Você está na página 1de 14

Allegories of Reading

---Figural Language
incR~ietzsche, Rilke, and Proust

( 9 \ .{,--- : l -~ ~ (. '-

( n l ;.






Paul de Mal)j

J ''





' ' 'U-iv.J-:>

~--\ ~ '~~t- 20 J,--~






New Hav~n and London

Yale Umversity Press


c - --,



Published with assistance from the Kingsley Trust Association Publication Fund established by the Scroll and
Key Society of Yale College.

Quand on lit trap vite ou trap doucement on n'entend rien.

Copyright 1979 by Yale University. All rights reserved. This

book may not be reproduced, in whole or in part, in any form
(beyond that copying permitted in Sections 107 and 108 of the
U.S . Copyright Law and except by reviewers for the public
press) , without written permission from the P.ublishers.
Set in Zapf International type.
Printed in the United States of America by The Murray
Printing Company, Westford, Massachusetts.

Library of Congress Cataloging in Publication Data

De Man, Paul.
Allegories of reading.
Includes index.
1. French literature-History and criticism.

2. Rousseau, jean jacques, 1712-1778-Style.

3. German literature--History and criticism.
4. Figures of speech.


5. Allegory.

ISBN 0--300--02322-7
ISBN 0--300--02845-8 (pbk.)


I. Title.








;v~) '\2.-C.-<'' ~ '


Ut D ""

mon that they share referential reading-moment xplicitly built in

f within the spectrum o
e1 ca
matter how deluded
this moment may be in its mode as well as in its thematicCOi1ferrr:
the deadly "horn of the bull" referred to by Michel Leiris in a text that
is indeed as political as it is autobiographical. 1 But whereas the
relationship between cognition and performance is relatively easy to
grasp in the case of a temporal speech act such as promise-which,
in Rousseau's work, is the model for the Social Contract-it is more
....--,..-. complex in the confe~sional mode of his autobiographies. By reading
a central passag~ from the Confossio;Is, f"attempt to cl;ri:ty the relationship between critical procedures that start out from the discourse of the subject and procedures that start out from political
Among the various more or less _~ul and embarr(:!ssing
scenes from childhood and adolescence relatea in the first three
books of the Confessions, Rousseau singled out the episo~(~~on
and the ribbon as of particular affective significance, a t!J1ly primal
( scene of lie and deception strategically placed in the narrative and
told with special panache. We are invited to believe that the episode
was never revealed to anyone prior to the privileged reader of fne
1 { Confossions "and . . . that the desire to free myself, so to speak, from
I\ this weight has greatly contributed to my resolve to write my confes~ sions" (86). 2 When Rousseau returns to the Confossions in the later
Fourth Reverie, he again singles out this same episode as a paradig-

1. "De Ia litterature consideree comme une tauromachie," in Michel Leins,

L'age d'homme (Paris: Gallimard, 1946). The essay dates from 1945, immediately
after the war.
2. Page numbers are from].]. Rousseau, Oeuvres completes, Les confessions,
autres textes autobiographiques, ed. Bernard Gagnebin and Marcel Raymond (Paris:
Gallimard [Bibliotheque de Ia Pleaide], 1959), vol. 1. The passage concludes Book II
of the Confossions and appears on pp. 85-87.


matic event, the core of his autobiographical narrat~e-seleetion.

is, in itself, as arbitrary as it is suspicious, but it provides us with a
textual event ofunaeniable exegehc mteresC the juxtapositiffil,oftwo
confessional texts linked together by an expliS!...rs:.Retition, ~onfession, as it were, of a confession ..-----1ne epiSode itself is one in a series of stories of petty larceny, but
with an added twist. While employed as a servant in an aristocratic
Turin household, Rousseau has stolen a "pink and silver colored ~
ribbon." When the theft is discovered, he accuses a young maidservant of having given him the ribbon, th_e irnR.lication being _!?at she
was trying to seduce him. In publ1c confrontation, he obstinately
cl~sstoiy, thus casting irreparable doubt on the honesty and
the morality of an innocent girl who has never done him the slightest
bit of harm and whose sublime good nature does not even flinch in
the face of dastardly accusation: "Ah Rousseau! I took you to be a )
man of good character. You are making me very unhappy but I
would hate to change places with you" (85). The story ends badly, \
with both characters being dismissed, thus allowing Rousseau to
speculate at length, and with some relish, on the dreadful things that 1
are bound to have ~~mtfieSubsequent career of the haples.:_)
The first thing established by this edifYing narrative is that the ~
Confossions are not primarily a confessional text. To conJeSSis to
overcomeguiitand shame in the name of truth: it is an epistemological use of language in which ethical values of good and evil are
superseded by values of truth and falsehood, one of the implications
being that vices such as concupiscence, envy, greed, and the like are ,
vices ~imarily because they_c~.12el one to lie. By_sJ.!J}ng things as
they are~ ethical balance is restored and redemption
can start in the clarified atmosphere of a truth that does not hesitate ,
to reveal the crime in all its..b.orr.or. In this-case, Rous;au even adds
totfue horror by conjuring up, in the narrative of the Confossions as l
well as that of the Promenade, the dire consequences that his action "
may have had for the victim. Confessions occur in the name of an
absoluturuth ~c_h is said to exist "for itself'' ("pour elle seule,"
[:t028]) and of which pa~ticular truths are only derivative and secondary aspects.
But even within the first narrative, in Book II of the Confossions, Rousseau cannot limit himself to the mere statement
of what "really'' happened, although he is proud to draw attention










to the fullness of a self-accusatiOn whose candor we are never supposed to suspect: "I have been very thorough in the confession I have
illaae,aild iteould certainly never be said that I tried to conceal the
blackness of my crime" (86). But it does not suffice to tell all. It is not
IJ enough to conftss, one also h~s to exc"Use: ''But (would not fulfill the
purpose ofthis booK. ifi did not reveal my inner sentiments as well,
and ifi did not fear to excuse myself by me~p~what confo.2:...11}.S.tothe truth" ("que je [ne] craignisse de m'excuser en ce qui est conforme
~ la verite" [86, my italics]). This al~o happens, it should be
noted, m the nam_s: of truth and, at first sight, there should be no
con~i~ween conf~ssion a~d ex:_use. Y~~he 1a-ng~~- rev~-~~
~nswn m thmpresswn~ cramdre ae m 'excuser. The only thmg one
hasta-fea1 fi om t~atifWTI1-i!llieea exculpate the confessor, thus making the confession (and the confessional text) redundant as it originates. Qui s'accuse s'excusf!i this sounds convincing
and convenient enough, but, in terms of absolute truth, it ruins the
seriousness of any confessional discourse by making it .rl.Gdgs.t.m&-..
thre . Since confes~i;;n is not a rep:;-wtion in tnGa~ of pract~
justice but exists only as a verbal utterance, how then are we to
knowtha twe are!ndeeo"Clealing-with- a true confession, since the
reGogniti6i1 Of guilt implie~ i ts exonerati;n in the name of the same
transcendental principle of truth that allowed for the certitude of
guilt in th_e first place? -- - T;fact, a far-; eaching modification of th~ or~izing _principle
of truth occurs between the two sections of the narrative. The truth
in whose name the excuse has to be stated, even at Rousseau's as1\, sumed "corps defendant," is not structured like the truth principle
that governs the confession. It does not unv~~-oLh~t
,-. states a suspicion, a possible discrepancy that might lead to an impossi0ili1y to know. The discrepancy, of course, is between the "sentiment interieur" that accompanied (or prompted?) the act and the
act itself. But the spatial inside/outsid.e metaphor is misleading, for it
articulates a gifferent!tio.u..thaLiSJlQ.!_J,Eatial at all. Th"'e"disfffic-tieR--bctween the confession stated in the mode-ofreveaied truth and the
'( confessi?n stated .i~1 th~ ~od:J:f exeusejs_that t~~ evid~ns;~_fu!:.Jh.e.___....-1
former IS referentiai-(ihe nbbon), whereas the eVIdence for the latter
carrOnly be verbal. Rousseau ~om-;ey'hi~ "in-;;;feel:Ing'... ttn1soruy
/ i~, his word for ~whereas t~e for his
~ft is, at least in t~_le2...Whether we bel1evef1im

J 3.

