METAPHOR AND COGNITION
STUDIES IN COGNITIVE SYSTEMS
VOLUME 13
EDITOR
James H. Fetzer, University of Minnesota, Duluth
ADVISORY EDITORIAL BOARD
Fred Dretske, Stanford Universiy
Ellery Eells, Universiy of Wisconsin, Madison
Alick Elithom, Royal Free Hospital, London
Jery Fodor, Rutgers Universiy
Alvin Goldman, University of Arizona
Jaakko Hintikka, Boston Universiy
Frank Keil, Cornell University
William Rapaport, State University of New York at Bfalo
Bary Richards, Imperial College, London
Stephen Stich, Rutgers University
Lucia Vaina, B osto n University
Terry Winograd, Stanford Universiy
The titles published in this series are listed at the end of this volume.
METAPHOR AND COGNITION
An lnteractionist Approach
by
BIPIN INDURKHY A
Computer Science Department.
Boston University, Boston, MA. .S.A.
�·
''
KLUWER ACADEMIC PUBLISHERS
DORDRECHT I BOSTON I LONDON
Library of Congress Cataloging-in-Publication Data
InoJrkhy!.
B1p1n .
Metapho�
'ncunhva.
� .
Incluoes
ISBN
3.
5.
1.
anc
1959-
cogn1t1on
--
c• .
IStud1es
o>�11ogr3ph1ca1
0-7923-1687-8
Symbcl1sm
1n
I.
Cogn1t1on.
1992
approach
cogn•t•ve systems
alk .
<hara
T1tle.
Jnterac:1on1st
references
<Psychologyl
Analogy--Psvchciog;cal
Bl58.l53
an
2.
paperl
;
v .
4.
B1p1n
131
1ndexes.
�etaphor--Psycho1oglca1
aspects.
II.
and
1
S•m•larlty
aspects.
<Psychology!
Ser1es.
153--oc20
92-7189
ISBN 0-7923-1687-8
Published by Kluwer Academic Publishers,
P.O. Box 17, 3300 AA Dordrecht, The Netherlands.
Kluwer Academic Publishers incorporates
the publishing programmes of
D. Reidel, Martinus Nij hoff, Dr W. Junk and MTP Press.
Sold and distributed in the U.S.A. and Canada
by Kluwer Academic Publishers,
101 Philip Drive, Norwell, MA 02061, U.S.A.
In all other countries, sold and distributed
by Kluwer Academic Publishers Group,
P.O. Box 322, 3300 AH Dordrecht, The Netherlands.
Prited on acid-"ee paper
On the Cover: Other World by M.C. Escher.
© 194 7 M.C. Escher I Cordon Art - Baam- Holland
Collection Haags Gemeentemuseum, The Hague
All Rights Reserved
© 1992 Kluwer Academic Publishers
No part of the material protected by this copyright notice may be reproduced or
utilized in any form or by any means, electronic or mechanical,
including photocopyng, recording or by any information storage and
retrieval system, without written permission from the copyright owner.
Printed in the Netherlands
L tke
memory of my grandmother
and
to Xela
SERIES PREFACE
This series will include monographs and collections of studies devoted to the
investigation and exploration of knowledge, information, and data-processing
systems of all kinds, no matter whether human, (other) animal, or machine.
Its scope is intended to span the full range of interests from classical
problems in the philosophy of mind and philosophical psychology through
issues in cognitive psychology and sociobiology (concerning the mental
powers of other species) to ideas related to artificial intelligence and com­
puter science. While primary emphasis will be placed upon theoretical,
conceptual, and epistemological aspects of these problems and domains,
empirical, experimental, and methodological studies will also appear from
time to time.
The nature of metaphor and the nature of cognition are both illuminated in
this stimulating study by Bipin Indurkhya. Beginning with a distinction
between conventional metaphors, similarity-based metaphors, and similarity­
creating metaphors, he elaborates the idea that similarity-creating metaphors,
which affect an interaction between the source of the metaphor and its target,
fulfill a fundamental role in human cognition. In addition to the development
of his own account, lndurkhya thoughtfully examines alternative theories and
evaluates
their strengths and weaknesses.
By placing the problems of
metaphor within the framework of cognition, this work makes an exception­
ally valuable contribution to understanding the nature of the mind.
J. H. F.
vii
Contents
Acknowledgments
XV
Pro logue
The Pro b l e m
I
1
11
Characterizing Metaphor
13
1.1
1 .2
Introduction . . . . . . .
Some Examples of Metaphors
13
14
1 .3
Characteristics of Linguisti. Metap hors
17
1 .4
Degrees of Metaphoric Conten t :
T h e Conventional vs. the Metaphorical
Metaphors i n Non- Linguistic Domains
Metaphors, S i m i les, A nalogies and Models
1 . 6 . 1 Metaphors and S i m iles . .
