Confessions, Excuses, and the Storytelling Self:
Rereading Rousseau with Paul de Man

Recall the details of Rousseau’s self-described worst confession, the incident of Marion and the stolen ribbon (pp. 78-85).  He writes that his “conscience is still weighed down” by the event decades later, that its “bitter knowledge, far from fading, becomes more painful with the years,” and goes so far as to suggest that the need to confess it specifically (he claims to have never told anyone before) “has greatly contributed to the decision I have taken to write my confessions” (pp. 82-3).  Having been expelled from Mme. Basile’s household and shop in Turin upon the return of her husband, Rousseau, now sixteen years old, finds new work as a servant for the Comtesse de Vercellis, a childless widow dying herself of breast cancer.  She does soon die, and Rousseau, having been so recently hired, is not on the official list of staff and is thus not to be left anything (the other servants were each left a year’s wages, he tells us).  Mme de Vercellis’s nephew gives him a small sum of money and allows him to keep his uniform, however.  Nonetheless, Rousseau steals a ribbon, which, slight as it might seem, was a significant enough object to be included on the house’s official inventory and so found to be missing.

Rousseau’s theft is discovered, and he stands accused.  He blames Marion, a young girl who works in the kitchens on whom he transparently (to us as readers, but seemingly not the other servants at the time) has a crush.  She is held in high regard, but Rousseau suggests that this actually hurts her case.  She denies, simply and calmly, that she has anything to do with the incident.  Eventually she cries, but she refuses to counterattack Rousseau.  “The contrast between her moderation and my decided tone worked against her,” Rousseau writes, drawing attention as he so often does to the problem of reading the relationship between inner sentiments and outer expression (p. 83).  No further evidence is available, and so no final judgment is reached.  Both are fired, though Rousseau suggests that his innocence was generally believed.

In a coda to his chronological narration of these events, which closes the second book of the Confessions, Rousseau insists that he meant Marion no harm:

I have never been less motivated by malice than at this cruel moment, and when I accused this unfortunate girl, it is bizarre, but it is true, that it was my fondness for her that was the cause of it.  She was on my mind, and I simply used as an excuse the first object that presented itself to me.  I accused her of doing what I wanted to do, and of having given me the ribbon, because my intention had been to give it to her. (p. 84)

Rousseau suggests that this revelation of goodwill, of good intentions, excuses his otherwise hideous-seeming behavior.  He argues that his daring to confess such an incident, which would usually remain unspoken, is itself praiseworthy and that his own harsh judgment of his actions proves that “no one could think that I have in any way sought to mitigate the infamy of my crime” (p. 84).  Yet in the very next breath, Rousseau does just that and insists that “I would not be fulfilling the purpose of this book if I did not at the same time reveal my own innermost feelings, and if I were afraid to excuse myself, even where the truth of the matter calls for it” (p. 84).  We thus have a knot of confessions and excuses, revelations and mitigations, that needs to be unraveled.

Rousseau’s Confessions is a peculiar work.  As Paul de Man writes, “the Confessions are not a primarily confessional text” (p. 279).  Rather, Rousseau tries to evade responsibility for his frequently abhorrent actions, rationalize his misdeeds, and explain his life in any way that would leave him faultless.  Here I work through de Man’s provocative and difficult reading of the incident of the ribbon.  At least implicitly, Rousseau’s autobiography is already essentially modern, abandoning as it does the epistemological structure of the confession for the performative one of the excuse.  This leads de Man to give up on any notion of us as responsible or hermeneutic beings: he argues that signifiers are detached from signifieds, and that we are finally manifestations of language, not users of it.  I reject de Man’s last moves and argue that the meaning of Rousseau’s action is to be found not in a slippage of signifiers, but in an intertextual web of differing and possible narratives.  The conception of the self at work in the Confessions is thus not simply discursive, but specifically one in which we are the products of narrative and its conventions.