'"Ql.~ is

v. } /
( '\



c. ' ( Q


so even within the immediate situation, when no actual text is

-, '-'


..._ '.







lur~<\ 0 .Q,






(L'{...? 1 V





~ L~








Q~ ~ ~Y'






it is !he verbal OL.Ilonverbal

evidence that makes the difference, not the sincer._ity of the spe<!!se~
t~ g~llibili!)c o_f theJistener. T.he disti~tion is that the latter process "\
necessarily includes a moment of understanding that cannot be
equated with a perception, and that the logic that governs this moment is not the same as that which governs a referential verification.
What Rousseau is saying then, when h e insists on "sentiment int~
ieur," is that confessional language can be considered under a double

eP.istemor6gica.l pe__:~
. ect.iv.e.:::Jt T::nctioris as a veri!latlle !ef~~8
<;:o'gmtii5i1,1Jufl.f..?lso functions as a statement whose reliability can""
not be verified bLfi?:,lpiric_!.!l _m~ans ._ Tne convergence . of the two
modes is not a priori given, and it is because of the possibility of a
discrepancy between them that the possibility of excuse arises. The \
excuse articulates the discrepancy and, in so doing, it actually asserts
ira'Sla'Cltwhereas-rl- i-s-on-ly-a--suspicion) . It believes, or pretends to
believe, that the act of stealing the ribbon is both this act as a
physical fact (he removed it from the place where it was and put it in
his pocket, or wherever he kept it), ._,____
as well as a certain "inner feeling"
that was somehow (and this "how" remains open) -;;nnecteerwrrmr.Moreover, it believes that the-fac t :ma tlie leelfng are- nor the -same.
Thus to complicate a .Q.tct certainly is: to act. The difference between ,. ~
the verbal excuse and the referential crime is not a simple opposition

between ~_9ction and a mere utterance about an action. To steal is

to ~d include's nonecessa ry Verbal ere'inerits.~ To confess is dis- C
ciir'sWe.,-but- the discourse -1s.governea by a rinciple of referential
verificatio~ that includes an extraverb~Y mo~;;;-t: eve n 1f we c~
tfiat wesard something (as ~posedt;did): t'il'everificaflon oltnis
verbal event, the d"E"Cision -about .the.Jr:vih or falsehood of its occur- 1
~nee, i~ot ve.:_~al b~t ~actual, the knowl;{jge th;ti'he- Utterance , <
actually too"Kplace. No such lJns-s-ibili ty of verification exists for the
, excuse, which is verbal in its utterance, in its effect and ih its authority: its purpose is not to state but to convince, it'~lf an "inner" process
to which only words can l?~aJ_witness. As is well knowrrat-l-east..s-inc~l
'!rrlstitV--excuses--an:-a'complex instance of what he termed perform-


present. Someone's sentiments are accessible only through the medium ofmiU:i_siY\
of gestures that require deciphering and function as a language. That this deciphering is not necessarily reliable is clear from the fact that the facial expression of;
say, a thief at the moment he is caught red-handed is not likely to weigh heavily as
evidence in a court oflaw. Our own sentiments are available to us only in the same
4 . See, for example,.). L. Austin, "Performative Utterances" and "A Plea for



ati;ve utterances, a variety of speech act. The interest of Rousseau's

text is that it explicitly functions performatively as well as cogm( tively, ana thus gives indications about the structure ofperformative
etoric; this is already established in this text when the confession
ls to close off a discourse which feels compelled to modulate from
e confessional into the apologetic mode. 5
Neither does the performance of the excuse allow for a closing
off of the apologetic text, despite Rousseau's plea at the end of Book
II: "This is what I had to say on this matter. May I be allowed never
( to mention it again" (87). Yet, some ten years later, in the Fourth
Reverie~ tells ~e story all over again, in the context of a
meditation-ttmr11asto do~e possible"t"'excusability" of lies.
Clearly, the apology has not succeeded in becalming his own guilt to
the point whePe he would be allowed to forget it. It doesn't matter
much, for our purpose, whether the guilt truly relates to this particular act or if the act is merely made to substitute for another, worse
crime or humiliation. It may stand for a whole series of crimes, a
general mood of guilt, yet the repetition is significant by itself: what-

Excuses," in Philosophical Papers, ed. ]. 0. Urmson and G. ]. Warnock (Oxford,


.--.---- -

5. ~~lyvay_~fdealing with t~ recurreo~ttern in Rousseau's writings

is by stressing the bad faith of his commitment to a morale de !'intention , fhe ethical
stance for which he wa~ taken severely to task by Sartre. In his commentary on the
passage, Marcel Raymond, though less severe, takes the same approach: "By revealing his 'inner feelings' ['dispositions interieures'] which were good . . . it apR~a]]>
that after having stigmatized his misdeed he gratlually begins to justifY it. The same
gliding- ana swerving motion can be observed more than once . i~ the Conftssions,
especially when Rousseau accounts for the abandonment of his children. He is
always led to distinguish the intent from-the-a.t'' (1273-74). It can, however, be
shown that Rousseau's ethics is muctl r ather a morale de pratifJue than a morale de
]'intention, and that this analysis therefore does not account for the genuinely preKantian interest of his ethica l language and theory. 'T he extensive possibilities of bad
faith engendered by the distinction between the actual event and the inner feeling
are abundantly present throughout Rousseau, but they don't govern the mo~ puzzling and interesting movements and coinages ofthe text. Whether the link between
'- "inner" feeling and "outer" action can be called intentional is precisely the burden
of the interpretation and cannot be asserted without further evidence. If we are
righ t in saying that "qui s'accuse s'excuse," then the relationship between confession
and excuse is rhetorical prior to being intentional. The same assumption of intentional apologetics, controlled by the narrative voice, underlies the recent readings
of the Confessions by Phillippe Lejeune in Le pacte autobiographifJue (Paris, 1976)
and "Le peigne casse," PoetUjue 25 (1976): 1-30.

ever the content of the criminal act may have been, the excuse pre1 \
sen ted in the Conftssions was unable to satisfy Rousseau as a judge
Jean-Jacques. This failure was already partly inscribed within the
excuse itself and it governs its further expansion and repetition.
Rousseau excuses himself from his gratuitous viciousness bJ'
identifYing his inner feeling a_s shame about himself rather than any
hostility_ tQwards his. victim: " . . . the presence of so many people
was stronger than my repentance. I hardly feared punishment, my
only fear was shame; but I feared shame more than death, more
than the crime, more than anything in the world. I wished I could
have sunk and stifled myself in the center of the earth: unconquerable shame was stronger than anything else, shame alone caused my
impudence and the more guilty I became, the more the terror of
admitting my guilt made me fearless" (86).
It is easy enough to describe how "shame" functions in a context
that seems to offer a convincing answer to the question: what is
shame or, rath_eG_~hat is one ashamed of? Since the entire scene
.stands under the aegis of theft~ it hasto do with possession, and
desire must therefore be understood as functionini;'afleastat times,
as a desire to possess, in all the connotations of the term. Once it is
~moved from its legitimate owner, the ribbon, being in itself devoid
of meaning and function, can circulate symbolically as a pu~e _signifier and become the articulating hinge in ~chain of exchanges and
pCYssessions. As the ribbon changes-hands it traces a circuit leading to
the exposure of a hidden, censored desire. Rousseau identifies the
desire as nis desire for Marion: "it was my intention to give her the
ribbon" (86), i.e., to "possess" her. At this point in the readrn
suggested by Rousseau, the proper meaning of the trope is clear
enough: the ribbon "stands for" Rousseau's desire for Marion or,
what amounts to the same thing, for Marion herself.
Or, rather, it stands for the free circulation
the desire b~
tween Rousseau and Marion, for the reciprocity which, as we know
from ]J.Llie, is for Rousseau the. vecycondi_tion oflove; it stands for the
substitutability of Rousseau for Marion and vice versa. Rousseau desires Marion as Marion desires Rousseau. But since, within the atmosphere of intrigue and suspicion that prevails in the household of
the Comtesse de Vercellis, the phantasy ofthis symmetrical reciproc- \ 1
ity is experienced as an interdict, its figure, the ribbon, has to be 1:
stolen, and the agent of this transgression has to be susceptible of 1
being substituted: if Rousseau has to be willing to steal the ribbon, '







