[Text: Roland Barthes, " Textual Analysis of a Tale By Edgar Poe," from Poe Studies, vol. X, no. 1, June 1977, pp. 1-12.]
Textual Analysis of a Tale
By Edgar Poe
Translated by Donald G. Marshall
University of Iowa
This essay is the third in our current series of translations sampling contemporary European responses to Poe (see Poe Studies, 9 , 1-6, 33-39). "Analyse textuelle d'un conte d'Edgar Poe" originally appeared in Semiotique narrative et textsuelle, ed. Claude Charbrol (Paris: Librairie Larousse, 1973), pp. 29-54. The translation is published by permission of the author and Librairie Larousse.
I think Poe would have thoroughly enjoyed Roland Barthes' essay. Its elaborate and playful rationalism is quite in the vein of the "Philosophy of Composition." Barthes has more formal schooling (in classics and theater) than Poe had, but is also something of a polymath and autodidact, well-read in linguistics, semiotics, mass culture and psychoanalysis. His structural Freudian study of Racine precipitated a heated controversy, and he has written on authors as diverse as Robbe-Grillet, Brecht, Loyola, Fourier, Sade, and Michelet. Now firmly entrenched at the Ecole Pratique des Hautes Etudes, he remains essentially a man of letters, rather than an academic. With immense curiosity and immense style, he combines an admirable capacity for and ability to communicate a sympathetic pleasure in texts. Any French critic worth his salt writes on Poe sooner or later (though Barthes says his conscious motive in choosing "Valdemar" was merely pedagogic). But anyone with a sense of Poe's own diverse interests and lively intelligence may suspect a deeper affinity (and Barthes admits his unconscious may have done the choosing).
For one thing, Barthes' erudition functions like Poe's "science." This essay deploys -- neither superficially nor incorrectly -- terms and notions from the linguists Roman Jakobson and Emile Benveniste, from philosophers as different as Jacques Derrida and J. L. Austin, and from a range of disciplines including psychoanalysis, medicine mathematics, and classical rhetoric. As with Poe, the erudition is aimed at an effect: in Barthes' case, pleasure. This idea has figured in his critical chinking from the beginning and takes stage center in The Pleasure of the Text (1973) . We should remember, I think, that Barthes' work aims at pure pleasure, at playing the game. We might think [column 2:] of the recreation in ingeniously constructed detective stories or of those vastly intricate conventions of description which help the love-struck poet concentrate intently on his (or her) beloved's minutes" particularity. Perhaps Barthes was inspired by classical philology -- this very typography may recall the commentaries and apparatus criticus of textual editing, though these blossom here under the warm double sun of semiotics and psychoanalysis. Certainly, the true philologist's devotion to detail is directly proportionate to his passion for the text. The reader in a hurry may find this frustrating, but to the text's true lover it is meat and drink.
Poe was similarly fascinated with detail: the "Marginalia" on punctuation were not written merely to fill space. Such fascination is, in fact, the mark of the artist. No one ever wrote a poem or novel vaguely. Poe may be the earliest of American authors who confounds our distinction between critical intelligence and creative imagination. He insisted his writings were not produced in an unconscious trance of inspiration, but resulted if anything from a hypertrophy of self-consciousness. Barthes similarly argues that the distinction between writer and critic, between writer and reader, is artificial. The critic reads a writing, but he writes his reading. It is notable that Barthes transcribes Poe's text absolutely literally, line by line, without any of the selection or rearrangement by which a literary critic customarily makes a text fit his interpretation, subordinating it to the coherence of his own thought. Barthes' commentary is interlinear: he writes his reading "between the lines" (the "lexias"). We may object that criticism is of course a parasitic discipline, that the writer's text is already there, a product of original creation. But in the "Longfellow Wars" that followed his charge of plagiarism, Poe himself concluded that the liability to borrowing was in direct ratio to the poetic sentiment: "for the most frequent and palpable plagiarisms, we must search the works of the most eminent poets." Barthes locates the borrowing in "codes." These are structured collections of citations which seem, "intuitively" as we say, to make sense. We could pick out the sentences of a story which carry the plot ("actional" code); those which rouse or satisfy the reader's interest (code of the "Engima"); sentences, phrases, or words which resonate with symbolic overtones; those which draw on particular knowledge ("cultural code"); or those which shape the implicit relation of narrator to reader ("code of communication"). A text "weaves together" all these codes and others, but does not invent them. Similarly, we unconsciously use the grammar of English in speaking; we didn't invent that grammar and cannot consciously change it, though paradoxically, with myriad speakers over long periods of time, [page 2:] it changes. So with the codes of mass culture, of fashion, and of literature. Barthes' criticism aims to catch the fleeting moment of reading, when the actual text meshes with the already-constituted codes according to which we read it and make sense out of it.
Barthes would readily admit that his own writings are also constituted by codes; he does not intend to disguise the fact, but to revel in it. If he does not explicitly name other writers he borrows from or alludes to, it is because the game is played, he assumes, before an audience of cognoscenti. He expects them to recognize the references and appreciate his subtle variations on them. Consider the essay's final observation: writing occurs when we can no longer say exactly who speaks, but only that "there begins to speak." In French, the phrase is ca commence a parler. "Ca" translates Freud's "id." Implicitly, it is the unconscious which speaks in writing. But the codes too are unconscious -- are the unconscious. (In translating, I tried to suggest "Here beginneth . . . ," the formula that introduces liturgical readings.) Thus a version of what we recognize as the "intentional fallacy" becomes for Barthes (as for recent French criticism generally) not just a negative principle ("do not inquire into the author's mind"), but a positive theory of the nature and origin of the text.
I would not, however, want to leave the impression that Barthes is playing a hermetic game based on the peculiar pleasure of manipulating esoteric knowledge. He neither attempts to lay down the "definitive" interpretation nor to impose himself as an "authority" whose comments must be dutifully registered in all future Poe criticism. He immensely enjoys his reading and finds that dissecting it minutely increases the pleasure. But he is also generous with his own reader. He does not want to tell you what to think, but rather set your thinking in motion; not put words in your mouth, but be midwife to your own speaking. Barthes' eminence among recent French critics is due in large part to the fact that however much or little one may know about the specialized fields he delights in, his writings are immensely entertaining and thought-provoking. The measure of his success is the productivity of his writing, its desire to nourish the continuous expansion and supplement of critical reading and writing, rather than the fixed reproduction of the "correct" reading.
A final word about the translation. I have tried to be accurate, but have sometimes yielded to the temptation to express an idea, as the phrase goes, in the way I think Barthes might have said it had he been writing in English. I have twice added a quotation from Poe's text where Barthes' words were dose to Baudelaire's translation, but the allusion might be lost in the English. "Significance" is throughout translated "signifying." I aimed at the active, verbal sense of the English present participle, and assumed Barthes did not use "signifiant" because that word had already become a technical term in linguistics (normally translated "signifier"). I translate "science" as "systematic knowledge" or "systematic study," because the overtones of the English "science" seem to me quite different. The French reflexive verb is somewhere between the English active and passive. When, for example, Barthes says the text "constructs itself," he should be understood in this "middle" voice.
D. G. M. [column 2:]
Structural analysis of narrative is currency in the process of being fully worked out. All the studies have a single scientific origin. semiology or the systematic study of significations. But already (and this is good) they show divergences from each other, according to the critical view each takes of the scientific status of semiology, that is, of its own discourse. These (constructive) divergences can be grouped under two main trends. According to the first, analysis, confronting all the narratives in the world, tries to establish a narrative model (formal, of course), a structure or a grammar of Narrative, beginning from which (once it is found) each particular narrative will be analyzed in terms of deviations. According to the second trend, the narrative is immediately subsumed (at least when it lends itself to being so) beneath the notion of "Text," space, process of significations at work, in a word, signifying (I will return to this word at the end), which one observes not as a finished, closed product, but as a production in process of making itself, "plugged into" ocher texts, other codes (this is intertextuality), thereby articulated with society, History, not along paths fixed in meaning, but citational. It is necessary, then, in a certain manner, to distinguish structural analysis from textual analysis, though I do not here wish to declare them antagonistic. Structural analysis properly so called is especially applicable to oral narrative (to myth). Textual analysis, which I will try to practice in the following pages, is applicable exclusively to written narrative."
