pdf Download PDF

Mallarmé’s Madness:
Poetry, Pedagogy, and Translation

Stéphane Mallarmé was a notoriously bad English teacher. Reports by school inspectors and former students criticize not only Mallarmé’s lack of attention towards his pedagogical duties but also his poor knowledge of the English language and his failure to improve.1 In an attempt to assuage his critics and maintain his teaching position, Mallarmé wrote a series of pedagogical works: Les Mots anglais, a philological study of the English language, complete with a dossier of two manuals of grammatical rules and translations, Thèmes anglais pour toutes les grammaires, and Recueil de “Nursery Rhymes.”2

These texts, however, did little in the way of reassuring his detractors. On the one hand, Les Mots anglais stops short of expounding the grammatical rules of the English language, opting instead to construct families of English words on the basis of seemingly fortuitous sound-sense associations, thereby offering itself more as the illustration of a theory of poetics than as a pedagogical manual. On the other hand, the book of Thèmes and the collection of nursery rhymes present not only an unconventional choice of materials for language teaching (old proverbs, out-of-date sayings, and children’s songs) but also an idiosyncratic method of translation (word-by-word back-translations, replete with errors, mistranslations, and misunderstandings) that constitutes an arguably questionable pedagogical tool.

Unsurprisingly, these works failed to mitigate the worries of Mallarmé’s employers and students; worse still, they fueled the already negative opinions about his teaching methods, eliciting disparaging judgments that connected his teaching failings to his famously obscure poetic productions and, ultimately, to speculations [End Page 141] about his mental health. According to class inspectors who assessed Mallarmé’s teaching, the manuals were deemed to be mad products of an unstable mind, filled with “niaiseries,” on a par with those “étranges élucubrations” or “productions insensées” that had brought him poetic fame—a sentiment that is summed up as follows: “On serait tenté de se demander si l’on n’est pas en présence d’un malade” (quoted in Mondor, Vie de Mallarmé, 371, 414).3

Madness would not only explain why Mallarmé was such a bad teacher but it would also explain why his poetry was as impenetrable as it was renowned to be. Indeed, madness would be responsible for the linguistic idiosyncrasy that was believed to be at the origin of that strange and foreignizing French that Jacques Scherer dubbed le mallarméen—an idiolectal construct, founded upon an non-idiomatic, yet systematized, reorganization of syntax, understandable in full only by its inventor and sole practitioner, Mallarmé.4 As such, the term madness would be synonymous, to a degree, with the more commonly used terms “obscurity” and “difficulty,” which have come to overdetermine the reception of Mallarmé’s poetry as being fundamentally resistant to interpretation.5

Yet, subsuming the term to these readings limits its potential for reassessing Mallarmé’s work and its reception. Indeed, the suggestion that there might be an enlightening correspondence between the “niaiseries” of his teaching and the “productions insensées” of his poetry is an evocative one.6 It implies bringing together two sides of Mallarmé that have typically been maintained as separate by critics—on the one hand, the traditional image of Mallarmé the poet, haunted by the pursuit of a pure language and author of some of the most hermetic poetry of all time; on the other, the less well-known figure of Mallarmé the teacher and translator, creator of what critics have called an “everyday poetics.”7

Bringing these two images together not only goes against much of the scholarship, but also runs counter to Mallarmé’s characterization of his work. Mallarmé clearly dismissed the works he did for pedagogical purposes, describing them as “concessions à la nécessité” and “des besognes propres et voilà tout . . . dont il sied de ne pas parler,” and commenting on the fact that teaching English to French schoolchildren was a painful chore that interrupted the noble activity of poetic composition.8 And, if he did on occasion express enthusiasm for the practice of translation, this mainly referred to his translations of poems by Edgar Allan Poe, whose work he admired to such an extent that he credited it with having motivated him to learn English in the first place—“ayant appris l’anglais simplement pour mieux lire Poe” (Œuvres, 1:788).

Why, then, should we even read the material that Mallarmé himself rejected, let alone read it in conjunction with the canonical poems? Furthermore, if the pedagogical works and the poetry can be thought to shed light on each other, to what extent does this occur through a reflection on madness? While we must remain skeptical of both Mallarmé’s dismissal of these works and the inspector’s comments, the latter’s accusations of madness resonate with certain uses that Mallarmé himself makes of the notion, thereby offering a compelling reason for examining the rejected material.9 Indeed, the notion of madness underpins the existential and literary crisis Mallarmé underwent in 1866, known as the “crise de Tournon,” out of which he developed a poetics of linguistic purification, theorized in the aptly-named text “Crise de vers.” [End Page 142]

Madness, in this context, affects both personal and public spheres. On a personal level, this moment is characterized by what Mallarmé refers to in his correspondence alternately as “folie,” “insanité,” “maladie,” “mal,” “malaise,” “anxiété,” “agonie,” “angoisse,” “hystérie,” and “épuisement nerveux”—terms that encompass both physical and psychological dimensions. While Mallarmé’s health was indeed a constant source of worry to him, the deeper source of his anxiety derived from his reflection on the nature of language and on the function of poetry. In expressing his fears, he writes “j’ai presque perdu la raison et le sens des paroles les plus familières,” thereby showing the extent to which madness and language are inextricably tied to one another: the risk of losing one’s mind is inseparable from the fear of losing the meaning of familiar words and the ability to use them (727).

The public dimension of this private affliction takes the form of a crisis of language, prompted by the realization that language is defective. Mallarmé refers to this as “le défaut des langues” and characterizes it in two ways: on the one hand, no one language is unique—“les langues imparfaites en cela que plusieurs, manque la suprême”; on the other, languages work by establishing an arbitrary relationship between words and their referents.10 Mallarmé illustrates this second point by showing that French contains what he considers to be an aberrant reversal: the word for night, “nuit,” sounds too bright when compared to the somber word for day, “jour.”11 Madness, then, is associated with linguistic plurality and with the enduring, yet always unsatisfied, need for translation that such plurality creates; but at a deeper level, madness is related to the realization that language is by nature representational and that it functions by breaking ties with the world it represents and by exposing its users to the experience of that rupture.

Such a characterization of madness reveals an important connection between Mallarmé’s poetry and his practice of translation. This occurs in the particular way in which madness defines our relationship to language by highlighting the fragility of the ties that exist between languages and between words and their referents. And because exploring these in-between spaces participates in a process of linguistic purification, which implies breaking the conventional ties that bind words to their referents in order to reveal the non-idiomatic nature of language as such, madness can be seen as residing at the very core of Mallarmé’s search for a pure language. Indeed, is it not the case that the purer the language, the further it moves away from idiom and the closer it comes to pure sound, at the risk of its users losing their senses and the sense of familiar words?

The study of English, and its relationship to French, plays an important role in understanding le mallarméen not because it might reveal that Mallarmé’s writing is influenced by his knowledge of this foreign language, but rather, because it suggests that the problem of linguistic default can be approached from the perspective of the inter-linguistic space that lies in the passage between languages. This, in turn, opens up new ways of understanding both the nature of Mallarmé’s private malady and the public crisis of linguistic insufficiency.