then Marion has to be willing to sul;?stitu_te for Rousseau in performing this act. 6 We have at least two levels of substitution (or displace[' m~ faking plac~: t!:~ rii:>bon sub_sti~ting for ~sire.which i~ itself
a desire for substitutiOn. Both are governed by the same desire for
specular symmetry wtrieh gives to the symbolic object a detectable,
univocal proper meaning. The sysie.Ill-wor.ks: "I accused Marion of
having done what I wanted to do and of having given me the ribbon
because it was my intention to give it to her" (86). The substitutions
(h--;;;e taken place without destroying the cohesion of the system,
reflected in the balanced syntax of the sentence and now understandable exactly as we comprehend the ribbon to signifY desire.
\s~ular figures of this kind are metaphors and it should be noted
that onthisstill elementary l~eTo:fjmderstanding, the introc!_u~
of the figural dimension in the text occurs first by ways of metaphor.
f1e_~ll~~~~1is m etaphor, revealed in- the ''confession" of
!ROusseau's desire for Manon, functions as an excuse if we are willing
I to take the desire at face value. If it is granted that Marion is desirable, or Rousseau ardent to such an extent, then the motivation for
the theft becomes understandable and easy to forgive. He did it all
out oflove for her, and who would be a dour enough literalist to let a
little property stand in the way of young love? We would then be
willing to grant Rousseau that "viciousness was never further from
me than at this cruel moment, and when I accused the hapless girl, it
is bizarre but it is true that my friendship for her was the cause of
my accusation" (86). Substitution is indeed bizarre (it is odd to take
' , a ribbon for a pe;son) but since it reveals mofives, causes, an
desires, the odditY is quickly reduced back to sense. The story may be
a rebus or a riddle in which a ribbon Ts""ill-adet~ signifY a desire, but
the riddle can be solved. The delivery of meaning is delayed but by no
means impossible.
This is nof the only wa , h
in which ~
possession allows for the al -important introduction ffioural displacement: things are not merely what they seem to
be, a ribbon is not just a n on, to steal can be an act oflove, an act
performed by Rousseau can be said to be performed by Marion and,
in the proc~ss, it becomes more rather than less comprehensible, etc.
Yet the text does not stay confine-d WiThl.n-Ulispattern of desire. For

one thing, to excuse the crime of theft does not suffice to excuse the
worse crime of slander which, as both common sense and Rousseau
tell us, is much harder to accept. 7 Neither can the shame be accounted for by the hidden nature of the desire, as would be the case
in an oedipal situation. 8 The interdict does not weigh very heavily
and the revelation of Rousseau's desire, in a public situation that does
not allow for more intimate self-examination, hardly warrants such
an outburst of shame. More important than any of these referential
considerations, the text is not set up in such a way as to court
sympathy in the name of Marion's erotic charm, a strategy which
Rousseau uses with some skill in many other instances including the
first part of julie. Another form of desire than the desire of possession
is operative in the latter part ofthe story, which also bears the main
performative burden of the excuse and in which the crime is no
longer that of theft.
~-The obvious satisfaction in the tone and the eloquence of the
passage quoted above, the easy flow of hyperboles (" . . . je la craignois [la honte] plus que la mort, plus que le crime, plus que tout au
monde. j'aurois voulu m'enforcer, m'etouffer dans le centre
de la terre . . ." [86]), the obvious delis!2.!_with which the desir~
to~j.Qg_!~vealed,_~!Lp.o.tnt to an~ther s~ii~-= , I
than mere possessiOn and mdependent of the par:ticular target of 1
the c.lesire. -One -is more ashamed ~fthe exposure of the desire to
expose oneself than of the desire to possess; like Freud's dreams of
~kedness , shame is primarily exhibitionistic. What Rousseau reglly
wanted is neither the ribbon nor Marion, but the public scene of
~~I11Cl1neactually gets. 'fhe faCt that he made no attempt to \.
eonceal the evidence confirms this. The more crime there is, the !
more theft, lie, slander, and stubborn persistence in each of them,
e better. The m ore there is to expose, the more there is to be
a~hamedOf;-the--~exposure, the mor~ satisfYing the
scene, and, especially, the more satisfYing and eloquent the belated
revelation, in the la ter narrative, of the inability to reveal. This desire




7. "To lie for one's own adva ntage is deceit, to lie for the benefit of another is
fraudulent, to lie in order to harm is slander ; it is the worst kind of lie" (Fourth
Reverie, 1029).
8. The embarrassing story of Roussea u's rejection by Mme. de Vercellis, who is
dying of a cancer of the breast, immedia tely precedes the story of Marion , but
nothing in the text suggests a concatenation that would a llow one to substitute
Marion for Mme. de Vercellis in a scene of rejection.

G. It is therefore consistent that, wh~ei

pas etre a vo tre place" (85).

~k~vould say: 'je ne voudrois





,: .


vl f'

/ r-0") .

,1. 1 ::o-rJ


is 1ruly shameful, for it suggests that ~~rion was destroyed, not for
the sake of RousseaU's saving face, nor for the sake of his desire for
her, but merely in order to provi9e him with a stage on which to
jparade his disgrace or, what amounts to the same thing, to furnish
l.---hJm with a good ending for Book II of his Confessions. The structure
, ' ~-s self-perpetu~tmg, en CiEfme,- a_s- is implied in its description as
exposure of the desire to expose, for each new stage in the unveiling
suggest~?._ a deeper shame, a greater impossibility to reveal, and a
;.r~facti~!1 in outwitting this jmposibillty.
The structure of desire as exposure ratherthan as possession
\ explains why shame functions indeed, as it does in this text, as the
most effective excuse, much more effectively than greed, or lust, or ..,
love. ~omie )s .:Oleptic, J:>u.!._ excuse is bela!ed and always occurs 'J
after the C!:_ime; s.ince the_5.~ime is expos~re, the ~xcuse co~ists in ' .'\
recapitulating the exposure intli:e gmse of concealment. The excuse
is a ruse which permits exposure inthe name of hiding, not unlikeBeing, in the later Reidegger, reveals itselfby hiding. Or, put differ- , ,..,.
1 ently, shame used as excuse permits repression to function as revelal tion and thus to make pleasure and guilt interchangeable. Guilj_is
r{fOrgl\ren because it allows for the pleasure of revealing its repn~S';;ion.
It follows that re2ressi~m is in fact an excuse, one speech ~ct among
"+--.,. ~@f'&-fur-t-her-pessibi-Hties...Ibe analysis of shame as
excuse makes evident the strong link between the performance of '-j
excuses and the act of understanding. It has Ted to the problema tics ",
Qf hiding and revealing, which are g~J'_P-f.9b~~.~~
Excuse occurs within an _e.pistern0logical twilight-..zane......be_tween Q)
1 ~~ng:f_~ot-~wingikthis is also why it has to be centered on
the crime of lying and why Rousseau can excuse himself for everything provided he can be excused for lying. When this turns out not
1 to have been the case, when his claim to have lived for the sake of
truth (vitam impendere vera) is being contested from the outside, the
closure of excuse C"qu'il me soit permis de n'en reparler jamais")
becomes a delusion and the Fourth Reverie has to be written.
The passage also stakes out the -l~ofhow ffiis understcrfiding ~
of understanding then is to be understood. For the distinction between desire as possession and desire as exposure, although it unde\ niably is at work within the text, does not .tructure its maiQ..JJJ.Q.Ve\ ~t 2 t could not be s~id,_f~r i!:lsfanC! , that the lat~r q~Qns.tructs .
' t,!1e former. Both converge towards a unified signification, and the
1 shame experienced at the desire to possess dovetails with the deeper


. ~..)

