Vous êtes sur la page 1sur 21

In Search of Verbal Mimesis

Author(s): William Wimsatt

Source: Yale French Studies, No. 52, Graphesis: Perspectives in Literature and Philosophy (1975),
pp. 229-248
Published by: Yale University Press
Stable URL: http://www.jstor.org/stable/2929756
Accessed: 26-06-2015 23:03 UTC
Linked references are available on JSTOR for this article:
You may need to log in to JSTOR to access the linked references.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

Yale University Press is collaborating with JSTOR to digitize, preserve and extend access to Yale French Studies.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

In Search of Verbal Mimesis*

4.016 In order to understandthe essential

nature of a proposition,we should consider
hieroglyphicscript, which depicts the facts
that it describes.
And alphabeticscriptdeveloped out of it
without losing what was essential to depiction.
LUDWIG WITTGENSTEIN, TractatusLogicoPhilosophicus,trans. D. F. Pears & B. F.
McGuinness(1961), p. 39.

In searchof a trulyverbalmimesisor iconicity,

I. Introductory.
This essay
mimesisor iconicity.
and especiallya trulyverbo-visual
noveland not evento push
attemptsnot to demonstrate
any theme to an ample illustration,but only to sort out certain
more or less well known relationsbetween words and what they
referto and in the process of sortingperhaps to arrive at some
improvedrealizations.In an essay publisheda few years ago under
the title "Laoko6n: An Oracle Reconsulted," I alluded to four
generic areas conspicuous in modern studies of visual art and
of visual art,notablyRenaissance paintliterature: 1. Interpretation
ings, in literaryterms; 2. the parallel or analogy between visual
and literaryart (ut picturapoesis, ut poesis pictura) as it flourished
notablyin the late Renaissance and Baroque eras; 3. the melange
or harmoniousunion of more or less analogous arts in various assembled arts (Gesamtkunstwerke)-as song, drama, opera; 4. the
question of medium-words compared to music and pictures1 It is arbitrarythat the presentessay
resemblancesand differences.
* It is our sad dutyto reportProfessorWimsatt'sdeath,whichoccurred
as this issue was going to press. We wish to expressour gratitudefor his
graciouscollaborationon this special issue. -Managing Ed.
1 "Laokoon: An Oracle Reconsulted,"in Eighteenth-Century
Studies in
Honor of Donald F. Hyde, ed. W. H. Bond (New York: The GrolierClub,
1970), pp. 347-64.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

Yale FrenchStudies
aims to moveas strictly
as possiblewithinthe limitsof the last
ofthosetopics,No. 4, medium-butarbitrary
onlyin thesensethat
it is alwaysarbitrary
to chooseany topic,in preference
to others.
TopicNo. 1,interpretation,
is a modernaspectofattention
to No. 2,
analogy.Nos. 2 and 3, analogyand melange,can logicallybe put
to one side here,forthe sake of our gettingahead withNo. 4,
medium,thetopicof the essay.Theywillintrudebriefly,
at pointswheretheyseem to me to makethe closestconnection
withwhatI am tryingto distinguish
I am taking"mimesis"in a narrow,intensified,
or saturated,
and I hopeusefully
precisesense.In theAristotelian
of a humanaction.But thisis a verybroadsense.
is an "imitation"
The mimesis,
or representation,
so faras itis notscenicor theatrical,
inheresin or dependson verylargemeasuresof expression
and of
Aboutexpressiona fewwordslater.Let me stresshere
the difference
between,on theone hand,mimesis(representationor iconicity)
and, on the otherhand,reference.
We will say
refersto the murmur
of a humanbeing,
thatthe word"murmur"
of running
water, hivingbees,
leaves,but thatalso,
in a qualifiedsense (See below III a) the wordsoundslike or representsthesenonverbalsounds.We will say thattheword"dog"
refersto a dogbut not,in anyveryconvincing
sense,thatit sounds
like,looks like,or feelslike a dog. The unquestionable
(if rather
parallels analogues
other arts2 and the fact of melangeor Gesamtkunstwerk,
of the arts' (grosslyevident such phe"perennialsyncretism"
2 JamesD. Merriam,"The Parallel of the Arts: Some Misgivingsand
a Faint Affirmation,"
Parts 1 and 2, Journalof Aestheticsand Art Criticism,
XXXI (Winter1972), 153-64,and (Spring1973), 309-22; JeanH. Hagstrum,
"Verbal and Visual Caricature in the Age of Dryden, Swift,and Pope,"
England in the Restorationand Early EighteenthCentury,ed. H. T. Swedenberg, Jr., (Berkeley: Universityof California Press, 1972), pp. 173-95;
Ronald Paulson, "The Pictorial Circuit and Related Structuresin 18thCenturyEngland," The Varied Pattern: Studies in the 18th Century,eds.
Peter Hughes and David Williams (Toronto: A. M. Hakkert,Ltd., 1971),
pp. 165-87.
3 Edward Stankiewicz,"StructuralPoetics and Linguistics,"forthcoming
in CurrentTrends in Linguistics,XII, ed. Thomas Sebeok.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