In his “Excuses (Confessions),” de Man rereads this event and Rousseau’s telling of it (pp. 278-301).  He distinguishes between the structure of a confession and the structure of an excuse.  “To confess,” he writes, “is to overcome guilt and 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” (p. 297).  A confession brings forward something previously unknown, at least publicly.  Questions of ethical judgment are initially bracketed, because before we can judge, we must know.  De Man’s first formulation here is, it seems to me, more accurate than his second.  Saying that guilt and shame are overcome, better than good and evil superseded, draws attention to the structural reversal that, as we will see, exists between confessions and excuses.  A confession is motivated by the sentiments, by inner feelings of guilt and shame; the confessor finds release and hopefully forgiveness and a restoration of ataraxia by entering a public forum.  Stating that one feels guilty and detailing the previously unspoken or even repressed story that gives rise to that guilt is a movement from the subjective to the objective.  This is obscured if we speak simply of good and evil (potentially objective) and not feelings or self-judgments of good and evil (guilt, shame, pride, and so forth).  In religious confession, one confesses to a priest or to god.  In a non-theistic context, we talk of disburdening oneself of guilt by telling someone, anyone, the truth.  In the barest case, where one confesses something to oneself, it seems that the double structure of self-consciousness is highlighted: we speak of having known “deep down” or in “some part” of ourselves and revealing that finally to our higher faculties.  Structurally, confession relies on finding an objective forum where judgments can rightfully be made, not merely in publicly pronouncing with certainty one’s own subjective judgments of oneself.  These judgments are motivational, not the essential content of a confession.  Thus de Man goes on: “By stating things as they are, the economy of 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 horror” (p. 297).  And further: “Confessions occur in the name of an absolute truth which is said to exist ‘for itself’ and of which particular truths are only derivative and secondary aspects” (p. 297).  Confessions, though they may exist in order to allow correct ethical judgment, are as a rhetorical trope essentially epistemological.  They begin with feelings, sentiments, and self-judgments and end in facts, knowledge, truth.  Once the facts are laid bare, we are equipped again to make judgments, but those judgments come later and are not written into the rhetorical structure of the confession itself.  Indeed we might speak of something like a disingenuous confession even when the content—the fact revealed—is the same.  If one confesses simply in hopes of receiving a beneficial judgment, then one violates the epistemological structure of the confession, perverting it to the end of absolution.  A disingenuous confession then starts to look more like an excuse.

As de Man suggests, it might seem that an excuse has a similar structure to a confession.  Rousseau stresses this when he speaks of the need to “reveal [his] innermost feelings.”  That is, he suggests that his confession is merely continued in the coda by a further unveiling of facts, these ones having to do with his sentiments and intentions.  We should immediately be put on guard, however, by noticing two distinctions.  First, Rousseau’s feelings, if facts, are facts of a different sort than facts about who stole the ribbon, who said this, and who was punished with that.  Second, as Rousseau’s claim that he cannot be “afraid to excuse [himself], even where the truth of the matter calls for it” highlights, whereas the confession is an essentially epistemological trope, the excuse is an essentially moral one.  Let us examine each of these distinctions in turn.

De Man distinguishes between confessions and excuses by highlighting the first distinction, that of the different kinds of facts revealed by each.  “Rousseau can convey his ‘inner feeling’ to us only if we take, as we say, his word for it, whereas the evidence for his theft is, at least in theory, literally available,” he writes (p. 280).  An excuse then becomes explicitly rhetorical, because its weight depends on its speaker’s ability to convince.  Our only access to Rousseau’s intentions and feelings comes by way of his self-report, and so the meaning—especially the moral meaning—of the situation, when structured by an excuse rather than a confession, becomes tied to his particular telling of it beyond physical evidence such as the ribbon itself: “No such possibility of verification exists for the excuse, which is verbal in its utterance, in its effect and in its authority: its purpose is not to state but to convince, itself only an ‘inner’ process to which only words bear witness” (p. 281).  The point is not that we must then tread carefully, vigilant under such limitations to find the real facts of the matter.  Rather, the situation is redrawn entirely, and in more slippery terms: “Whether we believe him or not is not the point; it is the verbal or nonverbal nature of the evidence that makes the difference, not the sincerity of the speaker or the gullibility of the listener” (pp. 280-1).  Throughout the essays of Allegories of Reading, de Man is interested in different kinds of performative, figural rhetoric.  His interest in the Confessions is in how it is essentially structured by just this kind of rhetorical performance (p. 282).

In the second distinction, that between confession as essentially epistemological and excuse as essentially moral, we might go further than de Man and say that excuse directly inverts the structure of confession.  Confession seeks to restore the possibility of objective moral judgment by admitting the facts of the matter.  It thus seeks to equip another to make that judgment and places such responsibility in someone else’s hands.  If Rousseau’s text were simply a confession, it would be left to the reader to judge Rousseau’s moral worth.  But he cannot leave things at that.  His excuse, like all excuses, begins instead with the certainty that he isn’t guilty (despite feeling so) and then seeks to redescribe the situation in a manner that would justify this claim.  So whereas confession, when extended into a moral context, moves from facts to judgment, the excuse inverts this, moving from judgment to “facts.”  In a confession the facts are certain and judgment requires discretion.  An excuse rhetorically presents the judgment as certain and then tries to pass off discretion as necessary to properly understand the facts.  Thus we might say that the essential form of the confession is: “this is what happened (and I feel guilty), therefore please forgive me.”  The essential form of the excuse is: “I’m not guilty (though I feel so), because this happened.”  The ordering of clauses in each case is not accidental, but reveals the respective epistemological or moral motivation of each trope.

