A synecdoche test

Peter Steiner

pp. 117-204

The three metaphors of Russian Formalist theory, decisive as they were in their proponents’ thinking, still do not account for perhaps the most fundamental Formalist conception: the notion of language as the material of poetry. “Insofar as the material of poetry is the word,” Žirmunskij wrote, “the classification of verbal phenomena provided by linguistics should be the basis for asystematically constructed poetics. Because the artistic goal transforms each of these phenomena into a poetic device, every chapter of theoretical poetics should correspond to a chapter from thescience of language” (Žirmunskij 1928, p. 39).Language thus generated a fourth Formalist model. But the trope underlying it was not a metaphor, as in the cases of the mechanistic, morphological, and systemic models. These posited a similarity between the literary work and a machine, organism, and hierarchical system, respectively, but the model described by Žirmunskij is a synecdoche, a pars pro toto relationship. It substitutes language—the material of verbal art—for art itself, and linguistics—the science of language—for literary studies.

The linguistic model, as this theoretical synecdoche might be termed, has its roots in the early Formalist preoccupation with “poetic language.” The importance of this notion for the entire Formalist enterprise cannot be overstated. Pavel Medvedev, a Marxist critic of the movement, quite correctly claimed that the “hypothesis of the distinctness of poetic language is the basis upon which the entire Russian Formalist method is built” (Medvedev 1928, p. 111).The Formalists themselves were aware of the privileged status of this concept. Indignant at the label of “Formalism” foisted upon them, these youngliterary scholars proudly presented themselves as students of poetic language and even as linguists. The names of their two original groups, the Society for the Study of Poetic Language and the Moscow Linguistic Circle, and the title of their first two collective publications, Studies in the Theory of Poetic Language, clearly indicate the image they strove to project at the inception of the movement (OPOJAZ 1916 & 1917).

The acceptance of any concept among the whole Formalist membership was never a simple matter and “poetic language” was no exception. Because of the inherent heterogeneity of the movement and the fluidity of its concepts over time, the Formalists never reached a general definition of either poetic language or the linguistic frame of reference for its description.Moreover, as Formalist theorizing unfolded, the fortunes of the linguistic model in general and the notion of poetic language in particular fluctuated widely. OPOJAZ’s initial infatuation with the two gave way to a sharp backlash in the early twenties. But just as the stock of the linguistic model was dipping in Petersburg, it was rising inMoscow. Obviously, the idea of a single theoretical synecdoche in Russian Formalism is an oversimplification. In fact, this fourth model encompasses several distinct theories, each of whichtreated literature as the art of language and used methods borrowed from linguistics. In the discussion that follows, I shall attempt to describe some of the most important currents among them.


The meaningless pursuit of meaning by our writers is quite astonishing.

Wishing to portray the incomprehensibility, the illogicality of life and its terror or mystery, they resort (as ever, as always!) to “clear, precise” common language. […]

We were the first to say that for portraying the new and the future completely new words and new combinations are necessary.

This striking newness will come through the combination of words according to their own immanent laws revealed to the poet and not according to the rules of logic or grammar, as has been the case before us.

— Aleksej Kručënych, “The New Paths of the Word”

“Poetic language” was already a loaded term by the time it entered Formalist discourse. Aleksandr Potebnja (1835-1891), the heir to the tradition of Humboldtian linguistics, was the first to introduce the distinction between poetic and prosaic language into Russian philology (Cf., for example, Jakubinskij 1923, pp. 113-14; Erlich 1981, pp. 23-26; Stempel 1972, p. xiv). The Formalists’ attitude toward their “precursor” was rather ambivalent, however. Their willingness to borrow from him implied a respect extended to no other nineteenth-century Russian philologist but Veselovskij. Still they criticized Potebnja’s work violently in order to differentiate themselves from him, and especially from his numerous epigones. The Symbolist literary critic D. Filosofov, the first reviewer of OPOJAZ’s 1916collective publication, described this dialectic relationship: “All the contributors to this new collection are in a sense Potebnja’s pupils. They know him by heart; they live off the late scholar’s ideas. But they are not arrested in them. They reexamine the mysterious correlations of sound and representation and in doing so they focus their entire attention on sound. But in the end they make clear that sound, even ‘non-articulated’ sound, generates representation. They speak of the magic of sound, the magic of words” (Filosofov 1916, p. 3).

The Formalist departure from Potebnja should not be viewed merely as a struggle for recognition. Though a powerful and prolific thinker, Potebnja was often more suggestive than clear, and in his elaborate handling of topics he often multiplied the definitions of even his most cherished concepts. The opposition between prosaic and poetic language is a case in point. Sometimes it is presented as a simple formal dichotomy between prose and poetry, and at others, as a psychological antinomy between prosaic and poetic thought. In the latter case, the presence of a mental image is the essential feature of the poetic, and “poetic thinking” is defined as “thinking in which the image is important” (Potebnja 1905, p. 98). When the distinction is made on formal grounds, the differential feature is a matter of function; prose is thus “language oriented solely toward practical aims or serving as an expression of scholarship” (ibid., p. 102).

Despite these inconsistencies, Potebnja’s poetics did rest on certain basic assumptions. The word and/or poetic work consists of three parts: the outer form (the perceptible aspect), the meaning (the intelligible aspect), and the inner form or representation (the tropological link between the two). The crucial member of this triad is inner form, a notion heavily dependent upon certain ideas from psychology. In agreement with the atomistic theory of association so popular in his time, Potebnja treated mental life as an aggregate of simple sensory elements. For him, the perceptual identity of an object (what he termed its “image”) was guaranteed by the persistence of a simple characteristic through whatevercontextual modification the object undergoes. Language follows this model when a distinctive characteristic motivates an object’s designation, that is, when the object is named according to this feature. But though thought and language coincide here, this is not a case of inner form proper. Only a single sensory image has provided the link between the outer form and meaning, whereas the inner form is an umbrella for a multitude of such images: in Potebnja’s words, “it is not an image of an object but an image of an image, that is, a representation” (Potebnja 1913, p. 117).

As a metaconcept, inner form is endowed with the power that a single image lacks: it links outer forms and meanings that were originally connected to diverse sensory images. In this respect, the inner form of language is the crossroads of the old and the new. As the “nearest etymological meaning of a word,” it stands for the linguistic past, but as the tertium comparationis that generates the figurative transformations of a word, it is the agent of the future (ibid., p. 146). Because of this creative potential, the inner linguistic form became the central category of Potebnja’s poetics. Without denying salience to the other two components of the word, Potebnja found the eidos of poetic language in its polysemy, the capacity of its inner form to evoke multiple meanings. Stated in quasi-mathematical terms, “the general formula of poetry (or art) is: ‘A (image) < X (meaning)’: that is, there is always an inequality between image and meaning because A is smaller than X. To establish an equality between A and X would destroy poeticity; i.e., it would either turn the image into a prosaic signification of a particular phenomenon devoid of relation to anything else or it would turn the image into a scientific fact and the meaning into a law” (Potebnja 1905, p. 100).

The Formalist redefinition of poetic language represents a considerable departure from Potebnja’s basic position. This departure, however, was not solely motivated by theoretical concerns but also by the current poetic practice. Some of the early Formalists entered the Russian intellectual scene not as disinterested observers or commentators, but as proponents and interpreters of Futurism, the most flamboyant artistic movement of their generation. The rise of Futurism in the earlyteens was directlylinked to the decline of another movement that had dominated Russian letters for nearly two decades—Symbolism. The great poet-theoreticians of this movement had exploited Potebnja’s philology as the theoretical springboard for their own poetics. In Roman Jakobson’s words, “the Symbolists canonized Potebnja” (Jakobson 1922, p. 223). Thus, the futurist onslaught against Symbolism involved at the same time a “de-canonization” of Potebnja, a search for new theoretical foundations upon which to construct their poetics.

Of the various groups in Russia calling themselves Futurists, the most iconoclastic was known as Hylaea and it was with this group that the Formalists were most closely allied, both personally and interms of a shared artistic sensibility. Hylaea’s membership included the Burljuk brothers, Chlebnikov, Kručënych, and Majakovskij. In the unceasing stream of public appearances, manifestos, and joint publications, all orchestrated to épater les bourgeois, the Hylaeans declared the art of the past dead and presented themselves as the only true championsof the artistic future. Their incompatibility with Potebnja’s system is obvious at first glance. Disdaining the cognitive function of art (“thinking in images”), the Futurists insisted on the shock effect. Artworks, according to Chlebnikov and Kručënych, ought to be “as if written with difficultyand read with difficulty, less comfortable than blacked boots or a truck in a sitting room,” their language “resembling if anything a saw or a savage’s poisoned arrow” (in Markov 1967, p. 53-56). Against the historicism of Potebnja’s poetic word (that is, its etymological meaning), the Hylaeans’ manifesto, “A Slap in the Face of Public Taste,” proclaimed the poet’s right to an “insuperable hatred for the language that existed before him” (ibid., p. 51).They ridiculed the entire psychologistic bias of the previous poetics. Poetry, they insisted, is not a mirror of the soul but “the unfolding of the word as such.” Or in more epigrammatic form, “The work of art is the art of the word” (ibid., p. 59).

This conception of verbal art was obviously reflected in the earliest, mechanistic model. Keynotions suchas de-familiarization or the absolute split between art and byt are direct projections of Futurist poetics onto Formalist literary theory. The notion of poetic language was most profoundly influenced by the Futurist concept of zaum’. Coined by Kručënych, the term designated a special tonguethat defiedthe rules of common sense: transrationallanguage, zaum’ attacked the very heart of Potebnja’s aesthetic system—the identification of poeticity with the inner form of language, since this “ultimate” language of verbal art was without inner form. Its two main exponents among the Hylaeans, Kručënych and Chlebnikov, disagreed about which of the remaining parts of the verbal parcel—outer form or meaning—was instrumental in zaum’. As Vladimir Markov has observed, “for Kručënych [zaum’] was a free, but often emotionally expressive, combination of sound, devoid of full meaning; for Chlebnikov, it was basic meaning expressed in the purest and most direct way” (Markov 1968, p. 303).

If, as in traditional aesthetic discourse, the term Formalism applies to theories asserting the primacy of artistic form over content, Kručënych’s zaum’ would seem to be distinctly Formalist. “A new content,” he proclaimed, “is born only when new expressive devices, new forms, are achieved. Once there is a new form, the new content follows. Thus, the form determines the content” (in Markov 1967, p. 72). Consequently, it was not the ideas or things presented by the literary work that were important, but the mechanism of this presentation itself. Because this mechanism is above all linguistic, Kručënych spoke of two types of language: rational common language governed by extra-linguistic requirements, a vehicle of meaning; and self-sufficient transrational language governed by its own rules, “whose words do not have a definite meaning” (Kručënych 1973, p. 55).

This indefiniteness of meaning in zaum’ is quite different from Potebnja’s poetic polysemy. The quantitative imbalance occurs not between the inner form of the word and its meaning, but between the meaning and the word as such, that is, its outer form. Moreover, transrational language reverses the ratio of Potebnja’s formula: in zaum’ sound always is greater than meaning. Kručënychwrote, “We declared in art: the word is broader than its meaning. The word (and the sounds composing it) is not merely curtailedthought, not merely logic, but above all the transrational(its mystical and aesthetic components)” (in Markov1967, p. 66). Hence, transrational language is literally language that goes beyond reason, that addresses the non-rational human faculties. To achieve this objective, the poet is free to dissolve language into elements that lack any logical meaning, or to combinethese elements into nonsensicalneologisms. The poet can also emulate the types of zaum’ existing outside verbal art. One especially favored by Kručënych was the glossolalia of religious sectarians speaking the “language of the holy spirit.” There was also children’s language18 or the sound patterns of foreign languages unknown to the poet.19

If Kručënych’s zaum’ privileged the outer form of the word, Chlebnikov’s privileged the meaning. To those familiar with the impenetrable hermeticismof Chlebnikov’s texts, this might come as a surprise. But understandingwas never an issue for him. “Verses,” he wrote, “may be comprehensible or incomprehensible, but they ought to be good, ought to be genuine [istovennyj]”(Chlebnikov 1933, p. 226). To be transrational for Chlebnikov meant to go beyond ordinary reason, but only to express the higher reason that he believed language inevitably embodied. Potebnja’s notion of inner form was thus suspect, for it posited merely a figurative link between sound and meaning. Chlebnikov’s zaum’ , in contrast, was a quest for the direct, unmediated meaning of sound.

In his study of Chlebnikovian zaum’, Ronald Vroon has identified four such linguistic structures: the languages of the “stars,” “Gods,” and “birds,” and “sound-painting.” He maintains that each of these tackles the issue of pure meaning in a different way (Vroon 1979, esp. pp. 30-34, 266-99). For instance, the “language of the stars”—Chlebnikov’s favorite—is based on the same kind of argument for the natural origin of names that Plato credits to Cratylus. It assigns a distinct spatiogeometric meaning to virtually all Russian consonants. This zaum’, Chlebnikov believed, was not an arbitrary construction but a faithful reconstruction of the original language of mankind, of which our present-day tongues are mere shadows.

In rough terms, the rift between Kručënych and Chlebnikov over zaum’ corresponds to the conflicting theories of poetic language in the Formalist movement. OPOJAZ’s early concern for poetic sound and its emotive qualities betrays Kručënych’s influence, whereas the Moscow Linguistic Circle’s insistence on the meaningfulness of linguistic sound reflects the logocentrism of Chlebnikov’s zaum’. We shall now look more closely at the earlier of these two tendencies, that of OPOJAZ.

In 1916, Viktor Šklovskij undertook the first direct critique of Potebnja’s poetics. “Imagery, or symbolism,” he insisted, “is not what differentiates poetic from prosaic language. Poetic language differs from prosaic language in the perceptibility of its structure” (Šklovskij 1919c, p. 4). Šklovskij’s strategy here is quite obvious: he is revising Potebnja’s dichotomy between poetic and prosaic language according to the specifications of his mechanistic model. The special perceptibility of poetic language drains ourmental energy and de-familiarizes our perception of language in general. This essential feature of poetic language is explained by its artistic telos. “If we study poetic speech… we encounter the same symptom of the artistic everywhere: that it was created intentionally to de-automatize perception and that the author’s goal was to call attention to this: that it was made ‘artificially’ in such a way that perception lingers over it, thus reaching its greatest possible intensity and duration.” The direct opposite of this ‘‘hampered and tortuous” speech is automatized prosaic language. “Prose is normal speech: economical, easy, regular (the dea prosa is the goddess of regular, uncomplicated delivery)” (Šklovskij 1919a,pp. 112-13).

It is significant that while Šklovskij’s treatment of poetic language rejects Potebnja’s, it retains his fundamental dichotomy of poetic and prosaic language. Here we witness yet another example of the peculiar contradiction in the mechanistic model mentioned earlier: namely, its propensity for merging the most radical stance with a traditional conceptual framework. This marriage of the old and the new tends to generate problems. The opposition between poetry and prose would appear to coincide with Šklovskij’s distinction between art and byt. But if this were the case, poetic speech, with its patent goal of de-familiarization, would be the only discourse to use language purposively, and prosaic speech, as a phenomenon of byt, would be governed purely by causality. As Šklovskij himself showed, “economical, easy, and regular” speech might also be used for the sake of de-familiarization.This is the case of artistic prose that renders extra-literary reality strange in the process of its verbal representation. In this way, one is forced to speak of not two but three types of language: the poetic, which makes strange our perception of language itself, the prosaic-artistic, which does the same to the perception of reality, and the prosaic proper, that is, normal everyday language. Yet according to the logic of Šklovskij’s model, the first two types are clearly different from the third. Whereas the two differ in what they de-familiarize, they are united through their common artistic goal. Normal everyday language, in contrast, belongs to byt. This fact, however, is completely lost in the simple opposition of prose and poetry that Šklovskij inherited from Potebnja. Thus, it was necessary to readjust Potebnja’s original opposition in such a way that the line between literature and non-literaturewould be drawn more clearly.

This task was performed by another OPOJAZ member, the linguist Lev Jakubinskij, who was responsible for introducing the distinction between poetic and practical language into literary theory. In linguistics, Jakubinskij argued, the opposition between the teleological and the causal can be suspended, because every utterance, whether poetic or not, pursues some objective. From this perspective, language can be conceptualized as a means-end structure serving particular goals. This view of language is similar to the functional classification of linguistic sounds advanced by the Kazan’ School to which Jakubinskij’s teacher, Baudouin de Courtenay (1845-1929), belonged (Cf., for example, Jakobson 1971, p. 399 or 524). It is also parallel to the thesis propounded by one of Franz Brentano’s follower, the philosopher Anton Marty(1847-1914),concerning the teleological origin of language as a means of human communication. Jakubinskij, however, avoided the psychologism of Marty’s teleology, which treated intention in terms of a conscious subject. For Jakubinskij, it was not the subjective intentions of the speaker but the objective correlation of linguistic means and ends that distinguished poetic from practical language.

“Linguistic phenomena,” Jakubinskij argued, “should be classified, among other ways, from the standpoint of the goal forwhich the speaker exploits the verbal material in a given case. If he uses it for the purely practical goal of communication, we are dealing with the system of practical language, in which linguistic representations (sounds, morphemes, etc.) have no value in themselves but serve merely as a meansof communication. Other linguistic systems are conceivable (and exist) in which the practical goal retreats into the background and linguistic combinations acquire a value in themselves... I conditionally call this system verse [stichotvornyj] language” (Jakubinskij 1919b, p. 37).20

Jakubinskij’s distinction between language as a means of communication and language as a self-valuable end should remind us of Kručënych’s distinction between common language and zaum’. This parallel becomes even more pronounced when Jakubinskij goes on to discuss the difference between practical and poetic language in terms of the opposition between sound and meaning. “In practical language the semantic aspect of the word (its meaning) is more prominent than its sound aspect... details of pronunciation reach our consciousness only if they serve to differentiate the meaning of words... Thus, various considerations compel us to recognize that in practical language sounds do not attract our attention. It is the other way around in verse language. There, one can claim that sounds enter the bright field of consciousness and do attract our attention” (ibid., p. 38).

This foregrounding of sound profoundly affects the structure of poetic language. Kručënych’s statement that zaum’ combines words “according to their immanent law... and not according to the rules of logic and grammar” is relevant. Jakubinskij too claims that poetic and practical language are demarcated by antithetical combinatory laws. He states that the liquidconsonants (r, l) tend to cluster in poetic language, whereas in practical language they are almost always randomly dispersed. If in practical language adjacent syllables contain the same liquid, this consonant will either be dropped altogether in one of them or replaced by another liquid, for the “clustering of the same liquidsimpedes pronunciation (even causing stammering) and violates the usual tempo of speech, thus willy-nilly directing the attention of the speaker toward the phonic aspect of the utterance... [It] violates the automatism which is so essential to practical language” (Jakubinskij 1919c, p. 52). Poetic language, on the other hand, which aims at focusing attention on sounds themselves, not only tolerates the clustering of the same liquids but deliberately produces such clusters.

Jakubinskij’s equation of poetic language with zaum’ goes even further. In his 1921 essay, “Where Does Verse Come From?” he argues that the concern for the sound of an utterance to the neglect of its content links poetic language to other types of discourse that defy normal reason. For example, “first of all, [in] the dream… the association of words according to their sound may determine the dream content. Second, in mental illnesssome patients utter entire tirades that are relatively unconnected in their content (as they ought to be) yet obviously linked in their sound, and often in meter. Third, in states of ecstasy, for instance among religious sectarians,” utterances often contain “sound repetition and meter” (Jakubinskij 1921, p. 23). In a ratherstartling move, Jakubinskij invoked Freud’s authority to claim that verse as well as the other three kinds of abnormal language are in fact the first stage of infantile language emerging from the subconscious in moments of weakened rational control. Thus, he answers the question raised in the title of his article by claiming that “verse comes from infantile babble,” providing a psychoanalytic explanation for Kručënych’s transrational language.

Jakubinskij was not the only Formalist to conceive of poetic language as a particular manifestation of zaum’, though the others usually did not invoke a psychoanalytic frame of reference. Not surprisingly, Viktor Šklovskij was one of the most powerful voices advocating the exclusion of semantics from verbal art. “We must ask,” he wrote, “whether words have meaning even in language that is not overtly transrational but simply poetic, or whether this belief is a mere fiction—the result of our inattentiveness” (Šklovskij 1919b, p. 25). In a speech to the Futurists at The Stray Dog, a Petersburg cabaret, Šklovskij spoke of transrational experiments in terms borrowed from Potebnja and the Symbolists. He compared zaum’, for example, to the foreign languages used in liturgical services. “The religious poetry of almost all nations was written in such a semi-understandable language: Church Slavonic, Latin, Sumerian (which died out in the twentieth century B.C. but was used as a religious language until the third century), and the German of the Russian Pietists [štundisty].”21Later Šklovskij dropped such “metaphysical” explanations and preferred to speak instead of the “sweetness of verse on our lips,” the de-automatized movement of our speech organs producing unusual phonicpatterns (Šklovskij 1923c, p. 8). “Maybe,” he mused, “the greatest part of the pleasure caused by poetry lies in its articulatory aspect, in the peculiar dance of the speech organs” (Šklovskij 1919b, p.24).