nomenaas drama,opera,and song)relateto verbalmimesismainly,
perhapsonly,in the verybroador Aristotelian
Languageas spoken,or speakable,maybe, but neednotbe, for
thepurposeof myexposition,
conceivedas theprimary
In the modernera the languageof literature
is usually
experiencedwiththe visual intervention
of graphemes.
We may
even suppose,if we wish,thatthe literarystatusdoes not arrive
4 In anyevent,we are talking
about a real distinction
and a difference.
A firstmajorsortingof
can plausiblydividetheminto1. thegraphemic
and 2. the phonetic-lexical.
(The reasonsforthe hyphenation
the lattertwo termswill appearas we proceed.)My sortingout
pursuesfirstsome topics under heading1. the graphemic(the
simpler,and perhapsmorespeciouslyattractive
if we have strong
eidetic propensities),
and then
2. the more treacherously
And thus:
II. a. The alphabet(an exercisein deconstructive
humanbeing,especiallythehumanbeingas child,or as artist,often
hoverson the vergeof reversing
the transitions
of his technical
history-theprimitiveinventions.Childrenhunt alley cats with
bowsand arrows,or withspears,pointedstickshardened
in flame.At summercamptheyalmostmakefirewitha bow drill,
a cedar slab, and shavings.A few expertTexan hobbyistsflake
excellentflintblades,as narratedin an exciting
I One
by the HoustonjournalistH. Mewhinney.
may well conceivethat he and his friendscould survivein the
A verysimilarflirting,
mountainsafteran atomic Armageddon.
on an invisiblethreshold
may said,I believe,to occur
in the presenceof certainchildishriddles.
A B C D goldfish.
L M N 0 goldfish.
O S A R goldfish.
4 Plato, Phaedrus 274B-278B; E. D. Hirsch, Validityin Interpretation
(New Haven: Yale UniversityPress, 1967), pp. 248-50, a review of HansGeorg Gadamer, Wahrheitund Methode, 1960.
5 Austin: Universityof Texas Press, 1957.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

Yale FrenchStudies
A giftedartistof the grotesqueand comic,WilliamSteig,throws
a littlebook of
offa byproduct
of his savageprimitive
C D B. A lessonin naturalhistory.Do you see it? The bee-on
theflower.I M A U-MB-N. U R N N-M-L. The bratlooksdown
smuglyupon his puzzled pooch.6A clipped,pinched,deformed
no doubt,as an exagdialectof AmericanEnglish(recognizable,
is rendered
gerationof whatcan actuallybe heardin Brooklyn)
lettersused not as actuallettersbut onlyas the names of letters.
We confront
the firststage in a morevivid realizationthatthe
of our alphabetcarrynames,whichare sounds,lively,
chargedwitha capacityforjoking,puns. Certainlettersin our
alphabetcan be used in a fictitious
value thatis sunk in the structure
a four-fold,
of graphemic
1. "I" (ai) is a visual or auditoryword, meaning"ego."
2. "I" (ai), the word meaning"ego," is a homophoneof a word "eye,"
meaning"organ of sight."
3. (ai) is the name of a grapheme-"I," and a homophoneof both 1. and 2.
4. "I" (the grapheme)is a writtenletter,the cue to utteringa syllable-as
in "i-o-nize,""i-o-dine."
5. "I" (the grapheme)is a writtenletter,the cue to utteringcertainless
than syllabicphonemes-as in "Ides," "in."

instanceof "B."
Or the slightlydifferent

"Be" (bi) is a monosyllabicword meaning"esse."
"Bee" (bi) is a homophoneof "be," and a word namingan insect.
(bi) is the name of a grapheme,"B," a homophoneof "be" and "bee."
"Be-" (bi) is a syllable,as in "be-fore."
"B" (the grapheme)is a writtenletter,the cue to utteringa less than
syllabicphoneme-as in "bid," "crib."

William Steig, CD B! (New York: Simon & Schuster,Inc., 1963),

pp. [1], [11].


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

William Wimsatt
Our writtenlanguage,the borrowedEnglishalphabet,did not evolve
in those ways-"eye" to "I" in "Ides," or "bee" to "b" in "bid."
The slight difficulty
one may experiencein holdingapart the four
or five values just now listed, the modest explosion of surprise
and fun in the inventionsof William Steig (I F-N N-E N-R-G.
-Be resolute in pronouncingthe names of the
letters,not the lettersthemselves)I may,however,help our realization of what happened when in the proto-Semiticto Phoenician
sequence, something like Egyptian hieroglyphics(pictures with
names) workedas puns or rebusesforwords and syllablesand then
as names of letters-when for the only time in human historya
sufficientalphabet was invented,in short when writtenlanguage
arrivedat an adequately intimateand fittedrelationwith the prior
spoken act. The alphabet-via the punning syllabic graphemebecomes the firstsystemof graphemeswhich are not pictures of
If we try next to imagine how the writtenlanguage began as
pictures,we will realize, I think,an even greater early step by
which it depictorializeditself.This must have occurredas soon as
any picturebecame in fact a pictogram-that is, when any definite
idea became attached to it. From then on, for practical purposes,
much of the picture would be irrelevantand mighteven be misleading. The supposed pictorialand poetic value of modern Mandarin characters,dreamed of by Fenollosa and Pound, commands
7 C D B!, pp. [36], [29].