The far-reaching implications of such an analysis go beyond Mallarmean criticism; they impinge on the study of translation and, more fundamentally, on our understanding of how language works. The space in which each language is taken out of itself [End Page 143] without, however, becoming other is defined by the peculiarities that are inherent in the process of translation. This process involves both a practical exploration of the grammatical structures of a given language in its relationship to another and a theoretical reflection on the limits that keep languages apart and on what it means to push beyond those limits. More specifically, translation, and the errors to which it gives rise, reveal the fine line between sense and senselessness, opening up languages to their points of rupture, thereby enabling an exploration of that elusive notion of linguistic fault. But, in revealing the tenuous nature of languages in their relation to the world and to the creation of meaning, translation also offers a potential remedy to this fault, by way of a process of linguistic purification. As such, it renders us acutely aware of what is involved in using language, of our obstinate yet always unfulfilled search for meaning, and of why we fear that the absence of meaning inevitably leads to the loss of our senses.

Mallarmé’s pedagogical manuals, then, are more than just failed teaching tools or secondary works; they offer new avenues for exploring how Mallarmé thinks about language in general. More specifically, they constitute, on a practical level, the counterpart to the more theoretical explorations into pure language that Mallarmé conducts in his poetry. Examining the pedagogical texts alongside the poems (and in particular those poems that have become emblematic of Mallarmé’s hermeticism and search for purity, “Le Tombeau d’Edgar Poe” and the “Sonnet en –yx”) not only confirms the importance of translation in Mallarmé’s work, by revealing how translation participates in the process of linguistic purification, but it also allows us to ask some important theoretical questions about the limits of translatability, the nature of meaning in relation to the translative process, and the problems inherent in language use.12

Madness and Method

The pedagogical works and the aforementioned poems stem from a particular moment in Mallarmé’s trajectory that places madness at the forefront of his preoccupations—namely, the crisis at Tournon.13 More precisely, his reflection on madness is inseparable from a preoccupation with linguistics and poetics. As per his own admission, the crisis was brought on by compulsive attempts to dig into the soil of poetry—“en creusant le vers à ce point, j’ai rencontré deux abîmes, qui me désespèrent”—and it was in turning to the study of linguistics that he managed to overcome writer’s block (Œuvres, 1:696). The scientific foundation for his poetry was to become his “refuge,” with the aim of, ultimately, returning him to poetry.14

The triad of linguistics, poetics, and madness connects Mallarmé to a tradition of literary logophilia or folie littéraire. This tradition exemplifies a particular relationship between language and literature, one that underpins much of modernist writing—namely, a quasi-scientific fascination with the materiality of the linguistic sign, the development of strategies to dismantle grammatical structures and recompose them according to idiolectal laws, and the exploration of multilingual spaces via the practice of translation.15 Loosely associated with this tradition are writers such as Jean-Pierre Brisset, the grammarian who traced the origins of the French language back to the [End Page 144] croaking of frogs; Raymond Roussel, whose proto-Surrealist work consists of a secret method of writing via imperceptible orthographic changes; Ferdinand de Saussure, the father of modern linguistics, obsessed with finding anagrams in Latin Saturninan verse; and Louis Wolfson, the American self-declared schizophrenic, who developed a multilingual translation system in order to destroy the English language.

As the etymology of the term logophilia implies (“love of the word”), this practice embraces the semantic field of fascination, passion, compulsion, extending into extreme and destructive forms of fervor. Thus, it is tied up with a certain conception of madness, as an experience of language that involves rupturing the linguistic sign—a connection that Michel Pierssens notes in his seminal study of the logophilic tradition: “Toute recherche de langage a donc pour rebord la folie. . . . Parce qu’il est articulé, rien de plus fragile que le signe, rien de plus angoissant que sa béance permanente et menaçante.”16

Pierssens qualifies his description of logophilia by linking it to the phenomenon of cratylism, which attempts to overcome the arbitrariness of the relationship between word and world: “Le signe est bien le maillon le plus faible du symbolique; le cratylisme chez chacun de nous, est l’effort infini de suturer la refente du signe pour y trouver sens et jouissance” (La Tour de Babil, 58). In Pierssens’s reading, cratylism and logophilia are two sides of the same coin; they share a fascination with the materiality of the linguistic sign. But, whereas cratylism aspires to resolve the scandal that words do not correspond to things, logophilia revels in exploiting the arbitrary relationship between signifier and signified.

Both tendencies are at play in the two “sides” of Mallarmé’s production: if his poetic explorations are motivated by the need to discover the supplement that will “remunerate” the default of language, his pedagogical works constitute the place in which he explores the limits and nature of that arbitrary space; and he does so by passing through another language, via the practice of translation.17 Translation, then, becomes central for thinking about the nature of the linguistic sign and, consequently, for understanding what is at stake in le mallarméen.

The Pedagogical Manuals

Both the book of Thèmes and the Recueil offer an idiosyncratic system for teaching English via translation. The first of the two manuals presents itself as a collection of sentences in French to be translated by students into English. These sentences are grouped together on the basis of particular grammar rules, which their translation is supposed to illustrate. This structure is not in itself unusual: translation is commonly used in foreign language teaching, whether it be translation from French into the foreign language (thème) or translation from the foreign language into French (version). Mallarmé’s text is unconventional, however, because he gives his students sentences in French that he has already translated from their original English; the aim is then for the student to re-translate, or back-translate, the sentences into English. Such a system may indeed constitute a “mad” method for teaching a foreign language because it is [End Page 145] not only far from simple, given the many steps involved, but it also creates ample room for the generation and propagation of mistranslations, misunderstandings, and errors.

In the introductory remarks to the book, however, Mallarmé claims that his system is founded upon the simplest of principles: “Il n’y a pas lieu à innovations dans un exercice aussi traditionnel que le Thème, et tout est là simplement: trouver des phrases qui, sous leur aspect français, sachent tout d’abord éveiller l’intérêt, puis rendues à l’Anglais originel, possèdent comme un droit à se fixer dans l’esprit.”18 The rule for the teacher: find sentences in English which, when translated into French, prove worthy of “interest” for the student, and which, when re-translated back into English by the student are entitled, “by law” and “rightfully,” to take up residence, become immobilized, and memorized, in the mind of said student. The task of the student: find the original form of the English sentence underneath the “calque” of the word-for-word translation and memorize the re-translation: “Dictons, proverbes, locutions spirituelles et imagées, toute la richesse intime de l’Anglais a été mise à contribution ici et voici que retraduits de leur calque français dans leur forme originelle anglaise par l’élève ils restent en sa mémoire, appris en leçons après avoir servi de thèmes” (Thèmes, 20).19

This is easier said than done. Firstly, the choice of proverbs is often very obscure. Mallarmé lifted the proverbs and sayings directly from an edition of a seventeenth-century English collection. If it is possible, though unlikely, that nineteenth-century English men and women might have been familiar with the meaning of an expression such as “She goes as if she cracked nuts with her tail” (which refers to a woman with a provocative gait), it is doubtful that nineteenth-century French schoolchildren would have been able to decipher the meaning of the expression “Elle va comme si elle écrasait des noix avec la queue” (not to mention its suitability for his audience) (278, 170). Secondly, the obscurity of the sentences appears to run counter to Mallarmé’s recommendation that they should elicit interest from the student—one is hard-pressed to imagine that children would be riveted by such expressions as “Il y a plus de filles que Maukin et plus de garçons que Michel” or “Aussi longue que Margot de Westminster” (31).20

Moreover, because Mallarmé’s own translations into French contain a number of mistakes, the chances of the student “finding” the original underneath those mistakes are somewhat slim. This is the case, for instance, with the sentence “Rire dans la manche de quelqu’un,” which is offered as the translation of “To laugh in one’s sleeve”: here, the shift in focus from one’s own sleeve to someone else’s loses the meaning implicit in the original of laughing secretly to oneself (127, 263). Other confusions emerge for a variety of reasons. Some derive from apparently inexplicable changes, such as “Un signe de tête pour le sage et une verge pour le fou,” which renders the English “A nod for a wise man and a rod for a school,” where “school” appears instead of “fool.” (43, 235). Some are the result of apparent inadvertencies: the English sentence “Seven may be company, but eight are confusion” is translated into “Sept peuvent composer une société, mais neuf font une confusion” (250, 91). Others are plain mistranslations: in translating “Sloth is the mother of poverty” into “La malpropreté est la mère de la pauvreté,” Mallarmé replaces idleness with uncleanliness (232, 35). [End Page 146]