1 . 6 . 2 Metaphors a n d Analogies .
1 . 6 . 3 Metaphors and Models
Conclusions . . . .
. .
1 .5
1 .6
1.7
2
1
.
.
.
.
Enter Similarity-Creating Metaphors
Introduction . . .
.
.
. .
.
. . .
.
.
.
19
21
26
26
28
34
36
39
2.1
2.2
2.3
Some Examples of S i m i larity- Creat i ng Metaphors
Psy chological S t u dies of t he Creation of S i m i larity .
40
2.4
Creation of S i m i larity i n Metaphor- Related P henomena
48
39
45
X
2.5
2.6
3
48
49
54
56
2.5. 1
S i m i lari ties Before and A fter the Metaphor .
57
2.5.2
S i m i l arities A fter but Not Before the Met aphor
59
2 . 5 .3 S i m i larities Before but not A fter the Metaphor .
Conclusions: The Problem of Sim i larity- Creati ng Metap hors
63
63
Approaches to Similarity-Creating Metaphors
65
3.1
3 .2
3 .3
Introdu ction
Max B lack .
P au l Ricoeur
65
68
74
34
Carl H ausman
75
3.5
W heel wright - M ac Cormac
76
3.6
3.8
The Lakoian A pproach . .
M y Earlier Approach . . . .
K i t t ay's Perspecti val Theory
3.9
Conclusions . . . . . .
78
84
86
90
Cognition as Interaction
93
4. 1
4. 2
93
94
3.7
4
2 .4 . 1 S i m i le .
2 . 4 . 2 A nalogy
2.4 . 3 Models .
S i m ilari ties and Creati ve P roblem Solving
Introduct ion . . . . . .
Empirical S upport for the Interaction View of Cogni tion
4 . 2. 1
4.2.2
4 . 2. 3
4 .2.4
4.3
4. 4
4.5
4.6
Concepts are M ore than Aggregates of Sense D ata .
Concepts can O rganize the World D i ferently . . . .
Concepts Cannot O rgani ze the Worl d Arbitrarily .
'Uni ver �als' and the Physiological Basis of Cognition
4.2.5 Summary . . . . . . . . . . . .
From Kant to Goodman: Worldmaking
P i aget's Constructivism . . . . . . . .
Lakof-Johnson: The Bodily Basis of Cognition .
Conclusions . . . . . . . . . . . . . . . . . . . .
95
1 00
1 04
1 05
111
111
116
1 24
1 27
xi
A Theory
II
5
129
An Interactionist Approach to Cognition:
Informal Overview
5. 1
Introduct ion .
5.2
A n Example .
5.3
5.4
5.5
5.6
5.7
6
131
1 35
Concept Networks .
151
Environments and Sensorimotor Data Sets
1 58
Cogni t i ve Relations and Coherency
161
A ccommodat ion and Projection
1 64
Cogni t i ve Models . . . . . .
1 69
5 . 7 . 1 G roupings on the Env i ronment
1 70
5 . 7 . 2 A ccom modation and Projection : A nother Perspect i ve . 1 74
5.7.3
5.7.4
5.8
5.9
131
Representation and Description . . . . . .
Layered Cogni t i ve System and M ulti p l e "Worlds"
1 76
1 78
1 79
S u mm ary . . . . . . . . . . . . . . . . . .
187
Some O ther M i scellaneous Notions . . .
.
An Interactionist Approach to Cognition:
Formal Concepts
6. 1
Introduction . . . . . .
6.2
6.3
Classes and G roupings
Relations and Induced G roupi ngs
6 . 3 . 1 P reli m i nary Defin i tions .
6.3.2
6.3.3
6.4
Functions and Operators
6.4.1
6.4.2
6.5
D i functional Relations .
Relations Within a C l ass
Functions
Operators
189
1 89
191
1 94
1 94
1 96
201
202
203
204
A lgebras and Structures
6 . 5 . 1 A lgebras . . . . .
6 . 5. 2 D escriptions and S t ructures
20.5
205
209
6.5.3
214
Closu res and Generating Classes .
xii
6.5.4
6.5.5
C losu re Over Operators
Computabi l i ty of O perators
6.6
Subalgebras a n d F i n i t e Generat i v i ty
216
2 16
2 17
6.7
G roupings on A l gebras: A lgebras o f Classes
220
6.8
Rel at ions Between A l gebras: Correspondences
6 . 8. 1 P roducts o f Algebras and Correspondences
6 . 8 . 2 G roupi ngs Induced b y Correspondences
6 . 8 . 3 Difunctional Correspondences
223
224
227
229
.
6.9
Cogn i t i ve Models . . . . . . . . . . .
6 . 9 . 1 Basic Dei n i tion . . . . . . .