~ ~{'>)~

~(;'!~ \





, ,r





shame felt at self-exposure, just as the excuse for the one conspired
~the exc~or the otl1er'inmutua!__:e_i1_2_forcement. This implies --J
t~at the mode of cognition as_hidingb:_e~~~:_~ fundamentally akin
~o th~ mode of cognitiQn as posses~ioE_ and t~at, at leasf!Jp"fill this
, poi~ t_Q.}>now and to own are structured in the same way. Truth is aJ)rO perty of entities, and to lie is to ~I, like Prometheus, this truth
a~om its owner. Int.Fi'ede~sness ofthe excuse pattern, the lie
is made kgftimate, but this occurs within a system of truth and ,
fa.l sehood th~!-~~g~_:I_2.ts ~alorization but not in its \'
~e.1T also implies that the terminology of repression and \'
exposure encountered in the passage on shame is entirely compatible
with the system of symbolic substitutions (based on encoded significations arbitrarily attributed to a free signifier, the ribbon) that
govern the passage on possessive desire (''je l'accusai d'avoir fait ce
que je voulois faire . . ." [86]). The ~~rhetoric of the passage;
whose ungerlying metaphor, encomp.ssing both possession and ex~-is....that of unveiling, combines with a generalized pattern of
t~opoiC2g!c;al substitution to reach a convincing meaning. What
seemed at first like irrational .b ehavior bordering on insanity has, by
the end of the passage, become comprehensible enough to be ins:Q!:( p_grated within a generaL ec.QQQillJLDf1lllrnau._aifr.lli;ity, in a theory
of desire, repression, and self-analyzing discourse in...wllich excus~
and_lg10wledge converge. Desire, now expanded far enough to in~ elude the }}iding(re_veaTing movem~ of the u_Dconscious ~-well as }
possession, functions as t he cause of the_~ntire scene (" . . . it is
bizarre but ~ fnendship f~r her was the cause of my
. accusations" [86]), and once this ?esi~~ has bee1;._made t~2.ear i!Y\
1/\\\ ali_its complexity, the action is understood and, consequently, ,
II\ e~~for itwas primarily its incongruity that ~~.2:_givable.
Knowledge-;-morality, possession, exposure, affectivity (shame as the \
~ synthesis of pleasure and pain), and the performative excuse are all
~ ultimately part of one system that is epistemologically as well as
i l ethically grounded and therefore available as meaning, in th~g1.900
- - \ of ~d~sta_nding. just as in a somewhat earlier passage of the Confessions the particular injustice of which Rousseau had been a victim
becomes, by metaphorical synecdoche, the paradigm for the universal experience of injustice, 9 the episode ends up in a generalized
economy of rewards and punishments. The i1~jury done to Marion is
9. See the episode of Mile Lambercier's broken comb i!l Book I of the Confessions, especially p. 20.






for b~ the subsequent suffering inflicted on Rousseau

y nameless avengers acting in her stead. 10 The restoration ofjustice
aturally follows the disclosure of meaning. ~e&-the_e~j
,, . u~e fail and why does Rousseau have to return to an enig.~t has
~ een so- well resolved?
-we~have, of course, omitted from the reading the other sentence
in which the verb "excuser" is explicitly being used, again in a somewhat unusual construction; the oddity of "que je craignisse de m'excuser" is repeated in the even more unusual locution: '~Je m'excusai
sur le premier objet qui s'offrif' ("I excused mysplf upon the first
thing that offered itself' [86]), as one would say '~je me VE'ngeai" or
'je m'acharnai sur le premier objet qui s'offrit."'' The sentence is
inserted, it is true, within a context that may seem to confirm the
coherence ofthe causal chain:" . . . it is bizarre but it is true that my
friendship for her was the cause of my accusation. She was present to
my mind, I excused myself on the first thing that offered itself. I
accused her of having done what I wanted to do and ofhaving given
me the ribbon because it was my intention to give it to her . . ."
(86). Because Rousseau desires Marion, she haunts his mind and her
nam~ i~-prono.2;1Eced almosLun_<~2_:msly, as if ~vere a slip, -;, I segment of the discourse ofthe other. But t11e use of a vocabulary of
conti;1gency C''~ pr~mier ol?jet qui ~it'') within an argument of
Wcausality is arresting and disruptive, for the sentence is phrased in If
~ ~
fl such a way as to allow .fl?r a comElete disjunction between Rousseau's desires and interests and the selection of this particular name. ~
t\M"arion just-happened to be the ' fir.t thing j_hat came to mind; any
I otEer na'me:-anyOth er word , any other sound or noise could ~e
done just as well and Marion's entry into the discourse is a mere
({effect of chance. She is a free signifier, metonymically related to the

part she is made to play in the subs~~l}t system <:_ f exchan~es and
substit~~!!.~t"s~e "i~, howeve~, in an en~irely dif~erent situation than ~
tn:eoTiler free s_l,gmfier, the nbbon, which also JUSt haJ2Pened to _b~ , I
seady-to-hand, but whj_ci!Js...n.ot in any way itself the object of a ''
_ desire. Whereas, in the development that follows and that introduces
the entire chain leading fiom desire to shame to (dis)possession to
concealment to revelation to excuse and to distribut ~ve j~stice, Marion can be the organizing principle because she is considered to be
~en_centerofanillgeto reveal. Her bo;-aage as target liberates
in turn the free play of her symbolical substitutes. Unlike the ribbon,
Marion is not herself divested of positive signification, since no revelation or no excuse would be possible ifher presence within the chain
were not motivated as the target of the entire action. But if her 1
no~!?~! presence is a mere coincidence, then we-a re . efltering anJt>
efi!}~ dl'f1eren::tsys'te~-in~wliic1i sucn"rerms as desrre;Shame, guilt, , ' I
exposure, ,aQd repressiOn no longer have any place.
>;- _ }1 _!h;;~JDLuUhtiext,. QJ]eshould resist all temptation to gi~~
.( any significance whatever to the sound "Marion." For it is only ifthe
b act that initiated th~ntire chain, the utterance o.f the sound "Ma.rarbitra.ci;ion," is truly without any conceivable motive that the total
~ ~ .. -- '""'' _,_....
erfonnative excuse of all. The estrangement between subject and
so r. adical tha..f it escapes any mode of camp. rehen- .
~, sian. ~n ev:ry!!:ing else fail~, one can always plead insanity. "Marion" is meaningless and powerless to generate by itself t~_hain of
causal SUbSfffi:iTtorrs~figures tfiat stri.ictti'?es-th?s~~unding text', I
WhiCfl IS a text- Of d'e sire as well as~aaesttefOr text: It Stands'venfirely
ut -~f the ~ystem-6ftruth:-virtue, and'""Oersrarrdi~1g (illordeceit,
evil, and er;Qr) that gives meaning to Hie passage, and to the Confossionys.a whole. The sentence: '~je m'excusai sur le premieP.ol?jet qui
s'offrit" is therefore an anacoluthon, 1~lement that disruE!-s








10. "If this crimt: can be redeemed, as I hope it may, it must be by the many
misfortunes that have da~ke~ed 'i11e Taler part -;r my life, by forty years of upright
~ and honorable behavior under difficult circumstances. Poor Marion finds so many
<Jven~rs_in this world that, no matter how considerably I have ofte;;deCflier';Tna ve
little fear tl~l~thi& gBiH-w-itl::wne.. ..Ihis is all I had to say on this matter.
May I be allowed never to mention it again" (87).
11 . The editor of the Pleiade Rousseau, Marcel Raymond, comments on the
passage and quotes Ramon Fernandez (De Ia personnalite, p. 77): "He accuses her as
if he leaned on a gi~e..offumtiture io.,a~jd falli.ng.~I~ond speaks of "aDariTIOsi
dreamlike movement dictated by an unconscious which suddenly feels itself accused
( and by which he transfers the 'misdeed' upon the other, on his nearby partner"






12. Classical rhetoric mentions anacoluthol"r--ss~c'ially with regard to the

structure b!rreriodicarSentences, wher;.-3~hift, _~J.:I~: ~her, i:~urs betw~~
the ffrst part of the period (protasis) and the second part (apodosis}.Hemrich
La usberg inHarzdbm:h der Literarischen Rhetorik (Munich, 1960), 1:459, 924, gives
an example from Vergil: "quamquam animus meminisse horret luctuque refugit,
incipiam" (Aeneid 2, 12). The following example from Racine is frequently quoted:
voulez que ce Dieu vous comble de bienfaits I Et ne l'aimer jamais." )~
Anacoluthon is not restricted to uninflected parts ofspeech};J.utcan involve.nguns-Gr- J
inflected shifters such as pronouns. It designates any grammatical or syntactical
discontl-;:;uity fn which a construction interrupts another befor_ejt is CO@p_le!e?. A

..........\ ~




l::Xl.:USl::S lCUNl'E:SI>IUN:SJ





\. V'>'

\ ' ~


the meaning, the readability of the apologetic discourse, and reopens

---rafthe excuse seemed to_hav~~losed offJ!-Iow are we to under_..D
stand-the-implications of this sentence and what does it do to the
very idea of understanding which we found to be so intimately
b~und up with and dep_:~dent upon _!he perfQr_mative fupctjon i~elf? \
Tbe question takes us to the Fourth Reverie and its imlicit shlti
. from repo_:ted guilt to the_guilt_o(repo.r:tmg,~~nce here the lie is no --;lj iiri"ger connOCt~e former misdeed but specifically with the
act o~iting the Confossions and, l;>y extens!g[J.,~hvri:tiftg+-QL
- course, we always were 1i1the realm- of writing, in the riarraf!ve of
the Confessions as wen as-irrthe-.Ri!7erte,t5lift'h~iwtion..o:f..thi
fact is now explicit: what can be said al5Ci'i:i'tthe interference of the
cognitive with theperformative function of_excuses i~ the Fourt. Reverie will disseminate what existed as a localized disruption in the


\.f Confossions.

