural languages (cf., in particular, Zipf 1935). Similarly, in the area
of adverbial conjunctions, Kortmann (1996, 1999) has shown that, cross-linguisti-
cally, dierent degrees of cognitive complexity correlate with dierent degrees
of morphological complexity.
In this connection it is intriguing to compare full verbs in English with their
auxiliary counterparts. We start by considering examples (1)(3) representing
Standard English.
(1) a. He had/d done it.
b. He had/*d three books.
(2) a. She has/s done it.
b. She has/*s three books.
(3) a. I came to the meeting but he did not/didnt.
b. I did my work but he did not/*didnt.
In (1a) the semantically much less specic auxiliary form had may be reduced
or contractedto the clitic d. In Standard English, this reduction is impossible in
the case of the corresponding full verb in (1b). The contrast is paralleled in
(2a,b) by the form has. A relatedcontrast is foundin (3a,b). The negatedauxiliary
did not may be contractedto give didnt as is seen in (3a). However, the contraction
of not is not possible with the main or full verb as shown in (3b). This brings us
to a similar contrast which is foundin most non-standardvarieties of English. In
fact, it is foundin all varieties using the form aint as a replacement of has not
or have not. As is seen in (4a,b), this particular kindof grammatical simplica-
tion (or contraction) is also conned to the auxiliary use of (the verb) have:
(4) a. He aint done it. (NStE)
b. *He aint three books. (NStE)
As is well-known, virtually all non-standard varieties of English have two past
tense forms of the verb DO, did and done, with the latter being conned to the
full or lexical verb (cf. e.g. Cheshire, Edwards and Whittle 1993: 7779). In
Rohdenburg (2002) it is argued that the distinction is also motivated by the
quantity principle. In addition, some non-standard varieties in Britain (and
elsewhere) (usedto) provide a related contrast in the present tense which is
clearly iconic (cf. e.g. Cheshire 1978: 5458; Ihalainen 1985: 70; Ihalainen
1987: 79; Cheshire, Edwards and Whittle 1993: 78). Compare, for instance:
(5) I still dos a bit of gardening. (Ihalainen 1985: 70)
(6) He do start at Brue. (ibid.)
While the full verb (usedto be or) may still be inected throughout the present
tense, the auxiliary (usedto be and largely) remains uninected. A close parallel
is foundin the same dialect spectrum with the verb have (Cheshire 1978: 5458;
Ihalainen 1985: 70; Ihalainen 1987: 79; Edwards 1993: 224226).
Unlike Standard English, most dialects possess one or several redundantly
markedcategories such as double comparatives (more stronger), multiple
negation or reinforcedd emonstratives (this here). To the extent that these
contrast at all with the simpler expressions they may at least occasionally be
usedfor purposes of intensication (cf. e.g. Quirk 1970: 304 andFai 1989: 29
on multiple negation or Peitsara 1987 on this/these here).
3. Some implications of the distance principle in English syntax
3.1 The transitivity principle
The distance principle represents a superordinate principle from which various
subordinate or domain-specic principles may be derived. The rst such
principle, whose validity has been demonstrated in a number of recent investiga-
tions, is the so-calledtransitivity principle. As is done in the seminal work of
Hopper andThompson (1980), the concept of transitivity shouldbe interpreted
as a global category relating to the way an action impinges on an object referent.
Transitivity principle
Indirectly coded objects, in particular prepositional objects, are indicative of a
lower degree of transitivity (Hopper and Thompson 1980; Moravczik 1978;
Haiman 1983: 790793; Givn 1984: 96; Dirven 1988; Dixon 1991: 271274,
278284; Thompson andKoid e 1987).
In English, the transitivity principle is manifestedby two large classes of
constructions involving nominal complements. In the rst case, we are dealing
with verbs of transfer in the widest sense, whose two objects may be represented
by two dierent syntactic constellations as shown in (7a,b).
(7) a.
I showed/oeredPeter the book, but he happenedto be asleep.
b. I showed/oeredthe book to Peter, but he happenedto be asleep.
In (7a), the beneciary object follows the verb directly, in (7b) the beneciary
is separatedfrom the verb andexpressed less directly by a prepositional phrase.
As Thompson andKoid e (1987) have shown, the greater or lesser syntactic
distance between verb and beneciary corresponds to a greater or lesser degree
of personal involvement. In (7a) the syntactic distance is small suggesting a high
degree of personal involvement. The overall sentence is odd, however, because
the context makes it clear that the required degree of involvement simply does
not exist. Incidentally, this also explains why announce, demonstrate and donate
do not allow the beneciary to be coded as a non-prepositional postverbal object:
The semantics of these verbs fails to provide for the requisite degree of personal
involvement. This is reected in the fact that the beneciary is generally left out
The second case concerns the literally hundreds of single-object verbs as in
(8)(10) allowing a choice between direct and prepositional objects.
(8) a. They approvedof the bill.
b. They approvedthe bill.
(9) a. They agreedto the pay oer.
b. They agreedthe pay oer.
(10) a. He beggedof Godto help him.
b. He beggedthe man to help him.
Clearly, from a syntactic point of view, prepositional objects represent a less
direct link between verb and object than do direct objects. As is to be expected,
the greater or lesser syntactic directness corresponds to a greater or lesser degree
of semantic transitivity. Unlike (8a) and(9a), the b-sentences imply that
owing to certain prior decisions the legal status of the bill and the pay oer
has changedd ramatically.
Another observation involves the choice of object NPs. To the extent that
prepositional objects are capable of indicating a less direct manipulation of the
object referent we wouldexpect them to favour referents which are more
powerful than the subject referents. This assumption is borne out, in particular
for earlier stages of English, by various object innitive constructions like that
in (10). For instance, compare the evidence in Table 1 dealing with eighteenth
century English. While God is almost exclusively preceded by the preposition in
this environment, most other NPs containing the denite article in initial position
function as direct objects.
Table 1. The distribution of selected nominal complements preceding innitival comple-
ments associated with the verb beg in the Eighteenth-Century Fiction Corpus (17051780)

percentage of I (of)
the N ()
3.2 The principle of clause integration
This takes us to the secondd erivedprinciple, the principle of clause integration.
It may be characterizedas follows:
Principle of clause integration
The degree to which the dependent clause is integrated into its superordinate
clause correlates (inversely) with the degree of its semantic independence
(Givn 1980; Givn 1990: 515561, 973975; Haiman 1983: 795800;
Gramley 1987; Givn 1991a: 9496; Kortmann 1991, 1995).
Informally speaking, a loosely integratedsubord inate clause (which preserves
much of its syntactic autonomy) is semantically more independent than a tightly
integratedsubord inate clause. So in theory we wouldhave to distinguish
various degrees of clause integration.
In the following, we take a brief look at two broad classes of dependent
clauses, complement clauses and adverbial clauses. In the area of verb-dependent
complements, we concentrate on two kinds of contrast. Consider rst examples
like those in (11a,b) or (12a,b) as discussed in Rohdenburg (1995a).
(11) a. John persuadedSusan that she shouldgo.
b. John persuaded Susan to go.
(12) a. John remindedSusan that she shouldgo.
b. John reminded Susan to go.
In (11a) the nite complement clause displays a relatively high degree of
syntactic autonomy reecting a relatively high degree of semantic autonomy. In
contrast to (11b), the inuence exertedby John on Susan in (11a) is compara-
tively weak. Huddleston (1971: 157) once characterized the situation in the
following terms: to persuade someone to do something is to get him to agree to
do it, whereas to persuade someone that he should do something is only to get
him to accept that he ought to. Or to put it more simply, (11b) entails that
Susan hadan intention to go: (11a), however, is not committedto this interpreta-
tion. A similar contrast obtains in (12a) and(12b): While in (12b) an obligation
to go has been imposedon Susan, (12a) needonly refer to a reminder of such an
obligation. In other words, in addition to functioning as typical directive verbs as
in (11b) and(12b), persuade and remind may also be usedas representatives in
cases like (11a) and(12a).
My secondset of examples in (13)(15) involves the so-called phenomenon
of subject to object raising (cf. Rohdenburg 1993).
(13) a. Requesting that more information (should) be given on the subject
/(should) be suppressed
b. Requesting more information to be given on the subject /*to be
(14) a. While conceding that this point is well argued /is misguided
b. While conceding this point to be well argued /*to be misguided
Comparing (13a) and(13b) or (14a) and(14b) we nd that the more basic
complement clause in (13a) and(14a) preserves much of its syntactic autonomy.
By contrast, in (13b) and(14b), the subordinate clause is tightly integrated into
the superordinate clause. As a result, there may be severe semantic restrictions
on the choice of the raisedobject NP in the object innitive constructions. As is
seen in (13b) and(14b), many of these raising structures presuppose some kind
of semantic compatibility between the (S)VO portion andthe overall sentence. In
the literature, such phenomena have been discussed under the heading of
apparent constituents (Bolinger 1967, 1977; Rohdenburg 1993).
In the area of conjunction-headedclauses we nd contrasts such as the
following, as analysedby Sidney Greenbaum (1973: 56):
(15) a. Although he does not receive any payment / Although not receiv-
ing any payment he is preparedto continue working for them.
b. Although she did not take part in last weeks demonstration /
though not taking part in last weeks demonstration she is joining
the march today.
(16) a. While I didnt want to seem obstinate / While not wanting to seem
obstinate I insistedon a denite reply.
b. While he didnt complain yesterday /
*While not complaining
yesterday he feels strongly about it.
(17) a. When they crossedthe bridge, they walkedin pairs.
b. When crossing the bridge, they walked in pairs.
Here we are dealing with fully explicit nite adverbial clauses introduced by the
conjunctions although in (15), while in (16) and when in (17). With these
conjunctions, it is possible, as shown in (15a) and(16a), to reduce the nite
adverbial clauses to non-nite ones, thus integrating them more tightly into the
overall sentences. However, this process cannot be carriedout in (15b) and
(16b). What then is wrong with these examples? In both (15b) and(16b), the two
clauses concernedhappen to involve dierent time spheres. Apparently, the
reductions carried out in cases like (15b) and (16b) are only acceptable to the
extent that both clauses involve the same time spheres.
We turn next to (17a,b). As is to be expected(17b) is restricted to an
interpretation involving a temporal overlapping of the two clauses. This reading
is also available in (17a), but this sentence has an additional interpretation. In
(17a), the two actions involvedmay be sequential rather than simultaneous. On
this interpretation, the past tense in the rst clause corresponds to the past perfect
(or pluperfect): when they had crossed the bridge. This reading is, of course,
excluded in (17b).
3.3 Verb + particle combinations
Another large domain in which the choice between grammatical alternatives may
be determined by the distance principle is provided by particle verbs as shown
in (20a,b):
(18) a. She put out the re.
b. She put the re out.
In most cases involving non-pronominal object NPs there are two positional
alternatives: The particle may either occur next to the verb it belongs to as in
(18a) or it may be separatedfrom it by the object as in (18b). Now the distance
principle wouldpred ict that in cases like (18a) the semantic closeness between verb
+ particle shouldbe greater than in cases like (18b). The principle wouldalso
predict that in cases like (18b) the particle should be more likely to preserve its
original full locative meaning (Bolinger 1952: 11271128). Both predictions are
in general conrmed by the available evidence (cf., in particular, Bolinger 1971).
With some particle verbs we nd that the two positions are clearly dierent
semantically andin the expectedd irection. Compare the data in (19) and(20):
(19) a. They carriedout the operation / the washing.
b. They carried*the operation / the washing out.
(20) a. She brought home the point / the washing.
b. She brought *the point / the washing home.
In (19b) and(20b) the particles (out and home) have remainedrelatively
independent of the verbs carry and bring. This makes them compatible with the
concrete object the washing but incompatible with abstract nouns like operation
or point. By contrast, (19a) and(20a) are ne with both kinds of object. But then
(19a) and (20a) provide two dierent kinds of interpretation, a concrete one in
the case of washing anda grammaticalized or abstract one involving operation
and point where verb + particle have become fusedsemantically and the particle
has lost its concrete meaning.
3.4 Miscellaneous phenomena
There are dozens of further syntactic variation phenomena in (generally) more
restricted domains which are sensitive to the distance principle. Of these, we can
only mention a few in this paper. For instance, consider the optional use of the
preposition at in (21a).
(21) a. Im (at) home right now.
While the preposition usedto be obligatory in cases like (21a) in traditional
British English (cf. e.g. Wood1965: 7274), American English has been able to
do without it for quite some time. However, as was pointed out by Charles
Fillmore (1992: 53), there are some cases like (21b) where the preposition is
obligatory even in American English.
(21) b. My computer is still *(at) home.
This curious contrast between (21a) and(21b) may also be motivatedby the
distance principle. In (21a) we are dealing with the relationship between a person
andhis or her home. In Anglo- Saxon culture, this is a relatively close andd irect
relationship. However, an entirely dierent andless direct relationship obtains
between the speakers computer and the speakers home. By normal standards
our possessions do not have homes. No wonder, then, that the lower degree of
involvement holding between the subject referent and the home in (21b) should
be reected by the greater distance produced by the obligatory preposition.
Turning now to (22), we nd that with most nouns there are two grammati-
cal constructions involving both as illustratedhere.
(22) a. both (of the) girls
b. both (of the) books
The distance principle would suggest that the relationship between the two
entities may be less close or direct in the longer phrase separating both andthe
plural noun. The prediction seems to be conrmed by the examples in (23a,b).
(23) a. both sides / *both of the sides
b. both ends / *both of the ends
As was pointedout by Frederick Wood (1962: 35) when of is usedthe two are
thought of individually and separately. This is why we never have of where one
cannot exist without the other, or where one implies the other.
Consider next the combination of measure phrase + adjective which follows
andmod ies a given noun phrase.
(24) It is importedin bundles (of) about 12 inches long.
In most cases like (24) we have a theoretical choice: The noun phrase and
measure phrase construction may optionally be separatedby the preposition of.
However, there are a few cases like those in (25)(27) where the preposition is
obligatory rather than optional.
(25) They were found to have reading ages of less than eight years old. (BBC)
(26) A spine on Santa Maria, Guatemala, reacheda maximum size of
500 m high and1300 m across the base. (Ollier)
(27) ; fourth, the chip is tiny an ARMGL has a die size of 5.9 mm
square (BNC)
Now the distance principle would predict a concomitant greater semantic distance
in such cases. What, then, is there in examples (25)(27) that points to a less
close semantic relationship between the noun phrase and its associated measure
phrase construction? The answer is suggested by the corresponding predicative
structures. Corresponding to (24), for instance, we have (24).
(24) in bundles which are 12 inches long.
Here a quality is being assigned to the referent denoted by the subject by means
of what Halliday (1967: 6670) calls intensive BE. The existence of such
perfectly natural paraphrases points to a close semantic relationship between the
noun phrase and the measure phrase construction. In cases like (25)(27), by
contrast, such paraphrases are not always completely felicitous, which suggests
a signicantly looser semantic relationship between the NP and the measure
phrase construction. Compare:
have reading ages which are less than eight years old.
reached a maximum size which was 500 m high and 1300 m
across the base.
has a die size which is 5.9 mm square
At any rate, we seem to be dealing here with what Halliday (1967: 6670) refers
to as equative BE, which arguably represents a less close relationship between
the subject and the associated predicative expression.
Finally, let me draw attention to the two kinds of (masculine) posses-
sive pronouns distinguished by (19th century) West Somerset English (Siemund
2001: 25):
(28) a. Bill cut-s vinger.
cut his
b. Bill cut ees vinger.
cut his
The (complementary) distribution of the two forms is clearly motivated by the
distance principle: In the reexive interpretation associated with (28a), the
semantic closeness between subject and object is mirrored by the clitic pronoun.
Conversely, the use of the independent form in (28b) is conned to non-reexive
readings, which involve a greater semantic distance between the subject and
object referents.
4. At the interface of morphology and syntax: The rivalry between
inectional and phrasal comparison
This section takes a brief look at monosyllabic adjectives like proud and full
which are compatible with both inectional andphrasal comparison. For
instance, consider examples (29a,b) and(30a, b):
(29) a. John had never been prouder / more proud.
b. John was even prouder / more proud of his rst cap / to be in the
rst team.
(30) a. The account was very much fuller / more full (than hadbeen
b. The account was much fuller / more full of intriguing remarks.
Up to a few years ago very little was known concerning the factors determining the
Table 2. The distribution of inectional and phrasal comparatives of proud in The Daily
Telegraph & The Sunday Telegraph for 19911994
more proud
percentage of
more proud
all examples
attributive uses
non-attributive uses
a. all examples
b. presence of complements
(of-phrase, to-innitive)
c. absence of complements

choice of the two options in such cases. But then Britta Mondorf (2002) made an
important discovery. Classifying the relevant predicative uses along the lines of the
a- andb-examples she found that adjectives followedby prepositional (or innitival)
complements as in (29b) and(30b) are much more likely to select the
more-comparative than the simpler uses without complement as in (29a) and(30a).
For the adjectives used in (29) and (30), the tendency is reected in the data of
Tables 2 and3. No doubt, these ndings wouldseem to lendthemselves to being
interpretedin terms of the complexity principle (cf. e.g. Rohdenburg 1995a,b, 1996,
1998): The additional complement in the b-cases increases the structural com-
plexity, which may be compensatedfor by the use of the more explicit phrasal
comparison. In essence, this represents the explanation adopted by Britta Mondorf.
However, there is another interpretation suggestedby the distance principle.
Table 3. The distribution of inectional and phrasal comparatives of full in selected British
corpora (d91-d94, M93-M94, BNC [imaginative prose])
more full
percentage of
more full
1. a. all non-attributive examples
b. full + of-complement
c. remaining cases
Arguably, what is being comparedin the b-examples is a close-knit unit consist-
ing of adjective and complement. As is well-known, the inectional comparative
is generally preferredwith monosyllabic adjectives. However, if this variant is
selectedin the b-examples, then elements that go together functionally and
cognitively wouldbe separatedby the sux. It is in cases like these that the use
of the more-comparative may preserve an iconic representation of the relevant
scope relations. The results in Tables 2 and3 suggest that the iconic impulse
may be an important factor in favouring the phrasal comparative.
This line of reasoning is supportedby another range of facts uncoveredby
Mondorf (2000). Compare the choices in (31a,b):
(31) a. more high-minded, more old-fashioned
b. broader-based, heavier-handed
With complex adjectives such as these the incidence of more-comparatives
corresponds very largely to the postulated degree to which the adjectival
constituent has become integratedinto the expression as a whole. The degree of
autonomy of the rst element can be gaugedby two parameters, (1) the frequen-
cy of hyphens, and(2) the frequency of parallel adjective noun collocations. Not
surprisingly, Mondorf (2000) nds that items which are frequently written solid
andwhich have few parallel adjective + noun structures (e.g. those in (31a))
show a greater tendency to adopt the more-comparison than formally related
expressions (like those in (31b)) in which the adjective displays a greater
autonomy in the sense of (1) and(2).
5. Iconically motivated root allomorphy
5.1 The relevance principle
In her survey of fty languages Bybee (1984) has shown that the most relevant
inectional categories do indeed occur closest to the verb stem while the least
relevant occur at the greatest distance from the verb stem. Bybee also points out
that the situation is paralleledin the area of nominal inections. As was shown
by Greenberg (1963: 112), the expression of number almost always comes
between the noun base andthe expression of case. According to Bybee this is
due to the fact that number has a direct eect on the entity referredto by the
noun while case is concernedinsteadwith clause-internal relations. In present-
day English, the expected ordering can only be seen in those few cases where
plural andgenitive suxes are not amalgamated. Compare, for instance:
(32) childrens, brethrens, oxens
In this section, I want to use the relevance principle to explain the divergent
evolution of fusedsingular genitives andplurals co-existing as late as Early
Modern English:
(33) Chaucer EME
a. wyf (sc) wife (sc)
b. wyues ( or ii) wiues/wives ( or ii)
Originally andcertainly as late as Late OldEnglish, the voiceless and voiced
fricatives representedtwo allophones of a single fricative phoneme which were
in complementary distribution. The alternation was due to a universal phonetic
tendency: The voiced variant occurredin voicedsurround ings, in particular when
preceding and following vowels, while the voiceless allophone was used
elsewhere. By Middle English times, and for various reasons, the two fricatives
had become two independent phonemes. This change resulted in the kind of root
allomorphy still foundin the case of wife/wives today. In Late Middle English
and Early Modern English, the schwa was dropped, thus removing the original
phonological motivation for the alternation between voiceless andvoiced
fricatives. This may have been an important factor responsible for the ensuing
reorganization of nominal morphology. Clearly, the more highly relevant plural
hada greater claim to being fused with the stem than didthe less relevant
genitive singular. An additional factor increasing the distance between the noun
andthe genitive may be seen in the development of the genitive marker into a
clitic as discussed by Plank (1985). Not surprisingly, the plural has been able to
resist the regularization pressures much better than the genitive. While the class
of fusedgenitives had virtually disappearedby the 18th century, quite a few
fusedplurals survive to this day.
5.2 Quantity correlations
Today, the class of fused plurals of the wives-type includes the following (Quirk
et al. 1972: 177):
(34) a. wives, wolves, shelves, elves, lives, calves, loaves, thieves, knives,
sheaves, selves, halves, leaves
b. scarves, hooves, wharves, dwarves, handkerchieves
The items in (34a) use the irregular plural exclusively, andmost of them have
done so in the standard language since Early Modern English times. By contrast,
the items in (34b) use the irregular plural alongside the regular one. In some
cases, as is shown in Tables 4 and5, the fusedplural has either become estab-
lishedor reasserteditself in more recent times.
For this paper I have selectedthe body-part noun hoof for a closer analysis.
Table 4. The distribution of the plural forms hooves/hoofs in the quotations of the OED
between the 18th and 20th centuries
percentage of hooves