8 David Diringer,The Alphabet,A Key to the Historyof Mankind,3d

ed., rev. with ReinholdRegensburger
(New York: Funk & Wagnalls,1968),
I, 160-69; Hans Jensen,Sign, Symboland Script,An Account of Man's Effortto Write,trans.GeorgeUnwin (London: GeorgeAllen and Unwin Ltd,
1970), pp. 50-53,264, 270, 280-83. Some of the "alphabetologists"(including
Diringer,I, 168) have arguedthat the names of the Semiticletterswere not,
like Egyptianrebussyllables,a punningpartof the alphabeticalor acrophonic
evolution,but were added only afterthe letterswere established.The older
and more commonview seems the easier to imagine.
All the consonantnames of the English alphabet are-unlike those of
the Hebrew or Greek-single-consonantmonosyllables,
and theyare divided
between eleven consonant-initials
(bi), (di) and eight consonant-terminals
(el), (em). These are circumstanceswhichgreatlyfavorMr. Steig's agreeable


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

Yale FrenchStudies
I In theoppositescale,thename
of a picture,"The Blue Boy,""The Age of Innocence,""Mrs. Siddons as the TragicMuse," does not represent
the wholemeaning
or the aestheticvalue of the picture.The cave bisonsof Lascaux
less formalseemmoreconcrete(ifonlyby clustering,
or herding),
ized, certainlyless conventionalized,
than the charactersof any
language.(We can witnessinstancesof amare stillused
biguoussurplustodayin contextswherepictograms
in highwaysigns,forinstance,wherethe arrowon the THRU
,but thehighwayTRUCK signpointsin one direction,
fortruckpointsin theopposite, )
abstract,and the difference
and not
ideograms difficult
On the one hand,the Chinesecharacter
fortree , theRomannumbersI, I, III, and theArabicnumbers
1, 2, 3. On theotherhand,thebarbed-wire
ox, thesignampersand
& (and-per-se-and),
and theArabicnumbers
as a visualword,or in so far
6, 7, 9. Whena picturecan function
as it does function
as a visualword,it has lost mostor all of its
needto be a picture.The nextstagein its abstraction,
pictureto formalized,
even arbitrary,
ideogram,is farless radical
than the transition
frompictureto pictogram.
The same kind of
and fixationmaybe observedin the difabstraction,
(birdor animalcalls)andtrueverbal
(see belowIII a). The depictorialization
et each stageaccordedwiththe deepestnatureof verbaldiscourse
of the alphabetfroma singleinand explainsthe wide diffusion
to the manifold
of fossilized
ventionand its superiority
II. b. Melange (Cum Pictura Poesis). There is a certain sense in
of themodernworld,especiallythepoetry,
featureof the Gutenberg
era is
gone heavilyvisual.A prominent
9 James J. Y. Liu, The Art of Chinese Poetry (Universityof Chicago
Press, 1962), pp. 3, 6, 14, 18-19; Achilles Fang, "Fenollosa and Pound,"
Harvard Journalof Asiatic Studies,XX (1957), 213-38,esp. 215-17.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

William Wimsatt
the linearand paragraphiclayoutof the printedpage-with a special
visual emphasis for lines and masses of verse. Modern poetry,or
poets collectively,seem to have been bent on regainingthe lost
thicknessof the pictorialword-but perhapsoftenin mistakenways.
We most oftenconsult our poetryin graphemesand oftenenough
in some sort of fancyor sophisticatedkind of graphemes.Much
pictorialor quasi-pictorialmanipulationhas been invited and has
taken place. It remains,however,an importantquestion whether
the visually aesthetic propertiesof graphemesare always in fact
very deeply tied in with the phonological-semiotic
are the reason for the occurrenceof the graphemes: whetherin
fact we encounterillustrationsof the supposed principleut pictura
poesis-or not rathercertainingeniouscontrivancesof a game that
ought to be called cum picturapoesis. This skepticalquestion,with
its implicitlysevere answer, I conceive as embracinga wide spectrum of speciously varied but closely akin phenomena of visual
melange, suggestionsof a living alliance between very disparate
media: viz., forone thingand mainly,all the "shaped" poems both
the more and less ancient and the modern, the technopaignia10
of the Greek Anthology(stigmatizedas a "Species of false Wit" by
Addison in Spectator 58), the formulationsin the Art of English
Poesie by Puttenhamin 1589, the freshrealizationsby Herbert in
typographicalwings and altar, and in our own time most notably
the Calligrammesof Apollinaire-e.g., the poems shaped like a
heart,a crown,a mirrorreflectingthe name of the poet, and that
which sprinkles the words in a shower of rain-and feeling(Il
pleut). 11All those are ways of disposingthe graphemesof the poem
itselfinto shapes that have some kind of superficialor cold relation
*10 For the poems, see the Greek Anthology,
Book XV, nos. 21-27: the
Axe, Wingsof Love, and Egg of Simias, the Altar of Dosiadas, and that of
Besantinus,the Pipe of Theocritus (Loeb Library,V, 127-33). The term
technopaignion,as applied to a shaped poem, is apparentlypost-classical.
The Technopaegnionof Ausonius (Works,Bk. XII) is a collectionof short
pieces illustratingthe trick of ending each verse with a stoppingmonosyllable.
11 GuillaumeApollinaire,Calligrammes,
Poeimesde la paix et de la guerre
(1913-1916),Preface Michel Butor (Editions Gallimard,1966), pp. 58, 64;
a second shower,p. 159.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