Textual analysis does not try to describe the structure of a work. It does not aim to record a structure, but rather to produce a moving structuration of the text (a structuration which displaces itself from reader to reader throughout the length of History). It aims to remain within the signifying volume of the work, within its signifying. Textual analysis does not seek to know what determines a text (what brings it together as the final term of a causality), but rather how it breaks out and disperses itself. We shall therefore take a narrative text, and we shall read it, as slowly as necessary, stopping as often as necessary (ease is a capital dimension of our labor), attempting to mark and to class without rigor not all the senses of the text (that would be impossible, for the text is open to infinity: no reader, no subject, no systematic study can stop the text) but the forms, the codes, according to which the senses are possible. We shall mark the approaches of sense. Our aim is not to find the sense, nor even a sense of the text, and our labor is no kin to literary criticism of a hermeneutic type (which seeks to interpret the text according to the truth it thinks is kept concealed in it), such as for example Marxist criticism or psychoanalytic criticism. Our goal is to come to conceive, imagine, live the plural of the text, the opening of its signifying. What is at stake in this labor is not limited then, I think, to the academic treatment of the text (even were it overtly methodological), nor even to literature in general. It concerns a theory, a practice, a choice which find themselves taken up in the struggle between men and between signs.
In order to proceed to the textual analysis of a narrative, we shall follow a certain number of working procedures (let us speak of elementary rules of manipulation, [page 3:] rather than methodological principles; the latter word would be too ambitious and in particular ideologically debatable, to the extent that "method" too often postulates a positivist result). We shall reduce these procedures to four seeps set out in summary fashion, preferring to let the theory run along within the analysis of the text itself. We shall say for the moment just what is necessary to begin as quickly as possible the analysis of the tale which we have chosen.
1. We shall cut up the text which I propose we study into contiguous and in general very short segments (a sentence, a portion of a sentence, at most a group of three or four sentences). We shall number these fragments beginning with 1 (for a dozen pages, there will be 150 segments). These segues are unities of reading, hence I have suggested calling them lexias (2). A lexia is evidently a textual signifier. But since our aim here is not to observe signifiers (our task is not stylistic), but sense, the cutting up need not be theoretically founded. Since it lies within discourse and not within language, we need not expect an easily perceived homology between signifier and signified. We do not know how one corresponds to the ocher, and consequently we ought to be willing to cut up the signifier without being guided by the underlying segmentation of the signified. In sum, the parcelling out of the narrative text into lexias is purely empirical, dictated by convenience: the lexia is an arbitrary product, it is simply a segment in the interior of which one observes the division of senses. It is what surgeons would call an operating field: the useful lexia is one through which there passes only one, two, or three senses (superposed within the volume of the piece of text).
2. For each lexia, we shall observe the senses which it sets in motion. By sense we evidently do not mean the sense of the words or groups of words such as the dictionary or grammar, in short acquaintance with the French language, would sufficiently account for. We mean the connotations of the lexia, the secondary senses. These senses of connotation can be associations. For instance, the physical description of a character, screeching over several sentences, may have only one connotative signified, namely, the "nervousness" of the character, although this word does not appear at the level of denotation. Connotations can also be relations that result from connecting two places in the text, sometimes ones very far apart (an action begun here can be completed, finished below, very much further on). Our lexias will be, so to say, sieves as fine as possible, thanks to which we shall "skim off" the senses, the connotations.
3. Our analysis will be progressive: we shall go through the length of the text seep by step, at least in theory, since for reasons of space we shall be able to give here only two fragments of analysis. That means we shall not seek to disengage the large (rhetorical) masses of the text. We shall not construct a map of the text, and we shall not seek its thematic. In a word, we shall not do an explication of the text, unless we give the word "explication" its etymological sense, to the extent that we unfold the text, leaf through its layers. We shall leave to our analysis even the pacing of the reading: simply, this reading will be, as it were, filmed in slow motion. This way of proceeding is [column 2:] important for theory: it signifies that we do not aim to reconstitute the structure of the text, but to follow its stucturation, and that we shall consider the structuration of the reading as more important than that of the composition (the latter being a rhetorical and classical notion).
4. Finally, we shall not worry ourselves out of measure if in our survey we "forget" some senses. The forgetting of senses forms in some sort part of the reading: what matters to us is to show the departures of senses, not the arrivals (at bottom, is sense anything other than a departing?). What founds the text is not an internal, closed, accountable structure, but the issuing of the text onto other texts, other codes, other signs. What makes the text is the intertextual. We begin to glimpse (by means of other systematic studies) that research ought little by little to become accustomed to the conjunction of two ideas which for a very long time have passed fat contradictory: the idea of structure and the idea of combinatory infinity. The reconciliation of these two postulations imposes itself on us now because language, with which we are beginning to be better acquainted, is at once infinite and structured.
These remarks are enough, I believe, to begin the analysis of the text (it is always necessary to cede to the impatience of the text, never to forget, whatever may be the imperatives of the study, that the pleasure of the text is our law). The text which has been chosen is a short narrative by Edgar Poe, in the translation by Baudelaire: "The Facts in the Case of M. Valdemar."3 My choice -- consciously, at least, for it is perhaps in fact my unconscious which has chosen -- has been governed by two pedagogic considerations: I needed a very short text to be able to master entirely the signifying surface (the succession of lexias); and one symbolically very dense, so that the analyzed text would concern us continually, beyond any idiosynctasy: who would not be concerned by a text of which death is the explicit "subject"?
I ought to add, frankly, this: analyzing the signifying of a text, we voluntarily abstain from treating certain problems. One will not speak of the author, Edgar Poe, nor of the literary history of which he forms part. One will take no account of the fact that the labor will be carried out on a translation. We shall take the text such as it is, such as we read it, without worrying ourselves to know if within a university department it belongs to English rather than to French or philosophy. That does not necessarily mean that these problems will not come up in our analysis. On the contrary, they will come up in the proper sense of the term: analysis is a crossing over the text; these problems can be pointed out under the heading of cultural citations that set out from a code, not from fixed meanings.
One last word, by way perhaps of conjuration, of exorcism: the text which we shall analyze is neither lyric nor political, it speaks neither of love nor of society, it speaks of death. That is to say that we must remove one particular censure: that which is attached to the sinister. We shall do so by persuading ourselves that every censure counts equally for others: to speak of death outside all religion is to lift at once the religious and the rationalist interdiction. [page 4:]
Analysis of lexias 1 to 17
(1) The Facts in the Case of M. Valdemar
(2) Of course I shall not pretend to consider it any matter for wonder, that the extraordinary case of M. Valdemar has excited discussion. It would have been a miracle had it not -- especially under the circumstances. (3) Through the desire of all parties concerned, to keep the affair from the public, at least for the present, or until we had farther opportunities for investigation -- through our endeavors to effect this -- (4) a garbled or exaggerated account made its way into society, and became the source of many unpleasant misrepresentations, and, very naturally, of a great deal of disbelief.
(5) It is now rendered necessary that I give the facts -- as far as I comprehend them myself. (6) They are, Succinctly, these:
(7) My attention, for the last three years, had been repeatedly drawn to the subject of Mesmerism; (8) and, about nine months ago, it occurred to me, quite suddenly, that in the series of experiments made hitherto, (9) there have been a very remarkable and most unaccountable omission: (10) -- no person had as yet been mesmerized in articulo mortis. (11) It remained to be seen, (12) first, whether, in such condition, there existed in the patient any susceptibility to the magnetic influence; (13) secondly, whether, if any existed, it was impaired or increased by the condition; (14) thirdly, to what extent, or for how long a period, t/se encroachments of Death might be arrested by the process. (15) There were other points to be ascertained, (16) but these most excited my curiosity (17) -- the last in especial, from the immensely important character of its consequences.
 "The Truth in the Case of M. Valdemar" [Baudelaire's tide; tr.].
