Syllepsis

Michael Riffaterre

Critical Inquiry, Vol. 6, No. 4. (Summer, 1980), pp. 625-638.

Stable URL:
http://links.jstor.org/sici?sici=0093-1896%28198022%296%3A4%3C625%3AS%3E2.0.CO%3B2-R

Critical Inquiry is currently published by The University of Chicago Press.

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained
prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in
the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/journals/ucpress.html.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic
journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers,
and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take
advantage of advances in technology. For more information regarding JSTOR, please contact [email protected].

http://www.jstor.org
Fri Sep 7 15:29:42 2007
Michael Riffaterre

Intertextuality inheres in the two fundamental and defining features of

the literary text: it makes the text what it is, a semiotic and a formal unit.
In a nutshell, the very idea of textuality is inseparable from and founded
upon intertextuality.'
Intertextuality is a modality of perception, the deciphering of the
text by the reader in such a way that he identifies the structures to which
the text owes its quality of work of art. Like all structures, these are
actualized in the form of variants. Such variants in a text must necessar-
ily be verbal shapes, that is, words considered at one or more of three
levels: phonetic, lexical, and syntactic. These words which serve to actu-
alize structures are not intrinsically different from other words. Rather,
they are now perceived within a particular grammatical sequence and
from a different viewpoint; hence they mean differently. This dif-
ference is what separates meaning from significance. Or put it this way: I
propose to describe how a literary text yields sense by distinguishing two
semantic systems for any verbal sequence forming such a text-meaning
and significance. I shall speak of meaning when words signify through
their one-to-one relationship with nonverbal referents, that is, their ref-
erence to what we know or believe we know as reality. I shall speak of
significance when these same words signify through their relationship

1. O n intertextuality, see Julia Kristeva's Simiiotikl.: Recherches pour une simanalyse

(Paris, 1969), esp. p. 255; Poetique 27 (1976), special issue devoted to intertextuality; Lau-
rent Jenny's "Semiotique du collage intertextuel," Revue d'esthdtigue 3-4 (1978): 165-82; and
my Semiotics of Poet~y(Bloomington, Ind., 1978), pp. 1 1 5 5 0 .

0 1980 by T h e University of Chicago. 0093-189618010604-0009$01.44

626 Michael Riffaterre Syllepsis

with structural invariants (no one-to-one relationship this time since

there must be two or more variants for one i n ~ a r i a n t ) . ~
Now variants can be recognized as such only by means of compari-
son: when it comes home to us that several statements are indeed con-
nected, despite their differences, through their identical relationship
with another statement, we realize that they are, so to speak, re-
formulations or translations into different codes of an archetypal mes-
sage. There are two possible (by no means mutually exclusive) ways of
reading comparatively, reading so that comparison forces itself upon the
reader's attention: retroactive reading and intertextual reading.
Retroactive reading occurs at every step of normal (from page top to
bottom) reading, growing more important as more textual space is cov-
ered. Working forward from beginning to end, the reader keeps re-
viewing and comparing backward, recognizing repetitions, recognizing
that some segments of the text are variations upon a semantic sameness
and therefore variants upon the same structure(^).^ Intertextual reading
is the perception of similar comparabilities from text to text; o r it is the
assumption that such comparing must be done even if there is no inter-
text at hand wherein to find comparabilities. In the latter case, the text
holds clues (such as formal and semantic gaps) to a complementary
intertext lying in wait somewhere.
We must be careful to avoid the confusion between intertextuality
and intertext that spoils most of the studies newly born of the current
craze for intertextuality. The intertext proper is the corpus of texts the
reader may legitimately connect with the one before his eyes, that is, the
texts brought to mind by what he is reading. This corpus has loose and
flexible limits. Theoretically it can go on developing forever, in accor-
dance with the reader's cultural level; it will expand as his readings
expand and as more texts are published that can be linked up to the
original point whence these associated memories took their departure.
The ability to connect or collocate texts does not, however, result from
2. The distinction I draw between meaning and significance is thus entirely different
from E. D. Hirsch's in Valzdity zn Interpretation (New Haven, Conn., 1967), p. 8. As I see it,
significance is the product of a second reading stage, and in the bipolar relationship
between meaning and significance, meaning appears as the continuously changing pole.
3. On retroactive reading, see my Semiotics of Poetry, pp. 5 , 9, 16566.