If we bring these two strands back together—confessions as verifiable and epistemological versus excuses as performative and morally motivated—we can agree with de Man’s assessment that the Confessions is not primarily confessional.  Rather, it prefigures something more modern.  While Rousseau’s title points explicitly back to Augustine, it wouldn’t be hard to connect the work instead to contemporary memoirs in all their apology, misrepresentation, and naked self-interest.  In moving from confession to excuse we have moved from the realm of empirical facts into the realm of storytelling.  As the first modern autobiography, the Confessions is already (at least implicitly) quintessentially modern, operating as it does with a more slippery notion of truth.  “Excuse occurs within an epistemological twilight zone between knowing and not-knowing,” de Man writes (p. 286).  That is, the realm of excuses is not one where nothing can be known, but where knowledge can no longer be foundationally secured, tracked back in justificatory chains to objective, publicly verifiable facts.  It is thus a realm of necessarily mediated knowledge, where facts are available only through performance and story.  It is no accident that the likes of de Man and Derrida have been drawn to reread the Confessions.  Essentially structured as it is by the trope of the excuse, it can be read as inaugurating an era where judgment is more certain than facts and where facts can be seen to be colored through and through by judgments.  Rousseau spins a self through storytelling rather than revealing pre-given facts about himself.  This is already implicitly true inasmuch as excuses move into a redescription, a retelling of events that would convince the reader of the writer’s worth.  But we need to follow de Man’s account a bit further (and, finally, deviate from it) to substantiate this claim.


“[I]n terms of absolute truth, [excuse] ruins the seriousness of any confessional discourse by making it self-destructive,” de Man writes (p. 280).  A confession is serious, indeed possible at all, only given a stable conception of truth.  Guilt is shed, moral culpability assumed, and leniency hoped for only inasmuch as one admits and reasserts the truth of what happened.  By refusing to let things so stand and leave his moral fate to another’s (the reader’s) judgment, Rousseau and his excuses negate the confession.  The “facts” are redescribed to justify innocence and so the very existence of stable facts is denied.  But this leaves his fate in our hands as readers in a new way.  Rather than being entrusted to examine the facts and level moral judgment, we are instead asked to validate the quality of his redescription, his storytelling.  If the facts are untethered, then we as audience, witnessing fictions and not truths, owe our allegiance not to Rousseau, the man on the scene, but to the best narrator available.  De Man (and any other critic) is then freed by Rousseau’s own abandonment of confessions for excuses to retell Rousseau’s story better than Rousseau tells it himself.  Not freed absolutely, of course.  But it seems that now there can be no final or definitive version of Rousseau’s story.  Certain facts remain public and must be accounted for in any telling or interpretation.  But other claims, of a type introduced by Rousseau in his move to excuses, are infinitely interpretable.  Thus we might say that the public facts are reduced to horizons within which the storytelling must operate, structuring and limiting, but not fully determining it.  Or, in a different metaphor, we might say that the public facts constitute certain necessary pieces in a puzzle that must be used, but that many final configurations (and supplemental pieces) remain possible.

De Man thus offers up a proliferation of readings.  First, he presents the ribbon as symbolic of the economy of forbidden desire present in the repressive household:

It stands for the free circulation of desire between Rousseau and Marion […] [I]t stands for the substitutability of Rousseau for Marion and vice versa.  But since, within the atmosphere of intrigue and suspicion that prevails in the household of the Comtesse de Vercellis, the phantasy of this symmetrical reciprocity is experienced as interdict, its figure, the ribbon, has to be stolen. (p. 283)