Errors and confusion also abound in the book of nursery rhymes. At first glance, the collection appears to present itself as a traditional book of versions, that is, a book of translations from English into French. Indeed, the book (or rather, the fragments that were collected together by Carl Paul Barbier, given that the book was never published during Mallarmé’s lifetime) consists of a series of leçons, which, in turn, are subdivided into three sections: the English nursery rhyme, a free-form translation into French (under the misleading heading of thème), and a series of notes intended to shed light on a particular set of grammatical problems and word-for-word translations of choice vocabulary. While this presentation seems straightforward, when we take a closer look at each leçon we are faced with problems of translation similar to those encountered in the book of Thèmes.

In the introduction to the Recueil, Barbier remarks on the various errors of translation that pepper the collection. For instance, he mentions a case in which Mallarmé’s own inadvertency caused confusion. In copying the story of the man from Newington who jumped into two hedges—the first time scratching his eyes out and the second time putting them back again—Mallarmé omits the second to last word (“in”), thereby finding himself at a loss when translating the song:

Leçon n.44 There was a man of Newington, And he was wond’rous wise, He jump’d into a quickset hedge, And scratch’d out both his eyes: But when he saw his eyes were out, With all his might and main He jump’d into another hedge, And scratch’d them again.


Un clown serait malin, s’il représentait l’homme de Newington!—Quel homme?—Vous ne savez pas? Celui qui saute dans une haie vive et s’arrache les deux yeux:—mais quand il vit ses yeux ôtés . . .—Avec quoi les vit-il, avec quoi?—C’est vrai, je n’y avais jamais songé; probablement qu’il les remit.—Mais alors, il ne les vit plus ôtés?—C’est pour cela que la chanson dit qu’il sauta dans une autre haie et les arracha de nouveau.—Malgré tout, voilà qui ne paraît pas très clair.

(Recueil, 71)21

The notes that Mallarmé attached to his translations of the nursery rhymes offer little in the way of clarification. Here the reader finds a peculiar interlanguage that Mallarmé uses in order to pass from one language to another, and that he offers to his students in the form of presumably helpful notes. These notes are typically divided into two sections: Leçon, which comprises notes relevant to the nursery rhyme itself, and Devoir, which refers to Mallarmé’s own translation into French.

In the latter section, Mallarmé provides translations of French syntactic units. His translations, however, are unusual in that they consist of French ungrammatical word-by-word copies of what the English equivalent would be. For instance, “que [End Page 147] lui arriva-t-il” becomes “quoi arriva à lui,” intended to then facilitate the work of the student who would ideally produce the English “what happened to him” (35). Other examples are equally ungrammatical: the French “une de ses amies” becomes “une amie des siennes,” in order to produce the English “a friend of hers”; “Que c’est joli!” becomes “combien joli c’est!” to be translated as “how pretty it is!”; “Sans vous réveiller” becomes “sans réveillant vous” for “without waking you”; “Ce bonhomme de paille immobile” is rendered as “cet immobile paille-bonhomme” in order to produce “this immobile strawman” (56, 42, 35, 51). And the list could go on.

It appears, from these fragmentary notes, that Mallarmé’s intention was for the student to translate back into English Mallarmé’s own “translation” of the nursery rhyme into French, as was the case in his book of Thèmes (this would, then, explain his choice of the term thème to refer to the versions—his translations of nursery rhymes into French); and the translation was to be aided by the insertion of this transitional language. Indeed, this intermediate language—which is neither French nor English, nor can it be defined as a language per se, but rather a mixture of both—is context-specific and makes sense only within the transition between Mallarmé’s French version of the nursery rhyme and the student’s English translation of it. In this movement, the meaning of the original is not lost; rather, these ungrammatical forms facilitate an unequivocal rendition of the expression that Mallarmé wants the student to produce.

The equivocation, however, comes at a later stage: this is because the system, in producing ungrammatical expressions, leads towards loss of idiom. Translation, no longer operating unidirectionally from a source language to a target language, but circularly or à rebours within a contaminated hybrid language, temporarily produces a “franglais” monster of literalism, the result of which is that we can no longer tell what is idiomatic in either language. From a pedagogical perspective, not only does this prevent learning English, but it also risks undoing the student’s mastery of French, because both languages are being rendered unfamiliar.

The broader implications are significant: through his idiosyncratic practice of translation, Mallarmé is exploring what happens to language when it becomes non-idiomatic, in order to break up the relationship between language and referent that is based on convention. As such, translation constitutes the practical counterpart to Mallarmé’s poetry and theoretical writings, insofar as it reveals the separation between ideal and conventional uses of language—“le double état de la parole, brut ou immédiat ici, là essentiel” (Œuvres, 2:212). In the breakdown of idiom, French and English lose their defining contours as languages recognizably different from one another; and, in losing familiarity with those particular languages, we are exposed to the defamiliarizing effect of any language—which, Mallarmé aims to show, constitutes the experience of language itself.

We are now able to propose a more nuanced definition of the madness that appears to afflict Mallarmé’s pedagogical work. In one respect, we could claim that this madness lies in Mallarmé’s choice of texts. Nursery rhymes, proverbs, and old sayings belong to a linguistic and cultural terrain that treads the fine line between nonsense and magic; here language veers towards babble, childish speak, and the language of childhood, [End Page 148] where sense is not absent, but is a by-product of the sounds and shapes of words. More precisely, however, madness lies in the very practice of translation with which Mallarmé engages, the word-by-word literalism that produces a non-idiomatic interlingual space in which the very nature of language is being exposed. In this respect, madness does not correspond to loss of sense because, as we saw with the nursery rhymes, the meaning of the original is maintained; rather, madness implies loss of idiom, thereby allowing us to experience language without being limited to the experience of a particular language. The errors, then, that occur in Mallarmé’s pedagogical works reveal a tension between sense and idiom that underpins his reflection on language.

Translation Errors, Sense, and Non-Sens

Errors play a significant role in foreign language acquisition: they shed light not only on the mechanisms by which a foreign language is learned but also on the internal workings of language more broadly. Indeed, “error analysis” refers to a scientific method used in applied linguistics to track and examine those errors that occur in the intermediary stage of second language acquisition known as “learner language.” This, in turn, constitutes the basis for “interlanguage,” a unique system that is between the learner’s mother tongue and the target language. Within this context, errors in translation are particularly revealing not only of how languages are learned (or not learned) but also of how languages work (or fail to work).

In the French context, Antoine Compagnon explores the different types of error that are associated with the practice of translation (in the form of both thème and version), by commenting on the scale according to which they are judged in the school system. According to this scale, the least offensive is the faux-sens, a “faked sense” that concerns a misunderstood word; then comes the contresens, which makes sense “but a wrong sense” and relates to a whole clause or sentence; and finally, the most serious is the non-sens, wherein, for instance, “one had not been able to make any sense out of” the given text.22 Translations are graded accordingly: one-half point is detracted for a faux-sens, one point for a contresens, two points for a non-sens. While this grading scale is just a “minor matter” in itself, Compagnon claims that it is however a “major” revealer of the French mind: “French petty rationality is more tolerant of a mistranslation that makes sense than of a mistranslation that doesn’t, more tolerant of stupidity than of folly” (“Somebody Will,” 71). Folly—or, madness—is inherent in senseless mistranslation, and non-sens defines the danger of loss of sense that can occur in the movement of passing between languages.