Table 5. The distribution of the plural forms scarves/scarfs in the quotations of the OED
between the 18th and 20th centuries
percentage of scarves
What, then, are some factors determining at present the distribution of the variant
plural forms hoofs and hooves? Before presenting the evidence in support of an
iconic motivation I wouldlike to turn very briey to a parallel set of data
discussed by Randolph Quirk (1970) in his classic article Aspect and variant
inection in English verbs. Analysing variable past tense forms like those in
(35) by means of forced-choice and judgement tests, Quirk is able to demonstrate
that the irregular -t variants show a striking anity for non-durative aspect in
both British andAmerican English.
(35) spilt/spilled, spelt/spelled, burnt/burned, dreamt/dreamed
While Quirk does not attempt a synchronic explanation of these facts, Bolinger
(1968: 110) suggests that the correlation may be iconically motivated. The regular
form, which is preferredin durative contexts, happens to be very much longer
phonetically than the irregular form. In other words, the longer form tends to be
selectedfor longer or habitual events. A similar correlation has been identied
by Samuels (1972: 161) concerning the distribution of long and short vowels in
a set of Middle English present tense verbs. While point action meaning is conveyed
by short vowels long vowels tend to denote states or duration of action.
Armedwith this fore-knowledge I tried to conceive of a similar iconically
motivatedcorrelation in the case of hoofs and hooves. This time, it is the
irregular fusedplural that is very much longer than the regular form. According-
ly, the longer variant might be interpretedas both an intensied and extended
plural, a superplural as it were denoting a quick sequence of dynamic hoof-events
over a relatively extended period of time. A preliminary analysis identied three
classes of context that might t the concept of an intensied and extendedplural:
1. explicit mentions of masses of hooves as in thousands of hooves
2. reports of fast movements involving hooves which can only be seen but not
heardas in a urry of hooves or ying hooves/hooves ying
3. descriptions of various sound impressions ranging from the clippety-clop via
the clatter to the rumble and thunder of hooves
Of the three classes mentioned, only the third provided suciently large
numbers of examples in the available subcorpora for a successful statistical
analysis. So I decided to concentrate on category (3) for the time being. Here it
was foundthat a slight extension of the category to cover any references to
soundimpressions was sucient to remove virtually all problems of classica-
tion. In other words, examples like (36) referring to the production of sound
eects in a studio would also have to be included in the general category of
(36) How do you do hooves?
So I endedup by comparing the general category of soundimpressions with all
other uses of hoofs and hooves in all of the corpora available at the time to our
project, something like 1300 million words of present-day British and American
Despite the potential shortcomings of the analysis, the results clearly bear
out our expectations for both British andAmerican English. For instance,
compare the evidence presented in Tables 6 and 7 for The Times andthe Los
Angeles Times. The two papers display the same overall tendencies despite the
fact that they happen to represent very dierent stages of irregularization.
Comparedwith the residual category of examples, the category of sound
impressions does indead show a striking anity for the phonetically longer
We can take it, then, that the superplural theory has been conrmed and that
the distribution of the two variants is at least in part iconically motivated.
The theory still has to be eshed out with much more descriptive detail. For
instance, can we expect to nd corresponding contrasts between more or less
salient/extendedsoundimpressions such as thundering hooves and clattering
hooves? Initial observations, though generally failing to reach statistical signi-
cance, point in this direction. And what contexts are most likely to preserve the
shorter regular plural? While I have not yet quantied the category, there is no
doubt that purely anatomical statements involving either individual animals or a
given species still show the greatest resistance to the adoption of hooves.
Table 6. The distribution of the plural forms hooves/hoofs in The Times and The Sunday
Times for 19901997
of hooves
1. all examples 327 45 372 87.9%
2. soundimpressions (thunde(ring)
[n., v., adj.])
3. remaining cases (cloven-) 186
5.3 Relations between derivationally and inectionally produced root/stem
Table 7. The distribution of the plural forms hooves/hoofs in the Los Angeles Times for
of hooves
1. all examples 42 158 200 21%
2. soundimpressions (thunder(ing)
3. remaining cases 18 136 154 11.7%
This section deals with adjective formation by means of the derivational sux
-ed as shown in (37a):
(37) a. loud-mouthed [t/d], stout-calv/fed, yellow-leav/fed, blue-
scarv/fed, cloven-hoof/ved
b. mouths [z], calves, leaves, scarves/scarfs, hooves/hoofs
Here, our main concern is again with the variable presence of the voiced
fricative, which typically shares its origin andearlier phonological motivation
with the plural type in (37b). Comparing these adjectives with their nominal
bases we nd that the derivationally produced root allomorphy (in (37a)) mimes
the inectionally produced root allomorphy (in (37b)) in a number of respects.
The rst observation concerns a simple relation of dependence between the
two classes. The presence of a voicedand longer variant in the derivedad jective
generally presupposes the existence of either an exclusively or variably used
Secondly, at least in the more frequently encountered cases, the -ed
Table 8. The distribution of the irregular plurals of leaf/scarf/hoof compared with the
corresponding -ed adjectives in a selection of British quality papers
1 leaves 100%
2 scarves 98%
3 hooves 88%
leaved 80%
scarved 65%
hooved 25%
derivatives weakly reect the distributional pattern of regular and irregular plural
forms. Compare the rough statistics given for leaves, scarves and hooves and
their relatedad jectives in Table 8 (given on p.279).
In addition, the scant historical evidence we have points in the same direction.
With scarf/ved and hoof/ved the (re)establishment of the voicedvariant seems to
follow in the wake of the increasingly usedvoiced plural. Couldit be that this
drag-chain relationship between derivationally and inectionally produced
allomorphs was also motivatedsemantically , at least in part? With the nouns in
question the voicedvariants mark the category plural in two dierent ways:
1. by means of the sux and
2. by means of the irregular voiced fricative, which produces a distinctly
longer root allomorph
So both on the morphological andthe phonological level, forms such as scarves
and hooves are markedmuch more explicitly for plurality than their regular
variants scarfs and hoofs. This leads us to hypothesize that the explicitly plural
root allomorphs leav-, scarv- and hoov- might preserve at least part of this plural
orientation when transferredto the corresponding -ed adjectives. With leav/fed
and hoof/ved I have failedto nd any striking dierences between contexts
referring to just one or more than one individual plant or animal. Clearly, even
a single hoofedanimal or a single silver-leaved plant generally has more than
one hoof or leaf. But what about scarfed and scarved? Since most individuals
usually wear at most one scarf at a time it seemedworthwhile distinguishing
between singular andplural contexts as in (38) and(39).
(38) while I climbedon in company with a headscarfed woman,
(The Times 1995)
(39) What I remember best about the match was that twenty minutes after
the nal whistle hadblown several thousand blue-scarved people
behindthe Rangers goal hadnot begun to leave. (The Times 1995)
In the rst case, only one person is being referredto while the secondandplural
context involves a (mostly unspecied) number of people wearing a scarf each.
In order to compile a suciently large number of authentic examples I hadto
analyze a corpus of 820 million words of British English. The data in Table 9
clearly conrm the hypothesis: The irregular root allomorph typical of plural
formation shows indeed a striking preference for plural contexts.
Similarly, the few examples of teethed which I have been able to discover
unambiguously focus on a persons set of teeth. Compare:
(40) But in the main those out there having a great time are the white-
Table 9. The root allomorphs scarv-/scarf- preceding the adjective-forming sux -ed in
selected British corpora (t90-t97, g90-g97, d91-d94, M93-M95, BNC)
percentage of scarved
all examples
singular reference
plural reference
teethed, clean-limbed, broad-shouldered o-spring (sic) of money.
(The Guardian 1996)
I have not exploredthe superplural hypothesis with respect to scarf/ved, but it is
just possible that references to a mass of scarves as in (39) favour the use of
scarved to a specially high degree.
6. Conclusion
This paper uses the iconically motivatedquantity andd istance principles to
account for a selection of grammatical variation phenomena which have either
not been dealt with at all or whose iconic motivations have been largely ignored.
The domains discussed cover a broad spectrum of grammatical variants ranging
from inectional andd erivational morphology via contraction andcliticization
phenomena to a wide range of syntactic alternations. The quantity principle
accounts for correlations between full or reducedwordforms (Section 2), the
presence or absence of inectional morphemes (Section 2), redundantly marked
categories andtheir simpler counterparts (Section 2), andphonetic (segmental)
quantity dierences (Sections 5.2, 5.3). Some of the general contrasts reecting
the distance principle include a greater or lesser distance between a given verb
of transfer andthe beneciary (Section 3.1), the presence or absence of a
specic preposition (Sections 3.1, 3.4), the placement of particles with phrasal
verbs (Section 3.3), dierent degrees of subordinate clause integration (Sec-
tion 3.2), a greater or lesser distance between a compared adjective and its
prepositional complement (Section 4), andfusedvs. unfused inectional
categories (Section 5.1).
1. The work reportedhere was carriedout within the research project Determinants of Grammati-
cal Variation in English, which is funded by the German Research Foundation (grant RO
2. The fact that there is an extra dimension of aectedness at work in those variants where the
beneciary occurs in (immediately) post-verbal position has been noticed in passing by a
succession of linguists. Compare, for instance, Larson (1988: 376), who observes that (i) is
perfectly acceptable as an utterance by a pregnant wife to her husbandwhile (ii) is decidedly
odd in this context.
(i) I knittedthis sweater for our baby.
(ii) I knittedour baby this sweater.
However, Thompson andKoid e (1987) were the rst to explicitly point out the iconic
motivation of this kindof semantic contrast.
3. A brief survey of the inuence of semantic connectedness on particle shift (and other kinds
of constituent ordering) is presentedby Wasow andArnold(forthcoming).
4. In a similar vein, Hogg (1988) suggests that the rise of irregular past tense forms like snuck or
struck is due to iconic (phonesthematic) causes.
Primary sources (Electronic corpora)
BNC = British National Corpus
d91-d94 = The Daily Telegraph & The Sunday Telegraph for 19911994
Eighteenth-Century Fiction Corpus (17051780)
g90-g97 = The Guardian for 19901997 (including The Observer for 19941997)
Los Angeles Times for 19921995
M93-M95 = The Daily Mail & The Sunday Mail for 19931995.
OED = The Oxford English Dictionary, 2nded ition
t90-t97 = The Times & The Sunday Times for 19901997
Bolinger D. 1952. Linear modication. PMLA 67: 11171144.
Bolinger, D. 1967. Apparent constituents in surface structure. Word 23: 4756.
Bolinger, D. 1968. Aspects of Language. New York: Harcourt Brace Jovanovich.
Bolinger, D. 1971. The Phrasal Verb in English. Cambridge, Mass.: Harvard University
Bolinger, D. 1977. Apparent constituents in surface structure. In D. Bolinger, Meaning
and Form, 124134. London: Longman.
Bybee, J. L. 1984. Diagrammatic iconicity in stem-inection relations. In Iconicity in
Syntax, J. Haimann (ed.), 1147. Amsterdam: Benjamins.
Bybee, J. L. 1985. Morphology: A Study of the Relation between Meaning and Form.
Amsterdam: Benjamins.
Cheshire, J. 1978. Present tense verbs in Reading English. In Sociolinguistic Patterns in
British English, P. Trudgill (ed.), 5268. London: Arnold.
Cheshire, J., V. Edwards and P. Whittle 1993. Non-Standard English and dialect
develling. In J. Milroy and L. Milroy (eds), 5396.
Dirven, R. 1988. From periphery to nucleus. Some eects of nucleus status in English
andGerman. In Essays on the English Language and Applied Linguistics on the
Occasion of Gerhard Nickels 60th Birthday, J. Klegraf andD. Nehls (eds), 276283.
Heidelberg: Groos.
Dixon, R. M. W. 1991. A New Approach to English Grammar, on Semantic Principles.
Oxford: Clarendon.
Edwards, V. 1993. The grammar of Southern British English. In J. Milroy and L.
Milroy (eds), 214238.
Fai, K. 1989. Nicht-standardsprachliche und dialektale Erscheinungsformen der
englischen Syntax unter Einbeziehung der Sprachgeschichte. In Sprachlicher
Substandard II. Standard und Substandard in der Sprachgeschichte und in der
Grammatik G. Holtus and E. Radtke (eds), 1440. Tbingen: Niemeyer.
Fillmore, C. 1992. Corpus linguistics or computer-aided arm-chair linguistics. In
Directions in Corpus Linguistics ed. by J. Svartvik, 3560. Berlin/New York:
Mouton de Gruyter.
Givn, T. 1980. The binding hierarchy and the typology of complements. Studies in
Language 4: 337377.
Givn, T. 19841990. Syntax: A Functional-Typological Introduction. I-II. Amsterdam:
Givn, T. 1991a. Isomorphism in the grammatical code: Cognitive and biological
considerations. Studies in Language 15: 85114.
Givn, T. 1991b. Markedness in grammar: Distributional, communicative and cognitive
correlates of syntactic structure. Studies in Language 15: 335370.
Gramley, St. 1987. The Innitive Forms of English as Verb Complements. Duisburg:
Linguistic Agency University of Duisburg
Greenbaum, S. 1973. Adverbial -ing participle constructions in English. Anglia 91:
Greenberg, J. 1963. Some universals of grammar with particular reference to the order
of meaningful elements. In Universals of Language, J. Greenberg (ed.), 73113.
Cambridge, Mass.: The M. J. T. Press.
Haiman, J. 1983. Iconic andeconomic motivation. Language 55: 781819.
Halliday, M. A. K. 1967. Notes on transitivity and theme in English. Part 1. Journal of
Linguistics 3: 3781.
Hogg, R. 1988. Snuck: The Development of irregular past tense forms. In An Historic
Tongue: Studies in English Linguistics in Memory of Barbara Strang, G. Nixon and
J. Honey (eds), 3140. London/New York: Routledge.
Hopper, P. J. and S. A. Thompson. 1980. Transitivity in grammar and discourse.
Language 56: 251299.
Huddleston, R. 1971. The Sentence in Written English. Cambridge: Cambridge University
Ihalainen, O. 1985. Synchronic variation and linguistic change: evidence from British
English dialects. In Papers from the 4th International Conference on English
Historical Linguistics, Amsterdam, 1013 April, 1985, R. Eaton, O. Fischer, W.
Koopman and F. van der Leek (eds), 6172. Amsterdam/Philadelphia: Benjamins.
Ihalainen, O. 1987. Towards a grammar of the Somerset dialect. Neophilologica
Fennica 45: 7186. Helsinki: Socit Nophilologique.
Kortmann, B. 1991. Free Adjuncts and Absolutes in English. Problems of Control and
Interpretation. London: Routledge.
Kortmann, B. 1995. Adverbial participial clauses in English. In Converbs in Cross-
Linguistic Perspective, M. Haspelmath and E. Knig (eds), 189237. Berlin/New
York: Mouton de Gruyter.
Kortmann, B. 1996. Adverbial Subordination: A Typology of Adverbial Subordinators
Based on European Languages. Berlin/New York: Mouton de Gruyter.
Kortmann, B. 1999. Iconicity, typology and cognition. In Form Miming Meaning:
Iconicity in Language and Literature, M. Nnny and O. Fischer (eds), 375392.
Amsterdam/Philadelphia: Benjamins.
Larson, R. K. 1988. On the double object construction. Linguistic Inquiry 19: 335391.
Milroy, J. and L. Milroy, eds. 1993. Real English: The Grammar of English Dialects in the
British Isles. London/New York: Longman.
Mondorf, B. 2000. Wider-ranging vs. more old-fashioned views on comparative
formation in adjectival compounds and derivatives. In Anglistentag 1999 Mainz, B.
Reitz and S. Rieuwerts (eds), 3544. Trier: Wissenschaftlicher Verlag.
Mondorf, B. 2002. The eect of prepositional complements on the choice of synthetic
or analytic comparatives. In Perspectives on Prepositions, H. Cuyckens and G.
Radden (eds), 6578. Tbingen: Niemeyer.
Moravczik, E. A. 1978. On the case marking of objects. In Universals of Language, J.
Greenberg (eds), Vol. 4, 249289. Stanford, Cal.: Stanford University Press.
Peitsara, K. 1987. On the uses of this/these here in the dialect of Suolk. Neophilo-
logica Fennica 45: 351361. Helsinki Socit Nophilologique.
Plank, F. 1985. The interpretation and development of form alternations conditioned
across word boundaries. The case of wifes, wives, and wives. In Papers from the
4th International Conference on English Historical Linguistics, Amsterdam, 1013
April, 1985, R. Eaton, O. Fischer, W. Koopman and F. van der Leek (eds),
205233. Amsterdam/Philadelphia: Benjamins.
Quirk, R. 1970. Aspect and variant inection in English verbs. Language 46: 300311.
Quirk, R., S. Greenbaum, G. Leech and J. Svartvik 1972. A Grammar of Contemporary
English. London: Longman.
Rohdenburg, G. 1993. Scheinbare Konstituenten und Verwandtes im Englischen.
Arbeiten aus Anglistik und Amerikanistik (AAA) 18: 249271.
Rohdenburg, G. 1995a. On the replacement of nite complement clauses by innitives
in English. English Studies 76: 367388.
Rohdenburg, G. 1995b. Betrachtungen zum Auf- und Abstieg einiger prpositionaler
Konstruktionen im Englischen. NOWELE 26: 67124.
Rohdenburg, G. 1996. Cognitive complexity andincreasedgrammatical explicitness in
English. Cognitive Linguistics 7: 149182.
Rohdenburg, G. 1998. Syntactic complexity and the variable use of to be in 16th to 18th
Century English. Arbeiten aus Anglistik und Amerikanistik (AAA) 23: 199227.
Rohdenburg, G. 2002. Grammatical parallels between English and Low German dialects
andtheir typological implications. In Anglistentag 2001 Wien, D. Kastovsky (ed.),
Trier: WVT.
Samuels, M. L. 1972. Linguistic Evolution with Special Reference to English. Cambridge:
C. U. P.
Siemund, P. 2001. Pronominal Gender in English: A Study of English Varieties from a
Cross-Linguistic Perspective. Postdoctoral thesis: Free University of Berlin.
Thompson, S. andY. Koide 1987. Iconicity andind irect objects in English. Journal of
Pragmatics 11: 399406.
Wasow, T. andJ. Arnold(forthcoming). Post-verbal constituent ordering in English. In
Determinants of Grammatical Variation, G. Rohdenburg and B. Mondorf (eds),
Berlin/New York: de Gruyter.
Wood, F. T. 1962. Current English Usage: A Concise Dictionary. London & Basingstoke:
Wood, F. T. 1995. A Remedial English Grammar for Foreign Students. Basingstoke &
London: Macmillan.
Wurzel, W. 1998. On markedness. Theoretical Linguistics 24: 5371.
Zipf, G. 1935. The Psycho-Biology of Language: An Introduction to Dynamic Philology.
Cambridge, Mass.: The M. I. T. Press.
</TARGET "roh">
AUTHOR "Wilhelm Ptters"
TITLE "Beatrice"
WIDTH "150"
or The geometry of love
Wilhelm Ptters
University of Wrzburg
The most gentle lady named nr:1icr is the poetical character with whom the
medieval Italian author Dante Alighieri (12651321) created one of the most
celebrated ctionalized women in the whole of literature. Together with the poet
himself, Beatrice appears as a protagonist in his two major poetical works: the
little book entitled Vita Nova (129293) andthe sacred poem called Divina
Commedia (13071321).
The name nr:1icr is a motivatedsign:
morphologically, the structure of
the wordcan be analysedas deriving from the basic morpheme beare, a verb
which means to bless, redeem, make happy, with the sux -(a)tore/-(a)trice
usedto form an agentive noun. Thus, semantically, nr:1icr can be dened as
the lady who brings bliss or salvation, i.e. beatitudo. In fact, the Vita Nova or
New Life is the chronicle of the love inspiredin the young poet by his gentilis-
sima donna. And later, in Dantes transcendental journey to the three lands of the
dead described in the Divine Comedy, it is Beatrice who leads the earthly pilgrim
through the whole universe, until he reaches the limits of space andtime, which
he will nally pass beyondto see the mystery of the Divine Trinity: the eternal
fountain of the First Love.
The object of the following paper is to show that, by means of the ctional
character of nr:1icr, Dantes two great love-stories standin a precise inter-
textual, metapoetical relation to each other. Their interdependence will be dened
as a particular type of iconic representation, in accordance with the concept of
iconicity as described, for instance, in the introduction to Form Miming Meaning
We will observe that one work evokes the other by means of geometric
designs, and that the dierent levels of textual organization form andmeaning
reveal isomorphism produced by mathematical strategies.
The paper is divided into three sections. Section 1 will describe the basis of
a new approach to Dante, especially some famous metaphorical numbers
attributedto Beatrice by her poet. In the secondsection I intendto show that the
numeric signs in Dantes poetry are interdependent elements of a semiotic system
concerning the central message of Dantes works. This part of the paper is a
summary of an article recently publishedin Italy, in the latest volume of
Letteratura Italiana Antica.
Proceeding to some new analyses in Section 3, I will
explain how the poetical numbers relating to Beatrice provide us with a key to
decipher a cryptic coding of the Commedia as a whole.
1. Beatrice and the poetical numbers: A new approach to Dante
1.1 General premises
With regardto the central elements of content andform, we can base our
interpretation on the following well-known facts.
1.1.1 Vita Nova: Beatrice and the number 9
The Vita Nova is a collection of verse arrangedwithin a framework of short
prose passages. In his 31 compositions, mostly sonnets, Dante sings the praise of
Beatrices noble beauty, whereas in the prose parts he introduces the events of
the love-story andcomments on the structure (divisioni) of his poems.
The story itself is simple enough. It begins with Dantes rst encounter with
Beatrice in May 1274, when the two young people are just under nine years of
age. The romance will last 16 years, until Beatrices decease in 1290. In a
mysterious atmosphere createdby obscure metaphors, dark allusions and
visionary encounters with the god Amor, the text treats of love throughout; more
precisely, it deals with a theory of love and with poetry as a medium of expres-
sion for the metamorphosis of amore. Dantes poetical doctrina amoris evolves
from a more courtly concept of love to lead nally to a praise of woman in
general. In this respect, the Vita Nova becomes poetry about poetry, a work of
the highest auto-referential andtheoretical signicance.
With regardto persons or events, there are very few concrete details
mentionedin the text. In general, Dante tends to give only unspecied information.
There is, however, one remarkable exception: the chronological circumstances of the
plot are elaboratedwith almost obsessive precision andritual repetition. This
instance of stylistic marking is relatedto a particular numerical sign, the number 9,
which dominates all indications of external circumstances, e.g.:
the age of the poet andthat of Beatrice when they rst fall in love:
Nove ate gi appresso lo mio nasci-
mento era tornato lo cielo de la luce
quasi a uno medesimo punto, quanto a la
sua propria girazione, quando a li miei
occhi apparve prima la gloriosa donna de
la mia mente, la quale fu chiamata da
molti Beatrice li quali non sapeano che si
chiamare. Ella era in questa vita gi stata
tanto, che ne lo suo tempo lo cielo stel-
lato era mosso verso la parte doriente de
le dodici parti luna dun grado, s che
quasi dal principio del suo anno nono
apparve a me, ed io la vidi quasi da la
ne del mio nono (V. N. II 12).
Nine times since my birth hadthe
heaven of light returnedto almost
the same point in its orbit when to
my eyes rst appearedthe glorious
lady of my mind, who was called
Beatrice by many who did not know
her given name. She hadalread y
been in this life as long as in her
time the heaven of xed stars had
movedtoward the East a twelfth
part of a degree, so that at about the
beginning of her ninth year did she
appear to me, andI saw her at
about the endof my ninth.
the number of years until their next reunion:
Poi che fuoro passati tanti die, che
appunto erano compiuti li nove anni
(V. N. III 1).
After many days had passed, so that
precisely nine years were completed

the exact time at the moment of her most sweet greeting:

Lora che lo suo dolcissimo salutare mi
giunse, era fermamente nona di quello
giorno (V. N. III 2).
The hour in which her most sweet
greeting reachedme was rmly the
ninth of that day
the moment of the vision of the consumedheart:
trovai che lora ne la quale mera
questa visione apparita, era la quarta de
la notte stata; s che appare manifesta-
mente chella fue la prima ora de le
nove ultime ore de la notte (V. N. III 8).
I foundthat the hour in which this
vision hadappeared to me hadbeen
the fourth hour of the night; so that it
appears manifestly to have been the
rst of the nine last hours of the
the position of nr:1icrs name in the listing of the names of the sixty
other beautiful women in town:
maravigliosamente addivenne, cio
che in alcuno altro numero non soerse
lo nome de la mia donna stare se non in
su lo nove, tra li nomi di queste donne
(V. N. VI 2).
wonderfully came to pass, which
is that in no other number wouldmy
ladys name stand except in the ninth
position among the names of these
the moment of another visionary meeting with Amor:
Onde io ricordandomi, trovai che questa
visione mera apparita ne la nona ora
del die (V. N. XII 9).
Then, thinking back, I realizedthat
this vision hadappearedto me at the
ninth hour of the day
the duration of Dantes disease after the death of Beatrices father:
onde io continuamente soersi per
nove d amarissima pena (V. N. XXIII 1).
from which continuously for nine
days I sueredbitterest pain
the moment of the visionary premonition of Beatrices death:
Io dico che ne lo nono giorno a me
giunse uno pensero lo quale era de la
mia donna (V. N. XXIII 2).
I say that on the ninth day I was
taken by a thought that concernedmy
the meeting with Beatrices blessedsoul:
si levoe un die, quasi ne lora de la
nona, una forte imaginazione in me,
che mi parve vedere questa gloriosa
Beatrice con quelle vestimenta sangui-
gne co le quali apparve prima a li occhi
miei (V. N. XXXIX 1).
arose in me one day, at about the
ninth hour, a powerful imagining:
that is, I seemedto see this glorious
Beatrice with those crimson vest-
ments in which she rst appeared
before my eyes
The number 9 is not only a metaphorical chronometer of the love-story. The poet
even goes so far as to identify the very nature of his gentle lady with the ritual
number. In Chapters XXVIII andXXIX he rst emphasizes the vital importance
the number has had in Beatrices destiny. Then he species the date of Bea-
trices death in several intricate occurrences of the numeral. All this leads him to
state that the number 9 was a friendto her. In the endhe even elevates the
abstract connection between the woman andthe number to a numerical denition
of Beatrice herself. He explicitly says: She was a nine:
Tuttavia, per che molte volte lo numero
del nove ha preso luogo tra le parole
dinanzi, onde pare che sia non sanza
ragione, e ne la sua partita cotale numero
pare che avesse molto luogo, convenesi
di dire quindi alcuna cosa
(V. N. XXVIII 3).
Nevertheless, since the number nine has
many times occurredamong the previous
words, which evidently is not without a
reason, andbecause in her departure that
number has clearly gured repeatedly, it
is tting, therefore, to say a few things