Or again, as Rousseau’s penchant for exposure is another of his great revelations in the Confessions (one that Derrida makes rather more of; see 1976, p. 141ff), de Man reads Rousseau’s need to be laid bare in shame as the real story: “shame is primarily exhibitionistic […] What Rousseau really wanted is neither the ribbon nor Marion, but the public scene of exposure which he actually gets” (p. 285).  This “really” has to be read weakly, as asserting reality only within the frame of this one reading or hypothetical among others that de Man has opened.  Rhetorically, it is properly paired with the phrase “But the text offers further possibilities,” which de Man uses on the next page.  That is, de Man, in deconstructionist mode, has to be understood not as revealing the one true meaning of the events or text.  To do so would be to reassert the epistemological priority given up in the move from confessions to excuses.  De Man’s purpose is instead to embrace the mode of the excuse and demonstrate the manner in which meanings remain as much a matter of tellings and audiences as facts and agents.  Leo Damrosch’s attack on de Man’s “really” as “just too simple” (p. 438) is flatfooted then: of course de Man knows this.  But Damrosch’s larger point is exactly right: “What Rousseau does, really, is not to lay his guilt to rest but to make it ambiguous.  In traditional morality, either you have sinned or your haven’t.  In the stories Rousseau tells, the complicating factors are endless” (p. 438).  These further factors are interesting, though, only if they are necessarily complicating, not just complex.  That is, mere complexity, if it can finally be sorted and figured out, is not enough.  This would reduce the problem of understanding Rousseau’s situation to one of hard scholarly work or enough computation.  Only if they are necessarily complicating factors, qualitatively different than complexity in that they resist—even in theory—definitive sorting, is Rousseau’s treatment of his own life radically new in type.  Again, something more modern is birthed here.

But we notice, of course, Damrosch’s own “really.”  Should we not now conclude that “really” is itself not an epistemological but rhetorical use of language?  The reality asserted is not vertical, the hidden meaning uncovered by the critic.  Rather it is horizontal, a further reading established alongside all others, simultaneously feeding off and competing with them.  If de Man would happily assent to such a conception of criticism, Damrosch would not.  In this pair we find extreme cases of bold thinker and careful scholar.  Damrosch’s is a real “really,” uttered without irony.  Fair enough given the general, rather incontestable point he makes here.  But such a conception of criticism leads elsewhere to an unnecessary foreclosure of meaning.  Reviewing the range of psychoanalytic interpretations that have been put forward of the incident of Marion and the ribbon (and its companion piece of the broken comb), Damrosch asserts that “interpretations like these depend on ways of thinking that did not exist in Rousseau’s time” (p. 62).  Thus, Damrosch claims, they can’t be what was really going on or properly contribute to our understanding of the Confessions.  But inasmuch as one accepts Freudianism as a psychological rather than merely a literary or hermeneutic theory (admittedly, probably not much anymore), that is, inasmuch as Freudianism is supposed to explain what’s really going on in human action, of course such explanations could be valid even before their explicit thematization.  But this is to accept Damrosch’s conservative reality principle.  The wider rebuttal to be made then is that our interest is not simply in Rousseau’s life and the Confessions as a relating of it, but in the Confessions as a text.  Rather than simply recording his life, we see Rousseau struggling to find unity and form in its disparate events, and thus struggling to find a conception of his self.1  Damrosch as a biographer is interested in Rousseau’s autobiography primarily as a chronicle of the life of a famous philosopher.  But de Man, most critics, and I here am interested in it as something more: as a literary text, if a strange one given its at least supposed non-fictionality.

We might pause for a moment here and try to make more explicit the often opaque machinations of deconstruction.  I’ve noted that de Man proliferates interpretations of Rousseau’s situation to the end of revealing all meaning to be ultimately slippery.  But what exactly is the status of these interpretations?  Clearly de Man refuses the notion that Rousseau’s story (or texts generally) might have one final and definitive meaning.  But neither does he suggest that all meanings are equivalent and that we can’t make any judgment between them (a charge that has been leveled frequently enough at deconstruction).  Rather, thinkers such as Derrida and de Man, talented readers that they are, seem to take as quite basic that there are good and bad readings.  Or, perhaps, better and worse readings.  One wouldn’t expend such efforts in the attempt to say something if one could simply say anything.  As Jonathan Culler argues, the notion of “literary competence” becomes essential in later French theory (pp. 113-30).  One has to be initiated into the conventions by which literary texts are written and read in order to have any chance of interpreting them artfully, indeed in order to even know that they are the sort of things open to and in need of interpretation.  But better readings then seem to relate to their objects in a complicated manner.  It’s not that one can really say that Rousseau (or any other object of deconstruction) meant any of the things that Derrida or de Man reads into him.  After the death of the author, authorial intentions have been abandoned as the final arbitrator of a text’s meaning.  But neither can we say that any or all of these meanings are present in the text itself.  Even that seems a stretch—new criticism’s close readers didn’t (or don’t) turn out interpretations that look much like deconstructionist ones.  Deconstruction finally treats textual objects as occasions for meaning, sites to be used and misused toward the production of new thinking.  Derrida and de Man essentially reject the academic distinctions between primary and secondary texts, literature and criticism.  Reading becomes a kind of writing, interpreting a kind of original thought.  One might say that they see the world as a mess of meanings, and that resisting any attempt to sort them into stable categories, they instead simply hope to trace an interesting, idiosyncratic, and asymmetrical path through a number of them.  De Man writes elsewhere that “Literature as well as criticism—the difference between them being delusive—is condemned (or privileged) to be forever the most rigorous and, consequently, the most unreliable language in terms of which man names and transforms himself” (p. 19).  Such is typical of what readers find either infuriating or inspiring in deconstruction.