The metaphor of dance employed by Šklovskij is telling. Once poetic language is purged of meaning, verbal art can quite conveniently be described in terms of another non-thematic art. Music—the art of pure sound—is an obvious parallel; that is, if literature is nothing but a striking organization of phonic material, the poetic text, is very much like a musical composition. Osip Brik, another contributor to the early OPOJAZ collections, declared that “poetic language is musical language” and attempted to describe a major principle of the phonic organization of verse that had so far escaped the attention of other investigators (Brik 1919, p. 62).

Brik proceeded from the same assumption as Jakubinskij, namely, that poetic utterances are composed according to certain combinatory rules that are phonic in nature. For Jakubinskij, this was the clustering of liquids, but Brik went beyond this in two respects. First of all, he did not stop with liquids, but included all the consonants. Second and more important, he was not interested merely in isolated consonantal patterns but in the reiteration of these patterns throughout the poetic text. Traditional literary studies, according to Brik, merely paid lip service to the phonic aspect of poetic language and recorded only the most obvious cases of speech sound repetition: rhyme,assonance, alliteration, onomatopoeia. But these are merely an “obvious manifestation, a special case of fundamental euphonic laws,” and there are other cases that follow these laws but remain unnoticed (ibid., p. 60). Brik’s essay studied one of these—the recurrence of consonantal patterns—as it appeared, for example, in this Puškin line:

Vezuvij zev otkryl…
(Vesuvius opened its gorge) (ibid., p. 80)

Brik termed this type of consonantal reiteration “sound repetition” and attempted not only to provide a typology of such repetitions but also to relate them to the overall outer form of the poetic text (verse, stanza, rhythm). Using literally hundreds of lines from Puškin and Lermontov Brik demonstrated that sound repetition permeates even the most canonical of Russian poetry.

Another contributor to the early OPOJAZ volumes,Boris Kušner, argued that the treatment of poetic language in terms of the other arts, for example, dance and music, is natural because their materials have something in common. They are temporal rather than spatial media. “But,” Kušner warns, “despite their shared sound material, one can speak of verse music only metonymically. Here the term music no longer signifies a given art but the basic material of its works—sound” (Kušner 1916, p. 43). This figure of speech is therefore not productive for poetics, Kušner argues, for musical and poetic sounds are incompatible phenomena. The former are tones (tonirujuščie zvuki), sounds correlated according to precise scales and intervals, whereas the latter are merely sonorous sounds (sonirujuščie zvuki) whose actual phonic values are largely arbitrary. Music and poetry can, however, be related metaphorically, through the similarity of their artistic forms, that is, the precisely calculated organization of sound material. These “sonorous chords”—the repetitions of particular sounds and their groups in a poetic work—are what Kušner sets out to study.

But how do Kušner’s “sonorous chords” differ from Brik’s “sound repetitions”? First of all, in the way they are described: Brik presents his repetitions as objective phonic structures, whereas Kušner is concerned with the constitution of the “chords” in the perceiver’s consciousness. Second, Brik’s treatment of the poetic sound stratum is quite atomistic: he deals with a couple of isolated lines each time. For Kušner, on the other hand, the “sonorous chords” are the property of an entire poem. Of all the factors that create a rhythmical impression on the perceiver, Kušner focuses on two: the articulation of speech into syllables, and the segmentation of the continuous utterance into verse lines. In this way each speech sound is assigned a precise place within a two-dimensional grid based on its position vis-à-vis the other syllables of the line and vis-à-vis corresponding syllables in other lines. The resulting grid ofversepositions accounts for the distribution of all speech sounds in the poetic text, thus enabling Kušnerto detectany patterns that they might form—the sonorous chords.

The Formalists discussed so far tackled the category of poetic language as a primarily phonic phenomenon. Their preoccupation with poetic sound was chiefly inspired by Kručënych’s concept of zaum’—language contemptuous of everyday rationality and semantics. It must be stressed that even though Kručënych scoffed at language that merely conveys thought, he conceived of transrational language as something more than mere sound. The unfolding of the “self-valuable word” was only one aspect of zaum’, for the destruction of syntax and grammar still served a particular objective. A normally structured utterance, Kručënych reasoned, contains a logical meaning that transmits thought into words. The deformed zaum’, on the other hand, lacks such a definite meaning, but precisely because of this its words can express directly the non-cognitive components of the poet’s consciousness. “The clear and decisive evidence for the fact that until now the word has been in shackles is its subordination to sense. Until now it has been maintained that ‘thought dictates laws to the word and not the other way around.’ We have pointed out this mistake and provided a free language, transrational and ecumenical. The path of previous artists led through thought to the word; ours leads through the word to direct apprehension” (in Markov 1967, pp. 65-66).

Kručënych, however, failed to explain in any cogent way either the mechanism for this immediate apprehension or its object. His point might be expected to carry rhetorical weight within a poetic manifesto but certainly not elsewhere. Yet it caught the fancy of the Formalists, who argued against Potebnja’s identification of poetic language with inner linguistic form. From their point of view, Kručënych’s zaum’ was the best evidence that verbal art can do quite well without any images. To sustain this argument they had to translate Kručënych’s statements about the direct expressivity of outer poetic form into more scholarly terms.

Here they could turn to a theory of another member of the Kazan’ School, Mikołaj Kruszewski(1851-1887). In studying the universal laws of association operating in language, Kruszewski had argued that “the coexistence of the two aspects of the word—its external appearance and its meaning—rests on an association based on contiguity which binds these two aspects ‘into aninseparable pair.’ But to our memory ‘such a binding seems weak and insufficient; it must be supported by an association to an-other word based on similarity.’” This dual linkage of each verbal unit is the engine that drives linguistic change. Kruszewski depicts the process of linguistic evolution as “an eternal antagonism between a progressive force determined by associations based on similarity and a conservative one determined by associations based on contiguity” (Jakobson 1971, pp. 436-37).

Kruszewski’s two types of association correspond in turn to two figures of speech: metaphorand metonymy. The ingenious Šklovskij used this tropological distinction in attacking Potebnja. He claimed that not only poetic but prosaic language might involve inner linguistic form, that is, the figurative transference of meaning. But it is necessary to distinguish between two different figures of speech: the “conservative” metonymy, based on contiguity, and the “progressive” metaphor, based on similarity. Given the bias of Šklovskian aesthetics toward novelty in art it is not surprising that he considered the metaphor as the only truly poetic trope. Metonymy is merely the “practical means of thinking, of conceptualizing objects” and as such it characterizes prosaic language, but metaphoris the “means for intensifying perception” and hence the essence of poetic language. To illustrate, calling someone “a hat” [šljapa] simply because he happens to wear one is to evoke a prosaic image-trope, whereas the same designation for a helpless, languid fellow would be a poeticfigure (Šklovskij1919a, p. 103).

Despite Šklovskij’s criticism, however, he was still operating with Potebnja’s concept of inner form. The metaphoric designation that he described involved a cognitive tertium comparationis— a mental construct linking the outer form of the word with its figurative meaning, as in “helplessness” and “sloppiness” in his “hat” example. But the Formalists inspired by Kručënych’s zaum’ were not much concerned with traditional poetic tropes. Rather, they looked for cases of what Roman Jakobson aptly termed “negative inner form,” that is, “words which so to speak seek their meaning,” or, put differently, words with a directly expressive outer form (Jakobson 1921a, p. 67).

One hypothesis about the immediate emotive value of poetic sound was enunciated by Lev Jakubinskij. He approvingly quoted the observation of the famous French Indo-Europeanist Antoine Meillet (1866-1936) that in “practical language there is no inner link between the sound of the words and their meanings. Their link is determined by an association based on contiguity and is factual, not natural” (Jakubinskij 1919a, p. 44). This is so because in practical language sounds merely serve to differentiate meaning. The foregrounding of sound that is proper to poetic language, however, changes the picture. In such language, “because our attention is attracted by sounds, an emotive attitude is aroused toward them. This circumstance,” Jakubinskij stressed, “is very important for determining the interrelations of the phonic and semantic aspects of speech in verse language” (ibid.). Here the two are linked by the relation of similarity. Jakubinskij’s notion of similarity is, however, somewhat different from that of Kruszewski: what is similar in poetic language is the emotive charge belonging to the phonic and semantic aspects of the word. “The emotions evoked by certain sounds and their combinations can take various courses: ‘pleasure-displeasure,’ ‘arousal-satisfaction,’‘tension-resolution.’ It is also absolutely clear that the emotions triggered by sounds should not take a course antithetical to the emotions triggered by the ‘content’ of the poem (and vice versa)... Thus, the poet selects sounds and combinations that emotionally correspond to images valued by him for some reason, or, vice versa, he selects images that emotionally correspond to sounds and combinations that are significant for some reason in the given circumstances” (ibid., p. 45).

In addition to the emotive charge of sounds, the similarity of the phonic and semantic aspects of poetic language is provided by what Jakubinskij called the “capacity of the speech organs for expressive movements” (ibid.). There is, he believed, a curious juncture of emotions and language in our facial expressions. The movement of our facial muscles can be caused on the one hand by our emotions, and on the other, by the articulation of speech sounds. In practical language, where the phonè is just a means, speech sounds can be modified to accommodate the emotions. This is impossible, however, in language dominated by sound. Thus, in verbal art the poet is forced to “select words whose sounds are pronounced through movements of the speech organs corresponding roughly to given expressive movements… Broadly speaking, if the poet experiences emotions pertaining to a smile (a stretching of the lips sideways), then he naturally will avoid sounds articulated by pushing the lips forward (e.g., u, o) (ibid., p. 48).

Another theory of the direct expressiveness of linguistic sound was formulated by a specialist in Far Eastern languages, Evgenij Polivanov, in an essay dealing with a phenomenon that he termed “sound gesture.” This essay constituted a partial disputation of Jakubinskij’s views. Polivanov began by dividing all the means of linguistic expression into two by now familiar categories: the one completely arbitrary andconventional, for example, the phonic structure of the Russian word for table— s+t+o+l—which in itself does not suggest its meaning; the other motivated and natural, such as the intonation that expresses emotional states and seems to be immediately understandable to anyone, even to animals. Gestures—non-linguisticmeans of expression that often accompany emotive language—are prime examples of the latter category. They convey emotions in the most direct fashion.

Very soon, however, Polivanov undercut this simple opposition. As he argued, both motivated and arbitrary linguistic expressions are in fact conventional—deriving from the relation of contiguity between expressions and their meanings. “If we know that a given extralinguistic phenomenon is expressed through a particular intonation or gesture, the origin of this knowledge can be simply explained by the fact that we have always or often observed such an emotion accompanied by the given intonation or gesture. Thus, we have learned this link in precisely the same fashion as we learned the link between the phonic sequence s+t+ o+l and the representation of table, for this sequence was always used by the speaker when the thought of table was present” (Polivanov 1919, p. 30). Therefore, the difference between so-called natural and conventional linguistic expressions is not absolute but rather a matter of degree, an admixture of the two principles.

If all means of expression were placed on a scale from “conventional” to “natural,” the closest to the natural, in Polivanov’s opinion, would be mimetic gestures that copy objects or actions and seem spontaneously comprehensible to everyone. Well aware that the process of reproduction is always conventional, Polivanov calls these gestures “potentially natural.” The question, then, is whether language contains any “phonic sequences (combinations of vowels and consonants in a certain order) whose role is analogous to that of potentially natural gestures” (ibid., p. 31). The answer is yes, as Polivanov illustrates with numerous Japanese onomatopoetic words imitating sounds, and reduplicative words imitating the repetition of an action or the recurrence of a phenomenon. By analogy with “mimetic gestures,” Polivanov termed these imitative linguistic expressions “sound gestures.”

Polivanov departed from Jakubinskij both in denying that emotions are the vehicle of the direct expressiveness of linguistic sounds, and in not considering the connection between sound gestures and verbal art.22 In some respects, however, the two Formalists shared common ground. First of all, Polivanov claimed that sound gestures and children’s language were related. “Japanese ‘sound gestures’ can be regarded in general as the principle of a special, childish morphology that has retained its right to existence in the language of adults” (Polivanov 1919, p. 26). In addition, both Jakubinskij and Polivanov believed that a substantial phonic difference, which has its roots in pronunciation, existed between poetic language and sound gestures on the one hand, and practical or normal language on the other. For Jakubinskij, the clustering of liquids impedes pronunciation, thus attracting attention to the sounds themselves. Polivanov observed that in Japaneseonomatopoetic and reduplicative words the phoneme [p] occurred, which has disappeared from contemporary Japanese except in loan words; the “nasal g” [η]is also found in initial position in these words, though otherwise it occurs only medially or finally. Such aberrations, Polivanov believed, are caused by the fact that the “value of the particular phonic structure [of sound gestures] is greater than in other words. In normal words, as a matter of fact, it does not make any difference which phonic complexes express a particular idea...But obviously for ‘onomatopoeia’ words, some links between the expressed representations and particular sounds are important” (ibid., p. 34). Thus, Polivanov concludes, the p in normal language can easily be replaced by any other speech sound, but it must be retained in words imitating, for example, the puffing of tobacco smoke or thesound of a flute.

The Formalists, to be sure, did not claim originality in discovering the importance of oral articulation in language and verbal art.They referred to such nineteenth-century scholars as the German psychologist Wilhelm Wundt (1832-1920) or the Polish classical philologist Tadeusz Zieliński(1859-1944), whohadmade similar observations about pronunciation as the bridge between sound and meaning.23 But the Formalists’ interest in the articulatory aspect of language was most likely triggered by the rise of Ohrenphilologie in the German literary studies of Eduard Sievers (1850-1932), his pupil Franz Saran (1866-1931), and others. In contrast to traditional Augenphilologie, which analyzed the text primarily as a visual or graphic manifestation, Sievers’s “aural philology” emphasized the acoustic aspect of the text. Of particular interest were the involuntary motor reactions (movements of the diaphragm, bodily motions, facial expressions, and gestures) accompanying an utterance, which, in their opinion, were decisive in articulating the phonic substance of language.

There are several reasons that Ohrenphilologie was so attractive to the OPOJAZ Formalists. First of all, even though its overall outlook and goals were incompatible with those of the Russian Futurists whom the Formalists found so congenial, their respective views of literature coincided on one important point, namely, that sound is central to poetry. Earlier I mentioned that Kručënych attributed poetic value to foreign languages unknown to him as one source of his zaum’ (see note 20). In a striking correspondence Saran wrote: “the theoretician of verse…ought to adopt toward verse the attitude of a foreigner who listens to it without knowing the language in which it is written” (quoted in Jakobson 1923, p. 21). Jakobson considered this statement an epitome of the Ohrenphilological outlook.

Like Ohrenphilologie, OPOJAZwas essentially positivistic, attempting to establish a new science of literature that would “turn to the facts and push aside general systems and problems” (Èjchenbaum 1927, p. 120). Inthis “new fervor of scientific positivism,” sound was considered the only concrete reality of verbal art, for meaning, in its ephemerality, was only a subjective mental construct that could not be pinned down with any certitude. An earlier linguist and teacher of some of the OPOJAZ members, Lev Ščerba (1880-1944), had expressed this view in his introduction to a “linguistic commentary [tolkovanie]” on one of Puškin’s poems that was primarily a directive for the proper oral delivery of this text.24 In it, Ščerba argued that “all semantic observations can only be subjective,” whereas the analysis ofpoetic sound, especially in the oral reading of a text, can attain to some degree the objectivity of a laboratory experiment (Ščerba 1923, p. 17). This claim to scientific objectivity is reflected in the title of an informative article on Ohrenphilologieby Šklovskij’s brother Vladimir, which appeared in the second volume of the OPOJAZ Studies: “The Rhythmical-Melodic Experiments of Prof. Sievers” (Vl. Šklovskij 1917).This esteem for the methods of “aural philology” extended beyond the early stage of Formalism.25 Aside from the German literary theorist Oskar Walzel(1864-1944), Sievers and Saran were the only honorary members of the Section for Verbal Arts at the State Institute for the History of the Arts, the institution that absorbed OPOJAZ in the twenties (Otdel slovesnych iskusstv GIII 1928, p. 155).

Of the contributors to the OPOJAZ Studies, the closest to Ohrenphilologie was Boris Èjchenbaum. His affinity to this approach was most likely a function of his age. Born in 1886, Èjchenbaum began his literary studies before the advent of Formalism. Thus, he did not always share the Bohemian proclivities of some of the younger members of OPOJAZ, apparently more impressed by sober scholarship than the vague notions of Futurism. The scientism of Ohrenphilologie coincided with Èjchenbaum’s own orientation, as recorded in his diary entry of January 1919: “Proceeding from Rickert, one realizes that the methods of the natural sciences must be applied to the history of the arts... when we deal with the ‘nature’ of the material from which the work is made. In [this field] the construction of laws and definitions is quite conceivable” (quoted by M. Čudakova in Tynjanov 1977, p. 455). Moreover, Èjchenbaum held that the material of verbal art is the oral word.

We always speak about literature, the book, the writer. Written-printed culture has inculcated the letter in us.… We often totally forget that the word has nothing to do with the letter, that it is a living, ongoing activity treated by the voice, articulation, and intonation and joined by the gesture and facial expression [mimika]. We think that the writer writes. But it is not always so, and in the realm of the artistic word it is moreoften just the opposite. The German philologists (Sievers, Saran, et al.) began to argue a few years ago that the philology of the “eye” (Augenphilologie) must be replaced by its “aural” counterpart (Ohrenphilologie). This is an extremely fertile idea which has already yielded interesting results in the domain of verse… Such an “aural” analysis, however, is also fruitful for the study of artistic prose. The bases [of this form] are also marked by its origin in the oral skazwhich influences not onlyits syntacticstructure and the selection and combination of its words, but its very composition” (Èjchenbaum 1918, p. 10).

The untranslatable term skaz(akin to the Russian verb skazat’to tell) subsequently gained wide currency in Slavic literary studies. It was the focal point of Èjchenbaum’s Formalist debut, his analysis of Gogol’s short story, “The Overcoat.” Skaz designated a particularnarrative technique in which the elements of oral delivery play a crucial role. The structureof Gogol’s story, Èjchenbaum claimed, is not organized according to the laws of the plot but rather by a “certain system of varied expressive-articulatory facial gestures” (Èjchenbaum 1919, p. 151).In a later study devoted to the Acmeist poet Anna Achmatova, Èjchenbaum applied the notion of the articulatory gesture to poetry as well. His thesis was that Achmatova’s poetry “is oriented toward the process of pronunciation, of expressive [mimičeskij] pronunciation” (Èjchenbaum 1923, p. 87).This orientation is manifested in the frequent occurrences of what Èjchenbaum termed the “expressive quality of speech [rečevaja mimika].” He showed how the repetition of the same or similar vowels or the juxtaposition of contrasting ones forces the reader to move his lips in a particular way so that the “words come to be perceived not as ‘sounds’ and not as articulation in general but as an expressive [mimičeskij] motion” (ibid., p. 86).

The years separating Èjchenbaum’s study of Gogol’ and his monograph on Achmatova mark an important period in the development of OPOJAZ. As Èjchenbaum himself observed theteens for the Petersburg Formalists were “years of struggle andpolemics,” so that “many of the principles [they] advanced duringthese years of intensive struggle with their adversaries were not merely scholarly principles but paradoxical slogans exaggerated for polemical and contrastivepurposes. The failure to takethis into account, to treat the 1916-1921 works of OPOJAZ as strictly scholarly, would be to ignore history” (Èjchenbaum 1927, p. 132).The stock-taking thatfollowed this period of Sturm und Drang was to lead to an intensive reexamination of the earlier position. The linguistic approach to verbal art and the key notionof poetic language were among the first to undergo this scrutiny.