Io dico che, secondo lusanza dArabia,

lanima sua nobilissima si partio ne la
prima ora del nono giorno del mese; e
secondo lusanza di Siria, ella si partio nel
I say that, according to the custom of
Arabia, her most noble soul departed in
the rst hour of the ninth day of the
month; andaccord ing to the custom of
nono mese de lanno, per che lo primo
mese ivi Tisirin primo, lo quale a noi
Ottobre; e secondo lusanza nostra, ella si
partio in quello anno de la nostra indizio-
ne, cio de li anni Domini, in cui lo per-
fetto numero nove volte era compiuto in
quello centinaio nel quale in questo mondo
ella fue posta, ed ella fue de li cristiani del
terzodecimo centinaio. Perch questo num-
ero fosse in tanto amico di lei, questa
potrebbe essere una ragione: con ci sia
cosa che, secondo Tolomeo e secondo la
cristiana veritade, nove siano li cieli che si
muovono, e, secondo comune oppinione
astrologa, li detti cieli adoperino qua giuso
secondo la loro abitudine insieme, questo
numero fue amico di lei per dare ad inten-
dere che ne la sua generazione tutti e nove
li mobili cieli perfettissimamente saveano
insieme. Questa una ragione di ci; ma
pi sottilmente pensando, e secondo la
infallibile veritade, questo numero fue ella
medesima; per similitudine dico, e ci
intendo cos. Lo numero del tre la radice
del nove, per che, sanza numero altro
alcuno, per se medesimo fa nove, s come
vedemo manifestamente che tre via tre fa
nove. Dunque se lo tre fattore per se
medesimo del nove, e lo fattore per se
medesimo de li miracoli tre, cio Padre
e Figlio e Spirito Santo, li quali sono tre e
uno, questa donna fue accompagnata da
questo numero del nove a dare ad inten-
dere chella era uno nove, cio uno mira-
colo, la cui radice, cio del miracolo,
solamente la mirabile Trinitade. Forse
ancora per pi sottile persona si vederebbe
in ci pi sottile ragione; ma questa
quella chio ne veggio, e che pi mi piace
(V. N. XXIX 1 4).
Syria, she departed in the ninth month of
the year, for there the rst month is Tixryn
the rst, which for us is October; and
according to our custom, she departed in
that year of our indiction, that is, of the
years of our Lord, in which the perfect
number nine [
] hadbeen completed nine
times in that century in which she had
come into this world, and she was a Chris-
tian of the thirteenth century. As to why
this number was so much her friend, this
couldbe one reason: since, according to
Ptolemy andaccord ing to Christian truth,
nine are the heavens that move, andac-
cording to common astrological opinion,
the saidheavens inuence the earth below
according to their reciprocal relation, this
number was a friendto her to make it
understood that at her conception all nine
moving heavens were in the most perfect
relationship to one another. This is one
reason for it; but if we consider more
subtly andaccord ing to infallible truth, this
number was she herself; I speak by simili-
tude, and I explicate it thusly. The number
three is the radical of nine, since, without
any other number, through itself it makes
nine, as we see self-evidently that three
times three is nine. Therefore, if three
multipliedby itself is the factor of nine,
andthe factor of miracles multipliedby
itself is three, that is, Father, Son, and
Holy Spirit, who are three andone, this
lady was accompanied by this number
nine to make it understood that she was a
nine, that is, a miracle, whose root that
is, of the miracle is solely the won-
drous Trinity. Perhaps a still more subtle
person might see in this matter a more
subtle reason; but this is the one I see in
it, andwhich most pleases me.
The poet gives two reasons to explain the strange identity between his lady and
her number. As he points out, Beatrice can be dened as the number 9:
1. because the 9 is the structural principle of the ninefoldsystem of the
cosmos, whose concentric circles were in the most perfect harmony at the
exact moment of Beatrices conception.
2. Beatrice was a nine, the author continues, because that number is distin-
guishedby having a square root of highly symbolic value: the number 3. As
he emphasizes, the number 3 can be regarded as the numeric sign of the
Divine Trinity andtherefore be recognizedas the wondrous root of the
miracle (= miracolo) calledthe nine or nr:1icr.
If we consider the character of nr:1icr a poetical device in which the central
ideas of the poets message are condensed to a semiotic nucleus, then Chapter
XXIX of the Vita Nova teaches us that the rst question of our exegesis should
be the following: why andto what aesthetic end didthe poet create a numerical
identity for his donna?
Given the particular nature of Beatrice andthe special rhythm of the love-
story, it is worthwhile to look at the temporal framework of the romance as a
whole. From the moment they fall in love, in May 1274, until Beatrices death
in 1290, their relationship lasts 16 years:
1265 May: birth of Dante
1266 birth of Beatrice January:
1274 May: first meeting
1290 June: death of Beatrice
age of Dante and Beatrice
duration of the love story
age of Dante and Beatrice
On examining this chronological scheme, we are ledto ask another important
question: why is the love-story as a temporal event ruledby the set of the three
numbers 91625?
Let us now sum up some basic facts of the Commedia.
1.1.2 Commedia: The number 9 and the order of the cosmos
Ten years after Beatrices death, in the Holy Week of the Holy Year 1300, the
35-year-old Dante is going through the darkest moment of his life: a deep moral
and intellectual depression. This kind of medieval midlife-crisis is the starting
point for the metaphysical undertaking dened as the sacred poem, on which
heaven andearth have put their hand.
The text starts with Dante nding
himself lost in the dark wood a landscape that is a metaphor for his alien-
ation from God in the decade after Beatrices death.
Fortunately, he meets the
ghost of Virgil, his literary hero andexample. The Roman poet arms that,
after having seen the horrors of Hell andthe pains of Purgatory, Dante will be
puried and his soul be savedaccord ing to the plan of three celestial ladies: the
Virgin Mary, Saint Lucy andthe donna of his heart, nr:1icr. While Virgil has
to accompany him on his pilgrimage through the rst two realms of the hereaf-
ter, Beatrice will be Dantes secondguid e andtake him across the nine spheres
of Paradise up to the largest and outermost circle of the cosmos.
It is known that Dante conceivedthe structure of the entire cosmos andthe
three parts of the transcendental space in accordance with the theory of the
Greek astronomer Ptolemy (2ndcentury AD).
In this cosmological vision of the
universe, the centre is occupiedby the planet Earth, dened as an immobile
globe or ball. Aroundthe Earth, the nine spheres of heaven revolve in
perpetual movement. The globe itself is divided into three quarters ocean and one
quarter land: the great dryness, or quarta habitabilis, as Dante calls it, which
rises out of the water in the northern hemisphere. In the middle of the quarta
habitabilis we nd Jerusalem, the hub of the world.
On the other side of the
globe, diametrically opposite Jerusalem, the mountain of Purgatory looms out of
the ocean desert as the only body of land in the southern hemisphere. Dante
describes it as having the form of a steep cone, with a rugged cli base (= the
Antipurgatorio), divided into two levels (1st and 2nd balzo), anda top of seven
concentric belts containing the areas of purication. Thus we see that the
Purgatorio is also arrangedas a system of 9 circular units.
In geometrical
analogy to the cone of the mountain of Purgatory, Hell is formedlike a crater,
with nine concentric circles. The whole circular system bores into the Earth,
descending down to the centre of the globe, where we nd the darkest point of
the universe: the area of Lucifer.
As a counterpart to the lowest point of evil and darkness, we nd, beyond
the nine spheres of the cosmos, the Empyreum or Sphere of Pure Light. This is
the place of non-space andnon-time, the area of unlimitedextension, of God.
Dantes vision of the Creator is that of a point of extreme luminosity, surrounded
by the nine circles of the angels. The metaphysical rays which emanate from the
eternal light cast their reections onto the outside edge of the cosmos, thus
producing a gigantic rose consisting of luminosity, the Rose of the Blessed
Souls. In its formal structure, the White Rose looks like the circular edice of
a classical arena.
Taking into consideration the major details of Dantes poetical universe, we
come across one formal pattern over andover again: the fundamental principle
underlying the design of the cosmos and of innity nds its expression in the
geometrical gure of the circle. Generally regarded as the epitome of formal
perfection, the circle has been usedby the Creator to design the macrosystem of
his cosmos using several ninefoldcombinations of concentric circles or spheres.
At this point we are able to grasp one rst evident aspect of the relation between
the nature of Beatrice andthe structure of the cosmos: the donna of the Vita
Nova is identied with the number 9 because Gods work as described in the
Commedia is organizedin terms of 9.
1.2 Further numeric signs relating to Beatrice
In addition to the numeric sign for Beatrice, the 9, we nd two other mystic
numbers relating to Dantes donna: 515 and61.
1.2.1 The number 515: Beatrices prophecy
The rst of these items is a cryptic code which we nd in the last canto of the
Purgatorio: a ve hundredten and ve. This puzzling combination of numbers
is ascribedto the messenger of God, announcedby Beatrice as the future
saviour who will come andkill the forces of Evil, namely the wickedgiant and
the whore of Babylon:
nel quale un cinquecento diece e cinque,
messo di Dio, ancider la fuia
con quel gigante che con lei delinque
(Purg. XXXIII 4345).
In which a ve hundred ten and ve,
A messenger of God, shall kill the whore
Together with the giant who shares her sin.
Beatrice herself calls her prophecy an enigma forte which Dante has been
designated to explain to mankind after his return to Earth. Dante scholars refer
to this strong puzzle as the DXV-problem, about which pages andpages
have been written without producing any convincing solution.
1.2.2 The number 61: Beatrices name
A further numerical metaphor concerns the 61. This number is a code for the name
nr:1icr. In Chapter VI of the Vita Nova, Dante tells us that he hadonce felt the
desire to compose a lyric in praise of Beatrices name along with the names of 60
other beautiful ladies in town. He adds that he nally abandoned his idea. The
strange conclusion of this chapter emphasizes the hidden sense of its message:
Dico che in questo tempo che questa don-
na era schermo di tanto amore, quanto da
la mia parte, s mi venne una volontade di
volere ricordare lo nome di quella gentilis-
sima ed accompagnarlo di molti nomi di
donne, e spezialmente del nome di questa
gentile donna. E presi li nomi di sessanta
le pi belle donne de la cittade ove la mia
donna fue posta da laltissimo sire, e com-
puosi una pistola sotto forma di serventese,
la quale io non scriver: e non navrei
fatto menzione, se non per dire quello che,
componendola, maravigliosamente addi-
venne, cio che in alcuno altro numero
non soerse lo nome de la mia donna
stare se non in su lo nove, tra li nomi di
queste donne (V. N. VI 12).
I say that during that time when that lady
was a screen to so much love as came
from my part, the desire arose in me to
recollect the name of this most gentle one
andto accompany it with names of many
ladies, and especially with the name of this
gentle lady. And I took the names of sixty
of the most beautiful ladies of the city
where my lady was placed by the most
high Lord, andI composedan epistle in
the form of a serventese, which I will not
inscribe; andI would not have mentioned
it, except to say that which, as I composed
it, wonderfully came to pass, which is that
in no other number wouldmy ladys name
standexcept in the ninth position among
the names of these ladies.
Years ago, ManfredHard t successfully revealed a central semantic aspect of
Chapter VI of the Vita Nova, showing that the number 61 (=60 +1 = the sixty
ladies + Beatrice) can be read according to the rules of medieval gematria, i.e.
the transposition of letters by numbers.
On the basis of the equations A=1,
B=2, C=3 and so on, Hardt identied the number 61 as the numerical (gema-
tric) value of the name nr:1icr: B=2, E=5, A=1, T=19, R=17, I =9, C=3,
E=5 tot. =61.
2. First results: The numeric signs as elements of a semiotic system
It is commonly maintainedby all Dante scholars that the above-mentioned
numbers 9, 61 and515 must have a central importance for the inner order of
Dantes symbolic universe. Our approach to these enigmatic numbers is basedon
a very elementary question: what kindof logical or aesthetic relation can be
the poetical numbers relating to Beatrice, on the one hand, and
the transcendental journey across the cosmos, on the other?
2.1 The product of 61 and 515
Let us begin our analysis with the two values 61 and515, the gematric name
of Beatrice and the numeric coding of the name of Gods messenger.
2.1.1 The circle as basic design
If we consider the numeric signs in their original sense, i.e. as numbers, we can
subject them to a series of simple mathematical tests by using the four funda-
mental operations of arithmetic. By multiplying them, we obtain a really amazing
result. The product of the factors 61 and 515 is 31415. This number is a
surprisingly goodapproximation to the value 10000 =31415.926 The
precision provided by the product of the factors 61 and 515 is so good that we
can exclude the possibility of an accidental coincidence. The number
=3.1415926 is one of the most famous values in mathematics: it is the
general key to resolve all problems concerning circular gures. It denotes the
constant ratio of
circumference andd iameter or
semicircumference andrad ius.
In any given circle the ratio dened by can be illustratedwith the basic circle
in which all dimensions refer to the radius r =1:
c : d =
: r
c c
_ _
2 2
r = radius
d = diameter
c = circumference
= semicircumference
A = Area
r = 1
d = 2r = 2
c = 2r = 2
A = r =1 =
p p
p p p
As we have seen, Dantes very goodapproximation to 10000 can be dened
by means of the poetical numbers 61 and515 considered as factors of the
product 31415 10000. Bearing in mindthe particular structure of Dantes
universe, let us now consider the value 10 000 as a measure of the area A of
a circle. First of all, we can calculate its radius according to the formula
: r

r =

With this radius we draw a circle which we can identify as Dantes basic model:
A = .926...
31 415
= 515
61 31 415
r = 100
From this gure we can easily derive the circle with radius 1. The only operation
we have to perform is to divide the product of 61 and 515 (which is 31 415) by
10 000 (or 100
). The obtainedresult, 3.1415 precisely, is an excellent approxi-
mation to the value of itself: 3.1415926
The number is as mathematicians say irrational, which is why it can
only be calculatedby the scientic method of approximation. Archimedes (240
BC) was the rst to dene the question of as a problem of approximation. With
his methodof measuring the circle by calculating inscribedandcircumscribed
polygons, he indicated the interval of the two gures with two fractions:
3 1408 3 14159 3 1429
< <
< <

. . ... .
Modern mathematics has determined more than 6 billion decimal places without,
however, having discovered any recurrent scheme in the series of numbers.
2.1.2 The circle in medieval geometry
In medieval geometry, at least in the ocial academic treatises, scholars do not
seem to have had any knowledge of Archimedes scientic method, and apparently
they were also unaware of the ndings of Indian and Arabian mathematics.
Almost all the authors of geometrical texts in medieval Europe use the fraction
as if it were an absolute value. Thus the corresponding model
3 1429 .
presentedin medieval treatises is always basedon the fraction :
= 22
Comparing this gure with Dantes basic design, we notice that the fraction
=3.1429 is considerably less accurate than the poetically expressed value
=3.1415, as seen in the following table:
61 515
100 100


= =

3.1415 ...
This observation shows that our historical knowledge of the state which the
sciences hadreachedin the Middle Ages is rather incomplete. The most surpris-
ing aspect from the methodological point of view is the fact that, using his value
, Dante seems to have masteredd ecimal fractions. This revolutionary
31 415
10 000
method stemming from Indo-Arabian mathematics, introduced to Europe by the
great Leonardo Pisano (called Fibonacci) in his Liber abbaci (1202/28), remained
a scientic novelty for almost three centuries, known only to a small number of
As we have seen, Dantes circle with the radius 100 is a decimal variant of
the standard model of the circle, dened by r =1. This circle is the model of all
Figure 1. Deus geometra.
possible circles, andthus of all circular gures which constitute the cosmos, the
physical cosmos of Ptolemy andthe ctional one of Dante. The circle with
radius 1 is the geometric eikn of the universe in Ptolemys theory; according to
ancient and medieval thought it represents the divine idea of Creation. The
metaphysical dimension of the circle may explain the fact that medieval scholars
do not seem to have spoken in public about their real scientic research and
ndings. Presumably, they kept their studies secret because the gure of the
circle was considered to be a central element of the mystery of Creation. Any
attempt to penetrate this domain of Gods omniscience was likely to be regarded
as a sign of human superbia. The spiritual quality which is attributedto the
circle can easily be recognizedin medieval manuscripts. For instance, in the
Bible moralise of 1220 the Creator is representedas a deus geometra, holding
the compass in his handat the very moment he is about to create the world in
accordance with the idea of circularity.
We can come to a rst conclusion. As Gods own circle with the radius
r =1 is the basis of Dantes poetical circle with r =100, the model of the divine
Creation can be determined by values which are exclusively dened by Dantes
poetical numbers 61, 515 and100 (= number of cantos):
r = 1
c = 2
A =
r 1
c 2
61 515
61 515
r = 1
r 1
100 100
100 100
(number of cantos)
BEATRICE Gods messenger
The correspondence of the two gures is much too striking to be a mere
2.2 The product of 9, 61 and 515
Let us now take Beatrices number, the 9, andcombine it with 61 and515.
Multiplying the three factors, we obtain 282735 90000 .
2.2.1 The geometric archetype of the cosmos
The new product can be regarded as the area of another circle, dened by the
radius r =300:
r = 300
A = 90 000p
= 300
90 000p
90 000p
90 000
The two areas A=90000 andA =10000 can be interpretedas the outermost
andthe innermost unit in a system of nine concentric circles. The whole
geometric ensemble is dened by a set of values in which all dimensions can be
derivedfrom Dantes poetical numbers relatedto his belovedBeatrice:
r = A:
1 (61 515) 10000
2 (61 515) 20000
3 (61 515) 30000
4 (61 515) 40000
5 (61 515) 50000
6 (61 515) 60000
7 (61 515) 70000
8 (61 515) 80000
9 (61 515) 90000
1 100 = 100
2 100 141.421
3 100 173.205
4 100 = 200
5 100 223.607
6 100 244.949
7 100 264.575
8 100 282.843
9 100 = 300
This system can be simplied by taking away the powers of 10:
A = r
r = A:

1 = 1
2 1.41421
3 1.73205
4 = 2
5 2.23607
6 2.44949
7 2.64575
8 2.82843
9 = 3
With this simplied set of relations it is possible to draw the most abstract
system constructedby means of 9 concentric circles. We thus obtain an archetyp-
al model of the whole cosmos, a perfect eikn of the Creation as Dante con-
ceivedit on the basis of Ptolemys theory:
Drawing a line which connects the radii of the nine concentric circles, we obtain
nine right-angledtriangles. According to the theorem of Pythagoras (a
the short sides of each of these triangles are equal to one. Thus we obtain a
spiral gure. This gure seems like a geometrical scheme of the poets long way
to God, an abstract representation of the three stages our pilgrim had to master
during his spiritual journey through the realms of the hereafter:
2.2.2 The mathematical order of the cosmos
Let us now take another look at the archetypal design of the Ptolemaic cosmos.
We observe that a certain order governs the relation between the circles. This
order can be described in the form of a mathematical law:
If the ratio of the areas of two circles is n : 1
the ratio of the corresponding radii is n : 1
For example, the relation between the inner andthe outer circle can be described
as shown in the following diagram:
outer circle :1io inner circle
:r: A
= 9 9:1 A
= 1


9 =
=3 3:1


1 =
The two numbers 9 and3 standfor the law that dominates the inner and the
outer circle of Ptolemys cosmos. By recognizing this order, we can nally
understand the secret sense of Chapter XXIX of the Vita Nova, where Dante
himself explains Beatrices numerical metaphor, the number 9, by referring to its
cosmological function andits theological dimension as a square power of the
divine number 3. This, indeed, is the reason for her being a miracle conceived
by Godhimself: the Holy Trinity is dened as the square root of the gentilis-
sima donna.
3. Poetry and geometry: The iconic diagram of Dantes Commedia
3.1 The circular design of the Commedia
3.1.1 The principle of the three-in-one
For the next step of our analysis we have to introduce a new numerical item, a
number which stands for the work as an aesthetic whole. This numeric repre-
sentation of the Commedia is the number 14233, the total number of verses in
the poem.
As a rst andmajor principle of the formal structure of the text we can
identify the division into three parts or canticles. Dante applies this principle of
tripartition to all levels of his work. It is even present in the basic unit of the
text, the terza rima, a metrical scheme basedon stanzas of three lines rhyming
aba bcb cdc etc.
In accordance with the fundamental idea of tripartition reected in the entire
structure of the Commedia, we can divide the number 14233 by 3, obtaining in
this way the following set of equations:
I: 1/3 14233 = 4744.333
II: 2/3 14233 = 9488.667
III: 3/3 14233 = 14233.000
Analysing these results in the light of our circular designs, we discover that the
equation 2/3 14233 =9488.667, which marks the beginning of the ight across
the nine spheres of Paradise, is an excellent approximation to

=9488.531016 This observation leads us to a geometrical denition of the
Commedia as a metrical whole: 14233