- ---..__,_
-- _.._..r ~_vyith t~e complicity of the casual, ambli~g,_andfr~e-associating _J
mode of the Reverie, the text atl<JWSifSeTtap~ling lack 6f~u- \
Cast in the tone of a pietistic self-examination, it sounds
1 '\i
\ - \}
severe and rigorous enough in its self-accusation to give weight to the \ ~ \ _,.
exoneration it pronounces upon its author-until Rousseau takes it '~\ /\
[aif'back in the penultimate paragraph
~ d_ecrees him to be \hl\\
"inexcusable" (1038). There is a1so a strange unbalance between the
~~drifC of the argum~nt, which proceeds by fine distinctions and
ratiocinations, and the drift of the examples, which do not quite fit
their declared intent. The claim is made, for example, that, in the
Confessions, Rousseau left out several episodes because they showed ~~him in too favorable a light ; when some of these incidents are then
-being told in ~rckr to make the disfigured portrait more accurate,
th_e.}:' .!_urn o~ tc:_b~us!): irrelevan.!:_ They do not show-Rouss-eau \
.. in all that favorable a light (since all he does is not to denounce
playing companions who harmed him by accident and from whose
\ denunciation he would, at the time, have stood to gain very little),~-~
"""? \ ~n_9_!_hey are, moreover, mosi...unple.asanLs..taries of p.h.}!sica-1 ass:rl.i11,
l.?ody mutilation, and crushed fingers, told in such a way that one
'-1:' ,, remembers the pain and the cruelty much better tha .
e \ .\
~suppl:lst:~d-rcriTiustrate. Al.Itnis adds to the somewhat ncanny \






striking instance of the structural and epistemological implications of a nacoluthon

occurs in Proust in the description of the lies used by Albertine ("La prisonniere ,"A
la recherche du temps perdu [Paris: Pleiade, 1954], 3:153). Fof Rousseau's own
description of an anacoluthon-like situation, see note 16.

UrlL t:vr!~




obliqueness fa slightly delirious text which is far from masteri~

" e e ec s 1t pretends-to produce. -~
The impT~~he rando..!;!lli~ in the Marion episode C')e
m'excusai sur le premier objet qui ..-/
s'offrit") are distributed, in the
Fourth Reverie, over the entire text. The performative power of the lie 1
as excuse is more strongly marked here, and tied specifically to the
a15Sence of rel':erenhal significat1on; it .aisocarnes, in.this literary ', \
context, a more-fm'ffihar aiiE!=Peptffii'bTe name since it is now called
tion: "To lie without intent and without harm to oneself or to
==101Te: it is not aTie but a fiction" (1029TT1ie notion of
fi'chon Is introduced in the same way-th~ -Q.le-exsuse ofrandof!lness
functions -in the Confessions. Within- the airtight system of absolute
t"f'Ufliitproauces the - almost imperceptible crack of the purely '.z. ,
gratuitous, what Rouss-eau -calls un fait oiSeill:-;-fndifferent a taus
eg~et sans consequence pour personne
("a fact that is
totally useless, indifferent in all respects and inconsequential fi
anyone" [1027]). There is some hesitation as to whether such "perfectly sterile truths" are at all conceivable, or if we possess the necessary judgment to decide authoritatively whether certain state~
can be to that extent devoid of any significance. But although the text
vacillates on this point, it nevertheless functions predominantly as if
the matter had been settled positively: even if such truths are said to
be "rares et difficiles," it is asserted that the "truth" of such "useless
facts" can be withheld without lying: "Truth deprived of any concciVahle kind of usefulness can therefore not be something due [une
chose due], and consequently the one who keeps it silent or disguises
it ....______
does not lie" (1027). Moreover, "I have found there to be actual
instances in which truth can be withheld without injustice and disguised without lying" (1028). Some speech acts (although they might
better be called silence acts) therefore escape from the closed system
in which truth is property and lie theft: " . . . how could truths
entirely devoid of use, didactic or practical, be a commodity that is
-due [un bien du], since they are not even a commodity? And since
OWnership is only based On USe, there Can be no property where there
can be no use" ("ou il n'y a point d'utilite possible il ne peut y avoir de
propriete" [1026]. Once this possibility is granted, these free-floating
"truths''_g_r "facts," utterly devoid ofvalue ("Rien ne peut etre dude
cequf n'est bon a rien" [1027]) are then susceptible ofbeing "used"
as.-an excuse for the embellishments and exagg~rations that were
innocently added to the Confossions. They are mere "details oiseux"







from the slander back to the theft itself, which was equally unmotiand to call them lies would be, in Rousseau's words, "to have a
.. ..,
vated: he took the ribbon out of an unstated and anarchic fact of
'ConscienceTiiat is more delicate than' mine" (1030). The same para- --...
. "'>
graph call..s~these weightless, airy non-substances fictions: '\-vhatever,
Jl ~ pl'bximity, without awaren~iili pf 9ny l~w of ownership, Not the
' \ fiction itself is to blame for the consequences but its falsely referenalbeit contrary to truth, fails to concern justice in a ny way, is mere
''1 tial reading. As a fiction, the statt:;_ment js_ inn~C}l.QYS and the error )
ha rmless; it is the misguidedreading ofthe error as theft or slander,
L~fiction as If It were a he has a consCience tha t_I more delicate
~mine" (1030). Wharmakesancflcir1a -fi~tic";"n is not some polart~fliSal to 'cld mit that fiction IS ficfion, the SfUOOOrnr esistance -fO /l

the "fact," obvwus 6y rfSe1f;lnaf1anguage is entirely free with regard ct.