6. 9 . 2 Local Coherency and Coherency .
6. 9 . 3 Some Characteristics of Cogni t i ve Models
6 . 10 Cogni t i ve Models Over an Environment .
235
236
6.11 P roj ecti ve and A ccommodating Models
6.12 F i n i te Representab i l i ty and Coherency
239
24 1
An Interaction Theory of Metaphor
245
.
7
232
234
7. 1
7.2
7.3
7.4
I n t roduction . . . . . .
M etaphor as P rojection .
omenclat u re Associ ated with Metaphor
Modes of Metaphor . . . . . . . . . . . .
7. 4 . 1 S i m i l ari ty-Based ( Comparat i ve) Metaphors .
246
253
256
256
7.4.2
271
7.5
Summary .
III
8
.
232
Si m i l ari ty-Creating (Proj ective) Metaphors .
The Implications
Some Metaphor-Related Issues
8.1
8.2
Introduction . . . . . . . . . .
T h e Thesis ' A ll K nowledge is M etaphorical'
8.2. 1
8.2.2
Version 1: All Knowledge i s Projecti ve
Version 2 : A l l Thought i s Comparati ve
245
279
283
285
285
286
287
289
xiii
8.2.3
8.2.4
8.3
8.4
9
Version 3 : A ll Conventional Mean ings A rise By Way
of M et aphor . . . . . . . . .
L akoff- Mac Cormac Debate
Metaphor and Correctness . . . . .
8 . 3 . 1 Correctness, Trut h and Coherency .
8.3 . 2 U nderstanding v s . Correctness . . .
8 . 3 . 3 Conventional a n d M etaphorical Correct ness
A p tness ( Q u ality ) of Metaphor
. . . . . . .
On Predictive Analogy and Induction
9.1
9.2
I ntroduction . . . . . . . . . . . .
Predictive A nalogy and Metaphor
9.3
The Search for Logical Justi ication o f P red i t i ve A nalogy
9.3. 1
9.3.2
9.4
9.5
9.6
9.7
P redictive A nalogy as an Inductive Process . . . . .
P redictive A nalogy a s a First O rder Generali zation
9 . 3 . 3 P redictive A nalogy as a Second O rder General ization
The Search for Empi rical J ustiication of P red i c t i ve Analogy
290
292
30 1
301
305
306
309
315
. 315
318
322
323
325
327
329
9 .4 . 1 Evidence from C l assroom Experiments . . . . . . . . 329
9 . 4 . 2 Evidence from Real-World Problem-Solving Activit i es . 332
The ' D ar k Side' of P redictive A nalogy
334
339
P redi c t i ve A nalogy and Cogn i tion . . .
The Problem of Induction . .
343
9.8
The S am p l i ng Princip le, Randomness, and the Generalized
Grue Paradox . . . . . . . .
9.9 The ' D ark Side' of Induction .
9 . 1 0 Induction in Cogni t io n . . . .
344
350
352
Computational Approaches to Metaphor and Analogy
357
10 On
10.1
Introduction
.
. . . . . . . . . . . . .
357
1 0 .2 Comp u t at i onal Approaches to Linguistic Metaphors
1 0 . 3 Com p ut ational A pproaches to Predi c t i ve A nalogy
. . . . . . .
360
365
1 0 . 4 A Computat iona] Model of Creati ve Analogies:
Douglas H ofstadter . . . . .. . . . . . . .
376
xiv
1 0 .4 . 1 A n Aside: Context- Sens i t ivity of Descriptions i n Evan s '
. . . . . . . . . . . . . . . . . . . .
. 378
Approach
10 . 4 . 2 Resumption: Hofstadter and M i t chell ' s Copycat
. 384
1 0. 5 P rojecti ve ( S i m ilarity- C reati n g ) Metaphor in Artiicial
Intelligence . . . .
. . . . . . . .
392
10 . 5 . 1 Projection as ' Top- Dow n ' G rouping
392
. 395
1 0 . 5 . 2 Novel vs. Conventional P rojection .
.
.
.
.
.
.
1 0 . 5 . 3 The C reation of S i m ilarity . . . . .
10. 6 Model ing Met aphor as Change of Representat i on
10 . 7 Conclusions . . . . . . . . . . . . . . . . . . . . .
. 400
.. 401
. 408
B ib l iograp hy
411
N ame Index
433
Subject Index
439
Acknowledgments
I wou l d l i ke to express my thanks and appreciat ion to the following people
who h ave contribu ted to this manuscript in various ways:
•
To Remko Scha for providing encouragement, i deas , and much con­
structi ve criticism t h rough the ent i re project .
•
To M ark Joh n son for providing an encouragi ng feedback on a long essay
t h at I wrote as an abridged version of t he book i n 1 988.