De Man makes a great deal of Rousseau’s claim, quoted earlier, that Marion was only on his mind and that he “simply used as an excuse the first object that presented itself to [him].”  He argues that this should not be understood as part of a normal causal chain leading from Rousseau’s infatuation with Marion to his blaming of her.  Rather, he asks us to understand Marion’s name as detached from any signification and as representative of the gap between signifiers and signifieds:

She is a free signifier. […] In the spirit of the text, one should resist all temptation to give any significance to the sound ‘Marion.’  For it is only if the act that initiated the entire chain, the utterance of the sound ‘Marion,’ is truly without conceivable motive that the total arbitrariness of the action becomes most effective, the most efficaciously performative excuse of all.  The estrangement between subject and utterance is then so radical that it escapes any mode of comprehension. (pp. 288-9)

In this way de Man’s methodological deconstruction—of proliferating interpretations to reveal the uncertainties and absences glossed over in any reading—merges with the substantive program of deconstruction—most centrally the notion that in a post-Saussurian understanding of linguistics, meaning is said to be established not by the vertical relationship between signifiers and signifieds, but the horizontal difference between signifiers and other signifiers.  As standardized by Ferdinand de Saussure in his attempt to make linguistics scientifically rigorous, a sign is the unification of a signifier and signified.  A signifier is a “sound-image,” a spoken sound or a written word that is conjoined to a signified, a concept.  As Saussure argues, the link between the two is arbitrary.  There is nothing in the word “tree” as either spoken sound or written set of letters that has any necessary connection to the concept of a tree (pp. 65-70).  Derrida, de Man, and others radicalize this claim, arguing that if the conjunction between signifier and signified is arbitrary, then any signifier receives its meaning only in its difference from other signifiers, and that in this horizontal slippage, meaning is always ultimately deferred.2  “Marion” is a sound, related to other sounds, or a word related to other words, not a display of interior intention, sentiment, or meaning, not something that represents something else behind it.  The utterance of “Marion” then becomes a site of contested meaning, not a revelation of it.  De Man’s radicalized conclusion is that “Marion” is not merely contingently linked to the girl it names, but semantically and hermeneutically detached, not referring to her meaningfully at all.

De Man ends his argument impressively but dangerously far from its impetus in introducing various metaphors of language and texts as machines.  He refers to the “machinelike quality of the text” and writes:

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 be generated.  There can be no use of language which is not, within a certain perspective thus radically formal, i.e. mechanical, no matter how deeply this aspect may be concealed by aesthetic, formalistic delusions. (p. 294)

De Man argues that such aesthetic and formalistic “delusions” rely on a misleading or outdated metaphor of the body, a metaphor which suggests that texts posses an organic unity: “The text as body, with all its implications of substitutive tropes ultimately always traceable to metaphor, is displaced by the text as machine and, in the process, it suffers the loss of the illusion of meaning” (p. 298).  If the body is organic, whole, and naturally meaningful, a text as machine is instead cold and contingent.  De Man thus seeks to invert our normal understanding of the relationship between ourselves as subjects and language as communication.  It’s not that we exist and use language for our purposes, sending meanings to one another, meanings that arise internally and are translated into transmittable form.  Rather, language is primordial and uses us, de Man suggests:

Far from seeing language as an instrument in the service of psychic energy, the possibility now arises that the entire construction of drives, substitutions, repressions, and representations is the aberrant, metaphorical correlative of the absolute randomness of language, prior to any figuration or meaning.  It is no longer certain that language, as excuse, exists because of a prior guilt but just as possible that since language, as a machine, performs anyway, we have to produce guilt (and all its train of psychic consequences) in order to make excuse meaningful.  Excuses generate the very guilt they exonerate, though always in excess or by default. (p. 299)