Michael Riffaterre, Blanche W. Knopf Professor of French Litera-

ture and chairman of the department of French and Romance literatures
at Columbia University, is the editor of Romanic Review. His most recent
books are Semiotics of Poetry and La Production du Texte. "Syllepsis" devel-
oped out of seminars he led at the Irvine School of Criticism and Theory
and at Johns Hopkins University. His forthcoming books, Typology of
Intertextuality and A Grammar of Descriptive Poetry, are scheduled to appear
next year.
 Critical Inquiry Summer 1980 627

merely superficial similarities of wording or topic; two or more literary

passages are collocable and comparable as text and intertext only if they
are variants of the same structure. Intertextual connection takes place
when the reader's attention is triggered by the clues mentioned above, by
intratextual anomalies--obscure wordings, phrasings that the context
alone will not suffice to explain-in short, ungrammaticalities within the
idiolectic norm (though not necessarily ungrammaticalities vis-a-vis the
sociolect) which are traces left by the absent intertext, signs of an in-
completeness to be completed elsewhere. These, in turn, are enough to
set in train an intertextual reading, even if the intertext is not yet known
or has been lost with the tradition it reflected.
This kind of zero intertextuality (unfulfilled or unactualized) makes
it somewhat paradoxical to speak of comparability where comparison is
no longer possible or has been postponed. I shall therefore use pre-
suppositions instead and talk of significance where words signify by pre-
supposing an intertext either potential in language or already actualized
in literature. Thus one of the basic components of a text's literariness
(along with catachresis, overdetermination, and closure) is that the text is
not simply a sequence of words organized as syntagms but a sequence of
presuppositions. In literary writing every lexical element is the tip of an
iceberg, of a lexical complex whose whole semantic system is compressed
within the one word that presupposes it. T o put it otherwise: the literary
text is a sequence of embeddings with each significant word summariz-
ing the syntagm situated elsewhere. So far as I can make out, there are
three types of intertextuality: first, the complementary type (every sign has a
reverse and an obverse; the reader is forced to interpret the text as the
negative, in the photographic sense, of its intertext); second, the mediated
type (where the reference of text to intertext is effected through the
intercession of a third text functioning as the interpretant that mediates
between sign and object, in Charles S. Peirce's terminology); and third,
the intratextual type (where the intertext is partly encoded within the text
and conflicts with it because of stylistic or semantic incompatibilities).
These three distinctions should help check the tendency, now all too
general, to see intertextuality as nothing more than a newfangled name
for source or influence. Influence from text to text, or the linkup of text
with source, is a "vertical" relationship of recurrence and sameness,
whereas intertext is related to text "laterally": there is a simultaneity and
otherness, a contiguity, a mutual solidarity, so that the text functions as a
literary artifact only insofar as it complements another text. The text's
ungrammaticality is but a sign of a grammaticality elsewhere, its sig-
nificance a reference to meaning elsewhere. Again, it would be wrong
to confuse the intertext with allusion or q ~ o t a t i o n for, ~ the relation

4. On the poetics of quotation, see Antoine Compagnon's La Seconde mazn; ou, Le Tra-
vat1 de la n t a t z m (Paris, 1979); see also "Allusion Studies: An International Annotated Bib-
liography, 1921-1977," comp. Carmela Perri, Style 13 (Spring 1979): 178-225.
628 Michael Rgaterre Syllepsis

between these and the text is aleatory-identification depends upon the

reader's culture-while the relation of text to presuppositions is obliga-
tory since to perceive these we need only linguistic ~ o m p e t e n c e . ~
Having posited my basic principles, I now propose to examine more
specifically what relationship exists between intertextuality operating as
a constraint upon reading (as a set of restrictions upon the reader's
freedom, as a guide for him in his interpreting) and ambiguity, or the
kind of obscurity that prevents the reader from quite discerning which
of a word's pertinent meanings are equally acceptable in context.
Ambiguity is generally, if empirically, recognized as a feature typical of
literary discourse. As I see it, ambiguity exemplifies the idiolectic
ungrammaticalities that warn the reader of a latent intertext. Text and
intertext alike derive from these ungrammaticalities.
Ambiguity is not the polysemy most words display as dictionary
entries but results from the context's blocking of the reader's choice
among competing meanings, as when, to use an example from Derrida,
a French context hinders the reader from deciding whether plus de
means "lack" (no more) or "excess" (more than).6 In this case, the un-
decidability is due entirely to the fact that the reader is playing a score,
the syntax, that will not let him choose. This must be because the score is
badly written; yet it is precisely this sort of willful neglect that critics have
labeled poetic licence, thereby underlining its literary nature. Un-
decidability has become a central feature in Derrida's analyses of literar-
iness, and it is also the main underpinning of his creative ~ r i t i n gBetter
.~
still, his own critical discourse has put undecidability to use, not a
rare case of metalanguage imitating the very devices of the language it
purports to analyze. My examples are therefore drawn from Derrida on
the assumption that his conscious practice of tcriture, backed up by a
sophisticated theory, will be particularly illuminating. For my own
analysis of these phenomena, I shall be using a special word that Derrida
has adopted and adapted from the terminology of ancient rhetoric. He
proposes it in his commentary on this sentence of Mallarme's: "La scene
n'illustre que l'idee, pas une action effective, dans une hymen . . . entre
le desir et l'accomplissement, la perpetration et son s ~ u v e n i r . Our
" ~ critic
5. Compare Jonathan Culler's criticism of Harold Bloom in "Presupposition and
Intertextuality," Modern Language Notes 91 (1976): 1380-96.
6. See Jacques Derrida's La Dissiminatia (Paris, 1972), p. 307.
7. Because Derrida is a philosopher by trade, one would expect his undecidability to
reflect the very precise logical and mathematical concepts of that discipline-which is to
say, the limitations inherent in the axiomatic method. Kristeva, for example, tries to d o this
in her Le Texte du roman: Approche simiologzque d'une structure discursive transformatianelk
(The Hague, 1970), pp. 7&78. So far as I can make out, however (as a layman 1 have
hardly been able to go beyond the relatively simplified but highly instructive exposition of
the problem in Ernest Nagel and James R. Newman's Godel's Proof [New York, 1958]),
Derrida's critical theory and reading practice d o not pack more into the word "undecid-
able" than does the definition I offer in this paper.
8. Stephane Mallarme, Mimigue, Oeuz~rescomplites (Paris, 1951), p. 310: "the scene [a
drama, o r rather a pantomime] bodies forth only an idea, not an action: it is like a hymen
 Critical Inquiry Summer 1980 629