•
To Erica Melis for providing many usefu l comments on the abridged
version of the book an d on earlier drafts of some chapters.
•
To Sylvia Candelari a de Ram for a t horough readi n g of early drafts
of Chapters 1 , 5, 6 and 7 . Her numerous comments , especially on
Chapter 6 , were i nstrumental in add i n g to the clarity of the i nal draft .
•
To George Lakoff for providing an encouragi ng feedback on an earlier
draft of Chapter 7 and for providi ng val uable com ments .
•
To Melan ie M itchell for pai nstakingly goi ng through a very rough ver­
sion of the enti re manuscri p t . Her detai led comments were an invalu­
able aid in revising the manuscript .
•
To M argo G uerti n and Scott O ' H ara for read i ng through coun tless
drafts and catchi ng m any syntactic and semantic errors .
•
To M ary A t k i n s , Regina Bl aney and Beryl Nel on for read ing parts of
the i n al draft and ofering comments .
•
To Doug Hofstadter and Melanie Mitchell for
ments on parts of the inal manuscri p t .
•
To Spyridon B raoudakis a n d Dan Solis for carefully going over the
entire i n al draft and mak i ng n umerous comments.
•
To Marie- Dom i n i q ue G ineste and an anonymous
Academic P ublishers for wri t i ng very positive and
providing extensive
com­
reviewer for Kluwer
en c o u rag i n g revi ews
of the m a n u s c r i p t, an d for mak i ng comments on var i ous parls of i llhaL
were q u i te helpfu l d u r i n g rev i si on s.
XV
xvi
•
To S i mon Ross for doing a wonderfu l j ob as the edi tor i n get t i n g t h e
manuscript reviewed prompt ly, a n d i n oferi n g much useful advice i n
get t i ng the copyright permissions a n d i n the p reparation of the camera­
ready copy.
•
To the N at ional Scien ce Foundation for grant No. I R I - 9 1 05806, w h i ch
supported in part the p reparation of t h e manuscrip t .
I would also l i ke t o express m y grat i t ude to various i nd i v iduals and i nstitu­
t ions for allowing me to i nclude some copyrighted m aterial in t h i s book:
•
Oth e r World by M . C . Escher ( on the cover) : @ 1 947 M . C . Escher /
Cordon Art - Baarn - Holland, Collection aags Gemeentemuseum ,
The H ague. Reproduced w i t h the k i n d permission o f Cordon Art a n d
H aags Gemeentemuseum.
•
Fog by Carl Sandburg ( page 2): From C H I C A G O POEMS by Carl
Sandburg, @19 1 6 by H ol t , Rinehart and Winston , I n c . and renewed
1 944 by Carl Sandburg, repri nted w i t h the permi ssion of H arcourt
Brace Jovanovich, Inc.
•
Pa rk bei Lu(zern) (Pm·k n ea r L u(cerne)) ( 1 93 8 ) by Paul K lee ( plate 1 ) :
Paul Klee- S t i ft ung/Kunstmuseum Bern , @ 1 9 9 1 by VAG A , New York ;
reproduced w i t h t h e k i n d permission o f Kunstmuseum Bern .
•
•
•
•
Th e MmTiage of Giovan n i (?), Arnolin i and Giovan n a Cen a m i (?),
( 1 434) by J an Van Eyck ( plate 2 ) : The National G al lery, London,
reproduced w i t h the k i n d permission of the National Gallery.
Quotat ions from A l fred H i tchcock ( pages 24 and 44 ) : @ 1 96 7 by Fran­
cois Tru faut , revised @ 1 984 by Francois Trufaut , reprinted i n the U.S.
and C anada w i t h the permission of S imon & Schuster, and elsewhere
w i t h the k i n d permission of Seeker and Warburg .
While Hawthorn i n the West of Irela n d b y Eavan Boland (page 41):
@1989 by Eavan Boland, original ly published in The New Yorker,
reprinted w i t h t he kind permission of Eavan Bolan d and The New
Yorker .
Excerpt from Seascape by Stephen Spender (page 42) : from C O L­
LECTED POEMS 1 928-1 985 by Stephen Spender, @ 1 946 by S t e p h e n
Spender, repri nted i n the British Commonwealt h (excluding Canada)
with the k i n d permi ssion of Faber and Faber Limited, and elsewhere
with the permission of Random House, Inc.
XVII
•
Composition with Blu e and Yellow, (1935) by P iet Mon drian ( plate
3): Hirshhorn Museum and Sculpt u re Garden , Smit hsonian Instit u­
tion, G i ft of Joseph H. H i rshhon Foun d at i o n , 1 972, ( photograph by
Lee Stalsworth ) ; reproduced w i t h t he k i n d permission of H i rsh hon
M useum and Sculptur e Garden .
