You are on page 1of 15

Roland Barthes and the Limits of Structuralism

Author(s): Paul De Man


Source: Yale French Studies, No. 77, Reading the Archive: On Texts and Institutions (1990), pp.
177-190
Published by: Yale University Press
Stable URL: http://www.jstor.org/stable/2930153 .
Accessed: 10/08/2013 17:49

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
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.

http://www.jstor.org

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions
PAUL DE MAN

Roland Barthesand the Limitsof


Structuralism'

Despite the refinements of modernmeans of communication,the


relationshipbetweenAnglo-American and continental-especially
French-literarycriticismremainsa star-crossed story,plaguedbya
varietyof culturalgaps and time lags. The Frenchhave only just
gottenaroundto translating an essayby Empson,2and by the time
Americanworksofliterary theoryorliterarycriticismappearin Paris
they often have lost muchof theiryouthfulfreshness.Thereis more
goodwill and curiosityin theotherdirection,yetheretoo a mixture
of misguidedenthusiasmand misplacedsuspicionblursthe actual
issues. Evensome ofthemostenlightened ofEnglishand American
criticskeep consideringtheirFrenchcounterparts with the same
suspicionwithwhichEnglish-speaking touristsmightapproachthe
cafeau lait theyare beingservedforbreakfast in FrenchProvincial
hotels: they know they don't like it but aren't entirelycertain
whethertheyarebeingimposeduponorif,forlackofsomeritualistic
initiation,theyare perhapsmissingout on a goodthing.Othersare
willingto swallow Frenchculturewhole,frombreakfastcoffeeto
Mont Saint Michel and Chartres,but since intellectualfashions
changefasterthanculinarytastes,theymayfindthemselveswearing

1. [Themanuscript ofthisessayappearsto datefrom1972.It was commissioned


bytheNew YorkReviewofBooksas a reviewofextanttranslations ofBarthes's
work
intoEnglishbutwas neverpublished.Correspondence indicatesthatthereweredif-
ferencesbetweende Man andtheeditorsoverwhether thevocabulary oftheessaywas
tootechnicalfora generalreadership.
All textinbrackets
is suppliedbythetranscriber
andeditor,ThomasPepper, whoalsowishestoexpresshisthankstoE. S. BurtandJanie
Vanp6efortheirhelpandpatience.Thanksarealso duetoMrs.PauldeManformaking
thisarticleavailable.]
YFS 77,ReadingtheArchive:On TextsandInstitutions,
ed.E. S. BurtandJanie
Vanp6e,
X)1990byYale University.

177

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions
178 Yale FrenchStudies

a beretand drinkingPernodat a momentwhen the Frenchavant-


gardehas longsinceswitchedto a dietofcashmeresweatersandcold
milk.The CriticalEssays3ofRolandBarthesthathavejustbecome
availablein excellentEnglishtranslationsdate from1953 to 1963;
Mythologiesgoes backto 1957and appearsin a regrettably abridged
version.4I cannothelpspeculatingaboutall thethingsthatcouldgo
wrongin thereceptionoftextsthatnow combinea nostalgicwitha
genuinebut out-of-phase revolutionary quality.Perhapsthe most
usefulfunctionforan American-based viewofRolandBarthesmaybe
to tryto anticipateunwarranted dismissalas well as misplaceden-
thusiasmfortheaspectsoftheworkwithwhichBartheshimselfmay
no longerbe so pleased.Bartheshas beenintroduced to Americansas
possibly"themostintelligent manofourtime"5and anymanneeds
and deservesprotectionfromtheexpectationsraisedbysuchhyper-
bole.
Fordespitethe emphasison structure, code, sign,text,reading,
intratextualrelationships,
etc.,and despitetheproliferationofa tech-
nical vocabularyprimarilyderivedfromstructurallinguistics,the
actual innovationsintroducedby Roland Barthesin the analytic
studyofliterary textsarerelativelyslight.Evenin his moretechnical
works,unfortunately not yetavailablein English,such as S/Z (the
analysisofa briefnarrativetextbyBalzac),andthevariousarticleson
narrativetechniquepublishedin Communications,6 the contribu-
tionto practicalcriticismis not as extensiveas the methodological
apparatuswouldlead one toexpect.The workof"pure"structuralists
such as the linguistGreimasand his groupor of some amongBar-