3.1.2 The macro-circle of the poem
On the basis of the observation that 2/3 14233 =9488.667 is surprisingly near
to 9488.531016=
, we can obtain a new approximation to by taking the
8th root: . The result shows that Dante must have
14 233 8 = . ... 3.14159
known a value of which is correct to the fth decimal place, whereas the
fraction used in medieval geometry has only two correct places.
If we consider14 233 as the length of a diameterd, we can draw the
following gure dened by a circumference c which is equal to 3/2

d = 14 233
c =
= 9

Dened by a diameterequal to the total numberof verses of the Commedia, this

gure can be identied as the macro-circle based on the metric corpus of
Dantes poem. In the equation indicating the circumference of this design,
, we observe a really striking constellation of Beatrices metaphor 9,
c = 9

of her divine square root =3 and of the mathematical mystery of the divine 9
Creation, the number .
3.2 The spherical design of the Commedia
All the geometric designs proposed in the preceding chapters of our analysis are
gures of plane geometry. However, if we want to obtain a more realistic view
of Dantes ctional world, we have to bear in mind that the realms of his
cosmos are three-dimensional and that the whole cosmos is, in fact, not a two-
dimensional circle, but takes the form of a sphere.
3.2.1 The volume of the largest sphere
This observation leads us to construct the macro-sphere by means of the given
diameterd =14233.
The method usedby medieval geometry in calculating spheres is basedon
the ratio between a sphere andits circumscribed cube, which is V :V =:6.
V : V = : 6
In medieval geometry, the ratio :6 0.523599 is expressedby the fraction
11/21 0.52381. On the basis of the ratio /6, the volume of a sphere can be
measuredby means of the following formula: V =d
/6. Using the number
14233 andits equivalent 3/2
as the diameter of a sphere, we can execute our
operation in two ways, with normal numerals andwith symbolic language using :
V = V = d

a = d = 14 233
6 6
d 14232.79652 d= 3/2
V (14232.79652)
V = (3/2
V 2883173130000 V = 27/8
V = V /6 V = V /6
V 2883173130000 /6 V = 27/8
V 1509625921000 V = 9/16
The result of our operation is a spherical design which we can dene by its
volume. The dimension is, with stupefying precision, V =9/16
V =
d =
3.2.2 The iconic function of the geometric design
If we take the fraction 9/16
as the volume of the outer sphere of a complete
Ptolemaic cosmos, we can derive a corresponding system of nine concentric spheres
by carrying out a very simple operation. The numerator, whose rst element is
equal to 9, has to descend from 9 down to 1, as listed in the following table:
V V d c
IX 9
9 (

p )
3 2
) 9
3 2
8 (

p )
3 2
) 8
3 2
7 (

p )
3 2
) 7
3 2
VI 6
6 (

p )
3 2
) 6
3 2
V 5
5 (

p )
3 2
) 5
3 2
IV 4
4 (

p )
3 2
) 4
3 2
3 (

p )
3 (
3 2
) 3
3 2
II 2
2 (

p )
( 3
) 2
3 2
I 1
1 (

p )
3 2
) 1
3 2
Continuing our analysis of the fraction 9/16
, we see that the sign is
surroundedby the numbers 9, 16 and25. We remember that, in the love-story
toldin the Vita Nova, the rst meeting of Dante andBeatrice takes place in May
1274, when they are just under 9 years of age; andthat the romance lasts a total
of 16 years, until the relationship comes to an endon Earth, with Beatrices
death in 1290, when the lady and her poetic admirer are in their 25th year.
From this, we can conclude that Dante has used the same set of numbers for
the chronological structure of the romance in the Vita Nova as for the geometri-
cal conception of his Commedia. Or, expressedin terms of iconicity: the
mathematical form of the Commedia mimes a central part of the meaning of the
Vita Nova andof the poem itself.
The correspondence of the two poetic works expressed in the set of numbers
91625 reveals that Dantes approximation of in the volume of the sphere
must really have been extremely advanced for its time, otherwise the
volume V =9/16
wouldnot leadto a diameter corresponding to a total
number of 14233 verses. The surprising extent of Dantes mathematical know-
ledge shows that he was well aware of the central problem of cosmography.
Obviously, he was convincedthat the measuring of the cosmos andthe geo-
graphic exploration of the globe hadto be foundedon a precise solution to one
of the most basic scientic problems: the calculation of the circle andof the
sphere as such. The irrational number provides an elementary scientic key to
cosmology, geography andnavigation with all their concomitant economic and
cultural implications.
Let us now come to our last question in this paper: why did Dante select the
particular triplet 91625 as the determinant factor in the poiesis of his works?
As squares of 345 respectively, 91625 evoke one of the earliest andmost
famous mathematical discoveries in the history of occidental science, the theorem
of Pythagoras:
a + b = c
9 + 16 = 25
2 2 2
3 + 4 = 5
2 2 2
3 4
Employing the elements of Pythagoras theorem as the arithmetical basis for the
composition of his poetry, Dante has conceived his works as a hidden homage to
the great Greek thinker. Dantes poetry represents a semiotic adaptation of one
of the most famous examples of human science. In the Middle Ages, Pythagoras
was known as the father of human thought in so far as it is dened as philo-
sophia, which literally means love of wisdom. In his philosophical treatise,
Convivio, Dante puts special emphasis on the eminent historical role which
Pythagoras plays as the creator of the philosophical concept of mans search for
scientic truth. He says in Convivio III xi 3:
Dico adunque che anticamente in Italia,
quasi dal principio de la costituzione di
Roma, che fu [sette]cento cinquanta anni
innanzi, poco dal pi al meno, che l Sal-
vatore venisse, [] viveva uno losofo
nobilissimo, che si chiam Pittagora.[]
Questo Pittagora domandato se egli si
riputava sapiente, neg a s questo voca-
bulo, e disse s essere non sapiente, ma
amatore di sapienzia. E quinci nacque
poi, ciascuno studioso in sapienzia che
fosse amatore di sapienzia chiamato, cio
losofo; ch tanto vale in greco philos
com a dire amore in latino, e quindi
dicemo noi: philos quasi amore, e
soph[os] quasi sapien[te]. Per che vedere
si pu che questi due vocabuli fanno que-
sto nome di losofo, che tanto vale a
dire quanto amatore di sapienza.
I declare, then, that in antiquity in Italy
shortly after the founding of Rome, which
took place approximately 750 years before
the coming of our Saviour [] there lived
a most noble philosopher called Pythago-
ras. [] When Pythagoras was asked
whether he considered himself a wise man,
he refusedto be describedby this term and
declared himself to be not a wise man but
a lover of wisdom. From this there arose
the custom of calling anyone devoted to
wisdom a lover of wisdom that is a
philosopher, for the Greek wordphilos
is the equivalent of love in our language
andso we say philos signifying love
andsophia signifying wisdom. It is
clear, then, that these two words make up
one term, philosopher, which is equiva-
lent to a lover of wisdom.
From these words about Pythagoras and the origin of philosophy we can get an idea
of the historical andspiritual background to the geometrical conception underly-
ing Dantes poetical tales of love. Thus we can come to a provisional conclusion.
By means of a limitedset of cryptic numbers relating to Beatrice, Dante hands
us the key to the understanding of the mathematical coding of his poetry. In fact,
with her poetical numbers 9, 61 and515, Beatrice instructs us to decipher the
hidden meaning of the Vita Nova andthe Commedia. The truth underlying [the]
ne lies
of the poetry has its origin in Dantes scientic aim of unveiling the
true principle and the inner order of the cosmos. The ndings of our interpreta-
tion can be summarizedin the following scheme:
3 4

volume of
the poetical
Vita Nova:
story of Dantes .. LOVE
to Beatrice

temporal frame: . 9-16-25

story of Dantes
journey across
the whole COSMOS
towards God the
fountain of Eternal LOVE
= founder of the
philosophy = . LOVE
of wisdom