ity of fact and representation . Fiction has nothing to do with repreI'
to referential meaning and can po-sifwhatever its grammar allm~i?_it to
~entat ion but iS" the abse~ce of any link between utterance and a
referent, regardless of whether this link be causal, encoded, or govr-~ sa~f;;~;tg_ the transformation of random error into inju~ti~e>
The radical irresponsibility of fiction is, in a way, so obviol:!_S, that it
v erned py any ot~er conceivable relatio_g~hip that could leRd" itself t90
hardly necessary to caution aga-inst its misreading. Yet its
-s.ystema tization. In fiction fhli'SCOnceived the-conecessary link" of the
assertion, within the story of the--eonfessions, appears paradoxical
metaphor has been metonymized beyond the point of catachresis, . _
and the fiction becomes the disruption of the na rrative's referential \
nd far-fetched to the point of absurdity, so much so that Rousseau's /
,' illusion. This is precisely how the name of Marion came to b e uttered
own text , against its author'~ interests, p~e~ers being suspected of!~ ~
1 in the key sentence in the Confessions: ']e m 'excusai sur le premier
, and slander rather tha n of mnocently lacking sense. It seems to be
--:-o ' esBiblrtorsolate the moment in whicn the fiCtion stands free of
oqjet qui s'offrit," a sentence in which aQJ-_authropomorphic conn~.
]_ation of seduction implied by the verb "s'offrir" has to be resisted if
. , ' '; my si_g::~cation; in the ve~o:r:_e&a"C~]iich it is positea! as w~ll as
) in the contexf fhat it generates, it gets at once misinterp~eted into a
I the effectiveness of the excuse is not to be undone and replaced by
determination which is, ipso factO, overdetermined. Yet w1t1iout this
the banality of mere bad faith and suspicion. Rousseau was making
l) 11 ,. moment, never allowed-to e xrs! a~-su~,n9 su]:Ii llllng as a text is
whatever r~jseJ:!? ppened to com e into his head; he was saying noc~
\ conceivable . We know this to be the case from empirical experience
_).bing a t all, leas_t_of_alLsomeone's name. Because this is the case the
statement can function as excuse, just as fiction functions as a~
- as weT!: iTis always possible to face up to any experience (to excuse)
excuse for the disfigura tions of the Conftssions.
any guilt), becaus-e the experience always exi~ts sim_!-!ltan~c:_msly as
:(ic.ti.ouaLdis.c~ and as emRirical _~~en ~nd it i~ never possible t?_.j
__:__ It will be objected tha t ficti"on in the Reverie and the denun
ciation of Marion~ miles apart in that the_ former is without ~ ~
Q-ecide which m_:e _?f the two_possibilities ~the ~ight one. 'fhe indeciconsequence whereas tEe 1atter results- in considerable damage to
siarr-m.akes 1tpossible to excuse the bleakest of crimes because, as a
fiction, it escapes from the constraints of guilt and innocence. On the
' others: Rousseau himself stresses this: "whateveris cor-1trary to truth
other hand, it makes it equally possible to accuse fiction-making
a~h urts justice in an conceivable way is a lie" (1030), and also
. . _. \
"the absence of a purposefully ar
ntent -dues not suffice to " ....1
which, in Holderlin's words;Is-'~1he mo~t innoc~~t of all a_ctivities," of Illbeing
the most cruel. The-Kilowledge_of radicaLinn.o.c.ence also permake a lie innocent; one must a lso be assured that the error one ""'
~e harsiiesl ""1nutilations. Excus~s not only accuse but they
inflicts upon one's interlocutor can in no conceivable way harm him ~
c;rry out t he ~ictimplicit-iH ~~he~usation.
or anyone else" (1029). But the fiction , in the Confessions, becomes"-. j
This other aspect ~se i~l
nveyed by the text of
liar~mful only because i~ .is not~~n~~rstood for wha~is~b~c~use
fictional statement, ""ITS It generates tbe syste!11'"'Uf"Shame, desrre, and
the Reverie, though necessarily in a mo~lique
nner. In telling /
ar\,other instance of a situation in which he lledouf of shame-a less ""'
: repression we described earlier, is a t gnce caugh!_ ;~md enm~ in a
interesting example than the ribbon, because ther s nothing enigof
a b.out a_lie which, in this case, is only ra efe~e__ :3-Rousseau
se;m's accusersh"ad realized that Marion's name was "le premier
'-. ."'"
1\, /\
.'J Jf
l f'vWUv"'-'"l , c
oqjet qui s'offrit," they would have understood his lack of guilt as
\.S}- 13 . In this case he is being provoked into lying by the""ha lf-teasil(_g, 1halfmalicious questions of a woma n inquiring whether he ever had children.
well as Marion's innocence. And the excuse would have extended_






vf .





u /') \


'- "'"""'


r )



r! l





\) ~




V V'



writes: "It is certain that neither my judgment, nor my will dictated

my reply, but that it was the_!utoma!ic result, [l'rjfet machinal] of my c
embarrassment" (1034). The machin:~g_uali_!Lofthe text ofthe lie
is more remarkable still when, as m the Marion episode, the dispro- \ \
portion between the crime that is to be confessed and the crime \..-performea oy tbe li~ adds a deiirious _dement_to the si~_lti<;m. By
saymgmar the excuse is not only a fiction but also a machine one f'
I ~
adds to the co~ of_:e[erential c!etgch!llept, ?fgratl!_itous_irn-_ ~
provisation, that of the implacable repetition of a preordained pat~1_!1 Like Kleist's marionettes, the machine is both "anti-grav," the -..J~
anamorphosis of a form detached from meaning and capable of "-'
taking on any structure whatever, yet entirely ruthless in its inability
J ~dl_:ty its own structural design for nonstructural Leasons. The
machine is like the grammar of the text when it is isolated from its
, rhetoric, the merely formal element without which no text can bfJ~'~)
generated.-I.l2_e~:~can be no use of language which is not, -~n..a_ '.
certain perspective thus-radlcaiiyrorffiai, 1.e. mec~o- ~atter
how deeply this aspect may be concealed 6y aesthetic, formaliStic
delusions-:- - ---~
- ~
The machine_ not only gen;rates, but also s~ppresses, a,nd not
al~ys in an inn~~ or balanced wax- The economy of the Fourt
...Reverie is curiously inconsistent, although it is strongly thematized in
a text that has much to do with additions and curtailments, with
- -- - - -------
"filling_hole;;_::_.("remplir les lacunes" [1035]) and creating them. The
part; of the text which are destined to be mere~gn;:and exem-,{ ~
r.lifications _acqui~_utonomo~p~wer of signification to the point \\ ~
!Iere they can be saiQTc)reduce t~n_!Igument to impotence.
0{ '
The addition of exampks leads to the subversion of the cognitive \ ' y
1 affirmation of innocence which the examples were supposed to ill ustrate. At the end of the text, Rousseau knows that he cannot be

excused, et the .ex~~r~"iliel[fro~o~_l?Y- the~~ of its r icaf1iC'fio~lity.:. _

'!'he _liteal 'censorsnlg) md curtailment of texts appears prominently in several places. A quotation from Tasso provides a first
example: Rousseau compares his own resolve not to denounce his
playing companion t~ S..ophronie's -sa-crificial lie when_, in order to
save the life of the Christians, she confessed to a crime (the theft of a
religious icon) th!_!t did not take place. The comparison bQ~ders on
the ludicrous, since Rousseau's discretion is in no way equivalent to a
sacrifice. BU1tnequotation -wltrch""'Roosseau now inserts- into'"th-e- text


















14. The translation is available in several of the early Rousseau editions, for
example in Oeuvres completes de].]. Rousseau (Aux deux ponts: chez Sanson et ,
Compagnie , 1792), 4:215-47. It is rrinted in bilingual version and evet'l the ear:ly ' \
e01tors had_o.Qs~rved and indicated the -absence of the passage which ~as later to be ,
quoted in the Fo~rth ~erz'ZliQid ., 229).
. ~
15. On Rousseau and Tasso, one finds general observations, not very informative in this context, in several articles, mostly by Italian authors, mentioned by
Bernard Guyon in his notes to the Pleiade edition of the Nouvelle Heloise (2:1339}.
16. The statement is not a quotation from Rousseau but is report~d _b)' CQrancez in De].]. Rousseau (Extrait dujournal de Paris,# 251, 256,259,260,261, An 6,
42"=43) . The sequel of the statement, in which Rousseau describes the one exception
to the organic integrity of T~s;o's work, is equally interesting for our purposes and
c~i!d a~_Eousseau'p description of an anacoluthon:" . . . sans que le poeme
\ en tier ne s'ecroule, tant(leTasse)" etait precis erne mettatf"l'ien que de necessaire. Eh
bien, ote~ la strophe ~re dont je vous par~; rien n'en soutrre, l'ouvrage reste
,..J parfait. Elle n'a rapport ni ace qui precede, ni a ce qw suit; c'esru:!1_e ptece absolu\ t,m:_riT inutile. II est a presumer que le Tasse I' a faite involontai~ement et sans l~
~omprendre lui-meme; mais elle est claire." Corancez could not remember the I
'stanza Rousseau quoted;-but it ha;;been tentatively ider.ilifte_d..as.stanza.77 of Canto..._
XII of jerusalem Delivered~~~ de].]. Rousseau (Paris: F.
Alcan, 1923) , p. 327 and Oeuvres completes, 1:1386-87, which Rousseau chose to
i 1 ead as the prefiguration of his own p~rsecutions . Corancez tells the story as an
~r ', in. stance
-:onsse~'s growing paranui~nd, in the-same article, he r~ports RousI seau's death as suicide. His article is vvntte)} in defense, however, of Rousseau's
; '
17. Social Contract (3:306).