The function of titles has not been thoroughly studied, at least from a structural point of view. What one can say immediately is that society, for commercial motives, needing to turn the text into a commodity, a piece of merchandise, has need of operators that mark the brand. The title has as function to mark or brand the beginning of the text that is to constitute the text as merchandise. Every title therefore has many senses at once, among them at least two: 1) that which it states, tied to the contingency of what follows it; 2) the very announcement that a piece of literature will follow (that is to say, in fact, a piece of merchandise). Otherwise put, the title always has a double function: enunciative and deiceic (4).
a) To announce a truth is to stipulate that there is an enigma. The posing of the enigma results (on the level of signifiers): from the word "truth" [Poe: "faces"; tr.]; from the word "case" (that which is exceptional, hence masked, hence signifying, and consequently that the sense of which muse be found); from the definite article the (there is only one truth, hence the whole labor of the text is necessary to enter in at this strait gate); from the cataphoric (5) form implied by the tide: what follows will actualize what is announced, the resolution of the enigma is already announced; note that the English says: "The Facts in the Case . . .": the signified envisioned by Poe is of the empirical order, that envisioned by the translator into [column 2:] French (Baudelaire) is hermeneutic: the truth refers therefore to the exact facts, but also perhaps to their sense. However that may be, one will code this first sense of the lexia: Enigma, posing (the enigma is the general name of a code, the posing is only one of its terms).
b) One could speak the truth without announcing it, without making reference to the word. If one speaks about what one will say, if one doubles language into two layers so that the first caps the second in some way, one does nothing other than have recourse to a meta-language. Hence the meta-linguistic code is present here.
c) This meta-linguistic announcing functions like an "appetizer": the point is to give the reader an appetite (a procedure akin to "suspense"). The narrative is a commodity, which is brought forward after some "puffery." This "puffery," this "appetizer," is one term from the narrative code (rhetoric of narration).
d) A proper name ought always to be questioned carefully, for the proper name is, so to say, the prince of signifiers. Its social and symbolic connotations are rich. One can read into the name Valdemar at least the two following connotations: 1) presence of a suxio-ethnic code: is it a German name? Slavic? In any case, not Anglo-Saxon. This little enigma, which is here implicitly formulated, will be resolved at no. 19 (Valdemar is Polish); 2) "Valdemar" is "the valley of the sea"; oceanic abyss, marine depth is a theme dear to Poe: the gulf refers to what is doubly outside nature, beneath the waters and beneath the earth. Hence, from the point of view of analysis two codes leave their trace here: a socio-ethnic code and a (or the) symbolic code (I will return to these codes a little later).
e) To say "M(onsiesur) Valdemar" is not the same thing as to say "Valdemar." In many tales, Poe uses simply first names (Ligeia, Eleonora, Morella). The presence of this Monsieur carries an effect of social reality, of the historically real: the hero is socialized, made part of a definite society, in which he is provided with a civil title. We must therefore write: social code.
 "Of course I shall not pretend to consider it any matter for wonder, that the extraordinary case of M. Valdemar has excited discussion. It would have been a miracle had it not -- especially under the circumstances."
a) This sentence (and chose which follow immediately) evidently function to stimulate the reader's expectation, and it is for that reason that they are apparently insignificant: what one wants is the solution of the enigma posed by the title (the "truth"). But one delays even the statement of this enigma. This must then be coded: delay in posing the enigma.
b) Same connotation as in (1) c: the point is to excite the reader's appetite (Narrative code).
c) The word "extraordinary" is ambiguous: it refers to what departs from the norm, but not necessarily from nature (if the case remains "medical"). Bur it can also refer to what is supernatural, has passed over into transgression (it is the "fantastic" in the tales precisely "extraordinary" ones which Poe recounts). The ambiguity of the word here signifies: we are concerned with a horrible tale (beyond the limits of nature) and nevertheless one covered by the scientific alibi (connoted here by "discussion," which is a word used by learned persons). This [page 5:] alloy is in fact cultural: the mixing of the strange and the scientific had its apogee in this part of the nineteenth century to which Poe roughly belongs. One was stimulated to observe the supernatural scientifically (hypnotism, spiritism, telepathy, etc.). The supernatural took a scientific, rationalist alibi. Such is the cri du coeur of this positivist age: if only one could believe scientifically in immortality! This cultural code, which, in order to simplify, I have called scientific code, will have a major importance in the whole narrative.
 "Through the desire of all parties concerned, to keep the affair from the public, at least for the present, or until we had farther opportunities for investigation -- through our endeavors to effect this -- [. . .]"
a) Same scientific code, resumed by the word "investigation" (which is also a word pertaining to a policeman or detective: one knows the fortunes of the detective novel in the second half of the nineteenth century, starting, precisely, from Poe; what matters, ideologically and structurally, is the conjunction of the code of the detection enigma and of the code of science -- of scientific discourse -- which proves that structural analysis can work together very well with ideological analysis).
b) The motives of the secret are not stated. They can proceed from two different codes, present together in the reading (to read is also, silently, to imagine what has been left silent): 1) the scientific-deontological code: the doctors and Poe, out of loyalty, prudence, do not wish to make public a phenomenon which has not been scientifically elucidated; 2) the symbolic code: there is a taboo on living Death: one keeps silent because it is horrible. It must be said immediately (although the point will be stressed in what follows) that these two codes are undecidable (one cannot choose one over the other), and that it is this very undecidability which makes a good narrative.
c) From the point of view of narrative actions (this is the first we shall meet), a sequence is here enticingly initiated: "to keep hidden" implies in effect, logically (or pseudo-logically), consequent operations (for example: uncovering). Here must then be posed the first term of an actional sequence: to keep hidden, whose sequel we shall find later.
 "C. . .] a garbled or exaggerated account made its way into society, and became the source of many unpleasant misrepresentations, and, very naturally, of a great deal of disbelief."
a) The demand for truth, that is to say the enigma, has already been posed twice (by the word "truth" and by the expression "extraordinary case"). The enigma is here posed a third time (to pose an enigma means in structural terms: to state: there is an enigma), by adducing the error to which it has given rise: the error, posed here, retroactively justifies the tide ("The truth about . . .") by anaphora. The redundance brought about by the posing of the enigma (one repeats in many ways that there is an enigma) has its value as an "appetizer": the point is to excite the reader, to procure customers for the narrative.
b) In the actional sequence "To hide," a second term appears: the effect of the secret, distortion, false opinion, the charge of mystification. [column 2:]
 "It is now rendered necessary that I give facts -- as far as I comprehend them myself."
a) The emphasis placed on "the facts" presumes the entanglement of two codes, between which, as in b, it is impossible to decide: 1) the law, scientific deontology subjects the scientist to servitude, the observer to the fact; the opposition of fact to rumor is an old mythic theme. Invoked in a fiction (and invoked in emphatic fashion by a word in italics), the fact has the structural function (for the real import of this artifice dupes no one) of authenticating the tale, not to make believe that it really happened, but to keep the discourse to reality, and not to fabling. The fact is thus grasped in a paradigm in which it is opposed to mystification (Poe recognized in a private letter that the tale of M. Valdemar was a pure mystification: it is a mere hoax). The code which structures the reference to fact is then the scientific code, with which we are already acquainted; 2) however, any more or less high-sounding recourse to Fact can also be considered as the symptom of a conflict between the subject and the symbolic. To clamor aggressively in favor of "Fact pure and simple," to clamor for the triumph of the referent, is to bring into doubt signification, is to mutilate the signifier which displaces fact, is to refuse the other scene, that of the unconscious. By repressing the symbolic supplement, the narrator (even if in our eyes he does so by a narrative feint) takes on an imaginary role, that of the scientist. The lexia's signified is then the asymbolism of the subject of the enunciation: I presents itself as asymbolic; the denial of the symbolic evidently forms part of the symbolic code itself.