2. It shouldbe added,in all fairness,thatPoetique has also publishedrecent


Americanwork.[See,forexample,thefollowingessays,all publishedin Poetique 2
(1971):SeymourChatman,"HenryJameset le stylede l'intangibilit6"; WilliamEmp-
son, "Assertionsdansles mots";and Northrop Frye,"Litterature et mythe."]
3. [RolandBarthes,Essais critiques(Paris:Seuil, 1964),trans.RichardHoward,
CriticalEssays(Evanston:Northwestern University Press,1972).]
4. [RolandBarthes,Mythologies (Paris:Seuil, 1957),trans.AnnetteLavers,My-
thologies(New York:Hill and Wang,1972).]
5. BySusanSontag.[Shewritesin herprefaceto Writing DegreeZero(NewYork:
Hill and Wang,1968),vii: "Still,I wouldarguethatBarthesis themostconsistently
important,
intelligent, andusefulcritic-stretching thatterm-to haveemergedany-
wherein thelastfifteen years."]
6. [S/Z(Paris:Seuil,1970),trans.RichardHoward,S/Z (NewYork:Hill andWang,
1974).Some oftheessaysfirstpublishedin Communications havebeenreprintedin
RolandBarthes,LAventuresemiologique (Paris:Seuil,1985),andin Englishin Roland
Barthes,Image-Music-Text, trans.StephenHeath,(New York:Hill andWang,1977).]

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions
PAUL DE MAN 179

thes's most giftedassociates,such as GerardGenetteor Tzvetan


Todorov,is more rigorousand more exhaustivethan Barthes's
thoughit is onlyfairto pointout hereits avowedindebtednessto
him.Hence theriskofdisappointment oroverhasty dismissalforthe
wrongreasons.Barthesis primarily a criticofliteraryideologyand,as
such,his workis moreessayisticandreflective thanit is technical-
perhapsmost of all when the claim to methodologicalprecisionis
most emphaticallybeing stated. The close integrationof meth-
odologywithideologyis an attractive characteristic ofEuropeanin-
tellectuallifeeversince structuralism becamea publicissue in the
sixties-and, forbetterorworse,Frenchwriterson literature arestill
much closerto beingpublicfigures,committedto articulateposi-
tions,thantheirAmericancounterparts. Barthesplayedan activepart
in therecentBattlesoftheBooksandhis workbearsthetracesofhis
involvements.It has to be read and understoodas an intellectual
adventureratherthanas thescientifically motivateddevelopment of
a methodology. He is at leastas interestedin thereasonsforadvocat-
ingcertaintechnicaldevicesas in theiractualapplication;hencethe
polemical,proselytizing tone ofmanyofhis essays,hence also the
manyinterviews, pamphlets, positionpapers,etc.His workshouldbe
readwithinthecontextoftheparticularsituationwithinwhichit is
written,thatof the ideologicaldemonsunderlying the practiceof
literarycriticismin France.Thissituationis idiosyncratically French
and cannot be transposedtel quel (c'est le cas de le dire) to the
Americansituation.It does notfollowhoweverthatthestoryofBar-
thes'sintellectualitinerary is withoutdirectinterestforAmerican
readers.Americancriticismis notoriouslyrichin technicalinstru-
ments(butmuchpoorerin understanding therationalefortheiruse);
but it is frustrated,as well it mightbe, in its attemptsto relate
particularstudies and findingsto largerhistorical,semantic,and
epistemological questions.Thatsuchdifficulties existis byno means
a signofweakness;it onlybecomesone iftheveryawarenessofthe
largercontextis lost or if the broaderinferencesof a methodare
misunderstood. Regardless ofitsregionalpeculiarities, theconfigura-
tion of Barthes'senterpriseis of wide enoughsignificanceto have
paradigmaticvalue forall studentsof literaturewillingto put the
premisesoftheircraftintoquestion.