This recalls a claim repeated frequently in the later Heidegger, in “… Poetically Man Dwells …” for example: “Man acts as though he were the shaper and master of language, while in fact language remains the master of men” (p. 213).  Clearly, one of the great difficulties of stomaching the sorts of claims that de Man and Derrida make is that it’s not clear how or if many of their metaphors can be naturalized.  Their roots in Heidegger are important in this respect, then, as it seems necessary first to accept Heidegger’s claim, made already in Being and Time but more central in his later writings, that science’s description of the world isn’t final or primordial, but founded on more basic levels of experience that scientific thinking passes over.3  One of my goals here is to domesticate some of de Man’s and Derrida’s seemingly mystifying metaphors, but this involves rejecting their final claims.  De Man’s ultimate conclusion is that “The main point of the reading has been to show that the resulting predicament is linguistic rather than ontological or hermeneutic” (p. 300).  Questions of existence, meaning, and interpretation, it is suggested, are epiphenominal.  What’s really going on is revealed by a proper understanding of signifiers and Derridian différance.  With this “really” de Man opens himself to the objection that his methodological and substantive deconstructions conflict.  That is, if deconstruction’s method is to proliferate interpretations of any text toward the suggestion that all meaning is ungrounded, how does it then ground the positive certainty that everything is really a matter of linguistic difference?


The notion that the meaning of words is generated not through vertical reference to things  signified, but their horizontal difference from other words might seem deeply implausible.  Certainly this is true if one takes nouns as paradigmatic of language, and “medium-sized dry goods” as paradigmatic of things in the world.  We don’t teach a child what “cat” means by comparing it to “bat,” “mat,” and so on.4  Rather, we point at a cat (or a picture of one), and say or write the word.  Such an example substantiates Saussure’s point that the connection between a signifier and signified is initially arbitrary, merely stipulated; it doesn’t, however, lend credence to the radicalized claim that meaning is found in horizontal difference, or that meaning is forever deferred from presence.  This claim becomes more plausible if one takes simple nouns as fringe rather than paradigm cases: cats and bats are among the few things that we can define ostensively.  Consider instead a cluster of terms like “always,” “sometimes,” and “never.”  One can’t point to them.  They can only be learned together and in a richly linguistic context where it is indeed their place among other words that gives them meaning.  But “Marion” then seems a badly chosen example for de Man to use in illustration of his substantive program.  Even more than nouns, proper names are brutely stipulated, pressed into the service of referring to their bearers and nothing else.

De Man’s invocation of differential meaning here might be salvaged (though domesticated) if we take it to be found not in the horizontal difference between words, but stories.  Rousseau claims right at the beginning of the Confessions that he learned the sentiments from his childhood reading, and not just any reading, but of genre romances:

I have no idea what I did before the age of five or six: I do not know how I learned to read; all I remember is what I first read and its effect on me; this is the moment from which I date my first uninterrupted consciousness of myself.  My mother had left some romances.  We began to read them after supper, my father and I.  Our first intention was simply that I should practise my reading with the help of some entertaining books; but we soon became so engrossed in them that we spent whole nights taking it in turn to read to one another without interruption, unable to break off until we had finished the whole volume. (p. 8)

Indeed Rousseau seems to credit his very discovery and consciousness of himself as a self to this reading.  Somewhat later Rousseau describes the effect of his reading on his thinking:

This was to dwell on situations that had appealed to me during my reading, to recall them, to vary and combine them, and to appropriate them in such a way that I became one of the characters I imagined and could envisage myself always in some role that was dear to my heart; I managed, in other words, to place myself in fictional situations that allowed me to forget the unhappiness of my real one. (p. 40)

Rousseau is an inveterate daydreamer, and he constantly imagines alternate lives that he might have lived had circumstances been slightly otherwise: as an engraver if he had been apprenticed to a better master, if he had married young, as a botanist, as a satirist, among many other examples (pp. 42, 142, 176, 188).  It seems to me that, at the end of Book Two—that is, in the reality there recalled—Rousseau understood his situation amidst a web of possible plotlines, most of them quite overdramatic given his past penchant for romances.  Each of these possible stories acquires its meaning from its difference from others, including the one Rousseau finally enacts.  But all this requires that we resist de Man’s final barring of our semantic and hermeneutic selves in the name of linguistics alone.