b) The actional sequence "To hide" unfolds: the third term states the necessity to rectify the distortion marked in (4)b. The rectification counts as: to wish to disclose (what was hidden) . This narrative sequence "To hide' evidently constitutes a stimulus to narrative. In one sense, it justifies it, and by the same token alludes to its value (its counting-for), in fact a commodity: I relate, says the narrator, in exchange for an exaction of counter-error, of truth (we are in a civilization where truth is a value, that is to say a commodity). It is always very interesting to try to disengage a narrative's counting-for: in exchange for what does one relate? In the Thousand and One Nights, each tale counts for a day of survival. Here, we are apprised that the tale of M. Valdemar counts for truth (presented first as a counter-distortion).
c) The I appears explicitly for the first time -- it was already present in the we of "our endeavors" (3) 6 The enunciation carries in fact three l's, that is to say three imaginary roles (to say I is to enter into the imaginary):]) an I who is a narrator, an artist, motivated by calculations of effect; to this I corresponds a You which is that of the literary reader, he who reads "a fantastic story by the great writer Edgar Allan Poe"; 2) a witness 1, who has the capacity to witness to a scientific experiment; the corresponding You is that of a jury of scientists, of serious opinion, of the scientific reader; 3) an I who acts, an experimenter, he who will hypnotize Valdemar; the You is then Valdemar himself. In the two latter cases, the motive of the imaginary role is "truth." We have here the three terms of a code we shall call, possibly provisionally, the code of communication. Without doubt between these [page 6:] three roles there is another language, that of the unconscious, which expresses itself neither in science nor in literature. But that language, which is literally the language of inter-diction, does not say 1: our grammar, with its three persons, is never directly that of the unconscious.
(6) "They are, succinctly, these:"
a) To announce what follows returns to meta-language (and to the rhetorical code); it is the border which marks the debut of a tale within the tale.
b) "Succinctly" conveys three mixed and undecidable connotations: 1) "Have no fear, this will not be too long": in the narrative code, this is the mode of the phatic (marked out by Jakobson) (7), whose function is to hold attention, to maintain contact, 2) "This will be brief because I will stick strictly to the facts"; it is the scientific code which permits the statement of the scientist's "denudation ' of the superiority of the claim of fact over the claim of discourse; 3) to plume oneself on speaking briefly is in a certain fashion to lay a claim against speaking, to limit the supplement of discourse, that is to say the symbolic; it is to speak the code of the asymbolic.
 "My attention, for the last three years, had been repeatedly drawn to the subject of Mesmerism;"
a) In any narrative, it is necessary to keep an eye on the chronological code. Here, in this code ("the last three years"), two values are mingled. The first is in some sort naive. One notes one of the temporal elements of the experiment which is going to be conducted: the time of its preparation. The second has no diegetic, operating function (this is evident from the proof by commutation: if the narrator had said seven years instead of three that would have had no impact on the ,tale). The point then is to give a pure effect of reality: the number emphatically connotes the truth of the fact: what is precise is adjudged real (an illusion, moreover, since there exists, well recognized, a delirium of numbers). We note that linguistically the word "last" is a "shifter" (8), a coupling-gear: it refers to the situation of the speaker in time; it therefore reinforces the presentness of the testimony which will follow.
b) Here begins a long actional sequence, or at very least a sequence well furnished with terms. Its object is to get an experiment under way (we are under the alibi of experimental science). This getting under way, structurally, is not the experiment itself. It is an experimental program. This sequence counts in fact as the formulation of the enigma, which has already been posed repeatedly ("there is an enigma"), but which has not yet been formulated. In order not to overburden the report of our analysis, we shall code the "Program" separately, it being understood that the whole sequence, by proxy, counts as one term in the code of the Enigma. Within this sequence "Program," we have here the first term: posing of the scientific field of the experiment, mesmerism.
c) The reference to mesmerism is drawn from a cultural code, one very prominent in this part of the nineteenth century. Following Mesmer (in English, "hypnotism" can be called "mesmerism") and the Marquis Armand de Puysegur, who had discovered that hypnotism could induce sleep-walking, hypnotists and societies for hypnotism multiplied in France (around 1820). In 1829, one had been able, it seems, to remove a tumor painlessly [column 2:] under hypnosis. In 1845, the year of our story, Braid of Manchester systematized hypnosis in ,the form of inducing a nervous fatigue by the contemplation of a shiny object. In 1850, at the Mesmeric Hospital of Calcutta, painless deliveries were achieved. Subsequently Charcot classed the hypnotic states and delimited hypnotism from hysteria (1882), but since then hysteria as a clinical entity has disappeared from hospitals (starting from the moment when one stopped observing it). 1845 marks the summit of the scientific illusion: it was believed that hypnosis had a physiological reality (moreover, Poe, pointing to the "nervousness" of Valdemar, can let the subject's hysteric predisposition be understood).
d) Thematically, hypnotism connotes (at least at that ePoeh) an idea of fluid: something passes from one subject to the other. There is a "said-between" (an interdiction) between the narrator and Valdemar: this is the code of communication.
 "and, about nine months ago, it occurred to me, quite suddenly, that in the series of experiments made hitherto, [. ..]"
a) The chronological code ("nine months") falls under the same remarks as those made in  a.
b) Here is the second term of the sequence "Program": a domain has been chosen in b, hypnotism; it is now divided up: a particular problem will be isolated.
 "[. . .] there had been a very remarkable and most unaccountable omission :"
a) The structure of the "Program" continues to be stated: here is the third term: the experiment which has not already been done -- and hence, for any scientist concerned with research, is to be done.
b) This lack of experiment is not a simple "forgetting;' or at least this forgetting is extremely significant: it is quite simply the forgetting of Death. There had been a taboo (which will be lifted, in the most profound horror). The connotation belongs to the symbolic code.
 " -- no person had as yet been mesmerized in articulo mortis."
a) Fourth term of the sequence "Program": the content of the lacuna (evidently this sets up in advance the relationship between the assertion of the lacuna and its definition, within the rhetorical code: to announce/to state precisely) .
b) Latin (in articulo mortis), the language of law and medicine, produces an effect of scientificity (scientific code), but also, by the intermediary of a euphemism (to say in a little known language something one dares not say in everyday language), designates a taboo (symbolic code). It seems that in Death what is essentially taboo is the passage, the threshold, the "dying." Life and death are states classed relatively. Moreover, they come into paradigmatic opposition, responsibility for them is taken by sense, which is always soothing. But the transition between the two states, or more exactly, as will be the case here, their encroachment, undoes sense, engenders horror: there is a transgression of an antithesis, of a classification.
 "It remained to be seen,"
Details of the "Program" are announced (rhetorical code and actional
sequence "Program"). [page 7:]
 "first, whether, in such condition, there existed in the patient any susceptibility to the magnetic influence;"
a) In the sequence "Program;' this is the first coin minted from she announcement made in  : what is in question is a first problem to elucidate.
b) This first problem itself gives a title to an organized sequence (or sub-sequence of the "Program"). We have here its first term: formulation of the problem. Its object is the very being of hypnotic communication: does it exist, yes or no? (the affirmative reply will come in : the very great textual distance which separates the question from the reply is determining for narrative structure: it authorizes and even obligates us to construct sequences carefully, each of them being a strand interbraided with its neighbors).
 "secondly, whether, if any existed, it was impaired or increased by the condition;"
a) In the sequence "Program," here occurs the second problem (note that problem II is tied to problem I by a logic of implication: if yes . . . then; if no, the whole tale will collapse; the alternative, in accordance with the claim of the discourse, is therefore dodged).
b) Second sub-sequence of "Program" -- this is problem II: the first problem concerned the being of the phenomenon; the second concerns its measurement (all this is very "scientific"). The reply to the question will be given in . Susceptibility is increased: "In such experiments with this patient I had never perfectly succeeded before . . . but to my astonishment, [. . .]."
 "thirdly, to what extent, or for how long a period, the encroachments of Death might be arrested by the process."
a) This is problem III posed by the "Program."
b) This problem III is, like the others, formulated -- this formulation will be reasserted emphatically in . The formulation implies two sub-questions: 1) up to what point does hypnosis permit life to encroach on death? the reply is given in : up to the point which still includes language; 2) for how long? This question will not be replied to directly: the encroachment of life on death (the survival of the hypnotized dead man) will cease at the end of seven months, but this will be by the arbitrary intervention of the experimenter. One can then suppose: infinitely, or at very least, indefinitely within the limits of observation.