The errors found in Mallarmé’s manuals appear to belong to the more benign categories of the faux-sens or the contresens, because most of them derive from lack of attention or imprecision. It would thus seem far-fetched to claim that these translations are instances of madness, or “folly” as defined by Compagnon. Yet it would also be shortsighted to circumscribe the meaning of the term madness to refer only to the errors produced by mistranslations: rather, madness relates to the loss of idiom that occurs in passing between French and English which, originating in the error of [End Page 149] translation, extends to and affects the surrounding language, rendering it unfamiliar. Thus, while these errors do not necessarily undermine the readability of the text itself, their cumulative effect is to break up linguistic fluency.

Marina Warner experiences this effect as a form of melancholy, which she ties to the concept of translation failure. In describing the sentences that make up the book of Thèmes, she notes: “These adages or proverbs or whatever we might call them teeter on the verge of incomprehensibility. But their cumulative effect is melancholy: failure stalks them, regardless of syntactical exactitude.”23 In using the term “failure,” Warner is not referring specifically to the pedagogical potential of the manual (which, as we have seen, is questionable); rather, the term points towards an essential characteristic of the process of translation and, perhaps, it suggests a deeper failure—one that might be thought of as a failure of language, akin to Mallarmé’s conception of “défaut.” But how are we to understand this term “failure”? What is it that translation, and language broadly speaking, fail to do? And how does this failure relate to the notion of “incomprehensibility,” invoked in Warner’s comment?

Beyond truisms such as that a translation “fails” because it will never “be” the original, failure in and of translation is a nuanced concept that has the potential of reassessing the theoretical stakes of translatability. Paul de Man notes that the concept of failure in regards to translation derives from the way in which translation poses the question of meaning in language. Taking on Walter Benjamin’s legacy of a theory of translation that ostensibly defies a concern with semantics, de Man observes that the translator fails because, unlike the poet who aims to convey “a meaning which does not necessarily relate to language,” he is caught up in a movement that does not exit from the circuit of languages.24

The concern with meaning, and its loss, is at the heart of Benjamin’s appraisal of Hölderlin’s translations of Sophocles’s plays: madness, here, lies in the movement by which translation stretches out the German language beyond its grammatical limits, to a point where sense is almost evacuated—“meaning plunges from abyss to abyss until it threatens to become lost in the bottomless depths of language.”25 Yet, despite Benjamin’s apparent fascination with the risk of senselessness, critics remark that he does not advocate a “way of meaning” without a “meant” or a “meaning”; what is at stake is a particular conception of meaning.

For Benjamin, meaning does not refer to a stable or identifiable entity; rather, the term indicates a direction in which the translation is moving. This is because semantics is dependent upon syntactical spacing: the movement involved in translation would thus be syntactical—Mallarmé’s and Hölderlin’s word-by-word literalism. And the risk of senselessness coincides with the expansion and the alteration of the limits of language that result from the changes induced by translation to the ways in which semantics is bound up with syntax. Benjamin’s appreciation of Hölderlin’s “monstrous” translations, then, stems from the fact that they create a movement that puts pressure on the limits of the German language (by expanding them and modifying them almost to the point of rupture). [End Page 150]

This movement (or direction) potentially leads to the production of new meanings. Viewing sense in these terms, then, opens the way for conceiving of errors of translation as possible sites for poetic creativity.26 Jean-Michel Rabaté suggests this in his appraisal of the errors in Mallarmé’s translations of Poe’s poems: “Error becomes productive and positive since it assaults the language and fashions a new sort of French, halfway between a misunderstood original and an excessively faithful paraphrase.”27 The implications of such a reading carry with them the risk of overlooking the disruptive potential of the error: the aberrant force of the idiolect would be explained and brought back in line with the rules of the sociolect, and Mallarmé’s “new sort of French” would ultimately become unproblematic.

This is precisely what Mallarmé’s writing is warning us against; his language is destabilizing because it refuses to conform to conventional conceptions of sense-making. The idiolectal construct that is le mallarméen functions by placing us as readers in the uncomfortable position of attempting to find a balance between our need to reorganize Mallarmé’s strange and foreignizing syntax into a more conventional form that would enable us to get to what we think Mallarmé is “saying” and our recognition that this “content” is inseparable from the ways in which Mallarmé’s syntax is making us think this very dynamic.

In so doing, we become aware of how embroiled we are in our concern with meaning, either because we are constantly attempting to find it or because we are trying to identify why it might be absent; indeed, the search for meaning becomes all the more pressing where meaning is felt to be elusive or appears to be lacking. In this process, we realize that our attachment to finding meaning not only reflects a basic human need—as we believe it to be central to the process of understanding—but also constitutes the safeguard against the fear of senselessness (the lack or loss of meaning) and the feeling of anxiety that such a fear creates in us. The defamiliarizing effects produced both by the syntactical complexity of Mallarmé’s writing and by his practice of translation allow us to shift our focus from searching for meaning to asking ourselves why the search for meaning is so important. In deferring expectations of meaning, while still retaining a guarantee of intelligibility, Mallarmé is suspending the hermeneutic process; and the defamiliarization that derives from stripping away idiom exposes us precisely to that experience of suspension.28

This process is arguably what Benjamin is getting at in his discussion of pure language, and that he identifies as the result of Mallarmé’s particular use of syntax. Benjamin quotes Mallarmé’s definition of “défaut des langues” as deriving from the multiplicity of languages—“Les langues imparfaites en cela que plusieurs, manque la supreme,”—in order to provide an example of how pure language works.29 We might illustrate this in the following terms: the process of moving between languages, in the form of translation, reveals the arbitrary nature of the linguistic sign and the fragility of how individual languages construct meaning, while at the same time shifting our experience of language as such. In this shift, we move away from a conception of language as idiom towards the experience of language as lack—experienced, in Mallarmé’s terms, as the lack of a “supreme” language. Significantly, this process is a form of linguistic [End Page 151] purification, because it is only through translation that we can see so patently, across languages, the tenuous nature of idiomatic expression.

Familial and Familiar Ties: “la retrempe”

Mallarmé explores the relationship between pure language, purification, and translation in the third of his pedagogical manuals, Les Mots anglais, by playing on the notions of defamiliarization and familiarity. While presenting itself primarily as a diachronic and synchronic study of the English language, Les Mots anglais is an exploration of the familial ties that bind English to French and on the movement of passing between the two languages—and, as such, it provides an implicit theory of translation.30

The entire text is built on the vocabulary of the family. Describing the origin of English, and its subsequent development in relation to French, Mallarmé writes: “La greffe seule peut offrir une image qui représente le phenomène nouveau; oui, du Français s’est enté sur de l’Anglais: et les deux plantes ont, toute hésitation passée, produit sur une même tige une fraternelle et magnifique végétation” (Œuvres, 2:962). In a formulation that mixes kinship and botany, English is described as both “filial” by formation—Mallarmé distinguishes between “les langues mères” (Latin, Greek, German, Russian) and “les langues filles” (Italian, Spanish, and French)—and “brotherly” by growth, insofar as it is a product of the graft of French (an already formed language) onto the unsteady ground that was Anglo-Saxon after the Norman invasion of 1066 (1096, 951).