Yale FrenchStudies
withthesubjectof thepoem."The Poetry,"saysAddison,"was to
contractor dilateitselfaccordingto the Mould in whichit was
cast.In a Word,theVerseswereto be crampedor extendedto the
Dimensionsof the Frame that was preparedfor them; and to
undergothe Fate of those Personswhomthe TyrantProcrustes
used to lodgein his IronBed." (Variousexamplesinvolvearbitrary
disjunctionof letters-word-shredding-or
envelope,top-and-bottom,orderof lines.The typographic
dispersalsof E. E. Cummings
run the gamutof such devices.)12 Anothermaintypeof melange
introducesmoreor less subtly,and even beautifully,
variousadditives,the illuminated
capitalsand marginaland interlinear
thewoodcutsin Renaissanceemblem
books ("picturesthatforthe page atone"),the tendrilsand bird,
in amongthe
animal,human,and fairyfigures
wordsetchedat the centerof Blake's illuminated
plates.13 Somewherealongsidesuch instancesof the cum picturaand the fully
of wit
shaped wholepoem,we may perhapslocate a department
termed"Word Play," to be consultednowadaysappropriately
of devices,
the magazinePlayboy-where,throughan assortment
like autonyms.Thus
wordsor phrasesare made into something
s (
the word touc H down, and the word b
m. 14 A
Germanpoem (by ChristianMorgenstern)
about two funnelsthat
tapersin shorterand shorter
lines; theconcluding
phraseis und so weiter,u.s.-the invertedV in the centerof the


W beingadmiredas an imageof theopeningof the spout.15"Conof thelinguistic

in favor
crossof visual order,like repeating
wallpaperor carpetpatterns,
12 W (New York: Horace Liveright
Inc., 1931) showsthe extremesof two
opposite devices, "calligram-or picturewriting... word scramble... cryptogram" (Malcolm Cowley, "Cummings: One Man Alone," Yale Review,
LXII [Spring1973], 343).
13 Roman Jakobson,"On the Verbal Art of William Blake and other
Poet-Painters,"LinguisticEnquiry,I (January1970), 3-10, findsstructural
analogies betweenBlake's "InfantSorrow" and its illumination.
14 Robert Carola, "Word Play," Playboy, XIX (February 1972), 161;
XVII (April 1970), 163.
15 Quoted by John Lotz, "Elements of Versification,"
in Versification:
Major Language Types,ed. W. K. Wimsatt(New York: New York UniversityPress, 1972), p. 18.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

William Wimsatt
word puzzles, magic squares, or computerprint-outs,are a more
modern experiment; the real verbal element all but disappears.16
II. c. But it would be unfair,and it is not my purpose, to end
this section withoutindicatinganotherkind of visual possibilityin
graphemes.There are featuresof language structurewhich can be
accentuated by visual patterns, which perhaps fuse with these
patterns,yet also lie deeper in the linguistic grain. The simple
signals of a paragraphindentationand that of a capital letter are
easy graphicinstances.More complicated,but indubitableinstances
in English occur conspicuously in the kind of rhymed closed
pentameter couplets written by Alexander Pope. The stair-like
progressionsof the Russian poet Mayakovsky,closely related to
Slavic traditionsof oral and musical emphasis,belongin an adjacent
area. 17 We shall wish to distinguishdegrees of visual imposition,
degrees of dominance between the phonological-semioticand the
graphemic-and hence I believe degrees of depth and vitalityin
the union, and a gradationor shading of superficiality
toward the
arbitrarinessof the "shaped" poem. But the mattermay now convenientlyremainhangingfora while,untilwe reach the end of some
sortingin our second main series,the phonological-semiotic.
III. a. Onomatopoeia. C. S. Peirce, followedby Roman Jakobson,
has divided iconic (or mimetic)expressioninto image and diagram.
The latter inheres in relations,patternsof words (above III c and
below V); the former,image, is atomic and simple.18 Yet image
lb See the pot of FORSYTHIA thatsends up spraysof its repeatedletters,
the F spray,the 0 spray,etc., black ink against yellow,on the cover of
Concrete Poetry: A World View, ed. Mary Ellen Solt (poet of the Forsythia), Bloomington:Indiana UniversityPress, 1970.
17 See them displayedin Russian and English in VladimirMayakovsky,
The Bedbug and Selected Poetry,ed. PatriciaBlake (New York: The World
Publishing Company, 1970), pp. 173-235, "Brooklyn Bridge" and other
poems. These visual structuresare a way (perhapstoo easy) of separating
and emphasizingphrases,parallels, contrasts.Mallarm6's Un coup de des
jamais n'abolira le hasard (1897) combines steppingwith simultaneousor
shreddedand interspersedmessages in several kinds of type.
18 Collected Papers of Charles Sanders Peirce, ed. Charles Hartshorne
and Paul Weiss, vols. I and II in One (Cambridge: The Belknap Press of
Harvard UniversityPress, 1960); vol. II, Elementsof Logic, Book II, Ch. 3,
"The Icon, Index, and Symbol,"? 277-282,pp. 157-59(Icons "whichpartake
of simplequalities... are images;those whichrepresentthe relations,mainly