V""' :,;

serves a different function. It is a passage which he had omitted,

without apparent reason, from the translation he made of the Second Canto ofTasso's epic 14 ~ earlier date. Any ~e;tion ofTasso:
irtRousse~ always carfies a high affective charge and generates
stories clustering around-clttbtm:ts-trcmslcrtimTs; -liierary-falsifi.cafions, "
te?tlial" distortions, fallacious prefac~~ as~_ell as obs.eE_~OJlLof i~.,.,
t1fitation-trrvolvingeroficfantasi~nd anxieties ofinsanity. 15 Limit:
1ng-oneself;--in4his--cofftext, 1o the obvious, the insertion of'the-quota- )
tion must be an attempt to restore the in;egrity of a text written by -someone of w110m~ousseau hirriSelf had said "that one could not
~-from his work a single stanza, from a stanza a single line,
and from a lihe a smgle -word;-Withoutt he- entire poem- colla~
ing .. .. " 16 But the restoration- occursa s---ailentirely private and_ \
t U_~ike the Citizen Stealing "en Secret" the W~
Ord-- I
Secretive ge~-~' ??_
"-'-~-- ---" - --d thinking of himself when he votes for all. 17 Sue _a
enforces the sha_ll}efulness of the crime as well as
destroying any hope that it could be repaired. The mutilation seems










;;; ;/'C'




to be incurable an the protl1esls only serves .!_o...222_ark this fact more \

~over the cntir;;:'"Fourth Reverie
strongly. The accus tion t
-m;-d against which the excuse tries to defend itself seems to have to
d~-a fnrea} ofte;,:t_ualmu-ti~~.tJ_o.:_1,_ itselfl. i~ke~ to the organic and
totahzmg synechdoc.al language by means of wlll_<:;h Rousseau refers ~
to the 1J.l1ity.....of Tasso's work.
The omission and surreptitious replacem~..!l!:._9f_the Sop.b_ronie
_eassage iS"at most a symptom, all the more so since "Tasso," in _J.
ROtrssrnu, implies a. threat as well as a victim, a ~12on as well as a
~c!- The mutilatio~ i~- rio(just the excision of one SE_ec~fic piece of
e~t. Its wider significance becomes more evident in another literary -~
~usio~ in the Fourth Promenade, the reference to Montesquieu's ~~
y:onventionally deceptive preface toLe Temple de Gnide. J?y pretendj ~-'
l ing that his work is the translation of a Greek manusc~ie!;_the autho -~
/ shelters himself from the possible accusation of frivolity or licen{ tio~owing that the reader who is enlightened ~nough not
,(to hold his levity against him will also be sufficiently informed about
literary convention not to be taken in by the phony prefa:e. Rousseau ~
treats Montesquieu's hoax without undue S~ve1 ity-r'tould it have
occurred to anyone _to ~minate the author for this lie and to call
him an ir.n}2ostor?" [1030]), yet behind this apparent tolerance stands ) '
a much less reassuring question. As we-know from"The-'-'Prefcrc:-e- ~
diaioguee" to the Nouvezze-He1oi5e-;1h.e...p~e is th~pja~~jn...J.b_e...text~
where the questLon oftextual mastery and autbor_ity_is being decided .
in the
of]ulie, it is also
W1fh tb1s threatemng loss of control the possioi1rty anses of the \'
entirely gr~tuitous and irresponsible tex_!;.!!.qtjust (as was apparently
the case for Montesquieu m:-ror naive readers ofjulie) as an in ten,
tiona! denial of paternity for the sake of self-protection, but_as.....the \ \ )
radical annihilation of the metaphor ofselfhood and of the wi
. more than warrants the anxiety wit~ch Rousseau clui , rledge ~ ~ \'
the l_eth.al q.u~li~. of ~1. 1 wrlti1~g. Wri~iHg always i. nclug~ the rr:_o~eq_t )S'
. ~ d1s_pos~esswn 1_n favor. of the arb1 tra?' pow~r play oft -~
f _nd from_~~ew ~f th_:_ subject. , th~ ca'?- only b~expen-~
\ enced as a ai~ nberm t, a beheading or a castration. Behind I
t ~o_ntesquieu'~ar~less _lie, enying a'uthorship - otLe- Temple de
f Gnide by the mampulatwn of the preface that "}}~" the text, stands -tfie mucfi more dangerous ambiValence ofthe "beheaded"
author. JH



an_~ wh~re,




\ \'





fo~ ~=-~~~ec~dabl:_._\~



-S\, ~ ~-" \

t1, '


1 -

v. ~ ' 1



18. Tne same pnxiety is apparent in another referenc~ to prefaces in Rousseau,







t; '


~ But precisely because, in all these instances, the metaphor for
the text is still the metaphor of text as body (fro1n which a more or
less VItal part, inc-tudirrgthel'le-a:d, is being severed), the threat remains sheltered b~hind its metaphoricity. The possible loss of au- )
thorship is n~t without consequences, liberating as well as threaten- .
ing, for the empirical author, yet the mutilation of the text cannot be
taken seriously: the clear meaning of the figure also prevents it from
carcyi,ng out what this m eaning-1mplies. Thecundecidability ~1~
thorship is a ~ionof considerable epistemological importance
but, as a cognition, it remains ensconced within the figural delusion
~wi~g from -doing. OnTy wflen Rousseau no1onger confronts Tasso s orMontesquieu's but his own text, the Conftssions, J
does the ~etaphor of te_::t ~s body make way""f!M -e more directlY]
tl;lreatening alt~rnative of thq~xt as machine.
Unlike the other two texts, where the dis ortion had been a \.
s~th-Borif#s-&iblns...is...a.Lfus.Lguilty of disfiguring l;Jy excess, C
b.}LJ.h.g__?ddi.tion __of ..):!per:fluo~s 1Jictign_al embellishment~, ."I have
never said less, but I have sometimes said more . . . " (1035), but a C
few lines later it turns out that this ~a; -;:;~t ,the case either,
since 1 \~
~usseau . .admits..hamug-omi-tt-ed-sttme-e-t-:.his...u:.c.ollections fro_!Tl tpe
61\. narrative merely because they showed him in too favorable a light.
~ \} There is less contradiction between the two statements when it turns
out that what he omitted are precisely stories that narrate mutila:
)> tions or, in the metaphor of the text as body, suppressions. Botn
~ ~- la~e 1...99_ ~th mutilati~m a~~ehe:ding: he neaJ:.ly loses a
~\ hand in the first and comes close to having his i:Jfains knocked out in
the other. Thus to omit suppressions is, in a sense (albeit by syllepsis), to preserve an mtegrity;'"~Te-ja-m-a-i-s-dir:e-mo.ins." If-the-stories- that hav~itted tbreaien the integrity of the text, then it










interestingly enough also in connedlcm with risso. To deny authorship in a preface

in the name of truth (as Rousseau did in the case of]ulie) does not only mean that
one's authorship of all texts can b e put in question but also that all texts can be ~
attributed to one. This is precisely what happens to Rousseau when a malevolent (or
commercially enterprising) editor, in what reads like a transparent parody of the
"Pniface dialoguee," attributes to him a poor translation ofTasso's]erusalem Delivered, (see Oeu~res completes, 1:1740 for the text o~th e editor'~ prriface, and .also
Oeuvres completes, 1:1386). Rousseau mentwns the mCident w1th some degree of '
paranoid anxiety in a letter to Mme. de Lessert of August 23, 1774, and am6flg many
6il1e'rmsian~s-ufiah;e textual attribution, in the Dialogues (960). The chain that
le/a~om Tasso to translation, to prefaces, to authorship, to beheading, a nd to
insanity iS ready tbsUfface- in any context of anxiety about truth and falsehood_