points out that the grammar prevents the reader from choosing between
hymen as "marriage," a symbolic union o r fusion, and as "vaginal mem-
brane," the barrier to be broken through if desire is to reach what it
desires. Undecidability is the effective mechanism of pantomime as an
art form since from mimicry alone, without words, the spectator cannot
tell whether a dreamed, or a remembered, or a present act is being set
forth. This, in turn, Derrida shows to be fundamental to Mallarme's
concept of poetry. It is simply a pun or, as Derrida prefers to call it, a
the trope that consists in understanding the same word in two
"~yllepsis,"~
different ways at the same time, one meaning being literal or primary,
the other figurative.1° The second meaning is not just different from
and incompatible with the first: it is tied to the first as its polar opposite
or the way the reverse of a coin is bound to its obverse-the hymen as
unbroken membrane and as a breaking through of the barrier. The fact that
hymen is also metaphorical in both its meanings is irrelevant to its un-
decidability. What makes it undecidable is not that it is an image but that
it embodies a structure, that is, the syllepsis.
T o the three types of intertextuality I have listed above there corre-
spond three different roles played by the syllepsis. With the complemen-
tary type, the syllepsis itself suffices to presuppose the intertext and
by itself conveys the significance. With the mediated type, the syllep-
sis refers to the textual interpretant. With the intratextual type, the
syllepsis symbolizes the compatibility, at the significance level, between a
text and an intertext incompatible at the level of meaning.
The complementary type.-Mallarme's hymen illustrates syllepsis gen-
erating this type of intertextuality. It is rare, however, to find two
normal, current meanings of the same word activated s i m ~ l t a n e o u s l y . ~ ~
Yet a word may have only one meaning and still be turned into a syllep-
sis. For that to happen it suffices if the context's lexicon has semantic
features opposable to the semantic features of the word. For instance,
Derrida's own metalanguage builds upon undecidability a commentary
on the wellspring image in Paul Valkry's poetry-not the picturesque
wellspring, the natural forest fountain that is no more than a synecdoche
for the traditional locus amoenus; this spring is a metaphor for the origin
of a work of art, for an opposition between the poet's "I" and the "not I,"
the universe he is describing. Whereas that universe exists for the I, starts

between desire and its realization, o r between an act committed and the memory of it";
here and elsewhere, my translation unless otherwise cited. Derrida's commentary, "La
Double Seance," has been rpt. in La Dissimznation, pp. 199-317; see esp. pp. 240 ff.
9. See La Dissimznation, p. 249.
10. This definition has prevailed ever since Dumarsais' treatise, Des Tropes (Paris,
1730). That syllepsis must be distinguished from the so-called grammatical syllepsis o r the
zeugma is apparent in Heinrich Lausberg's Handbuch der literarischen Rhetorik (Munich,
1960), pars. 702-7; o n the acceptation chosen by Derrida, see pars. 7-8.
11. See La Dzssimination, p. 3 10 n.63.
630 Michael Riffaterre Syllepsis