The process of grafting French onto Anglo-Saxon implies the naturalization of a perfectly formed language into an unformed one. What, then, becomes of that portion of French that has been separated from its original source? What happens if the borrowed word gets lost—as Mallarmé asks in alarm, “Mais si le mot, anciennement prêté par nous, s’était chez nous perdu?”—(1027)? If these exiled words do return, do they come back in their original shape or in the haunting form of the memory of a loss? Mallarmé shows how, in borrowing words from French, English deploys a series of strategies, based on practices of alliteration, repetition, onomatopoeia, which ingeniously transform French words to such an extent that, once exiled, they never return: “Strict, il se plaît, comme épris lui-même de ce qu’on nomme la couleur locale, à respecter l’allure lointaine des sons, les traduisant par une orthographe ingénieuse et juste” (1078).

The observation that English “translates” the foreign word through an ingenious and just spelling implies that English possesses a cratylistic power that enables it to motivate words, which, in their own language, are arbitrarily defined. Mallarmé gives the following example: “oui, sneer est un mauvais sourire et snake un animal pervers, le serpent, SN impressionne donc un lecteur d’Anglais comme un sinistre digramme” (968). Artificially, the English language naturalizes that which is not originally its own, which means that, paradoxically, French words are more “natural” in the English language.31 Discovering the connection with English, then, might constitute a way of motivating the un-cratylistic French language, which, as we have seen, allows what [End Page 152] Mallarmé considers the scandalous sound-sense aberration of the heaviness of “jour” and the brightness of “nuit.”

The process according to which French might be able to recover its origins through English does not take the form of translation proper; rather, as Mallarmé prescribes, it implies a reimmersion—or, “retrempe”—into an original source, in order to return French to an anterior state of purity. This is arguably the main function of Les Mots anglais: an English lexicon that, in revealing the extent to which the origins of English words reside (partially) in French, allows French to envisage the possibility of going back to its origins. Mallarmé illustrates this by grouping words together into a series of tables (under the heading “Vieux mots”), which represent words that have been forgotten in French but that can still be found in English, albeit in a different form. For instance, the original French for pen, “penne” has been replaced by “plume” but it is conserved in the English form; “se souvenir” has supplanted the older form “remembrer,” which however remains in the English verb to remember; the old form for “témoigner” is “vocher,” which has become the English to vouch (1053, 1054, 1056). And so forth.

The proof that Mallarmé gives of the successful recovery of the origins of French is the feeling of pleasure that any French speaker presumably derives from saying these English words; they provide a feeling of being at home in French, because they reveal the bonds of familiarity that tie French to its past through English.32 However, from a broader perspective, the cumulative effect of the lists of words resembles that feeling of defamiliarization produced by reading the book of Thèmes. The lists rupture our sense of being at home in either French or English, insofar as they expose us not only to the arbitrariness inherent in the relationship between words and their referents across languages, but also to the fragility of the nature of memory.

The importance of memory in this context can be seen by juxtaposing two examples of etymology given by Mallarmé. The first is an example of what he calls a “jeu de mots heureux,” the successful result of a mistranslation, which retains within itself the memory of its origin: “le Chat Fidèle, sur l’enseigne antique des auberges de la Conquête, accepte du populaire un violon dont il joue avec les mots: The Cat and The Fiddle” (1045, 1046). Conversely, the second is an example of “maladresse,” caused by memory loss, which creates confusion and obfuscation: “voyez ce que donne écrevisse, crayfish ou crawfish, où fish est bien un poisson (en supposant qu’on puisse doter de ce nom générique un crustacé) mais cray n’existe pas, et craw, c’est gésier; ici l’on se borna à ne comprendre qu’une moitié du mot, dont l’autre plongeait dans les ténèbres” (1045).

With these examples, Mallarmé is pointing to the problems that emerge in the creation of new linguistic forms and, by extension, myths, when the origins of words cannot be recovered—he notes, “Mythologie, autant que Philologie” (1046). These reflections display echoes of Max Müller’s work on the relationship between mythology and linguistic pathology, which results from individual languages and cultures becoming amnesic. Mallarmé was familiar with contemporary studies on comparative mythology, having translated into French a text by Reverend George William Cox, A Manual of Mythology in the Form of Question and Answer, itself a rewriting of Müller’s work. In the text that he produced, Les Dieux antiques, which was intended to supplement his [End Page 153] pedagogical manuals, Mallarmé expounds Müller’s theory: the passage of time and the separation between tribes render cultures and languages forgetful of their origins; this, in turn, produces mythological anomalies that make us wonder whether language, or indeed, mankind has gone mad.33 In such a context, the role of translation becomes essential: translation offers a solution to Mallarmé’s (via Müller’s) fear that language itself has gone mad and that words have lost their memories in the Babelian dispersal of tribes across space and through time.34 Translation constitutes the vehicle for not only returning words to their original meanings, but, in so doing, also purifying (in the moral sense) the ways in which those words are being used.

Thus, Mallarmé’s interest in mythology is complementary to his pedagogical work insofar as it foregrounds the role of the error in language.35 But, rather than conceiving of the error as that which derives from the process of translation (the non-sens that emerges from the passage between languages), the work on mythology reveals that linguistic aberrations are caused by the lack of translation itself. The space of translation, then, incurs the loss of both idiom and the senses, but also offers a potential therapy for these losses. In both cases, translation participates in a process of purification and reveals a conception of madness as that which breaks down conventional understanding of how language relates to the world, either because it renders us aware of the arbitrariness of idiom, or because it shows us the dangers inherent in linguistic amnesia.

Purifying the Words of the Tribe

Mallarmé draws a connection between purity and malady in describing the crisis of Tournon, during which he almost lost his mind and the meaning of familiar words: “Et maintenant, arrivé à la vision horrible d’une Œuvre pure, j’ai presque perdu la raison et le sens des paroles les plus familières” (Œuvres, 1:727). While debates continue among critics about what Mallarmé intends by the expression “pure work,”—and whether the fragmentary and posthumous Livre comes close to incarnating it—the “vision” of purity is characterized as being dangerous and destabilizing. The only protection against such danger is to create a buffer of mystery and, significantly, forgetfulness: “Pour moi, voici deux ans que j’ai commis le péché de voir le Rêve dans sa nudité idéale, tandis que je devais amonceler entre lui et moi un mystère de musique et d’oubli” (726).36 If the pure work can never (or, should never) be attained, the movement towards such purity becomes a key motivator in linguistic production and, thus, poetic creation. This involves a process of willful amnesia, one that is distinct from the involuntary forget-fulness to which Mallarmé calls attention in his discussion of mythology; it requires a careful consideration of the functions of obscurity and of memory (or the lack thereof) within the process of linguistic defamiliarization.

The concept of linguistic purification has come to be seen as one of the tenets of Mallarmé’s poetic program—famously expressed by the line “donner un sens plus pur aux mots de la tribu”—and a contributing factor to the perception of Mallarmé’s writing as intractably hermetic. Significantly, this line appears in Mallarmé’s eulogy of Poe, “Le Tombeau d’Edgar Poe,” which is placed in the opening to Mallarmé’s book of [End Page 154] translations of Poe’s poems—thereby highlighting a connection between purification and translation (Œuvres 2:727). From this famous line emerges a characterization of linguistic purification that ties it to hermeneutics, by way of a reflection on obscurity. This emerges from a series of observations. Firstly, the purification that Mallarmé refers to hinges upon a relationship between the act of purifying and the act of giving—as we can see by looking at the numerous translations or versions that this line has given rise to, from T. S. Eliot’s rendition, “To purify the dialect of the tribe,” to MacIntyre’s “endow / with a sense more pure the words of the tribe,” Mary Ann Caws’s “Give a purer sense to the words of the tribe,” and Weinfield’s “Bestow a purer sense on the language of the horde.”37

Secondly, this reveals that at the heart of the poetic work lies a tension between excess and lack. On the one hand, the poet adds something to what already exists (an increase that is present in the use of the comparative of superiority “plus pur”), but on the other, the poet takes away the impurities that obfuscate the sense of words (purification is the act of extraction and removal intended to liberate something from extraneous matter). The role of the poet, then, comprises both giving and taking, and the “purity” of the pure work must also be an act of obfuscation induced by accumulation (visible in Mallarmé’s use of the term “amonceler”). Poetic purity thus appears to be inherently related to a form of obscurity that transforms the familiar words of the tribe.