De Man’s most compelling reading of the event of the ribbon comes before he reduces everything to linguistic concerns.  He writes: “Marion 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 provide him with a stage on which to parade his disgrace or, what amounts to the same thing, to furnish him with a good ending for Book II of his Confessions” (pp. 285-6).  With theorists of the narrative self from Alasdair MacIntyre on, we need not think that Rousseau at sixteen knew he would write a famous autobiography and thus consciously staged such a dramatic event.  Rather we can see the way in which he might in the moment, consciously or unconsciously, have been responsive to narrative concerns and seen himself as existing within a certain kind of story.  He thus might have been unable to act in any other way than he did, because those actions best completed this story.  His sense that he was superior to the status granted him in the household and the fact that this was unrecognized, his unstated love for Marion, his desire for attention and simultaneously for shame: these conditions, present in the household, created a narrative configuration that would best be unified in the event—the story—of the stolen ribbon.  Rousseau shows himself to be thinking in explicitly narrative terms when he writes shortly thereafter that “my love-affairs never have a happy ending” (p. 94).  We can, perhaps, even understand de Man’s notion of a language machine which produces sentiments and events in this way.  If we take a story to be a discourse about events, it is normal to think that the events necessitate a certain presentation—hence the intelligibility of the accusation “you’re telling it wrong” even when no facts are misrepresented.  Peter Brooks argues that, just as often, we can only explain a literary text by suggesting that its manner of telling generates certain events as necessary (p. 28).  Such a reversal takes place in our understanding not just of literary texts, but of our lives.  Events don’t just give rise to understanding; in the reverse direction, understanding ourselves to be part of certain narrative configurations, we create appropriate events.  But this makes sense only if we explicitly understand language as a machine structured not just by linguistics and horizontal relationships between free-floating signifiers, but also by narrative conventions, which seem to be of a higher semantic order and explicitly call on our hermeneutic selves, which de Man finally bars from the proceedings.  The conception of the self at work in the Confessions is not just linguistic, nor discursive, but specifically grounded in storytelling and its conventions.

Accepting that something about de Man’s treatment of detached signifiers is right, the meaning of a text is largely constituted not (just) in terms of its mimetic character, in how it represents the world.  Rather, the meaning is largely constituted in terms of its horizontal relationship to other texts and conventions.  But under the notion of texts here we have to include episodes in our own lives, not just novels, stories, films, and so forth.  So my understanding of a moment in my own life, both in that moment and after the fact, is largely constituted in understanding the place such a moment holds in a network of narrative conventions relating to other similar moments.  And my understanding of a literal text is no doubt in large part constituted in seeing its parallels to stretches of my own life.  Brooks writes:

The source of the codes is in what Barthes calls the deja-lu, the already read (and the already written), in the writer’s and the reader’s experience of other literature, in a whole set of intertextual interlockings.  In other words, structures, functions, sequences, plot, the possibility of following a narrative and making sense of it, belong to the reader’s literary competence, his training as a reader of narrative.  The reader is in this view himself virtually a text, a composite of all that he has read, or heard read, or imagined as written. (p. 19)

And similarly, Paul Ricoeur: “I will say that, for me, the world is the whole set of references opened by every sort of descriptive or poetic text I have read, interpreted, and loved.  To understand these texts is to interpolate among the predicates of our situation all those meanings that from a simple environment, make a world” (p. 80).  I’m suggesting that we take the structure of differential meaning that Derrida and de Man make so much of, but in the place of signifiers, drained of their meaning, we plug in narratives in all of their significance instead.  The resulting configuration is thus not a language as a structure of linguistic difference, but a network of overlapping, similar, contrasting, and contradictory plots.  The meanings of the most basic of these are stipulated, taught to us as children.  Others have no meaning in isolation, integral to themselves, but only held up against others.  I’ll close then, by adapting an image of Rousseau’s:

Have you ever been to the opera in Italy?  While the scene is being changed in the great theatres there, an air of disorder prevails, which is disagreeable and lasts for quite a while: the sets are all muddled together; on every side there is a heaving and a pulling, which it is disturbing to watch; you are afraid it is all going to topple over.  And yet little by little everything finds its place, nothing is missing, and you are astonished to see emerge from all this tumult a delightful spectacle. (p. 111)

Rousseau offers this as an image of his thinking when he sits down to write.  It is just as much, I would claim, an image for his thinking as he acts.  A confusion of sets, stories, characters, and plotlines provides a network of meaning that eventually closes down to one which is enacted.  The self is thus not erased or always deferred from presence but always—and only—interpretable through the narratives in which it finds itself.