Èjchenbaum himself launched this critique. He commended the recent confluence of poetics with linguistics as a healthy counterbalance to the traditional domination of poetics by psychology or sociology. “But,” he warned,

a rapprochement with a neighboring discipline can be genuinely fruitful only ifit does not lead to a new submission. Inassociating with linguistics, poetics ought to retain its independence. For linguistics, a poetic work is a “phenomenon of language” that furnishes interesting material for the study of phonetic, syntactic, orsemantic issues. Linguistic observations about poetic language enrich the general science of language with new phenomena that occur only rarely in normal “practical speech.” The literary theoretician, however fruitful he may find linguistic methods to be, should pose his questions in a completely different way. What emerges here is the distinction between the concepts of language and style, linguistic phenomenon and stylistic device. Linguistics belongs among the natural sciences, poetics among the humanities [nauki o duche]. Linguistics classifies poetic language as one of its varieties; it differentiates among them according to their goals merely to classify the phenomena of language as functions. Poeticsbegins with the separation of poetic language from other linguistic phenomena as an activity set toward a particular goal. And even though this goal cannot be defined with any precision, its symptoms are apparent. In this way, poetics is built on the foundation of a teleological principle and thus proceeds from the notion of the device; linguistics, like all natural sciences, deals with the categoryof causality and therefore proceeds from the notion of the phenomenon as such (Èjchenbaum 1922b, p. 14).

Linguistically oriented Formaliststended to dismiss this statement of Èjchenbaum’s as a relic of nineteenth-century scholarship. Viktor Vinogradov, for example, claimed that “both the inclusion of linguistics among the natural sciences and the disregard for the teleological principle in it are widespread but incorrect, narrow-minded ideas” (Vinogradov 1923, p. 206).

Èjchenbaum was not the only Formalist in the early twenties to clash with the concept of poetic language and the linguistic approach to literature so central to OPOJAZ. In a proposal for a monograph on Evgenij Onegin, Jurij Tynjanov listed as one of his topics, “Why poetic language is not a poetic dialect and does not belong completely within descriptive linguistics” (Tynjanov 1977, p. 416). Thus, Èjchenbaum’s (and Tynjanov’s) dissent from the other Formalists cannot be simply swept aside.

Any characterization of Èjchenbaum’s position will depend on what we make of his concept of the device. At first glance, his contrasting of poetic teleology with linguistic causality mayappear to be another version of Šklovskij’s mechanistic model. However, Èjchenbaum speaks of stylistics and linguistics, and contrasts the stylistic device with the linguistic phenomenon. In this respect his polemics recalls Žirmunskij’s critique of the mechanistic metaphor discussed in the preceding chapter. It was precisely through the notion of style that Žirmunskij reformulatedthe functional definition of the device. From his standpoint, style is a principle of unity determined by the overall artisticgoal, which ascribes to each device a specific role within the artistic whole. The device is thus not an a priori, independent monad of artistic form for the morphologists, but a functionally integrated element of the work. In the same way, though from a different theoretical perspective, Tynjanov argued against an atomisticapproach to the device. In his systemic metaphor, the identity of each element is a function of the hierarchical relations within the work and the higher systems in which the element participates.

It is obvious that Èjchenbaum’s rejection of the linguistic model was motivated bysimilar considerations. For the linguist, he believed, poetic and practical language are nothing but abstractions. In separating the two, the student of language might classify them, “among other ways” (Jakubinskij’s words), accordingto their respective goals. To do so, however, is only a heuristic procedure, amatter of choice, as Jakubinskij himself demonstrated when in 1922 he rejected the goal as an inadequate criterion and proposed to classify utterances according to their actual forms (Jakubinskij 1923, p. 115-16). Students of literature, however, do not have this choice, for they deal with concrete literary works, that is,intentionally created poetic wholes. From their perspective, the ontological difference between poetic and practical language (forexample, the clustering of the same liquids) orbetween sound gestures and normal linguistic usage (for example, the occurrence of the speech sound p) is unimportant. It is not the presence or absence of these particular features that concerned Èjchenbaumas a literary scholar, but their functional place in the literary work. “Poetic language,” he argued, “is characterized solely by a particular set toward certain elements of speech and a specific utilization of them” (Èjchenbaum 1927, p. 250).

In more abstract terms, it might be said that the two factions in OPOJAZ used different “logics.” Those advocating the linguistic model were quite close to the mechanists in casting their categories in the form of polar oppositions. Their critics shunned this disjunctive stance, instead casting their categories in terms of a gradation, a relative difference. Thus, the linguist Jakubinskij, inspired by the Futurists’ zaum’, split all linguistic behavior into two incompatible classes: poetic language oriented solely toward the phonic aspect of speech, and its opposite, practical language set toward the semantic aspect of speech. His critic, Èjchenbaum,though considering this a powerful working hypothesis, claimed that it was not supported by the facts. Commenting on practical language, he wrote, “It is quite doubtful that there actually exists a type of speech in which our attitude toward the word would be totally mechanical, in which the word would be exclusively a ‘sign.’ Forms such as oratory, for instance, regardless of their ‘practical’ character, are in many respects quite close to poetic language” (ibid.).And Žirmunskij criticized the absolutism of theopposite category, poetic language, conceived as a purelyphonic structure. “If the poet really wanted to affect us by mere sounds he would take up music.” Poetry “does not affect the listener by sound as suchbut by sounding words, i.e., sounds tied to meaning” (Žirmunskij 1922, p. 149).

Though Žirmunskij and Èjchenbaum both conceived of style as the functional integration of elements in an artistic whole, they disagreed on the nature of this integration. Žirmunskij, faithful to his organic metaphor, favored a static notion of the whole in which elements were harmoniously related. Èjchenbaum, in contrast, prepared the way for the systemic metaphor byadvocating a more dynamic view. According to him, the unity of a work was a fragile equilibrium of elements struggling for domination. I dwelt on this difference in the preceding chapter and repeatit only to avoid the false impression that Žirmunskij andÈjchenbaum were speaking the same language. In fact, Žirmunskij’scriticism of those conflating literature and music was not addressed to the linguistically inclined OPOJAZ members at all but toÈjchenbaum, in a review of the latter’s book The Melodics of Russian Lyric Verse. In this work Èjchenbaum had formulated his dynamic notion of the poetic whole as a struggle between the organizing element (the dominant of the work) and the other subordinate elements constituting this whole. He illustrated his position with lyric poems wherethe dominant intonation deformed all aspects of language, including semantics, to its needs.

The deformation of semantics that Èjchenbaum discussed, despite Žirmunskij’s claims to the contrary, was quite different from that described by the early OPOJAZ members. We recall thatthey treated poetic language as sound that might but need not be accompanied by a cognitive meaning. Èjchenbaum was concerned not with the presence or absence of meaning in a particular verbal construction, but rather with its function there, a function determining its hierarchical position relative to the other elements of the construction. In other words, for him, meaning is always involved in a verbal construction, but sometimes it is subordinate to other elements and at other times it dominates them. Oratory, Èjchenbaumargued, may foreground the phonic aspect of language for the sake of persuasion, whereas artistic prose may be quite indifferent to sound if its goal requires this.

Joining Èjchenbaum against Jakubinskij’s separation of poetic from practical language was Boris Tomaševskij, who wrote, “Instead of the clear, though perhaps terminologically unfortunate opposition of the old scholastic theory, ‘poetry’ and ‘prose,’ we, following a linguistic path, have advanced another opposition: ‘practical’ versus ‘artistic’ language. This opposition, however,does not cover all aspects of a verbal composition. It pertains solely to the sphere of language and, secondly, does not coincide with the bounds of ‘poetry’ and ‘prose.’ For the ‘prosaic’ perhaps as much as the ‘poetic’ should be contrasted to ‘practical language’” (Tomaševskij 1924, p. 140).

Earlier in this chapter I described how Jakubinskij arrived at his frame of reference. He proceeded from Potebnja’s original opposition between poetic and prosaic language, but replaced the second element with “practical language,” which he considered more appropriate. His critics proceeded in the opposite fashion; they retained “prosaic language” and replaced the other element of the opposition with what they claimed to be the more accurate concept of “verse language.” In the introduction to his pioneering 1924 monograph, The Problem of Verse Language,Tynjanov explained this step: “The notion of ‘poetic language’ put forth not so long ago is today in a crisis which is undoubtedly caused by the broad and diffuse character of this psychological-linguistic concept. The term ‘poetry’ that had long existed in our language and scholarship has now lost its concrete scope and content and gained an evaluative tinge. In this book I shall analyze the specific concept or verse (in opposition to the concept of prose) and the specific features of verse language” (Tynjanov 1924, p. 5).26

These conceptual shifts were not solely a matter of terminology. By substituting the notion of practical language for Potebnja’s “prose,” Jakubinskij was redefining the category ofthe poetic. The same was true of his critics. Their opposition between verse and prose is not equivalent to the earlier dichotomy of poetry and prose. The early OPOJAZ members ignored verse, considering verse rhythm just one of many artistic devices that de-familiarize the sound stratum of language, whereas their critics argued that verse and prose occur in both literature and byt.27 What these two forms represent is not the opposition of art to non-art, but two different principles of verbal construction, or what Tynjanov called functions. Tomaševskij wrote in his comprehensive Russian Versification that “the difference between prose and verse rests in the fact that in verse the phonic imperative [zvukovoe zadanie] dominates the semantic one and in prose the semantic dominates the phonic one. Everything boils down to the relative role of these two origins” (Tomaševskij 1923, p. 8). Similarly, Tynjanov argued that “it would be premature to conclude that verse form differs from prose form merely because in verse the external sign of the word plays the exclusive role whereas in prose such a role is performed by its meaning.” He concluded, “Prose and poetry, it seems, do not differ from each other in their immanent phonation and in the consequent set toward sound in poetry and semantics in prose, but rather in the way these two elements interact: how the semantic aspect of prose deforms its phonic aspect (the mental set toward the semantic) and how verse deforms the meaning of the word” (Tynjanov 1977, pp. 53, 54).

In short, the linguistic model and its fundamental concept, poetic language, underwent a criticism within OPOJAZ in the early twenties that entailed a significant shift in the scholarly endeavors of the group. Of course, this shift was not a total abandonment of the previous Formalist tradition. Those who rejected the “vague” and “inadequate” concept of poetic language followed the path established by their predecessors in one important respect. Theytoo focused theirattention on verbal constructions in which sound played the dominant role. However, they no longer carried out their research under the banner of the theory of poetic language but under that of metrics and verse semantics.


I would define, in brief, the Poetry of words as The Rhythmical Creation of Beauty. Its sole arbiter is Taste. With the Intellect or with the Conscience, it has only collateral relations. Unless incidentally, it has no concern whatever either with Duty or with Truth.

— Edgar Allan Poe, “The Poetic Principle”

Perhaps the most influential among the early Formalist studies of verse was Osip Brik’s 1920 lecture at OPOJAZ entitled “Rhythm and Syntax.”28 In it he coined the term “rhythmical impulse,” which became the “focal point of the Formalist and Structuralist conception of verse” (Červenka 1983, pp. 52-53). To appreciate Brik’s contribution fully it is necessary to sketch out its historical context.

The principles of Russian versification the Formalists inherited were those of the great poet-theoreticians of the Symbolist generation, A. Belyj and V. Brjusov. Though innovative in their approach to the study of verse (Belyj, for instance, was the first in Russia to apply statistics to metrics), their theories did not satisfy the young Formalists. Intheir eyes, the three major failings of Symbolist metrics were as follows: first, an atomistic approach to verse; next, the separation of meter from rhythm: and finally, prosodic egocentrism. The Symbolists considered the foot the minimal unit of verse. Unable to detect any overall Gestalt in the verse under study, they arbitrarily analyzed even the most regular verse into heterogeneous feet. This blindness to the holistic nature of verse stemmed from their divorce of meter from rhythm. They insisted that meter was an ideal scheme existing prior to verse, whereas rhythm was the actual pattern of deviations from this scheme. Because the Symbolists attributed aesthetic value precisely to such deviations, in their own analyses they purposely sought to segment verse into as many different kinds of feet as possible.

To avoid the problems of Symbolist metrics, Brik’s study did away with the concept of meter entirely. Instead it treated rhythm as the motoric or kinetic precondition of verse. “As a scholarly term, rhythm means a particular formation of the motor processes… motion shaped in a particular way” (Brik 1927, p. 16). Rhythmic shaping is a function of quantity (the increase or decrease in motion) and duration (the continuity or discreteness of motion). The projection of rhythm onto verbal material—the kinetic organization of an utterance in terms of stresses and intervals—constitutes what Brik terms the “rhythmical impulse.” This impulse organizes the verse as a whole, a fact that had eluded the Symbolist theoreticians. Only if we know the rhythmical movement of the entire poem can we correctly identify its smaller units. Brik takes as an illustration a line from Puškin, which in isolation seems dactylic but within the poem as a whole turns out to be trochaic. He concludes, “one should not speak of strong and light syllables [downbeats and upbeats]but of stressed andunstressed ones. Theoretically, any syllable can be stressed or unstressed; everything depends on the rhythmical impulse” (ibid., p. 17).

Brik’s statement clearly reflects the iconoclastic attitude of his Hylaean friends toward traditional accentual-syllabic versification. As they wrote in 1913, “we stopped seeking meters in the schoolbooks; every motion generates a new, free rhythm for the poet” (in Markov 1967, p. 52). To achieve such total rhythmical freedom, Futurist poets manipulated language in a particular way, as they themselves admitted. They “disregarded grammatical rules” and “shattered syntax.” It soon became evident to Brik, however, that the majority of Russian verse is written in more traditional language than zaum’, language whose words are units of meaning combined semantically as well as prosodically.

To account for the semantic constraint upon the rhythmical impulse in ordinary Russian verse, Brik returned to the concept of syntax disdained by the Futurists. “Syntax,” he wrote, “is the system of combining words in ordinary language. As long as verse language does not abandon the essential laws of prosaic syntax these laws are obligatory for it” (Brik 1927, p. 32).Thus, the relationship of sound and meaning in verse is necessarily complex: it is always a compromise between rhythmical and syntactic considerations. A verse line, in Brik’s opinion, is the minimal implementation of this rhythmical-syntactic compromise. It is a unit separated from the rest of the utterance to which it belongs on the basis of its prosodic features, but at the same time containing syntactic connections among its elements. “A rhythmical-syntactic word combination differs from a purely syntactic one in that it incorporates words into a fixed rhythmical unit (a line); it differs from a purely rhythmical combination in that it links words not only phonetically but semantically” (ibid.). Rhythmical and syntactic requirements may coincide in verse, as when a line is a complete sentence, or they may clash, for example, in caesuras or enjambments. In either case, words in verse are always subject to two sets of combinatory rules.

The value of Brik’s essay for Formalist metrics lay in its firm grasp of the structuring principle of verse. This grasp, however, was achieved only at the cost of considerable oversimplification, and all subsequent Formalist studies of the topic complicated Brik’s clear-cut picture. Its first limitation was its equation of the vehicle of rhythm with word stress alone. Obviously, in addition to the stress within an isolated word there are a variety of stresses belonging to higher syntactic units. Once this premise is accepted, syntax can no longer be seen in simple opposition to rhythm as meaning versus sound. Syntax actually consists of both phonic and semantic strata, furthermore, the phonology of syntax cannot be limited to the intensity of the voice (syntactic stress). The voice also has pitch, whose modulation creates syntactic intonation. It was this aspect of verse that Èjchenbaum examined in his study of the melodics of Russian lyric poetry.

Èjchenbaum divided the lyric into three categories according to the role played in each by intonation. In the declamatory (rhetorical) lyric, intonation supports the logical structure of the text; in the conversational lyric it serves to link the verse to everyday language. In both these types of lyric, intonation is subordinate to other verse elements. In the third lyric type, intonation performs a more significant function. This is the singable (napevnyj) lyric, which purposely imitates musical melody. In such poetry “we observenot a simple alternation of speech intonations but a developed system of intonation that determines the composition of the poem more than its verbal themes” (Èjchenbaum, 1922b, p. 9). Only such intonational schemes—symmetries, repetitions, or cadences—can in Èjchenbaum’s view be called melodics proper. Here intonation ceases to be a mere epiphenomenon and becomes the organizing principle of verse—its dominant.

The semantic aspect of syntax is subordinated to intonation in this type of lyric. For example, Vasilij Žukovskij, a Russian poet of the first half of the nineteenth century, exploited the syntactic patterns of emotive language for melodic ends. Some of his poems are merely a series of interrogative sentences combined with exclamations. Afanasij Fet (1820-1892), in contrast, built his melodics on intonational emphasis. To attain it he inverted word order, repeated lexical items in significant positions (anaphora, epiphora) and employed syntactic parallelism. With such cases in mind, Èjchenbaum concluded that the “analysis of the melodic style in which the role of intonation is obvious suggests the need for a study of the role it plays in verse in general” (ibid., p. 195).

The strength and the disadvantage of Èjchenbaum’s study lie in its specialization. His scheme convincingly illustrated the idea that verse is a hierarchical structure and called attention to one hitherto neglected element of this structure. But given its author’s mistrust of linguistics, the concept of syntax with which it operated was vague, to say the least. Furthermore, by focusing on intonation, it inevitably slighted other important factors. A study of melodics cannot substitute for a general theory of verse. The formulation of such a theory was left to the other Formalists.

In 1919, at a lecture before the Moscow Linguistic Circle, Boris Tomaševskij defined the role of rhythm in verse as the “distribution of expirational energy within the limits of one wave—the verse” (Tomaševskij 1929b, p. 182).This definition is broad enough to subsume both Brik’s rhythmical impulse and Èjchenbaum’s melodics. In addition to “lexical-accentual” (slovesno-udarnyj) and “intonational-syntactic” (intonacionno-frazovoj) rhythm, Tomaševskij spoke of “harmonic” rhythm (ibid., p. 25). Borrowed from the French linguist Maurice Grammont (1866-1946), “harmony” designates the relation between speech sound distribution and the rhythmical organization of the line. In verse, according to Tomaševskij, “harmony fulfills a twofold task: first, dissimilation—the segmentation of speech into rhythmical periods; second, assimilation—the evocation of the idea that the segments thus marked are analogous” (ibid., p. 22).

Rhyme is a good example of a harmonic correlation. On the one hand, it demarcates one rhythmical unit (a line) from the text, and on the other, it renders the two lines analogous through the repetition or sounds. But rhyme is not the only such phenomenon in verse. As Brik argued, verse is always marked by the orchestration of speech sounds. Using Puškin’s and Lermontov’s poems as examples, he showed how thoroughly poetry is permeated with sound repetition.

Tomaševskij’s attitude toward sound repetition differed considerably from Brik’s. Tomaševskij was not interested in repetition as a manifestation of the “fundamental euphonic laws” of poetic language, but as a functional element of rhythmically organized speech. In the Russian trochaic tetrameter, he argued, even feet carry stress more often than odd ones and the line tends to break into two colons each composed of one strong and one weak foot. This rhythmical partition of the line is underscored by the distribution of vowels in Puškin’s verse (where each downbeat is stressed):

On imel odnoviden’e
o—e o—e
(He had a single vision) (ibid., p. 23)

This, of course, is just one instance of the correlation of speech sound repetition and verse rhythm, and Tomaševskij provides many others to support his thesis that “verse ‘harmony’ belongs fully within the theory of rhythm” (ibid., p. 24).

Not only was Tomaševskij’s theory of verse rhythm more inclusivethan that of the other OPOJAZmembers, but it was constructed from the standpoint of the perceiving subject.29 In discussing harmonic rhythm, for example, he stressed its capacity for evoking the idea of analogy in the subject. In this respect he departed considerably from both Brik and Èjchenbaum. Brik arrived at his concept of the rhythmical impulse from the perspective of the creating subject. The kinetic organization of the verse (the regular distribution of word stresses in it) engenders motor processes that are present during its generation. The perceiver merely re-presents this original motion in his or her reading. It might seem that Ohrenphilologie had reversed this hierarchy in stressing the aural perception of verse, so that the perceiving subject was its point of departure as well, but this shift was purely a heuristic device. Sievers’s experiments with recitation in fact served as the basis for reconstructing what he took to be the correct authorial reading. And Èjchenbaum deliberately bracketed off the act of perception, seeking only the “objective” preconditions of verse melodics that he identified with syntax: “Independent of individual nuances in reading, syntactic structure is a totally objective fact and syntactic intonation, within the bounds of our requirements, is obligatory” (Èjchenbaum 1922b, p. 16).

This reduction of verse to its “objective” preconditions was clearly unacceptable to Tomaševskij.“We recognize verse through immediate perception,” he argued in the opening paragraph of Russian Versification. Yet, “‘verse-quality’ [priznak stichotvornosti] is generated not solely from the objective attributes of poetic language, but from the conditions of its artistic perception as well, from the hearer’s judgment about it based on his taste” (Tomaševskij 1923, p. 7). Thus, the starting point of metrics should not be rhythm as such but its constitution in the perceiver’s consciousness.