•
Quotation from Piet Mondri an ( p age 43 ): @1986 by Harry Holtzman,
reproduced w i t h the k i n d perm ission of M. Holtzman and G .K. Hall
& Co.
•
L e Cauter (Tea T ime), ( 1 9 1 1 ) by Jean Metzi nger ( p l ate 4 ) : P h iladel­
phia M u seum of Art : The Louise and Wal ter A rensberg Collection ,
reprod u ced w i t h the k i n d permission of P h i l adel p h i a M useum of A r t .
•
Figures from P. A . Kolers ' Aspects of Motion Pe 1·ception ( page 99 ) :
@ 1 972 by Pergamon P ress , repri nted w i t h the k i n d perm i ssion of Perg­
amon P ress.
•
Figu re from A . R. L ur i a's Cognitive Developm ent ( page 103): @ 1 976
by the P resident and Fel lows of H arvard Col l ege, repri nted with the
kind permission of H arvard Un ivers i ty P ress .
•
Q uotations and igures from T . G . Evans' " A Program for the Solution
of a Class of Geometric- A n alogy I n telligence-Test Questions" ( Section
10.4.1): @ 1 968 by MIT P ress, reproduced here with the kind permis­
sion of M I T P ress.
Prologue
This book i s about metaphor and cogni t ion , an d about the relations h i p
between the two. W h i le t here remain s much dispute about w hat metaphor
is, how it works, and what role i t plays in cognition , t here seems to be general
agreement that metaphor i nvolves two objects or s i tuations and some kind
of t ransference from one object or situation to the other. One object i s
referred to a s t he topic, t he tenor, the primary subjec t , or t he t arget domai n ,
and t h e other object as t h e vehi cle, t he secon dary subjec t , o r t h e source
domai n . For instan ce, in "The chairperson of the meeting plowed t h rough
the agenda," the target is the meeti ng, the source is the act of plowing, and
the transference exists i n being able to describe the meeting meaningfu l l y by
using the seman t i cally distant term of ' p lowi ng'.
I t i s also gen erally accepted that metaphors are not l i m i ted to individual
words, or even p hrases, but a whole poem or a novel can be metaphorical as
well . A nyone who h as read Dylan Thomas ' Th e fo·ce that th m ugh the g·een
fuse drives th e lower, Mel v i l le's Moby Dick, or Hem i ngway's Th e Old Ma n
a n d the Sea would know that these works become much more interesti ng
and i nsightfu l when given an i nterpretation other than the l iteral one. H ere,
while the source i s explicitly given , the target is not speciied at all ; it is
up to t he i ndivi dual reader to ind her own t arget and carry out her own
i nterpretation . The point i s that one needs a target domain and a meani ngful
i n terpretation i n order for the work to become metaphorical . For i nstance,
I once read an i nterpretation of Frank Baum 's Th e Wizm·d of Oz, a popular
fantasy book t hat described i t as a sati re of the A merican pol i t i cal situation
existi n g at the t ur n of the centu ry, when the book was first published.1 To
get the satirical fet, however, the characters of the book need to be given
an appropriate i n terpretation i n t hat context-t hat the Wicked Witch of th e
East be identiied with t h e i n an c i al i nst i t u t i ons of the N or t h east e n U.S., t h e
T i n Woodman be identiied wi th the industrial workers, t he Emerald City
be identiied w i t h Washington D . C., the cap i t al of the U.S., and so o n .
Classiication of metaphors themselves i s somewhat controversial. B u t for
our purpose, we need only make a d i s t i n c t i o n between conve n t i on al m e t a­
phors, similarity-based met ap hors, and si milarity-creating metaphors. Con­
ventional met aphors are t hose met aphors t h at are so much a part of everyday
speech that t hey seem hardly metaphorical . The 'plow ' me t apho r mentioned
above is a case i n poi n t . Conventional metaphors are e v i den ce to the fact
11 remember reading it in the Sunday edition of the Bufalo News sometime in December
1987
2
Metaphor and Cognition
that metaphoric t ransference is not someth i ng conined to poetry and lit­
erat ure, but is very m u ch a part of our everyd ay speech . T hey h ave been
studied extensively by George Lakof and his colleagues [Lakof & Joh nson
1 980; Lakof & Turner 1989]. In thi s study, however, convent ional metap hors
play only a minimal role. My main concern is not so much w i t h why our
everyday concepts are structured i n one way rather t h an another, but w i t h
how n e w concepts a n d mean ings emerge.
S i m i l ari ty-based metaphors invite the reader to make a comparison be­
tween the source and the target , as the t ransference of meaning is based on
some existing s i m i larity between them. For instance, in "The sky i s cry ing"
the reader i s d rawn into comparin g the rain falling down from the sky w i t h
t h e tears falling down from the eyes o f a person . T hese metaphors h ave been
captured in what is commonly known as the compari son theory of metaphor
[Henle 1 958].