existing when the poet gives it expression, "the I, the exception to and
the condition of everything that appears, does not appear itself."12Source
must therefore be defined as the borderline between water and "no
water." Instead of using this awkward negative compound, Derrida
pretends to find a syllepsis in an existing (perfectly unequivocal) word
which he transforms into an undecidable word: "le mot source: origine
d'un cours d'eau, de-part et point-d'eau, locutions qui sont tout pres de
virer, de facon nullement fortuite, vers les figures de la secheresse, du
negatif et de la ~eparation."'~Phrases about to veer away, literally: on the
verge of turning about or changing color. And that does indeed define,
or rather create, undecidability, and it does indeed produce a kind of
figurative code (in the sense of figurative versus abstract art). The
negativization of the French word source is not to be confused with pe-
joration. Pejoration would yield a dried-up well or a brackish spring-r
perhaps a snake in the fountain. Here we have only a glimpse of truth
through the complementary negative of what has been stated-the
translation of philosophical discourse into the language of echoing con-
notation: into poetic discourse.
Which leads me to a narrower scrutiny of what Derridajoyously says
about his own pun: this pun is by no means fortuitous. This puzzled me
at first because I was doing a linear reading, that is, I was assuming that
Derrida read point d'eau as a lexeme with two reversible or interchange-
able sememes, an obverse "spring" and its reverse, "no water." This
would be a sort of semantic coupling-now you see it, now you don't-
which I knew full well had to be his decision to fool or seduce the reader
by punning on point ("point") and point ("no"). He could not possibly
think that point, within a water context or within a compound word,
could turn into a negative.14 But the reader stops puzzling and can only
marvel with Derrida at the resources of language, for his pun is not
gratuitous. It makes explicit a presupposition o f p o i n t d'eau; it actualizes
the word's natural intertext because point d'eau in its "spring of water"
meaning is actually inseparable from drought and accommodates a der-
ivation leading to "no water." Indeed, the French point d'eau does not
belong to the same descriptive system as the regular word for spring,
source. Whereas source may appear in the mimesis of nearly any setting
and is surrounded by associations like fresh water, coolness, rushes,
12. Derrida, "Qual quelle,"Margesde laphilosophie (Paris, 1972), p. 334; cf. p. 339 n.8.
13. Ibid., p. 333: "the word 'wellspring': origin of a watercourse, point of departure
[Derrida separates root and prefix with a hyphen to emphasize the cutoff] and point d'eau
[in French point d'eau is literally a spot where water wells u p from the ground; but out of
context, if we take point as the old-fashioned, emphatic negative adverb, point d'eau can be
read as "no water," and this authorizes Derrida, though only just barely, to go on as
follows:], these are turns of speech about to veer-and absolutely not by accident-toward
images of drought, of negativity, and of separation."
14. Point in point d'eau is a noun and therefore must be preceded by an article; point
the negative adverb cannot take the article. Hence the completely different syntagmatic
distributions for the two.
 Critical Inquiry Summer 1980 631

reeds, the greenest kind of moss-in short, by positive associations,point

d'eau, by contrast, is surrounded by negative images of the desert. Point
d'eau, in a paradigm of source near-synonyms, conjures up, like oasis (but
without the latter's verdant connotations), arid landscapes, long, dusty,
thirsty treks from one rare water hole to another. You cannot say point
d'eau without generating a converse waterless context. The word con-
forms to the model of negative dialectic not because of any polysemy or
ambiguity but because it happens to presuppose a contrary intertext.
The mediated type.-Let me return now to Mallarme's hymen and
demonstrate how that syllepsis indirectly governs the intertextuality of
one of Derrida's own essays, "Tympan." The text of Derrida's essay is
derived from the title, the intertext from hymen. l 5 Tympan is equated with
hymen, o r rather both are made variants of the same syllepsis and thus is
created the intertextual relationship. The forcing of tympan into a syllep-
tic structure, although its natural structure rules out the syllepsis, is
accomplished by means of a textual interpretant that posits or implies
the equation tympan = hymen. l 6 Derrida's essay concerns the concept of
limitation in philosophy, especially the transgression of limits and, more
especially, the relationship between philosophy and its inseparable coun-
terpart, o r corollary, nonphilosophy o r even antiphilosophy (em-
pirical knowledge, for instance). In French, the (at least primary) mean-
ing of tympan, borrowed from the Greek for "drum," is "eardrum," mid-
dle ear. But more precisely and technically, tympan refers to the taut
membrane that receives sound vibrations and transmits them to the
inner ear. Hence this meaning is the synecdoche of the primary one.
The essay is not couched in abstract language but takes the form of a
long metaphor derived from the title. And in that title the eardrum
stands for boundaries, while sounds are facts and concepts translated
into philosophical discourse. These, in turn, are represented by the per-
ceiving and interpreting nervous system of the ear. The text even plays
on the word types through the Greek: here the word TVVOL, related to
tympan, may mean "form," "image," or "concept" but also the impression
made by a sound. Derrida's aim is to transcend the limits, to deconstruct
the hierarchies and distinctions of self-contained philosophical dis-
course, to make it communicate with other discourses and their intrinsic
features, normally filtered, annexed, and reconstructed by philosophy.
Translated metaphorically this produces: break through the eardrum
and yet somehow maintain the equilibrium between outside and inside
pressure enabling it to transmit sound. We recognize at once the un-
decidability of the hymen image. The final blow, in fact, that bursts the
philosophical eardrum is a pun once more-a readily understandable
pun but the sort universally greeted with a groan. What makes it tolera-
15. See Derrida, Marges, pp. i-xxv.
16. I adapt this textual model from Peirce's interpretant. See Umberto Eco's "Peirce
and Contemporary Semantics," Versus 15 (1976): 49-72; and my "Skmiotique inter-
textuelle: L'Interprktant," Revue d'esthitique 1-2 (1979): 12E-50.
632 Michael Riffaterre Syllepsis