The obscurity that derives from this tension between excess and lack defines a particular relationship to meaning. This is exemplified in Mallarmé’s most famously hermetic poem, “Sonnet en –yx,” which, significantly, begins with a reference to purity: “Ses purs ongles très haut dédiant leur onyx” (Œuvres 1:37). In this poem, the movement of purification revolves around the word “ptyx,” an absent center defined as “Aboli bibelot d’inanité sonore.” This definition creates a mirroring structure, in which the terms “aboli” and “bibelot” reflect and refract each other, inducing an internal breakage to the individual words and reducing them to their babbling syllables—“a,” “bo,” “li” / “bi,” “be,” “lo.” As such, the obscurity of the poem derives from the process of purification itself, which progressively introduces a rift between words and their meanings.38 What appears to be a loss of meaning, here, coincides with the presumed lack of meaning inherent in the word “ptyx”—a word that Mallarmé intentionally chose because of the fact that it did not exist in any known language (although it has been found in Greek, as Doric dialect for the word “fold”).39

However, despite the fact that absence is arguably at the heart of the poem, to the extent that it reflects on the negating function of language whereby words imply the disappearance of the object to which they refer, Mallarmé’s practice of purification does not lead to an absence of meaning.40 Rather, Mallarmé’s choice of the word “ptyx” stems from the desire that it can potentially evoke meaning by administering what he calls a “dose” of poetry: “le sens s’il y en a un (mais je me consolerais du contraire grâce à la dose de poésie qu’il renferme, ce me semble) est évoqué par un mirage interne des mots mêmes” (Œuvres, 1:731). Meaning derives from a mirage, which is both a reflection and an illusion, and which requires breaking up language. More significantly, it involves the therapeutic act of giving, implied in the term “dose,” typically used in the [End Page 155] medical context. Poetry would then involve a process of purification intended to compensate for (or, in keeping with the medical vocabulary, cure) the default of language.

This process of purification is precisely that Benjaminian aspiration towards pure language, as the end result of translation—the production of pure form, or pure sound, where sense is not evacuated but, rather, is “touched by language only the way an aeolian harp is touched by the wind” (“The Task,” 23). Because Mallarmé’s “ptyx” raises the question of whether it belongs to a foreign language per se, our understanding of it cannot involve a process of translation; rather, the presence of the word on the page opens up a space between languages in which they come into contact with one another in the search for familial ties that might bestow upon it a sense of familiarity.

Yet, such a sense of familiarity is never attained: the “ptyx” sets up a trap that exposes our unthinking urge to assign meaning to the unfamiliar word. As such, the “ptyx” generates a process of purification similar to the one we saw both in Les Mots anglais and in the manuals of translation. Whereas in the pedagogical works purification of one language is achieved by passing through another, with the end result of creating an interlingual space that deconstructs the binary between the two languages, in the poetry we take one step further: by starting from a place where there is no known language, we come to experience that interlingual space as fundamentally non-linguistic, as it redefines our very understanding of how language works.

The practice of translation, then, brings together the “inanités” of Mallarmé’s teaching and the “productions insensées” of his poetry (via a reference to the “inanité sonore” of the “ptyx”) because, in pushing against the limits of languages, it goes beyond the question of how sense is created in language and of what happens if sense is deemed to be absent. Mallarmé’s practice of linguistic manipulation explores the almost-areas where meaning is created, deferred, and potentially lost, with a view to exposing our ingrained need for interpretation and its reliance on language.41 We come to realize that these almost-areas, which we typically reject as detrimental to the process of understanding, are in fact fundamental to that very process. But not because they represent an empty center that needs to be filled—this would correspond to attempts to find a meaning for the elusive “ptyx” and, thereby, render it familiar; rather, these moments reveal to us the significance of that very emptiness and the extent to which that emptiness underpins the experience of language as such.

We can now clearly see the benefit of considering both sides of Mallarmé’s work together. Overcoming the division both in Mallarmé’s production and in the reception of his work encourages us to interrogate the relationship between our need for interpretation (manifest, for instance, in our tendency to want to assign meaning to the “ptyx”) and how we handle that which resists interpretation (visible in characterizations of Mallarmé’s work as hermetic or in dismissals of it as “mad”). Such an interrogation is important because it asks how the moments of interruption or resistance to our understanding shape the way we conceive of the process of understanding itself. If reading Mallarmé may indeed make us feel like we are losing our senses and the sense of familiar words, thereby maddening us in the process, it also allows us to understand why we are experiencing such feelings: Mallarmé’s work is not challenging us to find meaning [End Page 156] where meaning is ostensibly lacking; rather, it draws our attention to our pursuit of meaning, by showing us what happens when we suspend the hermeneutic gesture.

Alexandra Lukes

Alexandra Lukes is Assistant Professor of French and Translation Studies at Trinity College Dublin. Her work focuses on the relationship between literature and madness, on the concept of untranslatability, and on the theoretical implications of translating nonsense, both as a literary genre and as a marginal discourse. She has published on Antonin Artaud’s translations of Lewis Carroll and is currently editing a collection of essays, entitled Nonsense, Madness, and the Limits of Translation (forthcoming in Translation Studies).


1. For the most complete discussion of Mallarmé’s teaching career, see Bertrand Marchal’s essay “Mallarmé professeur d’anglais,” in Mallarmé et l’anglais récréatif: le poète pédagogue (Paris: Cohen and Cohen, 2014), 15–59. See also Charles Chassé, “Mallarmé Universitaire,” Mercure de France, October 1, 1912, 449–64; Paul-Gabriel Laserstein, “Stéphane Mallarmé, professeur d’anglais,” Les Langues Modernes 43 (1949): 25–46; Austin Gill, “Mallarmé fonctionnaire (I),” Revue d’histoire littéraire de la France, 68, no. 1 (1968): 2–25 and “Mallarmé fonctionnaire (II),” Revue d’histoire littéraire de la France, 68, no. 2 (1968): 253–84; Henri Mondor, Vie de Mallarmé (Paris: Gallimard, 1941).

2. Within the Dossier of Les Mots anglais, we also find Ce que c’est que l’anglais, a manual on the origins and formation of English, and a series of cards Mallarmé designed for teaching English, L’Anglais récréatif ou Boîte pour apprendre l’Anglais en jouant et seul. In addition to Les Mots anglais and its Dossier, the section entitled “Ouvrages pédagogiques” of the 2003 Pléiade edition of Mallarmé’s complete works comprises an anthology of English literature, Beautés de l’Anglais, and a translation of George William Cox’s study on mythology, Les Dieux antiques (which I will discuss later).

3. For comments made by Mallarmé’s contemporaries on the poet’s presumed madness, see Pascal Durand, “La folie Mallarmé,” Le Courrier du Centre International d’Etudes Poétiques 225 (2000): 7–27.