In that it treats at the same time of human love andof scientic eros, Dantes
poem reveals itself as a text of highly auto-referential character. Presenting his
love-stories with the aim of explaining a geometrically-basedtheory of the
cosmos, the poet establishes an iconic connection between content andform,
between the Vita Nova andthe Commedia, between poesia and losoa. In so far
as both levels of the text treat about the cosmos, the hidden form of Dantes
poem is a diagram creating an iconic image of the content.
In the aesthetic
unity of the poetical expression, the two methods of human understanding
poetry andphilosophy are combinedin perfect symbiosis in the text as the
supreme unit of form andmeaning.
By choosing for the lady of his heart the name nr:1icr, it was Dantes
intention to provide us with a concentrated summary of the whole poetical
message. Thus the name itself is ttingly used to illustrate Dantes own individu-
al iconic theory of the art of literature. In fact, his poetical concept of iconicity
is basedon the principle of linguistic form-meaning isomorphism. In the words
of the poet: Nomina sunt consequentia rerum The names express the nature of
the things they designate.
The lightning ash of absolute cognition, striking Dante at the very moment
he meets the primo amore, God, is considered by the poet as the highest
fullment of his desires, as the most sublime realization of his concept of love.
But we must see Dante as both, poet andphilosopher . The eros inspiring his
scientic quest has to be understood as love of wisdom, dedicated to the
unattainable lady, vri1:s. This kindof amore is the purest form of love, which,
in the end, leads to that bliss or salvation already promised to the young Dante
by the etymological sense of the name of his rst andeverlasting love: Beatrice.
I wouldlike to acknowledge the help andsuggestions given me by the following colleagues and
students: Sandra Ellena, Eva J. Ernst, Colin Humphrey, Christiane Schneider, Margaret Wirth and
Andrea Wutzer. I would also like to express my thanks to Silvia Feser for compiling the manuscript
of this paper.
1. Cf. Fischer andNnny 2001, 19.
2. Cf. Commedia, Paradiso XXIX 18 and Paradiso XXXII 142.
3. Cf. Nnny andFischer 1999: xvxxvi.
4. Cf. Ptters 2001.
5. Cf. Wehle 1986.
6. The English translation of Cervigni andVasta is wrong here by having the Italian numeral nove
nine refer both to the perfect number andto times. The correct English version of the
segment, It. in cui lo perfetto numero nove volte era compiuto, is the following: in which the
perfect number (i.e. ten!) had been completed nine times.
7. Cf. Paradiso XXV 2.
8. Cf. Inferno I 2.
9. See e.g. the synopsis in Marchese 1993. For Dantes cosmology cf. Boyde 1984.
10. Cf. Hausmann 1988.
11. Cf. Viti 1964: ii.
12. Cf. Viti 1964: i.
13. Cf. Mazzamuto 1970 andPalma di Cesnola 1995. Cf. also Ptters 1997.
14. Cf. Hardt 1973, 1989, 1995, 2000.
15. Cf. Ptters 2001.
16. Cf. Wedeniwski and Haenel 1996.
17. Cf. Tropfke 1940, Ptters 1987, Gericke 1992, Busard1997, Delahaye 1997, Ptters 1998.
18. For further information see Tropfke 1940, Gericke andV ogel 1965, Ptters 1987: 104114,
DAmore 1993 and1995, Ptters 1998: 4282.
19. Codex Vindobonensis 2554 (sterreichische Nationalbibliothek Wien).
20. Cf. 2.1.1.
21. See the references in note 17.
22. See our chart in 2.1.1.
23. See the title of Nnny andFischer 1999.
24. For the different aspects of Dantes cosmographia see especially trattato II andIII of Convivio
andthe authors last work, the treatise De situet forma aque et terre.
25. Cf. Convivio II I 4.
26. Cf. Nnny andFischer 1999: xxixxvi.
27. The iconistic theory of human language expressedin this quotation of Vita Nova (XIII 4) will
be modied by Dante in his later works to a more symbolic concept of the linguistic sign,
especially in De vulgari eloquentia (I III 3) andin Paradiso XXVI 130132, where we nd a
poetic denition of the concept of ad placitum, usedby medieval scholars to refer to the normal
arbitrariness of the formmeaning-relation.
1. Texts
Alighieri, Dante. 1932. La Vita Nuova. Edizione critica per cura di Michele Barbi, Societ
Dantesca Italiana. Edizione Nazionale delle Opere di Dante. Firenze: Bemporad e
Alighieri, Dante. 1996. Vita Nova. A cura di Guglielmo Gorni. Torino: Einaudi.
Alighieri, Dante. 1995. Vita nuova. Italian Text with Facing English Translation by Dino
S. Cervigni and EdwardV asta. Notre Dame andLond on: The University of Notre
Dame Press.
Alighieri, Dante. 1995. Il Convivio. A cura di Franca Brambilla Ageno. Firenze: Le
Alighieri, Dante. 1989. The Banquet. Translatedwith an Introduction and Notes by
Christopher Ryan. Saratoga: Anma Libri.
Alighieri, Dante. 1975. La Commedia secondo lantica vulgata. Testo critico stabilito da
Giorgio Petrocchi per ledizione nazionale della Societ Dantesca Italiana. Torino:
Alighieri, Dante. 1985. La Divina Commedia. A cura di Tommaso Di Salvo. 3 vols.
Bologna: Zanichelli.
Alighieri, Dante. 1997. Commedia. Con il commento di Anna Maria Chiavacci Leonardi.
Milano: Einaudi.
Alighieri, Dante. 1993. The Divine Comedy. Translatedby C. H. Sisson. With an Introduc-
tion andNotes by DavidH. Giggins. Oxford: OxfordUniversity Press.
2. Studies
Boyde, P. 1984. Luomo nel cosmo. Filosoa della natura e poesia in Dante. Bologna: il
Boyde, P. and V. Russo (eds). 1995. Dante e la scienza. Ravenna: Longo.
Busard, H. L. L. 1997. ber die Entwicklung der Mathematik in Westeuropa zwischen
1100 und1500. NTM (= Schriftenreihe fr Geschichte der Naturwissenschaften,
Technik und Medizin): 211235.
DAmore, B. 1993. Alcuni cenni sulla presenza della matematica nella Divina Comme-
dia. Cultura e Scuola 32: 145161.
DAmore, B. 1995. Probabilit, logica formale, geometria. Contributi allesegesi di
alcuni passi della Commedia. In Boyde and Russo (eds), 91108.
Delahaye, J.-P. 1997. Le fascinant nombre p. Paris: Belin.
Di Scipio, G. andA. Scaglione (eds). 1988. The Divine Comedy and the Encyclopedia
of Arts and Sciences. Acta of the International Dante Symposion. 1316 November
1983, Amsterdam and Philadelphia: John Benjamins.
Fischer, O. andM. Nnny (eds). 2001. The Motivated Sign. Iconicity in Language and
Literature 2. Amsterdam and Philadelphia: John Benjamins.
Gericke, H. 1992. Mathematik im Abendland. Von den rmischen Feldmessern bis zu
Descartes. Wiesbaden: Fourier.
Gericke, H. andK. Vogel (eds). 1965. De Thiende von Simon Stevin. Das erste Lehrbuch
der Dezimalbruchrechnung nach der hollndischen und der franzsischen Ausgabe
von 1585. Frankfurt: Akademische Verlagsgesellschaft.
Gorni, G. 1990. Lettera nome numero. Lordine delle cose in Dante. Bologna: il Mulino.
Hardt, M. 1973. Die Zahl in der Divina Commedia. Frankfurt: Athenum.
Hardt, M. 1988. Dante andarithmetic. In Di Scipio andScaglione (eds), 8194.
Hardt, M. 1989. I numeri nella poetica di Dante. Studi danteschi 61: 127.
Hardt, M. 1995. I numeri e le scritture crittograche nella Divina Commedia. In Boyde
andRusso (eds), 7190.
Hardt, M. 2000. Beatrice und die Vita Nuova. Italienisch 44: 214.
Hart, T. E. 1988. Geometric metaphor andproportional design in Dantes Commedia.
In Di Scipio andScaglione (eds), 95146.
Hart, T. E. 1995. Per misurar lo cerchio (Par. XXXIII 134) andArchimed es De
mensura circuli: Some thoughts on approximations to the value of . In Boyde and
Russo (eds), 265335.
Hausmann, F.-R. 1988. Dantes Kosmographie Jerusalem als Nabel der Welt.
Deutsches Dante-Jahrbuch 63: 746.
Istituto della Enciclopedia Italiana. 19701984. Enciclopedia Dantesca. 5 vols. +
Appendix, Roma 197078 (Appendix,
Maracchia, S. 1979. Dante e la matematica. LAlighieri 20: 5767.
Marchese, A. 1993. Guida alla Divina Commedia. 3 vols. Torino: SEI.
Mazzamuto, P. 1970. Cinquecento diece e cinque. Enciclopedia Dantesca, II: 1014.
Mazzoni, F. 1995. Dante misuratore di mondi. In Boyde and Russo (eds), 2553.
Moleta, V. (ed.). 1994. La gloriosa donna della mente. A Commentary on the Vita
Nuova. Firenze: Olschky andPerth: The University of W. Australia.
Nnny, M. andO. Fischer (eds). 1999. Form Miming Meaning. Iconicity in Language and
Literature. Amsterdam and Philadelphia: John Benjamins.
Ohly, F. 1982. Deus Geometra. Skizzen zur Geschichte einer Vorstellung von Gott. In
Tradition als historische Kraft. Festschrift fr Karl Hauck, N. Kamp andJ. Wollasch
(eds), 142. Berlin: de Gruyter.
Palma di Cesnola, M. 1995. Semiotica Dantesca. Profetismo e diacronia. Ravenna: Longo.
Picchio Simonelli, M. (ed.). 1994. Beatrice nellopera di Dante e nella memoria europea
12901990. Atti del Convegno Internazionale. 1014 dicembre 1990. Napoli:
Ptters, W. 1987. Chi era Laura? Strutture linguistiche e matematiche nel Canzoniere
di Francesco Petrarca. Bologna: il Mulino.
Ptters, W. 1997. Un cinquecento diece e cinque (Dante, Purg. 33, 43). Der Name des
Boten Gottes als Problem der Wortbildung und der poetischen Kosmologie. In
Latinitas et Romanitas. Festschrift fr Hans Dieter Bork, A. Bolle andJ. Kramer
(eds), 351376. Bonn: Hillen.
Ptters, W. 1998. Nascita del sonetto. Metrica e matematica al tempo di Federico II.
Ravenna: Longo.
Ptters, W. 2000. Il buon Gherardo (Purg. 16, 115145). irc1c: b:N1is cro-
xr1ic:. In Les Mots de la Tribu. Festschrift fr Gerhard Goebel, T. Amos e.a.
(eds), 4772. Tbingen: Stauenburg.
Ptters, W. 2001. Ella era uno nove. Rapporti geometrici fra la Vita Nuova e la Com-
media. LIA 2: 2760.
Ptters, W. 2002. la spera che pi larga gira. Spazio della poesia e disegno del
cosmo. irc1c: b:N1is croxr1ic: ii. LIA 3: 461505.
Tropfke, J.
1940. Geschichte der Elementarmathematik. IV. Geometrie. Leipzig: de
Vecce, C. 1994. Ella era uno nove, cio uno miracolo (V. N., XXIX, 3): Il numero di
Beatrice. In Moleta (ed.), 161179.
Viti, G. 196364. Guida alla Divina Commedia. Con tavole esplicative. 3 vls. Firenze: Le
Wedeniwski, S. and C. Haenel. 1996. Neue Runde. Die Kreiszahl undihre Berech-
nung. Ct 13: 348352.
Wehle, W. 1998. Dichtung ber Dichtung. Dantes Vita Nuova: die Aufhebung des Minne-
sangs im Epos. Mnchen: Fink.
</TARGET "pot">
AUTHOR "Masako K. Hiraga"
TITLE "How metaphor andiconicity are entwinedin poetry"
WIDTH "150"
How metaphor and iconicity are entwined in poetry
A case in haiku
Masako K. Hiraga
Graduate School of Intercultural Communication, Rikkyo University, Japan
there is clearly an iconic element in metaphor.
Paul Henle, Metaphor (1958: 177)
1. Introduction
The major problem in the treatment of metaphor andiconicity in the past
literature in semiotics andlinguistics is that metaphor andiconicity have not been
studied in a fully related manner within the same frame of reference. Although
Peirce (19311958, 1962 [1955, 1902]) dened metaphor as a subtype of iconic
signs along with image andd iagram, metaphor has been treatedquite separately
from image andd iagram in the major literature.
By contrast, this study emphasizes the importance of treating both in the
same framework because metaphor andiconicity are essentially entwinedas an
analogical mapping. On the one hand, metaphor entails iconicity. Imagic and
diagrammatic mapping of image-schematic structures resides in the creation and
interpretation of meaning in metaphor. On the other hand, iconicity entails
metaphor. The metaphors underlying the relationship of form and meaning
navigate the diagrammatic interpretation of linguistic forms. Likewise, a meta-
phorical reading of the text reveals the iconic structure in the text. In brief, two
types of the interplay between metaphor andiconicity are analysedandexplained
in this paper by the model of blending (Turner 1996, Fauconnier and Turner
1998, among others): (1) iconic moments manifestedas image-schemas in
metaphor; and(2) metaphor giving iconic interpretation to form.
The study presents, as an illustrative and foregrounded manifestation of the
interplay of metaphor andiconicity , a detailedanalysis of haiku, Japanese poetry
of 17-mora text form. With their brevity in form andrichness in meaning, haiku
are considered to oer an optimal example of how the human mindconnects
things with minimal linguistic resources but in a particularly subtle way (cf.
Blasko andMerski 1999). The analysis focuses on the following haiku text:
araumi ya Sado ni yokotou ama no gawa
rough sea: Sado in lie heaven of river
Rough sea: lying towardSad o Islandthe River of Heaven (Matsuo
1996 [1694]: 109).
The choice of this particular text was basedon three factors: (i) it displays a
metaphorical juxtaposition, in which two parts of the text are put in comparison
andcontribute to enriching the multi-layeredmeaning; (ii) the well-documented
revision of the text demonstrates that kanji (i.e. Chinese logographs) play a
cognitive role to strengthen the link between form andmeaning in the nished
text; and(iii) representations of the text in Japanese Sign Language are available
for further analysis (cf. Herlofsky this volume).
The analysis shows (1) the poems grammatical andrhetorical structure, (2)
local blends andrecruitment of backgroundknowled ge, (3) global blendand
emergent meaning, (4) iconicity of kanji as a cognitive medium, and (5) iconicity
of the soundpatterns. In my analysis, I hope to propose that metaphor and
iconicity shouldbe treated as an entwinedprocess, and that the model of
blending oers explanations of the dynamic creativity in the interplay of
metaphor andiconicity .
2. Cognitive approach to metaphor and iconicity
Before analysing the text in detail, this section presents major methodological issues,
which give a frame of reference to the present study. The cognitive approach regards
metaphor andiconicity as important theoretical issues in the exploration of the
operation of the human mind. The vagueness of the Peircean notion of metaphor
is compensatedfor by the denition given by the cognitive theory of metaphor.
2.1 Working denitions
In cognitive terms, iconicity (i.e., images andd iagrams in Peircean terminology)
deals with a mapping between form and meaning (Lako andT urner 1989) in
various degrees of abstraction, from concrete attributive resemblance to abstract
structural analogy. Metaphor, on the other hand, is a mapping between two
conceptual spaces of meaning, a projection of a schematisedpattern from a less
abstract source space onto a more abstract target space (cf. Lako andJohnson
1980, Lako 1987, 1993, Turner 1996).
It is often the case that the basis of similarity in the icons is derived from
visual, auditory and other formal traits of the object that they standfor , andthat
the iconic signs are often the visual, auditory or formal representations them-
selves. Metaphors, on the other hand, do not necessarily have such traits. Rather,
they manifest themselves as a heuristic device to mediate dissimilar concepts by
means of the similarity they yield(Ricoeur 1975). As Anderson (1984: 459)
correctly points out, a metaphor, like an image or an analogy, is what it
represents but not because of an antecedent identity or similarity, not as a
reminiscence, but in virtue of a similarity which it creates. Indeed, in icons, the
similarity relationship between the sign andthe object is taken to be pre-existent.
In contrast, a metaphor connects two entities, which are a priori dissimilar. The
connecting act via a metaphor establishes the similarity between the two
dissimilar entities, and thus creates a new meaning or interpretation.
Metaphor andiconicity often work together. There is an iconic moment in
metaphor. Iconic moments in metaphor manifest themselves as a mapping of
image-schematic structures in the generation of meaning. The term, image-
schema, itself suggests that both image andd iagram are to be relatedto this
cross-space mapping of metaphor. From a vast amount of information provided
by our bodily experiences, interactive perception, feeling, association, contextual
and background knowledge, we have an array of images that do not clash with
the entire cognitive process of metaphor.
In addition, there is a metaphorical navigation in the understanding of icons.
That is, metaphors in the text leadto an iconic interpretation of the form. Such
metaphors guide the way one interprets the forms as diagrammatic icons. This
interplay between metaphor andiconicity is not a static product of structure, but
a dynamic process, emergent, elaborated, and integrated with background
knowledge, contextual information, and feeling, at the time of meaning genera-
tion both in composition andinterpretation.
2.2 The model of blending
In order to explicate the interrelationship of metaphor andiconicity andto
analyse their actual manifestations in poetic language, this study introduces the
model of blending, developed by Turner and Fauconnier (Turner and Fauconnier
1995, Turner 1996, Fauconnier andT urner 1998). The basic claims of this model
can be summarisedas follows:
1.Metaphor is a cognitive process in which one set of concepts (a target) is
understood in terms of another (a source). According to the model of blending,
metaphor is a conceptual integration of four mental spaces.
Mental spaces are
small conceptual arrays constructed for local purposes of understanding. When
a conceptual projection occurs, two input mental spaces (source andtar get in a
metaphor) are created. There is a partial mapping of counterparts between the
input spaces, as representedby solid lines in Figure 1. These input spaces have
relevant information from the respective domains, as well as additional structure
from culture, context, point of view andother backgroundinformation.
2.There are two kinds of middle mental space in addition to the input spaces.
These middle spaces are a generic space and a blended space. The generic space
maps onto each of the input spaces, as indicated by dotted lines in the gure
below. It reects some abstract structure sharedby the inputs. The blended space
also receives partial projections from the inputs, as indicated by dotted lines. The
blendis a rich space integrating the generic structure, structures from each input
space andbackgroundinformation. Often the blendedspace has an emergent
structure of its own, as representedby the square in the gure. This emergence
occurs in ways of composition (in which new relations emerge from projections
from the inputs), completion (in which the composite structure projectedinto the
blendis completedinto the larger system by backgroundknowled ge, cognitive
andcultural models), andelaboration (in which the blendis further elaborated
according to its own logic) (Fauconnier and Turner 1998: 271).
3.Each mental space has an image-schematic structure that is consistent and
preservedthrough a conceptual projection of generic and input spaces. The
image-schemas are skeletal patterns in our sensory andmotor experience, such
as a container, a motion along a path, part andwhole, centre andperiphery ,
symmetry andso forth.
4.The blended space develops inferences, arguments, ideas and emotions, which
can modify the initial input spaces and change our views of the knowledge used
to buildthose input spaces (Turner 1996: 83). Therefore, the blend is a dynamic
mechanism of creativity.
2.3 Metaphor-icon links
How does metaphor relate to iconicity in the blending model? The model can be
elaboratedin the way in which the iconic mapping between form and meaning
is added in order to account for metaphor and iconicity as an entwined process.
Input (Source)
Generic Space
Input (Target)
Figure 1. The model of blending (Fauconnier andT urner 1998: 272)
Figure 2 is a graphic representation of the relationship of metaphor-icon links, in
which iconic mappings are indicated by arrows (cf. Hiraga 2000).
In the meaning box, the metaphorical mapping is illustratedaccord ing to
the model of blending as discussed in the previous section. The form box
represents the linguistic andother semiotic resources such as soundand visual
shapes. When the form directly mirrors the meaning as in the case of onomato-
poeia andvisible language (e.g. some logographs), the imagic mapping occurs in
the direction from form to meaning, as illustrated by an arrow. On the other
hand, when metaphors in the text give meaning to form, there occurs a diagram-
matic mapping in the direction from meaning to form illustrated by an arrow.
There are three types of occurrences of diagrammatic mapping: from the input
space, from the generic space andfrom the blendedspace onto the form. It is
these mappings that give iconic meanings to form at the level of text segments
as well as at the level of the macro-structure of the text.
The eectiveness of the elaborated model of blending shown in Figure 2
shows up here. For it is with this model that one can specify which part(s) of the
metaphorical process whether the input, generic, or blended spaces relate(s)
to the diagrammatic mapping of form and meaning. In theory there are four
possibilities: a mapping (1) from generic space onto form; (2) from input source
space to form; (3) from input target space to form; and(4) from blended space
to form. However, in practice, case (3) is unlikely because the target space in
Visual Structure
Visual Shape
Visual Structure
Metaphorical Mapping Iconic Mapping
Generic Space
Figure 2. Metaphor-icon links
metaphor, by its own nature, is where the mapping nalizes rather than initiates.
It shouldalso be emphasisedhere that the interplay between metaphor and
iconicity is a dynamic, creative process. The model of blending accounts for this
dynamism by way of an emergent structure as represented by the square inside
the blendedspace in Figure 2. Creation andinterpretation of metaphor andicon
is not a xed process which is pre-set prior to the text. Instead, meaning of
metaphor andicon is emergent at the time of the blend, in which all the mental
spaces are integrated, completed, and elaborated into the larger system by
inferences, emotions, backgroundknowled ge, cultural andcognitive models.
3. How a poem manifests metaphor-icon links
Through a detailed analysis of a prototypical example of haiku by Basho Matsuo
(16441694), this section demonstrates that the poetic text chosen indeed
foregrounds metaphor-icon links and thus serves as optimal data to support the
theoretical claims made in this study.
Haiku or hokku, as it was calledd uring the time of Basho, is the shortest
form of Japanese traditional poetry, consisting of seventeen morae,
divided into
three sections of 575. Originating in the rst three lines of the 31-mora tanka,
haiku began to rival the older form in the Edo period (16031867). It was
elevatedto the level of a profoundly serious art form by the great master Basho.
It has since remainedthe most popular poetic form in Japan. Originally, the
subject matter of haiku was restrictedto an objective description of nature
suggestive of one of the seasons, evoking a denite, though unstated, emotional
response. Later, its subject range was broadened but it remained an art of
expression suggesting as much as possible in the fewest possible words. Both
tanka and haiku are composedby people of every class, men andwomen, young
and old. As the Japanese language has only ve vowel sounds, [a], [e], [i], [o],
and[u], with which to form its morae, either by themselves or in combination
with a consonant as in consonant-vowel sequences, it is not possible to achieve
rhyming in the sense of European poetry. Brevity, suggestiveness andellipsis are
the life andsoul of haiku and tanka. The reader is invited to read the unwritten
lines with the help of his/her imagination (for further explanations given in
English about Japanese haiku, see Blyth 1952, Yasuda 1957, Henderson 1958,
andShirane 1998).
The haiku to be lookedat here is taken from Bashos Okuno Hosomichi,
one of the acknowledged masterpieces of Japanese literature:
Example 1
araumi ya Sado ni yokotou ama no gawa
rough sea: Sado in lie heaven of river
Rough sea: lying towardSad o Islandthe River of Heaven
(Matsuo 1996 [1694]: 109).
3.1 Metaphorical juxtaposition
(a) Grammatical andrhetorical structure
The poem at rst glance describes a natural scene. On the one hand, the sea is
rough; and on the other hand, over ones head, there is the Milky Way arching
toward the island of Sado. Even if one does not have much pragmatic knowledge
about Sado Island or the Milky Way in Japanese history and culture, one may
sense the grandness of scale depicted by this haiku. It is a starry night. The
Milky Way is magnicent. The grandeur of the Milky Way is put in contrast to
the dark rough sea. The waves are terrifying; the water churns and moans, as if
it wouldnot allow the boats to cross. It is dangerous andfearful in the night.
This dark sea does indeed separate the people living on the island of Sado
from the mainland. The islandis visible across the troubledwaves, perhaps with
its scatteredhouse-lights. Human beings (including the poet) are very small in
the face of this spectacular pageant of powerful nature. Andyet there are
thousands of human lives and stories embedded in the scene.
The rst ve-mora segment, araumi ya, consists of a noun, araumi (rough
sea), anda kireji (cutting letter), ya. Kireji, a rhetorical device, used in tanka
and haiku, consist of about a dozen particles and mark a division point in the
text. Although the functions of the division vary according to the particles, a
general eect of kireji is to leave room for reection on the feelings or images
evokedby the preceding segment. Ya in Example 1 is a kireji particularly
favouredby Basho andsaid to have something of the eect of a preceding
Lo! It divides a haiku into two parts andis usually followedby a description
or comparison, sometimes an illustration of the feeling evoked. There is always
at least the suggestion of a kindof equation, so that the eect of ya is often best
indicated by a colon (Henderson 1958: 189). That is, araumi (rough sea) and
the rest of the text, Sado ni yokotauama no gawa (the Milky Way, which lies
towardSad o), are juxtaposedto constitute a kindof metaphor in which the
feelings or images evokedby a rough sea are illustratedby the feelings or
images evokedby the Milky Way arching over the Islandof Sado (see Figure 3).
The next seven-mora segment, Sado ni yokotau ([which] lies towardSad o),
is an adjectival clause which modies the last ve-mora segment, ama no gawa
(the river of heaven). Sado is a place name, an islandlocatedabout fty miles
from the coast of mid-Honshu. Ni (toward) is a postpositional particle of
location. Yokotau (to lie) is a verb which normally has an animate agent and
describes the action of spreading ones body on something at (when used as a
transitive verb) or the state of a body spread out on something at (when used
as an intransitive verb). As the grammatical subject of yokotau in this poem is
ama no gawa (the river of heaven), an inanimate noun, the verb is used
The last ve-mora segment, ama no gawa, is a proper noun signifying the
Milky Way. The literal meaning of ama no gawa is the river of heaven. It
involves a metaphor in which the path-shapedset of stars (the Milky Way) is
seen as a river. The second andthe thirdsegments of the poem thus constitute
a local metaphor, in which the river of stars in the heaven spreads its body toward
the Islandof Sado. There are conventional conceptual metaphors behind this
local metaphor, namely, N:1cr is :Nix:1r (in this case ivr is :Nix:1r),
and : i:1n-sn:irb onrc1 is : ivr.
(b) Background knowledge
In addition to this local knowledge about the grammatical and rhetorical struc-
ture, it is indispensable to consider the background knowledge recruited at the
time of reading. The Island of Sado and the Milky Way have rich cultural
implications. Sado Islandhas a long history. The islandis geographically
separatedfrom the mainlandby the Sea of Japan. Because the rough waves
preventedpeople from crossing the sea by boat, the islandfunctionedas a place
of exile for felons andtraitors from the tenth century up to the endof the
nineteenth century. At the same time, goldmines were discoveredthere in the
early seventeenth century, andattracted all kinds of people. At the time of
Basho, the Tokugawa Shogunate hadcontrol of the goldmines, and the people
imprisonedin the islandwere forcedto serve as free labour there.
Thus, the metonymy of a rough sea with Sado Island activates the cultural
andhistorical meanings of the island. Also, the roughness of the waves is
consonant with the roughness of life on the island, which involves violence,
cruelty, andd espair. Another important point is that the name of this islandis
written in two Chinese logographs, and, which mean to help and to
cross respectively. The cognitive meanings of the logographs, particularly that
of crossing, seem to be mappedonto the image of a rough sea. One can
probably detect, in the alignment of these two image-schemas (rough sea and
island), workings of such salient conventional metaphors as iirr is : no:1
ocNrx and 1nr v:vrs :r :N ons1:cir 1o sccn : ocNrx. The diculty
of crossing is highlightedhere, andfurther reinforces the sadfeelings relating to
the diculty of reunion by separatedpeople, one group living on the island, the
other living on the mainlandacross the rough sea.
Moreover, there is a sadlegend about the Milky Way, which originatedin
China and was brought to Japan. The date on which this poem was composed,
the night before the seventh night of the seventh month of the lunar calendar,
suggests that the poet hadthis legend in his mind. For the seventh night of the
seventh month (i.e., the 7th of July) is known andcelebrated as the star festival
after the Chinese story. The two bright stars on either side of the Milky Way, the
star Vega andthe star Altair, are believedto be Princess Weaver andOxherd .
These two stars face each other across the Milky Way; but, because the Milky
Way is so wide and vast they cannot meet easily. One day the god of heaven
pitiedPrincess Weavers lonely life andarrangedfor her to marry Oxherd. After
they married, the Princess became too lazy to weave. The angry god punished
her and allowedher to visit her husbandonly once a year, the night of the 7th of
July, but only if the night was fair.
The separation of this legendary couple is mapped onto the separation of the
people imprisonedin Sado Island. Both the Milky way andthe Islandof Sado
with their cultural connotations share event frames for connement spatial
connement, limitedfreed om, limited means of travel, andthe mental state of
being separated.
In sum, this haiku text can be seen as a global metaphor in which several
images are comparedandpreserved to produce multi-layeredmeanings in the
blend. Most prototypically, such images include a rough sea and the river of
heaven (Milky Way) as an obstacle, a separating ow of water, andthe legend-
ary couple andthe prisoners as people prevented from reuniting with their loved
ones. They constitute the input spaces to be mappedonto one another, and
integratedinto the blendedspace, as illustratedin Figure 3. Furthermore, there
are other implicit meanings in the text. The text may evoke, for example, a
feeling of elegy or a realisation of the helplessness or nothingness of human
beings in the face of powerful nature, in this case, representedby terrifying
rough waves andvast starry skies. It may also imply varying kinds of contrast:
a contrast of motion between the violent waves andthe peaceful skies; a contrast
of colour andlight between the black andd ark sea andthe silvery andbright skies;
anda contrast of the real andthe legendary between life stories of people and
the love story of stars. These impliedimages and meanings can all emerge at the
time of reading of the text, as represented in the blended space in Figure 3.
3.2 Iconic interpretation navigated by metaphor
With the metaphorical reading analysedabove, the visual andthe auditory
conguration of the text shows up its iconic eects.
(a) Visual iconicity
Firstly, let us look at the visual elements. The Japanese language has a unique
writing system in which three dierent types of signs are usedto describe the
same phonological text: kanji (Chinese logographs), hiragana (syllabary for
words of Japanese origin), and katakana (syllabary for words of foreign origin
other than Chinese). In the context of the present discussion, logographs are of
particular importance because they function as a cognitive medium for poetry.
Basho revisedthis poem orthographically from Example 2 (a) to 2 (b) (Matsuo
1957 [1694]).
Example 2
Generic Space
Fusion of contrasts
Araumi ya Sado ni yokotou ama no gawa
Path of water
Milky Way
Weaver & Oxherd
Rough Sea
Rough waves
Gold mines
Weaver & Oxherd
Rough waves
Figure 3. Blending of Araumiya
araumi ya Sado ni yokotou ama no gawa
The poems three noun phrases, araumi, Sado and ama no gawa, were all spelled
in kanji in both the rst [Example 2 (a)] andthe revised [Example 2 (b)]
versions. The boxedpart, the verb of lying, was revisedfrom kanji, a Chinese
logograph, to hiragana, two moraic letters. The main eect of changing the
character type in the verb yokotou (to lie) from kanji to hiragana is to
make that part of the text a groundfor the conspicuous prole of (rough
sea) and(milky way). In general, because kanji, being logographic
characters, have a distinct angular form and semantic integrity, they dierentiate
themselves visually andcognitively as the gure while the remaining hiragana
function as the ground.
This dierentiation of kanji is particularly prominent in the words
(rough sea), (Islandof Sado), and (milky way). These three
nouns are all written in two kanji. All of them include kanji (underlined in
Example 2 (a) and(b)) such as (sea), (to cross water), and (river)
which are made up with the same radical signifying water. Both (rough
sea) and(milky way) relate to water, as described above. The semantic
similarity between (rough sea) and(milky way) in terms of
wateriness andthe obstacle (in the real life andin the legendexplainedabove)
andtheir dissimilarity (violence in the rough sea andpeacefulness in the river
of heaven) are also foregrounded. This is a case of diagrammatic iconic eect,
intensifying the meaning of the foregrounded elements by the repetitive use of
similar visual elements two-character nouns andthe same radical.
Besides, (to cross water) in Sado the name of the Island, seems
important, because this logograph means to cross. As the backgroundhistory
andthe legendshow , both rough sea and the Milky Way are obstacles for the
lovedones crossing for their meeting. This character is placedin the middle of
the poem as if it signalledthe crossing.
(b) Auditory iconicity
The soundstructure also exhibits interesting iconic eects navigatedby the
metaphorical interpretations. The following analysis illustrates three possible
iconic eects produced by the distribution of vowels, consonants, and the
repetition of adjacent vowels. Example 3 is a phonological notation of the
poems moraic structure.
Example 3
Line 1 a-ra-u-mi ya ([] = the division between mora)
Line 2 sa-do ni yo-ko-to-u
Line 3 a-ma no ]a-wa
The distribution of the vowels shows that the poem predominantly uses back
vowels such as [a] and[o]. As indicatedin Table 1, there are 8 [a] s (47%) and
5 [o] s (29%) out of a total of 17 vowels.
When one compares this distribution to an overall distribution of the ve
Table 1. Distribution of vowels
a o i u e Total
Line 1
Line 2
Line 3
vowels in the fty haiku poems by Basho in Okuno Hosomichi in Table 2, it is
clear that this haiku is characterizedby the high occurrence of [a] and[o] and
the low occurrence of [i].
[a] and[o] are pronouncedwith a wide passage between the tongue andthe
Table 2. Occurrence of vowels in the 50 Haiku in Oku no Hosomichi
a o i u e Total
roof of the mouth, andwith the back of the tongue higher than the front. These
two vowels have characteristics, which are contrastive to the vowel [i], a front
vowel, pronouncedwith a narrow opening of the mouth with minimum energy.
The backness andthe openness often create sonorous eects, which may draw
associations of something deep and large (cf. Jespersen 1964 [1921]). In this
poem, perhaps, these eects have something to do with the largeness of waves
in the rough sea andthe depth andwid th of the river of heaven.
The sonorous eects are also createdby the frequent use of nasals ([m],
[n], and[ ]]), andvowel-like consonants ([y] and[w]). Table 3 shows the
distribution of consonants.
The dominance of sonorants such as [m], [n], []], [r], [y], and[w] is
characteristic of the text. The sonorants often provide prolongation and fullness
of the sounds, and hence usually produce lingering eects (cf. Shapiro and
Table 3. Distribution of consonants
Position of Mora in Line # of
# of
1 2 3 4 5 6 7
Line 1
Line 2
Line 3
Total 9 4
Beum 1965: 1012). It couldbe arguedthat the back vowels andsonorant
consonants jointly reinforce a sound-iconic eect of the depth or the large-
ness of the image of water elements, i.e., a rough sea andthe river of heaven
expressedby the poem. Also note that the only line that has obstruents (i.e., non-
sonorants such as [s], [d], [k], and[t]), is Line 2, in which the islandis men-
tioned. If one can interpret sonorants as iconically associated with water
elements, then one can also infer that obstruents are associatedwith non-
water, namely, the islandin this text.
The last point is that the text seems to conceal very cleverly andwittingly
a key word, which is congruous with the metaphorical interpretation of the poem.
The prototypical soundsequence in Japanese is an alternation of a single
consonant anda single vowel such as CV-CV-CV. This general feature applies
to the haiku text, too.
Example 4
Line 1 a-ra-u-mi ya
Line 2 sa-do ni yo-ko-to-u
Line 3 a-ma no ]a-wa
A closer look, however, enables one to recognise that there are a few occurrenc-
es of two vowels (or semi-vowel) adjacent to each other such as [+Back] and
[Back], manifestedas [a-u], [o-u], and [a-w] as underlinedin Example 4. In
Line 3, there is a similar soundsequence, g a w a, as the Japanese semi-vowel
[w] is phonetically close to [u]. It couldbe said that each line of the poem has
a vowel sequence, [+Back] and[Back], which is poetically equivalent to [a-u],
hidden in the sound sequence of a word or two adjacent words. Very interesting-
ly, this vowel sequence, [a-u], is a verb in Japanese, which means to meet. The
hidden repetition of [a-u] in each line could be read as an echo of a hidden
longing between separatedpeople. Again, the iconic eect of this hidden
element supports the reading of the text as a global metaphorical juxtaposition,
i.e., separation of the two stars on either side of the Milky Way mapped onto the
people in the Islandof Sado separatedfrom their lovedones on the mainland.
3.3 Metaphor-icon links in haiku
It has been claimedthat the cognitive projection of metaphorical juxtaposition as
a result of the kireji (cutting letters) is to be explainedas a global blendwhich
integrates the input mental spaces of Milky Way andrough sea, which are at
the same time locally blendedspaces. This blendor conceptual integration occurs
as a dynamic process of making sense over the entire array of many mental
spaces under the recruitment from cultural and historical knowledge and other
backgroundcontexts, andthus creates emergent structures. The process of
integration is graphically representedin the meaning box in Figure 4.
The arrows connecting form andmeaning in Figure 4 illustrate how iconic
mappings occur. Firstly, the water radical repeatedly used in the three kanji
logographs is a mimetic image icon of three drops of water, and thus, is mapped
onto the input spaces of ama no gawa (the river of heaven or Milky Way) and
araumi (rough sea). This is representedby the two arrows of imagic mapping
in Figure 4. Secondly, there is a diagrammatic mapping between the generic
space (path of water as obstacles) andthe water radical in kanji. The mapping is
diagrammatic because the strength of water elements in the generic space is
mappedonto the repetitive structure in the form when one interprets an iconic
meaning of the repetition. Thirdly, when the blend forges emergent meanings for
the overall metaphorical interpretation of the text, the metaphorical reading
navigates diagrammatically some iconic interpretations of sound congurations.
The soundarchitecture of the text is now readas an analogical icon of the
semantic structure of the poem. This is illustratedby the arrows of diagrammat-
ic mapping from the blended space onto the form in Figure 4.
Workings of the interplay of metaphor andiconicity demonstratedabove by
the detailedtextual analysis andthe graphic representation are prevalent in
various texts of poetic andnon-poetic nature if not as prominent as in the one
discussed above (cf. Hiraga 2000). The point is that far from being subordinated
to meaning, the visual andsound shape of the text plays a crucial role in creating
andreinforcing meaning.
4. Conclusion
Milky Way
Generic Space
Rough Sea
Metaphorical Mapping
Iconic Mapping
(Water Radical)
[a-u]('to meet')
Figure 4. Metaphor-icon links in Araumiya
Interpretations of the literary text are constrainedin certain ways by the use
of conventional conceptual mapping, by commonplace knowledge and by
iconicity between structure andthe meaning. The analysis has demonstratedthat
the reading of haiku is also dependent on these factors. Basho used conceptual
metaphors, andexploitedalmost every possible resource in the lexicon, syntax,
andorthography to multiply the implications of the short poetic text, e.g., kireji
(cutting letters), kanji (Chinese logographs), allusions, andsoundpatterns. It
is indispensable to rely also on cultural and historical background knowledge to
understandthe enrichedmeanings of his texts. Finally, iconicity is of particular
importance in a short poetic text such as haiku because brevity seems to require
the form itself to participate in giving images, concepts, andfeelings. This has
been demonstrated by Bashos clever use of kanji andsoundstructure in visual,
auditory and cognitive terms.
The major methodological contribution of this study is that I have added to the
model of blending proposed by Turner and Fauconnier a few specications for
explicating the workings of metaphor andiconicity as an entwinedphenomenon.
The Turner-Fauconnier approach is exible andcomprehensive as it
assumes that meaning creation is an ongoing mental operation with an emergent
structure in the blendof input mental spaces. The model can therefore incorpo-
rate structures not only from the expressedentities (i.e., input spaces) but also
from the unexpressed(e.g., pragmatic contexts, backgroundprior knowledge,
cognitive models, inferences, and emotions). This makes it a very realistic and
eective instrument for literary analysis.
With regardto the relationship between form andmeaning in the creation
and interpretation of the text, this study has tried to dene, more explicitly than
the Turner-Fauconnier model, metaphorical and iconic mappings, which occur at
the time of blending. It has been shown through the sample analysis of haiku that
this elaboratedmod el can specify which parts of the metaphorical process,
whether the input, generic, or blendedspaces, relate to the imagic andd iagram-
matic mapping of form andmeaning.
Composing and understanding a text is a process of making sense out of the
stream of images, concepts, knowledge, and feelings, emerging together. The
interplay of metaphor andiconicity essentially concerns this dynamic relationship
of form andmeaning in action.
This is an enlargedand revisedversion of my paper, Rough Sea and the Milky Way: Blending in
a Haiku Text. Computation for Metaphors, Analogy andAgents, editedby Chrystopher L. Nehaniv,
2736. Berlin: Springer-Verlag, 1999.
1. Word-for-wordtranslation is given by the author andnot in Matsuo (1996 [1694]). My word-
for-wordtranslation is basedon Matsuo (1966 [1694]) and Matsuo (1996 [1694]). The in-text
reference with dierent years of publication indicates that the year in brackets is a source or
an original work andthe year in parenthesis is an access volume according to which the citation
is made.
2. Turner andFauconnier use the term mental space in contrast to the term conceptual domain,
employedby Lako, Johnson, andother cognitivists. Mental spaces are small conceptual arrays
put together for local purposes of action and understanding, while conceptual domain is a vast
structural array that couldnot be made active in thinking (Turner 1996).
3. A mora is a unit of timing. Each mora takes about the same amount of time to pronounce.
4. Some rivers have human male names such as Bando-Taro (place-male name) for Tone River.
Furthermore, rivers are prototypically metaphorizedas snakes in Japanese idioms, e.g., kawa ga
dakoo-suru (A river snakes), kawa ga hebi no yoo-ni magaru (A river curves like a snake),
Anderson, D. 1984. Peirce on metaphor. Transactions of the Charles S. Peirce Society:
A Quarterly Journal in American Philosophy, Vol. XX, No. 4: 453468.
Blasko, D. G., andD. W. Merski. 1999. Haiku: when goodness entails symbolism.
Metaphor and Symbol 14(2): 123138.
Blyth, R. H. 1952. Haiku. Tokyo: Hokuseido.
Danesi, Marcel. 1995. The iconicity of metaphor. In Syntactic Iconicity and Linguistic
Freezes: The Human Dimension, M. E. Landsberg (ed.), 265283. Berlin: Mouton de
Fauconnier, G., andM. Turner. 1998. Principles of conceptual integration. In Discourse
and Cognition, J. P. Koenig (ed.), 269283. Stanford: CSLI Publications.
Henderson, H. G. 1958. An Introduction to Haiku. NY: Doubleday.
Henle, P. 1958. Language, Thought, and Culture. Ann Arbor: University of Michigan
Hiraga, M. K. 1995. Vision as meaning: Iconicity in Bashos haiku in Oku no Hoso-
michi. Journal of the University of the Air 13: 109125.
Hiraga, M. K. 1999. Blending and an interpretation of haiku: A cognitive approach.
Poetics Today 20: 461481.
Hiraga, M. K. 2000. The interplay of metaphor andiconicity: A cognitive approach.
Unpublished Ph.D. dissertation. University of London.
Jespersen, O. 1964 [1921]. Soundsymbolism. In Language: Its Nature, Development
and Origin, 396411. New York: Norton.
Lako, G. 1987. Women, Fire, and Dangerous Things: What Categories Reveal about the
Mind. Chicago: Chicago University Press.
Lako, G. 1993. The contemporary theory of metaphor. In Metaphor and Thought, A.
Ortony (ed.), 202251. Cambridge: Cambridge University Press.
Lako, G., andM. Johnson. 1980. Metaphors We Live By. Chicago: Chicago University
Lako, G., andM. Turner. 1989. More than Cool Reason: A Field Guide to Poetic
Metaphor. Chicago: Chicago University Press.
Matsuo, B. 1957 [1694]. Oku no Hosomichi (Sora Zuikoo Nikki tsuki) [The Narrow Road
to the Deep North and Soras Travel Diary]. Annot. Shoichiro Sugiura. Tokyo:
Iwanami Shoten.
Matsuo, B. 1966 [1694]. The narrow roadto the deep north. In The Narrow Road to the
Deep North and Other Travel Sketches, trans. anded . Nobuyuki Yuasa, 97143.
London: Penguin Books.
Matsuo, B. 1996 [1694]. Bashos Narrow Road. Trans. Hiroaki Sato. Berkeley: Stone-
bridge Press.
Matsuo, B. 1997 [1690]. Basho JihituOku no Hosomichi.[Okuno Hosomichi by Bashos
Handwriting]. Eds. Yozo Ueno and Takejiro Sakurai. Tokyo: Iwanami shoten.
Peirce, C. S. 19311958. Collected Papers of Charles Sanders Peirce. Eds. Charles
Hartshorne andPaul Weiss. Cambridge: HarvardUniversity .
Peirce, C. S. 1962 [1955, 1902]. Logic andsemiotic: Theory of signs. In Philosophical
Writings, J. Buchler (ed.), 98119. New York: Dover.
Ricoeur, P. 1977. The Rule of Metaphor: Multi-Disciplinary Studies of the Creation of
Meaning in Language. Trans. R. Czerny with K. McLaughlin andJ. Costello.
Toronto: University of Toronto Press.
Shirane, H. 1998. Traces of Dreams: Landscape, Cultural Memory, and the Poetry of
Basho. Stanford: Stanford University Press.
Turner, M. 1996. The Literary Mind. Oxford: Oxford University Press.
Turner, M., andG. Fauconnier. 1995. Conceptual integration andformal expression.
Metaphor and Symbolic Activity 10(3): 183204.
Yasuda, K. 1957 The Japanese Haiku. Rutland, Vermont: Charles E. Tuttle Co.
</TARGET "hir">
TITLE "Intermedial iconicity"
WIDTH "150"
P:1 IV
Intermedial iconicity
</TARGET "p4">
AUTHOR "Werner Wolf"
TITLE "Intermedial iconicity in ction"
WIDTH "150"
Intermedial iconicity in ction
Tema con variazioni
Werner Wolf
University of Graz
1. Introduction and theme: The concepts of intermediality
and intermedial iconicity
Among the many denitions of iconicity the title of the rst volume of the series
Iconicity in Language and Literature, Form Miming Meaning (Nnny andFischer
eds. 1999) is perhaps the most concise. In this well-known and at the same
time iconic formula, as in other denitions of iconicity, meaning is usually
employedin the sense of referring to some kindof non-verbal or non-medial,
real-life phenomenon. What is less well known is the fact that in literature form
can also mime other arts and media: in other words that literary iconicity can
contribute to intermediality.
In its broadsense the term intermediality applies to any transgression of
boundaries between conventionally distinct media of communication; such
transgressions can occur either within one text, work or semiotic complex or as
a consequence of relations between dierent works, texts or semiotic complexes.
Yet, there is also a narrower sense of intermediality, which means, as I once
dened it, the participation of more than one medium of expression [or
communication] in the signication of a human artefact (Wolf 1999: 1). It is
with this kindof intracompositional intermediality that I am concernedhere.
Among the forms of intermediality in this narrower sense that involve
iconicity plays a role in only one variant: in the verbal imitation of
features of a non-literary medium or in the creation of analogies to it.
In such
intermedial imitation an implicit reference to a non-literary medium in general,
or to a particular work transmitted through such a medium, is made by shaping
the literary text so that it somehow becomes similar to the other, non-literary
medium. As a consequence, this other medium can, to a certain extent, be
experiencedas an imaginary presence while reading the literary text.
This intermedial iconicity as the characteristic trait of implicit intermedial
reference is my theme, which I will develop, considering various forms and
functions, in three variations dealing with painting, lm and music. All of these
variations focus on nineteenth- andtwentieth-century narrative ction as the
imitating medium.
2. Variation I: Iconicity in pictorialized ction (Hardy,
Under the Greenwood Tree)
The nineteenth-century realist novel has often been saidto be informedby a
remarkable pictorialism. In this context the term pictorialism tends to be used
in a purely metaphoric way (as is unfortunately often the case in critical dis-
course) and denotes an abundance of visual description, which does not imply
any genuine reference to painting (Armstrong 1999: 6).
There are, however,
some noteworthy cases, to which the term can, andind eedmust be appliednon-
metaphorically. Such a case is Thomas Hardys Under the Greenwood Tree
(1872), subtitled A Rural Painting of the Dutch School. The following extract may
help illustrate this. It is taken from the rst part of the novel, in which a country
choir pays its Christmas homage to the new school mistress Fancy Day by
singing some carols andshouting up to her a Christmas wish from under the
school house windows:
When the expectant stillness consequent upon the exclamation hadnearly died
out of them all, an increasing light made itself visible in one of the windows
of the upper oor. It came so close to the blindthat the exact position of the
ame couldbe perceivedfrom the outside. Remaining steady for an instant,
the blindwent upwardfrom before it, revealing to thirty concentrated eyes a 5
young girl framedas a picture by the window architrave, and unconsciously
illuminating her countenance to a vividbrightness by a candle she held in her
left hand, close to her face, her right hand being extended to the side of the
window. She was wrapped in a white robe of some kind, whilst down her
shoulders fell a twining profusion of marvellously rich hair, in a wild disorder 10
which proclaimedit to be only during the invisible hours of the night that such
a condition was discoverable. Her bright eyes were looking into the grey world
outside with an uncertain expression, oscillating between courage and shyness,
which, as she recognizedthe semicircular group of dark forms gathered before
her, transformeditself into pleasant resolution. (Hardy 1872/1985: 62) 15
This passage, which contains the rst appearance of the heroine on the stage of
Hardys ctional world, goes beyond the usual realistic emphasis on visualiza-
tion: it thematizes andat the same time displays a genuinely picture-like
quality, which was already praised by Hardys contemporaries as one of the
achievements of his arts.
In fact, Hardys narrator here seems to follow the
Horatian phrase ut pictura poesis (Horace 23 B. C./1980: 26) literally by trying
to imitate painting as far as this is possible in verbal art. As I will show,
narrative pictorialism thus turns into a genuine pictorialization of ction in
which descriptive form mimes painting as its intermedial referent, in this case
even a special kindof painting: the candle-light portrait, an example of which
can be seen in Figure 1.
The impression that Hardys description strives to imitate painting can be
Figure 1. Godfried Schalcken, 16431706 [Dutch School], Young Girl with Candle.
Florence, Galleria degli Uzi.
corroboratedby his use of two variants of intermedial iconicity. The rst is what
one may call intertextual iconicity with intermedial implications: an iconicity
that consists in the intertextual imitation of characteristic features of a verbal
discourse that in turn points to another medium. In our example this kind of
iconicity appears due to the fact that the concepts Hardy uses in Fancys portrait
do not merely refer to painting (as in l. 5) but create a similarity between his
descriptive discourse and ekphrasis, that is, a discourse employed for evoking
Iconic intermediality is thus introduced indirectly: by means of an
intertextual imitation of a discursive genre which by denition implies an
intermedial reference. The similarity between Hardys description and ekphrasis
on the level of discours thus parallels the similarity, on the level of histoire,
between Fancys appearance anda person portrayedin a painting.
The resulting pictorialization of Hardys text can be observed in various
details: To begin with, one can look at the numerous references to Fancys visual
appearance in terms of shape, darkness and illumination. According to Alastair
Smart (1961: 271), this device was inspired by Rembrandts chiaroscuro. In
addition, the pictorial quality of the description is underlined by Fancy being
virtually framedas [in] a picture by the window architrave (l. 5 f.).
Senn, in a chapter entitledThomas Hardys Frozen Scenes, has rightly
emphasizedthe frequency with which, [i]n the context of Thomas Hardys
pictorialism, the window serves as a frame to a picture-like view (2001: 98).
Especially important is the way in which the description is inserted in the
narrative context. This is done in the manner of a freeze (cf. also Senn
2001: 105), in which narrative time andmovement seem to come to a standstill
for a moment: we can surmise that, due to Fancys appearance at the window,
the original expectant stillness (l. 1 f.) of the group of choristers waiting outside
is re-established, and Fancy herself, whilst her eyes are adapting to the darkness,
can be imaginedto assume a portrait-like immobility.
The freeze is here particularly discernible, since it is framed by two
symmetrical activities: Fancys approaching the window and lifting the blind and
following a short dialogue with the singers her retreat into the house and
her lowering of the blind:
Together went the window quickly andquietly , and the blindstarted down-
wardon its return to its place. Her fair foreheadand eyes vanished; her little
mouth; her neck and shoulders; all of her. Then the spot of candlelight shone
nebulously as before; then it movedaway . (Hardy 1872/1985: 62)
As we can see here, Hardy pictorializes his text not only by intertextual iconicity
andby thematizing a painterly frame, he actually stages a frame through the
descriptive discourse itself: this discourse is focussed on Fancy as a static object
(see ll. 514 of the rst quotation) andframed by two narrative passages which
both centre on dynamic activities: Fancys appearance and disappearance (ll. 15,
and the secondquotation). Thus, Hardys intertextual iconicity is strengthenedby
a secondkind of iconicity: diagrammatic iconicity. In this case it consists in
imitating painting through the formal device of splitting a description into a
central freeze andsurround ing symmetrical movements: Fancys portrait thus
stands apart from its narrative context like an actual frame that separates a
painting from its spatial context.
Diagrammatic iconicity can be foundin Fancys verbal portrait in yet a
further sense. G. E. Lessing, who in his Laokoon proves to be keenly aware of
the dierences between poetry andpainting, hadpraised Homer for bridging
the gap between the two media by the narrativization of description, e.g. by
means of describing an artefact through narrating the process of its creation. This
device not only seems more natural to the temporal art of literature, it also
permits the creation of diagrammatic iconicity. In contrast to this, Hardy employs
a d i erent device, but one which also creates diagrammatic iconicity: to a
certain extent Hardys descriptive discourse and the sequence of the elements
mentionedin it appear to imitate the sequence of perception which would
probably occur in the process of viewing Fancy Days portrait as really painted,
for instance at a vernissage. This diagrammatic iconicity is a variant of experi-
ential iconicity, which I have discussed elsewhere (Wolf 2001) as occurring in
literary descriptions imitating in their structure the probable process of the
(visual) experience of a ctitious world.
This experiential iconicity provides a
means of shaping a descriptive discourse that follows the sequence of perceptual
experience (andis thus phenomenologically motivated, as is the case in our
text) and/or the natural order of its referent (and is then referentially moti-
vated [cf. Cobley 1986: 401405]);
it can generally be considered as a descrip-
tive analogy to the diagrammatic iconicity of the ordo naturalis as a way of
organizing a narrative discourse. In Hardys text the raising of the blind initiating
the description could correspond to the unveiling of the painting, whose identity
as such wouldperhaps rst be realizedby the rst elements mentioned: the
perception of a young girl representedin a frame (l. 5). Then our gaze would
presumably focus on the most important object in the painting that is next
evokedin the text, Fancys countenance (l. 6), before taking in details of her
overall gure: her robe andher hair (ll. 810). After this general scrutiny we
wouldin all likelihood return our gaze to what comes next in the text: Fancys
bright eyes (l. 11) andtheir expression as the most interesting objects. Eventu-
ally, our viewing would be terminated, like the descriptive discourse, by the
coming down of the veil in front of the picture. This iconic imitation of a
probable sequence of perception presupposes, of course, a point of view from the
position of someone present in the diegetic world. And even this internal
perspective is imitated in Hardys description: it resides in the perceptual relation
of Fancys portrait to the choristers.
They gaze at her from an outside position,
which pregures the position that a viewer of her paintedportrait wouldoccupy .
The intermedial iconicity of this passage is not the only example in this
novel. There are in fact a number of similar descriptions, which justify the
subtitle A Rural Painting andgive the novel an intense pictorial quality.
pictorializing his novel to this degree, Hardy obviously helps the readers to
visualize the narrative worldandto intensify their aesthetic illusion.
By doing
this, Hardy participates in a well-known development in nineteenth-century
cultural history which Nancy Armstrong described as the advent of mass
visuality (1999: 9 andpassim).
Yet, the participation in the ensuing pictorial turn (Armstrong 1999: 5) is
not the only function of Hardys intermediality, and it would hardly explain his
intermedial iconic reference to painting, since a general enhancement of the
visual quality of his novel wouldhave suced. In the light of Armstrongs
research one may ask in particular why Hardy evokedpainting andnot photog-
raphy, the visual medium which had become dominant by the publication date of
the novel, 1872: for according to Armstrong (see 1999: esp. 7f., 27f.), the
pictorial turn manifesteditself primarily in both the realist novel andphotogra-
One reason, apart from Hardys well-known familiarity with painting,
could have been the fact that photography retained an important visual disadvan-
tage in comparison to painting: colour prints hadnot yet been invented. Howev-
er, for our night scene with its absence of colours this handicap of contemporary
photography wouldperhaps not have been a major hindrance.
Another reason for choosing painting as an intermedial model might have
been that Hardy here continued an established tradition. For the pictorialization
of ction hadnot begun with nineteenth-century references to photography but
back in the eighteenth century by imitating the model of landscape painting, as
can be seen in Mackenzies The Man of Feeling (1771) andin numerous passages
of Radclies gothic novels.
Certainly of more importance is the partial hostility many nineteenth-century
authors, including Hardy himself, professed towards the undignied copyism
of photographic realism, which precludedpsychological insights andempathy .
In this situation, painting as an acknowledged high art appeared as a viable
alternative reference, since painting expressedthe dominance of the visual just as
well as photography but at the same time conferredaesthetic dignity even on
low subjects. This argument also forms the basis of George Eliots famous
programme of realism in Chapter 17 of Adam Bede. Eliots programme inciden-
tally also refers to the Dutch school as an aesthetic model, which is faithfully
followedin the pictorialization of the novel itself (cf. Mosthaf 2000: ch. 3.2).
Moreover, like photography, painting hadalso contributedto the aforementioned
advent of mass visuality, which resulted in an increasing mediatization of
perception through images (see Armstrong 1999, esp. p. 3).
By imitating painting, Hardy, like other realists, did not, however, intend to
foregroundthis subliminal mediatization, let alone its articiality, as Senn has
claimed(2001: 104). They rather employedthis device in order to authenticate
their ctional worlds by emphasizing their visual quality and by competing with
painting in the imitation of this visuality. The use of iconicity considerably
contributedto this, since iconicity is itself frequently employedas a means of
heightening the mimetic transparency of literary discourse.
However, such an illusion of authenticity is always precarious, since
attempting to imitate reality in ction by imitating painting can easily foreground
the inadequacy of the verbal medium and lead to a self-reexiveness that reduces
the impression of authenticity andrather lays bare the articiality of the text. As
Franziska Mosthaf has recently shown (2000: chaps. 3.1 and3.3.), this is in fact
an alternative function of the pictorialization of ction. It can, for instance, be
observedin Melvilles Moby Dick (1851), andto some extent also in Woolfs To
the Lighthouse (1927).
The imitation of a visual medium in the realist manner,
where pictorial iconicity appears as an armation of an image-basedculture and
is usedfor suppressing the mediacy of perception and literary representation (even
for suggesting the authenticity of narrative objects), is in fact only one option. A
further option is the iconic imitation of visual media as a means of criticizing
their dominance, of self-reexively foregrounding mediacy in general and of
reecting, in particular, on the position of literature in relation to other media.
3. Variation II: Iconicity in lmicized ction (Lodge, Changing Places)
This option is chosen in some postmodernist texts, such as Robert Coovers short
story The Babysitter (1969) or DavidLod ges novel Changing Places (1975).
The medium imitated here is, however, not painting but the medium towards
which Lodge (1977) himself said that Hardy had already striven as a Cinematic
Novelist avant la lettre, namely lm. In fact, the aforementionedd ynamic frame
of Fancys description in Hardys Under the Greenwood Tree couldalso be read
in terms of a quasi lmic intermediality.
In the following passage from the last chapter of Lodges Changing Places,
metactionally entitled Ending, such a lmic reading is not only a possibility
but imperative, although the iconic intermediality here serves a quite dierent,
non-realist function. The scene is a New York hotel room, where the novels
four protagonists have met: Morris, an American professor of literature, andhis
English counterpart Philip, who have changedplaces for six months not only
in academia but also in the beds of their respective wives Dsire and Hilary,
who are also present. The four of them are left with the unresolved conict of
how to handle their cross-marital interrelations, while on the TV screen a protest
march of Californian Students supported by rock bands and topless dancers is
being transmitted:
Cut to:
Interior: blue hotel room afternoon.
PHILIP: [] All Im saying is that there is a generation gap, andI think it
revolves aroundthis public/private thing. Our generation we subscribe to
the old liberal doctrine of the inviolate self. Its the great tradition of realistic
ction, its what novels are all about. [] Well, the novel is dying, and us
with it. [] Those kids (gestures at screen) are living a lm, not a novel.
HILARY: This is all very fascinating, Im sure, but couldwe discuss some-
thing a little more practical? Like what the four of us are going to do in the
immediate future?
DESIREE: Its no use, Hilary. Dont you recognize the soundof men talking?
MORRIS: (To PHILIP) The paradigms of ction are essentially the same
whatever the medium. Words or images, it makes no dierence at the structur-
al level.
PHILIP: (To MORRIS) I dont think thats entirely true. I mean, take the
question of endings.
thats something the novelist cant help giving away [] the tell-tale compres-
sion of the pages.
HILARY andDESIREE begin to listen to what PHILIP is saying, andhe
becomes the focal point of attention.
[] youre aware of the fact that theres only a page or two left in the book
[] But with a lm theres no way of telling, especially nowadays, when
lms are much more loosely structured, much more ambivalent, than they used
to be. Theres no way of telling which frame is going to be the last. The lm
is going along, just as life goes along [] andat any point the director
chooses, without warning, without anything being resolved, or explained, or
woundup, it can just end.
PHILIP shrugs. The camera stops, freezing him in mid-gesture.
(Lodge, 1975/78: 246, 250251)
There is no doubt that this passage is iconic. What could at best be questioned
is the intermedial nature of this iconicity. For the text obviously does not imitate
lm as such but a lm script, andthis, one couldar gue, points more to inter-
textuality than to intermediality. Yet, as also hintedat by the inclusion of a TV
report, the reference is nevertheless clearly to lm, though it is admittedly as
indirect as in the preceding case of Fancys portrait: again intermediality occurs
here through the intertextual imitation of a verbal genre which by denition (like
ekphrasis) implies the transgression of medial boundaries.
In addition to this intertextual iconicity with intermedial implications,
Lodges discourse shows elements of diagrammatic iconicity that also strengthen
the impression of an imitation of lm: it strictly follows the ordo naturalis. The
sequence of the dialogues, gestures etc. of the lm script we read is in fact the
exact counterpart to the sequence of events andhappenings in the lm we seem
to watch.
We must, of course, grant that intermedial iconicity can never really present
the medium referred to. This has been proved to be true of the pictorialization of
ction, andthis also applies here. However, the structure andform of Lodges
discourse its meticulous ordo naturalis suggesting a linear unfolding present,
its absence of external narratorial comments andits reduction of the text to a
mixture of dialogue and shooting directions that consistently use lmic vocabu-
lary seem to be the maximum that ction can do to approach the condition of
lm andto encourage the reader to imagine watching a screen rather than
reading a page of ction.
Or rather that would be the maximum if the content of the text, the meta-
medial discussion of the respective abilities of the lmic and the literary
medium, did not openly draw our attention to precisely the fact that we are
reading a book. While Hardys intermedial iconicity in Under the Greenwood
Tree, in the interest of achieving a realist impression of authenticity, concealed
the ctionality of the text by a persuasive visualization andexclusively aimedat
a similarity with painting, Lodges experimental use of lmic iconicity is part of
a strategy that leads to the laying bare of his novels textuality by also contrast-
ing it with lm. Lodge thus reveals the limits of the medium of ction, thereby
undermining both the impression of authenticity and the aesthetic illusion of his
novel. This brings a metareexive gesture to a brilliant climax, a gesture which
informs Lodges entire novel with its numerous metactional passages and
parodies of styles of novel-writing.
Part of Lodges metamedial strategy of contrasting lm and ction can be
seen in his conscious building up of a tension between the aforementioned
creation of an iconic similarity lm/ction as an eect of discours andthe
thematization of the dierences between these media on the level of histoire in
Philips and Morriss discussion about the media-specic use of end signals:
while the materiality of the lmic medium does indeed not necessarily entail end
signals, the tell-tale compression of the pages of a print medium inevitably
does so, as Philip points out with an intertextual reference to an equally self-
reexive passage in Jane Austens Northanger Abbey addressing the same issue.
As a further element of Lodges metamedial self-reexivity this media-
specicity is not only thematized but also illustrated: in the abrupt ending of the
novel in which the lmicization of Changing Places culminates. Obviously,
Lodge, by this conclusion medias in res, imitates the surprising inconclusiveness
of some lmic open endings after having himself given away the approach of
the ending not only by its metactional discussion but also by the tell-tale
compression of the pages of the very book one is reading.
Lodges unusual ending simultaneously indicates a distance towards the
implications of the traditional realist world-view, in which narrative closure often
indicates certainty of meaning (cf. Pfandl-Buchegger 1993: 305), and a self-ironic
distance towards his own narrative medium. Thus, the question as to what the
entangled couples will do in the immediate future remains as unresolved
the metamedial question of whether the novel is really dying and whether lm
is in fact the better medium to capture the inconclusiveness of life. On the one
hand Lodges text seems to suggest that the novel must imitate lm in order to
capture this aspect of the postmodern world-view (and necessarily will fail in
doing so), on the other hand the text is a novel itself andintelligently displays
the exibility of this medium. This medium, after all, still seems to be alive and
kicking last but not least owing to its ability of intermedial iconicity.
The very lmicization of Changing Places thus embodies a crucial ambiva-
lence which is typical of the muted postmodernism of Lodges and indeed of
several contemporary British authors (Broich 1993): while pointing to the limits
of the novelistic medium and its potential inferiority in comparison with todays
dominant medium lm, Changing Places, notably in its intermedial ending, also
celebrates the novel, even though the novel has become problematic.
4. Variation III: Iconicity in musicalized ction (Huston,
Les Variations Goldberg)
All the variations of intermedial iconicity in ction discussed so far extend only
over parts of novels. In the case of the pictorialization of ction such a restriction is
to be expected, since the static quality of painterly representation can hardly
inform an entire verbal narrative with its characteristic unfolding of a dynamic
ctional world. In my example of a lmicized ction the restriction to a single
chapter was, however, only accidental: given the medial parallels between lm
and ction especially in their common temporality, the lmicization of entire
novels is a denite possibility and has in fact already been attempted, for
instance by Dos Passos in his U. S. A. Trilogy (193036) (see Larsson 1980) or
in William S. Burroughss The Last Words of Dutch Schultz (1969), programmati-
cally subtitled A Fiction in the Form of a Film Script.
A similar extension of
intermediality over whole texts is possible in the imitation of an equally temporal
art, music. It will be exploredin the ensuing thirdvariation of intermedial
iconicity: in the musicalization of ction.
The iconic imitation of music in ction (andind eed in literature in general)
can take several forms which I have systematically discussed elsewhere (Wolf
1999: ch. 4.4). In the following, I will, however, concentrate on only one,
relatively frequent variant of musicalization (which, as for instance Hofstadter
1979/80 shows, is not restrictedto ctional literature).
It is basedonce again
on diagrammatic iconicity and consists in the suggestion of structural analogies
to musical forms through the shape and development of the verbal discourse. In
contrast to the two preceding examples from Hardy and Lodge, in which much
of the intermedial iconicity was indirect and used the detour of an intertextual
imitation of an intermedial verbal genre, we can here see an instance of direct
intermedial iconicity. My example of the exploitation of such formal intermedial
analogies in ction is the rst novel of the bilingual Canadian Nancy Huston,
Les Variations Goldberg (1981), a remarkable work, which she translatedherself
into English in 1996 as The Goldberg Variations andfor which she won a
Canadian Governor Generals Literary Award in 1997.
The overall diagrammatic iconicity in this experimental novel is not
dicult to recognize for anyone who knows that J. S. Bachs so-called Gold-
berg-Variationen (BWV 988, publishedaround 1741/42)
consist of an aria,
followedby thirty variatio[nes] anda nal repetition of the initial aria. This is
precisely the structure of Hustons novel, as indicated in the table of contents,
entitledDistribution (1981/94: 251). Her text is composedof thirty-two interior
monologues. With the exception of the framing chapters, which are both attributed
to the central character, middle-aged harpsichordist Liliane Kullainn, all the
variationes are dedicated to thirty people whom Liliane once loved or still loves and
whom she has invitedfor her one-and-a half hour harpsichordrecital of Bachs
Goldberg-Variationen. The recital, which extends over the entire story, takes
place in her and her husbands Paris house on midsummers night in the year
Apart from this structural analogy to the variation form, an additional
analogy is equally obvious: the correspondence between the cyclicity of Bachs
framing arias andHuston s framing chapters, which not only refer to the life of
the same character (Liliane) but centre on virtually the same issues: reections
on Bach, her instrument, her performance, her audience, and on time.
Still further analogies are less evident; nevertheless, they can be detected on
taking a closer look. Among these analogies one can discern some correspon-
dences between certain variations in Bachs Goldberg-Variationen andthe
correlatedchapters in Hustons novel. This applies, for instance, to the relation
between the melancholy variatio xxv in G minor andthe pervading loneliness
and dejection of the correspondent voice in Hustons text, the voice of an ageing
lesbian admirer of Lilianes who has lost contact with her.
A further detail among the structural analogies is the imitation, in most
chapters, of the bipartite internal structure of the aria andvariations in Bachs
music, which consists of twice sixteen repeatedmeasures (the repetitions, which
are sometimes omittedin performances and also in part in Lilianes recital, are
however, not imitatedin Hustons text, since a literary text wouldhard ly support
systematic verbatim repetitions throughout an entire novel).
Of particular importance for the structural correspondence between Bachs
music andHuston s text is the question of whether there is a unifying theme in
Hustons variations. In Bachs work such a theme is not easy to detect for all
those that are accustomedonly to the classic variations of soprano melodies
which aboundin musical history. However, Die Goldberg-Variationen follows
the pattern of a tema con variazioni nevertheless: as has already been remarked
by Albert Schweitzer (1908/72: 282), Bachs composition is not basedon a
variedsoprano melody but rather on a ciacconna-like, more or less constant
basso continuo with a recurrent harmonic pattern,
of which the rst eight
measures are reproduced in Figure 2.
In Les Variations Goldberg there are many recurrent andpartly interrelated
Figure 2. J. S. Bach, Goldberg-Variationen, bass of Aria, mm. 18.
literary motives, such as psychological disturbances and dierent forms of love
(in accordance with the subtitle of the novel Romance), political, feminist,
meta-aesthetic, metamusical andmetatextual issues (among which the relation
between words and music occupies a prominent place); also motives centred on
the body such as sex and food; and nally forms of the appearance of time.
Among all these motives, however, none occurs in such a systematic way as to
suciently become a unifying theme andto establish a correlative to Bachs
basso continuo except for the theme of time. Time, in fact, appears in all
variations in all possible variants: from the dedication to the deceased Roland
Barthes, who usedto be Hustons teacher (Pour celui qui est mort comme un
enfant), to the two epigraphs (Vous avez exactement quatre-vingt-seize
minutes andVous avez tout votre temps), from the rst lines of the main text
(Maintenant cest commenc, [] un temps sest dclench []13) to the last
line (cest la n maintenant [250]). Time occurs as a nite period(as a day on
the calendar, as the duration of a performance, as history; as the duration of life,
as ageing, as beginning or birth, and ending or death); and time also appears in
the form of an actual or wished-for negation of such niteness (in the numerous
references to cyclicity andto possibilities of experiencing a timeless present).
Les Variations Goldberg thus appears as a sequence of variations on the
theme of time in human experience. Together with the other formal elements
mentioned, the fact that this novel aims at forming, on the level of structural
analogies, an iconic correlative to a particular musical composition andthus
qualies as what Steven Paul Scher (1968) has calledverbal music
is easy
enough to recognize. Yet what is the idea, what are the functions of this
extensive musicalization of ction?
To begin with, the musicalization of Hustons novel is evidently meant as
a homage to Bach andhis music. Although this music was supposedly created
for a specic occasion, it is celebratedas being timeless as early as in the rst
aria (see Huston 1981/94: 18). However, there are also dissenting voices which
criticize Bachs music as out-moded, elitist and part of a bourgeois cultural rite.
One couldregardthis limitation of the celebration of Bachs music as an
equivalent to the limitation of the very iconicity informing Les Varations
A secondmotivation for Hustons musicalization is the exploitation of a
function which this intermedial device can often be seen to fulll in experimental
novels: the musical form allows experimental authors to abandon traditional ways
of storytelling while retaining aesthetic form andcoherence. In fact, Hustons
novel is most unusual, since there are practically no outer events on the diegetic
level except for the harpsichordrecital. The text rather meanders through
dierent stages of the pasts of various characters without coalescing to one
coherent story. As for teleology, a sine qua non of all traditional storytelling, it
is only marginally present in this mainly cyclical text,
namely in the fact that
Liliane seems to have learnt or discovered something about both Bachs music,
which she appreciates in a more profoundway at the end, andher audience. The
content of this lesson, the feeling of interconnectedness between herself as
interpreter, Bach and her audience (and hence the realization: il ny a pas de
moi qui soit que moi 247), can be linkedto another function of the musicali-
zation. Both the experimental departure from traditional storytelling and the
musicalization as its enabling device obviously aim at exploring the complexities
of that human faculty to which music is often saidto have a particular anity:
the unconscious. Hence the novels form of a series of interior monologues.
What strikes the reader in this context is the need to readjust a previous impres-
sion at the end: up to the last variation the dierent monologues seem to issue
directly from the consiousness of the respective characters listed in the table of
contents. Yet in the repetition of the aria in the closing chapter the suggestion is
made that all of these monologues in fact issue from Lilianes consciousness.
Liliane thus appears as the central controlling agency, not only as the interpreter
andre-creator of her belovedmusic, but also as an interpreter and (re-)creator of
the minds of those whom she loves or loved, and in this controlling creative
position she resembles the admired composer of the music she performs: Bach.
Through her performance she has discovered and brought to life both Bach
andthe thirty characters listening to her recital. In this respect the musicalization
of ction in Hustons novel fullls yet another function: it creates an equivalent
to Lilianes satisfying discovery of music and her friends, a discovery which she
is able to make through the torture, as she says (18), imposedupon her by the
interpretation of dicult, complex music. The equivalent of this is the satisfying
discovery of literary complexity by Hustons reader who undergoes the
tortures of reading and coming to terms with an extraordinarily complex and
dicult literary work, to whose diculty the musicalization itself contributes
to some extent. In both cases this discovery creates a proximity to the work and
yields pleasure, be it of a sensual or an intellectual kind. This pleasure, it is true,
has more of a sensual nature on the diegetic level: the reections triggered by
music in the novel frequently centre on sensual pleasures of the body and thus
seem to be an echo of the link which Barthes establishedbetween music and the
body ( [] dans la musique [] le rfrent [] cest le corps. [1982: 273]).
In contrast to this, the readers pleasure will presumably be of a more intellectual
kind. Yet in both cases the imagination, invoked in a reference to A Midsummer
Nights Dream (Songe dune nuit dt 249) is involved.
Part of the intellectual dimension of the musicalization of Les Variations
Goldberg forms an additional function of the musical iconicity. As has repeatedly
been remarked andas couldbe seen in Lodges Changing Places, intermediality
is frequently usedas a means of triggering meta-aesthetic or metamedial
reections (see Huber 1992, andLagerroth 1999). Hustons novel is no exception
here. In the course of the recital a number of characters, including Liliane
herself, are shown to reect on music andon its relation to words.
In the comparison of these media, characteristically, music seems to be
given a privilegedplace, which also sheds light on a nal function of Hustons
musicalization: music appears to create a synthesis between sensuality andthe
intellect, its beauty permits a suspension of the painful awareness of the negativi-
ty, the limitations andthe transitoriness of reality by immersing the listeners in
a timeless present even if paradoxically this present is itself limited in time
, andin an almost romantic fashion music appears as la chose la plus sublime
dumonde (85), as something which verges on the sacred (see p. 178). Admitted-
ly, these transcendental ights, which represent a variant of the age-old meta-
physical connection between music andcosmic harmony as instancedin the
Pythagorean myth of the harmony of the spheres, are relativizedin Les Varia-
tions Goldberg by the non-romantic, postmodern scepticism pervading the novel,
a scepticism which also aects the centredness of Lilianes very identity and
andby many, more secular metamedial reections. Yet ultimate-
ly the musical form of Hustons novel almost rearms the primacy of music
andits harmony in spite of everything, so that Les Variations Goldberg seems to
be a late realization of Walter Paters famous dictum, according to which All art
constantly aspires towards the condition of music (1877/1973: 45). This realiza-
tion is not limitedto a theoretical aestheticist armation of form: Hustons
musicalization is itself an illustration of aesthetic beauty andpermits the reader
to experience such beauty as perhaps a last andin itself problematic
remnant of positivity in a highly problematic andunharmonious world. As Anna,
the focalizer of Variatio XX, knows from her experience with music as a
therapeutic andnarcotic inuence in her work with mentally handicapped
children: La musique, a sert a; nous distraire de lhorreur. (168)
5. Coda: Problems of recognizing intermedial iconicity
and perspectives of research
In all three variations of intermedial iconicity which I have discussed, literary
form couldbe seen to imitate features of other media without, however, actually
turning into these media: in these imitations the other media are evoked by
implicit references that coalesce into an imaginary presence only in the recip-
ients mind. Although the examples chosen may seem fairly obvious to some,
others may still have doubts as to how one can be sure of the existence and the
precise nature of the alleged intermedial iconicity. Indeed: why, for instance, is
Hardys Under the Greenwood Tree not simply a visual but a pictorialized novel,
andwhy can Hustons novel be more convincingly called a musicalized text than
for instance RaymondQueneau s Exercices de Style (1947), a text that is also
basedon the principle of variation?
However, it will be notedthat in all three cases chosen here, and contrary
to Queneaus experimental text, the imitation of the respective medium is
reinforcedby thematizations of the media referredto. Hardy even plantedsuch
clues in his subtitle and conrmed them by the mention of Fancy appearing
framedas [in] a picture; Lodge insertedan extendedd iscussion on lm in his
lmicized chapter; and Huston didsomething analogous in her book, which, in
addition, even contains an intermedial reference in its title. Such clues not only
help the reader but are often a precondition for deciphering the medial reference
correctly. Yet, such brief explicit references in the mode of thematization are
only an initial incentive for a more extended intermedial reading. In addition, in
all three cases the texts couldbe seen to adopt features of the imitatedmed ia,
features that standout as relatively alien to the literary medium or at least as
unusual in the circumstances. Such unusual shaping of the narrative text couldbe
witnessedin Hardys literally framedportrait andits presentation as a prolonged
freeze, devices which go beyondthe conventions andmotivations of most
realist descriptions;
in Lodge we could notice the reduction of an entire chapter
into mere dialogue with editing directions; and in Hustons Les Variations
Goldberg we couldobserve a marked de-narrativization, which ledto the
fragmentation of an entire novel into thirty-two separate interior monologues
arrangedin a cyclical manner.
A trace of doubt may still remain, and perhaps necessarily does so, since the
very nature of iconic imitation implies the simulation of andan imitative
approach to, something which will always ultimately remain absent.
On the one
hand it is certainly advisable to bear in mind the dangers and pitfalls of
intermedial metaphors if applied in an impressionistic way for interpretive
purposes. On the other hand, when dealing with cases as discussed in this essay,
one shouldnot over-emphasize the ultimately unbridgeable gap between the iconic
literary signiers and the intermedial signieds or referents and thus lose sight of
what is also important: the fact that in some exceptional andperhaps all the more
signicant cases literature or in my examples, ction can be convincingly
shown at least to attempt to approach the condition of other media. Whether such
attemps are, or can be, successful or not is quite another question, andperhaps one
of secondary relevance, compared to the perspectives which the identication of
such intermedial gestures allows. Foremost among these perspectives are functional
considerations, or in other words questions such as: what is the eect of the
intermedial iconicity for the structure of the text, for the reader, for the position of
the text in cultural andmed ia history? It is this that I nd particularly interesting
about intermedial iconicity, and I hope to have shown to what extent the awareness
of intermedial iconicity can be rewarding for the reading of certain texts.
This variant of iconicity is in fact well worth further investigation. Much
remains to be done in the eld of a systematic examination of historical exam-
ples of intermedial imitation, in the eld of a general theory andtypology of
intermediality and, last but not least, in the eld of the various functions that
intermedial iconicity can serve. I could do no more here than state a theme and
reveal some of its aspects. Further variations or developments of this theme must
be left to future research.
1. For detailed general typologies of intermediality see Wolf 2002 and Rajewsky 2002a.
2. For the denition of intermedial imitation (which is opposed to explicit intermedial reference
or thematization) see Wolf 1999: 44 f.
3. An exception to this tendency towards metaphoric use is Byerly 1997: ch. 4, although her discus-
sion is not basedon any theoretical inquiry into the general phenomenon of intermedial iconicity.
4. See Byerly 1997: 151 (who cites critical phrases such as graphic pictures of rustic life or
watercolours in words); for the pictorial quality of some of Hardys descriptions see also
Smart 1961 (a thorough collection of evidence for Hardys visual sensitivity [264] and its
inspiration by painting; Smart [see 271 f.] also quotes part of Fancys description, but does not
discuss its painterly quality in detail nor does he mention iconicity); cf. also Berger 1990, who
rightly mentions Hardys style that demands crossing the line between the visual arts and
writing (6), without, however, exploring the technical details of this kind of literary visualiza-
tion from an intermedial point of view.
5. Cf., with reference to a parallel scene in Hardys The Woodlanders, Senn 2001: 105: a scene
that resembles the description of a painting.
6. For Hardys emphatic use of [painterly] terms and the precision and subtlety of his notations
of colour cf. Smart 1961: 271.
7. Senn (2001: 103 f.) also briey comments on the scene that I have singledout, yet, like the
authors mentioned above, in note 4, she does so without reference to intermedial iconicity.
8. Incidentally, Claus Clver (1989: 65) had observed the same phenomenon albeit without
mentioning iconicity with reference to the intermedial imitation of a painting in a poem
(Anne Sextons Starry Night as an ekphrasis of van Goghs La Nuit toile), in whose stanza
andverse structure he discovereda re-creation of the process in which a viewer perceives and
responds to the scene. Cf. also, with reference to non-ctional texts, Tabakowska 1999.
9. For these andother motivations of description cf. also Sternberg 1981.
10. For Hardys internal perspective as being conducive to perceiving some of his scenes as
pictures cf. also Jackson 1984: 95 (however, she unconvincingly tries to monopolize this use of
perspectivity as well as Hardys framing techniques for the imitation of photography).
With reference to these descriptions and similar cases Byerlys claim that Hardys pictorial
images [] are presented[] from the narrators point of view is clearly wrong (1997: 153).
11. Cf., e.g., at the very beginning of the novel, the description of the working villagers of the
parish of Mellstock in the manner of portrait[s] in black cardboard or some processional
design on Greek or Etruscan pottery (Hardy 1872/1985: 40), furthermore the description of Old
James as a well-illuminatedpicture (51), or of Mr Penny in his workshop, who is presented
like a framedportrait of a shoemaker (98).
12. Byerly (1997: 152) claims that Hardys frequent references to [] pictures are not illusio-
nist, but this is at best true of mere intermedial name-dropping (ibid.) and not of attempts
to pictorialize ction as in our example.
13. For the ambivalent relationship between Hardy and photography see also Jackson 1984
(Jackson, however, speaks of Hardys photographic style, while admitting that painting in
general [] shares some if not most of the [] qualities [] constituting the photographic
style [1984: 94]).
14. A symptom of this emulation of painting is the frequent explicit reference to painters, and
particularly to Salvator Rosa, in late eighteenth-century texts, e.g., in The Man of Feeling: [The
old man] was one of those gures which Salvator would have drawn; nor was the surrounding
scenery unlike the wildness of that painters backgrounds (Mackenzie 1771/1970: 85), or in
Radclies The Mysteries of Udolpho (1794): This was such a scene as Salvator would have
chosen, hadhe then existed, for his canvas (1794/1980: 30); for the beginnings of iconic
pictorialization in The Mysteries of Udolpho see also Wolf 2001.
15. See Hardys statement on realist copyism in The Science of Fiction (1891): A sight for the
ner qualities of existence, an ear for the still, sadmusic of humanity, are not to be acquired
by the outer senses alone, close as their powers in photography may be (quotedfrom Greiner
andKemmler , ed. 1997: 150). Cf. also the highly ambivalent evaluation of photography or
daguerrotype by the narrator of Anthony Trollopes Barchester Towers, who regret[s] that
no mental method of daguerreotype or photography has yet been discovered by which the
characters of men can be reducedto writing andput into grammatical language with an
unerring precision of truthful description, but who also admits: Let photographers and
daguerreotypers do what they will, [] they will never achieve a portrait of the human face
divine (1857/1983: 167).
16. Cf. also Rippl (1999), who in her interpretation of A. S. Byatts Still Life (1985) attributes such
a self-reexive function as well as a more traditional enhancement of the texts reality eect
(531) to the pictorialization of this novel.
17. Among these there is authorial narration, the epistolary novel andthe documentary novel.
18. There is, however, a continuation of the story in Lodges Academic Romance (subtitle) Small
World (1984), a kindof sequel to Changing Places (Authors Note).
19. The laying bare of this problematic nature and the inclusion of doubts about the medium,
according to Lodges own aesthetics (see Lodge 1969/71/77), is the hallmark of the self-
reexive problematic novel, of which some of Lodges own ctions are good examples (see
Wolf 1989 and1992).
20. A gesture towards lmicizing large parts of a novel is also to be found in Robert Coovers A
Night at the Movies (1987). For further studies on lmicized ction in English see also Edel
1974 and from a sceptical point of view Kellman 1987; cf. also with reference to
postmodernist Italian ction Rajewsky 2002b.
21. For further discussions of literary analogies to musical forms see, e.g., Petri 1964/84, Brown
1978/2000, Picard1995 andW olf 1999.
22. For the question of dating see the Preface to Bach 1741/n.d.: iii.
23. Cf. also a remark made to Bernald: [] les variations reprennent non pas la mlodie du
thme, mais seulement lagencement de ses harmonies (Huston 1981/94: 125).
24. For a discussion of this term see Wolf 1999: ch. 4.4.
25. Like most of the other eects of the musicalization this lack of teleology is metamedially
highlighted; see the reection attributed to Bernald, On ne progresse pas vers un apoge, une
rvlation du sens profond: il pourrait y avoir mille variations, nest-ce pas? et le centre vide
resterait le mme (Huston 1981/94: 125).
26. Tout cela, cest moi qui lai imagin, en eet. Sauf quil ny a pas de moi qui soit que moi.
Chaque variation, cest moi qui lai compose. Avec les notes de Bach. Avec les gens dans
cette salle. Toute seule dans ma tte. [] Jai prtendu parler pour trente personnes. Non pas
dans le sens franais de la prtention, mans dans celui, anglais, du faire semblant (Huston
1981/94: 247).
27. For the connoisseur of Bachs music the insistent references to the Goldberg-Variationen both
through thematization andiconic imitation can also be saidto constitute an acoustic and hence
sensory backgroundwhich accompanies at least some variations andcontributes intermittently
to a certain sensualization of the text if only in the imagination.
28. Cf. above, note 26, the quotation from p. 247, andalso the thoughts attributedto one of
Lilianes pupils: Et je sais aussi pourquoi le choix des Variations: par ce quelle-mme est
comme a. En fragments. (45) Thus music is here functionally relatedin an ambivalent way
to the phonocentric suggestion of the presence of a stable subject discussed by Meinhard
Wingkens (1996) with reference to the use of music in George Eliots novels.
29. For the deciphering of pictorialization, as opposed to mere visualization, see Torgovnick
1985: 2628; for the recognition of musicalized ction when reading one, see Wolf 1999: ch. 5.
30. As Cobley (1986: 403) has pointedout, such framing as a phenomenological [] motivation
of description usually centres on the subject of perception (a focalizer pausing in an activity)
and provides a means of following his or her gaze; but here it is the other way round: Fancy
Day is the object of perception.
31. For such doubts with respect to the musicalization of ction see e.g. Rabat 1980 or, with
reference to the notion of a cinematic novel, Kellman 1987.
Armstrong, N. 1999. Fiction in the Age of Photography: The Legacy of British Realism.
Cambridge, Mass.: Harvard University Press.
Bach, J. S. 1741/n.d. Goldberg-Variationen: Vierter Teil der Klavierbung/Goldberg
Variations: Fourth Part of the Clavier bung. Urtext der Neuen Bach-Ausgabe/
Urtext of the New Bach Edition. Christoph Wol (ed.). Kassel: Brenreiter.
Barthes, R. 1982. LObvie et lobtus: Essais critiques III. Tel Quel. Paris: Seuil.
Berger, Sh. 1990. Thomas Hardy and Visual Structures: Framing, Disruption, Process.
New York: New York University Press.
Broich, U. 1993. Mutedpostmod ernism. The contemporary British short story. Zeit-
schrift fr Anglistik und Amerikanistik 41: 3139.
Brown, C. S. 1978/2000. Theme andvariations as a literary form. Yearbook of Compar-
ative and General Literature 27: 3543. Rpt. in Musico-Poetics in Perspective:
Calvin S. Brown in Memoriam. WordandMusic Studies 2. J.-L. Cupers andU.
Weisstein (eds), 235252. Amsterdam: Rodopi, 2000.
Byerly, A. 1997. Realism, Representation, and the Arts in Nineteenth-Century Literature.
Cambridge Studies in Nineteenth-Century Literature and Culture 12. Cambridge:
Cambridge University Press.
Clver, C. 1989. On intersemiotic transposition. Poetics Today 10: 5590.
Cobley, E. 1986. Description in realist discourse: The war novel. Style 20: 395410.
Edel, L. 1974. Novel and camera. In The Theory of the Novel: New Essays. J. Halperin
(ed.), 177188. New York: Oxford University Press.
Greiner, W. F. andF . Kemmler (eds). 1997. Realismustheorien in England 16921919:
Texte zur historischen Dimension der englischen Realismusdebatte. Tbingen: Gunter
Hardy, Th. 1872/1985. Under the Greenwood Tree or The Mellstock Quire: A Rural
Painting of the Dutch School. Penguin Classics. David Wright (ed.). Harmondsworth:
Hofstadter, D. R. 1979/80. Gdel, Escher, Bach: An Eternal Golden Braid. A Metaphorical
Fugue on Minds and Machines in the Spirit of Lewis Carroll. Harmondsworth:
Horace (Quintus Horatius Flaccus) 23 B. C./1980. Ars Poetica/Die Dichtkunst: Lateinisch
und deutsch. Eckart Schfer (transl. anded .). Stuttgart: Reclam.
Huber, M. 1992. Text und Musik: Musikalische Zeichen im narrativen und ideologischen
Funktionszusammenhang ausgewhlter Erzhltexte des 20. Jahrhunderts. Mnchener
Studien zur literarischen Kultur in Deutschland 12. Frankfurt/M.: Lang.
Huston, N. 1981/94. Les Variations Goldberg: Romance. Collection Babel. Arles: Actes
Jackson, A. M. 1984. Photography as Style andMetaphor in the Art of Thomas Hardy.
In Thomas Hardy Annual 2. N. Page (ed.), 91109. London: Macmillan.
Kellman, St. G. 1987. The Cinematic Novel: Tracking a Concept. Modern Fiction
Studies 33. 46777.
Lagerroth, U.-B. 1999. Reading Musicalized Texts as Self-Reexive Texts: Some
Aspects of Interart Discourse. In Word and Music Studies: Dening the Field.
Proceedings of the First International Conference on Word and Music Studies at Graz,
1997. Wordand Music Studies 1. W. Bernhart, St. P. Scher andW . Wolf (eds),
205220. Amsterdam: Rodopi.
Larsson, D. F. 1980. The Camera Eye: Cinematic Narrative in U. S. A. and Gravitys
Rainbow. In Ideas of Order in Literature and Film: Selected Papers from the Fourth
Annual Florida State Univ. Conference on Literature and Film. P. Ruppert (ed.),
94106. Tallahassee: Univ. Press of Florida.
Lodge, D. 1969/71/77. The Novelist at the Crossroads, originally in: The Critical
Quarterly 11/1969. 105132. Rpt. in D. Lodge. The Novelist at the Crossroads and
Other Essays on Fiction and Criticism. London: Routledge & Kegan Paul, 1971, and
in The Novel Today: Contemporary Writers on Modern Fiction. M. Bradbury (ed.),
84110. Manchester: Manchester Univ. Press, 1977.
1975/78. Changing Places. Harmondsworth: Penguin.
1984/85. Small World. Harmondsworth: Penguin.
1977. Thomas Hardy as a Cinematic Novelist. In Thomas Hardy After Fifty
Years. L. St. John Butler (ed.), 7889. London: Macmillan, 1977. (originally:
Thomas Hardy and Cinematic Form, Novel 7/1974. 246254).
Mackenzie, H. 1771/1970. The Man of Feeling. Brian Vickers (ed.). Oxford: Oxford Univ.
Mosthaf, F. 2000. Metaphorische Intermedialitt: Formen und Funktionen der Verarbeit-
ung von Malerei im Roman Theorie und Praxis in der englischsprachigen Erzhl-
kunst des 19. und 20. Jahrhunderts. Horizonte 15. Trier: WVT.
Nnny, M. andO. Fischer (eds). 1999. Form Miming Meaning: Iconicity in Language and
Literature. Amsterdam: John Benjamins.
Pater, W. 1877/1973. The dialectic of art: The school of Giorgione. In W. Pater, Essays
on Literature and Art. Jennifer Uglow (ed.), 4347. London: Dent.
Petri, H. 1964/84. Literatur und Musik: Form- und Strukturparallelen. Gttingen: Sachse
& Pohl. Rpt. Form- undStrukturparallelen in Literatur undMusik. In Literatur
und Musik: Ein Handbuch zur Theorie und Praxis eines komparatistischen Grenz-
gebietes. St. P. Scher (ed.), 221241. Berlin: Schmidt, 1984.
Pfandl-Buchegger, I. 1993. David Lodge als Literaturkritiker, Theoretiker und Romanautor.
Heidelberg: Winter.
Picard, H. R. 1995. Die Variation als kompositorisches Prinzip in der Literatur. In
Musik und Literatur: Komparatistische Studien zur Strukturverwandtschaft. Europ-
ische Hochschulschriften. A. Gier andG. W. Gruber (eds), 3560. Frankfurt/M.:
Queneau, R. 1947/82. Exercices de style. Paris: Gallimard.
Rabat, J.-M. 1986. The silence of the sirens. In James Joyce: The Centennial Sympo-
sium. M. Beja, Ph. Herring, M. Harmon andD. Norris (eds), 8288. Urbana: Univ.
of Illinois Press.
Radclie, A. 1794/1980. The Mysteries of Udolpho. The Worlds Classics. B. Dobre and
F. Garber (eds). Oxford: Oxford University Press.
Rajewsky, I. 2002a. Intermedialitt. UTB 2261. Tbingen: Francke.
Rajewsky, I. 2002b. Intermediales Erzhlen in der italienischen Literatur der Postmoderne:
Von den giovani scrittori der 80er zum pulp der 90er Jahre. Tbingen: Francke.
Rippl, G. 1999. Visuality andekphrasis in A. S. Byatts Still Life andArt Work. In
Anglistentag 1999 Mainz. Proceedings. B. Reitz andS. Rieuwerts (eds), 519534.
Trier: Wissenschaftlicher Verlag Trier.
Scher, St. P. 1968. Verbal Music in German Literature. Yale Germanic Studies 2. New
Haven: Yale University Press.
Schweitzer, A. 1908/72. J. S. Bach. Wiesbaden: Breitkopf & Hrtel.
Senn, C. 2001. Framed Views and Dual Worlds: The Motif of the Window as a Narrative
Device and Structural Metaphor in Prose Fiction. Bern: Lang.
Smart, A. 1961. Pictorial imagery in the novels of Thomas Hardy. Review of English
Studies n.s. 12/47: 262280.
Sternberg, M. 1981. Ordering the Unordered: Time, space, and descriptive coherence.
Yale French Studies 61: 6088.
Tabakowska, E. 1999. Linguistic expression of perceptual relations: Iconicity as a
principle of text organization (A case study). In Nnny and Fischer (eds) 1999,
Torgovnick, M. 1985. The Visual Arts, Pictorialism, and the Novel: James, Lawrence, and
Woolf. Princeton, N. J.: Princeton University Press.
Trollope, A. 1857/1983. Barchester Towers. Penguin Classics. Robin Gilmour (ed.).
Harmondsworth: Penguin.
Wingkens, M. 1996. George Eliot undd ie Musik: Funktionsvarianten von Musik und
Stimme in ihrem Romanwerk. In Expedition nach Wahrheit: Festschrift zum 60.
Geburtstag von Theo Stemmler. St. Horlacher andM. Islinger (eds), 399424.
Heidelberg: Winter.
Wolf, W. 1989. Literaturtheorie in der Literatur. David Lodges Small World als kritische
Auseinandersetzung mit dem Dekonstruktivismus. Arbeiten aus Anglistik und
Amerikanistik 14: 1937.
Wolf, W. 1992. The ction of DavidLod ge in the eighties. Discontent with life and
literature expressedfrom a moderate, liberal andund ogmatically Christian point of
view. Anglistik und Englischunterricht 48 (A Decade of Discontent: British Fiction
of the Eighties): 1942.
Wolf, W. 1999. The Musicalization of Fiction: A Study in the Theory and History of
Intermediality. Internationale Forschungen zur Allgemeinen undV ergleichenden
Literaturwissenschaft 35. Amsterdam: Rodopi.
Wolf, W. 2001. The emergence of experiential iconicity andspatial perspective in
landscape descriptions in English ction. In The Motivated Sign. Iconicity in
Language and Literature 2. O. Fischer and M. Nnny (eds), 323350. Amsterdam:
John Benjamins.
Wolf, W. 2002. Intermediality revisited: Reections on word and music relations in the
context of a general typology of intermediality. In Word and Music Studies: Essays
in honour of Steven Paul Scher on Cultural Identity and the Musical Stage Proceed-
ings of the Third International Conference on Word and Music Studies at Sydney, NSW,
2001. Word and Music Studies 4. S. Lodato, S. Aspden and W. Bernhart (eds).
Amsterdam: Rodopi.
</TARGET "wol">
AUTHOR "Elz bieta Tabakowska"
TITLE "Iconicity andliterary translation"
WIDTH "150"
Iconicity and literary translation
Elzbieta Tabakowska
Jagiellonian University of Krakw, Poland
1. Preliminaries
The 2001 Jena Symposium on Iconicity in Language andLiterature the third
in a series provedonce again that iconicity, in all its manifestations, has
gainedan unquestionable position in both linguistic and literary studies. Theoreti-
cal discussions and practical analyses are based on the assumption that it is a
universal feature of human communication. Dened as a particular repertoire of
values stable anduniversal accross languages, iconicity is clearly functional in
nature. As a measure of functionality, it belongs among factors that enhance
communication via language.
Following the classical Peircean trichotomy, iconic signs have been tradi-
tionally divided into three subcategories: images, diagrams and metaphors. As is
well-known, linguistic signs belonging to the rst group are discussed under the
headings of onomatopoeia andsound symbolism, andthe sign is taken to
represent its object by imagic similarity to it. Thus by inspecting the sign we
may gain knowledge of the object. But while the principle itself is in all
probability universal, its manifestations in various languages obviously are not.
The problem has been dealt with too often to make it possible for yet another
discussion to provide anything but a repetition of trivial facts. It may suce to
remindthe reader of a single, well-known example. While for the French poet
Rimbaudin his poem Voyelles vowels symbolized the following basic colours:
A noir, E blanc, I rouge, U vert, O bleu: voyelles, the German philosopher August
Wilhelm Schlegel ascribedd ierent parts of the spectrum to German counterparts
of these vowels: A: redor white, O: purple, I: is the blue of the sky, E: grey, U: dark
blue (Schlegel, quotedin Mayenowa 1979: 428). Consequently, French and
German were considered to display dierent euphonic potentials (ibid.).
362 ELZ