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

Yale French Studies

embraces the greater number of subclasses more usually distinguished by linguists.If we argue that a simple verbal mimesis or
icon can properlybe only a mimesis of a sound, then we locate
onomatopoeiaas the most direct,most physicallytexturedinstance
of verbal image. Such words as murmur,whisper,hiss,pop do seem
indubitablyto sound like what they say. We speculate that they
have been developedinto wordsperhapsfromsimplymimeticnoises.
It may be necessaryto remindourselvesnow and then that,as true
idea-bearingwords, these all have, in addition to whateverdegree
of naturalfitness,a distinctcomponentof linguisticconvention.We
confrontin onomatopoeicwords a characteranalogous in the realm
of sound to that in the realm of sightwhich we note in the picture
partlysurvivingas pictogram.Dogs bark, we have oftenbeen told,
cats meow, pigs grunt,guns bang, machinesrattle,click, or chatter
noises in different
languages.One curious illustration
with different
of the principle occurs now and then in a certain self-betrayal
technique of the comic-stripartist. A boy carries a balloon on a
string; it breaks; the word "POP! " appears in the midst of the
explosion.In the next framethe boy's fatherlooks in fromanother
room. "Did you call me, Henry?" At the same time,it seems probable that any name of a sound we can thinkof in our own language
will strikeus with a degree of onomatopoeicforce.Consider,e.g.:
pop, fizz, bang, boom, murmur,whisper,groan, roar, tinkle,jingle,
scream,yip, screech,buzz, crack, growl.Try to thinkof the name
of a sound that does not have a degree of such suggestiveness.No
doubt a numberof principlesare at work in various words of the
above list. (Perhaps in the end we may conceive a very special
categoryforsounds made by the humanvoice?)
III. b. Root-formingMorphemes (Phonaesthemes).What used to
be thoughtof as a sort of vaguely extended and mysteriousonomatopoeia,a verbal magic,is betterexplainedby modernlinguistics
dyadic,or so regarded,of the parts of one thingby analogous relationsin
theirown parts,are diagrams....." ? 277); Roman Jakobson,"Quest forthe
Essence of Language,"Diogenes, LI (1965), 21-37.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

whichwordsin a givenlanguage,
echo,a property
as a paradigmatic
especiallythe nativewords,the oldest(oftenthe mostphysically
exhibitin virtueoftheirkinshipwithother
or emotively
(or phonaesthemof suchquasi-morphemic
the lexiconof the onlylanguage
ic) 19 affinities
It is a pleasantexerciseto lie halfwhichthiswriterspeaksfluently.
asleep and invitesuch sequencesto filterinto the mind. Run
by adding
to the emergent
You can hardly miss: bash, brash, clash, crash, dash, fash, gash,
gnash, hash, lash, mash, pash, (quash), rash, slash, smash, splash,

(squash),thrash,trash(especiallytherecentAmericanverb).A few

polysyllablescome to mind: calabash, succotash,mishmash,balder-

to knowwhetherwe have thoughtof themall.

dash. It is difficult
richlyimitativepoeticexpressionswe can
(In some of the
thinkof-for instance,Tennyson's"Moan of dovesin immemorial
bees-we maysee a unionof
of innumerable
elms,And murmuring
trueonomatopoeiaand a kindof momentary
ofonoma2?or a punning
We maycall this"orchestration"
of theportmanteau
topoeia.)The well-known
thegame.Flash,forinstance,is a juncturebetweenthe -ashfamily
and the small clusterthat includesalso flame,flare,flicker,flimmer.

An amblerwho rangesor rovesis a rambler.And surelya jilt is a

Jillwho jolts.

or emergently
The phonaesthemic
lexicaland conventional;theydo not reallydependon anyresemblancebetweenwordand thing.Yet suchfamiliesof wordshavea
kind of "natural"or felt forcethat makes themat least seem
Thustheydemanda kindofverbalattention
19Leonard Bloomfield,Language (New York: H. Holt and Company,
1933), p. 246; DwightL. Bolinger,"On Definingthe Morpheme,"Word,IV
(April 1948), 18-23, esp. 22; "Rime, Assonance, and MorphemeAnalysis,"
Word, VI (August 1950), 117-36, esp. 130, "phonaestheme"; Rulon Wells
and Jay Keyser, The Common Feature Method (New Haven: Sociology
Department,Yale University,1961), on the initial phoneme sequence (l,
esp. pp. 7-12.
20 Rene Wellek and Austin Warren,Theory of Literature(New York:
HarcourtBrace and Company,1949), pp. 163-64.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

Yale French Studies

not demanded by the literal and explicitlexical controlsso prominent in the verbal "paintings"of the traditionut pictura poesis21
and its progenythe numerous parallels and melanges of the verbal and visual arts.
In an essay of 1953 David Masson has delicatelysubtilizedand
extended this theme of the lexically implicit.He argues that the
Englishword nighthas a range of implicitor potentiallyexpressive
values that includebrilliance(frombright,white,light,sight,shine),
the numinousand the spacious (fromheight,might,rite,right,flight,
high),alarms and excursions(fromsmite,fright,bite, fight,excite,
flight).In contrast,the Frenchnuit has values that include glimmer
and glow (fromlueur, luir, lustre,lune), scream or hoot (fromcri,
aigu, huer,huee), ruin and danger(fromnuis, nuire,puni, fui,fini),
intimacy(fromnid, uni,benit,lui). And GermanNacht has splendor
(from Pracht, Strahl), violence, power, fear, awe, cruelty (from
Schlacht, Macht, Acht, Rache, Gefahr,Gram, Ahnung,Ahndung),
temporality(fromSucht, noch, nach), negation(fromnicht,nichts),
damp (fromnass, feucht).12
III. c. Kinaesthesis. There is somethingthat a few decades ago
used to be called "sound symbolism,"and somethingelse that
Samuel Johnson,in writing about Pope, called "representative
verse." Either mightbe mistakenfor,but neitheris, I believe, the
same as eitheronomatopoeia(III a above) or phonaesthemicechoing
(III b). Althoughone of these (sound symbolism)-can be more or
less located in the timbreeven of single syllables or in simple accumulationsof syllables,and the otherinheresin, or at least needs
the special aid of, sequences of syllablesor accents (especiallyverse
patterns),the two may be put correctly,I believe, togetherunder
the head of analogies generatedkinaesthetically(or internallyand
physically)3 in the articulation(even silent articulation)of verbal

See above note 3, recentstudies by Paulson and Hagstrum.