J J?' ):\




be even easier to excuse him for not having included them

~ to :xcuse ~im fm _the superfl_uous ornaments he added to the


\' ... ~
ollectwn of his happier memones.
-- ~~!\!elY are these narratives t~r_ratenin_g;? As instanc24s
of Rousseau's generosity they are, as we alrea_ciy pointed out, more
i~ than con"!_ncing. They seem to exist ~25im_ ily for the_sake of _
they describe. But
lw m, tn theic iilln, to be there moc_e .or th.oE._gf a11ow;'l!l_ f!':e ~ '>',
evocation of the machinethat causes them i h'l!l.for their own shock
,. .,_
.~ Rousseau lmgers complfcently~~T t~e descri2tion of tl_::e
machine that sedu_ces_ ~im into dangerous~ose contact: "I looked
at the metal rolls, my eyes were attracted by their polish. I was
tempted to touch them with my fingers and I moved them with ~
f pleasure over the pohshedsllrface 6fthe-cylinder . . . " (1036). In the
'general economy of the Reverie, the machine displaces all other sig'v nifications and becomes the raison d't~tre of ~_!ext. Its power of \ 1
~~ suggestion reache.s far beyond its illustrative puspose, especially if
/1 ~e bears in _m.i.nd the previous chara~teriz_ation_?f ~nmotivated, )
1 , fictiOnal language as "machi!:al." The underlying structural patterns
of addition a~pression as well as the figural system ofthe text all
converge trvards~) Barely conceal~d by it_s periphe~al f~n~~o~he l
t~ here ~g;s _!!}e _textual 1_31~c~I~e o_~s own ~Itutwn and _.._.
perrormance, Its o~ textual allegory. The threatening eleme
!.hese incidents then ~c~mes more apparent. The text as body, with
all its implications of substitutive tropes ultimately always retrace-=-
able to metaphor, is displaced by the text as machine and, in the
process, i~- the lc:ss ofthe illusion of meaning. The deconstrue==
tion of the figural dimension is a process that takes place indepen~ntly of any desire; ~s su::h it is not ~ncon;;-~io~s bu~ IT:echa~ical,\
systematic m itsperformance but arbitra m Its Q.QnCIle, 1IRea
grammar. T is threatens the autobiographical subject not as the lo~s
'l'~ing that o.nce was present-and tha:t it once poss essed , ~as ..
a radical- estrangement between the meanmg and the performance
of any text.-- - -- ----- -1
--=-In order to come into being as text, the referential function had
to be radically suspended. Without the scandal of random denuncia- ,.,.
{tion of Maricm-, -without. the "faits oiseux"~f the Confts~ions, ther~ \
f c~uld not have been a text; there would have been nothing to~excuse \'~
sm~e everything co~ld have !::_een_ ~xplajned a~ay by the cognitive
logic of understandmg. The cogmtwn would have been the excuse,


\ ('
this convergence is precisely what is no longer conceivable as \
sdon-aslhe metaphorical integrity of the text is put in question-;aS
'soona-s ihe text is said not --to be a _flgural body but a machine. Fa1
f~ seemg language as an instrume~ ii1Tiie service of a psychic
energy, the possibility now arises that the entire construction of
drives, substitutions, repressions, and....:::=presenta~ions is__.!_he aber1 rlint, metaphorical correlative of the absolute randomness Of1an~
\ guag~ to any figuration or meaning. It is no longer certain that
'language, as excuse, exists beZa~se of~ prior guilt but just as possible
that since language, as a machine, perform$ anyw~y, we have to
p~t.im:!Q all its train of psychic consequences) in order to
make the excuse meaningful. Excuses generate the very gullfli1ey
xonerate, though always .in excess or by default. At the end of the
Reverie there is a lot more guilt around Than we !-lad at the start:
Rousseau's indulgence in what he calls, in another bodily metaphor,
"le plaisir cfecrire" (1038), leaves him guiltier than ever, but we now
have also the two companions of his youth, Pleince and Fazy, guilty
~ault, brutality or, at the very best, of carelessness. 19 Additional
\ guilt means additional excuse: Fazy and Pleince now both have to \
apologize and Il_lay, for all '~e know, have written m~vin?_!exts_?-bout_
\ !.. qreadful thmgs they did to jearr-jacques who, m his turn , now
as to apologize for having possibly accused them arbitrarily, as he
ccused Marion, simply because their names may have happened to
occur to him for the least compelling ofreasons. 20 No excuse can ever
hope to catch up with such a proliferation of guilt. On the Other
hai1d, any guilt, including the g~ilty pleasure of writing the Fourth
Reverie, can always be dismissed as the gratuitous product of a
text.ual grammar or a radical fiction: there can never be enough guilt
around to match the text-machine's infinite power to excuse. Since
~il_h in this description, is a cognitive a nd excuse a performative
function oflanguage, we are restating the disjunction of the perforai~d

thes~ actual~



1 '\."SV
"""...J\) .


- ~

19.~escription of the .way in which Fazy i1~ured Rousseau is-ambiguous, \

since the narrative is phrased in such a way that he can be suspected of having done
it with deliberation: " . .. le jeune Fazv s'etant mis dans Ia roue lui donna un
demiquart de 1our si adroitement qu'il 1;'y prit que le bout de mes dem,: plus longs
doigts; mais e'en fut a~sez pour qu'ils fussent ecrascs . . . " (1036).
20. For example, the f~ct that their names may have come to mind because of
the![ Bhopic resemblance to the place names where the incidents are said to have
take1'pL3-ce: the one involving Fazy occurs at Paquis, the one involving Pleince at

. tropologica. l c. o. gnitio_n s, ! h_e sy.stematic un do. ing, in other words, of

. understanillng:-fto~ch, far ~closing_off_tlle tr_:_opological systeJ!1 2

-~c:_s(fh:_r:p_:~iti~~ ab~




t J\


21.- The s_!:nilarit v betwe~~t_!1 o n a nd p;Jrabasi,!;i_stell}!! from the fact

that both figures-~~~::~~ the expectati~ of ~ given ,&_am t~~tl m:r-I1efmica~

movcn-rent-. J\s d1grcsswn , asiae,-"mtelventwn dauteur, or-waus der Rolre;arr;,;,'
par<tbasis clea rlv involves the int errupt ion of a discourse. The quota tion fiom Friedrich Schlegel appears among the formerlv unavailable notes contemporarv wit h the
Lyceum and Atheneum Fragmenten. Friedrich Schlegel, Kritische Friedrich-Schlegels:sabe, ed. Ernst Behler (Munich, 1963), 18:85, !:i668. The use oft he term parabasis
(o~arekbasis) by Schlegel ec hoes the use of the device especially in the plays of




Friedrich Schlegel's [oi2!21Jl_ation,---i.t- becomes the permanent

para basis of an allegory (of figure), that is to say, irony. Irony is no
~ longer trope but the undoing of the deconstructive allegory of all

mative from the cognitive: any speech act produces an excess of

cognition , but it can never hope to know the process of its own
rPm~ction (the only thing worth knowing) .just as the text can never
stop apologizing for the suppression of glJilt that it performs, there
is never enough k.rfowl~dge available tO' account for the delusion of knowing.
-- ~ m ain point of the reading has been to show that the result_Predicament is liQguistic rather:Jh~tologic~.E_ermeneuti~.
~ A~ was clear from the Ma rion episode in the Confessions, the deconstruction of tropologica l patterns of substitution (binary or ternary)
em:~~ inclu. ded within. di~courses that lea:e the ass~mpt. ion ofintelligJbihty not only unquestwned but tha t remforce th1s ass\!mptJ<2n b
making the mastering of the tropological displacement the very burden of understanding. This prqject engenders its own na rrative
whi~h ~~n b e caiTed- a n allegory of figure. This narra tive b egins to
vacillate only when it appears that these (negative) cognitions fail to .,.
~ake' the performati~tio~ 2_f the discourse -~dictable _a nd
that, consequently, the linguistic model cannotbe reduced to a mere _)
system of tropes. Performative rhetoric and cognitive rhetoric, the 1
rl1etoric Of tropes, fail to converge. The chain of substitutions)
functions next to a nother, differently structure ~bar exists
independently of refe!>o.iia+--d
tion, in a system that is both
enJ.!.rely arbitrary cflld entirely repeatable , ke a gramr.::ar. The inter- (
section of the two systems can bemcared ii13- fexT as the disruption
of the figural chain which we identified, in the passage from the
Confessions, as a nacoluthon; )n the la nguage o_r....;:,;presenta~1~l
rhetoric , on:_:ould a lso callltparabasis, 21 a sudqen reveh!ti~e
disconti_~~ b_etween tkQ:Jlie.tuJ:ic.t~~.;r'-his isolated textua
-~~the reading of the Fourth Reverie shows, is dissemina ted
throughout the entire text and the anacoluthon is extended over a ll
the poir:ts_of
the figural line or allegory; in a slight exten sion of


~h : \