 "There were other points to be ascertained,"
The "Program" mentions under a global form other problems that can be appropriately posed to the anticipated experiment. The sentence is equivalent to et cetera. Valery said that in nature there was not et cetera. We can add: neither is there in the unconscious. In fact et cetera belongs only to seeming discourse: on the one hand it seems to play the scientific game of the grand program of experimentation, it is an operator of the pseudo-real; on the other hand by glossing over, by ducking the other problems, it reinforces the sense of the questions previously stated: the strongly symbolic has been pronounced, the rest is, under the claim of the discourse, only a trifle.
 "but these most excited my curiosity," [column 2:]
Within the "Program," the point is to recall globally the three problems ("recall" or "resume," like "announcement," are terms of the rhetorical code).
 "-- the last in especial, from the immensely important character of its consequences."
a) Emphasis (term from rhetorical code) is put on problem III.
b) Again two undecidable codes: 1) scientifically, what is at stake
is recoil from a biological fact, death; 2) symbolically, this is the
of the sense which opposes Life to Death.
Actional Analysis of lexias 18 to 102
Among all the connotations we have encountered or at least marked in this opening of Poe's tale, we were able to define certain ones as the progressive terms of narrative actions. I will return at the end to the different codes that have been put into play by the analysis, in particular to the actional code. But before undertaking ,this theoretical clarification, we can isolate these sequences of actions and make use of them to account economically (while nevertheless preserving the structural import of our remarks) for what follows in the tale. In fact, it will be understood, it is not possible to analyze minutely (still less exhaustively) the whole of Poe's tale: that would take too long. We intend, however, to take up again the textual analysis of some lexias from the conclusion of the work (lexias 103110). To join the fragment which we have analyzed and that which we shall analyze, on the plane of intelligibility, it will suffice that we indicate the principal actional sequences which arise and develop (but do not necessarily terminate) between lexia 18 and lexia 102. Unhappily, we cannot, for reasons of space, give the text of Poe which separates our two fragments, nor even the numbering of the intermediate lexias. We give only the actional sequences (moreover without even being able to set out the details term by term), to the detriment of the other codes, which are more interesting and certainly more interesting, essentially because these sequences constitute, by definition, the anecdotal armature of the tale (I shall make a negligible exception for the chronological code, by indicating by an initial or final notation the point in the narrative where the outset of each sequence is situated).
I. Program: the sequence began and largely unfolds in the analyzed fragment. Problems posed by the projecred experimenr are recognized. The sequence follows its course and closes with the choice of the subject (of the patient) needed for the experiment: this will be M. Valdemar (the proposing of the program occurs nine months before the moment of the narration).
II. Hypnotizing (or rather, if this very heavy neologism be permitted: hypnotizability). Before choosing M. Valdemar as subject for the experiment, P. has tested his hypnotic susceptibility. It exists, but the results are nevertheless deceptive: M. V.'s obedience carries some resistances. The sequence enumerates the terms of this test, which is made before the experiment is decided upon and whose chronological situation is not stated precisely.
III. Medical death: actional sequences are most often stretched out, interlaced with other sequences. By informing us of the bad stare of health of M. V. and of the fatal prognosis reported by the doctors, the narrative cuts up a very long sequence which runs the whole length of the tale and will be finished only with the last lexia (150), with the liquefaction of M. V.'s body. The episodes (of this sequence) are numerous, interspersed, but nonetheless scientifically logical: ill health, diagnosis, condemnation, deterioration, agony, mortification (physiological signs of death) [page 8:] -- it is at this moment of the sequence that our second textual analysis will be situated -- disintegration, liquefacation.
IV. Contract: P. proposes to M. Valdemar to hypnotize him when he arrives at the threshold of death (since he knows he is condemned to it) and M. V. accepts. There is a contract between the subject and the experimenter: conditions, proposition, acceptance, articles of agreement, decision to execute, official recording before the doctors (this last point constitutes a sub-sequence).
V. Cataleply (seven months before the moment of narration, a Saturday, at 7:55 a.m.): the last moments of M. V. having arrived and the experimenter having been apprised by the patient himself, P. begins the hypnosis in articulo mortis, conformably to the Program and to the Contract. This sequence can be titled Catalepsy. It includes, among other terms: hypnotic passes, resistances by the subject, signs of the cataleptic state, checking by the experimenter, verification by the doctors (the actions of this sequence take up three hours: it is 10:55).
VI. Interrogation I (Sunday, 3 a.m.): P. interrogates M. Valdemar under hypnosis four times. It is pertinent to identify each sequence of interrogation by the reply the hynotized M. Valdemar makes. In this first interrogation the reply is: I am sleeping (the sequences of interrogation contain canonically the announcement of the question, the question, the delay or resistance to replying, and the reply).
VII. Interrogation II: this interrogation closely follows the first. M. Valdemar then replies: I am dying.
VIII. Interrogation III: the experimenter interrogates anew the dying and hypnotized M. Valdemar ("do you still sleep?"). He replies by tying together the first two replies he has already made: I am sleeping, I am dying.
IX. Interrogation IV: P. tries to interrogate M. V. a fourth time. He repeats his question (M. V. will reply to it beginning from lexia 105, cf. infra).
We arrive then at the point in the narrative where we shall take up again the textual analysis, lexia by lexia. Between Interrogation III and the beginning of the analysis which will follow there intervenes an important term in the sequence "Medical death": this is the mortification of M. Valdemar (101-102) . M. Valdemar, under hypnosis, is henceforth dead, medically speaking. We know that recently, occasioned by organ transplants, the diagnosis of death has again been put in question: today the testimony of electro-encephalography is needed. To attest the death of M. V., Poe himself gathers up (in 101 and 102) all the clinical signs which at his epoch attest scientifically the death of a patient: eyes opened and rolled up, cadaverous skin, extinguishing of the hectic spots, fall and slackening of the lower jaw, black tongue, general hideousness, which causes the assistants to step far back from the bed (note once more the interbraiding of codes: all these medical signs are also elements of horror; or rather, horror is always given beneath the alibi of science: the scientific code and the symbolic code are actualized at the same time, in an undecidable way).
M. Valdemar being medically dead, the narrative will have to end: the death of the hero (except in the case of religious resurrection) closes the tale. The fresh thrust of the anecdote (setting out from lexia 103) appears then at once as a narrative necessity (in order for the text to continue) and a logical scandal. This scandal is that of the supplement: in order that there may be supplementing of the narrative, there must be supplementing of life: once more, narrative counts for life.
Textual Analysis of lexias 103 to 110
(103) "I now feel that I have reached a point of this narrative at which every reader will be startled into [column 2:] positive disbelief. It is my business, however, simply to proceed."
a) We know that the announcement of a discourse to come is one term of the rhetorical code (and of the metalinguistic code). We also recognize the "appetizer" value of this connotation.
b) Duty ("it is my business") to state the facts, without being concerned about disagreements, forms part of the code of scientific deontology.
c) The promise of an incredible "real" forms part of the field of narrative considered as a commodity. That raises the "price" of the narrative. We have here, then, in the general code of communication, a sub-code, that of exchange, of which every narrative is one term, cf. (5) b.
(104) "There was no longer the faintest sign of vitality in M. Valdemar; and concluding him to be dead, we were consigning him to the charge of the nurses, [. . .]"
In the long sequence of "Medical death," which we have signaled, mortification has been noted in (101): it is here confirmed. In (101), the state of death of M. Valdemar had been described (through a catalogue of indices); here it is asserted by means of a meta-language.