4. Jacques Scherer writes: “cette quasi-langue, créée par Mallarmé et qu’on appellera, si l’on veut, le mallarméen, peut être objectivement décrite par une grammaire, qui toutefois n’aura de réalité que par référence à la parole du poète en ce qu’elle a d’unique” (Grammaire de Mallarmé [Paris: A. G. Nizet, 1977], 8).

5. Shoshana Felman defines madness precisely as that which resists interpretation: “madness . . . can be defined as nothing other than an irreducible resistance to interpretation” (Writing and Madness [Literature/Philosophy/Psychoanalysis], trans. Martha Noel Evans and Shoshana Felman [Stanford: Stanford University Press, 2003], 254). See also Maurice Blanchot’s discussion of the connection between obscurity and interpretation in “La poésie de Mallarmé est-elle obscure?,” in Faux pas (Paris: Gallimard, 1943), 126–31.

6. Daniel Tiffany comments on this connection: “It appears, however, to have escaped the notice of either reviewer that there might be an instructive correspondence between the ‘stupidities’ Mallarmé dictated to his class from Mother Goose and the ‘senseless productions’ of his own poetry” (Infidel Poetics: Riddles, Nightlife, Substance [Chicago: University of Chicago Press, 2009], 222).

7. This encompasses everything that does not belong to the poetic canon, such as Mallarmé’s pedagogical writings, his translations (both literary and pedagogical), his writings on fashion, the notes he sent to friends and family on varied surfaces (bottles of Calvados, fans, Easter eggs). See Marian Zwerling Sugano, The Poetics of the Occasion: Mallarmé and the Poetry of Circumstance (Stanford: Stanford University Press, 1992); Hélène Stafford, Mallarmé and the Poetics of Everyday Life: A Study of the Concept of the Ordinary in His Verse and Prose (Amsterdam: Rodopi, 2000); and Roger Pearson’s Mallarmé and Circumstance: The Translation of Silence (Oxford: Oxford University Press, 2004).

8. Stéphane Mallarmé, Œuvres complètes, ed. Bertrand Marchal, vol. 1, Bibliothèque de la Pléiade (Paris: Gallimard, 2003), 789.

9. We should bear in mind that Mallarmé’s comments on his pedagogical work were made in the context of his autobiographical letter to Verlaine, a carefully crafted piece that was intended to constitute his literary legacy. This suggests that there might be more to these works than it would initially seem and that Mallarmé’s dismissal of them cannot be taken at face value.

10. Stéphane Mallarmé, Œuvres complètes, ed. Bertrand Marchal, vol. 2, Bibliothèque de la Pléiade (Paris: Gallimard, 2003), 208.

11. “[M]on sens regrette que le discours défaille à exprimer les objets par des touches y répondant en coloris ou en allure. . . . A côté d’ombre, opaque, ténèbres se fonce peu; quelle déception, devant la perversité conférant à jour comme à nuit, contradictoirement, des timbres obscur ici, là clair” (Mallarmé, Œuvres, 2:208). [End Page 157]

12. It is important to note that Mallarmé also produced a fair number of literary translations—notably of Poe, Whistler, Tennyson, among others—which offer an equally fertile ground for examining Mallarmé’s practice of linguistic purification. However, a full discussion of these translations lies outside the scope of this article, as does an in-depth analysis of the differences between pedagogical and literary translation. For the sake of argument, I am here focusing on the former.

13. The composition of the pedagogical works can be dated to the period 1870–1879; the first draft of the “Sonnet en –yx” began in 1868 (although the final version was not published until 1887); the “Tombeau d’Edgar Poe” appeared in 1877, but the project of translating Poe began as early as 1862.

14. “Retrouvant en face d’un livre toute ma pensée, je m’étais initié à des études (de linguistique), mon refuge au cas écheant” (Mallarmé, Œuvres, 1:749); “Pour ne faire qu’un effort du tout, j’ai choisi des sujets de linguistique, espérant, du reste, que cet effort spécial ne serait pas sans influence sur tout l’appareil du langage à qui semble en vouloir principalement ma maladie nerveuse” (Mallarmé, Œuvres, 1:751).

15. These idiolectal strategies are referred to as “le procédé.” For an illustration of the uses of this term in relation to logophilia, see Michel Foucault’s Sept propos sur le septième ange (Paris: Fata Morgana, 1986) and Gilles Deleuze’s “Schizologie,” his introduction to Louis Wolfson’s Le Schizo et les langues (Paris: Gallimard, 1970), 5–23.

16. Michel Pierssens, La Tour de Babil. La fiction du signe (Paris: Éditions de Minuit, 1976), 58.

17. In “Crise de vers” Mallarmé defines “le vers” as that which “philosophiquement rémunère le défaut des langues, complément supérieur” (Œuvres, 2:208).

18. Stéphane Mallarmé, Thèmes anglais pour toutes les grammaires (Paris: Gallimard, 1937), 25–26, emphasis in the original.

19. Mallarmé’s use of the term “calque” here is significant, and it requires additional examination in relation to the use he makes of the same term to talk about his translations of Poe: “voici un calque se hasarder sans prétention que rendre quelques-uns des effets de sonorité extraordinaire de la musique originelle, et ici et là peut-être, le sentiment même” (Œuvres, 2:771). If, in the case of the pedagogical works, the calque of translation functions through the visual route (as a translucent copy superimposed on an original), in the literary translation the calque works through sound (as an impressionistic rendition of the sonorous effects).

20. Carl Paul Barbier, in his introduction to Mallarmé’s collection of nursery rhymes, comments on the book of Thèmes: “Mais quels thèmes! Certains sont à dormir debout, d’autres contiennent de véritables énigmes qui doivent être expliquées avant qu’on puisse traduire. . . . On comprend aisément que Leroy ait refusé de publier un système hermétique et stérile qui préconise l’emploi de soi-disant thèmes pour aboutir à des proverbes parfois désuets” (introduction to Stéphane Mallarmé, Recueil de “Nursery Rhymes,” ed. Carl Paul Barbier [Paris: Gallimard, 1964], 19).

21. Mallarmé here seems not to have noticed that the last line was missing a syllable; this casts doubt on his knowledge of English prosody, unless he was pronouncing “scratched” with two syllables.

22. Antoine Compagnon, “Somebody Will Always Get It,” in “Nonsense: a special supplement in conjunction with the Whitney Museum of American Art,” ed. David B. Allison, John G. Hanhardt, Mark Roberts, and Allen S. Weiss, special supplement, Art & Text 37 (1990): 71–73, 71.

23. Marina Warner, “‘Who can shave an egg?’: Beckett, Mallarmé, and Foreign Tongues,” Raritan 27, no. 4 (2008): 62–89, 72. The term “melancholy” is worth further exploration, not only because of its historical relation to madness (see Michel Foucault’s discussion in Histoire de la folie à l’âge classique [Paris: Gallimard, 1972], 281–96), but also because it plays a significant role in Benjamin’s reflection on language and translation; see Ilit Ferber’s Philosophy and Melancholy: Benjamin’s Early Reflections on Theatre and Language (Stanford: Stanford University Press, 2013).

24. Translation is an intralinguistic activity that, Paul de Man notes, is not concerned with extralinguistic meaning: it relates to “what in the original belongs to language, and not to meaning as an extralinguistic correlate susceptible of paraphrase and imitation” (The Resistance to Theory [Minneapolis: University of Minnesota Press, 1986], 84).