Brooks, Peter, Reading for the Plot (Cambridge: Harvard University Press, 1984).

Culler, Jonathan, Structuralist Poetics (Ithaca: Cornell University Press, 1975).

Damrosch, Leo, Jean-Jacques Rousseau: Restless Genius (Boston: Houghton Mifflin, 2005).

de Man, Paul, Allegories of Reading (New Haven: Yale University Press, 1979).

Derrida, Jacques, “Différance,” in Margins of Philosophy, trans. Alan Bass (Chicago: University of Chicago Press, 1982), pp. 1-27.  First published in French in 1972.

_______, Of Grammatology, trans. Gayatri Chakravorty Spivak (Baltimore: Johns Hopkins University Press, 1976).  First published in French in 1967.

de Saussure, Ferdinand, Course in General Linguistics, ed. Charles Bally and Albert Sechehaye, trans. Wade Baskin (New York: McGraw Hill, 1966).  Compiled from student notes from lectures given from 1906-1911.

Heidegger, Martin, Being and Time, trans. John Macquarrie and Edward Robinson (New York: Harper Collins Publishers, 1962).  First published in German in 1927.

_______, “… Poetically Man Dwells …” in Poetry, Language, Thought, trans. Albert Hofstadter (New York: Harper Collins, 1971), p. 209-27.  First published in German in 1954.

_______, “The Question Concerning Technology” in Basic Writings, rev. and expanded ed., ed. David Farrell Krell, trans. William Lovitt (New York: Harper Collins Publishers, 1993), pp. 307-42.  First published in German in 1954.

Norris, Christopher, Deconstruction: Theory and Practice, rev. ed. (London: Routledge, 1991). Original edition published in 1982.

Ricoeur, Paul, Time and Narrative, vol. 1, trans. Kathleen McLaughlin and David Pellauer (Chicago: University of Chicago Press, 1984).  First published in French in 1983.

Rousseau, Jean-Jacques, Confessions, trans. Angela Scholar (Oxford: Oxford University Press, 2000).  First published in French in 1782.

1              Peter Brooks argues (in a book dedicated to de Man) that the very failure of Rousseau to find any final unity in his life forces him to create new texts and narratives, that the only way to unify the narratives he already finds himself in is to craft further ones, perhaps endlessly.  Such is perhaps to suggest that Rousseau is already implicitly postmodern (pp. 32, 53).

2              See especially Derrida’s essay “Différance.”  He writes directly about Saussure in 1976, pp. 27-73.  Another important background source for this view is section 17 of Heidegger’s Being and Time.

3              Heidegger, in the two most basic operative distinctions in Being and Time, distinguishes first between the being of entities as they are available to us as present-at-hand, the way things objectively look and are extended, formalized in empirical science, and as ready-to-hand, the way that tools are available to us in use, and second between both of these as ontic, or pertaining to the being of entities, and ontological investigations of the mode of being of Dasein (human beings’ being).  See especially section 15 for the first, sections 3 and 4 for the second.  He moves away from both distinctions in his later work, focusing instead on what he calls “enframing” (among other concepts) or the way in which scientific and technological thinking limit us to certain ways of conceiving of being, while simultaneously covering over the fact that they are so limiting us (see especially “The Question Concerning Technology”).

4              Christopher Norris, who played an influential role in explaining deconstruction to the English-speaking world, uses the example of “bat” and “cat” (p. 25).  Strictly speaking, there’s nothing wrong with it, but it can lead to the wrong idea.  The problem is that, according to Saussure’s analysis, one can’t use isolated instances like this to draw larger intuitions, as any particular speech act (parole) is dependent on an entire language as a formal system of differential relations (langue).

Re-thinking European Politics and History

IWM Junior Visiting Fellows’ Conferences, Vol. XXXII
© 2012 by the author
Readers may redistribute this article to other individuals for noncommercial use, provided that the text and this note remain intact. This article may not be reprinted or redistributed for commercial use without prior written permission from the author. If you have any questions about permissions, please contact Klaus Nellen at IWM.
Preferred citation: Roth, Ben. 2012. Confessions, Excuses, and the Storytelling Self: Rereading Rousseau with Paul de Man. In: Re-thinking European Politics and History, ed. A. Pasieka, D. Petruccelli, B. Roth, Vienna: IWM Junior Visiting Fellows’ Conferences, Vol. 32.