At the most abstract level, rhythm is experienced when a “phenomenon becomes arranged in ‘periods’ that are perceived as ‘isochronous,’ whereas in objective time they may be unequal” (Tomaševskij 1929b, p. 258). This is a generalization of Tomaševskij’s observations on the twofold task of “harmonic rhythm” discussed earlier. The constitution of rhythm in the perceiver’s consciousness has both dissimilative and assimilative aspects. It dissolves the utterance into distinct rhythmical periods and at the same time, by rendering these periods rhythmically equivalent, reconstitutes the utterance. In terms of the inner experience of time, this act can be described as a continuous interplayof expectations and fulfillments. The reading of a “long series of repeated, analogous lines creates a sort of rhythmical inertia in the perceiver, a scheme of ‘prosodic expectations’” (ibid., p. 142). Expectation alone is insufficient for the arousal of rhythm in consciousness: “Regularity distinguishes rhythmical speech from unorganized, unregulated speech only if the formed complex of phonic phenomena… recurs and is perceived as similar, thus enforcing in perception the sensation of this ‘regularity.’” The fulfillment of expectations, the “‘recognition’ at every moment of a recurring regularity,” must accompany the original expectation for the emergence of rhythm in the perceiver’s consciousness (ibid., p. 260).

Conceptualized so generally, however, the notion of rhythm clearly exceeds the sphere of metrics. The experience of rhythm as just described occurs not only in poetry but in the other temporal arts, as well as in extra-artistic areas. Second, “rhythm” in Tomaševskij’susage refers to the “objective” stratum of rhythmical experience, the real phonic sequence that the perceiver faces. In its actual physical heterogeneity, this stratum inevitably defies systematic description. According to Tomaševskij, “rhythm can only be concrete, can be based only on the elements of phonation that we hear or actually take into account in both rhythmical and non-rhythmical speech” (ibid., p. 13). In this respect, rhythm is a singular phenomenon: every utterance, every line, can have its own rhythm based on the repetition of any phonic element. In relation to verse, Tomaševskij prefers not to use the term “rhythm” but to speak instead of the “rhythmical impulse.”

As I pointed out earlier, the concept of the rhythmical impulse was introduced into Formalist terminology by Osip Brik. With Tomaševskij,however, itacquired quite a different meaning. Whereas Brik’s rhythmical impulse pertained to the motor process generating verse, Tomaševskij’s pertained to the process of interaction between verse and its perceiver. In this new meaning, the rhythmical impulse is an abstraction from the actual rhythm perceived by the subject. The isochronism of verse periods implies a selection among phonic features, the designation of those to be considered equal. Tomaševskij calls these “rhythm-creating elements.” Thus,verse, in “dissolving itselfinto periods that are subjectively evaluated as equivalent, maintains the law common to all periods and orders its rhythm-creating elements analogously” (ibid., p. 260). This reduction of all phonic data to those that arerhythm-creating, and hence regularly repeated throughout a poem, limits considerably the number of rhythmical possibilities and provides the perceiver with a grid or skeletal structure within which the interplay of expectations and fulfillments takes place. For under these conditions “rhythm is perceived against the background of an average rhythmical scheme, the most frequent, most expected one. We shall call this rhythmical expectation created in our perception by the aggregate effect of a series of recited lines, this ‘general idea’ about the rhythmical character of a poem, the rhythmical impulse” (Tomaševskij 1923b, p. 65).

It must be stressed, however, that Tomaševskij distinguished rhythm in general from the rhythmical impulse proper to verse not only on intrinsic criteria. The heterogeneous phonic elements whose repetition constitutes rhythm lack a social and historical dimension. As rhythm occurs outside language, virtually any phonic feature can serve as its vehicle, but verse language is a linguisticphenomenon and its repertoire of rhythm-creating elements is necessarily restricted by the social nature of language. “Language,” in Tomaševskij’s view, “is what links the speaker to the hearer. The speaker not only utters words but also listens to them, and the hearer is not absolutely passive in his listening. Language is apprehended because the hearer knows it. The sounds reaching his ears are signals for him to recognize the speech as an utterance that he could have made himself. The most passive listening is always accompanied by an activity— inner speech. Thus, reception and production inextricably compriseany linguistic fact. Only those features copresent in pronunciation and perception can be essential to language. Only this link—the consonance of the utterer and the hearer—is real language” (Tomaševskij 1929b, p. 30).In terms of the theory of verse, not every linguistic idiosyncrasy of interlocutors (such as a poet’s stammer), but only those that are obligatory for both utterer and hearer, can become rhythm-creating elements in verse. This premise was elaborated in detail by Jakobson in a book on Czech metrics written about the same time as Tomaševskij’s remarks (see below). It became the cornerstone of his phonological metrics, which Tomaševskij himself embraced in the mid-1920s.

The social nature of literature and the history of verse impose another constraint on the selection of rhythm-creating elements. In encountering a poem, for example, hearers or readers are usually not a tabula rasa, innocent minds exposed to verse for the first time. Almost always they carry with them the memory of their previous dealings with poems, a backlog of literary education, tradition, and so forth. The fact that they are willing to see the various lines of a poem as comparable, even if quite dissimilar, indicates that the constitution of the rhythmical impulse has at its basis some canonized set of rhythmical conventions. This for Tomaševskijis “meter.” Metrical norms function similarly to linguistic ones in the perception of verse rhythm. They “make the comparison (of verse units) easier by highlighting those features whose apprehension yields material for appraising the equivalence of speech periods. The goal of these norms is to provide a prearranged system for organizing the system of phonations, that necessary conventionality which links the poet with his audience and helps his rhythmical intentions to be perceived” (ibid., p. 11).

In using the concept of meter, however, Tomaševskij did not revert to the Symbolist dichotomy of meter and rhythm. For him, the two were not absolutely distinct: “It is clear that the study of a norm cannot be separated from the study of actual possibilities, the concrete forms of the phenomenon that are subject to this norm” (Tomaševskij 1929b, pp. 53-54). The actual implementation of a metrical norm is not a series of deviations from an untenable ideal but a set of tendencies complying to one degree or another with this norm. Thus, Tomaševskij’s 1919 study of Puškin’s iambic pentameter measures statistically the tendency of syllables to be stressed. As might be expected, odd syllables are stressed only exceptionally: even ones are much more frequently stressed, but even these are not stressed equally. Only the last syllable (or the penultimate one in feminine endings) carries an obligatory stress, because “this syllable is the boundary of the rhythmical series (the line) and subsequent syllables… do not continue this series but lie outside of it” (ibid., p. 141). On all the other even syllables, stress is distributed according to poetic style. Puškin’s iamb differs in this respect from the iambs of other nineteenth-century poets, and even the proportion of stressed syllables varies in different stages of his career.

Tomaševskij’s conception of meter also differs from the Symbolists’ in its relativism. Different languages inevitably employ different prosodic elements as vehicles of the “same meter.” And even within a single poetic tradition the metrical system changes in time. The change is triggered by shifts in the hierarchy of what Tomaševskij calls primary and secondary features of verse. A primary feature is a regular distribution of one phonic element canonized by a given metrical convention. “Thus, in classical [Russian] metrics, the canonized element of sound ordered according to the metrical norms is accent” (ibid., p. 8). Because verse language is a complex structure of correlated elements, the canonized ordering of one phonic feature entails the regular distribution of others. This patterning, though often vague or subliminal, creates the secondary features of verse, that is, its actual rhythm. Such a clear-cut distinction between primary and secondary features exists only at the moment when a particular metrical system is generally accepted as the only one possible. When its authority begins to be questioned, the secondary features come to the fore. Poets realize that “it is possible to write verse governed only by secondary features, that an utterance can sound like verse even without meter” (ibid., p. 9). Ultimately such a situation leads to the abandonment of the previous metrical norm and theestablishment of one of the secondary features as a rhythm-creating element.

Given the paramount role of meter in generating the rhythmical impulse, it is not remarkable that Tomaševskij considered it the “specific differentia of verse vis-à-vis prose” (ibid., p. 10). But insofar as he defined verse as the implementation of a specific metrical norm, he was unable to account for it’s overall unity. It was impossible for him to say what iambic and trochaic verse have in common, or, given the geographical and historical relativity of meters, what the connection is between, say, iambic verse in different languages or different historical periods. Therefore, Tomaševskij introduced the concept of verse language, which unites metrically disparate verse on the basis of other shared properties. For instance, “in contemporary European practice the custom was established of writing verse in even lines differentiated by capital letters, and to print prose in continuous lines without breaks. Despite the heterogeneity of graphia and living speech, this fact is significant, because there are specific linguistic associations with writing. The segmentation of the utterance into ‘lines,’ periods whose phonic potential is comparable or even identical in very simple cases, is evidently the distinctive feature of verse language” (ibid., p. 11).

This fact, however, does not imply that prose written as verse will always and everywhere be perceived as such, or vice versa. The customary graphic arrangement merely signals to the European reader one formal difference between verse and prose, but does not establish either of them. Only the projection of an utterance against the current metrical norm can do that. For Tomaševskij meter is a relative category; therefore, “there is no hard boundary between prose and verse” (Tomaševskij 1923b, p. 9).

Tomaševskij’s claim was almost immediately challenged by Jurij Tynjanov, who devoted an entire monograph entitled The Problem of Verse Language to discovering a factor capable of differentiating verse from prose. However, Tynjanov’s argument with Tomaševskij did not involve a radically different view of verse language. As I shall show, the two were quite close on many essential issues, but the logic of Tynjanov’s systemic metaphor and his insights into the semantic dimension of verse led him to different conclusions.

As I argued in the preceding chapter, the key concept of Tynjanov’s poetics was the literary system. Understood as a hierarchical set of variables, it consisted of a series of correlated subsystems (for example, genres), which in turn consisted of individual work-systems. Tynjanov related the interdependent variables through the concept of “function.” Thus, every work exhibits a particular function—a correlation of the dominant constructive factor with the subordinate material. This function, dubbed by Tynjanov the “principle of construction,” goes beyond the level of the single work. It unites individual works into literary subsystems—interdependent variables in the overall literary system. Such a system is not simply a logical construct; it has a historical correlate—the series of actual literary forms evolving in time.30 These forms are not just accidents of history that cannot be systematically studied; they are embodiments of specific functions and their continuity or change is indicative of relations among the variables within the literary system.

From this perspective, the Formalist Tynjanov held that verse language should not he treated as a form alone, but also as a function. The fact that poetry, unlike prose, has long been written in even lines betrays a fundamental functional difference between them. For Tynjanov, verse and prose were the two most general literary subsystems constituted through the inversion of their respective principles of construction. “In verse the pivotal constructive factor is rhythm and the material (in a broad sense) is the semantic grouping; in prose the constructive factor is the semantic grouping (the plot) and the material is the rhythmical (in the broad sense) elements of the word” (Tynjanov 1929, p. 15). The opposition between prose and poem is thus not absolute but a function of the literary system as a whole. As the system evolves, the “time may come when it will be inessential whether a work is written in verse or prose, but as long as the distinction between prose and poetry remainspalpable, their two contrastive principles of construction coexist within the literarysystem (ibid., p.29).

Because by definition the principle of construction is always a correlation of two elements—in the case of verse, rhythm and meaning—one cannot adequately describe verse by describing only its dominant component, rhythm. On this point Tynjanov departs significantly from Tomaševskij, who confined his poetic study to metrics, a “discipline… studying the principles that underlie the ordering of actual rhythm” (Tomaševskij 1923b, p. 253). Tynjanov believed the theory of verse language must also include verse semantics, which is a “discipline concerned with the meanings of words and verbal groups, and their evolution and shift in poetry” (Tynjanov 1977, p. 253).The deformation of meaning in verse distinguishes it from prose as significantly as the dominance of rhythm. According to Tynjanov, “prose and poetry are enclosed semantic categories; prosaic meaning is always distinct from poetic meaning, and consequently poetic syntax and even its vocabulary are also essentially different from those of prose” (ibid., p. 35). By systematicallyexamining the meaning of the lexical units that make up verse, Tynjanov went beyond the other OPOJAZ members who (beginning with Brik) claimed they were including semantics within their schemes. In fact, these Formalists reduced semantics to syntax, the rules for combining words into more complex meaningful wholes, and neglected the actual lexical content of the words involved.

Tynjanov’s treatment of rhythm, however, did not differ much from that of the other Formalists. In conceiving of it primarily as a “motor-energic” phenomenon, he was quite close to Brik. Instead of speaking of verse isochronism (whether objective or subjective), he treated rhythmical segmentation as a quantity of labor or energy expended (Tynjanov 1924, pp. 129-33). As we have seen, the idea of verbal art as energy-extensive language was the basis for Šklovskij’s conception of artistic de-familiarization, but in his purposive explanation of art the significance of rhythm lay in its effect upon the perceiver. Rhythmical irregularities were supposed to frustrate the reader’s expectations, thus requiring more effort on his or her part. In Tynjanov’s systemic metaphor, on the other hand, rhythm participates in the constructive function—a hierarchical correlation with other elements of the work. Here the labor involved in the rhythmical organization of verse seems to be the energy source for the ongoing strugglefor domination of its elements.

Tynjanov’s conception of rhythm was perfectly in keeping with his overall anti-substantialist position. As energy, rhythm cannot be identified with any of the phonic elements constituting verse. Rather, it is a system—a dynamic interplay of many factors: “‘Rhythm’ [is] the entire dynamics of the poem comprising the interactions among meter (accentual scheme), linguistic relations (syntax), and sound relations (repetitions)” (Tynjanov 1977, p. 341). Among these, Tynjanov claimed, meter plays the dominant role. Although this apparently echoes Tomaševskij’s belief in the paramount significance of meter for verse, a closer scrutiny reveals a difference. In Tynjanov’s view, what dominates rhythm is not meter as a system of regularly alternating prosodic features, but rather the “principle of meter,” in other words, the “dynamic grouping of verbal material according to a prosodic feature. Most elementary and basic to this is the singling out of some metrical group as a unit. This act also prepares dynamically for the isolation of a subsequent, similar group. If this metrical preparation is realized we get a metrical system” (Tynjanov 1924, p. 30). Even if this preparation is not realized in the subsequent group, even if the metrical system is absent (as in free verse), we are still dealing with verse language. “‘Unrealized preparation’ is also a dynamizing instance. Meter is preserved in the form of a metrical impulse. Every ‘nonrealization’ involves a metrical regrouping: either as a coordination of the two units (carried out progressively) or as a subordination (carried out regressively).… Here the meter as a system is replaced by meter as a dynamic principle, namely, the set toward meter, the equivalent of meter” (ibid.).

As the term “metrical impulse” indicates, Tynjanov’s “meter” covered what Tomaševskij perceived to be two separate categories. In the sense of “metrical system,” it coincided roughly with Tomaševskij’s notion of meter, but as the “equivalent of meter,” it overlapped with Tomaševskij’s “rhythmical impulse.” For Tomaševskij the rhythmical impulse alone could not constitute verse; for Tynjanov the principle of meter would.31 This variance reflects the difference in Tomaševskij’s and Tynjanov’s orientations. Tomaševskij proceeded from concrete verse forms, concentrating on their heterogeneity, whereas Tynjanov proceeded from the general category of the literary system. Striving to discover the identity of verse as a function within this overall system, Tynjanov concentrated on what poems have in common.

Naturally then, Tynjanov rejected features that were characteristic of verse at one point but later disappeared. Meter, in the sense of a prosodic system, was such a case. “In a certain literary system the function of verse was fulfilled by the formal element of meter. But prose diversified and evolved, and so did verse. The diversification of one type of [sound-meaning] correlation involves, or better, is linked to the diversification of another type of correlation. The rise of metrical prose (with Andrej Belyj) was connected to the transference of the verse function from meter to other features of verse that were often secondary or concomitant, such as the rhythm-demarcating verse units, particular syntactic forms, or vocabulary. The function of prose or verse remains, butthe formal elements fulfilling it are different” (Tynjanov1929, p. 59).

Thus, in a seeming paradox, Tynjanov reversed the hierarchy between central and peripheral features as markers of verse. Because central features are always the prime victims of historical change, the identity of a verse system lies in its peripheral features, in those elements that despite changes in the center continue to distinguish it from prose. “The principle of construction is revealed not in the maximum conditions comprising it, but in the minimal ones. For it is obvious that these minimal conditions are the ones intrinsic to the given construction and in them we should seek the key to the specific character of the construction” (Tynjanov 1924, p. 17). Free verse, then, belongs to the verse system despite the fact that it does not correspond to any metrical system. By segmenting a continuous utterance into rhythmical periods it transforms the verbal material according to the same principle as metrically regular verse.

There is, however, one important difference between free verse and more traditional verse forms. In metrically regular verse, recurrent rhythmical units tend to be smaller than those of free verse. They are the syllable, foot, and hemistich, whereas in verse organized solely by the metrical principle, the basic unit is the entire line. In the absence of any prosodic system, the only marker of such a unit is its graphic form. In free verse “graphics plays a special role, for it stands not only for the rhythm but for the metrical unit as well. Here graphics is the signal of a line, of rhythm, and by the same token of metrical dynamics—the indispensable condition of rhythm” (ibid., p. 31).For this reason, Tynjanov, unlike Tomaševskij, ascribed major importance to the graphic form of verse. Graphic form provides the minimal conditions for the rise of rhythm as the dominant factor of verse construction.

Tynjanov believed that not only rhythm, the constructive factor of verse, was reducible to its graphic form, but the subordinate material—that is, the semantic groups within it, was as well. In Puškin’s poetry, for example, a series of dots sometimes replaces a line or a group of lines, as in the original version of the thirteenth stanza of “To the Sea”:

The world has emptied . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .

Here, three and one-half lines of dots serve as the graphic equivalent of the same expanse of words. This substitution is purely graphic: no oral rendition is possible. The voice has at its disposal only a pause—a silence indicating the absence of words. The graphic equivalent signals the presence of this absence, and in doing so carries the metrical energy of the verse. “Obviously, the successive segmentation and reunification of metrical elements… does not occur [here]. The meter is given only as a sign, a potential that is hard to detect. To us, however, the fragments and the dots are equal to the entire stanza and we perceive the lines of the following stanza... precisely as the following stanza. That is, a stanza has elapsed between the fragment commencing the stanza discussed and the next stanza, and the fragment carries the metrical energy of the whole stanza” (ibid., p. 24). As long as the semantically empty dots serve the constructive principle and fulfill the function of actual words, they are a minimal equivalent of the material in the verse construction.

Earlier I suggested that Tynjanov conceived of verse rhythm as a system composed not only of the dominant meter but of other rhythmical factors. The most important of these are sound repetition and rhyme, which Tomaševskij included under the rubric of harmonic rhythm, as we have seen. For Tomaševskij, sound repetition and rhyme operate on the principle of expectation and fulfillment, thus performingthe twofold task of rhythmical dissimilation and assimilation. Tynjanov considered them only secondary rhythmical factors because the proportion of progressive and regressive forces they command differs from that of meter. In meter, the progressive force is most important. It in itself is capable of generating rhythm, as in free verse, where the regressive realization of the initial expectation is forever frustrated. The perception of sound repetition is just the opposite. It lacks all progressive force or as Tynjanov cautiously added in a footnote, it “is extremely weak” (ibid., p. 128). We usually do not expect a sound to be repeated. In rhyme, on the other hand, both forces—regressive and progressive—operate. Nevertheless, Tynjanov argues that here regression is the primary factor.

This claim may require some clarification. One could object that in a regularly rhymed and strophically organized poem the progressive force is paramount: the reader has every expectation of the recurrence of a rhyming ending. For Tynjanov, however, this situation merely shows rhyme under maximal conditions. In texts with looser rhyme and strophic schemes, the reader’s expectation that some subsequent lines will conclude with a group of sounds similar to those he or she is presently perceiving drops considerably. Tynjanov illustrates this claim with a poem of Tjutčev’s in which a rhyme separated by five verse lines passes by virtually unnoticed (ibid., p. 34). What accounts for the weak effect of this rhyme is the lack of expectation on the reader’s part, for he or she realizes it only regressively, and then only if he or she has retained the first rhyming ending over an interval of five lines. Rhyme, moreover, is secondary to meter because it depends on prior metrical segmentation: the rhyming sounds occupythe same positions within lines that have already been metrically delimited.