A somewhateuphoric,slightlymanic tone runsthroughBarthes's


writings,temperedby considerableironyand discretionbut un-

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions
180 Yale FrenchStudies

mistakablybracedbythefeelingofbeingon the thresholdofmajor


discoveries:"A new anthropology, withunsuspectedwatershedsof
meaningis perhapsbeingborn:the map of humanpraxis is being
redrawn,and the formof this enormousmodification(butnot, of
course,its content)cannotfail to remindus of the Renaissance."7
This statementdatesfrom1966,but one stillfindssimilartrumpet
blasts,onlyslightlymuted,in recentutterances.It is the tone of a
man liberated froma constrainingpast, who has "the earth . . . all
before(him),"and who looks about "witha heart/Joyous, notscared
at its own liberty."8The exactnatureofthisliberationcan best be
statedin linguisticterms,in a formulajustlyborrowed fromBarthes
himself:it is the liberationof the signifier fromthe constraintsof
referentialmeaning.In all thetraditional polaritiesused throughout
the ages to describethe inherenttensionthatshapes literarylan-
guage-polaritiessuch as content/form, logos (thatwhichis being
said)andlexis(themannerofsayingit),meaning/sign, message/code,
langue/parole,signifie/signifiant, voice/writing (and the sequence
couldbe continued)-theimplicitvalorization has alwaysprivileged
thefirsttermsand consideredthesecondas an auxiliary, an adjunct
in theserviceoftheformer. Languageitself,as thesignofa presum-
ablynonlinguistic"content"or "reality," is therefore devalorizedas
the vehicleor carrierof a meaningto whichit refersand thatlies
outsideit; in thepolarityman/language, it seems commonsensical
enoughforus humansto privilegethefirsttermoverthesecondand
torateexperienceaboveutterance.Literature is saidto "represent" or
"express"or at most,"transform" an extralinguistic entityor event
whichit is theinterpreter's(orcritic's)tasktoreachas a specificunit
of meaning.Whateverthe shadingsused in describingthe rela-
tionship(and theyare infinite), it remainsfundamentally best ex-
pressedby the metaphorofthe dependenceoflanguage(literary or
not) on somethingin whose serviceit operates.Languageacquires
dignityonlyto theextentthatit can be saidtoresembleortopartake
oftheentityto whichit refers.The Copernicanrevolutionheralded
byBarthesconsistsnotin turning thismodelsimplyaround(andthus
claimingthat,insteadofbeingtheslaveofmeaning,languagewould
now becomeits master)but in assertingthe autonomyofwhatthe
7. [See RolandBarthes,Critiqueet vdrit6(Paris:Seuil,1966),48; in English,Crit-
icism and Truth,trans.RichardHoward,(Minneapolis:MinnesotaUniversity Press,
1987),66.]
8. [See theopeningofWordsworth's Prelude,11. 15-16.]

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions
PAUL DE MAN 181

linguistSaussurewas thefirstto call thesignifier,i.e., theobjective


properties ofthesignindependent oftheirsemanticfunctionas code,
such as, forexample,therednessofa traffic signas opticalevent,or
thesoundofa wordas acousticevent.The possibility forthesignifier
to enterintosystemsofrelationship withothersignifiers despitethe
constraintoftheunderlying or,ifone prefers, over-standing or tran-
scendental,meaningprovesthat the relationshipbetweensign or
wordand meaningis not simplyone ofdependence;it suggeststhat
thismetaphoricallanguageofpolarizedhierarchiesandpowerstruc-
turesfails to do justice to the delicate complexityof these rela-
tionships.The sciencethatsetsoutto describethefunctionsandthe
interrelationships ofsignifiers
(includingreference, one amongoth-
ers)is calledsemiology,thestudyofsignsindependent oftheirpartic-
ularmeanings,in contrastto semantics,whichoperateson thelevel
ofmeaningitself.Barthesis one oftheleadingrepresentatives ofthis
science,not so muchas its initiator-he is thefirstto acknowledge
his debtto Saussure,Jakobson, Hjelmslevand others-but as one of
its mosteffective advocates.
Whyis it thatideasaboutlanguageleadingto thescienceofsemi-
ologyacquiredsuch a polemicalvigorin thehandsofBarthes?They
had beenaroundforquitea while,notonlyin thefieldoflinguistics,
butinvariousphilosophiesoflanguageandintheformalist schoolsof
literarycriticismthatdominatedthescenein mostcountries-with
thepossibleexceptionofFrance.It is truethattheFrenchhavea way
oftakinghold,oftenbelatedly,ofotherpeoples'ideas and suddenly
rediscoveringthem with so much originalenergythat they are
positivelyre-born;this happened,in recent years, with Hegel,
Heidegger, Freud,and Marxand is aboutto happennowwithNietz-
sche.In thiscase howeverthereis moretoit thanmereGallic energy.
Barthes'sdeliberateexcursionintotherealmofideologyis typicalof
the developmentsummarizedunderthe catch-alltermof "struc-
turalism"and ofall his books,theearlyMythologies is perhapsbest
suitedto illustratetheprocessI am trying to describe.
Barthesis a bornsemiologist,endowedwithan innatesense for
theformalplayoflinguisticconnotations, thekindofeyeand mind
thatnoticesat once how an advertisement fora brandof spaghetti
seduces the onlookerby combining,in thepictureofthered toma-
toes,thewhitespaghettiandthegreenpeppers,thethreecolorsofthe
house of Savoyand of the nationalItalianflag-thus allowingthe
potentialconsumerto tasteall thatmakesItalyItalianin one single