22 David I. Masson, "Vowel and ConsonantPatternsin Poetry,"Journal

of Aestheticsand Arts Criticism,XII (December 1953), reprintedin Essays

on the Language of Literature,ed. SeymourChatmanand Samuel R. Levin
(Boston: HoughtonMifflinCompany,1967), pp. 11-12.
23 Masson, in Chatman and Levin, p. 12, uses the term "proprioceptive
sensations."A categoryof "acoustic" associationsis forMasson wider than


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

The analogiesseemto includenotablytheheavyand the
light,the darkand the light,the slowand the fast,the roughand
the smooth,and perhapsthe hard and the soft.The difference
betweenE. A. Poe's celebratedrefrain"Quoththe Raven,'Never24
so effectively
by RomanJakobson)
of it, "Le Corbeaudit: 'Jamais
and Mallarme'scurioustranslation
plus,'" maysuffice
to suggestthedifference
heavy(or dark)and the light.The versemovement
I believe,
heavyor light,theslowor fast,maybe veryconveniently,
fromAlexanderPope's tourde forceof soundand sense
in the secondpart of his Essay on Criticism.The neo-classicor
baroqueage of ut picturewas indeedspeciallyconcernedto make
apt and lively.
A needless Alexandrineends the song,
That, like a wounded snake, drags its slow lengthalong.
When Ajax strivessome rock's vast weightto throw,
The line too labours, and the words move slow.
Not so, when swiftCamilla scours the plain,
Flies o'er th'unbendingcorn, and skims along the main.

of sound to sense, thoughtSamuel Johnson

These resemblances
dictatedto ourwilling
(in theRambler),were"chimerical"-fancies
beliefby theovertassertionofthesense.In hisLifeofPope he perof changingthesenseof a line,keepingapformedthe experiment
observedofcoursea quietusupon
A like operationhas beenperformed
the soundeffect.
by Mr. J.C.
Ransomupona line of Edna St. VincentMillay's,and he adds the
analogyof the chameleon,thattakes its color fromthe greenof
the traditional"onomatopoeia" (III a above) and very close to the "kinaesthetic." His categories are in general oriented toward the emotive or
"evocative" power of words ratherthan towardmimesis.His "kinaesthetic"
associations generallylie dormant unless activated by proximityto appropriatelexical forces.
24 "Language in Operation,"MelangesAlexandreKoyre (Paris: Hermann,
1964), II, 269-81.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

Yale FrenchStudies
2 (I myselfhave observedthat the bluish-greenish
the tree.
deep blue if she is
of a girlwill assumea marvelously
wearinga blue coat or even a largebroochof lapis lazuli.)But it
are reallydisablingto the fact,
seemsdoubtfulif such arguments
analogy.A delicate,perhapstenuous,
of kinaesthetic
or the theory,
is nonetheless realin thatit disappearswhenit has nothing
to fit-or perhapschanges,withinlimits,dependinguponwhatit
to fit.Dryden,doingbetterthanPope with
has the opportunity
roughSt. Cecilia odes, blendsa seriesof slow-fastmovements,
or at least
smoothwords,and perhapsonomatopoeias,
The trumpet'sloud clangor
Excites us to arms....
The double, double, double beat
Of the thund'ringdrum
Cries: "Hark! The foes come...."

II. d. Autonomy.(Auto-iconicity?)
And ten low words oft creep in one dull line.
Tho' oft the ear the open vowels tire.

graftsor blends-in
Pope's versepassageis fullof theseautonymic
virtueof thefactthatitsthemeis (it talksabout)suchspecialverse
but namesand assertsthem.
It not onlyillustrates
The autonomy
Wher'er you hear the cooling western breeze,
In the next line it whispersthroughthe trees.
That, like a wounded snake, drags its slow lengthalong.
The line too labours, and the words move slow.

Can the notionof autonomybe expanded,so as to includea

widersortof something,
a kindofauto-iconicity?
25 "The Poet as Woman," The World'sBody (New York: Charles Scribner's Sons, 1938), pp. 94-97.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