The other type of iconicity, known as diagrammatic (or metaphorical, since
the theoretical demarcation line between the two categories proves dicult to
establish see e.g. Hiraga 1994 andSection 4 below) is far less trivial.
Diagrammatic (or metaphorical) iconicity involves cases in which particular
relations between signs mirror relations between objects. Within this category
various authors speak of rst and second degree of iconicity. The former is
basedon a universal andgeneral principle of diagrammaticity; chief linguistic
evidence comes from pidgin communication, i.e. such grammatical systems
(proto-grammars) whose rules are extremely iconic, that is, cognitively
transparent andnon-arbitrary . On the level of intonation, proto-grammars
manifest direct correspondence between stress and semantic predictability,
between melody and pragmatic relevance of information, or between pauses and
rhythm of utterance and individual information chunks. Within an utterance,
spacing mirrors semantic segmentation (proximity vs. relevance, proximity vs.
scope), andsequencing reects ordering of information along the scale of relative
importance. Finally, the quantity of linguistic material reects the pragmatic
value of information (e.g. zero expression vs. predictability and relevance)
(Givn 1995: 406407); iconicity resulting from the general principle more form
more meaning is referredto as quantitative (as opposedto qualitative or
sequential). On the other hand, diagrammatic iconicity of the second degree
is basedon linguistic relations within a particular grammatical system, andthus
is non-universal, or language-specic (e.g.three two-syllable past tense verbs
beginning with same soundin the classical veni vidi vici example).
Since all types of iconicity are ultimately basedon the notion of similarity,
the pertinent question to ask is: What is similarity? The notion of similarity
underlies all categorization, and therefore also all perceptual and cognitive
processes of the human mind. Contemporary cognitive sciences (anthropology,
psychology, linguistics) acknowledge its basically subjective character: objects of
perception seem to be similar rather than are similar to each other (or one
another). In other words, similarity must be more or less consciously acknowl-
edged by a perceiver. Second, objects may seem to be similar to observers in
certain respects. The Wittgensteinian notion of family resemblance is now
believedto underlie the processes of categorization: human beings ascribe objects
to prototype categories on the basis of sharedfeatures which they consider
relevant (cf. classical Aristotelian categorization basedon nite sets of objective
critical attributes). Third, similarity is a scalar phenomenon: objects may be
judged to be more or less similar: within categories, objects are situatedat
varying conceptual distances from the category center, i.e. the prototype. Finally,
similarity is in general perceived andestablishedwith particular purposes in
view, the governing principle being that of pragmatic relevance. Thus it is
ultimately basedupon choice, andcond itioned by particular systems of values,
beliefs, attitudes, ideologies, etc. In short, it is culturally determined.
Since perception of similarity requires an interpreter, it is not self-explanato-
ry; what is at stake is not an objective property of being alike but the speakers
knowledge of and about relevant features of objects. Things become
similar when a particular observer conceptualizes them as such. Therefore
similarity, which constitutes the dening property of iconic signs as representa-
tions of their objects, is a product of the process of conceptualization, and thus
implies the agency of an (individual) observer. In other words, iconic expressions
can only be dened as iconic my way or iconic your way. Hence their
subjectivity, ambivalence and language-specicity. Like metaphor, diagrammatic
iconicity comes to exist only when it is perceived, andit becomes perceivedonly
if the perceiver possesses the necessary knowledge and necessary skills.
The relation between universal cognitive mechanisms andprinciples that
underlie linguistic phenomena on the one hand and particular uses of those
mechanisms and principles which surface in individual linguistic conventions on
the other has for long been the moot point of linguistic theories. The Chomskyan
idealized speaker has re-appeared in the cognitivist guise as Langackers
generalizedobserver : an embodiment of conventionalizedperceptions. Icons
undergo conventionalization like all other elements of language. Each iconic
sign involves some more or less transparent conventional principles that underlie
its production andreception. Those principles may be entrenchedin our culture
(Mayenowa 1973: 48, transl. E. T.). Our understanding of iconic signs is thus
based on both our knowledge of the worldandour knowledge of conventions
(Wysouch 1985: 208, transl. E. T.). But it also relies on perceptual abilities
of each individual observer: the continuum marking a transition from an individ-
ual observer to a conventional observer runs parallel to the route from
linguistic convention to linguistic creativity. Novel signs appear as more or less
creative applications of well-entrenchedconventional principles.
Thus, when talking about iconicity, it is important to distinguish between
two types of signs: those created ad hoc andthose that are sanctioned by
linguistic convention. Limiting the discussion to diagrammatic iconicity, we shall
illustrate both types with particular reference to literary translation.
2. Iconicity and translation
Conventional iconicity undergoes lexicalization and/or grammaticalization, and
in the process the transparency of the underlying principles naturally diminishes
364 ELZ