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions
182 Yale FrenchStudies

bite of cannedpasta.9He has used thisgiftedeye not onlyto scru-


butsocial andculturalfactsas well,treating
tinizeliterature, themin
the same mannerin which a formalistically orientedcriticwould
treata literarytext.Mythologies,a book thatremainsremarkably
freshalthoughthefactsit describesbelongto thebygoneera ofpre-
GaullistFrancein theearlyfifties,undertakespreciselythiskindof
semiocriticalsociology.The undisputedmastersofthegenreareWal-
terBenjaminandTheodorAdornoand,althoughBartheswas an early
exponentoftheworkofBrechtin France,I doubtthathe knewBen-
jaminorAdornowell whenhe wrotetheMythologies. The common
ancestryis neverthelessapparentfromreference, in the important
concludingessayon historyand myth,to Marx'sGermanIdeology,
themodeltextforall ideologicaldemystifications.
AlmostanyoftheMythologiescan be used to illustrateBarthes's
maininsight.Take,forinstance,theopeningessayon catch-as-catch-
can wrestlingas an exampleof the contrastbetweena referential,
thematicreadingandthefreeplayofsignifiers. The pointis notthat,
in theworldofcatch-as-catch-can,all thefightsarerigged;thiswould
not make the eventless referentialbut merelydisplacethereferent
fromthe theme "competition"to thatof "deceit."Whatfascinates
Barthesis that actorsas well as spectatorsfullyacquiesce to the
deceitandthatall pretenseat opencontesthas beenabandoned,thus
voidingthe eventofcontentand ofmeaning.Thereonlyremainsa
seriesofgesturesthatcan be highlyskillfulat simulatingthedrama
ofcompetition(thetriumphofwinning,theabjectionofloss,orthe
dramaofperipeteiaor reversal)but thatexistpurelyformally, inde-
pendentlyofan outcomethatis no longerpartofthegame.Catchis
not a game but a simulacrum,a fiction;Barthescalls it a "myth."
Mythsofthistypeaboundin thefabricofanysociety.Theirattrac-
tion is not due to theiractual contentbut to the glitterof their
surface,and thisglitterin turnowesitsbrillianceto thegratuity, the
lack ofsemanticresponsibility, ofthefictionalsign.This playis far
frominnocent.It is in the natureoffictionsto be morepersuasive
than facts,and especiallypersuasivein seemingmore "real" than
natureitself.Theirorder,theircoherence,theirsymmetry is possible

9. [See "Rh6toriquede l'image,"in Communications8 (1964),trans.Stephen


Heath,"RhetoricoftheImage,"in RolandBarthes, Image-Music-Text(NewYork:Hill
andWang,1977).De Man inflects in "SignandSymbolin
thesameexampledifferently
in Paulde Man,AestheticIdeology,
Hegel'sAesthetics," (Min-
ed.AndrzejWarminski,
neapolis:MinnesotaUniversity Press,forthcoming).]

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions
PAUL DE MAN 183

because theyare accountableonlyto themselves,yettheseare pre-


ciselythequalitieswistfully associatedwiththeworldofnatureand
necessity.As a result,the most superfluousof gesturesare most
likelyto become the hardestto do without.Theirveryartificiality
endowsthemwitha maximumofnaturalappeal.Fictionsor myths
areaddictivebecausetheysubstitutefornaturalneedsbybeingmore
naturalthanthenaturetheydisplace.The particularshadeofperver-
sityand bad conscienceassociatedwithfictionstemsfromthecom-
plicityinvolvedin partialawarenessof this ambivalencecoupled
with an even strongerdesireto resistits exposure.It followsthat
fictionsare themostsaleable commoditymanufactured byman; an
adman'sdreamofperfectcoincidencebetweendescriptionand pro-
motion.Disinterestedin themselves, theyarethedefenselesspreyof
any interestthatwishes to make use ofthem.Whentheyare thus
beingenlistedin the serviceof collectivepatternsof interest-in-
cludinginterestsofthe "highest"moralormetaphysicalorder-fic-
tionsbecomeideologies.One can see thatanyideologywouldalways
have a vestedinterestin theoriesoflanguageadvocatingthenatural
correspondence betweensignandmeaning,sincetheydependon the
illusionofthiscorrespondence fortheireffectiveness
whereastheo-
riesthatputintoquestionthesubservience, resemblanceorpotential
identity betweensignandmeaningarealwayssubversive, evenifthey
remain strictlyconfinedto linguisticphenomena.Barthes'sMy-
thologiesare fullyawareofthis;theybringthe subversiveness into
theopenbyexposingthestructure ofthesocialmythsas wellas their
manipulation.The politicalresultsareclearlyvisibleas Mythologies
movesfromtherelativelyinnocentmystifications ofcatchas catch
can or the Tourde Franceto consumergoods(e.g.,the CitroenDS,
steak pommes fritesor the singingstyle of the baritoneGerard
Souzay,etc.) to reachfinallythe domain of the printedwordand
imageas itappearsin themoviesorinParis-Match. Afterhavingbeen
thetargetofa heavyhandedandviciousattackbyRaymondPicard,a
SorbonneProfessor ofFrenchliteraturewhosefieldofspecialization
is thelifeofRacine,Bartheswroteperhapshisbestmythologie in the
firstpartofthe counterattacking pamphletentitledCritiqueet ve-
rite,in whichtheideologicalinfrastructure ofFrenchacademiccrit-
icism is revealedwithmasterful economyand withoutan ounce of
personalspite.
The demystifying powerofsemiologyis botha sourceofstrength
and a danger.It is impossibletobe so completelyrightat theexpense