William Wimsatt
autonymis not an image of itself,but an instanceof what it names
or refersto. Thus: word,noun,polysyllable.26Autonymycomes out
of the peculiar abstractive,universalizing,and reflexivepower of
We may speak loosely of the "auto-icon" of Jeremy
Benthamkept in a cabinet at UniversityCollege, London. But only
the wax head is an icon. The restis a clothedmummyor skeletona relic. Perhaps a typerwriter
on a pedestal in a certainkind of art
show, and a tube of toothpastein a pictureframe,may be thought
of as mounted,framed,or foregroundedin such a way that they
become icons of themselves.Or perhaps not. Perhaps they are just
what they are. I saw a movie once in which William Powell, who
had lately returnedfromWorld War II duty in the Pacific,acted
the part of a homecomingveteran.The effectwas amusing for an
admirerof Powell. There is a sense in which every famous actor
becomes a partial auto-icon-only partial! Himself now mimes
himselfthen. Mimesis involves difference.Long ago Plato disposed
of the idea of the complete auto-icon. "We should have not an
image of Cratylus,but two Cratyluses."This mayhelp us to evaluate
a celebratedthesis of Lessing. Lessing thoughtthat the poetic aim
of making words, "artificial"signs (workingby convention),into
"natural" signs(workingby likeness)was best realizedin the theater,
where words, presumably,were direct and natural imitationsof
words. But direct mimicry,whetheron the stage or in life, must
be distinguishedfromdrama and poetry.So far as Samuel Foote,
on the stage, mimickedFaulkner a Dublin bookselleror Apreece a
Welsh aristocrat,he did indeed mimic them or take them off.So
far as Garrick, in company, mimicked his schoolmaster Samuel
Johnsonmakinglove to his wife Tetty,he did indeed mimic him,
in action and word. But so faras Garrickacted Hamlet afraidof his
it is
26"Since the name of a given object may be chosen arbitrarily,
quite possible to take as a name forthe thing,the thingitself,or, as a name
fora kind of thing,the thingsof this kind. We can, forinstance,adopt the
rule that, instead of the word 'match', a match shall always be placed on
the paper.But it is more oftena linguisticexpressionthan an extra-linguistic
object that is used as its own designation.We call an expressionwhich is
usd thisway autonymous"(RudolphCarnap,The Logical Syntaxof Language,
trans. Annette Smeaton [London: Routledge & Kegan Paul Ltd., 1967],
Part IV, ? 42, pp. 156-57). Cf. pp. 211-12,Grelling'santinomy.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

Yale French Studies

father'sghost, he created an instance of a fictiveuniversal. The
words indeed were not a mimesis.They were Hamlet's words. (So
far as any actor's words and gesturesreferto an inner state, they
are not mimesisbut an expression27-either of somethingpretended,
as in the view of Diderot or Charles Lamb, or of somethingreal, as
with the school of Stanislavskyand Boleslavsky.)
IV. Metaphor. Under the Saussurian head of "motivation,"the
reasons why an expressionmeans what it means or means it aptly
or vividly,linguistscorrectlyinclude metaphor.28 And metaphoris
a near equivalentof poetry,an aspect of the drama of words as the
response of human consciousness to context. (Drama does not
depend on the physicalpresenceof theaterand acoustic utterance.)
Metaphor is always more than conventionand more than simple
reference.Here let us rememberthat visual art too can be metaphoric,or symbolic,and perhaps always is when it is "art." This
happens in ways that verge on conventionbut are not simplyconvention. Visual art, like verbal art, is imaginative."The eye," as
Leo Steinberghas put it, "is a part of the mind."29And here are
deeper and richer materials for sortingand realization than this
essay will attempt. A book that I once published, entitled The
Verbal Icon, was about iconicityin the widersense.
V. Diagram. Ludwig Wittgenstein'swidelycited "proposition
a picture of reality... a model of reality"(Tractatus 4.01) refers
to a kind of "projection" or diagram which may be difficultfor
27 "Inner or psychic reality cannot be imitated, only expressed....
'WMknow about [feelings]in othersonly by way of signs.... Drama and
painting express... the activity of the soul only indirectly
.... [They]
imitatethe naturalsigns of these innerrealities'" (David J. Gordon,"Form
and Feeling," quoting Elias Schwartz, The Form of Feeling: Toward a
Mimetic Theoryof Literature,in Yale Review, LXII (Summer 1973), 592.
a StephenUllmann,Semantics,An Introductionto the Science of Meaning (Oxford,Basil Blackwell,1964), pp. 91-92. Cf. Peirce,Elementsof Logic
(above note 18), ? 277. "Those [icons] which representthe representative
character of a representamenby representinga parallelismin something
else, are metaphors."
29 "The Eye Is a Part of the Mind," Partisan Review, XX (1953), 194212. Or see Wallace Stevens,"The RelationsbetweenPoetryand Painting,"
The NecessaryAngel (New York: VintageBooks. 1951), pp. 159-76; E. H.
Gombrich,Symbolic Images, London: Phaidon, 1972; Rudolf Arnheim,
Visual Thinking,Berkeley: Universityof CalifoeniaPress, 1969.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

William Wimsatt
the literaryscholar to assimilate. Compare the epigraph to the
present essay with, for example, Tractatus 4.0311: "One name
stands for one thing,anotherfor anotherthing,and theyare combined withone another.In this way the whole group-like a tableau
vivant-presents a state of affairs."Less esotericnotions,however,
are available. Distinctfromthe color or timbreof "image," a certain
logic of diagrammaticrelations is the second main division of
iconicityin the Peirce-Jakobsonaccount. In his article of 1965 on
the "Essence of Language," Jakobsonis interestedin the paradigmatic formsof diagram,those that lie in the structureof language
itself,and outside the syntacticpatternsof any one utterance.Thus,
the fact that the comparativedegree of an adjective nearlyalways
has more phonemesthan the positive,and the superlativemorethan
the comparative.But such diagrams can also appear as syntactic
diagramswithinthe singleutterance: "The balloon grewbig,bigger,
biggest-and burst." Or a diagramof lexical irreversibility
may be
framedin a classic instance by repeatingmorphology:"Veni, vidi,
vici." Jakobsonin his celebrated analyses of short poems has bestowed much loving attention on very complicated syntacticdiagramsthat have hithertopassed unnoticed,perhapseven unfelt,by
many readers. Simple formswill sufficefor the presentargument.
Simplest or commonestof all is the mere parallel of elements in
prose enumeration,pairing,or antithesis.And this is often most
keenly noticed by the grammarianstylistwhen he encountersits
curious distortionand obliterationin the work of student writers,
the most common example perhaps being the enumerationof three
; 2.
or four subdivisionsof a topic: 1.
; 3.
; 4.
, in
such a way that the syntaxof each membervaries (eitherpainstakinglyor carelessly) fromthat of all the others. In the absence of
an example, it requires some pains, and is painful,for the grammarian to constructone. But I once improved(or deteriorated)a
student sentence to the followingextreme: "Let us beware the
falsificationinvolved in focusingone's attentionon a narrow segment of experience,interpretedon the basis of a limited set of
inclusive."My purpose
shallow principles,whichare not sufficiently