or disappears altogether. Lexicalization and grammaticalization are dened as
results of the evolution of iconic features into symbolic ones. Unless emphasized
by the speaker, lexicalizedandgrammaticalizediconicity generally goes unno-
ticed both by the speaker andby the hearer. The process, which some authors
refer to as de-iconization of signs (cf. Wysouch 1985: 207), raises doubts as to
whether they shouldbe still describedas iconic, at least from the synchronic
perspective. The question parallels the problem concerning the status of conven-
tionalized (dead) metaphor (cf. Cienki 2000). Creative iconicity, on the other
hand, retains transparency and, like any act of creativity, tends to attract the
attention of the receivers of the message.
The interest of translation theory (andtranslation practice) like that of
contrastive linguistics lies in the non-universal character of conventional
iconicity. As was said, in spite of the universality of underlying principles, it is
language (or culture) specic. For instance, although all languages display
grammaticalizediconic wordord er whereby a more important information
chunk is fronted (Givn 1995: 438), it is evident that not all languages imple-
ment the principle in all cases or in the same way.
However, I wish to claim that in interlingual translation andliterary
translation in particular iconicity may become problematic only when it
constitutes (an element of) a particular goal-orientedstrategy on the part of the
original author, i.e. when it becomes instrumental in achieving a particular
communicative purpose. In other words, it has to be intentional, thus going
beyondthe scope of mere communicative functionality. The intentionality may
apply to both types of iconicity: the conventional and the adhoc, andthe
problems may stem from various sources, which the following two case studies
are meant to illustrate.
3. Case studies
3.1 Extension of conventionalized diagrams: A Medieval miniature
by Wisawa Szymborska
The crucial part of the poem in an English translation by the Polish poet
Stanisaw Baranczak andhis English-speaking colleague Clare Cavanagh (1996)
is given below.
Up the verdantest of hills,
in the most equestrian of pageants
wearing the silkiest of cloaks.
Towards a castle with seven towers,
each of them by far the tallest.
In the foregrounda duke
most atteringly unrotund
by his side his duchess
young andfair beyondcompare.
Behindthem, the ladies-in-waiting,
all pretty as pictures indeed,
then a page, the most ladsome of lads,
andperchedupon his pagey shoulder
something exceedingly monkey-like
endowed with the drollest of faces
Following close behindthree knights,
all chivalry andrivalry ,
so if the rst is fearsome of countenance,
the next one strives to be more daunting still,
andif he prances on a bay steed,
the thirdwill prance upon a bayer,
andall twelve hoofs dance glancingly
atop the most wayside of daisies.