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions
184 Yale FrenchStudies

of otherswithoutsome dangerto oneself.The perfectconvergence


betweenBarthes'ssocialcriticism, includingthecriticismofacadem-
ic traditionalism,and the meansused in accomplishingthishighly
desirableaim engenders its ownmystification,thistimeon thelevel
ofmethodratherthanofsubstance.The verypoweroftheinstrument
used createsan overconfidence thatgeneratesits ownsetofcounter-
questions.In this case, the questionshave to do withthe claim of
havingfinallygroundedthe studyofliteraturein foundationsepis-
temologicallystrongenoughto be called scientific.The headytone
alludedto earlieroccurswheneverthisclaimappearson thehorizon.
It is accountableforsome ofhis mostpowerful influence.Puttingit
intoquestionnowisemeansa desiretoturntheclockback-a foolish
wish at best,fortherecan be no returnfromthedemystifying power
ofsemiologicalanalysis.No literary studycan avoidgoingthrougha
severesemiocriticalprocessand thereis much to be said forgoing
throughthese fireswith as urbane,surefooted, and entertaining a
guide as RolandBarthes.At stake is the statusof structuralism, a
methodological blueprint forscientific
researchthat,likeRousseau's
stateofnature,"no longerexists,has perhapsneverexistedand will
probablynevercome intobeing"butwhichwe nevertheless cannot
do without.10
As in Barthes'ssocial myths,thereferential, representational ef-
fectiveness ofliterary languageis greaterthanin actualcommunica-
tion because, like his catch-as-catch-can wrestlers,it is so utterly
devoidof message.Literatureovermeans,as we say of bombsthat
theyoverkill.This referential suggestiveness, whichaccountsforthe
factthatone respondswithstronger emotionto a fictionalnarrative
thanto an actualevent,is ofcourseillusoryandsomething forwhich
a scienceofliterature(whetherwe call it stylisticsor literarysemi-
ologyorwhatever)shouldaccountwithoutbeingtakenin byit. The
classical way of dealingwiththe questionis to bypassit, as when
RomanJakobson rightfullyassertsthat,in literature,
thelanguageis
auto-telic,i.e., "focus[sed]on themessageforits ownsake"11rather
than on its content.By gettingrid of all the mess and muddleof

10. [Jean-JacquesRousseau,Du contratsocial,in Oeuvrescompltes (Paris:Gal-


Cf.,also Paul de Man,Allegoriesof
limard,1964),vol. 3, 123. De Man's translation.
Reading(New Haven:Yale University Press,1979),136.]
11. [See RomanJakobson, "ClosingStatement:Linguisticsand Poetics,"in his
SelectedWritings, "PoetryofGrammarand GrammarofPoetry," ed.,StephenRudy,
(The Hague: Mouton,1981),vol.3, 25.]

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions
PAUL DE MAN 185

signification, theformulaopensup a heretofore undiscovered world


of scientificdiscourse,coveringthe entirefieldof literarysyntax,
grammar, phonology, prosody,and rhetoric.Withthe inevitablere-
sult,however, thattheprivilegedadequationofsignto meaningthat
governstheworldoffictionis takenas theidealmodeltowardswhich
all semanticsystemsareassumedto tend.This modelthenbeginsto
functionas a regulatory normbymeansofwhichall deviationsand
transformations of a givensystemare being evaluated.Literature
becomes,to borrowa phrasefromthe titleofBarthes'sfirstbook,a
degreezero ofsemanticaberration. Weknowthatit owes thisprivi-
legedpositionto thebracketing ofitsreferentialfunction,dismissed
as contingencyor ideology,and not takenseriouslyas a semantic
interference withinthe semiologicalstructure.
The seductionofthe literarymodelhas undoubtedly workedon
Barthes,as ithas toonall writers endowedwithliterary sensitivity.Up
through it
Mythologies takesat times rather naiveforms, when,in
as
theconcludingessayofthatbook,literature is heldup,inoppositionto
ideology, as a "transformationofthesignintomeaning:itsidealwould
be ... to reach,not the meaningofwords,but the meaningofthe
thingsthemselves."12 Moretechnicalversionsofthesame mythap-
pearin varioustexts,as when,in an articleonnamesin Proust(whose
plannedtitlefortheconcludingsectionofhis novelis knownto have
beenatsomepoint,"TheAgeofThings")hespeaksofliterature as [that
which"wouldbedefinedbya Cratylianconsciousnessofsignsandthe
writerwouldbe themouthpieceofanage-oldmythwhichdecreesthat
languageimitatesideas and that,contraryto the specificationsof
linguisticscience,signsaremotivated"]. 13Unqualifiedassenttosuch
propositionswould be an exampleofmisplacedenthusiasmforthe
mostdebatableaspectofBarthes'senterprise.
In themanifestoCritiqueet verite(1966)inwhichthevocabulary
is moretransformational thanstructural,closerto Chomskythanto
Jakobson, thepositionis morecomplexbutnotessentiallydifferent.
It now takes the formof a three-pronged, hierarchizedapproachto