In a similarity-creat i ng metaphor, however, t here are no similarities be­
tween the source and the target when the metaphor is irst encountered .
Yet , after the metaphor is assi m i lated , ( i f i t i s assimilated at all , ) t here are
sim i l arities between the two. T hus, the metaphor c rea t es t he s i m ilarit i es
between the source and the target . To appreciate the force of a s i m ilarity­
creating metaphor, t ry the following experi ment . For each of t he fol lowing
pai rs of words, t ry to enumerate the simi lari t ies between the referen t s of
the two words: snake-eel, dog-computer, snowsto1·m -autom ob ile, cat-fog, fog­
clock. For i n stance, for the snake- eel pair, you might say t h at t hey are both
long, slithery, etc. I f you cannot ind any similarities , j ust say so. G i ve
yourself, say about ive m inutes, for each pai r .
Now consi der Carl San d b u rg's beaut i fu l poem Fog:
The fog comes
on l i t t le cat feet .
It sits looki ng
over harbor and c i ty
on silent h au n ches
and then moves on .
After readi ng the poem , the fog at once appears similar to the cat! T hey
both creep up on you ever so silently. Moreover, I would be very surprised i f
t his s i m i larity was inclu ded i n your initial comparison o f fog and cat, u nless
you peeked ahead, or you were al ready fami liar with t he poem, and the words
fog-cal remi n ded you of i t . Thus, i t wou l d be appropriate to acknowledge that
Prologue
3
t he metaphor created the s i m i larit ies . I f you need more convincing to accept
the phenomenon of creation of similarity, I present m any more examples i n
C h apter 2 , since similari ty-creati n g metaphors are m y pri mary concen i n
th i s book .
The existence of s im ilarity-creating metaphors has not been u n i versally
acknowledged . The reason i s that i n a s i m i l ari ty-creati n g metaphor, t here
are always s i m i lari ties after t h e metaphor is p resented and u nderstood. So if
a person is presented w i t h a metaphor, and then asked to explai n her under­
stan d i ng of i t , she would i n variably gi ve a similarity-based accoun t . But t h is
fails to address the questi on o f whet her the simi larit i es were t here before the
metaphor or not . ( O ne notable exception i s the i nteresting study of Cam ac
and G lucksbu rg [ 1 984] , w h i ch is d iscussed i n C h apter 2 . ) Consequently, i t
might n o t b e surprising t hat almost all t h e research on metaphor i n cogni­
t i ve science has been in the p u rsu i t of similarity-based metaphors. E m p i ri cal
studies and theories h ave sought to articulate exactly w hat k i n ds of si milari­
t i es u nderlie metaphors [ Gentner, Falkenhai ner, & S korstad 198 7 ; Gentner &
S tuart 1 983; M algady & Johnson 1 980; O rtony 1 979] . Following these leads ,
the comp u t at i on al models h ave focused either on b o w t h e s i m i l arities might
be comp uted , gi ven a source and a target ( Weiner 1 984 , 1 985; Fass 1989); or
on how the similar i ties might be used i f t hey are expl icitly given (Carbonell
1 982; Marti n 1 988) . S i mi larity-creati n g metaphors have been left out i n t he
cold .
I nteres t i ngly, however, a n approach to metaphor i n t he phi losoph ical
trad i tion-an approach t h at is widely known as the intera ction th eor--w as
created expressly to account for similarity-creating metaphors. Originat i ng i n
I.A. Richards' Philosophy of Rhetoric, the i nteraction t heory owes its present
for m l ar g ely to the works of such scholars S Max Bl ack [1962; 1979], Carl
H ausman [ 1 983 ; 1984; 1989], and Paul Ricoeur [1977]. Hars h l y cri ticizing
the c o mp ari s on theory of metaphor for its i n ab i lity to address t h e creation of
s imil a r i t y interaction theory p ro p oses that every m et ap h or involves an i nter­
action between i ts source and its target, a process in w h i ch the t arget (and
possib ly the source) i s reorganized, and new s i m i l arit ies between the source
and the target emerge. This account, w h i c h is exam i ned i n Chapter 3, is,
however, quite vague, and even paradoxical at times.
,
For i nstance, it neither pins down what exactly thi s mysterious 'interac­
tion' i s , nor speciies exactly how the new similarities e me r ge There is a
crucial problem here: the creation of similari ty i s obviously not an arbitrary
p ro cess I f it were, then anyt h i ng would be mean i n gfu l and there would be
n o way to maintain a disti n ction between wh a t is genui nely metaphorical
.
.