ble nonetheless is a kind of literary accolade bestowed upon it: the reader
sees it as a reference to a textual interpretant of its own, a Mallarme
passage once again. Indeed, the blow is called a coup de doncl' ("the blow
struck by thereforeH)-so we have a pun stacking one expression on top
of the other: first, an auditory metaphor, coup de gong ("gong stroke");
second, the adverb that best summarizes deduction, donc ("therefore")
--donclgong; and third, an oblique allusion to the title of Mallarme's prose
poem, Igztur, Latin for donc, which, once translated, slips easily into the
mold of Mallarme's use, in another title, ofcoup de de's ("roll of the dice").18
The trouble with the eardrum metaphor, however, is what Derrida
himself remarks on: the logic of its representation, its logic as mimesis.
How can any sound break through the eardrum and still be heard?
Logically, philosophy should be deafer. Once the coincidentia oppositorum
of the hymen image is inserted into that of tympan ("eardrum"), the solu-
tion is provided by a mediating interpretant, the parallel text of excerpts
from the poet Michel Leiris' autobiography, Biffures.lg These form a
narrow column running along Derrida's text-a significant margin that
conveys meaning, instead of the white barrier of an ordinary margin
(another oblique allusion to the opposing functions of hymen). In that
text Leiris describes melody, especially the operatic voice, as a steel blade
piercing the listener with pleasure-an obvious phallic image that takes
on the dual aspects of hymen. And better yet, Leiris uses two images for
his dreaming upon the inward convolutions of the ear: a tiny bug, the
earwig, commonly called perce-oreille (literally, "ear-pierce?); and Per-
siphone, the Greek Proserpine, goddess of the dead. Leiris rationalizes
this name as "piercing" and "phone"' ("voice") and speaks of the voice
descending, like the goddess herself, into the Underworld, the "deeper
subterranean reaches of hearing . . . where the caves still echo the faint-
est murmurs."20 Thus the two complementary features of hymen are
transferred to tympan through Leiris' image. If we are to read the "Tym-
pan" text correctly, that is, with a logic not of its metaphor but of what
Derrida's reasoning might be if expressed literally, we must interpret it
intertextually. Tympan, then, presupposes Mallarme's hymen, or rather
Derrida's commentary on it (the intertext), while Leiris' interpretant
gives the reader guidance, as it were, mapping out the path that leads

17. Marges, p. xxv.

18. Coup de dis, then, is like a further interpretant connecting a "stroke ofigztur" and a
"stroke of donc." See Mallarme, Igitur; ou, La Folie d'Elbehnon, Oeuvres compl?tes, pp. 433-43.
Without laboring the point, we can say Igitur'scoup de dts coincides with the gongish stroke
of midnight, and Mallarm6 comments upon it as if it were a symbol of undecidability:
"Minuit sonne-le Minuit oh doivent &re jetes les des. . . . Igitur . . . fait le geste. In-
difference. . . . L'infini sort d u hasard [Midnight strikes-midnight, the moment for rolling
dice. . . . Igitur makes the throw. Indifference. Out of the random flows Infinity]" (p. 434;
see also p. 442).
19. See Jeffrey Mehlman's A Structural Study of Autobiography: Prowt, Leiris, Sartre,
Lhi-Strauss (Ithaca, N.Y., 1974), pp. 149, 113-17, 135-41.
20. Marges, pp. vi-ix.
 Critical Z n q u i ~ Summer 1980 633