12. [SeeMythologies (Paris:Seuil,1957),241.]


13. [QuotationsuppliedbytheeditorfromRolandBarthes,"Proustetles noms,"in
Le Degr6zerode l'6crituresuivide Nouveauxessaiscritiques(Paris:Seuil,Collection
Points,1953,1972),136.FortheEnglishtranslation oftheessay,see RolandBarthes,
New CriticalEssays,trans.RichardHoward,(NewYork:Hill andWang,1980),68. De
Man referstoa previouspassagefromthesameessayin "TheResistancetoTheory," in
The Resistanceto Theory(Minneapolis:MinnesotaUniversity Press,1986).]

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions
186 Yale FrenchStudies

literaturein which a distinctionis made betweenliteraryscience,


literarycriticism,and literaryreading.The controlling
authority of
thefirstdiscipline,theonlyone to be freeoftheerrorofsemantiza-
tionandto layclaimto truth,is beyondquestion."Ifone is willingto
admitthe textualnatureofthe literarywork(and drawthe proper
conclusionsfromthis knowledge),then a certaintypeof literary
science becomes possible.... Its model will undoubtedly be lin-
guistic.... Literarysciencewill have forits object,not to explain
whya certainmeaninghas to be accepted,norevenwhyit has been
accepted(thisbeingthetaskofthehistorians), butwhyit is accept-
able.Not in termsofthephilologicalrulesofliterary meaning,butin
termsofthelinguisticrulesofsymbolicmeaning."[Critiqueetverite
57-58, de Man's translation.] Byemphaticallydrawingattentionto
its own methodologicalapparatus,S/Z, Barthes'smost systematic
piece ofliteraryanalysisto date,allows itselfto be takenas a first
exemplarymovein the elaborationofsuch a science.The impactof
thisprogramon literarystudieshas been and will remainconsider-
able. It will notdo to dismissthemethodological claimsas a hoax or
as ironicwindow-dressing used bya writerofmoretraditionalliter-
aryvirtues.We cannotreassureourselvesby stressingthe elegance,
thesensitivity, thestrongly personal,evenconfessionalelementthat
is partofBarthes'sdistinctive toneandthatmakeshimintooneofthe
"best" writersat worktodayin any genre,in the most traditional
senseofthisqualitativeepithet.The theoretical challengeis genuine.
Ithas to be takenall themoreseriouslysincetheparticular qualityof
Barthes'swritingis due tohis desireto believein itstheoretical foun-
dationsand to repressdoubtsthatwouldbreakits stability.
The unresolvedquestionis whetherthe semantic,reference-ori-
ented functionof literaturecan be consideredas contingentor
whetheritis a constitutive elementofall literary language.The auto-
telic,nonreferential aspectofliterature stressedbyJakobson cannot
seriouslybe contested,but the questionremainswhyit is always
againand systematically beingoverlooked, as ifit werea threatthat
had to be repressed.The firstquotedpassagefromCritiqueet verite
layingdownthe directivesforthe literaryscienceofthe futureis a
goodexample:one can see Barthesfluttering aroundthequestionlike
a motharounda flame,fascinatedbutbackingawayin self-defense.
All theoreticalfindings aboutliterature confirm thatit can neverbe
reducedto a specificmeaningorsetofmeanings, yetitis alwaysbeing
interpreted reductively as ifit werea statementormessage.Barthes