25. Walter Benjamin, “The Task of the Translator,” trans. Harry Zohn, in The Translation Studies Reader, ed. Lawrence Venuti (New York: Routledge, 2000), 15–25, 23. [End Page 158]

26. Jean-Jacques Lecercle notes that the results of failings in translation can be both entertaining and inventive: “the slip of the tongue is a form of wit, the mistranslation may end up as a successful metaphor” (“Translate It, Translate It Not,” Translation Studies 1, no. 1 [2008]: 90–102, 97). He cites the concept of “parapoetics” developed by Clark Lunberry, an American teacher of English in Japan, to account for the “poetic force conveyed by the broken English of his Japanese students” (97). Lunberry talks of “error enjoyment” rather than error analysis in “Broken English: Deviant Language and the Para-Poetic,” Kyoto Journal 29 (1995): 90–99, 97.

27. Jean-Michel Rabaté, The Ghosts of Modernity (Gainesville: University Press of Florida, 1996), 115.

28. Mallarmé describes himself as “profondément syntaxier,” defending his use of syntax in response to Proust’s denunciation of the Symbolists’s obscurity. In “Le Mystère dans les lettres,” he writes: “Quel pivot, j’entends, dans ces contrastes, à l’intelligibilité? Il faut une garantie—la Syntaxe” (Œuvres, 2:232–33). Barbara Johnson notes that the guarantee that Mallarmé seems to be proposing here is not one of intelligibility itself, but of the “pivoting” of intelligibility—what she calls “pivotal intelligibility” (The Critical Difference: Essays in the Contemporary Rhetoric of Reading [Baltimore: Johns Hopkins University Press, 1980], 67).

29. Here is the whole quotation: “Les langues imparfaites en cela que plusieurs, manque la suprême: penser étant écrire sans accessoires, ni chuchotement mais tacite encore l’immortelle parole, la diversité, sur terre, des idiomes empêche personne de proférer les mots qui, sinon se trouveraient, par une frappe unique, elle-même matériellement la vérité” (Mallarmé, Œuvres, 2:208; quoted in Benjamin, “The Task of the Translator,” 20, emphasis in Benjamin).

30. It is important to note that “translation” here follows Benjamin’s usage, which does not seem to distinguish between the question of translating a language and that of translating a text. Samuel Weber explains this in the following terms: “what is at stake here is a relationship not between individual works so much as between singularly dividual languages. In short, translatability defines language as the medium of singularly dividual events, rather than of universally meaningful works” (Benjamin’s -abilities [Cambridge, MA: Harvard University Press, 2008], 69).

31. “Naturel, factice et naturel à la fois, factice, il est tout” (Mallarmé, Œuvres, 2:1093).

32. “Qui n’éprouverait autre chose qu’un charme délicat à proférer sciemment, au cours d’une récitation à haute voix ou d’une conversation en Anglais, des paroles séparées de lui par un nombre de siècles important, aurait déjà tiré quelque bénéfice de la lecture de la première moitié de chaque Table” (Mallarmé, Œuvres, 2:1057).

33. Mallarmé (via Müller) ascribes this madness to language: “Il ne peut y avoir d’erreur plus grande que de croire à des nations entières soudain saisies d’une folie étrange qui les pousse à inventer toutes sortes de contes ridicules et tristes, et que chaque nation eut, à son tour, son heure de cette démence” (Œuvres, 2:1459). Mallarmé cites the myth of Oedipus to illustrate his point: “Mais quand les Grecs eurent oublié ce que signifiait le nom d’Œdipe, ils dirent de ce personnage qu’après avoir frappé le Sphinx, il se maria avec sa propre mère, et que des maux terribles s’ensuivirent” (1458–59).

34. “[C]omme le temps marcha, et que les peuples se séparèrent, le vieux sens s’oblitéra, totalement ou partiellement” (Mallarmé, Œuvres, 2:1456).

35. In a study on Mallarmé’s treatment of myths, Pierre Renauld draws a connection between the madness of language, the disease of myths, and the errors made by children falling into the traps of grammar: “[les mythes] ce sont spécifiquement des erreurs, les erreurs d’un âge postérieur, d’un âge de décadence—quelque chose comme les contresens et les divagations d’écoliers tombés dans un piège grammatical” (“Mallarmé et le mythe,” Revue d’Histoire Littéraire de la France [1973]: 48–68, 51).

36. See Blanchot’s comments on the relationship between language and forgetting in “L’oubli,” in L’Entretien infini (Paris: Gallimard, 1969), 289–91.

37. T. S. Eliot, “Little Gidding,” in The Complete Poems and Plays: 1909–1950, 138–48 (London: Faber and Faber, 1969), 194; Stéphane Mallarmé, “Le Tombeau d’Edgar Poe,” in Selected Poems, trans. C. F. MacIntyre (1957; rpt., Berkeley: University of California Press, 1971), 89; Stéphane Mallarmé, “The Tomb of Edgar Poe,” in Selected Poetry and Prose, ed. Mary Ann Caws (New York: New Directions, 1982), 51; Stéphane Mallarmé, “The Tomb of Edgar Poe,” in Collected Poems, trans. Henry [End Page 159] Weinfield (Berkeley: University of California Press, 1994), 71. Perhaps the insistence on the role of giving may be an overinterpretation, insofar as the French expression implies a transformation of the object in question (the meaning of words) rather than the act of endowing someone with something, an act that does not change the object of the donation itself. However, underscoring the role of the gift in this definition of the poet’s work relates to Mallarmé’s own poetic practice as a giver of poems, the “gift poems” or the Vers de circonstance (writings on fans, boxes of candied fruits, small stones, empty jugs of Calvados, and Easter eggs) that constitute the “side” of Mallarmé’s production to which the pedagogical works belong.

38. As Patrick McGuinness explains, “the ‘Nothing’ swallows up and brings forth the ‘Thing’”: “In terms of sound therefore ‘Aboli’ gives birth to, produces, ‘bibelot,’ but in terms of meaning (as derived from word-order) it cancels it pre-emptively in a process of emptying (‘inanité’) which takes place with a particularly rich (‘sonore’) sound-pattern reversal” (“‘Beaucoup de bruit pour rien’: Mallarmé’s ptyx and the Symbolist ‘bric-à-brac,’” Romanic Review 86, no. 1 [1995]: 103–13, 108).

39. McGuinness writes: “It has to be pointed out that “ptyx,” despite not existing in French, occurs in Greek dictionaries as Doric dialect for “fold,” and it is not a hapax legomenon (a word that occurs only once in extant literature) either in Greek or strictly speaking in French (since it occurs in Victor Hugo’s poem). In Mallarmean exegeses it has been taken to mean many things, among them seashell, writing tablet, derivative of the English word “pyx” etc. and the above ideas do not seek to contradict these explanations. We are concerned here with detecting the particular pressures that necessitate its appearance in the sonnet” (“‘Beaucoup de bruit pour rien,’” 107).

40. In relation to the negating function of language, Mallarmé writes in “Crise de vers”: “Je dis: une fleur! et, hors de l’oubli où ma voix relègue aucun contour, en tant que quelque chose d’autre que les calices sus, musicalement se lève, idée même et suave, l’absente de tous bouquets” (Œuvres, 2:213).

41. Roger Pearson remarks on this “nearness”: “[Mallarmé] knew that what he wrote was nearly nonsense, and that in that nearness lay at once its point and his integrity” (“Review of Mireille Ruppli and Sylvie Thorel-Cailleteau, Mallarmé: la grammaire & le grimoire,” French Studies 61, no. 2 [2007]: 234–35, 235). [End Page 160]