him from the essay's title back to the syllepsis. The significance of Der-
rida's essay remains lacunary unless we read it along with Leiris'
marginal text. Similarly, coup de donc, as a pun on coup de gong, continues
to look less than felicitous as long as we do not see that it is "verified,"
"justifiedn-that is, motivated-by its Mallarmean interpretant.
The intratextual type.-A perfect, if complex, example of this type is
Derrida's baffling book G l m ("Death Knell").21Glas is an enormously ex-
panded commentary on Hegel that proclaims itself a monument to
intertextuality, divided as it is into two texts. The left-hand column on
each page is a direct paraphrase of Hegel; the right-hand column is at
once a paraphrase of Jean Genet and a spin-off from the Hegel column.
I am examining Glas also because the book is offered as a fragment: it
begins with the truncated end of a sentence and ends with the unfinished
beginning of a sentence. And yet, paradoxically, intertextuality does
impose upon this fragment the closure or clausula without which it
would be hard to distinguish between text and discourse.
The truncated start of Glas, "quoi du reste aujourd'hui, pour nous,
ici, maintenant, d'un Hegel? . . . Ces mots sont des citations [anyway--or
anyhow-what is Hegel for us here, today? or: What is a Hegel the way
we see things here and now? . . . These are his words I quote]," alludes to
Hegel's criticism of the idea of sense-certainty: Can we acquire absolute
knowledge from sensory experience? Hegel begins with an analysis of
the meaning of the pronoun "this" (German dies, French ~ a )which , we
use to point to what we imagine is reality. Hegel divides it into its seman-
tic components "Here" and "Now" and finds them equally difficult to pin
down:

[Let's] take the "This" in the twofold shape of its being, as "Now"
and as "Here." . . . T o the question: "What is Now?" let us answer,
e.g. "Now is Night." . . . We write down this truth; a truth cannot
lose anything by being written down. . . . If now, this noon, we look
again at the written truth, we shall have to say that it has become
stale.22

This "Here" and "Now" we recognize in Derrida's beginning and again

in his ending of Glas:

Ce que j'avais redoute, naturellement, deja, se reedite. Au-

jourd'hui, ici, maintenant, le debris deZ3[Naturally, what I had

21. Derrida, Glas (Paris, 1974). For commentary on Glas, see Geoffrey Hartman's
"Monsieur Texte: On Jacques Derrida, His Glas," Georgza Review 29 (1975): 759-97; his
"Monsieur Texte 11: Epiphony in Echoland," Georga Review 30 (1976): 170-94; and Jef-
frey Mehlman's Revolutim and Repetztion (Berkeley, 1977), pp. 105-7.
22. G. F. W. Hegel, Phenomenology ofspirit, trans. A. V. Miller (Oxford, 1977), par. 95,
pp. 59-60.
23. Glas, p. 291.
634 Michael Riffaterre Syllepsis

been dreading all along is starting all over again (or, more literally:
"is already appearing again." This minor image is borrowed from
the language of publishing: "is already in its second printing.")
Today, here, now, the fragment left over from]

The book ends the way it started, back at square one of the game. In
between, the critique of sense-certainty runs through the parallel glosses
on Hegel and Genet. Just before he lets the book trail off into a repeat of
the initial sentence, Derrida describes periphrastically what Geoffrey
Hartman calls a "laborious phallic erection as just another-yet an-
other-raising up (re-leve), equivalent to the re-telling-re-publishing
(re-edite) alluded to in the next and last sentence of the HOW-
ever valid the phallic interpretation may be, I think we must first take it
for what it says it is: an erection in the monument or edifice sense-the
construction of an Egyptian obelisk:

C'est tres aride, sur l'esplanade immense, mais Ca ne fait que

commencer, le travail, ici, des maintenant. Des que Ca commence a
ecrire. Ca commence a peine. Ne manque plus qu'une piece.
Ca grince. Roule sur les troncs d'arbre couches. Poulies. Les
cordes graissees se tendent, on n'entend qu'elles, et le souffle des
esclaves plies en deux. Bons a tirer. Fouet cinglant du contremaitre.
Regain de force liee. La chose est oblique. Elle fait angle, deja, avec
le sol. Remord lentement son ombre. sBre de soLZ5

Which is another way of indicating a return to the beginning of the text

since at the outset the matrix Hegel quotation is followed by a discussion
of the philosopher's views on religious buildings in the history of ar-
chitecture. Although the passage is in truth a kind of rebus, its empty
center, the key word, left unspoken, must be "obelisk" rather than
"phallus," for the entire description plainly evokes an Egyptian scene.
First a vast expanse of sand, already landscaped into a monumental
plaza ("C'est tres aride, sur I'esplanade"). Then something sliding slowly
over a bed of huge, rounded logs, dragged forward by slaves, then raised
up toilsomely: What can this object be but a huge stone rolled from a
quarry to the space left in front of the temple, where it is to stand as a
sacred pillar? The reader can have no hesitation: every detail fits into a
stereotyped descriptive system about engineering in ancient Egypt, part