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions
PAUL DE MAN 187

grantsthe existenceofthispatternoferrorbutdeniesthatit is the


objectofliterary scienceto accountforit; thisis saidto be thetaskof
historians,thusimplyingthatthereasonsforthepattern'sexistence
arenotlinguisticbutideological.The further implicationis thatthe
negativelaborofideologicaldemystification will eventuallybe able
topreventthedistortion thatsuperimposes uponliterature a positive,
assertivemeaningforeignto its actualnature.Bartheshas neverre-
nouncedthishope; in a recentinterview, despitemanynuancesand
reservations,he stilltalksabout"theultimatetransparency ofsocial
relationships"14 as thegoalofthecriticalenterprise. Yet,inthemean-
time,his methodologicalpostulateshave begunto erodeunderthe
impactofthequestionwhichhe delegatedto other,moreempirical
disciplines.
That literaturecan be ideologicallymanipulatedis obviousbut
doesn'tsuffice toprovethatthisdistortion is nota particular aspectof
a largerpatternoferror.Soonerorlater,anyliterary study,no matter
howrigorously andlegitimately formalisticitmaybe,mustreturnto
theproblemofinterpretation, no longerin thenaiveconvictionofa
priorityofcontentoverform,butas a consequenceofthemuchmore
unsettlingexperienceofbeingunableto cleanseitsowndiscourseof
aberrantly referential implications.The traditionalconceptofread-
ingused byBarthesandbasedon themodelofan encoding/decoding
processbecomesinoperative iftheoriginalmastercode remainsout
ofreachoftheoperator, who thenbecomesunableto understand his
A
owndiscourse. scienceunableto readitselfcan no longer be called
a science.The possibilityofa scientificsemiologyis challengedbya
problemthatcan no longerbe accountedforin purelysemiological
terms.
This challengereachedBarthesfromthe somewhatunexpected
quartersofphilosophy, a disciplinethatearlierstructuralistshad dis-
missed in favorof the so-calledsciences of man: psychology, an-
thropology, and linguisticsconsideredas a social science.This dis-
missal provedto be premature,based as it was on an inadequate
evaluationofthespecifically philosophicalabilityto putthefounda-
tionsofphilosophy intoquestion,ina self-destructivemannerthatno
sciencecouldeverdareto emulate.The workofMichelFoucaultand
especiallyofJacquesDerrida-whose determining impacton literary

14. [See RolandBarthes,"Reponses,"in Tel Quel 47 (Fall 1971),specialissue on


RolandBarthes,107.]

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions
188 Yale FrenchStudies

theoryis confirmed by the recentlypublishedbook La Dissemina-


tion-thematizes the problemof linguisticdelusion in a manner
which semiologicalcriticsof Barthes'spersuasioncannotafford to
ignore,all themoresinceitrevealsthatthechallengehadneverceased
tobe presentina philosophicalandliterary activitythatstructuralists
triedto ignore.One thinksofcertainrecurrent misreadings ofRou-
sseau's, Hegel's,and especiallyNietzsche's(as well as Heidegger's)
attitudetowardsliterature andalsoofBarthes's crypticremark duringa
recentdiscussionthat"A criticismofLautreamont, for example, is
probablynotpossible,"a remarkthatcouldbe readas an abdicationof
semiologywhenit confronts thelanguageofpoetry.15
Barthes'sintellectualintegrityis apparentin his reactionto this
philosophical challenge. For the moment,it has takentheformofa
retreatfromthe methodologicaloptimismthatstill inspiredS/Z.
Morerecenttheoretical papers(nototherrecentbookssuchas L'Em-
piredes signes,inspiredbya tripto Japan, orSade, Fourier,Loyola,in
whichthe semiologicaleuphoriais allowedto reignundisturbed)16
sketchout a muchless ambitiousprogramthatsoundslike a return
to a pragmaticcollectingofliterary data,andaresharplyawareofthe
inabilityof semiologyto accountforthe stylistictensionbetween
writtenandspokenlanguage.One ofthesepapersavailableinEnglish
translationinvitesus to embarkon "the searchformodels,ofpat-
terns:sentencestructures, syntagmatic cliches,divisionsand clau-
sulae ofsentences;andwhatwouldinspiresuchworkis theconvic-
tionthatstyleis essentiallya citationalprocess,a bodyofformulae, a
memory(almostin thecybernetic sense oftheword),a culturaland
not an expressiveinheritance.... These models are only the depos-
not
itoriesofculture(eveniftheyseemveryold).Theyarerepetitions,
essential elements; citations,not expressions;stereotypes,not
archetypes."17Traces of the readingof Derrida,Gilles Deleuze,
Foucault(and perhapsalso of the Columbia-basedstylist,Michael

15. [See, forexample,the secondpartof Critiqueet veritW,(65),whereBarthes


addressestheproblemoftheconfrontation ofcriticismwithlexicallyesoterictexts,
and refersto Lautreamont.]
16. [RolandBarthes,L'Empiredes signes (Geneva:Skira,1970),trans.Richard
Howard,TheEmpireofSigns(New York:Hill andWang,1982);Sade, Fourier, Loyola
Loyola(New York:Hill and
(Paris:Seuil, 1971),trans.RichardMiller,Sade, Fourier,
Wang,1976).]I suspectthatsomeoftheseessaysmayin factbe ofearlierdatebuthave
no information to confirmthis.
17. [See RolandBarthes,"Styleand its Image,"in LiteraryStyle:A Symposium,
SeymourChatman,ed. (London:OxfordUniversity Press,1971),9-10.]