All utterances organized according to the constructive principle of verse just outlined exhibit, according to Tynjanov, four essential features (ibid., p. 47). The first he calls the unity of the verse sequence [rjad], which is created by metrically isolating a particular segment from the continuous speech chain. Through this segmentation the second property of verse language arises, namely, the density of the verse sequence. The isolation of a metrical segment from its linguistic context brings its constitutive elements closer together: new connections among them, non-existent before this segmentation, are established. This explains “why the quantitative content of a verse sequence must be limited. A unit that is quantitativelyexcessive eitherloses its boundaries or itself becomes segmented into other units. In both cases, however, it ceases to be a unit” (ibid., p. 39). The unity and density of the verse sequence generate the third feature of verse construction—the dynamization of the verbal material. The segmentation of an utterance into recurring rhythmical units makes the semantic units similar to each other not only because of their meanings but also because of their phonic and grammatical features, position in the line, and so forth. In the progressive-regressive buildup of the line, words and their groupings cease to be mere carriers of infinitely repeatable meanings and turn into heterogeneous entities whose multiple facets are constantly foregrounded in the ongoing process of rhythmical permutation.

The most difficult to grasp of Tynjanov’s four features of verse construction is the successivity of its verbal material. In the first place, he opposes it to the simultaneity of the verbal material of prose. Language is a temporal medium, so the verbal material of any speech construction must be successive. In Tynjanov’s usage, however, the words “successivity” and “simultaneity” refer not to the medium itself but to the mode of its perception. In prose, the dominant set toward semantics prevents us from perceiving the utterance as a process. The successivity of its elements is there merely to help us grasp the meaning of the utterance in its totality. This perception of wholeness occurs only after the utterance is finished and we retain all of its elements in our consciousness as a simultaneous whole. In verse, with its dynamized verbal material, the goal sought is not a simultaneous meaning but the sequence itself, the rhythmical unfolding of the verbal material. Such speech is perceived as a process—a continuous correlation of different facets of language whose heterogeneity resists any final semantic summation.

But amazingly, at the same time Tynjanov claims that in prose “time is perceptible,” whereas in verse “time is not perceptible at all” (ibid., p. 119). Here we are confronted by apparent oxymorons: the “temporal simultaneity” of prose and the “atemporal successivity” of verse. This contradictory notion arises from the fact that Tynjanov was really talking about two different temporal strata: the temporality involved in the perception of the artistic medium and the temporality of the extralinguistic semantic groupings that are represented in it. This extralinguistic temporal stratum is especially important in prose, where such groupings are the dominant constructive factor. Through a series of gradual semantic buildups, the reader constitutes characters and events whose causal temporal relations (the story) present one temporal flux. In addition to the indirect experience of temporal flow presented in the story (fabula), the reader experiences directly the flux of the plot (sjužet). That the reader is simultaneously aware of both of them is apparent in Gogol’s short story “The Nose,” in which the “decelerated narrative about the barber Ivan Jakovlevič eating bread and onions produces a comical effect because too much of the (literary) time is devoted to it” (ibid.). In verse language dominated by rhythm, semantics (in the broad sense) is merely a subordinate material. The constitutive elements of verse construction are organized primarily through their rhythmical permutations, and the experience of timein the story—plot interaction is largely missing. Moreover, as these permutations are an ongoing process, there are no breaks in its perception dividing the temporal continuum into “now” and “then” points. Every moment is simultaneously a function of its future (progressive preparation) and its past (the regressive realization of a previous preparation). Tynjanov’s claim about the imperceptibility of time in poetry refers therefore to the fact that the unfolding of an entire verse construction takes place in a single perceptual “now” suspended from the temporal flow.

The discussion of temporal perception in prose and verse occurs in the second half of Tynjanov’s monograph, which is concerned with the effects of verse construction on lexical meaning. The fact that he originally planned to call his book The Problem of Verse Semantics indicates how crucial he considered this part to be. The nearly six decades that have passed since its publication have rendered Tynjanov’smany revolutionary insights about verse semantics commonplaces in modern literary scholarship, but within the context of Russian Formalism their value is unquestionable. And though Tynjanov’s metrics often depended upon discoveries made by other members of the movement, his study of verse semantics is without any doubt an original contribution to Formalist poetics.32

Tynjanov’s analysis of verse meaning was firmly rooted in his systemic metaphor, according to which every phenomenon is relational. For semantics this meant that “it is not necessary to proceed from the word as the single indivisible element of verbal art, to regard it as the ‘bricks with which an edifice is built.’ This element is analyzable into much finer ‘verbal elements’” (Tynjanov 1924, p. 35). Hence, as with rhythm, verbal meaning is a system of hierarchically correlated factors—semantic features.

The first distinction Tynjanov drew was that between the “basic feature” and the “secondary features” of semantics. A basic feature is a general lexical category common to all the usages of a word and hence guaranteeing its semantic identity. This identity is purely semantic, for though homophones share the same outer form, they do not share their basic semantic feature. Drawing a parallel with phonology, Tynjanov saw the “concept of the basic feature in semantics as analogous to that of the phoneme” (ibid., p. 52; p. 134).

The secondary features of meaning can be divided into the “vacillating” and the “steady.” The former are a function of the immediate linguistic context in which the word appears. Every speech construction semantically colors the words which compose it by furnishing them with (slightly) different connotations. Steady secondaryfeatures are a function of a broader socialcontext: the milieu from which the word comes (slangs, dialects, and so forth). Tynjanov calls it the “lexical coloring of the word” and claims that it is “discernible only outside the activity and situation which it characterizes.” Finally, in synthetic languages like Russian, words are usually composed of two parts: the “referential” (veščestvennyj) part that carries the semantic charge of the word, and the “formal” part—the vehicle of its grammatical meaning (ibid., pp. 56-58).

The domination of rhythm in verse tends to realign the hierarchy of semantic features in its words according to their verse function. The unity and density of the verse sequence is perhaps the most obvious cause of such a semantic shift. In a verse construction the rhythmical and semantic divisions need not coincide and syntactically related words may be separated by metrical boundaries. Enjambmentis acase in point. A word separated from its context and incorporated into a metrical sequence gains strong new connotations because of the density of the sequence.

An interaction of rhythm and semantics also occurs within segments smaller than the line, for example, feet and syllables. If a line is composed of words whose boundaries coincide with foot boundaries, every word turns into a rhythmical unit (a foot) and its syntactic relation to other words weakens. Such word-feet tend to be perceived as if in isolation, so that their basic semantic features are intensified (ibid., p. 71). Caesura, an obligatory word boundary after a particular syllable, is another rhythmical division capable of interfering with semantics if, for example, the concomitant intonational pause divides words that are syntactically closely related. Thus, in Lermontov’s line

No ne s toboj / ja serdcem govorju
(But not to you / with my heart I speak)

such a pause (accompanied by a seeming parallelism of the two hemistychs) even leads to a misreading (a “secondary semasiologization” in Tynjanov’s terms), attested to by the fact that two years after the poet’s death this line was printed as:

No ne s toboj, / —ja s serdcem govorju
(But not to you,/ to my heart I speak) (ibid., p.63)

The lexical coloring of words (a steady secondary feature) enjoys a special position in the semantics of Russian verse. It results from the strong influence of liturgical Church Slavonic on literary Russian. Lomonosov’s linguistic reform of the eighteenth century and his theory of three styles identified the high style with the use of Church Slavonic vocabulary. Although in modern Russian this factorhas decreased considerably, there are still many cases in which a poet can playon the synonymity or homonymityof Russian and Church Slavonic words. Lexical coloring can even become a dominant semantic feature when the Church Slavonic word is no longer understandable tothe reader butstill carries the lofty liturgical connotations belonging to that tongue. Vocabulary drawn from other foreign languages, proper names characterizing foreign cultures, or even Russian words connected to a particular region, trade, or milieu fulfill a similar function. All of them foreground secondary features in the words with which they comprise a verse sequence.

In addition to the semantic features that I have discussed so far, the word consists of referential and formal parts. Their relation, or more precisely, the change in this relation caused by rhythm, is equally important for verse semantics. Here secondary rhythmical factors—sound repetition and rhyme—play a central role. Needless to say, for Tynjanov these devices are complex phenomena, and in studying them he takes into account the proximity of repeatedsounds and rhymes, their relationship to meter,the quantity and quality of the sounds utilized, the part of the word in which they occur, and the general character of the word (ibid., pp. 102; 109).

Sound repetitions affect lexical meaning in many ways, for instance, through the mimetic and expressive sound patterns that the early Formalists found especially intriguing. Tynjanov, however, was less interested in this direct link between the phonic and semantic aspects of individual words than in their relationship in words interlocked in a verse sequence. For example, his commentary on the line

Unylaja pora, očej očarovan'e
(Doleful time, the charms of eyes)

provides a good explanation of this phenomenon. “‘Očej očarovan'e’ is a group united both metrically and phonically, and we perceive the sounds očej, oča- as comparable. This perception involves two successive moments: the recognition in the word očarovan'e [charm] of an element from the previous word and the uniting of the two words into a group. In this, the referential part of the word očarovan'e becomes colored through its strong linkage to the referential part of očej [eyes]. It is as if the first stage in the redistribution of the referential and formal parts… had taken place, in this case, as though we derived očarovan'e from the root oči”(ibid., p. 107).

Obviously, sound repetitions need not be limited to contiguous words. Theymay permeate an entire verse construction; by rendering words phonically similar they dynamize their verbal material, and through a regressive movement make this material successive. Summing up the role of sound repetition in verse semantics, Tynjanov wrote that its “evocation of the vacillating features of meaning (through the redistribution of the referential and formal parts of the word) and transformation of the utterance into an amalgamated, correlated whole, cause me to view them as a particular kind of rhythmical metaphor” (ibid., p. 108).

The role of rhyme in verse semantics is to some degree similar to thatof sound repetition. There are, however, certain differences between the two, the stronger progressive force of rhyme being the most important, because of the anticipation raised by the first rhyming member, rhyme is capable of deforming not only the meaning of the rhymed words but also the “direction of the utterance itself.” Put differently, the very play on the fulfillment or frustration of expectations in an actual rhyme can of itself motivate the unfolding of a lyrical “plot” outside of any story. The poem seems to come about only as an exercise in rhyming. Moreover, because of their fixed positions, rhyming words tend to retain their relative independence: theydo not interpenetrate or amalgamate as do words in a sound repetition. “The moment of juxtaposition, comparison,” wrote Tynjanov, “is so important that I view rhyme as a particular kind of rhythmical simile with a partial change in the rhyming member’s basic feature or the foregrounding of its vacillating features. Its significance as a powerful semantic lever is beyond any doubt” (ibid., p. 109; p. 117).

Tynjanov’s The Problem of Verse Language was the most significant criticism of the early OPOJAZ notion of poetic language and the linguistic model that underlies it. Yet, despite such formidable opposition, the linguistic model and its key notion of poetic language did not vanish from Formalist discourse. Quite the contrary: this synecdoche not only survived the movement that spawned it, but after receiving a powerful boost from Prague Structuralism during the thirties and forties, continued into the present day. The re-emergence of this theoretical model after its OPOJAZ critique was the work of the second wing of the Formalist movement whose institutionalized center was the Moscow Linguistic Circle. In particular, the genius of the vice-chairman of this group, Roman Jakobson, invested the linguistic model with a depth and sophistication that it had lacked in the early days of OPOJAZ. We now arrive at the complex topic of Jakobsonian poetics.


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.

— Stéphane Mallarmé, “Crise de vers”

Within the limits of this study of Russian Formalism, Roman Jakobson’s theoretical model poses a special problem. In July of 1920 he left Russia for Czechoslovakia, and with the exception of a handful of articles all his major works were published outside his native land. His stay abroad, which only subsequently turned into permanent exile, did not in the beginning preclude scholarly or personal contact with the Formalists he had left behind. His works were read in Russia and his ideas had an impact on several members of the movement. But the scholarly and political situation in Bohemia was quite different from that of Russia, and as time passed the difference grew. By the late twenties all the other Formalists discussed so far had yielded to official pressure and abandoned either their scholarly careers or their earlier theoretical views, whereas Jakobson’s intellectual history does not contain any such caesura. This is not to say that his ideas stood still. In fact, as his research progressed, his approach to linguistics and poetics evolved into a wholly new scholarly paradigm that in 1929 he christened “Structuralism” (Jakobson 1929, p. 11). This development, unlike that of his former comrades, was not the result of an abrupt leap that negated an earlier position; rather, it was a series of gradual changes—an expansion of intellectual horizons and a shift in theoretical emphasis.

It is this very continuity in Jakobson’s thought that makes my account of it rather difficult. Because of its organic development, it is impossible to pinpoint with any precision the momentat which Jakobson’s Formalist period ended and his Structuralist phase began. And while it is obvious that his linguistic model was an integral part of the Russian movement, it is equally indisputable that a refined version of it informed Structuralist poetics as well. For these reasons, if I am to remain within the strict limits of my topic of Russian Formalism, I cannot treat Jakobson’s approach to verbal art adequately: yet any serious attempt at a full analysis will lead me far astray.

To escape this dilemma, my treatment of the Jakobsonian model will be somewhat more arbitrary than that of the others I have discussed. As a way of stressing the Formalist quality of Jakobson’s notion of poetic language I shall focus on his booklet on Chlebnikov “written in May, 1919, in Moscow as an introduction to Chlebnikov’s Collected Works in preparation” and published some two years later in Prague (Jakobson 1975, p. 17). The Chlebnikov book contains in nuce most of Jakobson’s ideas about verbal art, but as it is not a full-fledged theory of literature but only a preliminarysketch (nabrosok), I shall extract from it the basic principles that came to underlie Jakobson’s “literary science.” At the same time, because many of the notions vaguely hinted at in the Chlebnikov pamphlet are much more fullypresented in Jakobson’s works of the twenties and early thirties, I shall turn to them whenever they clarify the earlier principles of his linguistic model, though I shall make every effort to respect the diachronic development of his thought.

As Elmar Holenstein has argued persuasively, among the intellectual movements that shaped Jakobson’s theoretical outlook, Husserlian phenomenology occupied an especially prominent position. Jakobson’s acquaintance with this subject dated back to his student days at Moscow University in the mid-teens, as shown in the epistemological assumptions behind his earliest project in the new literary science. His conception of literary studies closely parallels the procedures of eidetic phenomenology, which in Holenstein’s account “is concerned with the grasp of the essential features common to objects of the same category” (Holenstein 1976b, p. 4).Accordingly, Jakobson believed that the literary scholar should bracket off the phenomenal heterogeneity of poetic works and focus on the underlying essence that endows them with their categorical identity. As he succinctly put it, “the object of literary science is not literature but literariness, i.e., what makes a given work a literary work” (Jakobson 1921a, p. 11).

Jakobson’s conception of this eidos yielded the first principle of his new poetics. It is the “set [ustanovka] toward expression,” he wrote, “that I designate as the only factor essential for poetry” (ibid., p. 41). Holenstein calls this principle “phenomenological” in that it defines poetry in terms of a perceiver’s mental set, thus following the basic premise of phenomenology that no object can be studied “in itself” but only as it is apperceived by an experiencing or observing subject (Holenstein 1976a, p. 9). As we have seen, however, both Tomaševskij and Tynjanov advocated that the study of verse must begin with the particular mental set with which a perceiving subject approaches rhythmically organized speech. Thus, if I were to follow Holenstein’s suggestion fully I would have to extend the label “phenomenological” to designate their metrical studies as well. However, for me, what is phenomenological in Jakobson’s formulation is not the mental set alone but its qualification as the “set toward expression.” That Jakobson himself considered this qualification crucial is obvious from his suggestion that his method of literary study be called “expressionist” (Jakobson 1921a, p. 10).It would seem vital, then, to approach the phenomenological nature of Jakobson’s poetics through the concept the expression.

The expression (Ausdruck), a notion that Husserl advanced with great rigor in “Investigation 1” of his Logical Investigations, served as the cornerstone of his search for a universalist semiotic theory. For Husserl, only a repeatable sign, a sign that retains its essential self-sameness under all circumstances, can serve as a vehicle of logical thought capable of embodying truth. The psychologistic and physicalistic doctrines of representation prevalent in his day failed to account for the ideal nature of the logical sign. By reducing it to a mere representation of the mental states it indicates or the objectivities it denotes they opened the sign’s identity to the vicissitudesof the phenomenal world. Radically stated, if every significative act posits the sign in a new and unrepeatable spatiotemporal nexus, each of these acts inevitably turns the sign into a unique, nonidentical event.

To avoid the relativism inherent in all naturalistic semiotics, Husserl divided signs into two incompatible categories: (1) the expression, identified as “each instance or part of speech” and “each sign… essentially of the same sort” that are capable of remaining self-same regardless of the actual context; and (2) the indication (Anzeichen), which is any sign lacking such identity and hence merely representing a fluctuating state of affairs (Husserl 1970, p. 275). This scheme, however, was merely taxonomic and did not in anyway explain whywords (and this is what expressions primarily are) can remain unaffected by the context of the speech event. Thus, Husserl was forced to analyze the internal structure of the expression to discover a factor resistant to contextual change. “In the case of a name [for example], we distinguish between what it ‘shows forth’ (i.e., a mental state) and what it means. And again between what it means (the sense or ‘content’ of its naming presentation) and what it names (the object of that presentation)” (ibid., p. 276). Both the “showing forth” and the “naming” are contingent upon empirical reality and thus cannot retain their sameness in repetition. Only the “content of an expression’s naming presentation,” the “meaning” (Bedeutung) of the linguistic sign, is independent of the phenomenal context. It is therefore this lexical meaning inherent in the word prior to its representing other entities that endows the expression with its identity and distinguishes it from the indication.

This distinction has a direct bearing on Jakobson’s probe into the essence of verbal art. To the three functions of the name— showing forth, naming, and meaning—correspond Jakobson’s three goal-oriented verbal activities, or more precisely, functional dialects—the emotive, the practical, and the poetic. He argued against the claims of F. T. Marinetti, the leader of the Italian Futurists, that their experiments in poetry were in fact perfect vehicles for the modern sensibility. Jakobson agreed that “in both emotive and poetic languages, linguistic representations (both phonetic and semantic) attract attention to themselves; the bond between sound and meaning in them is closer, more intimate.” However, “these facts exhaust what emotive and poetic language have in common” (Jakobson 1921a, p. 10). For Jakobson, emotive language was a clear-cut case of the communicative use of language. By intimating a speaker’s mental state, an emotive utterance refers to a phenomenal entity very much as practical language speaks of an objective state of affairs. But “the poetic word is to a certain degree objectless”; it “lacks what Husserl terms dinglicher Bezug” (ibid., p. 47). Poetic language stands apart from the other two functional dialects because the “communicative function inherent in practical and emotive language is minimal in it.” Thus, “poetry, which is nothing but an utterance set toward the expression, is governed by its own immanent laws” (ibid., p. 10).

The suspension of representation in verbal art profoundly affects the way the poetic utterance operates. Whereas in its communicative function the word is a mere transparent vehicle for the signification of other, non-linguistic entities, in poetry the word itself, its internal structure, occupies center stage. Grigorij Vinokur—another influential member of the Moscow Linguistic Circle—drew attention to this fact. “A poetic creation,” he claimed, “is work with a word that is no longer a mere sign but a thing endowed with its own structure, whose elements are reevaluated and regrouped in every new poetic utterance.… if the communicative function makes social intercourse possible through the word, the poetic function informs the perceiver about the very structure of the word, shows him the elements that compose its structure, enriches his mind with knowledge of a new object—the word. The poetic function tells us through the word what the word is, whereas through the other functions of the word we learn about objects ontologically different from the word: other functions tell us through the word about something else” (Vinokur 1923, pp. 109-10).

Jakobson’s conversion of the Husserlianexpression from a logical to an aesthetic category was unorthodox, to say the least, and generated certain problems that had to be solved as his expressionist model developed.33 Within its immediate historical context the rationale for his move was quite clear. By rendering the expression the key notion in his poetics, Jakobson staked out the territory of this discipline beyond the two opposing camps of contemporary Russian literary study. The expressionist model rejected the transrational theory of poetic language propounded by early OPOJAZ, but avoided slipping into the pre-Formalist notion of the literary work as an undistorted mirror of either the poet’s soul or the social reality it depicted. With the expressionist model Jakobson could deny that the artwork was a mere psychological or sociological document without implying that it was therefore devoid of meaning. If poetry, as another critic of the transrational model, Jurij Tynjanov, wrote, “does not operate… with the word but with the expression,” meaning is still always a component of its structure (Tynjanov 1929, p. 509).