4
Met aphor an d Cognition
and w h at i s certai nly nonsensical . B u t t hen what constrains t h is creation
process? I t cannot be the similarities between the source and the target , for
this would t u rn the i n teraction theory into nothing b u t a variant of the com­
parison theory, and all the crit i cism that the interactionists d irected against
t he comparison theory can be d i rected at t hemselves . So t he question i s s t i l l
unanswered: w here do t he created s i m ilarities come from ?
A few scholars have tried to elaborate the i nteraction t heory and t ackle
some of t hese i ssues head on, but again , by and large, either t hese elabo­
rations are quite vague t hemselves , relying on metaphors and analogies to
communicate the key concepts involved i n their explanations, or t hey t urn
the i nteraction theory into a variant of comparison theory. H ausman [1983],
for instan ce, postulated ' u n iqueness ' and 'extra- l i nguistic' con d itions to ex­
plai n how metaphors can create new mean ings and similarit i es , but t hese
conditions t hemselves are not spelled out w i t h much clarity. Verbrugge [ 1 980]
proposed t hat a metaphor works by ' transform i ng' the target i nto the source,
t hereby mak i n g it s i m i l ar to the source, but then i t i s not speciied exactly
what this t ransformation p rocess is, and what constrains i t so t h at arb i trary
transformations are ruled out . Tourangeau and Sternberg [1982] proposed
a 'domains-interaction' view , w h i ch is p urported to be a ' more speciic for­
mulat ion of the i nteraction view , ' but w h i ch , i n essence, turns out to be t he
comparison view i n d i sguise, because i t assumes an u n derlying analogy to
be the basis of every metaphor. A notable except ion is provided by K i t t ay's
[ 1 987] perspec t ival theory, which comes quite close to prov i d i ng a reasonable
explanation of the creation of similarity. B u t , for the most part , we see t h at
the i nteraction t heories have s t i l l retai ned t heir fuzzy character, which m ay
wel l be the s ingle most i mportant reason cogni t i ve science researchers h ave
shied away from them [ Waggoner 1990].
There i s also a minor i nconsistency i n the interaction view t hat is some­
times overlooked. At one place in his classic essay 'Metaphor,' Black re­
m arked, "If to call a man a wolf is to put h i m in a s pe c ial l ight , we must
not forget t h at the metaphor m akes t h e w ol f seem more human t han he oth­
erwi se wou l d . " This clearly implies a symmetry i n the interaction between
the source and the t arget However, at o t h er places in his d is c uss i on t here
is a clearly implied as y m m e tr y For e x am p l e in his later essay 'More about
Metaphor,' in a deeply insightfu l sect ion t i t led 'Thi nkin g in Metaphors ' Black
cons iders how different concepts can organize the igure of the S tar of David
differently. I n organizing the igure of Star of D av i d as t h ree parallelograms
wi t h t heir axes one hundred and twenty degrees apart , the process does not,
at the same t i m e organize our concept of parallelograms as a part of t he
igure of the Star of Dav i d-t hat i s , t he parallelogram does not app e ar to be
.
.
,
,
Prologue
5
Star-of- D av i d-li ke after the i nteraction.
W h i le a few i nteractionists, such as Verbrugge an d 1\it t ay, see the i n ­
teraction view a s essentially asymmetr i cal , others , Hausman for i nstance,
have emphasized the symmetry aspect of the i n teract ion . H ausman [1989,
p . 67] went as far as to argue that one need not disti ngu ish between the
source and target of a metaphor. A result of this con fusion bas been that
many scholars now accept the symmetry property to be a key aspect of the
i nteraction t heory [Waggoner 1990], an d some, such as Lakoff and Turner
[1989, pp. 131-133], use arguments agai nst the sym metry property as a way
to d iscred i t the whole i nteraction theory. B lack 's i n s i ght fu l observations i n
' T h i n k i ng i n M etaphors ' seem t o have been grossly overlooked .