24. Hartman, "Monsieur Texte," p. 791.

25. Glas, p. 291: "The vast expanse of the plaza is utterly arid, but it is only beginning,
the work, here, right now. As soon as it begins to write. It is hardly beginning. Only one
piece [part] is missing.
It is creaking. Rolls along over tree trunks lying side by side. Pulleys. T h e greased
ropes grow taut, all you hear is the noise they make except for the heavy panting of the
slaves with their low-bent backs. Good for nothing but dragging loads. T h e slashing whip
of the foreman. T h e redoubling of harnessed strength. T h e thing is oblique. Already it is
at a n angle to the ground. Slowly now it is gnawing into its own shadow, steadily."
 Critical Inquiry Summer 1980 635

of French schoolroom lore, the answer to the classroom query: How on

earth could they move such heavy masses without the help of cranes or
mechanical power? Of course the reader knows right off that the obelisk
stands metaphorically for some meaningful sign (the text speaks not of
primitive engineering but of "precapitalist writing": "la machine est en-
core trop simple, le mode d'ecriture capitaliste"). Nevertheless, every
realistic detail familiar to schoolchildren is mentioned here, down to the
resonance of the taut ropes vibrating, the slaves panting, the foreman's
lash cracking.
This realism functions first to disambiguate the description beyond
all doubt; but, revealingly, its culmination as an index of reality is also the
point where a new pun undermines the whole archeological tableau so
that the description turns out to be nothing but a verbal booby trap. The
reader is yanked back from the right-hand column on the last page to
the left-hand column on the next to last page-the page facing the last.
By the same token, the pun tells him that the phallus is only a first
metaphorical level; it then yanks him still further back to the book's first
left-hand column three hundred pages upstream. For this pun reveals
that the obelisk is actually a meta-metaphor for the retelling, the return
to the incipit. And all this happens to the one detail that is really just too
much, where Derrida overdoes his archeology: the slaves pictured as
mere beasts of burden, good for nothing but shouldering their loads.
The text says "bons a tirer." This is pathos-the life of the toiling masses
in Pharaoh's land-but it is also a compound technical term of printing
and publishing jargon: bons a tirer is press-proof, the galleys in their last
stage before the book is actually printed: Note also that at the very mo-
ment he finishes his retelling Derrida is also speaking printers' language:
" ,,
re-edite" means to publish a second edition. In its technical usage, bons a
tirer may be seen as facilitating production of the image, as a kind of
Freudian Bahnung through the resistant density of more expectable ver-
bal associations. But as a pun, bons a tirer bridges the gap between the con-
clusion and the page preceding, which contains the word tirage ("second
printingM)-from bons a tirer to tirage. And thus the reader comes ret-
roactively to understand a feature still now nonsensical: in the left-hand
column of the page before, stands quoted, quite unaccountably, a long
business letter from Hegel to his publisher anent a second printing
which, says Hegel, will contain no emendations or additions. In other
words, the last two pages, left and right, repeat the same message in
architectural (or Egyptian) code and in publishing code, both subdialects
of "Hegelian," the language that has now been substituted for French.
The message: "I shall be saying the same thing all over again." That is, of
course, "I am going to read and meditate anew upon the Hegel sentence
I began with."
But in spanning the gulf between left page and right, the pun
rounds off the book in another way too. It links the left-hand column,
the Hegel column, to the right-hand, the Genet column, through a ver-
636 Michael Riffaterre Syllepsis

bal, lexical bridging like that on the first page of Glas, where a variation
on one verb saturated both columns. The two-columned page becomes a
single page once more. The Mallarmkan fold is closed, the pli, Mal-
larmk's basic image for the structure of the book-page piled on top of
page-an image Derrida likes to play with.26Better yet, since here the
left- and right-hand columns, punningly connected, are not on the same
page but on facing pages, the folding up shuts the book. Again, and still
better, take another step: since this fold was first unfolded by the dis-
semination of the book's first sentence into a Hegel and Genet text and
now folds back, the last words at bottom right of the last page, by taking
u p again the first words at top left of the first page, in effect reduce to
naught the whole mass of the book's intervening three hundred pages.
They are all just a gloss on that initial sentence. And with exemplary
circularity the gloss itself has brought us back full circle to its first exam-
ple too, since the erection of the obelisk illustrates the folding over, the
shutting of the book, because the image resumes the first comments on
Hegel, quoting from his Esthetik on India's phallic columns and Egypt's
giant statues. In so doing, the obelisk plays, curiously, the same role as
Jean Cocteau's image of the factory smokestack repeated at start and
finish of his film Le Sang d'un pobte. The first shot in the film is of the
smokestack beginning to crumble; the last shows its total collapse. Which
is to say the whole movie in between, with its long dream sequences and
its plot involvements, all really took only a few seconds of objective time;
the dreamer's subjective imaginary experience has given it duration. So
it is with Glas, the book-length Talmud of a very short Hegelian dictum,
inserted by the reader into the narrow slit between the Here and the
Now into which Hegel divided This. O r else the whole of Glas equals the
"and" conjoining the two instances of This.
It is obvious from the above that the letter to the publisher and the
obelisk tableau are non-sense in a linear reading (in which case, tirer
would be the trope contrary to the syllepsis, namely, the antanaclasis-
repeating the same word but with different meanings). The letter and
the tableau make sense only together, in a comparative reading, and
there would be nothing to dictate such a reading if it were not for the
undecidability of bons a tirer and for ungrammaticalities that corroborate
the effect of the syllepsis on the reader and further insure that he will
read the two texts together and interpret them as two variants of one
invariant. There are such ungrammaticalities. They even provide fur-
ther objective proof that the obelisk does indeed complement (redite) or
re'tdite the letter and is a metaphor for the return to the book's beginning.
First this: that the obelisk, centerpiece of the Egyptian scene, remains
unnamed. This disappearing act would be passing strange if the subject
of the description were really an exotic tableau. In truth, however, the