(105) "when a strong vibratory motion was observable in the tongue. This continued for perhaps a minute. At the expiration of this period, [. . .]"
a) The chronological code ("a minute") underpins two effects: an effect of reality-precision, cf. (7) a, and a dramatic effect: the laborious rising up of the voice, the delivery of the cry, recalls the combat between life and death: life tries to extricate itself from the lime-twig of death, it struggles (or rather here it is death which does not manage to extricate itself from life: let us not forget that M. V. is dead: he has to hang on to not life, but death) .
b) A little before the moment we have reached, P. has interrogated (a fourth time) M. V.; and before he replies, he is clinically dead. However, the sequence Interrogation IV has not dosed (it is here that the supplement of which we have spoken comes in): the movement of the tongue indicates that M. V. is going to speak. The sequence must therefore be constructed thus: question (100) / (medical death) / effort to reply (the sequence will continue further).
c) From all evidence, there is a symbolism of the tongue. The tongue is the word (to cut the tongue is to mutilate language, as we see in the symbolic ceremony of punishing blasphemers). More, the tongue has something visceral (of the interior) and at the same time phallic. This general symbolism is here reinforced by the fact thee the tongue which moves is opposed (paradigmatically) to the black and swollen tongue of medical death (101). It is then visceral life, profound life which is assimilated to the word, and the word itself is fetishized under the specific case of a phallic organ which starts to vibrate in a kind of pre-orgasm: the minute-long vibration is the desire of enjoyment and the desire of speaking: it is the movement of Desire to arrive at something.
(106) "[. . .] there issued from the distended and motionless jaws a voice, [. . .]"
a) The sequence interrogation IV continues little by little, with a great detailing of the global term "Reply." [page 9:] Certainly, the delaying of reply is well known from the grammar of Narrative. But it has in general a psychological value. Here, the delay (and the details it brings with it) is purely psychological: it is the rising up of the voice, filmed and recorded in slow motion.
b) The voice comes from the tongue (105), the jaws are only gates. It does not come from the teeth: the voice which is getting itself ready is not dental, external, civilized (stressed dentalism in pronunciation is a sign of "distinction"), but internal, visceral, muscular. Culture values the dean, skeletal, distinct, dear (the teeth). The voice of the dead sets out from the "thick" and pasty, from the internal muscular magma, from depth. Structurally, we have here one term of the symbolic code.
(107) "[. ..] -- such as it would be madness in me to attempt describing. There are, indeed, two or three epithets which might be considered as applicable to it in part; I might say, for example, that the sound was harsh, and broken and hollow; but the hideous whole is indescribable, for the simple reason that no similar sounds have ever jarred upon the ear of humanity."
a) The meta-linguistic code is here present, by a discourse on the difficulty of maintaining a discourse. Whence the use of frankly meta-linguistic terms: epithets, to define, to describe.
b) The symbolism of the Voice is deployed: it has two characteristics: internal (hollow) and discontinuous (harsh, broken): this prepares a logical contradiction (guaranty of the supernatural): the contrast between the broken and the glutinous (108), even though the internal authorizes a sensation of distance (108).
(108) "There were two particulars, nevertheless, which I thought then, and still think, might fairly be stated as characteristic of the intonation -- as well adapted to convey some idea of its unearthly peculiarity. In the first place, the voice seemed to reach our ears -- at least mine -- from a vast distance, or from some deep cavern within the earth. In the second place, it impressed me (I fear, indeed, that it will be impossible to make myself comprehended) as gelatinous or glutinous matters impress the sense of touch.
I have spoken both of 'sound' and of 'voice.' I mean to say that the sound was one of distinct -- of even wonderfully, thrillingly distinct -- syllabification."
a) Here there are many terms of the meta-linguistic code (rhetorical): the announcement ("two particulars"), the resume ("I have spoken of") and the oratorical precaution ("I fear, indeed, that it will be impossible to make myself comprehended").
b) The symbolic field of the Voice spreads, by taking up again the in part of lexia 107: 1) the distance (absolute distance): the voice is distant because/in order that the distance between Death and Life is/may be total (the because implies a motive belonging to the real, to what is "behind" the paper; the in order that refers to the necessity of discourse which wants to continue, to survive insofar as it is discourse; in the notation because/in order that we accept the twisting together of two claims, that of the real and that of discourse, we attest the structural duplicity of all writing). The distance (between Life and Death) is affirmed in order to be better denied: it permits [column 2:] the transgression, the "encroachment," the description of which is the very object of the tale; 2) the subterranean; the thematic of the Voice is in general double, contradictory: sometimes it is a light thing, the bird-thing which flies away with life, sometimes it is a heavy thing, cavernous, which comes from below: it is the attached voice, anchored like a rock; the latter is an old mythic theme: the chthonian voice, the voice from beyond the grave (this is the case here); 3) discontinuity is the foundation of language; there is then a supernatural effect in understanding a gelatinous, glutinous, pasty language; the notation has double value: on the one hand, it underscores the strangeness of this language which is contrary to the very structure of language; and on the other hand it sums up the discomforts, the dysphoria (9): the riven and the sticking, the gluing (cf. the suppuration of the eyelids at the moment when the dead man is led back from hypnosis to waking, that is to say is going to enter into true death, 133); 4) the distinct syllabification constitutes the proximate speech of Death as a language full, complete, mature, as an essence of language, and not as a language that is stammering, approximative, babbling, a childish language, encumbered with non-language; whence the frightful and the terrible ("wonderfully, thrillingly distinct"): there is a gaping contradiction between Death and Language; the contrary of Life is not Death (that is a cliche), but Language: it is undecidable if Valdemar is living or dead; what is sure is that he speaks, though one cannot refer his speech to Death or to Life.
c) We note an artifice which belongs to the chronological code: "I thought then, and still think"; three temporalities are here co-present: time of history, of the diegesis (10) ("I thought then"), time of writing ("I think at the moment I am writing"), time of reading (carried along by the present of writing, we ourselves think at the moment we read). The combination produces an effect of reality.
(109) "M. Valdemar spoke -- obviously in reply to the question I had propounded to him a few minutes before. I had asked him, it will be remembered, if he still slept."
a) Interrogation IV is still in course: the question is here recalled (cf. 100), the reply is announced.
b) The speech of the hypnotized dead man is the reply precisely to problem III, posed in (14): up to what point can hypnosis arrest death? Here is the reply to this problem: up to the point of language.
(110) "He now said: 'Yes; -- no; -- I have been sleeping -- and now -- now -- I am dead."'
From the structural point of view, this lexia is simple it is the term "reply" ("I am dead") to Interrogation IV. Nevertheless, beyond the diegetic structure (presence of the lexia in an actional sequence) the connoration of the word ("I am dead") has an inexhaustible richness. Certainly there exist numerous mythic narratives where death speaks; but that is in order to say: "I am alive." There is, here, a veritable hapax of narrative grammar, a staging of a speech that is impossible insofar as it is speech: I am dead. Let us try to unfold some of its connotations:
1) We have already taken up the theme of encroachment (of Life on Death) . Encroachment is a paradigmatic [page 10:] disturbance, a disturbance of sense. In the paradigm Life/Death, the bar is normally read "against" (versus). It would suffice to read it "over," in order to produce encroachment and destroy the paradigm. That is what happens here. There is an undue nibbling of one space on another. The interesting thing is that the encroachment comes here at the level of language. The idea that death might continue to act once dead is banal. That is what the proverb "The dead seizes upon the living," that is what the great myths of remorse or of posthumous vengeance, say. It is what the witticism of Forneret says comically: "Death teaches incorrigible people how to live." But here, the action of death is a pure action of language and, what is an acme, this language serves for nothing, it does not come with a view to action on living persons, it says nothing if not itself, it designates itself tautologically. Before saying "I am dead," the voice says simply "I speak." It is a little like an example of grammar which refers to nothing other than language. The uselessness of the utterance is part of the scandal: the point is it affirms an essence which is not in its place (the displaced is the very form of the symbolic).
2) Another scandal of the enunciation is the turning of the metaphoric back into the literal. It is in effect banal to utter the phrase "I am dead!": that's what a woman says who has spent the whole afternoon shopping at Bloomingdale's, who has gone to her hairdresser, etc. The reversal of metaphoric into literal, for just that metaphor, is impossible: the utterance "I am dead," strictly literally, is precluded (even though "I am sleeping" remains literally possible within the field of hypnotic sleep). The point here then, so to speak, is a scandal of language.