Then t here i s the problem of explai n i ng the role of metaphor i n cogni ­
t i o n . I t h as been recognized for quite some t i me now t hat metaphor i s not
j ust a phenomenon of language, but pervades all aspects of cogn i t i o n . t
h as been claimed that metaphors play a key role i n lean i ng and education
[ Holstein 1970; Petrie 1979; Sticht 1979], and t hey are an i nval uable ai d to
p roblem solvi n g [Schon 1963]. In the early stages of form ulat i ng a scient i i c
t heory, met ap hors are often i n d ispensable [ Gruber 1978; Hesse 1980; M i l ler
1978; Roth bart 1984, p p . 611-612]. Religious schol ars have emphasi zed time
and again t h at religious symbols ( scriptures, rituals, etc . ) derive thei r sig­
n i i cance due to t hei r metaphorical n at u re, and should not be taken l i terally
[ B rown 1983; Soskice 1985; T i l l i ch 1961; Wheel wright 1954]. For i n stance,
the Christ i an ritual of tak i ng communion i s meani ngful only i f one u nder­
stands it metaphoricall y. Hone does not bel ieve in the transsubstant i ation
doctrine, then neither i s the bread li terally the body of the Christ , nor i s
the w i ne l iterally H i s blood, a n d a metaphorical i n terpretation i s requi red
to render the ritual meani ngful . Even if one adopts the transsubstan t i ation
doct r ine, a metaphorical i nterpretation i s st i l l requ i red to att ach signiicance
to what would l iterally be a can n i b alisti c act. The wel l- k nown mythologis t
J oseph Campbell [1949; 1986; 1988] argued t h roughout h i s pol i i c career
t h a t m y t h s are met aphorical ways of capturing the very essence of the expe­
rience of l iving, and, with an appropri ate metaphorical i n terpretation , various
ancient myths are still as relevant as they m i ght have once been.
Besides literature, met aphors pervade various other art forms. The ab ­
stractionism p revalent i n contemporary arts parti c u l arly requ i res a metaphor­
i cal i nterpret at i on for a work to be meani ngfu l . W i l lem De Kooni ng's Exca­
vations, Barnett Newman 's Achilles, and J ackson Pollock 's Cathedral are all
examples of pai n t ings that req uire metaphorical i nterpretations to be u nder­
stood. Joh n Cage's compos i t ion 4'3311 which is essentia.ll y 4 m i n u tes and 3 3
6
Met aphor and Cogni tion
seconds of silence, is another case in poi n t . ( See Cage's ' Lectu re on Not h ing, '
i n Cage [ 1 96 1 ) , p p . 1 08-1 27 . See also Rowel l [ 1 983] for the role of metaphors
in music, including an i nterpretation of 4 ' 33" [p. 220] . ) Fox [ 1 982) , in i n t ro­
ducing the works of six modern sculptors, Vito A cconci , S i ah A rmaj an i , A l i ce
Aycock , Lauren Ewing, Robert Morris , and Dennis Oppenheim , h as further
emphasized the reli ance of contemporary art on metaphor to commun i cate i t s
meaning. Whi t tock [ 1 990] has analyzed several diferent types of metaphors
i n feature i lms, m any of w h i ch were consciously i nt roduced by the directors .
For i nstance, the classic shower scene from A l fred H i tchcock 's Psycho i s seen
as an act of spirit ual cleansing: Marion C rane ( J anet Leigh ) , hav ing decided
to go back and return the money she absconded w i t h , is not j ust washing
away her body, but i s also ridding herself of the gui l t [Wood 1 989, p . 1 46] .
This identiication enhances t he shock of her subsequent murder a great deal
[ W h i ttock 1 990, p. 53] .
Fi nal ly, ant h ropologists have poi nted out t hat many of our own social,
cultural and moral val ues result from the metaphors t h at are p revalent in our
society and cult u re [Kempton 1 987; Lakoff & Kovecses 1 98 7 ; Q u i n n 1 987;
Reddy 1 979; Schon 1 9 7 9 ; Tuner 1 9 74] . For i nstance, Q u i n n 's study found
t h at t here are different metaphors u n derlying t he modern A merican concept
of marri age: 'a manufactu red product , ' 'an ongoing jouney, ' 'an i n vestment , '
a n d s o o n . She also fou n d t h at a n i n d i v idual's perception o f whether h i s o r
h e r marr i age i s a success or a fai l ure, whether t here i s some problem facing
the marriage, and i f so, how i t might be corrected , etc. are all determined
by the u nderlying metaphor.
G i ven t h i s overwhelm i n g evidence for the variety of roles metaphors play
i n cogni t i o n , one would expect a t heory of metaphor to shed some light on
what it i s about metaphor that m akes i t pervade so many d i ferent fac e t s
of cogn i t i on . In fac t , the evidence is so strong as to suggest that a t heory
of metaphor should perhaps be set within a fr am e work of cogni t ion . That
i s , t h ere should be a general account of cogni t i o n , and metaphor should be
presented as one of the mechanisms u sed in cogn i tion.
J n exploring t h i s hypothes i s , w e i n d that there exi s ts a problem i n cogni ­
t i on that is remarkably parallel to the problem posed by similarity-creati n g
metaphors . T hi s problem has to do w i t h a v i ew of cogni t i on according to
w h i ch the world view of a cogni t i ve agent does not relect some pre-exis t i ng
structures i n the external world , but is created by the cogni t i ve agent . Yet ,
t h i s creation is not arb i trary, but i s constrained b y t h e external world . Per­
haps not surprisi ngly, this view is referred to as th e interaction view of cog­
nition, sin c e it sees cogn ition as a process of i nteracti o n between a cognitive