Earlier I suggested that the theoretical gulf between the Petersburg and Moscow Formalists on the issue of poetic language corresponds in some degree to the two notions of zaum’ among the Futurists. For Kručënych, who inspired the founding members of OPOJAZ, transrational language was an attempt at liberating linguistic sound from the yoke of rationality; for Chlebnikov, the subject of Jakobson’s first book, it was a return to an original language of pure rationality. “It is possible to say,” Chlebnikov argued, “that everyday language is the shadow of the great laws of the pure word fallen on an uneven surface” (Chlebnikov 1933, p. 230). And the proper domicile of this pure word, as he observed elsewhere, is the human mind: “besides the language of words there is a mute language of concepts composed of mental units (a tissue of concepts governing the language of words)” (ibid., p. 187). Or, in an anthropomorphic metaphor, “the word is a face with a hat tilted over it. The rational [myslimoe] in it precedes the verbal, the aural” (ibid., p. 191). From Chlebnikov’s standpoint, therefore, verbal art as the art of the word is forever caught in the conceptual web generated by rationality, is always permeated with cognitive meanings.

Such a view of verbal art, however, has in recent years become somewhat unpopular. It exhibits what the French philosopher, Jacques Derrida, calls a “logocentric” bias, for it conceives of the linguistic sign as an instrument of reason. This bias, Derrida argues, has its roots in the “Western metaphysics of presence,” which forged the image of the sign as an instance of logos, the signification of Truth. Whatever Chlebnikov’s reasons might have been for elevating rational meaning in the verbal parcel, Jakobson’s seem somewhat less metaphysical than Derrida would suggest. They stem from another conviction of his—unexpected perhaps, given his phenomenological orientation— that literature is a social institution, a consensus among the members of a particular collectivity. Jakobson believed that poetic works are inter-subjective signs involving some form of rationality which he conceptualized as the (imperfect) sharing of cognitive meanings. The OPOJAZ theorists who emphasized the transrational components of poetic language (the emotive and so forth) had in Jakobson’s opinion lost sight of the social nature of verbal art. His 1922 comparison of the Moscow and Petersburg branches of the Formalist movement makes this point unequivocally: “Whereas the former [the Moscow branch] argues that the historical development of artistic forms has a sociological basis, the latter [the Petersburg branch]insists upon the full autonomy of these forms” (Bogatyrëv & Jakobson 1922, p. 458). Thus, accepting Derrida’s notion of the “instituted trace” (Derrida 1974, p. 46) as a substitute for the concept of the sign (damaged beyond repair by its millennia-long marriage to the Western metaphysics of presence), one might say that Jakobson’s “logocentrism” stems at least in part from his taking too seriously the fact that the trace is instituted. For what else does the act of instituting a trace achieve but some form of presence, that is, a consensus among those whose vested power or interest enables them to promulgate one trace as opposed to anotherand those who recognizethe others’ efforts as an accomplished fact? Anyone like Jakobson, who had experienced revolution and civil war, would be well aware of the brutal force by which such a consensus is brought about.34 Physically enforced presence hardly qualifies as metaphysical.

Rejecting the social determinism of pre-Formalistliterary theory, but maintaining nevertheless that literature is essentially social, Jakobson formulated a rather unusual view of “literary sociology.” Here the second, “linguistic,’’ principle of the expressionistic model becomes relevant. This principle projects the social dimension of literature into its linguistic material. If verbal art, in contrast to communicative discourse, directs our attention to the internal structure of language, poetic forms are above all linguistic forms. Thus, in Jakobson’s words, “poetry is language in its aesthetic function” (Jakobson 1921a, p. 11). Because language is for him the social institution par excellence—a set of rules obligatory for the members of a particular speechcommunity—verbal art cannot be asocial. “The theory of poetic language.” Jakobson declared, “can be developed only if poetry is treated as a social fact, if a poetic dialectology of its own kind is established” (ibid., p. 5).

To appreciate Jakobson’s linguistic model fully it is necessary to introduce his overall concept of language. The great Swiss linguist, Ferdinand de Saussure, exerted the most decisive influence on the young Jakobson. As Jakobson recollected in 1956, he gained his first insights into Saussurean linguistics through Saussure’s student Sergej Karcevskij, “who in 1917-1919, during his short-lived return to Russia, fired the young generation of Moscow linguists with the Cours de linguistique générale” (Jakobson 1956, p. 10).

Jakobson would have found Saussure’s theory of language especiallystimulating because the main problem it tackled—the identity of the linguistic sign—was also the central theme of Husserl’s “Investigation I.” The solution the Swiss linguist had to offer was, however, quite different. We have seen that Husserl found the meaning of the expression to be the vehicleof its sameness. But this step only raised the further question, “what is the nature of that meaning?” To answer it Husserl was forced, first, to come up with a situation in which the word would function as a pure meaning free of any indicative relations, and then to account for the self-sameness of meaning in repetition, its identity in every actual situation belonging under this heading. He fixed on the mental soliloquy to meet the first condition. In an interior monologue the subject knows what he means; his words do not serve him as indicators of his thought. Instead, in the directly experienced unity of the significatory act, the meaning of the expression merges with the subject’s meaning-intention. This, however, does not imply that for Husserl meaning was a totally subjective entity: if it were, meaning would dissipate into a multiplicity of meaning-intending acts and so would lack any essential sameness. In addition to its intuitive presence in one’s consciousness, meaning had to exist intersubjectively for Husserl as a universal object (like numbers or geometrical figures) prior to and independent of its actualization. All subjective meaning-intentions would thus be merely tokens of a type, their identity being the ideal self-sameness of the members of a class.

What connects the Swiss linguist to the German philosopher is Saussure’s mentalist stance. Thestarting point of Saussure’s “semiology” was not the word in its physical existence but its representation in the subject’s consciousness. The two constitutive elements of the linguistic sign (the signifier and the signified) are not the actual sound and referent whose materiality renders them unique, but instead infinitely repeatable mental representations—the “sound image” and the “concept.” Like Husserl, Saussure is not a subjectivist, for such a stance would subvert the issue of semiotic identityfrom the very start. But unlike Husserl, who relegated the intersubjective sameness of the sign to the ideal realm of universal objects, Saussure sought it in the social nature of language.

The basic postulate of Saussurean linguistics is that every phenomenon of language has a strictly dualistic existence. On the one hand, it is a component of an actual utterance by an individual speaker (parole), and on the other it is an element within the potential and socially shared system of language (langue).Concrete utterances are nothing but particularized instances of the preexistent system, its implementations in physical, heterogeneous matter. In its purely material being every utterance inevitably differs, if ever so slightly, from any other one; therefore the sameness of a linguistic sign cannot be a fact of parole.

The situation, however, is radically different with langue. It is a homogeneous system of purely linguistic relations devoid of any physical substance, sheer form articulating sound images and concepts into linguistic units. Thevalue of every such unit is precisely circumscribed by its incorporation into this differential grid. Moreover, though entirely conventional, at any given moment langue is fixed and obligatory for all users of a language. Thus the sameness of the linguistic sign, which cannot be discerned in its manifold material manifestations, is a function of the linguistic system. The sign retains its identity through repetitions only because each of its occurrences is an embodiment of the self-same unit of langue. Needless to say, for Saussure the science of language should concern itself solely with this internal system of language. Eschewing the traditional preoccupation of linguists with cultural or natural phenomena contingent upon language, he declared: “the true and unique object of linguistics is language studied in and for itself” (Saussure 1959, p. 232).

Though he accepted Saussure’s postulate of the social nature of language, Jakobson was quite uneasy about the abstract character Saussure ascribed to it. From his point of view, the trace—to return once more to Derrida’s terminology—is never instituted at random but rather for some particular purpose. In other words, language, he believed, is pre-eminently a means-end structure allowing the user to achieve particular goals.

Earlier I mentioned a similar notion of language advocated by the Petersburger Jakubinskij. The two Formalists differed in an important respect, however. For Jakubinskij, the classification of utterances according to telos was only a heuristic device, possible but definitely not the exclusive possibility. For Jakobson, in contrast, language existed in no other mode than as a means to a particular end, so that the teleological view was the only one possible. Furthermore, in accord with the strict binary structure of his transrational model, Jakubinskij recognized only two functional dialects—practical language, in which sounds are mere means, and poetic language, in which they are ends. This bifurcation of language was unacceptable to Jakobson because it juxtaposed sound and meaning as two incompatible phenomena. His own classification (inspired by Husserl) proceeded from the actual speech situation, allotting an appurtenant function to each of the indispensable components of the situation—the speaker, the referent, and the sign. For Jakobson, sound and meaning coexist in every functional dialect; only their relationship is a variable.

Maintaining his means-endmodel for betteror for worse, Jakobson challenged Saussure’s credo that linguistics was concerned solely with language “in and for itself.” “Language,” Jakobson argued, “according to the correct definition of contemporary French linguists, is a system of conventional values, very much like a pack of cards. But because of this, it would be wrong to analyze it without taking into account the multiplicity or possible tasks without which the system does not exist. Just as we have no rules for a universal card game valid equally for rummy, poker, and card-house building, linguistic rules can be determined only for a system defined by its goal” (Jakobson 1925, p. 1). What is under attack here is not langue per se, but Saussure’s notion of it as a homogeneous system uniformly governing each and every utterance. Instead, Jakobson conceives of language as a set of functional dialects each with its own system of rules structured in the way best suited to its specific goals.

Of course, the division of langue into functional dialects presents some problems of its own, the unity of the national language being perhaps the most important. It would seem reasonable to argue, for instance, that a Russian poem has more in common with utterances belonging to other Russian functional dialects than with a poem, let us say, in English. To account for this unity in variety, Vinokur proposed a modification of Saussure’s rigid dualism of langueand parole.Between the socialsystem of a language and its individual utterances, he posited sets of “stylistic” norms, each governing one particular type of goal oriented verbal behavior. These norms, pertaining only to specific usages of a language, are less general than the norms of langue, but at the same time they are shared by at least some speakers of a language. Like langue, they are social.

Viewed through this conceptual prism, every utterance, poetic or otherwise, is simultaneously governed by two normative systems: a general langue and a particular style. “The word taken as a thing [i.e., the poetic word], insofar as it is a word, remains liable to all the laws that determine the life of a word in general, that rivet every kind of superstructure belonging to the sphere of the utterance to the firm, normative basis of language proper” (Vinokur 1923, p. 109). At the same time, a poetic word, Vinokur argued, is not just an utterance but a poetic utterance that belongs in the specific class of utterances united by the pursuit of the same goal. “Taken in itself, of course, each empirically concrete utterance (poetic ones included) is asocial. But the point is that stylistics in general and poetics in particularstudy these concrete utterances as elements of a specific system that is superimposed upon the system of language proper. An utterance is an individual, creative, volitional act. But several of these acts are no longer merely a sum total of individual acts but a system endowed with a purpose, a significance, that is generally valid within perhaps narrow, yet surely social, limits. This system of poetic utterances is, in fact, the genuine object of poetics” (ibid., p. 111).

Although Jakobson did not at first discuss the unity of functional dialects as fully as Vinokur, his occasional statements on the subject reveal a more critical attitude toward Saussure. He rejected the notion of a homogeneous langue equally implemented in every utterance, instead conceiving of a national language as a “system of systems,” a hierarchically organized structure of functional dialects each with its own langue.Within such a structure, each dialectis only relatively autonomous. Practical language is the most basic or, according to Jakobson’s later terminology, the unmarked dialect. Every member of the speech community is inevitably competent in it, for through it one communicates one’s everyday business. As the most universal functional dialect, practical language creates the background against which the utterances of all other dialects are perceived. As Jakobson argued in the Prague Linguistic Circle’s 1929 “Thèses,” “From a synchronic standpoint, poetic language has the form of a poetic utterance (parole) and hence of an individual creative act evaluated both against the backdrop of the immediate poetic tradition (poetic langue) and against that of the contemporary practical [sdělovací] language” (in Vachek 1970, p. 47).35

A poetic utterance is perceived against the background of practical language because the two are functional dialects and not foreign languages. They share most of their linguistic elements and mechanisms, differing only in their methods of exploiting them. Arguing against Jakubinskij’s definition of poetic language as a particular phonetic feature, Jakobson wrote, “The clustering of liquids is possible in both practical and poetic language, but in the former it is causal whereas in the latter... [it is] goal-oriented; i.e., they are two essentially different phenomena” (Jakobson 1923, p. 17). At this point, the concept of the “device” enters Jakobson’s critical vocabulary. The clustering of liquids and other striking organizations of verbal material in poetry are not, as in other linguistic processes, mere accidents, but means to a specific end. They disrupt the communicative function of the verbal sign and in this way redirect attention from the subjective or objective realities signified to the internal structure of the sign itself. The langue of poetic language, the “immanent laws” governing this dialect, can thus be seem as a system of poetic devices. Hence Jakobson’s oft-quoted slogan that “if the science of literature wishes to become scientific it must recognize the ‘device’ as its sole ‘hero’” (Jakobson 1921a, p. 11).

This statement obviously suggests Šklovskij’s mechanistic metaphor, in which the device, if not the sole hero, was definitely one of the main protagonists. The affinity between Šklovskij and Jakobson here is undeniable; however, there are several important differences between them as well. One of these, concerning the linguistic versus extralinguistic nature of the artistic device, was discussed in the preceding chapter. A second difference is that for Šklovskij, the device functioned to de-familiarize, and hence was crucial to the process of artistic perception. For Jakobson, however, the device was important to the process of artistic signification: a poetic utterance de-familiarizes language because of its peculiar semiotic status, because it does not refer in the manner of communicative utterances. Finally, the two Formalists approached the device with different epistemological economies. Šklovskij clearly multiplied the entities designated as devices, cataloging as many different varieties as possible. Jakobson, in accord with his general phenomenological orientation, was decidedly reductivist. Rather than describing the manifold heterogeneity of poetic devices he strove to isolate a few elementary structuring principles implemented in all of them.

What are these basic principles that govern every poetic utterance? Because the “set toward expression” renders prominent the internal structure of the word, verbal art operates with the constitutive elements of this structure—phonic and prosodic factors, morphemes of all types, semantic features—which play only a subsidiary role in communicative language. From this point of view, poetic praxis is the restructuring of an utterance to bring to the foreground the constitutive elements of language. This goal is achieved through two correlated processes: the uncoupling of the speech chain into its basic linguistic elements and their reassemblage into new patterns determined by some form of equivalence. As Jakobson wrote, “in poetry, the role of mechanical associations is minimized, for the dissociation of verbal elements is the exclusive goal. The dissociated fragments are [then] easily regrouped into new combinations” (ibid., p. 41).This view was subsequently reiterated by Vinokur, for whom the “specificity of the poetic tendency” in language “ultimately boils down to the dissolution of a linguistic structure into its elements, which are then recombined. But here, in contrast to the language system proper, the relations among the parts are reshuffled and displaced and thus the very significance, the valency, the linguistic value of these constitutive parts are laid bare and precisely calculated” (Vinokur 1923, p. 109).

This dual process of analysis and resynthesis operates, according to Jakobson, at all levels of poetic language. Analysis occurs in such devices as the rhythmical splitting of the word, poetic etymologizing, and “accentual dissimilation,” that is, the reaccentuation of a word or the juxtaposition of accentual doublets. Resynthesis is implemented in such devices as “rhyme, assonance and alliteration (or repetition)” and “all forms of parallelism: partial parallelism—the simile; parallelism unfolding in time—the metamorphosis; [and] parallelism reduced to a point—the metaphor” (Jakobson 1921a, pp. 47-48).

The poetic restructuring of an utterance not only affects the individual strata of language, but establishes new relations among them. Most importantly it realigns the link between sound and meaning. Throughout this chapter I have noted the keen interest of the Formalists in the similarity of poetic sound to what it signifies, Jakobson was noexception. He also believed that in poetic language “the link between sound and meaning is closer, more intimate… insofar as thehabitual associations based on contiguity retreat to the background” (ibid., p. 10). Earlier we saw OPOJAZ’s preoccupation with expressive and mimetic sound metaphors and Tynjanov’s study of the semantic amalgamation of similar sounding words within a verse line. Jakobson described yet another similarity between the phonic and semantic aspects of poetic language, which might be characterized as the thematization of sound. It occurs when the phonic structure of several semantically disparate words is repeated in the key word of an utterance. A Russian proverb mentioned by Jakobson is a good illustration;

Sila solomu lomit
(Power breaks the straw)

Here “two members of a construction intersect in the third one” (ibid., p. 51).36 The key word soloma (straw) contains both the consonants of the initial word sila (power) and the root of the final verb lomit(breaks). This sound equivalence creates a semantic rapprochement among the words composing the sequence.

The poetic restructuring of an utterance not only disrupts its communicative function, but affects poetic perception through a third principle of expressionist poetics, which might be called “Futurist” (in accord with Holenstein’s terminology) (Holenstein 1976a, p. 18). The non-referential poetic word transforms our attitude to language; it makes what seemed intimately familiar into something strange and unknown. According to the “Futurist” principle, the distinctivefeature of verbal art as a type of linguistic behavior is that it de-familiarizes language and renders its forms unusual.

Like Šklovskij, Jakobson insisted that poetic “form exists only insofar as we feel it, as we sense the resistance of the material, as we wonder whether we face prose or verse” (Jakobson 1921a, p. 5). Hence de-familiarization is a historical process in which all three dimensions of time interpenetrate. As the “unknown is comprehensible and striking only against the background of the known” (ibid., p. 30), so de-familiarization necessarily involves the past: the old automatized forms that serve as a backdrop to the new perception. At the same time, the novelty of the present poetic forms is merely transitory. “There comes a time,” wrote Jakobson, “when traditional poetic language ossifies, ceases to be palpable and becomes outlived like a ritual or a sacred text whose very lapses are considered holy.… the form masters the material, the material becomes fully dominated by its form, the form turns into a stereotype and dies out” (ibid.). New, unusual forms must at this point be created to rejuvenate poetic language. Yet this future de-familiarization is contrastively related to the forms now becoming automatized, and these present forms, as the cause of the subsequent development, contain the seeds of the future within them.

Moreover, Jakobson held that de-familiarization takes place not among isolated poetic phenomena but among phenomena integrated into structures corresponding to literary schools, groups, movements, or even individuals. Thus, like langue, the system of poetic language is not homogeneous. Rather, it comprises various subsystems interlocked in an ongoing historical struggle. Jakobson describes this process in terms of geographical linguistics. “From this point of view, Puškin is the center of the poetic culture of a particular time with a particular zone of influence. The poetic dialects of one zone gravitating toward the cultural center of another can be subdivided, like the dialects of practical language, into: transitional dialects, dialects with a transitory tendency, and mixed dialects. The first have adopted a group of canons from the center toward which they gravitate; the second have adopted certain poetic tendencies from it; and the third, only individual heterogeneous elements—devices. Finally, one must take into account conservative archaic dialects, whose centers of gravity belong to the past” (ibid., pp. 5-6).

The de-familiarization of language is not fully exhausted by the interaction of old and new poetic forms. I noted previously that in the expressionist model, poetic language is closely related to another functional dialect—practical language. Within this dialect too a historical clash goes on between the conservative tendency of standard literary language to preserve traditional forms and the innovative tendency within living colloquial speech to generate new ones. Russian poets, according to Jakobson, have always exploited the creative potential of colloquial speech for the sake of de-familiarization. “From Simeon Polockij on, through Lomonosov, Deržavin, Puškin, Nekrasov, and Majakovskij, Russian poetry has continuously adopted newer and newer elements of the living language” (ibid., p. 30). Raw, uncultivated colloquialisms replace old poetisms-turned-clichés to render the medium of verbal art vivid once again.

The Futurist principle introduces another facet denied by Saussure into the language system: time. In Saussure’s Course, langue is defined as atemporal and linguistic change as asystemic. The motivation for this decision is obvious: concern over the identity of the sign. Once different stages of langue are included in one system, the precise value of linguistic units is compromised. By functioning simultaneously in different relational grids, their identity becomes ambiguous. Moreover, Saussure maintained that the impulse for change came not from within the homogeneous system of language, but only from without it, through theaccidental destructive intervention of extra-linguistic factors. Therefore he split the science of language into its synchronic and diachronic branches and identified the study of langue solely with the former.

But can we actually purge a linguistic system of its history? The Jakobsonian de-familiarization of poetic language would argue against it. This process inevitably brings together past, present, and future states of the system. Moreover, the resulting mutations are not caused by accidents external to the system but by its immanent need for constant rejuvenation, True, in Jakobson’s opinion, the impulse for change is greater in poetic language than in other functional dialects, but synchrony and diachrony interpenetrate in other linguistic systems as well. Hence, a langue devoid of temporality would be a fiction. In every synchronous linguistic system “there are styles of pronunciation, grammatical variants, phrases, which are interpreted by a collectivity of speaking subjects as belonging to and appropriate to a generation of older people, and others which are considered the prerogative of youth, the latest fashion.” Besides these time-marked variants,Jakobson argues, diachrony mingles with synchrony because of the functional heterogeneity of the linguistic system. “The most characteristic form of the projection of diachrony into synchrony is the attribution of a different function in the two terms of a change: thus, two phonological stages are judged as attributes of two functional dialects, two ‘styles.’ The characteristic form of the projection of synchrony into diachrony, on the other hand, is the generalization of a style; two styles become two [developmental] stages” (Jakobson 1929, p. 15).