3) The point is also a scandal of the structure of language (and no longer of discourse). In the ideal sum of all possible utterances in the language system, the coupling of the first person (1) and of the attribute "dead" is precisely what is radically impossible: it is the empty point, the blind spot of language structure which the tale will occupy very exactly. What is said is nothing but this impossibility: the phrase isn't descriptive, it isn't constative (11), it yields no other message than its own utterance: one can say in a sense that it's a matter here of a performative, but of a kind, surely, that neither Austin nor Benveniste foresaw in their analyses (let us recall that the performative is the mode of utterance according to which what is uttered refers only to the fact of its being uttered: I declare war; performatives are always, by necessity, in the first person, otherwise they slide toward the constative: he declares war). Here, the infelicitous phrase performs an impossibility.
4) From a properly semantic point of view, the phrase "I am dead" asserts two contraries at the same time (Life, Death): it is an enantioseme (12), but once again, itself unique: the signifier expresses a signified (Death) which contradicts its utterance. And nevertheless, it is necessary to go yet further: it isn't a matter of a simple negation, in the psychoanalytic sense, "I am dead" meaning then "I am not dead," but rather an affirmation-negation: "I am dead and not dead." There is here the paroxysm of transgression, the invention of an unheard-of category: the true-false, the yes-no, the death-life is thought as a whole, indivisible, uncombinable, non-dialectic, for the antithesis [column 2:] implies no third term. It isn't a two-faceted entity, but a new and single term.
5) On the "I am dead," a psychoanalytic reflection is still possible. I have said that the phrase effected a scandalous return to the literal. That means that Death, as primordial repressed, erupts directly into language. This return is radically traumatic, as is shown by the image a little further on of the explosion (147: "amid ejaculations of 'dead!' 'dead!' absolutely bursting from the tongue and not from the lips of the sufferer . . ."): the speech "I am dead" is an exploded taboo. Now, if the symbolic is the field of neurosis, the return of the literal, which implies that symbols are precluded, opens the space of psychosis: at this moment in the short story, every symbol ceases, as does every neurosis. It is psychosis which enters into the text, by the spectacular precluding of the signifier: the extraordinary of Poe is very much that of madness (13).
Other commentaries are possible, notably that of Jacques Derrida (14). I confine myself to what can be drawn from structural analysis, trying to show that the unheard-of phrase "I am dead" is not at all an unbelievable utterance, but much more radically an impossible uttering.
Before coming to some methodological conclusions, I shall recall, on the purely anecdotal level, the end of the story: Valdemar remains dead under hypnosis for seven months; with the agreement of the doctors, P. then decides to wake him; the (mesmeric) passes succeed and a little color returns to Valdemar's cheeks; but while P. tries to accelerate the subject's waking by intensifying the passes, cries of "dead! dead!" explode on his tongue, and at one stroke, his whole body shrinks, crumbles, rots beneath the hands of the experimenter, leaving no more than a "nearly liquid mass of loathsome -- of detestable putridity."
The remarks which will serve to conclude these fragments of analysis will necessarily not be "theoretical." Theory is not abstract, speculative: the analysis itself, although bearing on an accidental text, was already theoretical, in the sense that it observed (that was its goal) a language in process of forming itself. That is to say -- or to recall -- that we have not proceeded to an explication of the text: we have simply tried to seize the narrative insofar as it constructs itself (which implies simultaneously structure and movement, system and infinity). Our structuration does not go beyond what reading accomplishes spontaneously. The point here then is not, by way of conclusion, to yield up the "structure" of Poe's tale, still less of every narrative, but only to retur4 in a freer way, less attached to the progressive unrolling of the text, to the principal codes which we have marked.
The word code itself ought not to be understood here in the rigorous, scientific sense of the term. The codes are simply associative fields, a supra-textual organization of notations which impose a certain idea of structure. The occurrence of the code is for us essentially cultural: codes are certain types of deja-vu, of already seen, already read, already made: the code is the form of this already which is constitutive of the writing of the world.
Although all the codes are in fact cultural, nevertheless there is one of those we have encountered which we shall call preferentially cultural code: this is the code of [page 11:] knowledge, or rather of human knowings, of public opinions, of culture such as is transmitted by the book, by teaching, and in a more general, more diffuse way, by the whole of society. This code has as its reference knowledge, insofar as this is a body of rules elaborated by society. We have met many of these cultural codes (or rather sub-codes of the general cultural code): the scientific code, which rests (in our tale) at once on the precepts of experimentation and on the principles of medical deontology the rhetorical code, which gathers up all the social rules of saying: coded forms of narrative, coded forms of discourse (announcement, summary, etc.); meta-linguistic statement (discourse speaks about itself) forms part of this code; the chronological code: the "dating" which today seems to us natural, objective, is in fact a highly cultural practice -- which is normal since it implies a certain ideology of time ("historical" time is not the same as "mythic" time); the ensemble of chronological points of reference constitutes then a strongly cultural code (an historical manner of cutting up time for the ends of dramatization, of scientific appearance, of the effect of reality); the socio-historical code permits the mobilization, within the utterance, of the whole innate knowledge we have of our time, our society, our country (the fact of saying M. Valdemar -- and not Valdemar -- we recall, has its place here). One must not be disturbed by the fact that we might constitute as codes utterly banal notations: it is on the contrary their banality, their apparent insignificance which predisposes them to the code, as we have defined it: a body of rules so worn with use that we take them for natural features, but if the narrative departs from them, it would very quickly become unreadable.
The code of communication could also be called the code of the destination. Communication should be understood in a restricted sense. It does not cover all the signification in a text, still less its signifying; it designates only every relation which, in the text, is uttered as an address (this is the situation with the "phatic" code, which bears the burden of accentuating the relation between narrator and reader), or as an exchange (the narrative is exchanged for truth, for life) . In sum, communication should here be understood in an economic sense (communication, circulation of commodities).
The symbolic field ("field" is here less rigid than "code") is, to be sure, very large; just as we here take the word "symbol" in the most general possible sense, without burdening ourselves with any of its habitual connotations. The sense to which we refer is close to that of psychoanalysis: the symbol is in sum that feature of language which displaces the body and lets another scene than that of the utterance "be glimpsed," so that we believe we are reading it. The symbolic armature in Poe's tale is evidently the transgression of the taboo of Death, the disturbance of classification, what Baudelaire translated (very well) as the encroachment of Life on Death (and not, banally, of Death on Life). The subtlety of the tale comes in part from the fact that the utterance seems to begin from an asymbolic narrator, who has donned the role of the objective expert, riveted on facts alone, a stranger to the symbol (which does not fail to return forcefully in the short story).
What we have called the code of actions supports the anecdotal armature of the narrative. The actions, or the [column 2:] statements which denote them, are organized into sequences. The sequence has an approximative identity (one cannot determine its contour with rigor nor in an unexceptionable way). It is justified in two ways: because one is drawn spontaneously to give it a generic name (for example a certain number of notations, bad health, deterioration, agony, mortification of the body, its liquefaction, are naturally grouped under a stereotyped idea, that of "Medical Death"), and then because the terms of the actional sequence are tied to each other (from one to another, since they succeed each other throughout the length of the narrative) by an appearance of logic. We mean by that the logic which institutes the actional sequence is very impure from a scientific point of view. It is only a semblance of logic, which comes not from the laws of formal reasoning, but from our habits of reasoning, of observing: it is an endoxal logic, cultural (it appears "logical" to us that a severely correct diagnosis follows the assertion of a bad state of health). Moreover this logic is confused with chronology: what comes after appears to us as caused by. Temporality and causality, although in the narrative they are never pure, seem to us to found a sort of naturalness, intelligibility, readability of the anecdote: they permit us for example to summarize it (what the ancients called the argument, a word at once logical and narrative).