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions
PAUL DE MAN 189

are noticeablein these sentences.They cannothowever


Riffaterre)
representa definitive position.The mindcannotremainat restin a
mererepertorization its ownrecurrent
of aberrations; it is boundto
systematizeits ownnegativeself-insight intocategoriesthathaveat
leasttheappearanceofpassion,novelty, anddifference.Thereis every
reasonto supposethatBarthes'sfutureworkwill participatein this
development in thedevelopment
as he.participated thatled up to it.
The avant-garde reviewTel Quel, whose attitudetowardorthodox
structuralism has alwaysbeenhealthilyuncomplacent, recentlyde-
votedan entireissue to RolandBarthes, 18thuscreatingthemislead-
ingimpressionthattheyweretryingto erecta monumentofa man
who is aboutas monumentalas a Cheshirecat.I doubtthatTel Quel
was tryingto kick Barthesupstairsinto some kindof Pantheonof
unchangingforms;whoeverassumes this to be possiblewould se-
riouslymisjudgetheresilienceofone ofthemostagileandresource-
fulmindsin thefieldofliteraryand linguisticstudies.
As faras Americancriticismis concerned,itsreactionto Barthes
is not yetclear.The recenttranslations are a usefulbut still inade-
quate firststep in introducing his workinto English.The Critical
Essays,mostlyprefaceswrittenforcommercialeditions,stemfrom
the period that precedesthe developmentof semiology-roughly
1963-and are mostlyinteresting in showingBarthes'sdiscontent
with the prevailingmethodsof literarystudyduringthe fiftiesin
France,and his delightat discovering thenewperspective openedby
his readingsin linguistics.They createthe somewhatmisleading
impressionthat his main interestsare confinedto the theaterof
Brechtand to the novelsofRobbe-Grillet and theyshouldcertainly
notbe takenas a fairsampleofhis accomplishments. Thereis more
semiologicalfinesseto be gatheredfromtheMythologies, including
severalnotincludedin thisselection,thanfromtheCriticalEssays.
How the availabilityof his moreimportanttheoreticalbooks (On
Racine, Critiqueet verite,S/Z, varioustheoreticalpapers)19might
influenceAmericancriticismcan be inferred fromthe reactionsof
someAmericanspecialistswhoarefamiliarwithhisworkandshowa
fundamentalresistance.In a recentessay entitled"On Defining
Form,"evenas knowledgeablea scholaras SeymourChatman,who

18. ITel Quel 47 (Fall 1971).]


19. On Racine,whichwastranslated in 1962,raisestheentirequestionofBarthes's
to psychoanalytical
relationship criticism,a questiontoo complexto be treatedhere.
[See RolandBarthes,On Racine (New York:Hill andWang,1964).]

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions
190 Yale FrenchStudies

has alreadydone a good deal in bringingtogethercontinentaland


Americanstudiesofliterature, takesBarthesto taskforputtingthe
referential functionofliterary languageintoquestion:"It is difficult
to understand," he writes,"whyone shoulddenythatthereare,ulti-
mately,contentsorsignifiesreferred to.... The contentofa literary
workis not thelanguagebutwhatthelanguagestandsfor,its refer-
ence. . .. The languageis a mediating formbetweentheliteraryform
(structure-texture) and theultimatecontent."20 Barthes'spointnever
was thatliterature hadno referential functionbutthatno "ultimate"
referent couldeverbe reachedandthattherefore therationalityofthe
criticalmetalanguageis constantlythreatenedand problematic.I
have suggestedthatBartheswas beingall too hopefulin havingbe-
lieved,fora while,thatthe threatcould be ignoredor delegatedto
historians.At least,the scientificself-assurance thusgainedis pro-
ductiveand has a negativevalidity, as faras it goes-and now thatit
seemsto knowitshorizons,itremainsa necessaryfactofanycritical
education.To returnto an unproblematic notionofsignificationis to
taketwostepsbackward,a stepbackwardintoa pseudo-sciencein a
domainin whichno scienceis possible,and a stepbackwardinto a
pseudo-sciencethat,unlikeBarthes'ssemiology, is too remotefrom
its objectto be demystified byit. As longas the "liberationdu signi-
fiant"is beingresistedforthewrongreasons,Barthes'scriticismwill
have littleto teachAmericanstudentsofliterature.

20. [InNew Literary 2 (1971),218-26.]


History,

This content downloaded from 206.212.0.156 on Sat, 10 Aug 2013 17:49:26 PM


All use subject to JSTOR Terms and Conditions

You might also like