The difference between Saussure’s and Jakobson’s notions of the linguistic system might be represented as follows:


Saussure’s diagram contains the following coordinates: “(1) the axis of simultaneity (AB), which stands for the relations of coexisting things, from which the intervention of time is excluded, and (2) the axis of succession (CD), on which only one thing can be considered at a time but upon which are located all the things on the first axis together with their changes” (Saussure 1959, p. 80).The system of language, then, is the geometric point at which the two axes intersect.

For Jakobson, as my diagram suggests, the axis of simultaneity (ab) is impregnated with history, for in language at every moment a number of time-marked variants (archaisms, modernisms) always co-occur. By the same token, the axis of succession (cd) contains more than one element at a time. Language consists of several systems of functional dialects each involving a number of subsystems linked both synchronically and diachronically. Thus, rather than a hic et nunc point, the Jakobsonian linguistic system is a field comprising homogeneous and heterogeneous elements.

What my diagram omits, however, is the profoundly dialectic nature of Jakobson’s linguistics, which makes any separation of the system from its history impossible a priori. According to this view, language is not a harmonious, symmetrical whole but an ongoing struggle between revolutionary tendencies aiming to alter the status quo and their conservative counterparts set on preserving it. At any moment the system is both balanced and imbalanced: it is simultaneously a state and a mutation. The ruptures in previous equilibriums coexist with the equilibriums that mended these ruptures, and all of them point to subsequent changes that will redress this situation in the future. This dialectic conception of language also contradicts Saussure’s claim that the causes of linguistic change are necessarily extrasystemic and hence accidental. For Jakobson, linguistic development is triggered by internal contradictions within language, and as such is subject to the rules of the system. External factors, therefore, are neither accidental nor destructive to langue. They are able to penetrate and affect it only if they satisfy some of its internal demands, that is, only if they correspond to the developmental tendencies of the system itself.

To return to verbal art, de-familiarization there according to Jakobson operates on three planes. “We perceive every fact of contemporary poetic language in necessary relation to three factors: the current poetic tradition, contemporary practical language, and the prior poetic tendency” (Jakobson 1921a p. 4). In the case of Chlebnikov—the poet with whom Jakobson’s booklet was concerned the poetic tendency of the immediate past was Russian Symbolism. Whereas Symbolist poetry strove to emulate music, Chlebnikov considered the word the only proper material of verbal art. His zaum’—speech transcending the utilitarian rationality of practical language—had no counterpart in Symbolist poetry. Equally new was his penchant for what the Formalists termed the ‘‘laying bare of devices,” that is, the pure unfolding of verbal material in poetic constructions lacking any psychological, natural, or metaphysical motivation (ibid., p. 28). And in contrast to the predominantly lyrical mode of Symbolist poetry, Chlebnikov returned to the epic genre. In Jakobson’s assessment. “Chlebnikov gave us a new epos, the first genuinely epic creations after many decades of drought” (Jakobson 1931, p. 8).

Central to Chlebnikov’s rebellion against the Symbolists was his use of the Russian vernacular. “Most of Chlebnikov’s work,” Jakobson observed, “is written in language derived from colloquial speech” (Jakobson 1921a, p. 30). This introduction of colloquialisms into poetry was a deliberate challenge to the Symbolist dogma that the profane language of the mob is incompatible with the sacred language of poets. According to Vjačeslav Ivanov, an outstanding poet-theoretician of this movement, “in all ages in which poetry has flourished as an art, poetic language has been contrasted to the colloquial, common language. Both singers and the people loved its differences and peculiarities—singers, as their prerogative, a liturgical or imperial robe; the crowd, as a national treasure and cult” (Ivanov 1909, p. 355).

Naturally, Chlebnikov was not the only Russian poet reacting against the Symbolist canon in the second decade of this century. There were at least two antipodal tendencies within Russian post-Symbolist poetry: the archaizing of Acmeistssuch as N. S. Gumilëv and O. E. Mandel’štam, who sought inspiration in the poetic tradition of past ages, and the iconoclasm of Futurists such as Chlebnikov, who claimed that they were inventing the art of an epoch yet to come. And even within Futurism there was a distinct struggle between the old and the new, as manifested in the writingsof its three leading figures, Majakovskij, Pasternak, and Chlebnikov. Jakobson described this conflict as follows: “In the evolution of Russian post-Symbolist poetry Majakovskij personifies the Sturm und Drang, Chlebnikov the most clear-cut, characteristic conquests, and Pasternak the link of this new art with Symbolism” (Jakobson 1969, p. 387).

This sketch of the Futurist movement is a good illustration of Jakobson’s dialectic conception of the linguistic (poetic) system as a synchronous state containing conservative tendencies pointing toward the past and revolutionary tendencies pointing to the future. The de-familiarization of language in verbal art is not a simple unilateral progression in which every new work leaves all previous ones automatized. The interaction of old and new is instead an oscillation, a seesaw movement, as the contemporary literary reception of the three Russian Futurists verifies. “Despite the fact that Chlebnikov’s poetic personality crystallized prior to Majakovskij’s and, in turn, Majakovskij’s before Pasternak’s … the reader brought up on Symbolism was willing to accept Pasternak first, then he stumbled over Majakovskij, and only after conquering him was he ready to begin the strenuous siege of Chlebnikov’s fortress” (ibid.)

Chlebnikov’s belated critical recognition, eloquently described by Jakobson, is a function of what might be called the dialogic nature of the literary process: the spatiotemporal gap between the author and reader. Here we reach a crucial contradiction, for to conceive of the poetic utterance as dialogic is utterly inconsistent with Husserl’s or Saussure’s semiotic concept underlying the expressionist model.

Husserl’s contempt for the dialogic form of language was absolute. Once the word is addressed to someone and leaves the safe haven of a single consciousness, its identity is totally compromised, for “all expressions in communicative speech function as indications” (Husserl 1970, p. 277). Saussureanlinguistics is equally monological. It relegates any actual verbal intercourse to the sphere of parole and focuses solely on langue—the set of all linguistic elements at a given moment which are uniformly internalized by the speech community. And because Saussure deemed language prior to thought, the linguistic system is not merely a seamless semiotic web connecting all individual minds but their identical content as well. Thus, even though his Course begins with a discussion of the speech-circuit through a schematized dialogue between Mr. A. and Mr. B, ultimately these gentlemen are nothing but two identical instances of a hypostasized social consciousness, two interchangeable voices in a single monologue, two terminals whose semiotic input and output are one.

Saussure’s postulate that the linguistics of langue is possible only if the distance between the interlocutors is obliterated had repercussions among the Russian Formalists. Those who paid attention to the dialogic form of language turned against the notion of system, whereas those concerned with system ignored the dialogic. Jakubinskij and Tynjanov are the two most obvious representatives of these opposite tendencies. In his 1923 essay “On Dialogic Speech,” Jakubinskij rejected the teleological view that divided language into functional dialects according to their respective goals (though he himself had earlier propounded one variant of this view) because he considered it too abstract for the classification of concrete utterances. This classification, Jakubinskij insisted, must proceed from the linguistics of parole, actual discourse. Accordingly, he drew the criteria for his classification from the two characteristics of every human interaction: the type of contact between the subjects (immediate/mediated) and the directionality of the information flow (alternating/continuous).

To the immediate (face-to-face) form of human interaction correspond immediate forms of verbal interaction. These are characterized by the immediate visual and aural perception of the speaker. To mediated interaction corresponds, for example, the written form of an utterance. Correlated with the alternating forms of interactions involving a relatively quick exchange of actions and reactions between interacting individuals is the dialogic form of linguistic intercourse. And for the continuous form we have the monologic form of utterance (Jakubinskij 1923, pp. 116-17).

Jakubinskij believed that in contrast to the “artificial” monologue, dialogue is the “natural” form of language and that the “dialogic form is, in fact, almost always linked to the immediate form of interaction” (ibid., p. 117).37 As a result he concentrated on the oral dialogue and described various linguistic, paralinguistic, and social features of such exchanges.

Fruitful as it might be for the study of dialogue in general, Jakubinskij’s approach was incapable of dealing with what I have termed the dialogic quality of the literary process. From a purely formal standpoint, the literary work is nothing but a mediated continuous communication, the monologue of an absent author read by a passive audience. Yet the curious delayed reaction to Chlebnikov’s work (as described by Jakobson) suggests that the relationship between the author and reader is much more complex. The literary audience is not merely a sounding board for the poet’s words: its choice of reading matter, the timing of its choice, and so forth are, in fact, the audience’s replies to the author’s poetic message. Clearly, such replies are a function not only of the actual literary discourse but of the socially shared literary system as well, the poetic tradition that conditions the reader’s interaction with the text.

Tynjanov opposed Jakubinskij in that he built his model specifically on the notion of the system. In doing so, he, like Saussure, collapsed the space between author and reader. Because he believed that a work’s identity is determined by its evolutionary position within a literary system, this gap is irrelevant to its identity and is simply another name for the work’s alienation, its inauthenticity. Paying attention to it merely subverts the systemic metaphor and leads to subjectivism and psychologism, to a “naive evaluation,” which instead of viewing the ”‘value’ of a given literary phenomenon... in its ‘developmental significance and character,’” arbitrarily “transfers the value from one era-system to another” (Tynjanov 1929, pp. 31-32).

Despite the fact that Tynjanov conceives of the era-system as a diachronic lamination of several contrastive principles of construction, there seems to be no gap between the author and the reader within it. A “literary fact” is identical for everybody. “Whereas a hard definition of literature is more and more difficult to make,” Tynjanov claimed, “every contemporary can point his finger at what is a literary fact” (ibid., p. 9). But once again, the reaction of the Russian reading public to Chlebnikov’s experiments contradicts this assertion. By refusing to read them, the majority of Chlebnikov’s contemporaries indicated that for them his works belonged among the facts of byt, somewhere between infantile babble and the ravings of a madman; only a miniscule minority considered them literary works. Thus, even within a single era-system,one person’s literary fact is not necessarily another’s. The de-familiarization of poetic language takes place among a multitude of individuals whose reactions will differ considerably.

Jakobson’s expressionist model stands between Jakubinskij’s and Tynjanov’s. It acknowledges the dialogic relationship between the artist and audience but accommodates it within a shared system of artistic conventions. Jakobson explored the difference between the subjects involved in the artistic process in his essay “On Realism in Art,” published the same year as his Chlebnikov pamphlet. The notion of realism, because of its apparent simplicity, offered especially fertile ground for debunking the monologic view of art. According to the simplest definition, realism is an “artistic movement that strives for the closest possible representation of reality, for maximal probability.” But within a dialogic context, “representation” and “probability” acquirea curious duality: “On the one hand we deal with an intention, a goal; that is a work is realistic if the author conceived of it as probable (meaning A); on the other hand, a work is realistic if I, the judging subject, perceive it as probable (meaning B) (Jakobson 1921b, p. 301).

The difference between the author and the perceiver described by Jakobson need not, however, lead to the subjectivism that Tynjanov feared. The degree of realistic probability is not totally idiosyncratic; it is measured against the background of a given artistic tradition, the socially valid norms for representation in art. Thus, authorial realism can be subdivided into “A1 = the tendency to deform a given artistic canon, interpreted as an approximation to reality” and “A2 = the conservative tendencywithin the bounds of a given artistic tradition, interpreted as faithfulness to reality.” The same holds for the perceiver. In the “meaning B1 [he] is a revolutionary vis-à-vis the given artistic conventions, who comprehends their deformation as an approximation of reality.” In “meaning B2 [he] is a conservative who sees the deformation of the artistic conventions as a short-changing of reality” (ibid., p.302).

Jakobson succeeded in accommodating the spatiotemporal gap between the participants in the artistic process within the concept of system because of his dialectic outlook. As I argued earlier, he conceived of the system not as a homogeneous langue but as an ongoing struggle among antithetical tendencies and heterogeneous elements. Moreover, the system was not internalized uniformly and totally by every subject. Rather, each individual appropriated only a particular segment of it. From this perspective, the author is neither identical to nor absolutely distinct from the reader. Despite the fact that the two are separate, insofar as they share a similar attitude to past artistic canons they are closer to each other than two contemporaneous authors who represent opposing artistic tendencies. At the same time, conservatives and revolutionaries are not unrelated either, although their connection is a negative one. They embody the thesis and antithesis of a single artistic state and as such they are inseparably bound to each other within the given system.

One important problem arises with this argument. The interplay of sameness and differenceoccurs within the limits of a system. But what are the limits of the system, or in other words, how far apart can an author and reader be before they cease to share anything (whether positive or negative)? This problem is aggravated by the particular modality of literary discourse—its written form. Once a work is fixed in a permanent substance, it can transcend the moment of its origin and become available to a distant reader, projected against a poetic system that is radically different from the one that generated it. When Jakobson and another Moscow Circle exile in Prague, Pëtr Bogatyrëv, compared high literature and folk poetry, they discovered that the primary difference between the two is their respective utilization of permanent and transient linguistic substances—writing and speech.

Their findings, published in “Folklore as a Special Form of Creativity,” can be summarized as follows. A literary work is usually written, so its existence does not coincide with its acceptance by the reading public. It can be ignored by the author’s contemporaries and become popular decades or even centuries later. This fact explains the considerable freedom of the writer in respect to the poetic canon of his or her time. The writer may not only emulate or reject it, but ignore it totally. “In the domain of political economy,” Bogatyrëv and Jakobson wrote, “so-called production for the market provides a close parallel to the relationship of literature to the consumer.” In folklore, on the other hand, this relationship “is closer to ‘production on demand’” (Bogatyrëv & Jakobson 1929, p. 906). A folkloric work, framed in the transient oral medium, comes into existence only when accepted by the community. In fact, it is nothing but a potential set of norms, a living artistic tradition, which persists in the collective memory of a given group to be actualized in every individual performance. Any innovations introduced by these performances can survive only ifthey correspond to the immanent developmentaltendencies of the normative structure, and fulfill a collectivedemand. Asocial aberrations are rejected and, unrecorded, they vanish without a trace. The performer’s attitudetoward his or her creation reflects this state of affairs. The performer exercises a “preventive censorship” and voluntarily conforms to the collective tradition.

Projected into Saussurean terminology, the difference between oral and written poetic works thus corresponds to the opposition of langue and parole. Bogatyrëv and Jakobson wrote:

The role of the performer of folkloric works may not be identified with that of either the reader, the reciter, or the author of literary works. From the folklore performer’s standpoint, the work is a fact of langue, i.e.. an extra-personal, given fact independent of the performer, even if the fact allows for deformation and the introduction of new poetic and quotidian material. To the author of a work of literature, the work appears as a fact of parole. It is not given a priori, but is subject to an individual realization. There is simply a set of artworks effective at a given moment. The new work of art is to be created and perceived against the background of their formal requisites (in that the new work of art appropriates some forms, transforms others, and rejects still others) (ibid., p. 905).

This is a radical statement; indeed, it is seemingly at odds with Jakobson’s notion of verbal art as a social institution. “As a fact of parole,” the poetic work is above all a unique and individual product definitionally exceeding the linguistic system of a given collectivity. One could argue that I am reading too much into Bogatyrëv and Jakobson’s essay. As its title suggests, it does not pretend to deal with the entire literary process but only with its production. Such a reduction is possible because the principal topic of the piece is folklore, in which creation and reception coincide. Had the two authors dealt with literary reception, the issue of poetic langue would have inevitably emerged.

This objection does not invalidate the point I made earlier, however. If, as Bogatyrëv and Jakobson argue, high literature is unlike folklore in the separation of its production and reception, then written literary texts must eventually outlive the system that spawned them, only to be “misread” by later audiences subscribing to totally different poetic canons. And considering the actual conditions of the literary process, one might wonder how things could be otherwise. This was, of course, in part the point Šklovskij made in his article on Puškin, discussed in the preceding chapter. Jakobson, like most of the other Formalists, rejected the radical relativism of Šklovskij’s Rezeptionsästhetik. In his booklet on Chlebnikov, he assailed the aesthetic egocentrism of old-fashioned critics who “usually impose upon the past current modes of poetic production” for negating the social nature of verbal art (Jakobson 1921a, p. 5).

To check the relativism creeping into his expressionist model, Jakobson had to deal with two problems: the need to bridge the gap between author and reader, that is, to find a system obligatory for the two parties; and the need to neutralize the written substance of literature, whose permanence opens the identity of the literary work to the vicissitudes of history. These problems turned out to be two sides of the same coin, and a single solution proposed by Jakobson took care of both of them.

The path that led Jakobson into this difficulty is worth considering. The starting point of his poetics, we recall, was the concept of the expression—a sign whose self-sameness was absolute. Jakobson departed from Husserl, however, in conceiving of this semiotic identity in terms of a Saussurean “social consciousness”; he then further relativized it by breaking up langueinto historically changing functional dialects. Among these, poetic language, propelled by its need for incessant de-familiarization, exhibited the highest degree of change and thus, ironically, was the least reliable guarantor of long-term semiotic identity.

To mitigate the tension between phenomenological stability and Futurist instability in the aesthetic sign Jakobson’s strategy was to turn to language itself, a move I call the “linguistic principle” of his poetics. For him, the literary work is always perceived against the background of contemporary practical language. Poetic language, to appropriate Vinokur’s “Marxist” lingo, is a mere superstructure built upon this normative basis: the aesthetic efficacy of a literary work is founded upon it. Once the author and audience cease to share the system of practical language, the text can no longer function poetically. Jakobson made this point clear in the introduction to his edition of two Czech poetic compositions of the early fourteenth century: “Literary works are so connected with language, they are linguistic facts to such a degree that the distance between the linguistic structure giving rise to the medieval poem and modern linguistic usage is a serious hindrance to its living perception. It is easier to perceive the aesthetic value of an ancient painting or building than to live out the linguistic consciousness of a writer or a reader from that period. This explains why the resurrection of the medieval poetic work as an aesthetically experienced fact lags behind oursurmounting of the merely archaeological approach to the medieval visual arts” (Jakobson 1927, p. 9).

This argument, of course, has its roots in Jakobson’s “logocentric” notion of language, according to which every linguistic fact is a vehicle for intersubjective, cognitive meanings. A poetic utterance, as an expression, is a prime example of such a sign. The same holds for communicative utterances. They differ from expressions not in being without such meanings, but in subordinating themto the referential functions they carry out. This intrinsic bond between the two functional dialects limits the possible spatiotemporal displacement of the literary work that threatened its identity. Therefore, the writer and the reader cannot be totally isolated from each other as long as they share a language. They might subscribe to different literary canons, but the more conservative system ofpractical language is still common to them. Thus, the “misreading” of a work, its projection against a set of poetic norms totally alien to it, still implies that the work makes sense as an utterance. Such a misreading is qualitatively different from the simple incomprehension that occurs when a work is produced in a language unknown to the reader.

At this point, however, it might appear that Jakobson would like to have it both ways. He claims that “every word of poetic language is in essence phonically and semantically deformed vis-à-vis practical language” and thus he can speak of “language in its aesthetic function” as a specific dialect “governed by its own immanent laws” (Jakobson 1921a, p. 47). Yet at the same time he maintains that in some respects poetic utterances are not totally unlike those fulfilling a communicative function.

Ultimately though, Jakobson comes down on the side of non-uniqueness. If verbal art is the “organized violence of poetic form upon language,” such violence is necessarily circumscribed by certain limits, and these are the limits of language itself (Jakobson 1923, p. 16). A poetic form cannot distort its material to such a degree that it loses its linguistic nature, Thereis a sacrosanct structure underlying all functional dialects and rendering articulated sounds (even the Futurist zaum’) linguistic facts. This structure is the phonological system of a given national language. It was thus phonology that Jakobson chose as the key to the self-sameness of the literary sign.