This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

Yale French Studies

was to illustrate a clear opposite to the prose style of Samuel
Johnson."Accordingto the inclinationsof nature,or the impressions
of precept, the daring and the cautious may move in different
directionswithouttouchingupon rashness or cowardice." "If you
are pleased with prognosticksof good, you will be terrifiedlikewise
withtokensof evil." That kind of extremeprose diagram(sometimes
diagramin excess of semantics)vergeson verse and calls out forit.
'Tis hard to say, if greaterwant of skill
Appear in writingor in judgingill;
But, of the two, less dangerousis th'offence
To tire out patience,than mislead our sense.

And with this example we arriveat that junctureof (1) the visual
page with (2) the auditorylogic which we alluded to at the end of
our firstsection. It is my notion that so far as alphabetized graphemes can have any close union with the phonological-semiotic
stratumwhich they are employed to betoken, this is the way in
which theyhave it. Some complicationsoccur. The union and support may be in a straightlogical pattern,such as we have just seen.
Or it may occur in various subtlerways, such as I have once called
"counterlogical": the metrical equalities of syllable and accent
marchingalong with the disparatephrasal patterns,the paradox of
antitheticor disparate words bearing the similarityof rhyme,the
special secondaryemphasisof caesural pause, or the emphasiswhich
falls upon line-terminalwords in secondary syntactic positions
throughthe artificeof enjambment.
Of man's firstdisobedience and the fruit
Of that forbiddentree....

Such counterlogicaldiagrammingin meterand rhyme,and also in

the stronglymarked vowel and consonantpatternsso well known
in poetry,doubtless picks up various elusive elementsof onomatopoeic, morphemic,and kinaestheticintimation(See above III a.b.c)
-and thus the "music" or the "magic" of the poetic sound pat246

This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

William Wimsatt
terns.30In all such patterns,as we experiencethemperhapsfirst,and
throughthe visual medium,we can ask whetherthe
visual does in fact join and accentuatea diagrampresentindependently in the semiotic and prosodic structure(Pope's syntax and
couplets,Milton's syntaxand blank verse,the concentrichexameters
of Horace can be recognizedby the attentivemind-eareven if written out as prose), or whetherin fact the visual only imposes, or
attemptsto impose, a diagram which otherwisehas no existence,
so that we have in fact an instance of poesis cum picture,where
poetrymerelyattemptsto become a visual art. This latter is what
happens, I believe, in the technopaigniaor shaped poems of the
classical tradition,in the Calligrammesof Apollinaire, in certain
other merelytypographicalmeasures of modernverse (the columnwidth lines of William Carlos Williams, for example, the approximate syllabics of Marianne Moore), and certainlyin some, though
not in all, of the typographicarrangementsof E. E. Cummings.But
it is not a dutyof this essay to adjudicate such examples.
VI. Conclusion-in the sense of a terminal remark. In Plato's
Cratylus,the chief inquiry into mimetic linguisticsthat survives
from ancient times, Socrates the fugleman,delivers a prolonged
etymologicaldiscourse in which he attempts,in large part playfully,
to convincethe "conventionalist"Hermogenesthat words reallydo
embodyreasons for meaningwhat theymean. Then he turnsto his
otherfriend,the "naturalist"Cratylus,and says in effect:"Yes, like
paint in pictures,words do bear some kind of resemblanceto the
thingstheymean. But afterall, the resemblanceis not perfect.If it
were, we should have not an image of something,but two identical
things.The principleof verbal imitationis really ratherweak. To
firmit up, so that we have a usable language, we must call in
generousassistance fromthe complementarylinguisticprincipleof
convention."It is my own notion that today we stand at almost
30 See the subtlydetectedexamplesin Masson's essay of 1953 (Chatman
and Levin, 1967, pp. 3-10). See JohnHollander,"'Sense VariouslyDrawn
Out': Some Observationson EnglishEnjambment,"in LiteraryTheoryand
Structure,ed. Frank Bradyet al. (New Haven: Yale UniversityPress, 1973),
pp. 201-26.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions

Yale French Studies

studies-though coming to it
the same juncturein linguistic-poetic
froman oppositedirection.Afterthe sweepinglysuccessfulassertion
of the primacyof conventionin language by the fatherof modern
linguistics,Ferdinandde Saussure, studentsof the presentera, and
most notablyand perhaps initiallyRoman Jakobson,reach back to
the earlier insightand authorityof C. S. Peirce, to renew and improve an awareness of the "natural" powers of language, both
imagisticand diagrammatic,on a wide front.


This content downloaded from on Fri, 26 Jun 2015 23:03:29 UTC
All use subject to JSTOR Terms and Conditions