26. See La Dissimination, pp. 28 1-306.

 Critical Inquiry Summer 1980 637

obelisk is merely the figurative substitute for Hegel's Ca, and thus stamps
with a final image the circularity of Glas, rolls up the scroll of exegesis
upon that pronoun. The text says so: "Ca grince, Ca se penche [This is
creaking, this is about to fall]." But actually the clincher must be a curi-
ous contradiction in the Egyptian scene: "Ca commence a peine. Ne
manque plus qu'une piece [This has hardly started. (At first glance,
"this" must mean the job of erecting the monument.) Only one piece is
missing]." If the work is just starting, how is it only one piece is missing,
unless that piece be both the essential one and a simple one within a
simple mechanism, unless it be our missing pronoun? And lo and be-
hold, the pronoun is missing: "Ne manque plus qu'une piece" has been
shorn of its subject. This subject should have been the neutral il; not il as
pronoun for a person-noun but a mere grammatical abstraction for the
third person, a mere tag for a function and hence an apt symbol for the
universal Dies in Hegel's German, Ca in Derrida's French.
Second corroborating ungrammaticality: the paronomasia on
redite-redoute (Derrida confessing his fear that when the whole Glas has
been gone through, everything will have to be said over again) simply
dramatizes Hegel's own dispirited statement. Having demonstrated that
the universal is attained through the negativizing of particulars, of the
multiple Heres and Nows, Hegel remarks: "natural consciousness, too, is
always reaching this result, learning from experience what is true in it."
But then, almost sadly, he adds: "it is always forgetting it and starting the
movement all over again."27This sense of helplessness is exactly echoed
by Derrida just before he quotes Hegel's letter telling his publisher he is
going to repeat himself; and again when Derrida speaks of advancing
slowly, painfully, along a gallery of images, a gallery of things, of particu-
lars, to reach at last the Sa, S-a, "savoir absolu" ("absolute knowledgeH)-
that is, Ca (this) translated into "universal" code by means of a pun.
Slowly, painfully indeed, since the "gallery" of phenomenological
stopping points is compared with the stations of the cross. Thus
the text is derived a second time from the tirer syllepsis, now through
the translation of the syllepsis into a "depression" or "obsessive fear"
code. The derivation confirms once again that the significance of
the syllepsis is, first, that it functions as a clausula and, second, that it
makes sure the reader correctly interprets that clausula as an analogon
of the incipit.
T o conclude, I shall modify (the better to adapt this trope to the
concept of undecidability) the definition of syllepsis as follows: syllepsis
consists in the understanding of the same word in two different ways at
once, as contextual meaning and as intertextual meaning. The contextual
meaning is that demanded by the word's grammatical collocations, by
the word's reference to other words in the text. The intertextual mean-

27. Hegel, Phenomenology ofSpirit, pars. 10%9, p. 64

638 Michael Rffaterre Syllepsis

ing is another meaning the word may possibly have, one of its dictionary
meanings and/or one actualized within an intertext. In either case, this
intertextual meaning is incompatible with the context and pointless
within the text, but it still operates as a second reference-this one to the
intertext. The second reference serves either as a model for reading
significance into the text (e.g., point d'eau read as "no water" in a text
where it should mean "water," and seems to) or as an index to the
significance straddling two texts (e.g., tirer as "reendeavor," halfway be-
tween a reference to reprinting and a reference to an engineering
feat).2sThus undecidability can exist only within a text; it is resolved by
the interdependence between two texts. And now for a final rephrasing
of my definition: Syllepsis is a word understood i n two dfferent ways at once, as
meaning and as signzjicance. And therefore, because it sums up the duality
of the text's message-its semantic and semiotic faces-syllepsis is the
literary sign par excellence.

28. Bons ci tirer cannot be interpreted as a metaphor since there is no content common
to the subverted and subverting meanings (no ground, in I. A. Richards' terminology; no
shared semes, in semantic parlance).