For Saussure, the study of phonology outlined in his Course was an important aspect of linguistic science. This discipline was charged with the task of examining the verbal signifier outside its historical development. It proceeded from the premise Jacques Derrida has termed “phonocentrism,” namely, the view that the spoken word is the original, authentic form of language (see, for example, Derrida 1974, pp. 27-44). Phonocentrism counters the infinite spatiotemporal dislocation of the sign that relativizes its identity by eliminating the cause of this slippage—written language. In Saussurean linguistics, the absolute self-sameness of the verbal sign is guaranteed by its participation in the synchronic system of langue internalized uniformly by every member of the speech community. Because of its transience, intangibility, and absolute proximity to the speaking subject, voice is much better suited to embodying the signifierthan writing is. Fixed in permanent and tangible matter, the inscription falls outside the purely mental langue, and hence is subject to the vagaries of external forces. The numerous discrepancies between pronunciation and spelling, in Saussure’s opinion, reveal the inability of the graphic substance to represent faithfully the internal system of language. Whether flukes of history, geography, orfalse etymology, they lead to a single conclusion: “writing obscures language; it is not a guise for language but a disguise” (Saussure 1959, p. 30).

It is, however, obvious that in its raw physical heterogeneity the phonè cannot participate in the system of language as Saussure conceives of it. Its amorphous multiplicity must be reduced to a limited inventory of elements which can be incorporated into the relational grid of langue. Here Saussure’s phonocentrism merges with his logocentric view, according to which voice is solely the vehicle of reason and has novalue outside this relationship. “Sound,” as he sees it, “is only the instrument of thought; by itself it has no existence” (ibid., p. 8). The phoneme—the minimal unit of the signifier—is, therefore, defined through its relation to the signified—the rational meaning it expresses. “The important thing in the word is not sound alone but the phonic differences that make it possible to distinguish this word from all others, for differences carry signification” (ibid., p. 118). Thus, the phoneme is nothing but a speech sound that is capable of differentiating morphemes. Saussure illustrates this point with the Russian speech sound t.It can be pronounced in a number of ways: aspirated, palatalized, and so forth. The aspirated variant, though acoustically quite distinct, does not differentiate meaning in Russian and, hence, is not an element of its phonological system. The palatalizedt, on the other hand, as a verbal desinence, signals an infinitive form of the Russian verb, in contrast to the nonpalatalized t which in the same position indicates the third person singular form, and therefore is a phoneme.

This approach to phonology provided Jakobson with a solution to the two possible sources of relativism within his expressionist model. By proclaiming the voice to be the original substance of language, phonology eliminated one cause of the spatiotemporal dislocation of the literary work: its written form. As a mere secondary representation of sound, the written text must always relate to the primary substance—voice—whose basic structure is provided by the phonological system of a given language.

Phonology also takes care of the second cause of semiotic slippage—the distance between the participants in the literary process. Of the multitude of norms making up language, the phonological system is the most obligatory, the one the interlocutors must share if any intercourse at all is to take place. This postulate stems from the Saussurean conception of language as a semiotic system whose significatorymechanism is by definition double-tiered. Full-fledged signs or signifiers that carry meaning require the existence of smaller sound elements which do not signify in themselves but serve to differentiate the signifiersof unlike meanings. These meaning-differentiating elements, or phonemes, thus constitute the most elementary linguistic system, which is indispensable to the semiotic functioning of language. In other words, according to this view there can be no language without a phonemic system. Therefore,poetic violence cannot deform this system in any significant way, or verbal art would lose its linguistic nature and become a “variety of less than perfect vocal music” (Jakobson1921a, p. 48).38

The phonological conception of the linguistic signifier enabled Jakobson to treat poetic sound in a way radically different from the transrational model of OPOJAZ. Even if the phonic stratum of poetic language is deformed on purpose, its relation to cognitive meaning is not eliminated, for verbal art “operates not with sounds but with phonemes, i.e., acoustic representations capable of being associated with semantic representations” (ibid.). Even utterances that “deliberately strive to avoid any relationship with a given practical language” (such as the zaum’of the Russian Futurists) cannot escape the constraints of phonology, “for insofar as [a given practical language] exists and a phonetic tradition is present, transrational language is as distinct from prelingual onomatopoeias as a nude contemporary European is from a naked troglodyte” (ibid., p. 67).

Jakobson’s claim extended not only to poetic production but to reception as well. Once a subject internalizes the phonological system, he or she perceives every linguistic sound in terms of it. Here the expressionist model diverges from the purely acoustic approach to poetry characteristic of Ohrenphilologie and its Formalist followers. Arguing against one of its basic postulates propounded by Saran—that the “theoretician of verse… ought to adopt toward verse the attitude of a foreigner who listens to it without knowing the language”—Jakobson wrote: “Not a single person perceives the sound form of poetry in his native langue, its rhythm in particular, as Saran’s foreigner does. Indeed, even this foreigner is fictitious; even his perception would not be purely acoustic. He would merely approach the foreign utterance from the standpoint of his own phonological system, with his own phonological habits. He would, so to speak, transphonologize this utterance” (Jakobson 1923, p. 21).

Jakobson did not stop at criticizing older conceptions of poetic sound: he advanced his own theories about the phonic organization of poetry. The most ambitious was the project of a “phonological prosody” launched in a comparative study of Czech and Russian verse in 1923. Earlier we encountered the polarization of the OPOJAZ on the issue of poetic language as opposed to verse language. Those on the side of poetic language considered rhythm just one among many devices characterizing poetic language, and hence largely ignored it, whereas the others rejected the notion of poetic language as too vague, and focused on the specific problems of verse rhythm. The expressionist model, however, managed to bring these perspectives together by integrating versification into the overall study of poetic language.

Jakobson’s discussion of verse was indirectly a polemic against Brik’s identification of verse with the device of rhythm. According to Jakobson, the mere presence of rhythm in an utterance does not render it poetic, for rhythm may equally occur in practical language. It is the role rhythm performs in these functional dialects that differs. “The dynamic rhythm of practical language is a process that automatizes exhalation during an utterance. In contrast, poetic rhythm is one of the ways to de-automatize the utterance. It is the prerequisite of the [mental] set toward the time of the utterance, what the German psychologists call the experiencing of time (Zeiterlebnis). The division of an utterance into subjectively equal segments, the rhythmical inertia that makes us expect the repetition of a specific signal at a specific moment, the repetition of this signal that foregrounds that signalized sound vis-à-vis its neighbors, all of this is missing in practical language, where time is not experienced” (ibid., pp. 17-18).

This delimitation of practical and poetic rhythm proceeds from the same principles that Jakobson employed to distinguish poetic language from the other functional dialects. What is involved, first of all, is the phenomenological principle: verse triggers a particular set toward the utterance in the perceiving subject. Its temporal dimension, which in communicative discourse is irrelevant, becomes the center of attention in verse language. According to Jakobson, “poetic time is a typical Erwartungszeit; after a particular period expires we expect a particular signal. This time superimposed upon the utterance subjectively transforms it” (ibid., p. 19). What is in question here is another manifestation of the Futurist principle. By foregrounding a feature which in practical language is merely a means toward a communicative end, verse de-familiarizes the verbal medium and renders prominent the internal structure of the verbal sign. Moreover, this transformationemploys a particularvariant of the two basic devices operating in every poetic utterance. The speech chain is dissolved into rhythm-creating elements only to be reassembled on the basis of their regular repetition.

At this point, it might appear that Jakobson’s phonological prosody does not differ significantly from the other Formalists’ positions on verse. Viktor Šklovskij, for example, arguing against Spencer’s conception of rhythm as an energy-saving mechanism, had already pointed out the difference between prosaic and poetic rhythm—between the regular rhythm of a work song, which by automatizing movements tends to save labor, and the violation of this rhythm in art for the sake of de-familiarized, difficult perception (Šklovskij 1919a, p. 114). One could also draw a parallel between Tynjanov’s and Jakobson’s discussions of the temporality of verse language. Tomaševskij’s redefinition of the “rhythmical impulse” is also quite close to Jakobson’s understanding ofverse perception as the pendulum-likeprocess of expectations and fulfillments aroused in a perceiver’s consciousness by the regular recurrence of rhythm-creating elements.

As I observed earlier, the Formalists never reached an agreement as to what those rhythm-creating elements were. Tynjanov’s graphic approach, according to which the ultimate source of poetic rhythm is the visual property of the verse line, was rejected by other Formalists as too simplistic to have any explanatory value. Tomaševskij wrote that in verse, “graphics is merely a sign, not unlike punctuation, that expresses other linguistic correlations but only sometimes is the sole objective evidence of these correlations (as when it happens that only punctuation makes a sentence understandable). For often the other factors are so powerful that graphics becomes redundant and merely accompanies an utterance that is understandable without it. Thus, Puškin’s classic verses will remain such even if printed as prose” (Tomaševskij1924, p. 267). It is the various prosodic features existing in language, Tomaševskij asserted, whose regular alternations create the rhythmical impulse. But even though he seemed intuitively aware of what these features were, he failed to specify them, and went on to embrace Jakobson’s phonological prosody, which provided a coherent and simple hypothesis about the nature of the rhythm-creating elements in verse (see especiallyTomaševskij 1929b, pp. 39-42).

Jakobsonian metrics evolved from the linguistic principle of poetic language, according to which, as we have seen, verse is an utterance with a particular organization of its sound stratum. This organization, moreover, must be rooted in the phonological system of a particular language. Given the resistance ofthis system to poetic violation, the linguistic principle leads to twoconclusions: first, verse deforms above all the extra-phonemic elements of language, and second, it is the inviolable phonological elements that provide the organizational base for this violence. That is, phonological features are those hitherto elusive rhythm-creating elements.

Earlier I tried to show how the differences among the various Formalist theories of verse were conditioned by their points of departure. The same applies to Jakobson. The other members of the movement dealt primarily with Russian verse: Jakobson’s orientation was comparative. As an exile in Prague, he was in fact Saran’s foreigner forced to experience poetry in an alien language. Indeed, it was this experience that convinced him of the intimate link between verse and language. While the other Formalists considered prosodic features such as stress non-problematic, Jakobson, transplanted into a foreign linguistic milieu, directly witnessed their relativity. This relativity was especially evident because Czech and Russian are so similar. At first glance Puškin’s line appears almost identical to its Czech translation:

Russian: Burja mgloju nebo kroet
Czech: Bouře mlhou nebe kryje
(Jakobson 1923, pp. 46-47)39

And yet an actual reading reveals a tremendous prosodic difference between them. This difference, Jakobson argues, results from the dissimilarity of the Czech and Russian phonological systems. While both languages contain dynamic stress, only in Russian is it a phonological element: for example, múka (torment) differs from muká (flour) just in the position of its stress. Czech stress, on the other hand, always falls on the initial syllable of the word and, therefore, is nonphonemic. But vocalic length differentiates words in Czech, for example, byt (apartment) and být (to be), something it cannot do in Russian, where vocalic length is obligatorily bound to stress.

Given Jakobson’s premise that the rhythm-creating elements must be phonologically based, it might appear that the difference between Czech and Russian prosody lies in the fact that the former is quantitative (tied to vocalic length), whereas the latter is accentual (tied to word stress), but, with the exception of the earlynineteenth century when a few attempts at quantitative metrics appeared, modern Czech verse, like Russian verse, has been based on the regular alternation of stressed and unstressed syllables. Accordingly, Jakobson introduced another phonological element into his theory—word boundary. Just as Russian quantity always coincides with stress, Czech stress (fixed on the initial syllabic of the word) always coincides with word boundary. Therefore, Jakobson concluded, not stress but word boundary is the rhythm-creating element in Czech verse.40

The foregoing discussion yields the following typology of phonic phenomena that play a role in the constitution of verse: “(1) the phonological basis of rhythm, (2)concomitant extra-phonemic elements, and (3) autonomous phonological elements, or more precisely, phonological elements that in a given poetic language are not a factor in the rhythmical inertia” (Jakobson 1923, p. 46). The profound difference between the Czech and Russian systems of versification becomes obvious if we superimpose this grid upon the prosodic features with which they operate: stress, quantity, and word boundary.


Returning to Puškin’s line and its Czech translation, it is obvious now why the two are so different despite their surface similarity. First of all, Jakobson argues, they differ in their distribution of quantity. In the Russian original, following the regular trochaic alternation of stresses, every odd syllable is long, whereas in the Czech version only the first and fourth syllables with the diphthong ou are quantitatively different from the rest. Second, because Russian stress is free, the fact that every word in Puškin’s line is disyllabic is “perceived as an episodic coincidence of the normally autonomous word boundary with the rhythmical inertia.” In Czech, on the other hand, with its fixed stress and with “word boundary the basic rhythm-creating factor… the quoted line in respect to its word boundaries is canonical” (ibid., p. 47).

This example illustrates the plausibility of Jakobson’s hypothesis about the close link between the prosodic and phonological systems. It would be wrong to interpret this link in a totally deterministic fashion, to say that one particular phonological system inevitably gives rise to one particular system of versification. On the contrary, as the history of Czech verse has shown, the early nineteenth century witnessed a struggle between quantitative and accentual (or, more precisely, accentual-syllabic) prosodies, both based on different phonological elements coexisting in Czech. Thus, the actual victory of accentual-syllabic verse cannot be explained in terms of phonology. This was the conclusion Jakobson reached in the final paragraph of his study: “I think that a versification system can never be totally deduced from a given language. If a versification system is theunknown and what is given to us are only the prosodic elements of the language, we can arrive merely at an indeterminate equation, i.e., thepossibility of several values for the X. An explanation for the historicalchoice of this or that solution from among the several possible ones involves factors that are outside the phonetics of the given language, namely,the present poetic tradition, the relationship of the given poetic movement to this tradition, and cultural influences” (ibid., p. 118).

This conclusion was not at variance with the universalistic thrust of the expressionist model. Despite its possible heteromorphism, the essence of verse is still provided by the phonological system of the language underlying it, the ultimate system connecting the participants of the literaryprocess. But,as Stephen Rudy has observed, Jakobson’s conclusion contains the seeds of the full subsequent development of Jakobsonian poetics: “It anticipates his later realization that literature is part of a ‘system of systems’ and its studynecessitates a ‘correlation between the literary series and other historical series’” (Rudy 1978, p. 493). The quotation within Rudy’s passage is taken from the nine-point thesis written in 1928 by Tynjanov, the leading theoretician of the then-defunct OPOJAZ, and Jakobson, the vice-chairman of the newly established Prague Linguistic Circle. These theses are generally recognized as marking the end of the Formalist era and the beginning of a new stage of literary studies that emerged in Prague under the name of Structuralism (Cf., for example, Erlich 1981, p. 135 or Matejka & Pomorska 1978, p. viii). Alas, a description of thetransformation of Jakobson’s linguistic poetics into Structuralism lies beyond the scope of this book.


18 According to the title page and a note inside, Kručënych’s 1913 collection Piglets (Porosjata) was co-authored by an eleven-year old, Zina V.

19 Another 1913 collection of Kručënych’s works, Explodity (Vzorval’), con­tains three poems written in “Japanese,” “Spanish,” and “Hebrew.”

20 It is important to stress that Jakubinskij himself conceived of “verselanguage”simply as a special case of poetic language,” (Jakubinskij 1919b, p. 54). As I shall show later, this seemingly subtle difference de­veloped into an important argument against the entire linguistic model.

21 This speech was published separately in 1914 as The Resurrection of the Word, see Stempel 1972, p. 14.

22 Nevertheless, this connection is implied by the fact that Polivanov’s essay appeared in the OPOJAZ Studies in the Theory of Poetic Language. Viktor Šklovskij wrote, “The observation that in Japanese poetic language there are sounds which do not exist in practical Japanese was most likely the first actual indication that these two languages are divergent” ( Šklovskij 1919, p. 104). Still, it seems far-fetched to claim, as Ladislav Matejka does, that Polivanov wrote about Japanese poetry (Matejka-Pomorska 1978, p. 282).

23 In an appendix to the first volume of the OPOJAZ Sborniki appeared a Russian translation of segments of M. Grammont’s Le vers français and K. Nyrop’s Grammaire historique de la langue française that discussed the expressive quality of linguistic sound stemming in part from its articulatory properties (see OPOJAZ 1916, pp. 51-71).

24 It was from Ščerba’s monograph on Russian vowels, Russkie glasnye v kačest­vennom i količestvennom otnošenii (St. Petersburg, 1912), that the Formalists drew their conclusions about the nature of sound in practical language (see, for example, Jakubinskij 1916, p. 38 or Jakobson 1921a, p. 9).

25 For a list of Formalist articles pertaining to Sievers’s school, see Mayenowa 1970, p. 18.

26 In this passage Tynjanov insists on a subtle but untranslatable difference between two synonymic adjectives stichotvornyj and stichovoj, both rendered in English as “verse.” His preference for stichovoj most likely can be attributed to the fact that Jakubinskij who conceived of “verse language” as a mere sub­category of “poetic language,” used stichotvornyj (see note 22 above). For this reason, it is quite surprising that sitchotvornyj, rejected by Tynjanov, should have appeared in the very title of his book. Tynjanov’s correspondence reveals, how­ever, that this title was chosen by his publisher who was apprehensive of the original title Problema stichovoj semantiki (see Tynjanov 1966 [1924], p. 142).

27 In the conclusion to his “Art as Device,” Šklovskij promised to devote a special book to the problems of rhythm. This plan never materialized, however, perhaps because Šklovskij considered poetic rhythm nothing but a deformation of prosaic rhythm, a deformation that must remain unpredictable and hence unsystematizable in order to carry out its de-familiarizing function (Šklovskij 1919a, p. 114).

28 Although quoted in the early twenties by many Formalists, this article was not published until 1927, when it appeared in four installments in the journal Novyj Lef.

29 For a detailed discussion of this topic, see Červenka 1983, pp. 73-84.

30 For Tynjanov’s discussion of the relation between form and function, see especially Tynjanov 1929, pp. 38-41.

31 Apparently in the mid-twenties, perhaps under Tynjanov’s influence, Toma­ševskij modified his position somewhat. Thus in 1925 he was willing to concede that “Majakovskij’s verse is constrained merely by its rhythmical impulse” (Tomaševskij 1929, p. 59).

32 This, of course, does not mean that Tynjanov’s semantic theory is without any intellectual predecessors. As the footnotes to his book indicate, he adopted some of his most important notions from French and German students of language: M. Bréal, C. Bally, J. Vendryes, H. Paul, A. Rosenstein, and W. Wundt, to name a few.

33 Although Jakobson’s definition of verbal art proved to be quite workable for distinguishing poetic language from its emotive and practical counterparts, because of its origins in logic it tended to obliterate the difference between poetic language and another functional dialect which Jakobson later termed “meta-language.” Viktor Šklovskij, for example, when analyzing authorial meta-discourse in Don Quixote,viewed it as a manifestation of the “set toward ‘expression’ which is so typical in art” (Šklovskij 1925, p. 85). Thus Jakobson and some other Prague Structuralists were eventually forced to come up with a secondary criterion to distinguish the metalinguistic from the poetic set toward expression: see, for example, J. Mukařovský 1940, pp. 114-15: or Jakobson 1960, p. 358.

34 See, for example, the joke that Jakobson quotes in his review of André Mazon’s Lexique de la guerre et de la révolution en Russie about a peasant asking the direction to Ljubljanka (a squarein Moscow where the headquarters of the Soviet secret police arelocated). The answer he got was: “Start to sing the Czarist anthem and you will get there quite quickly” (Jakobson 1920, p. 111).

35 To maintain Jakobson’s earlier nomenclature I have translated “sdělovací” as “practical” instead of the more correct “communicative.” By the late twenties, however, Jakobson expanded his functional dialectology and “practical language” became a subcategory of the more general “communicative language."

36 For a more detailed discussion of this proverb see Jakobson 1965, pp. 32-33.

37 The concept of dialogue gained a rather prominent status in the subsequent development of Russian intellectual life as a rallying point for the scholars connected with Michail Bachtin. But with their negative attitude toward Formalism, the Bachtinians approached dialogue from a different perspective. They saw it primarily as a metalinguistic phenomenon—a chain of utterances commenting upon each other from different points of view. Thus, for the Bachtinians, dialogue was a predominantly ideological phenomenon.

38 A special problem that deserves more attention than I can devote to it here is the historical changeability of phonemic systems. In contrast to Saussure, Jakobson maintained that these systems evolve. At the same time, he regarded this change as purely phenomenal, not affecting their “deep structures”—the universal and absolute inventory of hierar­chically correlated distinctive features that in one way or another is im­plemented in every actual phonological system.

39 In English the line means “The storm covers the sky with haze:” it is from Puškin’s poem “Zimnij večer” (Winter Evening).

40 This is the most controversial point in Jakobson’s theory. For the opposing view, which maintains that stress rather than word boundary constitutes the prosodic basis of Czech verse, see, for example, Mukařovský, 1926, p. 217-20; or Červenka 1981, pp. 260-65.

This document is unfortunately not available for download at the moment.

Not implemented yet !
First Edition
(1984) "A synecdoche", in: Steiner Peter, Russian formalism, Ithaca, Cornell University Press, pp.138-241.