Thus they proceedmost pleasantly

through this feudalest of realisms.

An object of scholarly interest to several Polish literary scholars (cf. Zarebina

1991, Pajdzinska 1998), the original poem is described as a successful attempt at
intersemiotic translation: a verbal rendering of a non-verbal picture. To anyone
familiar with medieval painting the analogy is striking: what comes to mind is
one or all of the well-known miniatures, oeredto Duke de Berry by the
Limbourg brothers. It recreates all objects characteristic of the worldaccord ing
to the Limbourgs: the prince andhis lovely wife, the ladies-in-waiting, the
knights andthe pages making up their entourage, the castle with its tall towers,
and the verdant hill. To those familiar with the style of medieval painting, the
text of the poem is diagrammatically iconic in at least four ways (Pajdzinska
1998). The highly conventionalizedpainting style, which requiredmeticulous
representation of the everlasting essence of things at the cost of realistic
faithfulness to idiosyncratic detail resulted in high schematicity, where objects
were representedas prototypes of their categories. The reverse of impression-
ism, this overall conventionality is mirroredin Szymborskas poem by the
366 ELZ

conventionality of linguistic expression, notably the abundance of set idiomatic
phrases (most of which, however, disappear in the English translation). The
naivet, which contemporary viewers tendto ascribe to medieval miniatures,
nds its verbal counterpart in simple, unsophisticated syntax, which only
exceptionally departs from parataxis. Finally, the static character of a pictorial
image, which freezes all movement in the depictedscene, is paralleledby almost
total lack of nite verbs (cf. Pajdzinska 1998: 86). All these formal features
comply with the rules of Polish grammar, andtheir particular function in the
poem becomes apparent only when the poets intention is recognized. Thus the
readers knowledge of the world (viz. of medieval painting) combines with the
knowledge of linguistic convention to reveal the hidden iconic meaning of the
text. Once noticed, at least some of it can be rendered in translation at least
in the case of an English rendering, which may employ equivalent formal devices
in the target language (cf. e.g. the limitednumber of nite verbs in the Baran-
czak-Cavanagh version).
However, at this point I wouldlike to concentrate on the last of the four
iconic aspects of Szymborskas poem: the use of adjectives in the superlative
degree. Naturally, it has not escaped the attention of both literary critics and
linguists (cf. notably Zarebina 1991, Pajdzinska 1998); my own analysis (made
from the theoretical perspective of cognitive linguistics) was presentedin
Tabakowska 1998. The poet repeatedly introduces superlative forms of adjectives
that normally occur only in the positive degree. When discussing their distribu-
tional limitations, Polish grammars propose various criteria, both formal and
semantic. It seems, however, that non-gradable adjectives refer to features that
dene non-prototypical, or Aristotelian, categories of the either-or structure,
where elements are merely considered to belong or not to belong to a category.
Consider the following fragment (line 3):
(WS1) w paszczach najjedwabniejszych
in coats most-silky
wearing the silkiest of cloaks
In normal reasoning, a courtiers cloak like any other garment can either
be or not be made of silk. Further down, in line 12 we nd
(WS2) paz najpacholetszy
page most-ladsome
page, the most ladsome of lads,
andin line 24, there are
(WS3) stokrotki najprzydrozniejsze
daisies most-roadside
the most roadside of daisies
A young male can be referred to as a lad when he is considered to have
properties conventionally associatedwith the word, andd aisies simply grow, or
do not grow, on the side of the road. But the superlative (like the comparative
degree) changes an Aristotelian category into a category based on the notion of
prototype, thus implying the existence of elements which are considered as its
good, better or best members. The comparative and the superlative pick up
those members within a category that display a higher or the highest degree of
the categorial feature, setting them against other elements, i.e. they involve an
implicit comparison within a set of objects. Szymborskas superlatives perform
a d i erent function. While still implying the highest level of intensity of the
feature, they erase the notion of comparison by removing from the visual eld
all other members of the category; for instance, the cloaks of those riding in the
pageant are most-silky in an absolute sense. The interpretation is enhancedby
placing the adjectives in postposition, which Polish reserves for predication, in
opposition to the merely attributive function of adjective preposition. Postposed
adjectives place things which they modify within categories whose categorial
attributes are properties capturedby the adjectives (in cognitive linguistics, such
categories are referredto as quality spaces; for discussion, see e.g. Tabakowska
2001). Objects are thus presentedas displaying the largest number andthe
highest conceivable degree of properties associated with the adjective silky
(softness, lustre, delicate texture, etc.) without any reference to other similar objects.
They are situated, so to say, in the very centre of the quality space and thus become
prototypical for the category. Such a prototype, however, is an abstract schema
rather than a concrete instantiation: the very essence of the quality. Thus the
superlative adjectives become diagrams of the essence of things as a property
of medieval painting, which strives to represent the stable, the most perfect
qualities. The most striking example comes in line 5 of the poem, where the
castle is described as having seven towers, of which each is the tallest:
(WS4) z ktrych kazda najwyzsza
of which each the highest
Each of them by far the tallest
The description, defying the normal state of aairs, where out of a set of objects
(seven towers) the perceiver wouldpick up one as surpassing the others in
height, mirrors the painters endeavour to depict a prototype of a castle rather
than just a castle.
368 ELZ

The unconventional superlatives conspire to create an overall diagram of the
painting style:
(WS5) w tym realizmie najfeudalniejszym
in this realism most-feudal
through this feudalest of realisms
The feudalest realism is an ironic label attached not to an object but to an
abstract prototype of an object.
As was said, instances of diagrammatic iconicity based on grammatical
convention are in general less noticeable, or considered less relevant, than its
creative ad hoc counterparts (cf. 3.2. below). The intention of the speaker
becomes visible within the context of the whole poem, where a pattern emerges
through a systematic repetition of signs which conspire to evoke the hidden
meaning that wouldnormally be ignored.
The translators, who are both masters of their trade, must have noticed
Szymborskas diagrams. As someone for whom English is not the native tongue,
I wouldnot dare ascribe the non-equivalence of their translation to their lack of
sensitivity to linguistic nuance but rather to an objective lack of equivalent
formal devices in the target language. However, the fact remains: in this
particular respect, the translation lacks equivalence. Although some of the
English superlatives are just as unconventional as their Polish counterparts
(ladsome, feudalest), the formal pattern usedis the (this) Adj-superlative of
N-plural: the silkiest of cloaks, the most ladsome of lads, the most wayside
of daisies, this feudalest of realisms. The objects are situated within their
respective sets, and described as possessing the highest degree of a particular
property in comparison with other category members. They become the best
representatives of their categories (or, as a cognitive scientist might put it,
exemplars) rather than schematic category prototypes. The diagrammaticity is
thus shiftedto a lower level: the poem becomes an icon of a token (a medieval
miniature) rather than an icon of a type (the miniature), in the generic sense.
3.2 Creating new diagrams: The Isles and Europe by Norman Davies
The second case study involves diagrams based on properties of signs that are
not conventionally used as iconic devices andwhich have been createdby the
author for the sake of a single message. They are basedon a general principle,
whereby formal parallels in a text imply the existence of semantic parallelism
(Mayenowa 1979: 168). This universal principle, well-known to classical rhetoric,
may have various language-and culture-specic realizations. The degree of this
type of iconicity may also vary, providedit goes beyondandabove what is
intuitively felt to be the normal level. For instance, accumulation of nouns
within an utterance, unless enhanced by some additional formal properties of the
text, wouldnormally fail to be consideredas a case of quantitative iconicity. The
case is dierent if all these nouns begin with the same letter (sound). As the initial
element of a word does not under normal circumstances carry information, an
emerging pattern attracts the attention of the readers or listeners, making them look
for hidden meanings. In this case, knowledge of the world involves the hearers
(or readers) expectations: sound parallelism is a conventional poetic device (in
rhyme or alliteration) andas such its aesthetic value will be taken for granted.
The unconventional iconic value of soundcomes to the fore only when there are
no grounds for such expectations. The extracts quoted