A final code has traversed (from the beginning) our tale: that of the Enigma. We have not had occasion to see it at work, because we have analyzed only a very small part of Poe's story. The code of the Enigma gathers the terms by the linking of which (as a narrative sentence) one poses an enigma, and after some "delays," which give all its spice to the narration, one reveals the solution. The terms of the enigmatic (or hermeneutic) code are highly differentiated: it is necessary to distinguish, for example, the posing of the enigma (any notation whose sense is "there is an enigma") from the formulation of the enigma (the question is put forward in its contingency). In our tale, the enigma is posed in the very title (the "facts" are announced, but one doesn't yet know on what question), formulated from the opening (in the scientific exposition of problems linked to the projected experiment), and even, from the opening, delayed: every narrative evidently has an interest in delaying the solution of the enigma which it poses, since this solution will sound its own death insofar as it is a narrative: we have seen that the narrator uses a whole paragraph to delay putting the case, under cover of scientific precautions. As to the solution of the enigma, it isn't here of a mathematical order. It is in sum the whole narrative which replies to the question at the opening, the question of the facts (these facts can however be condensed to two points: the uttering of the "I am dead" and the abrupt liquefaction of the dead man at the moment of his hypnotic awakening). The facts are not here the object of a revelation, but of a revulsion.
Such are the codes which have traversed the fragments which we have analyzed. We voluntarily do not structure them any more than this, voluntarily do not attempt to distribute the terms, within each code, according to a logical or semiological schema. For us the codes are only the points of departure of the already-read, the enticing beginnings of inter-textuality: the unravelled character of the codes is not what contradicts structure (as, it is believed, life, imagination, intuition, disorder contradict [page 12:] system, rationality), but is on the contrary (this is the fundamental affirmation of textual analysis) the integrating part of structuration. It is this "unravelling" of the text which distinguishes structure -- the object of structural analysis properly speaking -- from structuration -- the object of the textual analysis which I have tried to practice here.
The textile metaphor which I have just employed is not haphazard. Textual analysis insists in effect on representing the text as a tissue (this is moreover the etymological sense), as a braid of different voices, of multiple codes, at once interlaced and unfinished. A narrative is not a tabular space, a plane structure, it is a volume, a stereophony (Eisenstein laid much stress on counter-point in his mises-en-scene, thus intriguingly taking the lead in identifying film and text): there is a field of hearing of a written narrative; the mode of presence of the sense (except perhaps for the actional sequences) is not development, but outbreak: calls for contact, for communication, posings of contract, of exchange, outbreaks of references, of gleams of knowledge, duller, more penetrating strokes that come from "the other scene," that of the symbolic, discontinuous with the actions which are attached to a single sequence, but in a cowardly way, ceaselessly interrupted.
This whole "volume" is drawn forward (in the direction of the end of the narrative), thus provoking the reader's impatience, under the influence of two structural procedures: a) distortion or twisting out of shape: the terms of one sequence or of one code are separated, braided with heterogeneous elements; a sequence seems to be abandoned (for example the deterioration of Valdemar's health), but it is taken up again later, often very much later; an expectation is created; now we can even define sequence: this floating micro-structure which constructs not a logical object, but an expectation and its resolution; b) irreversibility: despite the floating character of structuration, in the classic readable narrative (such as a tale by Poe), there are two codes which maintain a vectored order, the actional code (founded on a logico-temporal order) and the code of the Enigma (the question crowned with its solution); thus the narrative is made irreversible. It is evident that modernist subversion will bear down on this point: the avant-garde (to stick to a useful term) tried to render the text reversible from part to part, to expel the logico-temporal residue, to attack the empirical (logic of behaviors, actional code) and the truth (code of enigmas).
However, we must not exaggerate the distance which separates the modern text from the classical narrative. We have seen, in Poe's tale, that a single phrase very often refers to two codes at once without our being able to choose which is the "true" one (for example, the scientific code and the symbolic code): what is peculiar to the narrative, from the moment it attains to the quality of a text, is to constrain us to the undecidability of the codes. By what authority would we decide? In the name of the author? but the narrative gives us only a speaker, a performer who is caught in his own production. In the name of this or that criticism? all are open to objections, swept away by history (which does not mean that they are useless: each participates, but for one voice only, in the volume of the text). Undecidability is not a weakness, but a structural condition of narration: there is no univocal determination of the utterance: in an utterance, many codes, many voices, are there, without prevalency. Writing is [column 2:] precisely this loss of origin, this loss of "motives" to the profit of a volume of indeterminations or overdeterminations: this volume is precisely signifying. Writing arrives just exactly at the moment when speaking ceases, that is to say sets out from the instance when one can no longer point out who speaks and when one only asserts that there begins to speak.
(1) I have attempted the textual analysis of a whole narrative (this will not be possible here, for reasons of space) in my book S/Z (Seuil, 1970) [Barthes' note. S/Z has been translated by Richard Miller (New York: Hill and Wang, 1974)].
(2) For a stricter analysis of the notion of lexia, as also of the working procedures which will follow, I must refer to S/Z [Barthes' note].
(3) Histories extraordinaires, tr. Charles Baudelaire (Paris: NRF, Livre de Poehe, 1969), pp. 329-345 [Barthes' note. All quotations from "The Facts in the Case of M. Valdemar" are given here from The Complete Works of Edgar Allan Poe, ed. James A. Harrison (1902; rpt. New York: AMS Press, l9G5) VI, 154-166].
(4) The technical term derives from a Greek word meaning "to point out, to point out by words, show, prove."
(5) From a Greek word meaning "to carry along", Barthes' phrase following the colon explains the technical term.
(6) "I" appears for the first time in Baudelaire's translation, though not in the English text.
(7) "Phatic" messages establish, prolong, or discontinue communication. They are therefore directed at the person addressed. See Roman Jakobson, "Linguistics and Poetics," in Style in Language, ed. Thomas A. Sebeok (Cambridge, Mass.: MIT Press, 1964), pp. 355-356.
(8) A word whose meaning is determined in relation to the message in which it occurs. For example, the last three years" means the three years preceding the uttering of the sentence in which occurs the phrase "last three years."
(9) Medical term, "state of uneasiness or anxiety."
(10) Originally a Greek rhetorical term for the narratio or "relation of facts" in a judicial speech; recently current in film criticism. The diegesis" is the world or events reported in the fiction hence, a fictional world treated by the story as if real.
(11) "Performative" and "constative" are terms taken from "speech-act" analysis in philosophy. See J. L. Austin, How to Do Things With Words (New York: Oxford Univ. Press, 1965) and John R. Searle, Speech Acts: An Essay in the Philosophy of Language (Cambridge: University Press, 1970).
(12) "Enantiosis" has three main meanings: in grammar, a kind of antithesis; in philosophy, each of the ten oppositions, which, according to the Pythagoreans, were the source of all things; in medicine, treatment by contraries. The suffix "--eme" here designates a unit in an analytic system based on significance.
(13) Barthes' terms are from Jacques Lacan, an influential and sometimes idiosyncratic interpreter of Freud. "Primordial repressed" is that initial repression which establishes the unconscious. That repression has the character of a negation, a lack or absence, of which "Death" is obviously exemplary. The "symbolic" is a realm of mental phenomena structured like a language, the structure reflecting the congruence between the unconscious and the social systems within which the "individual" emerges. By maintaining the difference of words and ideas from things, this realm permits symbolic representation, along with the deviations or substitutions of symbols (words and ideas) which characterize neurosis. The psychotic, Freud asserted, confuses words and things, treating words as things. In this sense, psychosis is a kind of literal-mindedness. See Jacques Lacan, The Language of the Self: The Function of Language in Psychoanalysis, tr. with commentary by Anthony Wilden (Baltimore: The Johns Hopkins Press, 1968); and French Freud: Structutral Studies in Psychoanalysis, ed. Jeffrey Mehlman, Yale French Studies no 48 (1972), especially pp. 151-63 and the excerpts from Vocah$slary of Psychoanalysis by Lacan's students J. Laplanche and J.-B. Pontalis.
(14) Jacques Derrida, La voix et le phenomene (Paris: Presses Universitaires de France, 1967), pp. 60-61 [Barthes' note. Derrida analyzes the expression "I am mortal," not Poe's tale].
~~~ End of Text ~~~
[S:0 - PSDR, 1977]
~~~ End of Text ~~~
[S:0 - PSDR, 1977]